IEEE Cement 2002 PDF

IEEE-IAS/PCA
2002 CEMENT INDUSTRY

TECHNICAL CONFERENCE
CONFERENCE RECORD
May 5-9, 2002

Jacksonville, Florida
ISBN 0-7803-7255-7
IEEE Catalog Number 02CH37282C
Copyright ©2002 IEEE. Personal use of this material is permitted. However,

permission to reprint/republish this material for advertising or promotional
purposes or for creating new collective works for resale or redistribution to
servers or lists, or to reuse any copyrighted component of this work in other
works must be obtained from the IEEE.
FOREWORD
2002 IEEE – IAS/PCA Cement Industry Technical Conference
May 5 – May 9, 2002
Adam’s Mark Hotel, Jacksonville, Florida, USA
Sponsored by:
The Industry Application Society’s (IAS), Cement Industry Committee of the Institute of Electrical and Electronics
Engineer, Inc. (IEEE) and The Portland Cement Association (PCA), Manufacturing Technical Committee.
This year’s conference is the 44th Cement Industry Technical Conference, the 7th jointly sponsored
by the IEEE and PCA and the 4rd time the conference is held in Florida.
With this year’s theme “Cement Technology Shining in Florida” we recognize and are emphasizing
the important and excellent work of the Cement Industry Committee (CIC) of the IEEE/IAS.
The IEEE/IAS CIC Working Groups which are comprised of Drives and Related Products; Power
Generation; Maintenance and Safety; Automation; and General Practices continue to enhance the
conference and the industry by their efforts to present the latest developments and issues affecting
the cement industry.
The IEEE and PCA represent the leading cement technical manufacturing groups in North
America. Through their participation and sponsorship, the conference is recognized as the
premiere technological presentation in the world.
It is the objective of the CIC and PCA to strive for continuous improvement in the quality of the
conference by building on past success and looking to the future. We believe this year’s
conference continues the tradition of excellence that is now the standard. Those who attend will
benefit from all the conference has to offer and that this benefit will be realized by their
organizations.
We thank you for attending the Jacksonville Conference and trust your future plans include the
IEEE – IAS/PCA Cement Industry Conference in Dallas, Texas in 2003. We would also like to
thank all of those who contributed to the successes of this conference, which include but is not
limited to; the working group chairs, the paper presenters, the IEEE/IAS CIC Full Committee, the
Adam’s Mark Hotel and most importantly the local organizing committee. It has been a tremendous
experience and pleasure to host the conference in Jacksonville, Florida. We hope you take home
many wonderful memories.
Earl Bouse, Jr., Chair Cary O. Cohrs, Chair

IEEE – IAS/PCA Cement Industry Committee 2002 Organizing Committee
IEEE-IAS CEMENT INDUSTRY COMMITTEE
Executive Committee
Chair Vice Chair

Earl F. Bouse, Jr. Roderick H. Simmons
Secretary IEEE Relations/Nominations

Craig Gotro Satish H. Sheth
Bylaws and History Committee Members

Ronald F. Palmer Ib Bentzen-Bilkvist (Life Fellow)
Neal W. Biege, Sr.
Conference Liaison Russell J. Meta
Paul J. Riley
Working Group Chairs
Automation Paper Review

Jeffrey P. Kemmerer Anya M. Gill
Drives and Related Products Conference Site Selection

Reinhold A. Errath Robert J. Sagmeister
Environmental, Maintenance and Safety Paper Selection

Christopher Macey Roderick H. Simmons
General Practices East Coast Subcommittee Chair

R. Joseph Pospisil Diana L. Inglis
Power Generation, Distribution Gulf Coast Subcommittee Chair

Mark A. Huber Ed Buehler
Awards and Recognition West Coast Subcommittee Chair

Brian P. Keefe Harley O. Biggs
Conference Chairs
2003 Dallas, Texas 2004 Lancaster, Pennsylvania
D. Randall Jones David Roush
PORTLAND CEMENT ASSOCIATION
MANUFACTURING TECHNICAL COMMITTEE
Chairman
F.A. Nelson
Hanson Aggregates West
SUBCOMMITTEES
Subcommittee 1 Subcommittee 2
Industry Liaison Cement Manufacture
(Vacant), Chairman James H. Melick, Rotating Chairman
John G. Wheeler, Vice Chairman Lafarge North America Inc.
Capitol Aggregates, Ltd.
Subcommittee 3 Subcommittee 4
Communication Education
Mark Terry, Chairman Daniel R. Fritz, Chairman
Polysius Corp. Polysius Corp.
Louis J. DeHayes, Vice Chairman Hans E. Steuch, Vice Chairman

California Portland Cement Company Ash Grove Cement Company
TASK COMMITTEES
Task Committee 1 Task Committee 2
Materials Preparation and Finishing Pyroprocessing
C. Thom Martin, Chairman Joel R. Filius, Chairman
BHA Group, Inc. Baker Refractories
Stuart Pryor, Vice Chairman Mark Morgan, Vice Chairman

North Texas Cement Company, L.P. Holcim (US) Inc.
Task Committee 3 Task Committee 4

Auxiliary Automation
Rob B. O’Conor, Chairman James H. Melick, Chairman
BHA Group, Inc. Lafarge North America Inc.
Max L. Frailey, Vice Chairman Jeffrey P. Kemmerer, Vice Chairman

Capitol Aggregates, Ltd. FLS Automation
IEEE-IAS/PCA
2002 CEMENT INDUSTRY TECHNICAL CONFERENCE
Jacksonville Organizing Committee
Chairperson Plant Tour

Cary Cohrs Jim Kern
Florida Rock Industries, Inc. Florida Rock Industries, Inc.
352-472-4722 352-472-4722
[email protected] [email protected]
Vice-Chairperson Promotion/Publicity
Mark Collett Jim Wiles/Jo Lobsiger
Roberts & Schaefer Company Skamol Inc.
801-364-0900 704-544-1015
Secretary Special Projects

Chris Waller Mark Mueller
Rockwood International Systems Supply, Inc. Polysius Corp.
770-849-0100 ext. 104 770-980-5029
Treasurer Publications
A.J. Pont Terry Brooks
Florida Rock Industries, Inc. Menardi
352-472-4722 800-321-3218
Exhibits Registration
Robin Riester Rick Shields
FLS Automation Florida Rock Industries, Inc.
410-771-0850 352-472-4722
Renee Williams Chris Horner

Roberts & Schaefer Company Florida Rock Industries, Inc.
801-364-0900 352-472-4722
Facilities Scheduling
Lisa Spiegler Brian DeVos
Florida Rock Industries, Inc. Siemens Applied Automation
352-472-4722 918-662-7473
Hospitality Spouse Activities

Bob Baldwin Shelley Cohrs
F.L.Smidth Inc. Florida Rock Industries, Inc.
770-509-2322 352-371-1232
PCA Liaison Advisor

Garth Hawkins Anya Gill
Portland Cement Association Polysius Corp.
847-966-6200 ext. 319 770-980-5071
IEEE-IAS/PCA CEMENT INDUSTRY
TECHNICAL CONFERENCE HISTORY
National Committee
Year Location Chairman Chairman Reg.
st
1 1959 Allentown, PA W.A. Allan A.H. Huelsman 298
2nd 1960 Milwaukee, WI W.A. Allan V.B. Murdock 316
3rd 1961 Detroit, MI J.F. Hower I.E. Swanson 307
4th 1962 St. Louis, MO J.F. Hower J.B Woodward 437
5th 1963 Cleveland, OH R.J. Jager G.E. MacDonald 304
6th 1964 Pasadena, CA R.J. Jager D.B. Carson 433
7th 1965 Allentown, PA H.P. Cassel C.A. Zimmerman 601
8th 1966 Denver, CO H.P. Cassel W.A. Walking 436
9th 1967 Albany, NY I.E. Swanson J.R. Kelly, Jr. 479
10th 1968 St. Louis, MO I.E. Swanson A.C. Lordi 515
11th 1969 Toronto, ON A.C. Lordi J.A. Allan 501
12th 1970 Indianapolis, IN A.C. Lordi G.F. Messinger 453
13th 1971 Seattle, WA J.A. Allan F.J. Bauer 370
14th 1972 Detroit, MI J.A. Allan I.E. Swanson 464
15th 1973 Miami, FL R.P. Kistler I.W. Copple 503
16th 1974 Mexico City, Mexico R.P. Kistler R.J. Plass 676
17th 1975 Montreal, PQ F.J. Bauer M.S. Jackson 583
18th 1976 Tucson, AZ F.J. Bauer J.A. Vidergar 587
19th 1977 Omaha, NE R.C. White F.E. Staples & 570
R.F. Palmer
20th 1978 Roanoke, VA R.C. White K.C. Wiles 677
21st 1979 Tarpon Springs, FL J. Warshawsky F.W. Cohrs 829
22nd 1980 Toronto, ON J. Warshawsky M.E. Wrinkle 839
23rd 1981 Lancaster, PA L.L. Warner N.W. Biege 841
24th 1982 Vancouver, BC L.L. Warner B.T. Price 595
25th 1983 San Antonio, TX N. Roistacher U.K. Alsguth 622
26th 1984 Anaheim, CA N. Roistacher J.A. Vidergar 563
27th 1985 New Orleans, LA R.J. Krekel L.C. Cockrell 595
28th 1986 Salt Lake City, UT R.J. Krekel R.W. Riegel 490
29th 1987 San Francisco, CA R.F. Palmer C.D. Maars 557
30th 1988 Quebec City, PQ R.F. Palmer A. Morneau 560
31st 1989 Denver, CO lb Bentzen-Bilkvist E.D. Parker 546
32nd 1990 Tarpon Springs, FL lb Bentzen-Bilkvist E.A. Buehler 629
33rd 1991 Mexico City, Mexico N.W. Biege J.R. Trevino 430
34th 1992 Dallas, TX N.W. Biege P.H. Caldwell 528
35th 1993 Toronto, ON R.J. Meta J.R. González 536
36th 1994 Seattle, WA R.J. Meta S.E. Sheridan 505
37th 1995 San Juan, PR P.J. Riley R. Di Cristina 546
38th 1996 Los Angeles, CA P.J. Riley J.G. Brown 623
39th 1997 Hershey, PA M.W. Jasberg R.J. Pospisil 632
40th 1998 Rapid City, SD M.W. Jasberg D.L. Rydquist 604
41st 1999 Roanoke, VA J.R. González A. Papadopoulos 682
42nd 2000 Salt Lake City, UT J.R. González J.R. Sommers 783
43rd 2001 Vancouver, BC E.F. Bouse, Jr. A.S. Moore 836
44th 2002 Jacksonville, FL E.F. Bouse, Jr. C.O.Cohrs
2002 CEMENT INDUSTRY TECHNICAL CONFERENCE
Table of Contents
Power Generation, Distribution and Related Products

“Options for the Cement Industry in the Deregulated Power Era”
By: Greg J. Miranda, P.E., GE Industrial Systems
“Recovery of Industrial Heat in the Cement Industry by Means of the ORC Process”
By: Hilel Legmann, Ormat Industries Ltd.
“Cogeneration Systems in Material Drying Applications”

By: Dwayne R. Walker, Cummins & Barnard, Inc.
“Higher Efficiency – What is the Real Cost?”

By: John J. Stroker, Penta Engineering Corporation
“Energy Efficiency Improvement in the Cement Industry”

By: Brahm Segal, Power Correction Systems, Inc.
Drives and Related Products

“Dealing with Motor Winding Problems Caused by Inverter Drives”
By: Mark Fenger, Steven R. Campbell, Iris Power Engineering Inc. and Jan Pedersen,
Techwise A/S
“Spotting Drives for Large Motors”

By: Barry Dick, GE Toshiba Automation System
“Replacement of a Wound Rotor Motor with an Adjustable Speed Drive for a 1400 KW Kiln
Exhaust Gas Fan”
By: Roman Menz, Holcim Group Support Ltd. and Felix Opprecht, Holcim Zement AG
“Selecting the Right Preheater Fan and Drive”

By: Victor J. Turnell, P.E., Penta Engineering Corporation
“Alternative Drive Solutions for the Cement Industry”

By: Ashok B. Amin, Hagglunds Drives Inc.
General Practices
“Commissioning a 2 MT/Y Cement Plant in Midlothian, Texas”
By: Marc A. Wurman, F.L. Smidth Inc. and William Brown, TXI
“Solving Raw Material Challenges”

By: Linda M. Hills, Vagn Johansen and F. MacGregor Miller, Construction Technology
Laboratories, Inc.
“The Kosmosdale Expansion Project”

By: Alan Rowley, Cemex and Detlev Babel, Humboldt Wedag, Inc.
“Environmental Considerations for Preheater Design”

By: Steven W. Miller, F.L. Smidth Inc.
”Finish Grinding with Vertical Roller Mills – Operating Data”
By: Chris Oesch and Blaz Jurko, Loesche America, Inc.
“Cement Mill System Upgrade”

By: Donald A. Longhurst, Polysius Corp.
“Increasing Cement Grinding Capacity with Vertical Roller Mill Technology”

By: Gary R. Roy, F.L. Smidth, Inc.
“Existing Mill Capacity Doubled Developments in Grinding Process”

By: Siegfried Strasser, KHD Humboldt Wedag
Environmental, Maintenance and Safety

“Large Project Effectiveness using Plant Personnel”
By: John Sommers and David Larson, Holcim (US) Inc.
“NOx Formation in Rotary Kilns Producing Cement Clinker – Applicable NOx Control Techniques and
Cost Effectiveness of these Control Techniques”
By: Gerald L. Young, Penta Engineering Corporation
“Staged” Combustion for Low-NOx Calciners

By: Brian P. Keefe and Robert E. Shenk, F.L. Smidth, Inc.
“Low NOx Rotary Kiln Burner Technology: Design Principles & Case Study”
By: Max H. Vaccaro, PILLARD E.G.C.I.
“Staged Combustion for NOx Reduction Using High Pressure Air Injection”
By: Eric R. Hansen, Cadence Environmental Energy Inc.
Automation
“Plant Automation at Signal Mountain Cement”
By: Michele Maranzana, RC Cement Co., Inc.
“Upgrading the Process Control System at GCC Rio Grande’s Tijeras, New Mexico Plant”
By: Oscar Aragonez, Grupo Cementos de Chihuahua and Orlando Aguilar and
Leroy Cockrell, GCC Rio Grande, Inc.
“A Real World Innovative Concept for Plant Information Integration”

By: Urs Herzog, Holcim Group Support Ltd. Switzerland and Rauli Hantikainen, Marnix Buysse,
ABB Switzerland Ltd.
“Total Integrated Device Operation Network (TIDON)”

By: Michael Voigt, F.L. Smidth
“The Use of Mining Software to Optimize the Limestone Mineral Deposit at the Colton Quarry”
By: James Chadwick, California Portland Cement Company and John Morrison, Datamine North America, Inc.
Automation
Working Group
Chair
Jeffrey P. Kemmerer
FLS Automation
Vice Chair
Stephen J. Coppinger
California Portland Cement Company
Members
Michael Bertsch Ronald F. Palmer
Siemens AG CEMEX
Leroy C. Cockrell, Jr. Andrew Robinson

Rio Grande Portland Cement Corp. Ash Grove Cement Company
Craig Gotro Jay Shah

JSHA Engineering
Andrzej Janik Gregory Witte

Asea Brown Boveri Inc. Rockwell Automation
system sets a certain range of the above parameters and in addition performs a fine tuning of the
amount of kiln feed according to the level of certain process parameters like oxygen and kiln
torque.
For every system of this type a lot of fine-tuning is always necessary and Signal Mountain is still
working on the optimization. Up to this point the automatic control system does not run all the
time as yet. The main reason is that the operators training program is still going on and that
requires the kiln to be operated in manual.
CONCLUSION
We are very satisfied with the level of automation reached at the Signal Mountain plant, but also
we realize that further improvements can and should be achieved. Our main objectives for this
year is to complete the implementation of the kiln control system and to implement the
“Production Controller” module of the software control package, a higher level of enterprise
management that can execute different commands according to different equipment conditions.
One of the most important lessons we learned during this project is that the utilization of the best
software package is not guarantee to end up with a good control system. In fact, none of the
above could be accomplished without the dedication and the expertise of everybody involved with
this project
18
UPGRADING THE PROCESS CONTROL SYSTEM AT
GCC RIO GRANDE'S TIJERAS, NEW MEXICO PLANT
By
Oscar Aragonez of Grupo Cementos de Chihuahua and Orlando Aguilar and Leroy
Cockrell of GCC Rio Grande, Inc.
For Presentation at the

IEEE-IAS/PCA 44th Cement Industry Technical Conference
May 2002, Jacksonvile, Florida
1
INTRODUCTION:
The Tijeras, New Mexico plant is located about 6 miles east of Albuquerque just south of
Interstate 40 nestled in the southern rocky mountains. The plant is in Tijeras (Spanish for
Scissors) canyon at an elevation of 6,300 feet or 1920 meters above sea level. It was
originally built by Ideal Cement Company, and the first clinker from the new plant was
produced on May 6, 1959. When it was built, it was a state of the art central control room
plant with one long dry kiln 11 feet in diameter. The control room is located in the heart of
the plant with views of the kilns and mills, and all the controls were hard wired to the control
panels using 440 volt control voltage to make them reliable. Since that time, the plant has
undergone many additions and upgrades. The plant now has two kilns each with a two stage
preheater, two ball type raw mills, and three finish mills giving it a capacity of 500,000 short
tons per year. Grupo Cementos de Chihuahua purchased it from Holnam, now Holcim, in
1994 through its wholly owned US subsidiary Rio Grande Portland Cement Corporation
which is now GCC Rio Grande, Inc.
RECOGNITION OF THE PROBLEM:
In 1995 the company was faced with doing something with the aging control system that had
some isolated Programmable Logic Controller (PLC) control, but most of the controls were
still the original 440-volt hard wire controls. The newest PLC in the plant was 10 years old.
The plant people recognized that it was becoming very difficult and even impossible in some
cases to get spare parts and support for aging process control equipment including field
devices and instrumentation. Furthermore Y2K was a concern with the old computer
equipment. The plant agonized with the problem of modernizing the controls at a reasonable
cost and went through several scenarios before settling on a direction for the future of the
plant.
2
THE BEGINNING OF A PROJECT:
The first thing we did was to ask ourselves what kind of budget we would be willing to
propose to the new company to address the control problems in the plant. The people at the
plant tended to be rather conservative, and for that reason, it was decided to try to keep the
cost as low as possible by using an “Open System” and utilizing internal plant resources as
much as possible to control the costs. The main focus was on keeping costs low rather than
looking at the whole project and deciding a scope and a detailed plan of what we wanted to
achieve together with a proposal to management to get approval of a budget with a defined
project. The result was that the project floundered while we went through the process of
deciding how we would work together to develop a proper project definition and a project
plan that could be approved by management.
As is the case with most projects to upgrade existing facilities, one of the first problems was
deciding where to start the project scope definition and where to stop. In this case, it was
decided that we would have to start with addressing the problem of the 440 volt control
voltages that were impossible to use by directly connecting digital input and output signals to
any modern control system. There were two possible solutions to this problem. One solution
was to replace the transformers in every breaker in the plant to convert to 110-volt control
voltages. The other potential solution was to use interposing relays. It was decided to
convert all of the starters since this would be the cleanest and most professional approach,
and we started on the first phase of the project a short time later since this needed to be done
regardless of the other project definitions. We also knew that this would be a rather long and
tedious process since we also had to keep the plant running and could not arbitrarily take
equipment down to work on the breakers. We did have spares for most breakers, and this
allowed us to replace breakers with short outages until we worked our way through all the
breakers. The process took over a year to complete using in house labor.
In the beginning, the project was defined as a plant project, but as time moved on, we began
to realize that we had resources in the new company that we had not considered in the initial
project definitions. Our new company had an Engineering department that had built a very
successful plant in Samalayuca, Mexico; and, they had resources that had designed and
started a very successful control system as a part of that project. The plant realized that we
were having trouble documenting the electrical drawing changes associated with the project,
and we began to start working with our Engineering group in Mexico to solve this problem.
Initial Scope and Plan:
1. Replace all of the controls and panels in the control room with new computerized
controls.
2. Use the remote Input / Output (I/O) capabilities of the modern control systems to
locate I/O close to the I/O devices and eliminate the need for more room in the
substations.
3. Use Computer Aided Drafting (CAD) software to document the control system
using local people augmented by in house engineering people.
3
4. Provide Uninterruptible Power Systems (UPS) systems for reliability and
minimize field sensor and wiring upgrades to keep costs low.
5. Minimize the use of outside engineering and contractors to keep costs low and
stimulate internal "Ownership" of the system.
As time moved on and we began to nurture the relationship with our new partners, it became
obvious that we were not heading in the right direction. In the beginning, one of the most
important concepts we used was to remain as flexible as possible and keep our options open
so we could change the upper levels of the control system with as little effort and with as
little change in the hardware as possible. By this time, we had studied the problem in
significant detail that we had a pretty good idea of how we needed to define the scope of the
project.
We also recognized that the 6 phases of most typical projects were:

1. Enthusiasm
2. Disillusionment
3. Panic
4. Search for the guilty
5. Punishment of the innocent
6. Praise and honor for the non-participants
We did not like this approach; so, we decided to make a change.
TIME FOR A CHANGE:
At this point we discussed the problem with management and decided to make a change in
the direction of the project. We were fortunate enough to have a group of very talented
internal engineering people who have even be used by other companies to help solve their
control problems. We put together a new scope of the project and a new and larger budget
that allowed the use of outside engineering and construction resources along with our
company engineering resources to do a more complete job of automating the plant. The plan
proposed to management was approved, and we began the process of making the required
changes.
Revised Scope and Plan:
1. Replace all of the controls and panels in the control room with new computerized
controls.
2. Provide new process control substations near the existing mill and cooler
substations to house new I/O and instrumentation and utilize new wiring trays for
new wiring required. Remove obsolete wiring.
3. Upgrade the weigh belt feeders to get adequate signals and controls and add
variable speed cooler fan drives. Add instrumentation and controls to the crusher
to help automate the controls. Provide UPS for system reliability.
4. Develop the software tools to automatically generate the CAD drawings,
simulator code to test the software, and PLC Code for the system using standard
templates.
4
5. Utilize Company Engineering resources and outside engineering and contractors
to keep work loads reasonable while achieving professional results.
6. Revised the budget to increase it to three times the original budget
The new plan included more work on the field devices, a new substation to locate I/O devices
in a clean environment, more emphasis on developing and using software and hardware
templates and naming conventions to simplify the software and hardware engineering tasks.
It also included the development of an internal development tool that automatically generates
electrical drawings, PLC code, and simulator code to be able to test the logic without being
connected to live I/O. The tool uses Project Definitions in two Excel Spreadsheets and
multiple CAD templates as inputs to generate the outputs as shown in the GCC Tools
diagram below (See the detailed figures in Appendix A).
GCC TOOLS
Device Drawing
See Figure 4
Excel Spreadsheets
See Figures 2 & 3
GCC TOOLS
PLC Program Code
See Figure 1
See Figures 5
CAD Template
See Figure 3
Simulation System
See Figures 6
The whole approach resulted in a more complete development of a philosophy that can be
used as an approach to future projects as well as a solution to the problem faced at the Tijeras
plant. This new philosophy required a little up front investment to develop, but it greatly
simplifies engineering of new projects, increases reliability, simplifies start-up, and it
provides a basis for uniform support of the control systems used through the company. The
tool also supports several hardware configurations.
5
THE FIRST STEPS:
The initial steps in the project had included identification of the I/O points, introduction to
the point naming schemes, and an introduction to defining templates and standard
subroutines. This allowed the people to slowly become more familiar with these initial ideas
as planning documents containing I/O definitions and drawings were developed to expand
the scope of the project. The additional rooms for substation equipment were added to the
mill and cooler substations as planning documents were developed. At the same time, the
feeders were overhauled and brought up to speed with modern instrumentation. Control was
moved from the old control systems to the new one at a time during normal scheduled
outages since shipping and production schedules had to be maintained throughout the project.
Finally, all of the existing controls that had been partly automated were all moved into the
new system, and there were only two control systems in the plant. These were the new
control system and the old hard-wired systems.
It was really important for everyone to understand that to achieve the goals of automating the
plant, we had to really address the big picture that included the field wiring and sensors.
That is, we had to upgrade the field sensors, the wiring and all wiring documentation. This
included items such as upgrading the weigh feeder balance beams to new load cells,
replacing pressure sensors, and replacing some magnetic motion sensors with new encoders
for instance. Wiring changes included communications fiber optics from the control room to
the crusher 1/4 mile away and for remote I/O on top of the silos and at the coal receiving area
for example. It also included bringing stabilized power for the new I/O from more than one
source so a substation could be down without affecting the whole control system. Each
subsystem had to function properly and reliably before they could be integrated into the plant
control system. Adding a computer to a system that is a mess only achieves an automated
mess rather than a desirable solution. Adding the money to the project to fix the wiring
problems and correct any inadequate sensor problems should be a requirement of any project
plan to upgrade an existing system.
TEMPLATES AND PHILOSOPHIES:
If we look at a typical plant, it has many motors of many different sizes used for many
different purposes. However, when we looked at the way we control the motors, we could
see that we could use a rather general approach and for the most part define all of the motor
controls in the plant using about three or four different definitions. For example, we defined
a standard motor, a reversible motor, and a synchronous motor as our basic standard three
motor types requiring different control definitions.
To define our templates, we needed to decide some issues regarding operation and
maintenance of these motors and set some standards that would be followed in the operation
of the plant. These standards become part of our templates and standard operational
philosophies, and they define how we operate our plant. These standards are useful not only
in programming the plant control system, but they are helpful in training people in the plant
on operational and safety matters.
6
A few examples of these items from our operations are as follows:
The safety lockout procedure for our plant requires that each person working on a piece of
equipment must lock it out at the breaker with his/her own personal padlock before beginning
to work on it with no exceptions. Furthermore, the person in charge of the work must go to
the control room and enter an equipment "Clip-Out" entry in the clip-out logbook and tell the
control room operator to enter the clip into the computer. Then, he/she is responsible to
check to see that the device and the "Group Start" are "Not Ready" for the proper system to
be sure the correct piece of equipment is clipped-out in the software. The responsible person
must then go to the local start switch and attempt to start the equipment. This is to be sure
the correct piece of equipment is locked out at the breaker. Then and only then can the repair
crew assigned to the job begin work on the equipment after it has been demonstrated that it is
not possible to inadvertently start the equipment locally or remotely.
The individual lockout procedure is rather common through the industry, but the "Clip-Out"
procedure is not. This is an example of an operational philosophy that we decided to include
in our equipment control designs to increase safety awareness and enhance safety procedures.
The procedure was practiced with the old hard-wired controls by trying to start the equipment
from the pushbutton in the control room after locking it out. Then the responsible person
physically put a clip across the button on the control panel in the control room. The clip
included a name of a responsible contact person on it. This way, the operators were informed
of what equipment was down for repairs and the name of a contact to check on it was
supplied to the operating personnel. This also helped to communicate to new operators
coming on duty at the change of shifts so they would not attempt to start equipment that was
locked out. With the new computerized controls, we wanted to use the same procedures, but
there were no longer any push buttons in the control room. So, we devised a software "Clip-
Out" procedure that prevents the software from trying to start equipment that is clipped-out.
It also displays a small skull and crossbones next to the equipment display list as shown
below so the operators know what is clipped-out.
This procedure and operational philosophy is integrated into all our software and all of our
people know that the procedure is to be followed without exception. The philosophy is
included in all the PLC and graphics templates used to assemble a new control system as a
standard procedure not only at Tijeras but throughout the company.
7
There are two other operating philosophies that are integrated into these procedures as part of
the safe operating practices. The first of these is the standard that says that no systems will
be operated remotely out of interlocks. The second is the standard that prevents any
individual equipment start buttons from being available on the computerized control system.
The system design criteria says that each system and each device will have a "Local Test"
software switch. The control room operator can only put a system or a device into Local Test
Mode or take it out of Local Test Mode. If the system is in Local Test Mode, the individual
pieces of equipment in the affected system can be started and stopped from the local switches
located near the motor but not from the control room. The person at the local switches is in a
position where he can see that everyone is clear and that there is no problem to leave the
equipment running for a short test period normally associated with working on it. If the
system is not in "Local Test" Mode, it is automatically in interlocks to assure the operator
that other required equipment in the system is operating properly when the subject equipment
is in operation.
There are a couple of other things that need to be understood in this example for all the
pieces to come together. Our motor starting philosophy includes a "Ready" signal from each
starter that indicates that power is on the breaker and any required internal signals are OK so
the equipment is "Available" to start when the start signal is given. Furthermore, a "System
Ready" computer flag is set "True" when each piece of required equipment in the system has
a Ready Signal that is "True". If a system is Not Ready, it means that some required piece of
equipment in the system is Not Ready to start. This helps prevent failed sequence start
routines for complicated systems involving many pieces of equipment that must operate
together. It is also required that the local area warning horn must be sounded for 15 seconds
followed by 20 seconds of silence before starting any equipment remotely so people in the
are have time to move away before the start is initiated. There are also "Quick Stop"
software buttons that immediately stop all equipment in a system without normal timing
sequences, and there are *Emergency Stop" buttons in the control room for each system that
will remove power to stop systems regardless of the other signals from the computerized
control system.
It is important that all persons working with or around equipment in the plant understand
these philosophies. It is also equally important that the people in the plant are able to trust
that these philosophies will be followed in all cases without exceptions. These types of
standards are the backbone of the control system that allows consistent controls to be
developed for all systems in the plant. The standards also include the way the various
conditions of equipment are displayed on the computer screens. The graphic standards also
need to take situations such as color blind control room operators into consideration so all
persons can see the various modes of operation to insure the safe operation of the equipment.
SIMULATION AND I/O CHECKOUT:
As the control programs for each system were completed, a simulation program was
completed to test the programming. The simulation software replaces the physical I/O so the
control software thinks it is attached to the real world. This way, the programming people
can set up situations that occur in the real world and try them without being attached to the
8
real world I/O. This permits trying to run the system entirely in software to catch any
programming mistakes including I/O assignment mistakes. This also permits the training of
the operators to try it before we try to go live and to get operator input into the design and
catch mistakes before we try to start the real system.
The other part of the checkout procedure is to check the physical connections for all of the
I/O. After the software has been checked, it can be used to check the hardware connections.
Since the software has been debugged, when problems show up when physically checking
each signal connection, it is usually a hardware problem. This eliminates much of the
confusion that happens during startup.
Using this procedure has been very successful and it has prevented almost all start-up
problems. As an example, the Samalayuca plant startup only required a few tuning changes
to get the plant up and running. The whole plant was started on time, on budget and was at
design capacity within one week. The use of the same procedures elsewhere in the company
has been just as successful. The main difference with the upgrade projects has been the
length of time necessary to complete the project since scheduling work periods has had to
revolve around normal maintenance outages. Because of shipping schedules, it has not been
possible to take equipment down to connect it on a schedule revolving around the upgrade
team's schedule. It's hard to change a tire on a vehicle moving at 30 mph. In spite of the
scheduling difficulties, following the procedures described above has resulted in very few
startup problems when systems have been converted. Furthermore, the conversions have
been on time as scheduled during normal outages for other items since most of the work
could be done so that the outages mainly involved switchover techniques to minimize the
amount of wiring that needed to be completed during the equipment outages. However, the
total time required for the project has been longer than anticipated originally because of the
change in scope and the availability of down time to go through the changes.
Documentation of the electrical systems is important for construction, but it is also

commonly neglected through the years of operation and changes unless the changes were
from an additional engineering project. Bringing the drawings up to date and migration to
CAD so they can be more easily maintained has been a help to maintain and troubleshoot
systems in the plant. It is often hard to attach a dollar value to this, but we estimate that we
save as much as 5 to 10 hours of operating time per month on each critical system through
improved troubleshooting methods inherent in the new control system.
LESSONS LEARNED:
The upgrade of the Tijeras plant control system has been a learning experience. In general,
we are in agreement that it is easier and faster to engineer and build a new system in a new
grass roots plant than to upgrade an aging system. The first thing you face in considering an
upgrade project is where you start the project definition and where it stops. That is, do you
include the field sensors and wiring in the plan? It takes time to go through the plant and
identify the field sensors and wiring, but this should be inventoried and you should get a
realistic view of the condition of all the wiring and sensors to be included. The next thing
that should be considered would be to look at any potential additional sensors that should be
9
added to achieve your objectives. Then you need to look at where the sensors and equipment
will be mounted to see if you will have to add space for new equipment or not. Generally,
you will need more space if you plan to keep the plant running and do the new construction
while you are in operation so you only need to move connections to switch over to the new
system. This means you will need more space than the space required for only one system.
In some cases we added new wiring and cabinets at extra cost just to achieve shorter
downtime for the conversion.
The next thing to consider is the type system you want to use. Generally speaking, the open
systems are less expensive than the DCS systems if we compare initial purchase costs.
However, you need to look at what you get and what resources you have available. Most of
the DCS systems have pre-engineered systems and solutions to common control problems
included with the system. You generally just need to connect the software modules supplied
by the vendor to achieve a system. If you have the budget, you can also get the vendor to
essentially do the system to a certain specification. By comparison, the open systems
generally have the tools to develop your own system and even do templates. You will
generally either need to do the work of developing the templates and philosophies yourself or
contract with a third party to do the work for you.
There is also the problem of ongoing support for your system that needs to be considered into
the equation. If you have a vendor who is knowledgeable in the Cement Industry and
includes time saving philosophies and modules in his system that you do not have to develop
or support over the life of the system, it can be of benefit to your company. All of these
points need to be worked out internally to select an appropriate vendor for the needs of your
company and your project.
CONCLUSIONS:
Adding a computer to a system that is a mess only achieves an automated mess rather than a
desirable solution. Adding the money to the project to fix the wiring problems and correct
any inadequate sensor problems should be a requirement of any project plan to upgrade an
existing system. It is also strongly recommended that good communications occur between
the project management, the control system stakeholders, and company management to
adequately define the objectives of the project, the budget, and the expectations of the
involved parties as early as possible in the project. Clear understandings develop long
friendships and help achieve goals.
10
The four pictures below show the changes in the control room from just before beginning
work on the project in 1996 to the concept view of the finished product. Pictures 1 and 2
show the old original hardwired panels and the old XT Computer terminals for the old PLC
systems. Picture 3 shows work in progress with only one of the old original hard wired
panels still in place, the new PLC cabinet, and the elevated computer flooring. Picture 4
shows the concept of the finished control room as it exists in the Chihuahua plant.
Picture 1 Picture 2
Picture 3 Picture 4
In our case, the project was a long hard project, but everyone participating grew and learned
together. We not only achieved our objective to automate the plant, but we participated in
enhancing the operating procedures for the company and developing the standards to help
make future projects easier, faster, and cost less. We also feel that safety procedures have
been improved through better definitions and better understanding by all of the plant people.
Finally, we feel that the people participating in the project grew personally to become better
and more effective employees through the learning experience.
11
Appendix A
Figure 1
12
Figure 2
Figure 3
13
Figure 4
Figure 5
14
Figure 6
Figure 7
15
A REAL WORLD INNOVATIVE CONCEPT FOR PLANT
INFORMATION INTEGRATION
By:
Urs Herzog, Holcim Group Support Ltd. Switzerland

Rauli Hantikainen, Marnix Buysse, ABB Switzerland Ltd.

May 2002, Jacksonville, Florida
Abstract
The drive for continuous improvement is probably one of the main constants in the cement industry. So it
is the case for our information systems. On both the business administration and the production
information systems, there is a strong push for tighter integration in order to cut down on overlaps,
multiple instances of similar data, expensive maintenance and so on. Excitement alone for (isolated) point
solutions is just not enough. The value of a point solution nowadays is only appreciated, if it fits the larger
picture of the total enterprise IT strategy. In the real world, it is not economically viable to replace all
systems at the same time to the latest and same technology platforms. Technically this means that the
focus shifts to ensuring consistency of information throughout a heterogeneous landscape of production
information systems and business systems.
In production management, production accounting is a tedious manual activity in reporting balanced

production figures. There are imperfections such as sensor errors, limited measurement accuracy,
calibration drift, transient accumulation, roundup, spillage and incomplete calculation. Production figures
inevitably require reconciliation and balancing before finalizing month-end closing.
This paper also shares the experiences and benefits of a large cement producer in the implementation of
modern, integrated solutions across multiple plants.
2
Importance of information
Our industry is under continuous pressure to optimise its operations. This means plants need to lower
production costs, minimise electrical energy consumption, optimise use of alternative fuels and raw
materials and increase maintenance efficiency and equipment availability. Furthermore the plants must
fulfil the requirements of quality regulations and environmental protection legislation.
To be able to react to this continuous challenge, it is critical for plant management and plant personnel of
all levels to have flexible access to accurate, consistent and relevant information about production, quality
and process equipment. This means that the right information should be viewed as a resource that is
indispensable for optimisation of the production and business process.
A good information management strategy and support system is an indispensable success factor in a
competitive market and accordingly it becomes a central task within all leading companies.
Data and information sources in cement production
For many years, modern IT (Information Technology) tools have been extensively used to manage data
storage and data workflow, resulting in a huge amount of data. Nowadays the market offers affordable
and tailored IT systems to handle the data for any purpose.
In the technical field of cement manufacturing the key systems, the process control systems are today
fully based on computer technology. Quality control and assurance are fully automated and the computer-
based systems are state-of the art. In the last two decades, for each specific requirement of
measurement and supervision, system manufacturers have developed new software applications running
on standard hardware such as office PCs. Examples of common systems are kiln cameras and
pyrometers, gas analyser systems, emission monitoring systems, energy management systems, on-
stream and on-line material analysers, automatic laboratory systems and automatic dispatch systems.
2000
1800
Measurements per Plant
1600
1400
1200
1000
800
600
400
200
0
1966 1979 1984 1992 2001
Year
Figure 1: Number of measurement points in a process for an average cement plant
All these systems measure and store data for immediate use and for later analysis. The amount of
technical data gathered, calculated and stored in a cement plant has grown significantly during the past
decades. Figure 1 illustrates this.
3
Data integration concept
The process measurements accumulated over time, by information source system, result in a huge
amount of data. Often these systems (or point solutions) are linked through a company-wide network and
access is provided to everybody by using, for example, web-browser technology. This alone is not
sufficient to achieve real integration. It does not bring coherency and consistency. Nor does it avoid
duplication of data, inputting, etc. Analysis on interdependent data is not possible.
10'000'000
1'000'000
Data Points per Day
100'000
10'000
1'000
100
10
Quality Control
Consumptions
Alarm & Event

Measurements
Runing Hours
Messages
Process
&
Figure 2: Type and amount of data to be stored daily
Figure 2 shows typical amounts for different data of a cement plant that are stored daily. The data is also
consolidated for uniform and easy access and analysis of stored data.
Business process versus manufacturing process
The administrative work of modern companies has already moved to a new era of IT systems. Integrated
ERP (Enterprise Resource Planning) systems have left behind the practice of numerous legacy systems
with high maintenance costs. Currently, integrated ERP systems are used almost exclusively, and the
advantages are significant. Data is entered and maintained only once. Duplication of data is avoided and
cross-functional use of data is an implied advantage in the concept (see figure 3).
The concept of managing the necessary information for manufacturing process optimisation is similar to
the established and proven solution for business process optimisation: the fully integrated ERP system.
The vital feature in this concept of production process data is the full integration of the system for the data
collection, storage and analysis. In the author’s company, this is called the Technical Information System
(TIS). The seamless integration of all data types from the different data source enables the system to
convert data to useful information with a larger context value. Correlation and interdependency can be
shown with powerful analysis tools. The TIS provides mechanisms and tools to validate, correct and
reconcile data. This enables fast and effective input to enterprise wide reporting (EIS) and to the ERP
system. In addition more effective support for daily production, quality, maintenance, environmental and
4
management functions is provided. The TIS bridges the gap between the real-time data environment
(instrumentation, control systems) and the transactional data environment (ERP).
To illustrate the solution, the complexity and the benefits of an integrated TIS, the example of production
accounting is further elaborated upon. It is just one of the functional areas that shows a clear case of the
significant benefits to be enjoyed from this integrated concept.
Enterprise Information
System
EIS
Optimising Technical Information Optimising Enterprise Resource

Production the Business
System Planing
and Process Process
Management TIS ERP
Internet,
Phone /
FI / Fax mail
PDH A&E REP PA LIMS PM MM SD HR
CO
Emission Energy Process Laboratory

Control
Warehouse Personnel Client

Spares Finance Customer
Planning Controlling Sales
e-business
Manufacturing Process Business Process
Figure 3: Integrated Information Management Concept for Business and Manufacturing related Processes
Production accounting
The practice of production accounting measures the inputs to a production plant in terms of raw materials,
energy, consumables and other resources. It then accounts for the distribution of these values among the
multiple products and outputs of the operation. In a convergent process like clinker and cement
production, a portion that is 'unaccounted for' is generally not acceptable and is therefore to be kept to
absolute zero. In the cement production industry, production accounting activities are usually focused on
the month-end closing and at the day-to-day administration of materials. Accounting is required for
incoming (shale, coal, oil, gypsum, bauxite, alternative fuels, …) , outgoing (clinker, cement) and
intermediate (shale, raw meal, clinker, …) materials, consumed and produced. Keeping accurate figures
on volumes and quality is of primary importance to the economics administration of the business but is
also vital for efficient and flexible daily production management. In addition, production plants are
constantly looking for improvements, and achieving them may require capital or operational expenditure.
Reliable and correct data are key to the investment justification. Production accounting usually on two
aspects:
- The production accounting for the purpose of month-end closing and reporting: production books are
established, reviewed and corrected, typically on a monthly basis. The effort for a single cement
5
production plant without integrated information systems easily varies from 5 to 20 man-days per
month.
- The production accounting for the daily management of the production plant focuses more on the on-
going, quasi real-time administration of production plant resources. The purpose is to support short
term planning, scheduling and optimisation. In practice, to minimize the work, many assumptions and
rough estimates are used as inputs to the production plant planning. The result is regular surprises
with costly measures to correct them, such as a long waiting time for bulk load carriers because silos
were less full than assumed.
Production accounting is a complicated, time-consuming, regular activity in production plant

management. In a non-integrated environment, the different point solutions produce result figures but
more often than not, they are inconsistent, incomplete and incorrect. The systems that produce these
figures are typically point historians, ERP, dispatch systems, sales admin systems, amongst others. This
is a typical heterogeneous environment: several and different systems, from different suppliers, all with a
different point solution focus. The difficulty comes when balancing or correcting these results. The
collected figures need to be matched with reality and need to be made coherent within and between
these different systems. After the results are corrected, reconciliation of the source data is required, in
order to maintain consistency.
Errors and inconsistencies
Where do the errors and the inconsistencies come from? Quality of human data input typically accounts
the majority. However even with full automation within the process space, significant errors ripple through.
These systems are not perfect and errors occur due to sensor errors, limited measurement accuracy,
calibration drift, transient accumulation, malfunctioning, human error, handling and rounding of data,
product spillage and incomplete calculation and implementation practices. Whether automated or not,
production results inevitably need balancing and reconciliation.
The difficulty mainly occurs with materials in a storage location where volumes and quality cannot be
directly, automatically or correctly measured. There are plenty of examples of these in a cement and
clinker production plant. Some examples are raw material stockpiles, raw mix silos, clinker silos, cement
silos, coal stockpiles, pre-blending beds.
Generally, all material storage points, whether physical or logical, and their information can be modelled
as shown in the Figure 4. Typically for one material storage point, five types of data are available. For the
purpose of this paper, we have established the following names for the types of data:
- 'Produced' figures represent the cumulated measurements of material going into the storage point
e.g. weigh belt feeders. The feeder measurements are collected continuously, typically in a process
control system and logged historically in a point historian or directly into the production information
management system. The cumulative over time period is a good but not exact indication of the
quantity of material that was brought into the storage point by a conveyor belt. The accuracy of such
an information loop, for the case of belt weighers is in the best case only around 1%.
- 'Received' figures are the transactional figures on the material that enters the storage point. In this
context, it is not relevant whether these figures represent material received from internal or external
suppliers. It could be shipped-in material from another company but it could also be material supplied
from another profit or cost centre within the same company. In any case, this transactional
information is for the purpose of production accounting 'correct' and usually not available for
correction once it is accepted.
6
Measured
Manual / automatic
Manual: occasional or regular
Produced
Typically measured & calculated
Continuous
Storage
e.g. Weighbelt feeders
Received
Typically transactional Clinker silo, cement
e.g. ERP silo, raw material
stockpiles, fuel tanks, Shipped
waste fuel storage,
grinding aids, additives Typically transactional
... e.g. Weighbridge, Dispatch, ERP
Consumed
Typically measured & calculated
Continuous
e.g. Weighbelt feeders, flowmeters
Figure 4: Material storage point
- 'Consumed' figures are identical in character to the 'Produced' figures but are for the cumulated
measurements of material leaving the storage location. Again in the cement and clinker production
industry, this is mainly information from belt weighers with typically a high inaccuracy.
- 'Shipped' figures are identical in character to the 'Received' figures but are for the cumulated
measurements of material leaving the storage location. Again, these could be physical shipments, but
they could also be material transferred to another profit or cost centre within the same company, local
or remote to the production plant.
- 'Measured' data is used when a material storage point cannot be directly and automatically
measured. The actual content of a storage point is then typically determined (measured or estimated)
manually and serves as reliable data to verify and correct data retrieved in other ways. The
'measured' data is established regularly e.g. once per week, month and/or occasionally. The data is
entered into the production management system manually. In most production plants, the stock taking
or 'measurement' is not done on a daily basis. In cases where permanent instrumentation is able to
measure the material volume (depending on silo size and type), resulting data provides excellent
redundancy for the reconciliation of production data.
A practical example for a storage point without permanent instrumentation is shown in figure 5.
7
In a typical modern cement plant, there are between 20 and 40 material storage points that are of
interest. The main variables influencing this number are the number of fuels used and the number of
cement types produced.
2 Shipment: single transaction e.g.

V from S&D admin system
c 3 Stock received transaction e.g.

from G/R admin
c
1 4 Weekly stock measurement for
verification & correction
c
Produced
c c Continuous update & tracking :
Consumed
stock increase as per production
rate and consumed rate
Measured Shipped Received Measured t
1 2 3 4
Figure 5: stock tracking on a non-instrumented material storage point
Industry practices for correction and reconciliation
There are different practices in the industry and they can be grouped into three categories.
1. Manual correction of the results only

This is the most common practice. It requires the least effort but it leaves the danger of accumulating
errors across multiple reporting periods until, for example, a year-end closing. The focus is on reporting
coherent figures for the closing reporting period. Typically this resulting data is used an input to the
business systems. Here, only the results are corrected to ensure coherence with the larger picture. No
effort is undertaken to reconcile the lower figures. The result is inconsistency between reported figures
and figures in the data sources. The problems show up as soon as there's some ad-hoc reporting to be,
where maintaining consistency quickly becomes very complex. In addition, since no correction was made
on the source data, the reliability and usability of the data for internal auditing, analysis, daily
management and so on, is limited.
2. Combined correction of results and source data and reconciliation within a limited scope
In a more automated production management environment, this is a feasible option. Here, the production
plant manager corrects the results or the source data at various levels or both. The automated system
takes care of reconciliation throughout the data hierarchy. For reconciliation down wards, usually a fixed
or a weighted factor is used. Reconciliation further down than hourly values brings in practice no relevant
advantage. The reason is twofold:
8
- For production management purposes a lower granularity data is not used, so reconciliation is not
really necessary.
- The factors, which are still artificially deducted for the purpose of correcting the high level data, have
no relation to the real error in the raw source data and the error model. In most cases, the error model
is not even known or established. Reconciling the raw source data to the lowest level would make
these data more inaccurate and even less useful.
This practice needs the support of a more complex, but above all integrated, information system. The
complexity and the volume of work are too high to do manually at a reasonable cost. Standard office
software such as spreadsheets fall short as well, since they are not suited to flexible handling of time
series data with bi-directional dependencies in the data hierarchy with circular referencing. The solution
requires an RDBMS based system.
3. Correction of raw (source) data

This is practice is widely used in other industries like utilities, petrochemical, mineral processing, etc.
Here, the quality of raw point data is improved before rolling up to high-level data. There are different
methods for improving the errors in the data: statistical methods, use of redundant data, tuning the error
using neural network techniques and process model based techniques.
The TIS uses a combination of the second and third practice, as this gives the most flexibility. The
outcome can be achieved in a customised solution, but for TIS, a productised solution with a standardised
configuration for easy rollout into the production plants was chosen. The economics of a productised
solution also allows for a lower purchase cost and a lower cost of ownership throughout the system's
lifecycle.
The solution in TIS
Daily
Point data Calculated data
(e.g. one weighbelt feeder) (e.g. total produced clinker) Reporting
Business
transactions Year
Business transactions
constraints Discrepancy
Actual at
Month end
Actual at
begin
Dispatch, ERP, Automatic
Week Sales Shipped reporting into
administration, ... ERP after
Received
reconciliation
Day and validation
Process Consumed Consumed
characterist Calculation
ics:
Produced Produced
physical, Shift
chemical, Roll up
thermodyna
mical, ...
Transactional
Hour
Importance of Data volume

reconciliation constraints
< Hour
Transactional data
Time series data
Security protected accessible entry point for

correction & reconciliation with possibility to
drill down to lower level figures and
entry points. Activities logged for audit purposes.
Process characteristics constraints
Real-time
Figure 6: Production accounting in the 'real-time to transactional' context
9
A summarised overview of the solution is shown in figure 6.
Over a reporting period, all available data (usually not coherent) are summarised and presented in an
overview and discrepancies are highlighted. Wherever discrepancies are detected, there is a need for
correction and reconciliation. If the results are corrected here, then the system automatically takes care of
down-wards reconciliation towards the source (point) data across all the time scopes down to hourly level.
Direct correction of the results is commonly too rigid. Often, in order to make a more sensible correction,
the production manager needs to know what source data and calculations delivered the results. From the
overview, the user can now drill down in each of the results to review the intermediate results, source
data, the calculations and the roll-up steps.
This provides full transparency and enables a more realistic and confident adjustment or correction of
data. This correction can be made on any level in the data hierarchy. The system takes care of automatic
reconciliation up-wards and down-wards. For systematic and reoccurring corrections, correction factors
applied to the source data are adjusted. If this is done systematically, then the relative discrepancy of the
results minimises over time. With any correction, the user can 'preview' the consequences of the
correction before saving.
The result is a coherent set of data within TIS. The results are now ready for validation and reporting, on
paper or directly and automatically into the business administration system (ERP). At this point, the result
figures are becoming transactional data. The bridge between real-time and transactional is made, and
accounted for.
Product Measured Stock Calculated Stock and all elements that contribute to this calculation. The Discrep
and last measure correction form for each contributing element can accessed directly from ancy
date here. factor
Discrepancy becomes ‘1’

after balancing the figures
i.e. reporting figures
correspond with actual and
booked transactions
Drill down into calculation and source Correction on source

elements of this production figure for data, intermediate or on
balancing and correction. end result or combined
Figure 7: User interface for production accounting corrections
This illustrates clearly the benefit of an integrated TIS. Production accounting is only one of the functional
areas where an integrated TIS brings extensive value to the manufacturing process management.
10
The benefits of using an integrated TIS
The first and easily quantifiable benefit of an integrated TIS of such scope is a lower manpower
requirement. Equally important but more intangible benefits are identified in its function as a solid
foundation for improving plant performance.
Automation of data acquisition and pre-processing of all types of data reduces the amount of manual data
input and manual data transfer drastically. Duplication of data and inconsistency is avoided. Minimizing
the number of databases and legacy systems reduces the overall cost for supporting production
management significantly. The savings are of the same magnitude as experienced with the introduction of
integrated ERP systems a few years ago. Easy and standardised interfacing between TIS and ERP
enables business synergies and even increases the efficiency of some ERP functions. In our industry, for
example, maintenance and spare part management can be optimised with accurate data from the
equipment and process.
Results and feedback from other industries confirm this finding. A Gartner Group study concludes that:
“Manufacturing businesses that make investments in ERP-directed manufacturing applications that fail to
provide for accurate real-time information from the process will achieve at least 50% lower ROI ...”
Gartner Group
An ERP expert in the chemical and pharmaceutical industry has stated that:
"If you implement traditional ERP and do not connect it to your plant, you can leave 40% to 60% of the
benefits on the table.”
Udo Edelmann, former Director of Chem/Pharm, SAP
Feedback
An inquiry of TIS users in more than 20 plants shows significant benefits and savings. The use, or value-
added, of the different modules of the TIS are shown in Figure 8.
Process value analysis and laboratory data management are used in all applications. Data consolidation
for short and long-term report data is very highly ranked. These functions are only available on integrated
TIS with interfaces to ERP and EIS systems. They yield the biggest reduction in manual data entries and
manual data manipulation, replacing the typically numerous spreadsheet reports.
Feedback from users illustrates the main benefits:
- Increased Manpower Efficiency: Minimizing manual data manipulation enables the plant personnel to
concentrate on evaluating the data. Timeliness of reporting is improved drastically. Manpower
savings are in the range 4 to 16 man-months a year. In one plant the duration to produce the monthly
report was reduced from 12 to 2 days. The fulltime job of production accounting was reduced to 4
man-hours per week.
- Support for production and maintenance management: the main advantages are fast and consistent
short-term report data about process and equipment. Changes in plant performance can be
monitored continuously. Consistent data is the key input for process optimisation, proper production
planning optimisation, energy management and storage facility management. TIS provides base data
for condition based maintenance and helps to detect problems before they arise. Production
increases of 2% and more have been reported.
- Proof of Performance and Compliance: Combined processing and logging of data from the burning
process, chemical composition of fuels, materials and stack emissions are legislative requirements in
most countries that permit burning of waste fuels. An integrated production management system
11
provides an excellent basis. TIS further provides the data to verify the process and performance
guarantees of process equipment suppliers and support damage liability cases. On several
occasions, TIS has supported claims worth more than 500'000 USD each.
Figure 8: Use of the TIS modules (results of more than 20 applications)
Payback
The value creation of an integrated TIS for an enterprise can be calculated with the Net Present Value
Method (NPV). Based on a plant capacity of 1 Mio tons cement, the savings achieved as mentioned
above and with a Weighted Average Cost of Capital (WACC) of 8%, the payback is less than a year with
an Internal Rate of Return (IRR) of around 60%.
Conclusion
To stay competitive in cement production in an ever more demanding environment, continuous

optimisation of the manufacturing process is indispensable. Modern integrated information systems are
the key to the necessary fast access to all relevant plant information. The example of production
accounting, shows the complexity but also the advantage and need for an integrated TIS. The practice of
scattered point solutions for information management is not sustainable and has been superseded.
Experience from users proves the clear benefits and savings of integrated TIS systems in cement
manufacturing.
12
Total Integrated Device Operation Network
(TIDON)
By
Michael Voigt
FL Smidth

ABSTRACT
Uniformity of the network in the area of automation has led to positive developments
not only in flexibility but also in minimising the initial installation costs and the
commissioning time.
The concept of “TIDON” Total Integrated Device Operating Network has major
advantages for the integration of all components together with an overall design of
communication interface between the devices.
The integration of all the components has now become possible between different
vendor’s products, however we must not neglect potential problems arising from this
opportunity. Even though it may seem attractive to use standard components from
different vendors to reduce purchasing costs, we most not ignore the complexity of
integrating components from different original purpose of design.
This Paper will highlight the concept that we are using in the process control system
for cement plants, underlining the benefit from this environment in the prospect of
technical and commercial issues.
INTRODUCTION
Data information technology is rapidly introduced in the business environment.
All modern company must have state of the art data information capability for
controlling the company business on all levels.
Systems like ERP, MES and Process control can become one common system with
advanced graphics and other new capabilities and features. Interface between these
levels has been in focus for some years now, however we have looked in another
direction. We have focused on the data information capability from operator system
level down through field level.
Process control system including the network design must be designed by engineers
or programmers with practical experience of real-life operational conditions at the
cement plant. This paper proposes a concept to ensure that, the chosen system
meets the needs of a well-operated data network with different vendor’s components.
WHAT IS “TIDON” AND HOW IS IT USED IN PROCESS CONTROL SYSTEM.

The concept of TIDON is that various brands of control equipment are implemented
into a common integrated network.
As the major manufacturers of control equipment are offering equipment with serial
interfaces based on open standards such as Profibus, Mod-bus, CAN-bus, etc. it has
become possible to integrate equipment from various manufacturers and still achieve
an integrated network for controlling the electrical devices.
This gives the advantages of selecting between the various brands of equipment to
meet specific client demands, achieve the most cost benefit solution, etc and still
establishing an integrated network.
2
SYSTEM CONCEPT FOR TIDON
Total Integrated Device Operation Network
Operator system level
Industrial Ethernet
PLC system level
Seriel bus
Device control & device

monitoring level
Instrument bus
Field level
As shown on the figure Total Integrated Device Operation Network the system
consists of four levels:
• Operator system level

• PLC system level
• Device control & device monitoring level
• Field level
HOW TO INTEGRATE THE EQUIPMENT FOR THE OPERATION NETWORK.
A NUMBER OF CRITERIA MUST BE TAKEN INTO CONSIDERATION IN ORDER TO SELECT

THE DEVICES AND DESIGN THE NETWORK.
1. Customer requirements
2. Vendor acceptance
3. Control system
4. Available standard equipment
5. Analysis of basic system functionality
6. Cost comparison of major vendors’ capabilities, installation and commissioning
7. Maintenance
8. Management Information level
3
1. CUSTOMER REQUIREMENTS
First of all the customer must have an overall idea about the communication
between the control system and the various types of equipment on the operator
level, device control level, device monitoring level and field level. We will
recommend the TIDON concept, unless the customer demands a classic wiring
I/O interface between PLC’s and the various devices. Depending on the
acceptance of equipment and brand it will be individual from customer to
customer how the network will be designed.
The design of the network will also depend on the type of project. If the project is
a new plant, the requirement will be different from a retrofit project where some
existing equipment will be reused.
2. VENDOR ACCEPTANCE
When a customer is considering accepting a vendor to his plant some different

criteria will be used. Some of the criteria can be the service capability or the
relation the vendor has in other electrical areas. The most important issue must
be the reliability for the customer.
3. CONTROL SYSTEM
The customer will always choose or have chosen a control system that is suitable
to run a cement plant. If the control system is designed only to a lower level of
controlling and monitoring devices in the plant, it can be difficult to get a benefit
out of the TIDON concept. The concept is in principle a device operation network
with a bus-system of some kind. The bus-system will be designed according to
the possibility that the control-system will offer. On the other hand if the customer
must select a new control system to his plant, then we will suggest that the
design of the network will be the guideline when choosing the control-system.
4. AVAILABLE STANDARD EQUIPMENT
We do not have any ambition to ask vendors to make new devices specifically for
the cement industry. We will always look into the standard components that are
available on the market and use the possibility from different vendors. We will
then choose the component that is most suitable for the project and best practice
in the cement industry. The benefit for the customer is that we are erecting from
3-5 totally new cement plants and many new single departments every year
where the experience is adapted from one project to the next. We are in a
position to be best in class because the amount of projects will give the best
practise in the cement business. Basic system functionalities with respect to
development to meet the requirements.
4
5. ANALYSIS OF BASIC SYSTEM FUNCTIONALITY
The analysis needed for choosing the right network will depend on the basic
system functionality. What level of information does the customer need for
operating the process? The customer most decide how many signals are required
to successfully operate a particular device.
Some of the basic information in the monitoring is available in the devices and if
access to the information must be flexible, it is important that the design of the
network is with a bus-system.
6. COST COMPARISON OF MAJOR VENDORS CAPABILITIES, INSTALLATION AND

COMMISSIONING.
As vendors have their strength and weakness and depending on the design it is
possible to choose between different vendors’ components. The point will be a
split of criteria that are important for an individual customer or project supplier like
us. For a cement plant supplier, it is the cost, installation flexibility and the
commissioning friendliness that will be in focus, but also the experience from
earlier delivered plants is an important factor for choosing the component.
7. MAINTENANCE
Maintenance is one of the most important issues for the customer. The best
service provider and the maintenance tool for the devices will be a considerable
initial benefit for the investment and life-cycle cost. If the devices can be equipped
with a mechanism that can indicate the service interval for other installations in
the plant the TIDON concept can be a tool for the maintenance program for the
total installation of the plant.
8. MANAGEMENT INFORMATION LEVEL
All modern companies will look into the information technology for running the
cement plant. Efficiency of a plant is very important for the return on investment
and the stockholders' satisfaction. The efficiency of a plant can be established
with an information technology and the “TIDON” concept will be an important part
of the task to create the modern cement plant. The ERP and the MES level will
be a part of the future modern cement works, but is not taken into consideration
in this paper.
With the TIDON concept we have prepared the new generation of modern
cement plants with new information technology.
SUMMARY:
The following section will be the best cases that we have in this area. We will
explain the requirement and the solution for the design of the network. The
benefit from our point of view of the selection will be highlighted.
5
CASE STORY I
THE HISTORICAL BACKGROUND

We started quoting for the Buxton/England project in 1998-99 and got the contract for
the project in the spring 2001.
During the final contract negotiations the concept for an integrated device network
was discussed and incorporated into the project as both the client and the supplier
saw the benefits of incorporating this technology into the project.
Operator level
For the Buxton project the operator platform is a windows based control system
implemented with standard industrial computers.
The operator system is linked to the PLC-system through a redundant Ethernet

connection.
Control room
OS OS
OS
Server Server
IndustrialEthernet)
Profibus DP
Electrical
rooms
Field
PLC LEVEL
During the procurement phase for the Buxton project two suppliers were offering PLC
hardware, but due to the fact that one of the supplier already had won the contract for
MV-panels, MCC and frequency drives. This supplier was able to give the most
economical quotation for the PLC-hardware,
For the Buxton project the PLC system is based on this vendor hardware and
standard software combined with our system software for implementing the PLC-
programmes.
6
DEVICE CONTROL & DEVICE MONITORING LEVEL
For the Buxton project all motor starters are equipped with intelligent motor starter
units. All medium voltage panels and all frequency drives are equipped with interface
for serial interface.
The serial bus between the PLC’s and the intelligent motor starter is implemented as
a Profibus DP, similar is the serial interface for the medium voltage panels and the
frequency drives etc. are implemented as Profibus DP.
If required all field instruments can be connected through an instrument bus suitable
for installation in the field.
BENEFITS FROM A INTEGRATED SOLUTION

Through the integration of the devices into the operation network a number of
benefits are achieved during the design phase, the erection phase, the
commissioning as well as the maintenance of the plant
DESIGN OF MOTOR STARTERS

Motors with a nominal current between 0,25 A to 100 A are covered by only three
types of intelligent motor starters: (0,25 A - 6,3 A, 6,3 A – 25A, 25 A – 100A), where
a standard solution would require 15 – 20 different types of overload protection to
cover the similar motor range.
An additional advantage is that a intelligent motor starter unit includes a number of

different motor protection functions and covers all various starting conditions
including heavy duty starting, which further minimises the required number of
different motor protection types compared with the traditional solution.
WIRING, CABLING AND TERMINATION

The saving in wiring, cabling and termination becomes obvious when comparing the
diagram for the control circuit for a conventional Star-delta starter and the similar
control circuit based on an integrated device control unit.
7
Conventional Star-delta control circuit
3/N/PE ~ 50/60 Hz 400/230V
L1
L2
L3
N -F1
PE
Automation level
-Q1
-Q1 Cu
-F2 Th rre
Ov er nt
erl mi val
-F3 Au oa
O Of st ue
-S3 Auto to d
n f or
Manuel
-S1 PLC
-K1 -K2 -K3
-S2 -K1
-K1 -K3
-F2
-K2
-K4
-K2 -K3 -K2
1U 1V 1W -K1 -K1 -S3 -F2 -F3
1 4-20
mA
1
M -K3 -K2 -K4 -K1 -F3
4-20
mA
3~ Controlgear
2W 2U 2V
Star-delta control circuit with integrated device control unit

3/N/PE ~ 50/60 Hz 400/230V
L1
L2
L3
N
PE
PROFIBUS-DP Transmission of control commands, diagnostic, operating

and statistic data
-K2 -K3 Manuel/Auto Stop Start

-K1
-S3 -S1 -S2
1
Overload Ready
A1 A2 T1 T2 1 2 3 4 5
Current unbalance Gen. Fault
Phase failure
Thermistor
Earth fault 3UF50 EIN
Blocking protection Test/

Current recording Reset AUS
1U 1V 1W
1
M
3~ -K1 -K2 -K3
2W 2U 2V
WIRING CHECK BEFORE COMMISSIONING

A traditional hardwired solution requires a time-consuming and costly wiring check of
all the total wiring before starting up the commissioning.
A system based on the use of integrated device control units with serial interface can
be factory tested for all functions except the serial plug connection to the serial
network. As the serial network consists of two wires only it becomes rather easy to
check for correct wiring of the network.
8
BENEFITS DURING ERECTION AND COMMISSIONING
A result of the wider range for each type of overload protection will be that the
number of elements to be changed due to divergence in the actual installed motor
sizes are decreased rapidly, this will course reduced cost and time during erection
and commission.
Compared with the traditional hardwired solution an integrated solution gives

extended possibilities for supervision and troubleshooting during commissioning as
all operation and diagnostic data can be made available for operator if required
during commissioning.
The figure below shows some of the features for vendors intelligent motor starter,
similar facilities are available for the frequency drives and the Medium Voltage
panels.
BENEFITS FOR THE MAINTENANCE OF THE PLANT

Due to the reduced number of different types of motor protection units, the number of
spare parts for motor starters can be significantly reduced compared to what would
be required for a traditional solution.

extended possibilities for diagnostics in case of trip due to overload, thermistor
protection, rotor blocking, etc.

extended possibilities for introducing preventive maintenance for critical equipment
based on statistical data of the numbers of starts, numbers of overload triggered and
operating hours for the selected equipment.
COST BENEFITS
Compared with the cost for a traditional hardwired solution, the cost for an integrated
solution is slightly higher for the hardware part.
9
As the traditional solution is more costly regarding cabling, wiring and termination,
the overall cost for the integrated solution is still in the same level as the cost for a
traditional hardwired solution.
The key figures are for the present as follows:
INCREASES IN COST
• Hardware for MCC increases ≈ 20 % in cost, but if current measurement of the

control system is required this increase will be covered by the saving in hardware
for the current measurement.
NO ADDITIONAL COST
• Hardware for MV-cubicles are almost unchanged as nearly all protective relays
are supplied with serial interface as standard.
DECREASES IN COST
• Cabling for signal interface to MV-panels are reduce by 66 %.
• Cabling for control interface to MCC-panels are reduced by 75 %
• Cabling for control interface to local control systems are reduced by 66 %
• Hardware for frequency drives decreases ≈ 10 % as an average.
• Hardware cost for the control system decreases ≈ 20 % as an average.
Further, the total erection time for the control and signal cabling will be reduced due
to the reduced number of control and signal cables.
As the hardware cost for the equipment with serial interface is expected to decrease
in the future due to increased use of this equipment, the overall cost for a integrated
solution will decrease thus making this solution more interesting not only from
a technical point of view but also from an economical point of view.
CASE STORY II
THE HISTORICAL BACKGROUND
We have worked out a retrofit concept for modernization of existing electrical

installations, especially in the control system area with related equipment. We have
delivered a retrofit solution to project Langkawi/Malaysia based on a device-operating
network.
The tender submitted by the client asked for a new control system and retrofit of the
Motor Control Centres (MCC) with intelligent components. There was also a wish for
monitoring of power in several places in order to better control power consumption,
more than 50 measurement points.
10
This is normal in the majority of MCC switchgears in the existing factory we are
asked to check the condition. Except for the contactors and thermo relays the MCC
in the factory were in a good condition. It is rarely necessary to completely replace
the MCC – if you have to replace the MCC switchgear, you normally also have to
replace the cables. The selected solution was to replace components in the MCC
switchgear with intelligent motor starters and additional components.
Another supplier (from the Buxton project) was selected for both the MCC
modification and for the PLC system hardware.
LANGKAWI/MALAYSIA CONTROL SYSTEM LAY-OUT.
11
The configuration we are using is Ethernet between the PLC system and the
Operator System. Control Net from the PLC to the individual I/O, and Device Net
between the PLC and the MCC, both the intelligent modules and the block I/O
module. For monitoring of the power consumption it was decided to use a Power
Monitoring unit which is also on the Device Net.
It was decided not to place any I/O's around in the factory, as there already
were junction boxes for collection of all the other signals and new shielded
multi-cables were installed from the boxes to the PLC cabinet. Another reason
for not installing I/O in the existing boxes was the temperature and the IP
rating of the boxes.
The experience today is to follow the design specification from the
manufacturer, and use a general, good engineering practice. All the
departments have been commissioned without any problems with the
network. The converting of the MCC was also faster than we normally see on
converting jobs. Generally our experience with the network connection is very
positive and we are suggesting this solution every time we have to do a
modification on a Motor Control Centre.
MCC UNIT BEFORE REWIRING.
12
Rewired MCC units, one with Digital I/O only and one with intelligent unit both
from same vendor
EXPERIENCE SO FAR AND CONCLUSION:
We are very focused on our new network concept TIDON. We are using standard
components from various vendors, but also use our best practical knowledge to
perform according to cement applications. The results from the different plants where
the TIDON concept has been used totally or partly in the projects have shown that
the technical and the commercial part can be approved from a business point of
view. In the future we will get even more experience with different types of vendor
equipment and have a full range of best practice for implementing the TIDON
concept to the benefit for the worldwide customers.
13
The Use of Mining Software to Optimize the
Limestone Mineral Deposit at the Colton Quarry
By:
James Chadwick, Mining Engineer, California Portland Cement Company

John Morrison, General Manager, Datamine North America, Inc.

Abstract
This paper addresses the implementation of a mining software package at the Colton
Quarry and discusses the advantages and benefits that have been realized by
management and engineering personnel in both day-to-day mining operations and long
term strategic planning. The paper describes how the implementation has progressed
from initial installation to its current state and continuing developments.
Introduction
California Portland Cement Company has mined limestone at its Colton Quarry for use
in the manufacture of cement for the last 111 years. The Colton plant produces different
types of cement for use in various market sectors. The quarry supplies the plant with
approximately one million tons of limestone per year. In addition, approximately one
million tons of waste material are removed per year to allow for the development of the
quarry to access further limestone reserves.
Geologically, the Colton Quarry consists of two main limestone units divided by an
intrusive granitic schist. The limestone and schist layers dip in an easterly direction at
approximately 45 degrees. The limestone unit lying to the west of the granitic schist has
many intruded granite dykes. This results in a variable low grade quality in the
limestone. The east unit is a high quality dolomitic limestone. The geology surrounding
the limestone units is a combination of granites and alluvium material. The intrusive
schist, alluvium and granite materials cannot be used in the cement manufacturing
process and are the main area of waste extraction.
The limestone is mined on 25 foot benches using front-end loaders and haul trucks that
transport the material to a primary crusher. The limestone is crushed to 6-inch minus
and conveyed to secondary crushers where it is further reduced in size to 1-inch minus.
This crushed material is conveyed to a storage building for the cement plant's use. The
crushed limestone material is sampled every hour and analyzed for its chemical
composition. Quarry management monitors these analyses to control the limestone
chemistry targets in the production.
Historically, the short term planning of the quarry on the day-to-day production cycle
relied on management reviewing chemical analysis of blast drill holes for a particular
shot and visually inspecting the mining area where the limestone was to be extracted.
The drill hole data was not used to predict future material extraction in areas and was
not plotted on any topography maps. The topography maps were updated using free
hand drawing and relied on aerial flyovers on a biannual basis to update quarry
development.
Historical Data
Before the implementation of the computer mining system, all of the mining data was
based around a paper system. The paper system consisted of mining plans, drill hole
logs, reserve estimation summaries using drill hole information and contour maps of
chemistry and topography. This information was updated by hand on an irregular basis;
therefore, regular reconciliation between actual versus planned data was problematic
and inconsistent.
2
The historical data, the only information available, was entered into a digital database as
the reference point for the computer mining system. The historical data can be
summarized into three main categories: survey system, drill hole data and mining plans.
Survey System
A survey control system of known points was established around the perimeter of the
quarry. These points were used as reference targets for an aerial survey that was
completed to produce a detailed digital topography of the Colton Quarry. This
topography was used as the initial reference for survey control of the quarry by the
computer mining system.
Drill Hole Data

The drill hole data was comprised of exploration drill hole information. This information
was stored on log sheets and contained details of drill hole coordinates, geologic
horizons and results of chemical analysis. This data was manually entered into a
spreadsheet.
Mining Plans
The mining plans consisted of geologic interpretation and mining boundaries on a
bench-by-bench basis. The interpretation and boundaries were established from the
geologic and chemical analysis of the exploration drill hole data. This information was
digitized directly into the computer mining system.
Initial Processing of Historical Data

Once the data had been converted into a digital format, it could be used in the computer
mining system. All captured data was imported and validated within the system where it
could be viewed in a three dimensional environment. The validation procedure
highlighted several problems with the drill hole information, such as missing or incorrect
coordinates, overlapping drill hole samples and missing or incomplete chemical analysis.
These problems were corrected prior to further processing by manual editing.
The next step of the data processing was to convert the geologic interpretations found
on paper into a three dimensional block model, as shown in Figure1. Block modeling is
a well established technique of representing insitu geology as solid three dimensional
shapes. Chemistry values from the exploration drill hole data were interpolated and
integrated in the block model using standard estimation techniques. The result produces
a new three dimensional block model that represents the insitu geology and chemistry
variations throughout that geology. This type of block model is commonly referred to as
a “resource model.” User-defined reports and bench plans were produced from the
resource model. The reports included limestone tonnage, associated waste and quality
on a bench-by-bench basis for the east and west units. The results for the entire quarry
compared favorably with the existing reserve estimates. This was to be expected as the
input data used to generate the results was the same for the traditional and computer
methods.
3
Figure 1. West-East geologic section through the Colton Deposit
Short Term Planning using the Computer Mining System

One of the main requirements of the computer mining system is to be able to perform
short term planning functions to meet tonnage and chemistry targets required at the
plant from limestone and other blendable materials. Normally, the resource model is
used as-is in the short term planning cycle. However, the current resource model did not
correlate favorably with actual production from the working areas. This is due to the
resource model being based on exploration drill hole data instead of localized drill hole
sample data. A new procedure was developed to incorporate the drill hole data and
update the resource model using the latest available information.
The first stage of short term planning is to perform the fieldwork to collect survey
information. This is collected using a Global Positioning System (GPS) that records the
perimeter of the blast area and the individual drill hole coordinates. Each drill hole is
identified by a unique identification number. This information is downloaded into the
computer in the office and plotted on existing topography maps of the quarry. These
maps are used by the quarry management to determine possible material extraction
areas and future drill locations. The drill operator samples each of the holes drilled in a
shot and sends the sample to the laboratory for analysis. The results of the laboratory
analysis are then combined with the survey data to produce a drill hole file that includes
both the positional and chemical information of the drill holes in the blast. The drill hole
data file is imported into the computer mining software where it can be viewed
graphically and queried in a three dimensional environment.
4
The drill hole information is a useful guide to the quality of the blasted material but does
not give the actual quantities of material. The software allows the user to estimate the
quantity and quality of the material within the specific blast perimeter using automated
modeling and estimation techniques. The results are output in both a graphical plot and
tabular format; see Figure 2. These results provide management with the ability to
identify areas of limestone and waste within the individual blast.
Figure 2. Blast Report and Contour Map
The individual blast information is stored in the computer mining system. It is also added
to a working drill hole database that is updated for each blast. This database of
information is used to create bench plans of drill holes that can be contoured on any
user-defined value, estimate future blast areas and reconcile actual versus planned
production. These functions allow management to identify geologic trends within the
quarry based on the latest information, which results in improved future mine planning.
Long Term Planning using the Computer Mining System

Historically, the long term plan was generated from existing mining plans. These plans
were used to calculate the total tonnage and average quality on a bench-by-bench basis.
Blending between the benches used weighted averages on quality variables to satisfy
production targets. The problem with using calculated quality averages is that it does
not allow for localized variations in both geology and chemistry. This problem can be
overcome using the computer mining system.
5
The long term planning cycle requires an up-to-date resource model. This model is
generated using updated survey and geologic information. The geology is modified
using the current drill hole data. This information is used to re-interpolate the chemistry
values in the resource model. The latest topographic information coming from the GPS
is incorporated into the previous topography to create a new mining surface. The
resource model is then trimmed against the new mining surface. The long term plan
then uses this updated resource model to predict the complete mineral extraction
sequence of the resource.
The first step to determine the extraction sequence is to use the resource model in the
computer mining system and apply economic constraints, such as mining and
processing costs versus a product value. The resultant file contains information on costs
versus potential income. The next step is to optimize the deposit using an algorithm that
maximizes the deposit while honoring user-defined chemistry targets. Geotechnical
parameters, including wall slopes and physical mining boundaries or limits, may be
incorporated into the optimization. The resultant file from the optimization contains
information on the largest possible quarry that honors the chemistry targets and the
physical constraints. Targets can be defined in terms of a production rate, quality values
or quality ratios and multiple targets can be used.
The remaining two steps involve generating practical mining shapes for the benches
based on the production rate and blending between available material using user-
defined targets. This creates an optimal extraction sequence that is both achievable in
terms of quarry geometry and chemistry for the life of the quarry. The computer mining
system also allows for stockpiling material either from the quarry for processing later in
the mine life or as an external source to assist with blending.
The computer mining system used for the long term planning has been designed
specifically for the cement industry. The fundamental difference between this system
and other software packages is that it allows for maximizing the resource based on
blending targets as opposed to traditional techniques that attempt to maximize the Net
Present Value (NPV) of the deposit.
6
Figure 3. Examples of graphical displays from the long term scheduler
The computer mining system can produce output in a variety of formats, such as bench
or section plots, spreadsheets, charts and three-dimensional animations of the quarry
development with time; see Figure 3. In addition, management can evaluate different
production scenarios quickly and accurately. Such scenarios may include variations in
product specification over the life of the quarry and minimizing costs of re-handling
material and/or purchasing of external additives.
General Utilities of the Computer Mining System

In addition to the surveying, modeling, short and long term planning and mine design, a
number of general utilities are available. The inventory control of raw material stockpiles
and other materials is an important function that requires monthly updates and
reconciliation. A stockpile can be surveyed using GPS and the data downloaded into
the computer mining system. This data can be used to generate the surface of the
stockpile and allow the software to determine the volume of the stockpile by comparing
the initial topography and stockpile surface.
The computer mining system can also be used to design ramps and benches within the
quarry as it develops. These designs can be viewed in any three dimensional
orientation and can be used to ascertain the practicality of the quarry design and the
environmental impact of the operation throughout the mine life.
Another important tool of the software is the ability to predict the cut and fill volumes for
roads. The computer mining system provides the tools required for road layout including
gradients, road width and turning radius. The road design is then converted into a
surface and, by using an angle of repose, cut and fill volumes between the road and
current topography can be determined.
7
Benefits
The current computer mining system provides all the functions necessary to assist in the
running of the Colton Quarry on a day-to-day basis. As a result, there are numerous
benefits that can be realized in adopting such a system. The benefits include, but are
not limited to:
• Full optimization and extraction of mineral resource

• Time and labor savings through automation of repetitive tasks
• Quick sensitivity analysis of quarry potential due to changes in parameters such
as market conditions, costs, production rates and quality targets
• Current information readily available for decision making by management
• High compatibility with other commonly used software packages
• Good reporting, visualizing and plotting tools
Future Developments
There will be continual development and improvement of the existing computer mining
system. Such improvements include data access via the company network and
incorporating short term blending into the current short term planning system. This
blending will use the results from long term planning and blend available faces on a
short term basis using constraints such as minimizing movement of mining equipment at
the face and quality targets. Scheduling on a short term basis within the long term plan
will ensure that both long and short term planning are fully integrated and targets are
met.
Summary
This paper highlights the benefits and methodology of moving from a paper based mine
planning system to the computer mining system at the Colton Quarry. By adopting this
technology, data is collected and used to produce a better and more consistent quality
limestone product allowing full utilization of the mineral resource.
8
Drives and Related Products
Working Group
Chair
Reinhold A. Errath
ABB Industries AG
Vice Chair
Michael C. Pendo
Dacotah Cement
Members
Terry L. Ball
The Falk Corporation
Jeffrey A. Nagel
RESCO Products Inc.
Ron Parks
BMH Americas
Douglas S. Phares
General Electric Company
Clayton H. Reid
Rockwell Automation
James L. Schaadt
RC Cement Co., Inc.
DEALING WITH MOTOR WINDING PROBLEMS
CAUSED BY INVERTER DRIVES
By:
Mark Fenger, Steven R. Campbell, Iris Power Engineering Inc.
Jan Pedersen, Techwise A/S
Abstract: Random wound stator windings operating in utility and industrial plants have failed
when exposed to the fast rise-time voltage surges coming from inverters. Studies show that the
failure is due to a combination of bad luck in specific motor installations (resonance phenomena
caused by power cable length and surge impedance ratios) together with the fact that modern
inverted-fed drives (IFD’s) create tens of thousands of surges per socond with rise-times as fast
as 50ns. Measurements on motors show that these surges create partial discharges (also called
corona) and these discharges may eventually destroy the turn-to-turn and/or phase-to-phase
insulation, resulting in premature motor failure.
The paper will discuss the specific mechanisms involved in the stator winding failure due to IFD’s
and present the measurements and analysis from surge monitoring installed on many different
motors. Although some motors may experience short rise-time, high magnitudes surges, most
motors experience either low magnitude and/or long rise-time surges, which are relatively
harmless. Usually, several different magnitudes and rise-times are present from the same IFD.
Thus, it seems that conventional motor stators can be safely used in many (but not all) IFD
applications. Methods will be presented to determine when special IFD duty motors are needed.
Voltage surge tests, as well as partial discharge tests, can help the user insure that motors can
successfully operate in severe applications.
Introduction: Researchers have understood for over 70 years that fast rise-time voltage surges
from a circuit breaker closing can lead to an electrical breakdown of the turn insulation in motor
stator windings [1]. If the turn insulation is of an insufficient thickness, or has aged in service, the
insulation will puncture when a short rise-time voltage surge occurs. Punctured turn insulation
allows for a very high circulating current to flow into the affected copper turn, rapidly melting the
copper conductors, which, in turn, results in a consequent burning/melting of the slot liner
insulation, thus leading to a stator winding ground fault [2,3].
Rapid advances in power electronic components in the past decade have lead to a new source of
voltage surges. Inverter-fed drives (IFDs) of the pulse width modulated (PWM) type that use
insulated gate bipolar junction transistors (IGBTs) can create tens of thousands of fast rise-time
voltage surges per second. Anecdotal evidence suggests that the huge number of voltage surges
from IFDs can lead to gradual deterioration and eventual failure of the turn insulation - both in low
voltage (less than 1000 V) and medium voltage (2.3 to 4.16 kV) motors [4-6]
This paper describes measurements of the surge voltage characteristics from a group of eight low
voltage motors driven by IFDs. As described in this publication, two motors have repeatedly been
subjected to unexpected stator winding failures. Inspection of the windings after failure indicated
that the main cause of the failures is the voltage surge environment applied to the stator winding.
Review of the impact of voltage surges on LV stator winding insulation: An investigation of

the surges applied to random wound stator windings by IFD’s show that these surges may have
frequency components up to 5 MHz or so. At such frequencies, the stator windings appear as a
complex ladder network with low impedance capacitive shunts to ground. The capacitive shunts
cause most of the applied surge voltage to be dropped across the first few turns in a stator
winding.
By injecting a 5V pulse from a variable
80.0
70.0 rise-time pulse generator (HP 8012) into
a stator winding, and by measuring the
Magnitude (%)
60.0
50.0 voltage across the first turn with a
40.0 differential very low capacitance probe,
30.0 measurements were conducted to
20.0 experimentally determine the amount of
10.0
0.0
voltage that appears across the first turn
0 100 200 300 400 500 600 700 800 900 100
in a stator winding, as a function of the
0 voltage rise-time – see Figure 1. As
much as 75% of the surge voltage
tr (ns)
applied to the terminals can be
distributed throughout the first coil.
Figure 1: Voltage drop across first coil versus Furthermore, Figure 1 shows the voltage
surge rise-time in a small random wound motor distributed across the first coil, relative to
the surge magnitude, is inverse
proportionate to the rise-time. The higher
the voltage across the first turn, the higher the risk of experience a Partial Discharge (PD).
Consequently, fast rise-time surges of higher magnitudes have a high risk of inducing PD in the
random wound stator winding.
Figure 2 and Figure 3 show the surge waveform measured at the terminals of a 10 HP, 440 V
squirrel cage induction motor fed by a 600 V pulse width modulated type of drive, which uses
IGBT’s. The waveforms were measured via an oscilloscope using low inductive resistive voltage
dividers attached at the motor drive and the motor terminals. There is about 30 m of shielded
triplexed power cable between the drive and the motor. This drive created 10,000 surges per
second. The recorded waveforms had risetimes as short as 80 ns. The highest magnitude
recorded was about 1200 volts, or about 3.3 per unit, with 1 pu corresponding to the rated peak
line-to-ground voltage of the motor. In Figures 2 and 3 the top trace is the A phase signal at the
drive. The other three traces are at the motor. The scope was in peak hold mode.
Figure 2: Surge Environment on a Motor. Figure 3: Zoom of an individual fast rise-

Ch. 1 is Aø at Drive, Ch. 2 is Aø at Motor, time surge. Ch. 1 is Aø at Drive, Ch. 2 is
Ch. 3 is Bø at Motor, Ch. 4 is Cø at Motor, Aø at Motor, Ch. 3 is Bø at Motor, Ch. 4 is
Vertical axis: 1: 2000. Cø at Motor, Vertical axis: 1: 2000.
Figure 2, which shows a full AC cycle of
applied voltage, shows that there is much "Calibrated" Surge Plot
more ring overshoot at the motor terminals
resulting in bipolar surges of varying 3.65
magnitudes. Thus, in order to characterize 3.30
the surge environment, surge measurement 2.89
Magnitude (pu)
must be carried out at the motor terminals and DIV Curve 2.55
not at the drive. 2.20
1.86
Furthermore, Figure 3 shows one surge from 1.45
the drive can create several surges at the 1.10
motor terminals of different rise-times and 0.76
magnitudes. Hence, the stator windings are
0.36
subjected to a distribution of surges.
0.00
1000
1100
1200
1300
1400
1500
100
200
300
400
500
600
700
800
900
0
Thus, in conclusion, Figure 1 shows the
shorter the rise-time, the greater the voltage
across the first turn and the more dangerous Rise-Time tr (ns)
becomes the surge to the motor. Therefore,
by fully characterizing the surge environment Figure 4: Example of a Calibrated Surge Plot
applied to a random wound stator winding, [12].
one can assess the risk of stator failure due to
electrical degradation.
Finally, as outlined in [12], the measured surge environment may be assessed quantitatively by
performing Partial Discharge Inception Voltage (DIV) measurements as a function of rise-time for
the stators examined. By superimposing the DIV curve on the measured surge plot, the surge
plot is essentially calibrated with reference to which surges may give rise to partial discharges.
Surge Measurement System: As described earlier, IFDs create tens of thousands of voltage
surges per second with varying magnitudes and rise times as fast as 50ns. As shown above, the
voltage surges from an IFD pulse applied at the stator terminals may be measured by
conventional means via a simple resistive voltage divider and a digital oscilloscope. This system
allows for the measurement of the waveform of a given surge.
However, as discussed in [11], digital oscilloscopes exhibit an inherent limitation when used for
recording surges: A vast majority of surges are ignored since a digital oscilloscope can only be
triggered from 1 to 10 times per second, while 20,000 surges may occur in the same interval.
Hence, only 1 in about 1000 surges can be recorded. In addition, the oscilloscope can normally
only be triggered on the largest magnitude surges. As moderate magnitude surges, with very fast
rise-times may be more damaging to the stator insulation than high magnitude/slow rise-time
surges, it is likely that the oscilloscope may not trigger on the surges which, in time, are most
likely to cause insulation failure.
In order to assess the severity of the electrical surge environment in which a motor operates, a
reliable measurement of the distribution of electrical surges must be performed. The distribution
of surges is defined by the magnitude, rise-time and repetition rate of each surge applied to the
stator. When the exact surge environment is known, the surge distribution is said to be
characterized. Thus, the surge environment cannot be characterized via conventional means.
To overcome the limitations outlined below, a special electronic instrument was developed. This
device, SurgAlert™, measures the magnitude and rise-time of every surge that occurs within a
given time interval. It also determines the total number of surges that a motor is subjected to
during the measurement interval. However, this instrument cannot record the entire waveform of
each surge. The monitor has the following specifications:
• Wideband (50 Hz to 10 MHz) resistive or capacitive voltage dividers, capable of operating
on motors rated up to 13.8 kV. For motors rated 600 V or less, a resistive voltage divider
is used. The dividers must be installed at the motor terminals.
• A portable electronic instrument, which is temporarily placed near the motor for the
duration of the measurement, which digitally records the rise-time and magnitude of each
surge, and stores this information in memory.
• A laptop computer that downloads a summary of the measured surges recorded in the
measurement interval, for display or printout.
The data acquired may be exported to a computer file, which is readable by Microsoft Excel. It is
thus possible to perform further processing of the data acquired. Reference 7 gives further
details of the surge monitoring system.
In-service Failures Due to IFD’s: Skaerbaek Power Station, Unit 3 is a combined power and
district heating plant. The net electrical output is 390 MW. The plant was put into operation in
July, 1997. A large number af variable speed drives have been installed for operation of pumps to
reduce the unit´s house load. Motors rated 90 kW (120 H.P.) and below are mainly supplied from
the 400 V busbar whereas motors rated above 90 kW (120 H.P.) are supplied from the 690 V bus
bars. A total of nine motors, of which seven participate in this survey, are supplied from the 690 V
bus bars. The size of these motors vary from 130 kW (175 H.P.) up to 1,890 kW (2,520 H.P.). All
seven motors and the belonging inverters are from the same supplier. The basic data for the
motors participating in this survey is given in Table 2
Since the commisioning two of the seven motors have been subjected to stator winding failures.
One motor (850 kW) failed three times within the first 36 months of service and another motor
(680 kW) failed once after 36 months. On the motor that failed three times, the third failure
occured after less than 6 days of service. Operating hours and number of starts for the 2 motors
that failed are given in Table 1.
After the first failure on Motor 1, the

motor manufacturer was confronted
with the suggestion that fast rise-time
voltage surges originating from the
IGBT inverter may have lead to the
premature winding failure. This
theory was rejected by the
manufacturer who came to the
conclusion that the failure was
accidential and that it was quite
unlikely that a similar failure would
occur again.
Following the second failure on the

same motor, the theory of fast rise-
time voltage surges being the cause
of failure was brought up again. Figure 5: Picture of Motor 1 After Failure
Once more, this theory was rejected
Motor Application Power Rating Date of fault Total operating Number of starts
[kW] hours
10-06-1998 2140 4
1 Main cooling pump 850 10-02-2000 13044 119
14-09-2000 13177 152
2 Condensation pump 680 13-10-2000 9440 753
Table 1: Failure Times for Motors 1 and 3
by the manufacturer. The conclusion
was, as before, that the failure was
accidental. The manufacturer
accepted however to perform on site
voltage measurements at the motor
terminals in order to assure the
customer that the failure was
accidental and not caused by voltage
surges. These measurements were
performed only two days before the
third failure on the same motor
occurred.
One month later, a second motor

failed. This failure lead to the decision
to perform an independent
measurement of the electrical surge
environment on all 690 V motors, Figure 6: Close up of Motor 1 Failure
using the surge monitor.
Although both motors that failed were still covered by the manufacturer guarantee, the
unforeseen failures have lead to considerable expence to the power station.
Motor/IFD Configuration: The basic data for the motors participating in this survey is given in
Table 2. The basic data for the motors participating in this survey is given in Table 2.
The following information on the specific winding design is given by the motor manufactor:
• The round wire is insulated with a quadruple build, class H enamel.
• The coils of different phases are

completely seperated with mica paper
on the overhangs.
• All the coils are separated with

mica paper on the nose area.
• In the slots, the coils are insulated

to ground and between each other
with layers of NOMEX
• The winding is VPI impregnated in

a special blend of class H, flexible and
thixotropic resin.
The resin was oven cured and the
stator turned while in the oven. Such a
process gives an even coverage on
the overhangs.
The motor manufactor insists that

random wound motors with the above
described specific design can be used
for this type of application.
Nevertheless motors supplied for
Figure 7: Enhanced Close Up of Motor 1 Failure. similar applications at another Danish
Local Melting of the Stator Winding Is Clearly power station in 1999 were delivered
Visible. with form-wound winding construction.
Power Rated
Drive Cabel Cable
Motor Application Rating Voltage
Model Type Length [m]
[kW] [V]
1 Main cooling pump 850 690 SVTL 1K2 XLPE 131
2 Condensation pump 680 690 SVTL 840 XLPE 114
3 Drain pump 130 690 SVTL 210 XLPE 75
4 Small pump 500 690 SVTL 600 XLPE 15
Main condensation
5 485 690 SVTL 600 XLPE 45
pump
Main condensation
6 485 690 SVTL 600 XLPE 53
pump
7 Big pump 1890 690 SVTL 2K4 XLPE 33
Table 2: Overview of tested motors.
Measurement Procedure: By attaching a clamp-on-type probe (a very low inductance resistive

voltage divider) to the motor terminal of a given phase, a measurement of the surge environment
of each phase could be performed. A three-dimensional surge plot characterizing the surge
environment applied to that phase could thus be created.
Using alligator clips to connect to the motor terminals provides an easy and quick way to perform
a measurement. The alternative is to temporally install low-inductive voltage dividers prior to
performing the measurements. This ensures that the protection equipment will not trip the motor
due to slight voltage imbalances between phases due to the increased load (from the instrument)
on one phase. This option is more time consuming, as it requires down-time to install each
voltage divider, and, in most cases, is not technically necessary. However, local plant regulations
may, or may not, allow for connecting a probe during normal on-line operation.
First, two measurements were performed on the same phase of a motor: A 5 second
measurement and a 10 second measurement. By normalizing the surge counts for each
measurement into surge counts per second, a direct comparison between the two tests can be
made.
This procedure allows for investigation of the consistency of the surge environment applied to the
stator winding. If the surge environment is consistent, only one test is needed per phase to fully
characterize the surges applied to the stator winding. This issue is discussed later in this
publication.
The output is plotted as a three dimensional curve, see Figure 8, with the left scale being the
magnitude of the voltage in p.u., the bottom scale is the rise-time of the surge in nanoseconds,
and the vertical scale indicating the number of surges per second for each combination of surge
magnitude and rise-time. Note that this is a log scale.
Often, a 2D representation of the three-dimensional plot is used – see Figure 9. A color scheme
thus provides information of the surge count rate. As described earlier, the surges most likely to
cause winding failure will have a short rise-time and high magnitude, that is, they will appear in
the lower right part of the three-dimensional plot. A two dimensional representation allows for
quick identification of these.
Results – The difference between phases: All phases were tested for all motors. The
measurements showed that the surge environment measured on one phase was very similar to
that measured on the remaining two phases.
Figure 8 shows the surge plots for phases U,
V and W of Motor 1. Phase U
As can readily be seen, there is no significant

difference in the surge environment applied to
each phase of the stator winding. This is 10000
hardly surprising as, from a theoretical point of
1000
view, the surge environment is defined by the
Surge Counts (pps)

output of the IFD drive. 100
0.00
0.36 10
The plots of Figure 8 are typical of those 0.76

1.10
1
obtained on all eight motors tested, i.e. no 1.45 0.1
1.86
significant differences in the surge
0
2.20
200
400
environment could be detected between Magnitude (pu) 2.55 3-4
600
2.89
800
phases of a machine. 3.30 2-3
1000
3.65 1-2
1200
tr (ns)
1400
0-1
Furthermore, both 5 and 10 seconds tests -1-0
were performed. Both tests were normalized
to a 1 second test. A comparison between Phase V
these tests showed similar surge distributions
for the 5 seconds and 10 seconds test, which
is indicative of consistent surge environment.
10000
Hence, for interpretational purposes, only one
measurement per phase is needed to address 1000
Surge Counts (pps)

the severity of the surge environment applied 100
to a given stator winding. 0.00

0.36 10
0.76
1
1.10
1.45 0.1
1.86
Results - Comparative Analysis: An initial
0
2.20
200
inspection of Table 3 clearly shows motors 1

400
Magnitude (pu) 2.55 3-4

600
2.89
800
and 2 to be subjected to the highest surge 3.30 2-3

1000
3.65 1-2
1200
environment in terms of pulse count rates and tr (ns)

1400
0-1
pulse magnitudes. The highest surge -1-0
magnitude measured on Motor 1 was 3.5 per
unit at a rise time of 1500ns whereas the Phase W
highest surge magnitude measured on Motor
2 was 3.1 pu at a rise time of 1400 ns.
Fortunately, these high surge magnitudes are
measured at relatively high rise-times. Such
10000
high magnitude surges should not be too
damaging to the insulation. 1000
Surge Counts (pps)
100
However, an investigation of the 2D plot for 0.00

0.36 10
Motor 1 reveal the presence of a series of 0.76
1
1.10
surges having rise-times of 100-650 ns with 1.45 0.1
1.86
magnitudes ranging from 0.25 per unit to 2.2
0
2.20
200
per unit and repetition rates of 1-10 and 10-

400
Magnitude (pu) 2.55 3-4

600
2.89
100 pulses per second – see Figure 9. The
800
3.30 2-3
1000
3.65 1-2
1200
presence of these pulses is more of a concern tr (ns)

1400
0-1
than the presence of the maximum magnitude -1-0
pulses of 3.5 p.u. at a rise time of 1200ns
having a repetition rate of 1 per second. As Figure 8: Surge Environment for Motor 1
3.65
Max. Fastest Max Slew
Motor 1 Motor Magnitude Risetime Rate
-1-0 0-1 1-2 2-3 3-4
3.30 [pu, ns] [pu, ns] [pu/µs]
2.89 1 3.51, 1500 0.24, 50 5.1
2 2.34, 900 0.24, 50 6.0
2.55
3 3.09, 1400 0.24, 50 5.3
4 2.20, 900 0.24, 50 5.3
Magnitude (pu)
2.20
5 2.41, 1100 0.24, 50 3.8

1.86
6 2.41, 1100 0.24, 50 4.0
1.45 7 1.44, 950 0.24, 50 3.7
1.10 8 2.41, 700 0.24, 50 5.8
0.76 Table 3: Result Summary of Tests

0.36 mentioned earlier, the higher the rise time,
0.00 the higher the electrical stress between turns
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
or phases. Hence, although not being of

Rise-Time tr (ns)
alarmingly high magnitudes, the faster rise
time pulses measured on Motor 1 may be of
3.65
Motor 2 a concern with respect to aging of the stator
-1-0 0-1 1-2 2-3 3-4
3.30 insulation.
2.89
Furthermore, the 2D plot indicates two types
2.55
of surges: Initial surges created by the IFD
Magnitude (pu)
2.20 and secondary surges possibly created by

resonance phenomena. This is indicated by
1.86
the general separation of surges into two
1.45 “islands” in the plot – see Figure 9. The
1.10
surge environment measured on Motor 1
constitutes the most significant surge
0.76
environment measured so far using the
0.36 SurgAlert technology. The data showed that
0.00
similar observations can be made for Motor
3.
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
Rise-Time t r (ns)
Table 3 show the highest slew rate measured
3.65 for Motor 1 is 5.1 and 5.3 for Motor 2. These
Motor 5
slew rates can be classified as being
3.30
-1-0 0-1 1-2 2-3 3-4 moderately high compared to the highest
2.89 slew rate of 8.1 per unit measured so far on
2.55
other machines elsewhere.
Magnitude (pu)
2.20
The surge environment for Motor 2 is also
1.86 given in Figure 9 and can be characterized
by a maximum pulse magnitude of 2.34 per
1.45
unit at a rise time of 900 ns. The measured
1.10 slew rate is 6 and that constitutes the highest
0.76
slew rate measured on these motors. As
such, it can be argued from a general point of
0.36
view, that this motor is subjected to the
0.00 “worst” surge environment of all the motors
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
0
measured in this test. An investigation of the

Rise-Time t r (ns) surge plot shows the initial surges for Motor 2
to have noticeable higher repetition rates
Figure 9: 2D Surge Plot for Motors 1, 2 and 5 than those measured for Motors 1 and 3.
The surge environments measured for the

Surge Magnitude & Slew Rate vs. Cable Length
remaining five motors are, without doubt, of
less concern. As can be seen from Table 3,
7
the maximum surge magnitudes measured are
6 noticeably lower than those measured on
5 Motors 1, 2 and 3. This is consequently
[per unit]
4 expressed in the 2D surge plots – an example

3 from Motor 5 is given in Figure 9. The slew
2 rates, however, appear to be moderately high
1
for the remaining 5 motors with Motor 6
0
yielding a slew rate of 4. This is the lowest
0 20 40 60 80 100 120 140 slew rate detected for the eight motors
Cable Length measured in this test.
Max Slew Rate Max. Surge Magnitude Linear (Max. Surge Magnitude)
Common for all motors, the lowest rise time

Figure 10: Surge Magnitude and Slew Rate
measured was 50 ns yielding surges having a
As a Function of Cable Length.
magnitude of 0.24 per unit. Although this rise
time can be categorized as short, it should be noted that the measured magnitude is relatively
low.
Results – The influence of cable length: Figure 10 shows the relationship between calculated
maximum slew rate and cable length for each measurement. It furthermore shows the
relationship between maximum measured surge magnitude as a function of cable length. Figure
10 suggests that a clear relationship between maximum surge magnitude and cable length exists:
The surge magnitude appear to increase with increasing cable length. Given the nature of
travelling wave theory, this is not a surprising result.
In the case of the maximum slew-rate, Figure 10 suggests that the maximum slew-rate does not
directly depend on the length of the cable connecting the motor to the IDF drive. The slew rates
is defined as the ratio between the surge magnitude and surge rise time. As documented in
Table 3, the rise-times measured for the maximum surge magnitudes range from 700 ns to 1500
ns thus giving rise to an erratic distribution of slew-rates for increasing surge magnitudes. Figure
10 thus suggest that the slew-rate is a secondary effect of the cable length and it can thus not be
concluded that the longer the cable length, the more damaging the surge environment is.
Partial Discharge Inception Voltage: The basic principle of the test setup is sketched in Figure
11. Via a Baker Surge Tester, Model D12000, a 50 ns rise-time surge voltage is applied to an
insulation sample or a stator winding. If of sufficient magnitude, the surge voltage will give rise to
a partial discharge. The partial discharge give rise to a high frequency current signal, which is
consequently extracted from the surge via specialized instrumentation, PDAlert™, connected in
series between the surge source and the insulation sample. The net output from the instrument
is a voltage signal originating from the partial discharge current pulse itself. An example from
Motor 1 is given in Figure 12. The leads connecting the various components of the test setup are
kept as short as possible.
Out Test Specimen
The test procedure is described
thoroughly in [12] but repeated in Stator
Oscilloscope
short here: Having connected a stator
In Out
to the surge source, the surge
Srg Out
Trg Out
PD Out
Surge Tester
magnitude was increased with
approximately 200 volts per second
from zero volts until a partial
discharge was observed. The surge PDAlert
magnitude was then quickly
decreased to zero volts. The Figure 11: Sketch of DIV Test Setup
DIV vs. Rise-Time For Motor 1
3.5
2.5
DIV [pu]
2
1.5
0.5
0
0 100 200 300 400 500 600
Rise-Time [ns]
Figure 13: Discharge Inception Voltage vs.

Figure 12: Example of PD on Motor 1 Rise-Time For Motor 1.
procedure was repeated seven times for each rise-time. Based on this, the mean (average) DIV
was calculated for each rise-time. In addition, the ambient temperature and humidity was logged.
The average discharge inception voltage was 1.73 per unit for Motor 3. For Motor 1, the average
DIV was 2.38 per unit. Table 3 shows that the maximum magnitudes - of surges having rise
times up to 1550ns – is below the DIV for these motors. That strongly indicates that these motors
are subjected to partial discharges, during normal operating conditions, due to the surge
environment applied to the stator windings from the IFD and the connection cables.
A curve of the DIV versus rise-time for Motor 1 is given in Figure 13. The curve shows the DIV to
decrease with decreasing rise-time. This is surprising as other curves obtained on new stators
prior to being put into service shows the opposite relationship, namely an increase in DIV with
increasing rise-time as explained by the distribution of voltage across the first turn as a function of
surge rise time.
Figure 14 shows the measured surge plot for Motor 1 with the DIV curve super imposed. Surges
above the curve may give rise to PD whereas surges below the curve will not give rise to PD.
As can be seen from [12], the DIV

"Calibrated" Surge Plot For Motor 13.65 measured on these motors are low
compared to those measured on virgin
3.30 stators. The results obtained on virgin
2.89
stators showed DIV’s of between 6 per
DIV Curve unit to 9.5 per unit – as seen from Figure
2.55 15 where Stators 1 and 2 are virgin
Magnitude (pu)
2.20
stators and Stator 3 constitutes the DIV
for Motor 1 [12].
1.86
1.45
Compared to the results obtained on
virgin windings and presented in [12],
Surges resulting in PD 1.10
these results indicate that for aged
Surges not resulting in PD 0.76 windings, i.e. windings subjected to real
operating conditions, DIV is more related
0.36
to the surge magnitude rather than the
0.00 rise-time coupled with the probability for
0
1000
1100
1200
1300
1400
1500
100
200
300
400
500
600
700
800
900
occurrence of a free electron, which

Rise-Time tr (ns) increases with increasing (slowing) rise-
time.
Figure 14: Calibrated Surge Plot for Motor 1
Discharge Inception Voltage Versus Rise-Time
- Stators -
10 50
9 45
8 40
Discharge Inception Voltage [pu]
Standard Deviation / Mean [%]

7 35
6 30
5 25
4 20
3 15
2 10
1 5
0 0
0 500 1000 1500 2000 2500 3000 3500
Surge Rise Time [ns]
Mean DIV (Stat 1) Mean DIV (Stat 2) Mean DIV (Stat 3)
St Dev/Mean (Stat 1) [%] St Dev/Mean (Stat 2) [%] St Dev/Mean (Stat 3) [%]
Figure 15: DIV’s for Motors of Different Design And Size [12]
Discussion: The measurements clearly showed Motors 1 and 2 to be subjected to a surge

environment which must be perceived as having a negative impact on the stator winding
insulation and its estimated life time.
Confronted with the measurements performed using the SurgAlert instrument, the motor
manufactor have accepted to supply and install filters on 5 of the 9 motors. Filters have already
been installed on the 850 kW motor that failed 3 times and will be installed on the remaining four
motors as soon as possible.
Filters are simple 20 µH reactors connected in series with the motor at the cable outlet from the
inverter. The impact of these reactors are not known at present. Preventive filters have been
installed on Motors 1 and 3. Additional surge measurements are to be be performed once
installed. The measurements will thus clearly document the effect of filters.
The Discharge Inception Voltage measurements strongly indicated that the surge environment
applied to Motors 1 and 3 gave rise to partial discharge activity.
Conclusions: The surge measurements clearly showed the two motors previously to have
experienced failures to be subjected to the worst surge environments of the motors measured
here. Furthermore, the DIV measurements performed on Motors 1 and 2 clearly documented that
the Discharge Inception Voltages, under surge conditions, were lower than the max surge
magnitudes measured on-line during normal on-line operations. This strongly indicate that the
root cause of the failures experienced were indeed the presence of partial discharges.
Also, the measurements showed little difference in the applied surge environment between
phases on the individual machines indicating when performing this types of measurements,
measuring one phase per motor should be sufficient.
Furthermore, when purchasing inverter drives intended for 690 V supply, the manufacturer should
be asked to indicate acceptable surge levels at the motor terminals, and should be required to
preform measurements of the actual surge inviroment when the motors are being commisioned.
Acknowledgements: The authors would like to thank Jens Aage Jensen of SV Production for
his cooperation on the work presented in this publication.
References:
[1] E.W. Boehne, Voltage Oscillations in Armature Windings Under Lightning Impulses, Trans
AIEE, 1930, p1587.
[2] M.T. Wright, S.J. Yang, and K. McCleay, General Theory of Fast-Fronted Interturn Voltage
Distribution in Electrical Machine Windings, Proc. IEE, Part B, July 1983, p 245.
[3] B.K. Gupta, et al, Turn Insulation Capability of Large AC Motors, Parts 1, 2, 3, IEEE Trans
EC, December 1987, p 658.
[4] A.L. Lynn, W.A. Gottung, D.R. Johnston, Corona Resistant Turn Insulation in AC Rotating
Machines, Proc. IEEE Electrical Insulation Conference, Chicago, October 1985, p 308.
[5] W. Yin, et al, Improved Magnet Wire for Inverter-Fed Motors, Proc. IEEE Electrical
Insulation Conference, Chicago, September 1997, p. 379.
[6] E. Persson, Transient Effects in Applications of PWM Inverters to Induction Motors, IEEE
Trans IAS, September 1992, p1095.
[7] G.C. Stone, S.R. Campbell, M.Susnik, New Tools to Determine the Vulnerability of Stator
Windings to Voltage Surges form IFDs, Proc. IEEE Electrical Insulation Conference,
Cincinnati, October 1999, p 149.
[8] L. Saunders et al, Riding the Reflected Wave, Proc. IEEE Petroleum and Chemical Industry
Conference, September 1996.
[9] E.P. Dick et al, Practical Calculation of Switching Surges At Motor Terminals, IEEE
Transactions On Energy Conversion, December 1988, Vol. 3, No. 4, p 864.
[10] C. Lanier, A Novel Technique For The Determination of Relative Corona Activity Within
Inverter-Duty Motor Insulation Systems Using Steep-Fronted Voltage Pulses, IEEE
Conference Record of The International Symposium on Electrical Insulation, Arlington, VA,
June 1998, p 229.
[11] G.C. Stone, S.R. Campbell and S. Tetreault, Inverter Fed Drives: Which Motor Stators are
at Risk?, IEEE Industrial Applications Magazine, p. 17 f.f. September 2000.
[12] M. Fenger, S. R. Campbell and G. Gao, “The Impact of Surge Voltage Rise Time on PD
Inception Voltage in Random Wound Motors of Different Designs”, 2001 Annual Report -
Conference on Electrical Insulation and Dielectrics Phenomena, pp. 352-355, October
2001.
SPOTTING DRIVES FOR LARGE MOTORS
By:
Barry Dick, Senior Application Engineer
GE Toshiba Automation System
Introduction: High power line-fed, constant speed motors sometimes must be operated at very
low speed for controlled positioning of the attached lo ad equipment. This process is referred to
as ‘spotting’ or ‘inching’. Examples include cement grinding mills (both raw and finish) and metal
processing rolls. These motors are usually medium voltage (>2300 V). Two methods of spotting
motors have been used in the past: 1) bumping the motor with the medium voltage motor starter,
and 2) an electromechanical inverter. This paper describes the application of a low voltage
(460V) PWM adjustable speed drive to medium voltage motors for spotting purposes. The drive
is capable of rotating a synchronous or induction motor at low speed under manual control to
position the machine attached to the motor to a desired position.
The Spotting Problem: “Spotting” is a process that refers to manually positioning a piece of
driven equipment to a desired position using the electric motor that powers the equipment. The
reasons for needing to spot a piece of equipment are varied, but they usually involve positioning
an opening in the equipment to a desired location or moving the equipment to a desired location
to remove it from an assembly. In any case, the requirement is to apply torque to the shaft
controllably to move the load machine to the desired location.
Previous Methods of Spotting: The spotting problem arises wh en the motor is a large medium
voltage unit powered from a fixed frequency power system. There is no effective way to apply
torque controllably to the motor directly from the power system for moving the load equipment
slowly. Over the years, two methods have been used to position large motors:
§ Bumping
§ Electromechanical Spotting Equipment
“Bumping” the motor means briefly connecting the motor to the power system to start it rotating,
then hoping by chance the drive train will stop where it is desired. This method has two
significant undesirable consequences:
1) The stopping position cannot be very well controlled, and the process would probably
have to be repeated more than once, and
2) The bumping process subjects the motor to full starting forces and currents repeatedly.
This can cause significant mechanical and electrical stress on the motor (with probable
reduction in motor life), driven equipment, and starting means. The bumping process can
also cause transient voltage dips on the power bus that can disrupt other equipment.
The other method that has been used for over 40 years is “spotting equipment”. This term refers
to an electromechanical inverter that is fed low voltage direct current from a motor-generator set
or rectifier. The spotting equipment contains six dc contactors and a motor driven relay that
sequences the contactors to produce low frequency currents in the motor stator. The low
frequency currents produced by the spotting equipment set up a revolving magnetic field in the
motor air gap that interacts with the rotor flux to produce torque. The low frequency insures the
shaft rotational speed will be slow.
While this traditional spotting equipment does work adequately, it does have disadvantages. The
technology is antiquated and maintenance intensive. The contactors and motor driven relay must
be maintained to keep the equipment operating well. Replacement of the contactor tips may be
required frequently, depending on the frequency of use and the current the contactors must
switch. Another performance shortcoming for the spotting equipment is its ability to operate at
only one speed. The output frequency is usually limited by the switching capability of the
contactors and the capability of the motor driven relay to operate at only one speed. Therefore,
no speed adjustment is possible. The most important disadvantage of the spotting equipment is
cost when compared to available drive technology.
A diagram of this type of spotting equipment is included as Figure 1. The low voltage constant
potential dc source has usually been an ac motor – dc generator set in the past. A solid state
rectifier with adjustable output is also usable. The dc source voltage is set to provide sufficient
current (torque) to move the motor and attached load. The contactors that form the
electromechanical inverter are numbered 1A to 6A. Reference is shown to the motor field supply
because the motor field must be energized for the spotting equipment to operate. The motor-
driven relay that sequences the contactors is not shown.
The spotting drive is an electromechanical inverter that changes the constant potential dc into low
frequency current to power the motor. The motor must have excitation applied while the spotting
drive is operating so that the motor is operating synchronously with the drive. Physically, this
means the rotor flux must be present for the current applied by the spotting drive to “push on” to
make torque in the motor.
Two main control quantities are present in the spotting drive:
1) The voltage on the constant dc bus, and
2) The output frequency of the spotting drive. The dc bus voltage and output frequency
were designed and adjusted to provide smooth operation of the drive train. Insufficient
voltage would result in too little current injected into the motor with resulting pole slippage
and/or stalling. The motor must start and stay synchronized from the first current pulse or
it is likely to remain stalled. Adequate torque must be generated to overcome any
overloads while running, a potential problem for grinding mill applications. Too low an
output frequency from the spotting drive results in “jerky” motion of the drive train. This
motion is the result of acceleration of the motor while the current is applied and
deceleration during the time the current is not applied. The correct frequency keeps the
motor and drive train moving smoothly and prevents wear and damage to gears,
couplings, and other drive train components.
In practice, the required motor current (and dc bus voltage) is estimated in the design phase of
the spotting drive from the motor characteristics and the load torque requirements. The final dc
bus voltage setting is then adjusted during the start -up phase of the project. The drive output
frequency was fixed by using a constant speed motor to drive the timing relay that sequenced the
contactors. By experience, the minimum frequency for grinding mills has been established to be
about 1 Hz. The maximum output frequency is limited by the switching speed of the output
cont actors. Therefore, the spotting drive is usually operated at or near the minimum frequency.
The PWM Spotting Drive: Recent developments in low voltage pulse width modulated (PWM)
insulated gate bipolar transistor (IGBT) drives make replacement of the electromagnetic spotting
equipment possible to consider. Low voltage (460 and 575 V) output drives with high current
capacities are available at attractive cost. Requirements for a PWM drive to replace the spotting
equipment include the capability to supply up to 300% of rated motor current and a drive control
capable of operating at low frequency.
If a PWM drive can be configured to be a source of low frequency current, it can replace the
electromechanical spotting drive. A comparison of the features of each spotting drive is included
in Table 1.
Electromechanical Spotting
Characteristic PWM Spotting Drive
Equipment
Power switches IBGT, no moving parts Mechanical contactors
Output current Near-sinusoidal, low Square wave, high harmonic
waveform harmonic distortion distortion
Starting Smooth ramp acceleration Fixed frequency operation.
from very low frequency Motor and load must
accelerate to spotting speed
in one current pulse
Speed adjustment Adjustable, within limits Fixed by design
Table 1 Spotting Drive Comparison
The advantages of the PWM drive derive from the solid state IGBT power switches and its design
as a motor drive with low output current distortion and digital control. The mechanical contactors
must interrupt the motor current by parting contacts, which results in sparking and contact wear.
The sparking usually gets worse as the contact tips wear. The IGBT’s produce no sparking and
are very quiet in operation.
Low distortion in the output current waveform is possible becaus e the IGBT’s are switched at a
relatively high frequency (1500 Hz or higher). With the drive output harmonics at multiples of this
switching frequency, the motor impedance is effectively a very high impedance to oppose the flow
of harmonic currents. The result is a nearly sinusoidal low frequency output current to the motor.
The advantage of the sinusoidal current is reduction of potentially damaging torque pulsations
created in the drive train.
For starting the drive train, the IGBT PWM drive starts at a very low minimum frequency (about
0.1 Hz) and ramps its output to the minimum frequency. If enough current is injected into the
motor, adequate torque will be available to line up the motor rotor with the stator magnetic field
(MMF), and rotate the motor and load. The result is a synchronized soft start for the motor and
drive train that minimizes the acceleration of the system and reduces mechanical stress. In
contrast, the electromechanical system can operate at only one speed, so the motor must
sync hronize with the first current pulse into the motor. This requires a high current pulses to
accelerate the drive train very quickly up to the operating frequency of the spotting equipment.
The result can be high mechanical stress on the motor, gearbox, and driven load.
Requirements for a PWM spotting drive: Typically, a motor that requires a spotting drive is a
medium voltage (2300 VAC or higher) synchronous motor. Induction motors are also be used,
and the PWM spotting drive can work with them as well. The application will determine how
much torque the motor must generate to move the driven equipment. If the load can be removed
from the driven equipment, the torque requirement will be reduced. For example, a grinding mill
motor may need to develop 130 – 150% torque to cause the material in the mill to cascade.
Conversely, spotting a roll in a metal rolling stand may require only 25 – 50% torque because it
can be unloaded and the only load is friction and acceleration of the roll inertia.
The basic operating principle of the low voltage PWM spotting drive is the establishment of a
revolving MMF in the motor air gap with a low frequency, low voltage power source. A medium
voltage motor can make rated torque at low speed and low voltage with proper excitation. At low
speed, the internally generated voltage of the motor (CEMF) is low, so the PWM drive does not
need medium voltage output to circulate rated current in the motor. Therefore, the spotting drive
output voltage can be rated a fraction of the motor rating and still deliver rated torque or more.
The current capacity of the drive is crucial. The voltage rating of the drive may limit the maximum
speed of the motor, but will not prevent its use.
Drive Rating
A potential PWM drive for a spotting application can be selected based on the torque required to
move the drive train and the low frequency current rating of the drive. Low frequency operation is
important because the drive output voltage will limit the motor frequency to low values. As an
example, a drive with a maximum output voltage of 460 V can operate 4160 V motor at a
maximum speed of 11%, assuming the motor is operating at rated flux. The corresponding drive
output frequency is 6.6 Hz. In practice, the motor may be operated in the 2 –3 Hz range. The
power conversion section of the PWM drive must be capable of supplying the required current at
the low frequency to be usable. Most drives must be derated at low frequencies because they
are not designed for continuous operation there. The operating requirements for current,
frequency, and time must be quantified for the application to select the correct drive rating.
Another requirement for the drive is to be able to operate at low frequency and near its rated
voltage. As noted above, if operated at rated excitation, a 4160 V motor will reach 460 V at 6.6
Hz. This is an unusual combination of operating conditions for a drive designed to operate closer
to 60 Hz at 460 VAC. Many drives cannot be programmed to operate at low frequency and high
voltage. The power converter is not a limitation, but the microprocessor-based control may not
accept this combination of rated voltage and frequency. Possible reasons for not permitting these
settings may include computation limitations in the drive control, shortcomings in the sensor
capability, or lack of planning to operate with this combination of settings.
Drive Control Strategy
The drive control strategy for a medium voltage synchronous motor while spotting is suggested
by the electromechanical inverter spotting equipment. That equipment received no feedback
from the motor, from either the motor terminals or a speed sensor. Employing a similar control
technique with a current design PWM drive generally requires selecting an open-loop Volts/Hz
control strategy. This form of motor control is also known as scalar control and is older than the
latest types of vector control for induction motors.
The fundamental concept of V/Hz motor control is to maintain a constant ratio of voltage to
frequency at the output of the drive. The speed reference at the drive input sets the inverter
output frequency and the output voltage is controlled in proportion to the frequency. Over the
years, enhancements have been made to the V/Hz motor control, such as slip compensation and
voltage boost, to make drives capable of better speed regulation. Slip compensation is not
applicable to spotting a synchronous motor, but voltage boost at low frequencies is useful. The
normal V/Hz ratio can be set manually as a ratio, or can be calculated by the drive control
processor from rated voltage and frequency data put into the drive.
However, at low frequencies, (usually <10 Hz) a V/Hz controlled drive usually has settings that
permit a higher than rated voltage to be applied to the motor. The reason for permitting this is to
overcome resistive voltage drops in the cabling and motor and ensure sufficient current and flux
are available to start a motor and load. The capability to manually adjust the voltage boost at low
speeds is key to setting up the PWM drive to spot a medium voltage synchronous motor.
The ideal control strategy for the solid state spotting drive is an open loop output frequency
command proportional to the speed reference. For best starting of the motor, the output
frequency should ramp from near-zero frequency to the minimum spotting speed. The maximum
output frequency can be limited as required by the process or voltage capability of the drive. The
output voltage of the drive should be adjustable so that sufficient current can be injected into the
motor to make the torque required to rotate the motor and attached load. With these attributes,
the PWM drive has the functionality to replace the electromechanical spotting equipment.
User Interface
The user interface with the PWM spotting drive can be a simple manual pushbutton control for
start and stop. A speed control device such as potentiometer or increase/decrease pushbuttons
can be added where the spotting process benefits from adjusting the speed of the motor within
minimum and maximum limits. If more than one motor is to be spotted, multiple control stations
or a wireless control link can be used to allow the operator to watch the driven equipment at it is
being spotted.
Spotting Multiple Motors
Multiple motors can individually operated from one PWM spotting drive in the same manner as
with older spotting equipment. The PWM drive is designed for operating at low voltage (less than
600 VAC), and so must be positively isolated from the normal medium voltage power supply to
the motor. There are standard medium voltage motor starter arrangements that can perform this
function. Typically, each motor is supplied with a medium voltage starter for normal operation
and a manual switch that connects the motor to the spotting bus. When a motor is to be spotted,
the medium voltage starter is opened and racked out. After locking out the starter, the manual
switch is closed to connect the motor to the spotting bus and the PWM drive. Closing the spotting
bus switch mechanically blocks the medium voltage starter from being closed. No means is
provided to prevent more than one motor at a time from being connected to the spotting bus. In
such a circumstance, the PWM drive will limit its output current to prevent damage to itself, but it
is likely neither motor would operate.
When spotting a synchronous motor, it is necessary that the field supply be providing current to
the rotor winding. This provides the rotor flux required to interact with the stator current to make
torque. A separate field current reference for the spotting operation should be provided. The
field current does not have to be regulated, but the reference should be adjustable so the best
level of rotor excitation can be established during the PWM drive commissioning process. A new
synchronous motor on which a spotting drive is to be used must have a slip ring excitation system
so the rotor current can be on while the motor is at standstill.
A PWM drive is also usable for spotting an induction motor if a drive with adequate current
capacity is available.
Experience: An Innovation Series Low Voltage Drive was supplied to a cement plant in Tulsa,
OK to replace electromechanical spotting equipment. A system diagram is shown in Figure 2.
The Innovation Series Drive was directly substituted for the old spotting equipment. All the
existing medium voltage motor starters and the manual switches for connecting the motors to the
spotting bus were retained. These starters and manual switches already had the necessary
lockout features to preclude connecting the medium voltage bus to the spotting equipment.
4160 V utility bus
Spotting bus
Drive isolation
transformer
Spotting switch
PWM
Medium voltage spotting
motor starter drive
(Multiple
motors)
Output
contactor
1750 HP, 4160 V 1500 HP, 4160 V

450 RPM 450 RPM
synchronous motor synchronous motor
Figure 2. Cement Mill Grinding Motor Power One-line Diagram
Four ball mill motors can be powered from the spotting bus. All the motors normally operate from
the 4160 V bus. Three of the motors are 1500 HP; the other is 1750 HP. The ratings of these
motors are close enough to permit the Innovation Series Drive to be setup with one configuration
suitable for all the motors.
The Innovation Series Drive has a V/Hz motor control pattern that permits essentially direct
control of the frequency output by the speed reference. The drive also has a manual boost
function that allows control of the current over the speed range of the spotting operation.
Only two days were required for drive startup and configuration on this first time application. The
total cycle time for removal of the old equipment, installation, and start up of the new PWM drive
was five days. The rotor field current was set for each motor to give consistent starting. The
performance of the Innovation Series Drive as a spotting drive was very good. The starts were
smooth and controllable. Speed adjustment was available to spot the motor at very slow speeds
or rotate the motors faster to remove grinding media from the mill. The old contactor-based
spotting equipment had aged to the point where contactor arcing was a significant problem, so
badly that some of the operators were concerned for their safety during operation. The new drive
completely eliminated the arcing and safety concerns.
One problem that was solved in the field was an intermittent dc bus over voltage that would occur
occasionally in the drive when it was given a stop command. The rollback of the mill from
operating position to rest position would cause the excited motor to generate sufficient voltage to
push the dc bus into over voltage. The voltages did not reach extreme levels, but the energy was
sufficient to raise the dc bus to levels that caused the drive to trip off. Two viable solutions were
studied. One was adding a dynamic braking resistor and switch to control the dc bus voltage; the
other solution was adding an output contactor to isolate the motor from the drive before the
energy can be pumped back into the dc bus of the drive. Either solution would work, but the
output contactor was chosen as the means to solve the problem in this case.
Conclusion: A low voltage PWM drive with the proper motor control algorithms and adjustability
can be retrofitted in place of old contactor-based spotting equipment. The PWM drive can also
be used for new installations using standard medium voltage lockout provisions. The motor
operation will be smoother and the PWM drive permits adjustment of the spotting speed within
limits. The solid state PWM drive also eliminates mechanical wear associated with contactors in
electromagnetic spotting equipment. The PWM drive eliminates motor and drive train stress
associated with bumping the motor from the power system. With the correct application
engineering and configuration, a PWM drive can provide a much better solution for spotting a
synchronous motor or an induction motor, even though it may have been originally designed only
for induction motor applications.
REPLACEMENT OF A WOUND ROTOR MOTOR WITH AN
ADJUSTABLE SPEED DRIVE FOR A 1400KW KILN EXHAUST GAS
FAN
By:
Roman Menz, Project Engineer, Holcim Group Support Ltd.

Felix Opprecht, Project Engineer, Holcim Zement AG
Abstract: The use of adjustable speed drives for process fan applications in our industry is already well
established. However, a variety of solutions are possible due to different requirements and influences in
a particular application.
The paper describes the project concerning a conversion of a fan drive from fixed speed to variable
speed. The different variants, which have been investigated, are stated and the final solution is shown
in detail. Furthermore, all relevant concerns for the final decision making such as costs, reliability and
efficiency were analyzed.
The energy savings and process advantages by means of applying an adjustable speed drive are
pointed out and a ROI calculation justifies the expenditures.
Introduction
General: Nowadays all new installed process fans in a cement plant are equipped with variable speed
drives. The technology especially in the field of frequency converters has in the recent years undergone
a strong development and is judged to be reliable. For fan applications the system efficiency if operated
with an adjustable speed drive is significantly higher compared to a damper controlled system.
Project
Initial situation: The Untervaz plant situated in Graubünden / Switzerland was founded in 1957. Today
the plant has two rotary kilns with a common daily capacity of 2700 mt.
Initially the kiln exhaust gas fan was equipped with a slip ring motor in conjunction with a liquid starter.
In order to control the gas flow a damper system has been installed. After several years of production
one decided to install a direct current motor mainly because of the higher system efficiency and hence
lower electrical energy cost. Furthermore the installation enabled an automatic control of the process.
After major problems with the DC drive it was decided to install the original slip ring motor at fixed
speed to ensure the availability of the plant.
Data of the existing motor:
Motor type Wound rotor

Shaft power [kW] 1400
Nominal speed [rpm] 1460
Stator voltage [kV] 6
Secondary starter liquid
Cooling Heat exchanger Air/ Water
M
M Kiln exhaust gas fan
M
M
Figure 1 shows the given situation in the plant
Objectives of the project: The project comprised the replacement of the old drive system with a
modern adjustable speed drive.
The following objectives have been set:
• Lower the specific electrical energy consumption by 2.5 kWh/tclinker

• High reliability
• Lowest "Total cost of ownership"
Design data
Operational data: Depending on whether the raw mill is integrated the values vary in a certain range.
3
Gas flow (direct) [m /h] 6600 at 95°C
3
Gas flow (compound) [m /h] 5538 at 165°C
Flap position (direct) [%] 70%
Flap position (compound) [%] 45%
Control of gas flow: The gas flow of the fan was controlled by means of a radial inlet vane. This
damper system is directly attached to the fan inlet and pre-spins the incoming air in the direction of the
wheel rotation. This results in a better efficiency compared to other damper systems. However, in
comparison with an adjustable speed drive the efficiency, especially in lower flow rates, is still low.
100
90
absorbed power (%)
80
70
radial inlet vane
60
50
40
variable speed
30
20
10
0
0 10 20 30 40 50 60 70 80 90 100
flow (%)
Diagram 1 shows the general comparison of flow rate versus power consumption for the two
configurations – radial inlet vane and variable speed -
If the savings are calculated one should not forget to consider the fan efficiency, which might drop if an
adjustable speed drive, is applied.
ROI: To justify the project and to predict the pay back period a ROI (Return On Investment) calculation
was done. Due to the fact that already reliable values were available from the time when the DC-motor
was in operation, there was no need to calculate the power savings again.
There are basically two different points were the fan is operated, one with the raw mill (compound) and
one without it (direct).
Fix speed with damper:
Damper pos. [%] Power [kW] Hours per year Cons. energy [kWh]
(estimated)
70 (compound) 1330 5080 6'756'400
45 (direct) 850 2440 2'074'000
Total 8'830'400
Adjustable speed drive:
Operation Power [kW] Hours per year Cons. energy [kWh]

(estimated)
Compound 1250 5080 6'350'000
Direct 350 2440 854'000
Total 7'204'000
The total amount of energy saved if an adjustable speed drive is applied is 1'626'400 kWh. As shown
below the investment is paid back after 29 months of operation. This figure may vary due to different
hour values or operational points and the local cost structure of the electrical energy. For the calculation
the actual energy costs of 0.042 US$ / kWh was taken.
ROI
ROI time is about 29
140 months
120
100
Costs [%]
80 Investment
60 Savings
40
20
0
0 6 12 18 24 30 36
months months months months months months
Time
Diagram 2 shows the ROI calculation as the base for the investment
Investigated variants: The following systems were investigated and compared:
Medium voltage: With the medium voltage variant it was foreseen to keep the existing slip ring motor
and to install a new frequency converter in medium voltage technique (6kV). The rotor slip - rings would
have been short circuited externally.
The variant with a new medium voltage squirrel cage motor (e.g. 3.3kV) and a frequency converter on
this level was not considered mainly because of cost reasons. Also the variants medium voltage
converter and a step up transformer, DC-drive or slip energy recovery (SER) were no envisaged
solutions.
Low voltage: The low voltage solution comprises a new low voltage squirrel cage motor in conjunction
with a frequency converter. The voltage level was chosen to be 690V to lower the currents and
therefore to reduce the cross section of the cables. From the very beginning it was intended to install a
12-pulse system to comply with the local regulation concerning THDU (Total Harmonic Distortion)
values.
Due to the fact that the plant made already very good experience with water-cooled motors the new one
was also envisaged to be of this type.
Comparison of variants
Technology
MV Drives: The medium voltage frequency converter technology is somewhat new especially on the
6kV level. There are only a few suppliers around the world, which manufacture a real 6kV converter
(see figure 2 below). The major advantage of this technology is that in case of an existing motor
(revamping) the motor might be reused. Beside this the same voltage level for motors (if fixed or
variable speed) can be maintained throughout the plant. However, the age of an existing motor plays an
important role while comparing variants and one should be aware that the motor might soon be in
condition for replacement.
The technology of the systems on the medium voltage level is fairly different from one to another
supplier and shall not be explained in this paper.
Rectifier 18-puls
6kV
M
~
Figure 2 shows the typical electrical diagram for a 18 pulse 6kV drive
LV Drives: The investigated variant with a low voltage drive on the 690V level is getting a quite popular
arrangement in our industry. The advantage compared to a lower voltage (e.g. 400V) is the lowered
current from the transformer onwards and this results in smaller cross sections of the cables. Many
manufacturers have the 690V converter in their program and therefore a certain competition is given. A
12-pulse arrangement was envisaged for the low voltage variant. With a three-winding transformer,
where the two secondary windings are 30° phase shifted, two parallel mounted rectifiers are fed. This
has the advantage that the 5 and 7 harmonic can be lowed or even eliminated.
Reliability / Availability: In terms of reliability it is too theoretical just to compare statistical figures.
However, the calculated value of MTBF (Mean Time Between Failure) is usually higher if a system is
set up with fewer parts.
The availability of equipment is one of the most important things in our industry. Therefore also the
MTTR (Mean Time To Repair) is crucial and may be influenced from the customer as well.
With the medium voltage solution the advantage is that the slip ring motor can remain and in case of a
failure in the converter the motor would be started with the liquid starter and run further at fix speed. A
prerequisite for this is of course that the damper system is still installed and operational.
Efficiency: The efficiencies of the two different systems were also compared (including transformer). It
resulted as follows:
LV- Variant 93,5 %

MV-Variant 92,6 %
Mainly due to the lower efficiency of the existing slip ring motor the MV-Variant has almost a one
percent lower overall efficiency.
Cost Comparison: The cost comparison in chart 1 shows the main items of the different variants. In
the installation costs all relevant expenditures are considered, such as:
• Cables cost (purchase, laying and connection)

• Installation of frequency converter unit
• Installation of new motor (LV-variant)
• Adjustment of coupling and foundation frame (LV-variant)
• Installation of new transformer (LV-variant)
• Civil costs (MV-variant)
If comparing the total expenditure one should be aware that in case of the MV-variant only a frequency
converter is calculated whereas in low voltage an entire new system including the motor was quoted.
Furthermore it should be considered that in case of the MV- variant the motor might be replaced in the
near future which would again cause significant costs.
Cost comparison
100
90
80
70
60 MV Variant
[% ] 50 LV Variant
40
30
20
10
0
r
sf. to tio
n ing tal
an Mo lla ain To
&Tr ta r
ter Ins . &T
er mm
onv Co
C
Chart 1 shows the detailed price comparison between the MV and the LV configuration.
Final solution: After comparing the different variants it was decided to go for the low voltage solution.
The main reason was the lower investment costs and the compact design of the low voltage inverter.
Due to the fact that the distances between transformer, frequency converter and motor are rather small
the installation costs for the low voltage variant are not significantly higher.
Circuit breaker
Three winding transf.
Load-break
switch
Main fuses
Contactor
Rectifier
DC-Bus
Inverter
Motor filter
M
3~
Figure 3 shows the electrical diagram of the applied solution

Frequency converter: Because of the very compact design the frequency converter fitted in an
existing electrical room just beside the motor and hence no civil work was required.
Supply cubicle Inverter unit Motor cubicle
2230mm
3600mm
Figure 4 shows the drives panel layout (depth of cubicle 600mm)
Usually frequency converters of this size do have two smaller units connected in parallel, which
accommodates half the power each. With the purchased system this is not the case anymore. The
rectifiers are still connected in parallel but there is only one inverter bridge installed.
Water cooled motor: Due to the fact that the existing slip ring motor was already water-cooled the new
low voltage motor is also of this type.
The ordered motor does not have a heat exchanger but is surface cooled. The major advantage is that
the heat is discharged very efficiently this means that no fan is needed for the internal air circulation.
This has the advantage that the motor efficiency is approximately 1.5% higher compared to motors with
an air/water heat exchanger.
Protection: For the protection of all the equipment the following measures were taken:
Motor
Danger Protected by
Over current Frequency converter

Over voltage Frequency converter
Over temperature windings and bearings PT100
To fast voltage rise (du/dt) Motor filter
Bearing current Insulated NDE bearing
Condensation Space heater
Frequency converter
Danger Protected by
Over current Itself

Over temperature Itself
Over voltage Itself and screened transformer
Earth fault between transformer and FC External earth fault detection
Worst case scenario: The project team was setting up a certain worst case scenario for the
occurrence of a failure.
If one of the parts failed the first action obviously would be to try to fix it. In case of an irreparable item
the old drive system should be installed again. Prerequisite for this is that all installations such as
cables, damper and starter remain but this is certainly justified if such an important drive is concerned.
Summary: For adjustable speed drives the solution with a 690V low voltage arrangement gains
popularity especially because of economical reasons. But one ought to compare the so-called "Total
cost of ownership", means all costs caused by the system over its lifetime, shall be considered.
Furthermore all different requirements and influences have to be investigated when a system is
purchased such as local support, reliability, efficiency etc.
The purpose of replacing the drive at the Untervaz plant clearly was to lower the specific electrical
energy consumption in conjunction with a high availability of the system. The fact that the investment
will be repaid in less than three years of operation time certainly justifies the expenditure.
In the near future the frequency converter technology will certainly develop especially in the medium
voltage field and one can be curious what the next generation might bring.
References: Slip Energy Recovery Paper, IEEE 1998

SELECTING THE RIGHT PREHEATER FAN AND DRIVE
By:
Victor J. Turnell, P.E.

Process Engineer
Penta Engineering Corporation
1 ABSTRACT
The preheater fan is a key piece of equipment in any cement plant. In many cases, this fan limits clinker
production by limiting the gas flow through the preheater. It consumes about 10 percent of the total
power used at a plant. Large gas flows, high gas temperatures, high static pressures, and the potential for
material abrasion and buildup on the impeller makes the fan design for this application complex.
Selecting the optimum fan and drive requires careful consideration of each application and of all design
options available.
Preheater fans proposed by vendors vary significantly. For example, some vendors propose single width,
single inlet type fans, while others propose double width, double inlet type fans. Some vendors propose
radial blade fans while others propose backward curved or airfoil fans. The degree of wear protection
differs significantly from one vendor to another.
This paper discusses various topics that are important when selecting a fan and drive for this application.
Topics discussed are: the process requirements and conditions, the options available for fans and drives,
the advantages and disadvantages of fans and drives available, and the impacts of each fan and drive on
capital and operating costs.
2 INTRODUCTION
The tendency today is to build cement plants with larger production capacities to minimize capital and
operating costs. Such plants are typically designed to use preheater exhaust gases in a vertical roller mill
for raw material drying. Preheater fans in these applications are designed for large gas volumes at high
gas temperatures and static pressures. Several variables must be considered while selecting a fan for
this application: wear caused by the dust loading in the preheater gases, material buildups on the fan
impeller, drive starting characteristics, drive power factor, and drive efficiency.
3 OPERATING CONDITIONS
Typically, preheater exhaust gases pass through a gas-conditioning tower before entering the preheater
fan. From there the gases go to either the roller mill or directly to the main dust collector before being
vented to the atmosphere. The preheater fan creates the induced draft that pulls gases through the
preheater and conditioning tower. During times when the raw mill is not operating, water is sprayed in the
conditioning tower to cool the gases to less than 200°C (400°F). During times when the roller mill is
operating, little or no water is sprayed in the conditioning tower. At this point, the gas temperature
entering the preheater fan ranges from 300 to 425°C (570 to 800 °F). Preheater fans are required to
generate static pressures in the range of 500 to 1000 mm WG (20 to 40 in WG) with large gas flows. The
3 3
dust concentration entrained in the preheater exit gases ranges from 20 to 60 g/m (9 to 18 grains/ft ) of
gas and will impact the recommended fan design.
4 GENERAL DESIGN PRACTICE
4.1 Process Considerations
The first step in selecting a preheater fan is to determine the expected operating conditions such as the
gas flow rates, temperatures, static pressures, and dust loading. These operating conditions are usually
determined using mass and energy balances, empirical equations, and assumptions based on past
experiences.
To ensure that the preheater fan does not limit the pyroprocess system to a level below the design
capacity, it is prudent to design the fan with a safety factor. One of the more commonly applied safety
2
factors is 115 percent of the expected gas flow at 132 percent (1.15 ) of the expected static pressure
requirement.
4.2 Mechanical Considerations
The design must minimize maintenance requirements caused by abrasion and material buildup on the
impeller. Several technical papers are available that provide potential solutions to buildup problems. One
paper, Cement Plant Preheater Fan Build-up Control, (Gutzwiller 1991) published in the IEEE
Transactions On Industry Applications provided recommendations which are summarized below.
• Reduce the gas operating temperature to below 250°C (482°F).
• Design the impeller for the smoothest possible flow lines to reduce the impact energy as dust
particles strike the fan blades. Smooth flow lines are best achieved by using backward curve and
airfoil shaped impellers.
• Design the fan to minimize the gas and dust particle velocity at the inlet of the fan impeller. This can
be achieved by selecting double inlet fans instead of single inlet fans, and selecting larger diameter
and lower speed fans instead of smaller diameter and higher speed fans. If possible, the peripheral
speed at the impeller inlet opening should be limited to 76 m/s (15,000 ft/min), and the inlet velocity
should not be in excess of 38 m/s (7500 ft/min).
• Oversize the impeller shaft, bearings, and supports to reduce the sensitivity to unbalance.
In general, the gas temperature entering the preheater fan is dictated by the roller mill requirements for
raw material drying; therefore, lowering the gas temperature to below 250°C may not be possible.
4.3 Example of Fan Selection Process Specifications
The examples used throughout this paper are based on a preheater fan designed for a 4500 mt/d clinker
pyroprocessing system with a modern 5-stage preheater and an inline vertical roller mill for raw grinding.
The preheater fan is assumed to be located after the preheater conditioning tower.
Table 1 contains both expected and designed gas flow conditions for the example of the preheater fan
with the raw mill operating and not operating. The fan design conditions are based on 115 percent of the
expected gas mass flow at 132 percent of the expected static pressure requirement. In all cases, the
fans are designed with double widths, double inlets, and a maximum speed of 900 rpm.
Table 1
Preheater Fan Operating Parameters
Metric Units U.S. Customary Units
Description Units Expected Design Units Expected Design
Raw Mill Operating
3
Inlet gas flow m /h 700,000 830,000 acfm 410,000 488,000
Inlet static pressure mm WG -790 -1040 in WG -31 -41
Temperature °C 310 310 °F 590 590
3 3
Inlet Density kg/m 0.61 0.59 lb/ft 0.038 0.037
Power at 100% fan kW 1508 2357 hp 2021 3160
and drive efficiency
Raw Mill Not
Operating
3
Inlet gas flow m /h 630,000 740,000 acfm 370,000 435,000
Inlet static pressure mm WG -790 -1040 in WG -31 -41
Temperature °C 210 210 °F 410 410
3 3
Inlet density kg/m 0.72 0.70 lb/ft 0.045 0.044
Power at 100% fan kW 1346 2104 hp 1800 2820
and drive efficiency
5 FANS
5.1 Fan Impeller Design Options
Six basic types of centrifugal fans are available which are distinguished by their impeller configurations:
radial blade, radial tip, backward incline, backward curved, forward curved and airfoil. Forward curved
impeller designs are not considered for the preheater fan because of its potential for material buildup;
therefore, it will not be considered further in this paper. Figure 1 shows the different impeller designs.
Each impeller type is generally more suitable than the other types for a particular application. There are
examples of preheater fans with each of the above impeller types that operate successfully.
Radial blade and radial tip fans are used primarily for gases heavily laden with dust applications because
they are simple to repair in the field (AMCA publication 201-90). These fans are considered the best in
resisting material buildup; however, in preheater fan applications backward curved and airfoil type
impellers are recommended. (Gutzwiller 1991)
Backward inclined and backward curved fans are used where airfoil blade designs are not acceptable
because of a corrosive or erosive environment. (AMCA publication 201-90)
Airfoil fans are used primarily for clean gas applications. (AMCA publication 201-90) The problem
commonly associated with airfoil fans in dust-entrained gas applications is the potential for material to
enter the inside of the airfoil and create vibration problems. This potential problem can be minimized in
the design, fabrication, and maintenance of the fan. It is important to note that several cement plants do
have airfoil fans operating successfully in preheater applications.
Figure 1 Fan Impeller Types
Table 2 provides some general information about the different fan impeller types.
Table 2
Fan Impeller Types
Description Radial Radial Tip Backward Backward Airfoil
Blade Incline Curved
Efficiency Lowest Lower Good Higher Highest
Peak efficiency 72% 75% 80% 84% 90%
5.2 Example of Fans
The Table 3 below compares some general data and capital costs for fans with different impeller types.
The data and costs are based on operating conditions of the example in Section 4.3. Following the table,
Graph 1 compares the fan performance of each fan type for the design condition with the raw mill
operating. At different flow and pressure combinations, the relative performance may be different. For
example, the efficiency of the airfoil fan may be higher than shown in Table 3.
Table 3
Fan General Design Data and Costs
Description Units Radial Blade/ Backward Backward Airfoil
Radial Tip Incline Curved
Maximum efficiency % 73 76 80 83
Wheel diameter mm 3,510 3,460 3,650 3,640
2
Wheel inertia kg-m 10,800 7,500 8,700 10,700
Fan cost US$ 332,000 329,000 331,000 343,000
The difference in costs between the various types of fan impellers is not substantial; therefore, the
recommended fan impeller types for typical preheater fan applications are backward curved or airfoil type
fans.
Graph 1, Comparison of Fan Performance
1,400
1,200
Work
Point
Pressure (mm wg)
1,000
800
Backward
600 Inclined
Backward
400 Curved
Radial
200 Tip
Airfoil
0 Point
0 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000
3
Volume (m /h)
6 FLOW CONTROL
6.1 Flow Control Options
Two options are available for flow control. One option uses a fixed speed fan with a damper in the
system while the other adjusts the fan speed. Figure 2 shows the different dampers and their locations in
the fan system.
Dampers provide control of gas flow by adjusting the system’s static pressure loss. Parallel-blade inlet
box dampers, variable inlet vanes, and radial inlet dampers provide an advantage as compared to the
other damper types available. As the parallel-blade inlet box damper, variable inlet vane, and radial inlet
dampers are closed, they pre-spin the gases entering the fan resulting in a more efficient fan performance
curve. Variable inlet vanes and radial inlet dampers are not used in preheater fan applications because of
the potential for material buildup on them.
Outlet dampers cost about the same as inlet dampers; however, outlet dampers are not as efficient. For
this reason, outlet dampers are usually not considered for preheater fan applications.
Opposed-blade inlet box damper, parallel-blade inlet box damper, and variable speed control are the flow
control options typically considered for preheater fans.
Figure 2 Fan Damper Control Options (Air Movement And Control Association, Publication 201-90, Figure
10-1)
6.2 Example of Different Flow Control Methods
Table 4 provides the static efficiencies for the different operating conditions described in the example of
Section 5.2.
Table 4
Fan Static Efficiency (%)
Description Opposed Blade Inlet Parallel Blade Inlet Variable Speed
Box Damper Box Damper Control
Airfoil
Raw mill on-design 82.7 82.7 82.7
Raw mill on-expected 69.1 72.3 83.1
Raw mill off-design 58.3 63.6 83.0
Raw mill off-expected 47.7 53.6 83.1
Backward Curved
Backward Incline
Radial Blade/Radial Tip
The most energy efficient method is variable speed control, followed by the parallel-blade inlet box
damper, and then by the opposed-blade inlet box damper.
There is little or no cost difference between the parallel and opposed blade dampers; therefore, if
selecting a fixed speed fan, the recommended damper for typical preheater fan applications is a parallel
blade damper.
7 FAN DRIVES
7.1 Fan Drive Options
Several options exist for fixed speed fans: AC squirrel cage induction motors, synchronous motors,
wound rotor motors, and synchronous induction motors. The options for variable speed fans include the
following: AC squirrel cage induction motor with a fluid coupling, DC motor and drive, and AC variable
frequency drive with an AC squirrel cage induction motor.
When selecting a preheater fan drive system, several factors must be considered such as the available
starting torque and inertia, the effect of the motor current inrush on the plant’s electrical system, the drive
efficiency, and the capital and maintenance costs.
Designing an AC squirrel cage induction motor capable of starting large inertia loads typical of preheater
fans can be difficult or impossible. Starting currents for these motors are between six and seven times the
normal full load current. This inrush of current causes severe heating of the stator and rotor. Excessive
heat can damage the motor windings, which would result in premature motor failure. Care must be taken
not to exceed the thermal limit of the motor. An AC squirrel cage induction motor can be customized by
adjusting the rotor resistance to allow starting with loads greater than NEMA moment of inertia minimums.
The resistance of the rotor of a synchronous motor, unlike the rotor of an AC squirrel cage induction
motor, cannot be changed significantly to allow the starting of large inertia loads. This limitation often
prevents a standard synchronous motor from being used on a large fan application.
A wound rotor motor is an AC induction motor with a wound rotor that is connected to a variable
resistance. In the past, wound rotors had received criticism because of the banks of wound wire resistors
that were used as rotor resistance. These banks were subject to frequent failures and choppy starts as
the different values of resistance were sequenced during start. A modern wound rotor motor uses a liquid
rheostat, which provides continuous instead of stepped rotor resistance changes. The liquid rheostat is
also more reliable than stepped rotor resistances.
During the startup of a wound rotor motor, an electrolytic salt solution is pumped out of a tank that
contains the motor rotor electrodes. The low level of electrolytic salt solution between the rotor
electrodes creates a high resistance that results in high starting torque for low starting current. In the
starting sequence of the motor, electrolytic salt solution is returned to the tank, raising the level of
electrolytic salt solution between the rotor electrodes and reducing resistance. Once the starting
sequence is completed, the rotor electrodes are shorted-out by a shorting contactor that reduces rotor
resistance to a minimum.
This motor has the best starting and operating characteristics from a rotor resistance standpoint. During
startup, the resistance is high resulting in lower current flow per unit torque supplied. During normal
operating conditions, the resistance is low resulting in better efficiency than during the startup.
Recently, the synchronous induction motor hybrid has been reintroduced as a fan drive option. This
motor starts as an AC induction wound rotor motor connected to an external rotor resistance. When
approaching synchronous speed, the external resistance is removed and a DC field is supplied to the
rotor causing the motor to operate as a synchronous motor. This motor can start large inertia loads and
then operate as a synchronous motor.
AC squirrel cage induction motors with fluid couplings were a common choice for fans in the past;
however, today they are not considered for variable speed fans because other methods to vary fan speed
exist that are more efficient at transferring power.
DC motors and drives were also common choice for fans in the past; however, they are difficult to
implement on large motor power applications. The commutating action of the armature limits the
armature voltage and results in a large current rating requirement. The losses associated with this large
current make the DC drive a poor choice for today’s preheater fans. The DC drive has an efficiency profile
that is about the same as an AC variable frequency drive (VFD). The DC drive requires more
maintenance because of the brushes and commutator.
Table 5 provides typical operating limitations of motors.
Table 5
Typical Operation Limits
Description AC Squirrel Synchronous Wound Rotor Synchronous AC Variable
Cage Motor Motor Induction Frequency
Induction Motor Drive and
Motor Motor
Starting current 600% 650% 150 to 200% 150 to 200% See the note
(percent of full below
load current)
Power factor 0.80 to 0.85 1.00 to 0.80 0.73 to 0.78 1.00 to 0.80 0.95 lag
lag lead lag lead
Note: Starting current is programmable and typically limited to the capacity of drive and motor.
Table 6 provides benefits of using a variable speed drive compared to using a fixed speed drive (ABB
Fan Group North America, November 2000).
Table 6
Benefits Of Using A Variable Speed Drive Versus a Fixed Speed Drive
Feature Benefit
High efficiency Reduced energy cost
Soft starting Reduced stress on fan, motor, and coupling, resulting in
extended equipment lifetime
Reduced impact on the electrical supply network (low voltage
drop due to starting)
Elimination of problems with starting high inertia loads
Unlimited number of starts per hour
Lower audible noise Improved working environment
Low fan erosion at lower speeds Extended fan lifetime
Reduced maintenance costs
Ability to select a fan with maximum Flexibility in design
efficiency at speeds other than
synchronous speed
7.2 Example Of Drives
For purposes of comparing the power transmission efficiency and capital cost of the different drives, the
fan from the example of Section 6.2 was selected.
Table 7 provides the expected efficiencies for different drive systems based on the fan systems in the
example of Section 6.2. This table does not consider power factor correction and the losses associated
with that correction. Table 8 provides an estimated cost of the different drives. This table does not
consider harmonic filtering that may be needed with the AC variable frequency drive.
Table 7
Typical Motor And Drive Efficiency At Full Load
Description Units Efficiency
Fixed Speed Motor efficiencies
AC squirrel cage Induction % 96.6 (see the note below)
Wound Rotor % 96.3
Synchronous Induction (1.0 power factor) % 96.5
AC Variable Frequency Drive and AC Squirrel
Cage Induction Motor
Transformer efficiency % 99.0
Drive efficiency % 98.1
Motor efficiency % 96.6
Effective efficiency % 93.8
Note: The efficiency of the AC squirrel cage induction motor depends on the rotor resistance of the
selected motor.
There is little difference in efficiency between three different fixed speed drives; therefore, the
recommended drive should be the one that has the lowest cost with the ability to start the fan while
satisfying the plant’s technical needs.
Table 8
Drive Systems Costs (US$)
Description AC Squirrel Synchronous Wound Rotor Synchronous AC Variable
Cage Motor Motor Induction Frequency
Induction Motor Drive and
Motor Motor
Transformer cost - - - - 78,000
Drive cost - - - - 235,000
Motor cost 160,000 247,000 240,000 435,000 145,000
Total cost 160,000 247,000 240,000 435,000 458,000
8 VARIABLE SPEED VERSUS FIXED SPEED FANS
The following is a comparison of a variable speed drive and a fixed speed drive based on the fan systems
used in the example of Section 7.2.
8.1 Capital Costs
Table 9 provides the capital cost comparison between installing a fixed speed fan and damper versus
installing a variable speed fan with a variable frequency drive.
Table 9
Variable Speed and Fixed Speed Fans
Capital Costs (US$)
Description Fixed Speed Fan and Damper Variable Speed Fan with a
Variable Frequency Drive
Fan 331,000 331,000
Damper 23,000 not required
Transformer cost not required 78,000
Drive cost not required 235,000
Motor cost 160,000 145,000
Drive installation cost 25,000 60,000
Total cost 539,000 849,000
Cost difference 310,000
8.2 Power Consumption
Table 10 provides the power consumption comparison between installing a fixed speed fan and damper
versus installing a variable speed fan with a variable frequency drive.
Table 10
Power Consumption
Fixed Speed Fan and Damper Versus
Variable Speed Fan with a Variable Frequency Drive
Description Units Raw mill Raw mill Raw mill Raw mill
on-design on- off-design off-
expected expected
Power at 100% fan and drive kW 2357 1508 2104 1346
efficiency
Fixed speed fan with
damper
Total input power (kW) kW 2823 2346 2853 2573
Specific power consumption kWh/mt 15.0 12.4 15.1 13.7
Variable speed fan with
variable frequency drive
Total input power kW 2907 1855 2606 1677
Specific power consumption kWh/mt 15.4 9.8 13.8 8.9
Difference between the
power consumption of the
two fan controls
Total input power kW -84 491 247 896
Specific power consumption kWh/mt -0.4 2.6 1.3 4.8
8.3 Fan Operating Scenarios
Two operating scenarios will be considered. Scenario One assumes the preheater fan operates under the
“expected” conditions as shown in Table 10. This operating scenario occurs when the preheater fan is
not the bottleneck of the pyroprocess system. Scenario Two assumes the preheater fan operates under
the “design” condition as shown in Table 10. This operating scenario occurs when plant personnel
operate their kilns at the maximum possible capacity, which usually results in the preheater fan being the
bottleneck of the pyroprocess system.
For the cost analysis of the two scenarios, the following assumptions were made:
• the kiln operates 8000 hours per year, equivalent to about 91 percent uptime
• the raw mill operates 7200 hours per year, equivalent to about 90 percent of the time that the
preheater operates
• the average power cost is US$0.05 per kWh consumed
In Scenario One, the power savings obtained by installing a variable speed instead of the fixed speed fan
is 491 kW when the raw mill is operating, and 896 kW when the raw mill is not operating. This results in a
potential annual savings in power costs of US$213,000, US$177,000 per year with the raw mill operating,
and US$36,000 per year with the raw mill not operating. The payback time for the additional capital cost
of installing the variable speed fan instead of the fixed speed fan is 1.5 years.
In Scenario Two, the power savings obtained by installing a variable speed instead of the fixed speed fan
is -84 kW when the raw mill is operating, and 247 kW when the raw mill is not operating. This results in a
potential annual loss in power costs of US$20,000, US$30,000 per year with the raw mill operating, and -
US$10,000 per year when with the raw mill not operating. The additional capital cost of the variable
speed fan versus the fixed speed fan is not justified in this operating scenario.
9 CONCLUSIONS
The difference in costs between the various types of fan impellers is not substantial; therefore, the
recommended fan impeller types for typical preheater fan applications are backward curved or airfoil type
fans.
There is little or no cost difference between the parallel and opposed blade dampers; therefore, if
selecting a fixed speed fan, the recommended damper for typical preheater fan applications is a parallel
blade damper.
There is little efficiency difference between the various fixed speed drives; therefore, the recommended
drive should be the one with the lowest cost and the ability to start the fan while satisfying the plant’s
other technical needs.
When purchasing a preheater fan, consideration should be given to the “expected” operation of the fan
and the safety factor applied to the fan design. As shown above, if the fan operates near its design
capacity, a variable speed-drive is not financially justifiable; however, if the fan operates significantly
below its design capacity, then a variable speed-drive is financially justifiable.
ACKNOWLEDGEMENTS
Mr. Jason R. Stoklosa at ABB Fan Group, Inc for providing information used to develop this paper.
ABB Motor Group for providing information used to develop this paper.
REFERENCES
Les Gutzwiller, Chief Engineer Robinson Industries, Inc.; H. Daniel Banyay, Manager Research Robinson
Industries, Inc.; Sidney M. Cohen, Director of Research and Development Fuller Company, “Cement
Plant Preheater Fan Build-up Control”, Published in the IEEE Transactions on Industry Applications, vol.
27, No. 3 May/June 1991.
ABB Fan Group North America, “SCR Retrofits Provide Opportunity for Variable Speed Fans”, published
in Power Engineering, November 2000, pages 82-86.
Air Movement And Control Association, Publication 201-90, Figure 10-1

ALTERNATIVE DRIVE SOLUTIONS FOR THE CEMENT INDUSTRY
By:
Ashok B. Amin,Marketing Manager, Hagglunds Drives Inc
Abstract; "In the Cement industry many critical applications like, Kiln drives, Apron and Belt feeders
drives, Feeder breaker drives, Reclaimers and Excavators drives, Inching drives, Grinder drives etc have
seen an increased popularity of a direct Hydraulic drive as an alternate to Electro Mechanical drive. This
paper explains advantages and comparison between various alternative drives based on performance,
efficiency, reliability, flexibility towards productivity"
Introduction; The Cement industry has always placed a high demand on equipment reliability, productivity,
efficiency and performance. In today’s competitive market place these demands are of even more interest.
Electro - mechanical drives have in the past been the traditional drive in this industry when variable speed
has been called for. Today the Hydraulic Direct Drive solution can in many cases be an alternative to the
electro-mechanical drives. The hydraulic drives has proved to be a real alternative on applications where
Load sharing is essential and in applications where high starting torque is needed and flexible productivity
at various capacity needed without sacrificing efficiency using modular flexibility. Hydraulic drives are used
in many applications with increased popularity and reduced maintenance. Hydraulic direct drives utilise
standard components and are very easy to install on the driven machine shaft without the need of gear
reducers or foundation.
Basic concept; A direct drive hydraulic drive system consists of
§ a low speed high torque hydraulic motor
§ an electric motor
§ an oil reservoir
§ a pump
§ and a control system , see fig.1.
In most systems, the equipment except the hydraulic motor is located in a sound insulated cabinet. The
hydraulic motor and the pump are connected to each other by flexible hoses and/or steel pipes.
Fig. 1: Direct
Hydraulic Drive
The hydraulic motor is a hydraulically balanced radial piston cam curve unit with a mechanical efficiency
approaching that of a roller bearing resulting in outstanding torque efficiency. A torque arm is utilised to
take out the reaction force whilst eliminating any other undesirable forces on motor bearings.
The standard power unit is designed as a closed loop hydraulic system for continuous high integrity drive
applications. The unit consists of a standard AC induction electric motor and a hydraulic pump. The electric
motor starts up in an unloaded condition and runs at constant speed, driving the pump. The pump is a
variable displacement unit, the displacement of which determines the flow and therefore the final speed of
the drive.
A simple electric signal can be utilised to control pump displacement and, therefore, the speed. This
method of controlling the speed means that, say in a feeder application, when feed is not required, it is not
necessary to shut down the main electric motor, simply the signal to the pump is zeroed and the feeder
stops. This feature allows an unlimited number of stops and starts of the drive. All electrical connections in
the power unit are made to a junction box and the complete power unit is housed in a sound insulated and
weather proofed cabinet ready for positioning on site.
Since the hydraulic system is able to maintain pressure at any flow, the hydraulic motor will maintain a
torque level which is unaffected by speed. Therefore, by allowing the system pressure to rise above the
normal operating level, very high torque levels can be sustained without time restriction. Furthermore due
to the fact that the motor is running at low speed, the moment of inertia in the drive is very low which
means that the drive is unaffected by shock loads, starts/stops and reversals.
For the higher power applications such as a kiln drive multiple pumps / electrical motors can easily be
installed in order to reach the maximum required speed.
Some of the unique features of Hydraulic Direct Drive are Low moment of inertia. The Fig. 2a shows a
hydraulic direct drive has approximately 1/500th of the moment of inertia on an equivalent Electro-
mechanical drive.
Fig. 2a. Comparison of Moments of Inertia

El
Fig. 2b: Natural Load sharing
concept, when multiple motors
Pump
and pumps are used
El
Pump
Comparison of variable speed - high torque drives

The different types of variable speed drives that exist on the market have different output characteristics of
torque and speed. It is important to know the differences when selecting the drive especially if a wide
speed range is called for, when the torque levels can vary or when a high breakaway torque is needed.
In Fig. 3a and 3b is a short comparison of drive alternatives between the DC, Variable frequency AC and
hydraulic drives.
Fig. 3a: DC/Variable

frequency drive
configuration
Fig. 3b: Hydraulic

drive configuration
Hydraulic Direct Drives;
A Hydraulic drive can operate continuously throughout the entire speed range up to 200 - 300 % of the
normal operating torque from zero to full speed, determined by the displacement of the selected pump and
the installed electrical power, see fig.4.
The drive can operate above the installed power rating with the same limitation as any fixed speed electric
motor with equivalent service factor.
This drive solution has a distinct advantage during lengthy acceleration cycles or during overload
conditions. A pre-set pressure limiting function, which destrokes the pump, effectively avoids loads that
exceed the maximum torque requirement.
The machine is also well protected from shock loads due to the low kinetic energy of the low speed motor.
Due to this low moment of inertia, unlimited starts/stops and reversals are accepted without any
restrictions. The standard drive can operate in both driving and braking mode, forward and reverse (4 -
quadrant drive) without any additional or special accessories.
Hydraulic Drive performance
Torque Pressure
bar
Peak 420
Hydraulic pump Hydraulic
Max 350 Installed rated power
motor
Driven shaft power
El
Average 175
Normal
operating
range
Max Oil flow/Motor speed
Figure 4. Hydraulic drives characteristics
Hydraulic drives with gear reducers;

The drives performance of a medium/high speed hydraulic motor in combination with a gear reducer is the
same as for a direct hydraulic drive with some exceptions. By using a gear reducer the kinetic energy is
higher compared to a direct drive. That means that the gear reducer and the driven machine have to be
dimensioned for the shock loads that can occur. The mechanical and overall efficiency are also lower than
for a direct drive especially if the gear reducer has to be over dimensioned due to shock loads or high
output power.
Electro-mechanical DC-drives;
A direct current (DC) drive consists of a Converter transformer, an AC/DC converter, a DC electric motor
and eventually a high-speed coupling. For a low speed high torque drive a gear reducer eventually with a
low speed coupling have to be added.
A DC - motor can operate continuously at constant torque up to 100% rated torque. Speeds above rated
are also possible but the motor is now limited to operate at constant power, see fig.3.
A DC-motor can operate intermittently above these limitations but only with time restrictions due to over-
heating of the windings. The shaft driven fan is insufficient to control the temperature at lower speeds and
usually forced cooling is introduced to improve lower speed performance.
A DC-motor is sensitive to harsh environments. In many applications an enclosed design is necessary and
forced cooling with filtered air is needed. Conductive carbon dust is generated internally from the brushes
and the drive must be stopped for cleaning and changing the worn out brushes on a regular basis.
Torque
(%)
Max
Output power gearbox

Fig. 5a: DC-drive
Installed rated power
characteristics
Intermittent
100
Forced cooling
Continuous
Speed %
0 100 Max
Electro-mechanical variable frequency drives:

A variable frequency alternating current (VFAC) drive consists of a Converter transformer, a frequency
converter, an electric AC induction motor and eventually a high-speed coupling. For a low speed high
torque application, a gear reducer with eventually a low speed coupling have to be added.
The VFAC drive can operate continuously at lower speeds up to 40 % of rated torque for standard drives
and up to 70 % of rated torque for drives with forced cooling. At synchronous speed the continuous
operating torque is approx. 90 % of rated torque and available within +/- 10 % of the nominal frequency,
see fig. 4.
The drive can intermittently operate up to 150 % of the rated torque for a few minutes and at starting
operate at 180 % of rated torque for a few seconds.
A VFAC selection must be based on its continuous torque capacity over the entire speed range. This is a
most important consideration due to limitations, especially at lower speeds.
Torque (%)
180
Peak Max output power gearbox
150
Intermittent Installed rated power
Rated torque %
Forced cooling
40
Continuous
Frequency /
0 50 Max Speed (Hz)
Fig. 5d: Variable Speed AC Drive performance (PWM)
Torque
200%
Peak
Inverter 3 sec.
150%
Gearbox Installed rated power
1min/
10min (Electrical motor shaft)
El.motor
100%
Driven shaft power
60%
Normal operating
range
50 Hz Motor speed
1500 rpm
Fig. 5e: Variable Speed AC-Drive performance (DTC)

Applications in the Cement industry;
Below you can find some features by using hydraulic direct drives on some of the applications in the
cement industry.
Apron and belt feeders;

A. Starting torque capacity
The hydraulic drive has the ability to provide full shaft torque without time restriction. This is an important
feature in applications, i.e. when the feeder is required to start from a full hopper condition. The ability of
the hydraulic drive to continuous overload torque without limitation means undesirable stalls of the feeder
are avoided.
Fig. 6a: TADS-200

hydraulic drives on
belt feeders
B. Shock load protection
A significant feature is the shock load protection afforded by the direct drive solution. A hydraulic direct
drive has approximately 1/500th of the inertia of an equivalent high-speed drive with gear reducer. This
feature is significant when considering the amount of down time experienced on feeders due to chain
related problems.
C. Start / stop frequency
The power unit’s prime mover is a standard AC induction motor, which runs at synchronous speed driving
the hydraulic pump. The hydraulic motor mounted on the feeder shaft turns only when the demand signal is
received by the pump and flow is generated. Therefore there is no requirements to start/stop the electric
motor, only to zero the signal to the pump. This important feature means that the high current and the
associated effects when starting electric motors are negated and infinite number of starts/stops and even
reversals are available.
Fig. 6b: Apron feeder
driven by 2 x MB 2400
developing approx.
1.7 M Nm in torque.
D. Low speed high torque capacity
The hydraulic motor is mounted directly onto the feeder drive shaft. There is no need for bedplates,
couplings or gear reducers between the motor and the driven shaft. As there are no high-speed elements
which need reducing in speed, the hydraulic motor can develop its exceptionally high torque from zero
speed. This allows exceptional controllability of the feeder speed in all materials conditions.
Fig. 6c: TADS-400,

Self Contained Torque
arm Drive system.
Fig. 6d: TADS-400,
Self Contained Torque
arm Drive system.
Crushers / Sizers;
A. Starting torque capacity

The hydraulic motor is installed directly to the crusher drive shaft without any gear reducers, couplings or
bedplate arrangements. The motors high torque capacity is a desirable feature in crushing operations,
particularly when starting up a machine in a stalled condition. If the crusher has to be started up in a stalled
condition, the high torque capability ensures the crusher to be started up.
The motor is bi - directional, and, therefore, in such situations full torque can be utilized in both forward and
reversed directions.
B. Shock load protection

The operating pressure of the hydraulic motor is directly proportional to the loads to be overcome, i.e.
material to be crushed.
By limiting the pump compensator to a certain pressure the actual torque transmitted is limited to any given
setting, i.e. crushing force required. This ability to limited forces in the machine along with the low moment
of inertia of the drive means any shock loading, i.e. stalling the crusher rolls, are absorbed by the hydraulic
system and not by the machine components. This also means a reduction of wear of the machine
components.
Fig. 6e. Mobile crusher
Two rollers are driven by 2 x MB 800 hydraulic motors.
Infeed plate feeder driven by 1 x MB 800
Discharge belt conveyor driven by 1 x MA 141
C. Variable speed
The variable displacement pump delivers fluid to the drive motor on receipt of a command signal.
The command signal can be via electric or hydraulic joystick control, and facilitates fully variable speed
control of the crusher rolls in both forward and reverse directions.
This feature can offer tremendous benefits in certain operating conditions, e.g. the variable speed capability
can be used to achieve a metering effect, whilst processing the material through a crusher. A particular
advantage on dump fed machines.
D. Automatic operation / reversal

Once the hydraulic system is set it will operate in an automatic mode. A reversal system can be introduced
which will automatically reverse the crusher rolls should a stall condition occur. If a feeder system is used
the feeder can be stopped and started automatically during this situation.
Cement Kilns;
A. Power (Load) sharing

The features of the hydraulic drive ensure that the drives always experience a common load. As the
hydraulic motors are part of a common system the power is always equally shared within the system. This
power sharing eliminates any problem with oscillations sometimes experienced when trying to balance
electro - mechanical drives.
The hydraulic drives remain balanced irrespective of load or speed requirements.
See figure 7a and 7b
Pinion/
roller
Cement kiln
Hydraulic
Power
unit
Fig. 7a: Dual pinion 4 Hydraulic Motors sharing the load on a Hydraulic
Direct Drive for kiln
Power unit Emergency
P Tank P P Tank P P
Roller or pinion
Fig. 7b, Multiple Electric Motors / Pumps share the load
B. Direct drive
Hydraulic direct drives are used on the recently developed 2 - base rotary kiln, which is supported by self -
aligning friction rollers. The main advantages are that the motors can be installed directly to the friction
roller shaft without any gearbox or foundation. This a very important feature because this type of drive is
unaffected by the kilns axis deformation (radial eccentricity and wobbling of the shell).
C. Starting torque capacity
The hydraulic drive has the ability to provide full shaft torque without time restriction. High starting torque’s
and overload torques can be designed into the system without over sizing installed electrical power, this is
due to the control of the drive system allowing a slow starting speed and controlled ramping up to running
speed. This is an important feature in kiln applications particularly if the kiln has to start up in a loaded
condition.
Fig. 8: Polro type kiln

2 x Friction rollers, 4 x
MB-800, 1000 kW
D. Low speed high torque capability
The hydraulic motor is mounted directly onto the pinion drive shaft. As there are no high-speed elements,
which need reducing in speed, the hydraulic motor can develop its exceptionally high torque from zero
speed, indeed it can maintain full shaft torque from stall to full speed condition.
Fig. 9: Cement kiln

driven by 2 x MB 566
motors trough a pinion
and a girth gear.
Fig. 10: Dual Pinion
and Girth Gear Drive,
2 x MB-1150, 400 kW
(536 Hp.)
E. Noise level
In the standard power units, the electrical motor and pump are installed in a sound insulated cabinet. With
this type of design the noise level is reduced compared to an electro - mechanical drive. The power unit
can also be installed wherever convenient for the user.
Fig. 11: Dual Pinion

and Girth Gear Drive,
2 x MB-1600
Fig. 12: Dual
Pinion and Girth
Gear Drive, 2 x
MB-1600
Belt conveyors;
A. Power capacity
In the past it has been more or less impossible to use hydraulic drives on belt conveyors because of the
limited sizes of motors available on the market. Today the situation is different. The biggest motor has a
capacity of up to 1500 kW intermittently or more than 1000 kW continuously for a single motor.
B. Power sharing
The features of the hydraulic drive ensure that the drives always experience a common load. As the
hydraulic motors are part of a common system the power is always equally shared within the system. This
power sharing eliminates any problem with oscillations sometimes experienced when trying to balance
electro - mechanical drives.
The hydraulic drives remain balanced irrespective of load or speed requirements.
C. Starting torque capacity

An important feature is the high starting torque capacity that can be utilized in an unlimited time when
starting up the belt conveyor. The high torque capacity can be achieved without over sizing the electrical
motor.
Fig. 13: Belt conveyor,
8500 tph, driven by 1
x MB 1150
Key Factors Comparison between VFAC / DC and DHD;

Investment cost;
• For lower range of output torque below 10 000 FT-LBf and installed power below 25 Hp VFAC drives
are substantially economical over DHD.
• For mid range of output Torque values between 10 000 FT-LBf and 100 000 FT-LBf investment cost for
DHD is higher by 10 to 25 %, however due to added features like flexibility in production rate
optimization and very low long term maintenance cost and reliability makes an attractive payback.
• For higher range of output torque over 100 000 FT-LBf and power range up to 2000 Hp cost of DHD is
lower compared to VFAC or DC drives due to the fact, that the gear reducers become customized and
bulky. Furthermore long lead items, costly down term and higher maintenance cost compared to
modular standard components for DHD.
There are several indirect savings like lower starting current service required due to cascading of smaller
E-motors, welding of the drum, inching during maintenance, space savings etc.
Drive Efficiency; As the application demands tougher duty cycle and higher safety factors for gear
reducers and the operating range lower than rated torque and speed, the reducer efficiency drops. VFAC
also suffers higher losses at lower speeds, DHD offers more energy savings, provided the system is
Engineered for the specific application. During past decade DHD have achieved overall efficiency of the
drive in the range of 75 to 79 %. Today with more advancements and improvements in Hydraulic motors,
Pumps and Synthetic hydraulic fluids overall efficiency of DHD is approaching over 82% level.
Maintenance; Direct Hydraulic drives are getting more and more user friendly and more and more
applications are utilizing the features and benefits of DHD in increasing the growing demand of productivity.
The experience from many of these applications have allowed DHD professionals to optimize new and
improved ways to provide long -term reliability and reduced maintenance. DHD does require some skills
and training, which is well within the capacity of a typical process plants.
Erection and Commissioning; DHD offers some unique features during commissioning. The hydraulic
motors do not require any foundation. Also the power unit can be placed at a convenient flexible location.
You can also use DHD for rotating Feeder or Kiln during assembly and welding etc.
Size of Equipments; DHD us es modular multiple Electric motors / Pumps and Hydraulic motors, these
items are standardized readily available produced in reasonable volume compared to a large electric motor
and Gear reducer used for VFAC drives. Even larger Hydraulic motors are made from many common parts
from several different motor sizes.
Energy Savings; Energy savings varies depending on operating parameters and application situation.
When operating over wider range of production capacity DHD offers significant savings due to modul ar
concept, which allows varying number of pump/ motor combination operated at a given time. Also in case
of future expansion to increase production capacity, DHD allows an economical upgrade by changing some
components compared to replacing large Gear reducer or Electric motor for VFAC drive.
Environmental Aspects; DHD does require handling Hydraulic fluid, which is an environmental hazard.
The design of DHD is such that it is a sealed system and uses mostly o-ring type fittings and connections,
which are made for industrial quality and for the process industrial. Mostly routine commonsense type
maintenance practice is needed. Also oil companies are coming up with environmentally friendly high
performance synthetic oils.
Conclusion;
The number of hydraulic drives in the Cement industry is increasing steadily. In applications where variable
speed is needed, high starting torque is required, power sharing is essential or where shock loads occur
frequently, hydraulic direct drives should seriously be considered. The close loop hydraulic system has also
proven to be insensitive for harsh environments which means a low maintenance cost for the drive.
Direct Hydraulic drives have made Technological advancements and are getting more user friendly with
increasing training efforts by industries.
These types of drives are used in other process industries like Mining & Materials Handling, Pulp and
Paper, Chemical, Rubber and Plastics, Recycling and Steel. All these industries want to have the highest
possible reliability and a low life cycle cost. Using a Hydraulic Gearless Drive (DHD) can fulfil these
demands.
Environmental, Maintenance and
Safety
Working Group
Chair
Christopher Macey
RESCO Products Inc.
Vice Chair
Christopher C. Rayner
Bechtel Corporation
Members
Neal W. Biege, Sr. Robert J. Sagmeister
Fuller Bulk Handling Suwannee American Cement Company
Ed Buehler Satish H. Sheth

Buehler Consulting, Inc. RMC Pacific Materials
Cary O. Cohrs William H. Siemering

Florida Rock Industries, Inc. Ash Grove Cement Company
Ann Dougherty Diana L. Inglis

Portland Cement Association Fuller Bulk Handling
Larry D. McConnell John R. Sommers

BHA Group Inc. Holcim Inc.
Allan S. Moore
Retired (Tilbury Cement Limited)
NOx FORMATION IN ROTARY KILNS PRODUCING CEMENT CLINKER
APPLICABLE NOx CONTROL TECHNIQUES
AND
COST EFFECTIVENESS OF THESE CONTROL TECHNIQUES
By:
Gerald L Young
Vice President
Process and Environmental Engineering
Penta Engineering Corporation
ABSTRACT
The Portland Cement Association commissioned a study of information available related to NOx
formation in rotary kilns producing cement clinker, potential NOx control techniques, NOx control
techniques applicable to the cement manufacturing industry, and the cost effectiveness of the applicable
controls. This paper is a summary of the PCA report and other articles and information concerning NOx
emissions and control related to rotary kilns in the cement manufacturing industry. Descriptions of
predominate NOx formation mechanisms and control techniques are provided. An assessment of the
technical feasibility, the NOx reduction potential, the potential effect on other pollutants, and a range of
cost effectiveness for control techniques applicable to the cement manufacturing industry are provided.
1.0 INTRODUCTION
In 1997, the Portland Cement Association (PCA) initiated a project to review literature available on
formation of oxides of nitrogen (NOx), potential NOx control techniques, NOx control techniques
applicable to rotary kilns in the cement manufacturing industry, and the cost effectiveness of these control
techniques. The purpose of that project was to provide a technical and cost review of various NOx
reduction techniques. This paper is a summary of the findings on NOx formation and control techniques
provided to the PCA.
2.0 NOx FORMATION IN KILN SYSTEMS
Three NOx formation mechanisms – thermal, fuel, and feed NOx – are typically cited as contributing to
NOx emissions from kilns. In kiln exhaust gases, more than 90% of NOx is NO, with NO2 generally
making up the remainder (Gardeik, et al, 1984; Nielsen and Jepsen, 1990).
2.1 THERMAL NOx
Most NOx formed in the high-temperature environment of the main combustion zone (burning zone) of a
kiln is “thermal NOx”. Thermal NOx is formed by oxidation of atmospheric nitrogen at high temperatures.
The literature cites threshold temperatures for the formation of thermal NOx ranging from about 1200 –
1600°C (2200 – 2900°F) (Gardeik, et al, 1984; Nielsen and Jepsen, 1990; Rother, 1997). Since the flame
temperature in a kiln is significantly above these temperatures, considerable amounts of thermal NO are
generated in the burning zone. The thermal reaction between oxygen and nitrogen to form NO is named
the Zeldovic reaction (Rother, 1997) and is simplified as follows:
O + N2 → NO + N
N + O2 → NO + O
NO formation increases exponentially as temperature increases, and increases as excess oxygen (O2)
increases. Above 1400°C (2550°F), small changes in temperature produce large changes in
concentrations of NO at a given oxygen concentration (PCA, 1999). Gas temperatures in kiln burning
zones are significantly above clinker material temperatures, which must reach about 1450°C (2650°F) to
form some clinker compounds (PCA, 1999).
Typically, the amount of O2 in kiln exit gases is held in the range of 1.0 – 2.0% for optimum kiln
performance and product quality. This is equivalent to about 5 – 10% excess air in the combustion zone,
an amount sufficient to enhance NOx formation.
Conditions in the burning zone of a cement kiln favor formation of thermal NOx. Secondary combustion
zones such as, calciners, preheater riser ducts and mid-kiln firing in long wet or dry kilns, operate with
temperatures of less than 1200°C (2200°F), which are not sufficiently hot for significant thermal NOx
formation (formation of fuel and feed NOx are more prevalent) (PCA, 1999).
2.2 FUEL NOx
NOx can also result from oxidation of nitrogen compounds in fuel (fuel NOx). Fuel nitrogen is only
partially converted into NO during combustion and this reaction occurs throughout the temperature range
of the combustion process relatively independent of temperature (Gardeik et al, 1984; Rother, 1997).
Since the combustion temperature in a calciner or in the kiln riser duct of a preheater kiln is significantly
below 1200°C (2200°F), NO formed in the secondary combustion zone is primarily fuel NOx (Nielsen and
Jepsen, 1990).
NOx formation from nitrogen contained in fuel occurs through a complex series of reactions whereby N2 is
oxidized into NO and NO is reduced to N2 by reacting with intermediate combustion products. Nitrogen
bound in fuel (coal, coke, heavy oil) is partially converted to intermediate combustion products, such as
HCN and NH2 radicals, in the early stages of combustion. In fuel rich flames, the deficiency of oxidizing
radicals increases the rate of reduction of NO to N2 while in oxygen rich flames, the formation of NO from
N2 predominates (Gardeik et al, 1984; Rother, 1997).
According to Nielsen and Jepsen (1990), NO formation in a secondary combustion zone will depend on:
• Nitrogen concentration in the fuel

• Oxygen concentration in the combustion zone
• Initial NO concentration in the combustion gas
• Volatile concentration in the (solid) fuel
• Temperature in the secondary combustion zone
Higher concentrations of volatiles in the fuel tends to reduce the percentage of fuel nitrogen converted
into NO. An increase in the temperature of the secondary combustion zone will reduce net NO formation.
At temperatures between about 815°C and 1100°C (1500°F and 2000°F), the following reactions may
take place (Nielsen and Jepsen, 1990):
“N” + O → NO (1)
“N” + NO → N2 + O (2)
Where: “N” = nitrogen compounds originating in the fuel.
The rate of reaction 2 increases more rapidly than the rate of reaction 1 as temperature increases.
Therefore, higher temperatures (between 815°C and 1100°C) may reduce NOx emissions from
secondary combustion zones (Nielsen and Jepsen, 1990). It is interesting to note that, as opposed to the
burning zone where higher temperatures result in much higher NOx formation, higher temperatures (up to
1100°C) in the calciner may actually reduce NOx emissions when a fuel containing nitrogen is used.
2.3 FEED NOx
NOx emissions also result from the oxidation of nitrogen compounds in the raw material feed to the kiln
(feed NOx). Table 1 lists data on the range of nitrogen concentrations in various kiln feeds (Gartner,
1983) and the potential contribution of feed NOx to total NOx emissions.
TABLE 1
POTENTIAL FEED NOx
Amount of Nitrogen in Potential Feed NOx

the Kiln Feed (Pounds NOx/
(parts per million) ton of Clinker)
20-1000 0.2 – 10
Nitrogen in raw materials is converted to NO or NO2 at temperatures of 300 – 800°C (570 – 1470°F). As
much as 50% of feed nitrogen may be converted to NOx if raw materials are heated slowly; with rapid
heating, the conversion rate is much lower (Gartner, 1983).
2.4 SUMMARY OF NOx FORMATION
The following conclusions can be qualitatively stated for rotary kiln systems (Young and von Seebach,
1999):
• Formation of thermal NOx in the burning zone is the major contributor to NOx emissions from the kiln.
• Fuel NOx is the major contributor in the secondary combustion zone of calciner and riser duct fired
preheater kilns.
• Feed NOx is usually a minor contributor to the total NOx generated in rotary kiln systems.
3.0 KILN OPERATIONS AND NOx EMISSIONS VARIABILITY
Proper clinker formation reactions require close and constant process control to account for process
variables. The dynamic nature of kiln operations can result in highly variable NOx formation.
Consequently, each kiln will tend to have unique NOx emission characteristics.
An important aspect of understanding NOx emissions variability from kilns is an understanding of the
inherent variability in the cement manufacturing process (PCA, 1999). Numerous physical and chemical
reactions occur as raw materials pass through a pyroprocessing system in the cement manufacturing
industry. Two areas are of interest because of their potential effects on kiln operation and hence on NOx
formation: the calcining zone and the burning or sintering zone (PCA, 1999).
As Young and von Seebach (1999) explained, calcination begins at a temperature of about 815°C
(1500°F), and is substantially completed at about 955°C (1750°F). In addition to these minimum
temperatures, an energy input of about 415 kcal per kg (750 Btu per pound) of CaCO3 is also required.
Further heating is required to raise the temperature of CaO (and the other feed materials), which then
chemically combines with other minerals in the raw materials to form clinker compounds. Formation of
cement clinker compounds in the main burning zone of a kiln system is a net exothermic reaction (as
opposed to the endothermic reactions preceding the burning zone) and liberates about 100 kcal/kg (180
Btu/pound of clinker). Depending on the chemical compounds being formed, temperatures ranging from
1200°C to about 1540°C (2200°F to about 2800°F) are required to initiate and sustain these reactions. If
the raw materials in the kiln are not heated to these temperatures, clinker compounds do not form. If the
clinker minerals do not form, the exothermic heat of reaction is no longer available to the kiln system and
the chemical processes stop. More importantly, the clinker produced will be of unacceptable quality for
the manufacture of portland cement (PCA, 1999).
3.1 OPERATING VARIABILITY OF KILNS
The following Section 3.1 has been taken from PCA publication R&D Serial No. 2227 (PCA, 1999) and
Young and von Seebach (1999).
A hypothetical calciner kiln that produces 100 mtph of clinker and consumes an average of 830 kcal/kg
(3.0 million Btu per ton) of clinker with 60% of the fuel fired in the calciner and 40% in the kiln burning
zone can serve as an example. Note that the fuel consumption of 830 kcal/kg is an average and the coal
is assumed to have a heating value of 6945 kcal/kg. The actual heat input required at any given time will
vary with:
• Kiln feed chemical variations, mass flow into the kiln, fineness, and moisture concentration;
• System heat losses;
• Ambient air infiltration;
• Heat recovery efficiency from the clinker cooler; and
• Other system parameters.
If, because of normal variations in any of the above parameters, the energy requirements of the system
were to increase slightly (e.g., from 830 kcal/kg to 845 kcal/kg or 15 kcal/kg) the required fuel firing rate to
the burning zone would increase from 4780 kg per hour to 4867 kg per hour (10540 pounds/hr to 10730
pounds/hr). Material in the kiln system will slowly begin to cool if the fuel rate is not increased. If the
temperature in the burning zone of the kiln dropped to a point where the clinker-forming reactions
stopped, the exothermic heat from clinker mineral formation would stop. The energy liberated by these
reactions, about 100-kcal/kg clinker, would no longer be available to the process and the burning zone
would lose 10,000,000 kcal per hour (36 mmBtu/hr) of heat energy. Therefore, instead of being deficient
by 87 kg/hr (190 pounds/hr) of coal, the system would be deficient by 1527 kg/hr (3360 pounds/hr) of coal
(87 kg/hr + 1440 kg/hr to replace the 10,000,000 kcal/hr in exothermic heat that was lost). To place this
in perspective, the heat input to the burning zone from fuel is about 33,200,000 kcal/hr and the loss of the
exothermic heat of reaction would have an impact similar to reducing the fuel rate to the burning zone by
30%. This would result in the burning zone temperature dropping precipitously.
To reinitiate the clinker-forming reactions, the kiln operator would have to make up the loss of the
exothermic heat of reaction by adding, at a minimum, an additional 1527 kg/hr (3360 pounds/hr) of coal or
reducing the production rate while maintaining the fuel rate. When the burning zone temperature returns
to the proper operating level, clinker-forming reactions would begin once again, and the exothermic heat
would once again be available. The kiln burning zone would then have 1440 kg/hr (3175 pounds/hr) too
much coal, which would cause the burning zone temperature to quickly increase.
This example is only given to demonstrate the sensitivity of a pyroprocessing system and the temperature
changes that may occur. As shown earlier, even small changes in kiln burning zone temperatures result
in rather large variations in NOx emissions. Consequently, it is not unusual for a kiln system to have a
large range in NOx emissions during normal operations.
3.2 NOx EMISSIONS VARIABILITY OF KILNS
Data are available from a tightly controlled, well-documented research project concerning NOx emissions
from a long dry process kiln in southern California (Young et al, 1986). These data demonstrated the
highly variable nature of NOx emissions. When these NOx data were subjected to statistical analyses, it
was determined that for the estimated mean emissions rate to be statistically accurate within 10 percent
of the true emissions rate, 30.7 days of emissions data were required. Similar tests conducted on a
preheater kiln determined that about 14 days of emissions data were required for similar statistical
accuracy (Young and von Seebach, 1999).
During the testing described above on the long dry kiln, 47 kiln operating variables were collected. These
data were subjected to rigorous statistical analyses in an attempt to determine which operating variables
affected NOx formation. The researchers found that 83% of the variation in NOx was explained by
burning zone temperatures. No other variable (of 14 variables examined) contributed more than an
additional 4% to the regression sum of squares (Croom et al, 1987).
NOx emissions from cement kilns are highly variable both on an hour-to-hour basis and on a day-to-day
basis. Figure 2 is a graph of 240 data points representing 24 hours of 6-minute average pounds of NOx
per short ton of clinker values. The hourly average emission rates range from a low of 2.24 to a high of
11.72 pounds NOx per ton of clinker. Figure 3 is a graph of 270 daily averages of the 6-minute averages.
These values range from less than 2.0 to more than 14 pounds of NOx per ton of clinker. Finally, Figure
4 is a histogram of the frequency of observations of the daily NOx averages. Most of the values fall
between 4.0 and 7.5 pounds NOx per ton of clinker; however, some daily averages are less than 2.0
while a few are greater than 13.5 pounds NOx per ton of clinker. Figures 2, 3, and 4 were taken from PCA
R&D 2227 (PCA, 1999).
As reported in PCA R&D 2227 (PCA, 1999), short-term testing for NOx will not provide a statistically
meaningful estimate to be made of the monthly or yearly mass emissions of NOx from a facility. To
develop a meaningful estimate of NOx emissions from a cement kiln with a statistical measure of
confidence, a continuous emissions monitor (CEM) must be used, and data must be collected for several
months, or longer if multiple fuels are used. Once data are collected, they must be subjected to rigorous
statistical analyses to establish the mean and the variability of the emissions. Without such analyses,
NOx emissions data are of little to no value in establishing emission factors or limits (PCA, 1999).
NOx emissions rates are also site and kiln-specific. NOx emissions from two apparently similar kilns may
not be similar. The causes are not fully understood; however, one cause may be due to the raw materials
used. Since cement kilns process large quantities of naturally occurring raw materials, minor chemical
variations in raw materials can result in different NOx emissions characteristics from cement kilns (e.g.,
less than a 1% change in SiO2 in the raw materials can significantly affect cement kiln operation and
therefore affect NOx emissions). Other causes for NOx emissions rate differences may result from
different types or classes of cement products being produced. Chemical variations between these
different products can influence cement kiln operating parameters and thus NOx emissions. Short-term
process transients such as kiln feed rates and fuel quality also affect NOx emissions. Such factors can
influence the applicability and costs of applying NOx controls.
4.0 NOx CONTROL IN ROTARY KILNS PRODUCING CEMENT CLINKER
There are several steps that should be taken when an effort is made to reduce NOx emissions. First, the
emissions rate and kiln operations must be adequately defined; second, process changes should be
developed and implemented; and finally, secondary control techniques can be investigated, if necessary.
4.1 DEFINE THE PROBLEM
The first step in developing a coherent, cost-effective plan to control NOx emissions is to accurately
characterize kiln operations, parameters that affect kiln operations, NOx emissions, and parameters
affecting NOx emissions. As previously stated, NOx emissions are dependent on the process type, are
site-specific, and are highly variable during normal operations on a given kiln system. Before spending
potentially large sums of money to reduce NOx emissions, it is usually beneficial to spend some money
up front to define the raw material and fuel characteristics, the pyroprocessing system, and the mean and
variability of the NOx emissions.
Once sufficient data are collected, they must be analyzed statistically to determine the mean NOx
emission rate and the variability of the emission rate. Additionally, the feed, fuel, and operating
characteristics of the pyroprocessing system must be defined. Once all this information is available, it
must be investigated to determine what, if any, process and/or operational changes can be made to
reduce NOx emissions.
4.2 NOx CONTROL TECHNIQUES
NOx reduction techniques for cement kilns are defined in two broad categories, process controls and
post-combustion (secondary) controls. Process controls, including combustion modifications, rely on
reducing or inhibiting the formation of NOx in the manufacturing process. Post-combustion controls rely
on treatment of the flue gases to remove NOx that has already been produced (PCA, 1999).
As stated in PCA R&D 2227:
“Combustion modifications are process controls that change how fuel is used in the
manufacturing process. Combustion modifications have achieved substantial NOx
reductions in several industrial categories. Combustion modifications for cement kilns
include low NOx burners, secondary combustion (mid-kiln firing, riser duct firing, and
precalciners), and staged combustion (low NOx precalciners). “ (PCA, 1999)
Pounds of NOx per ton of clinker for a single day
(240 data sets - 6 minute average values, hourly averages as indicated)
11.72
14.00 11.31
12.00
9.57
7.64 8.43
10.00
7.37 7.16
6.67 6.54
8.00 5.50
5.25 6.02 6.37
5.15 5.82 5.56
5.34
6.00
Pounds of NOx per ton of clinker

4.00
6.31 5.06
4.89
5.40
3.40
2.00
3.26
2.24
0.00
0 50 100 150 200
Number of 6 minute data periods
FIGURE 2
270 Daily Averages of 6-minute NOx (as NO 2 ) per Ton Clinker Data
16.00
14.00
12.00
10.00
8.00
6.00
Pounds of NOx per Ton off Clinker

4.00
2.00
6.27 5.43 5.90 6.45 5.73 5.54 6.70 6.26 6.76

-
0 50 100 150 200 250
Day (with 30 day averages indicated )
FIGURE 3
Frequency of Observations of Pounds of NOx (as NO 2 ) per Ton Clinker
(270 daily averages of 6-minute data points)
45
40
35
30
25
20
15
Frequency of Observations
10
0
.5 .0 .5 .0 .5 .0 .5 .0 .5 .0 .5 .0 .5 .0 .5 .0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
- 2 - 3 - 3 - 4 - 4 - 5 - 5 - 6 - 6 - 7 - 7 - 8 - 8 - 9 - 9 - 10 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1 - 1
0 5 0 5 0 5 0 5 0 5 0 5 0 5 0
2. 2. 3. 3. 4. 4. 5. 5. 6. 6. 7. 7. 8. 8. 9. 9.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
1 1 1 1 1 1 1 1 1
Pounds of NOx per Ton of Clinker
FIGURE 4
Industrial processes that require well-defined temperature profiles because of process requirements are
not generally as successful as boilers or gas turbines in reducing NOx with combustion modifications.
The chemical reactions that produce cement clinker require material temperatures of 1200 – 1540°C
(2200 – 2800°F). To achieve these material temperatures, gas temperatures must be substantially higher
(typically higher than 1760°C or 3200°F). An oxidizing atmosphere is also required in the main
combustion zone of a cement kiln because of product quality considerations. If the burning zone
temperature is too high, equipment damage will occur. If the burning zone temperature is too low, the
cement kiln will no longer produce a salable product (PCA, 1999).
4.2.1 Process Controls
Kiln feed chemical characteristics and kiln feed chemical uniformity can influence NOx emissions. Kiln
specific fuel consumption will also affect NOx emissions. An important statement regarding NOx control
is made in the introductory section of Chapter 5 of the EPA NOx Alternative Control Techniques
Document for the Cement Manufacturing Industry:
“For any given type of kiln, the amount of NOx formed is directly related to the amount of
energy consumed in the cement-making process. Thus, measures that improve the
energy efficiency of this process should reduce NOx emissions in terms of lb of NOx/ton
of product.”
The cement industry is always looking for ways to improve fuel efficiency and productivity, which in return
may reduce NOx. Based on extensive industry knowledge of how process modifications influence
cement kiln operation and NOx formation, the following is a list of some of the more common process
modifications that have been made to reduce NOx emissions from cement kilns (PCA, 1999):
• Changing fuel, such as from natural gas firing to coal firing.
• Improving kiln feed chemical uniformity, which results in more stable kiln operations:
- Modifications to quarry operations

- Raw material blending facilities
- On-line analytical control systems for raw material proportioning
- Kiln feed blending systems
• Modifications to improve thermal efficiency, such as:
- Reducing excess air infiltration.

- Increase efficiency of cyclones in preheater kilns.
- Reducing the amount of moisture in slurry (wet process kilns only).
- Revised kiln chain systems in long wet or long dry kilns.
- Modifying or replacing clinker coolers to improve heat recovery and cooler efficiency.
- Initiating operator training programs.
- Returning as much cement kiln dust (CKD) as possible to the kiln system (without adversely
affecting product quality)
• Installing or upgrading kiln system sensors and instrumentation.
• Installing or upgrading computer control of kiln systems to stabilize kiln operation and avoid burning
zone temperature variations.
Young and von Seebach (1999) said:
“Based on industry data and experience, control of burning zone temperature is the
primary process control method applicable to thermal NOx formation. Process
modifications that allow better control of the kiln burning zone temperature will result in
reducing thermal NOx formation and minimizing NOx spikes.
Although the heat requirements for producing clinker depend to some extent on the
chemical composition of the kiln feed, altering the chemical composition of kiln feed is
usually not a viable technique for NOx reduction. However, tight control of the kiln feed
chemical uniformity is a major contributor to stable kiln operations, resulting in overall
NOx reductions of 10% - 15%. Poor kiln feed chemical uniformity results in overfiring the
kiln, which produces higher NOx emissions.”
Switching from natural gas to coal will reduce NOx emissions in most cases (Young et al, 1986; Radian,
1994); however, most kilns that burned natural gas have already switched to coal. Although switching to
a lower nitrogen fuel in a calciner may reduce NOx, the nitrogen content of the fuel burned in the burning
zone has little or no effect on NOx generation (Miller and Egelov, 1980). Some researchers have found
no relationship between fuel nitrogen content and the NOx emissions from a cement kiln. Miller and
Egelov (1980) said:
“However, no relationship has been found between the fuel nitrogen content and the
corresponding NOx level. This can be attributed in part to the fact that flames in cement
kilns are very hot. In an ‘axial’ flame typical of cement kilns, the core is reducing, while
the hotter flame envelope is under oxidizing conditions. … Thus, the combustion gas, as
it leaves the hottest portion of the flame, will have approached the equilibrium NO
concentration regardless of the original source of the nitrogen. Thus, no influence of fuel
nitrogen on NOx from cement kilns is to be expected.”
In addition to process modifications, the following modifications to fuel firing systems have been found to
reduce NOx (Young and von Seebach, 1999):
• Conversion from direct-fired coal systems to indirect systems to reduce primary air quantity.
• Installation of a new burner (“low NOx” burner) to aid in combustion control and flame shaping. This,
however, is difficult with direct fired systems.
• Reduction in the amount of excess air (oxygen) used for combustion.
• Addition of fuel to the riser duct of preheater kilns to transfer some of the thermal load from the kiln
burning zone to the lower temperature environment of the preheater.
• Installation of mid-kiln firing of fuel in long kilns to transfer some of the thermal load from the kiln
burning zone to the lower temperature environment of the calcining zone.
• Use of staged combustion or “low – NOx calciners”.
Some combustion devices, such as utility boilers can use flue gas recirculation (FGR) or inject
water/steam into the main flame to reduce flame temperatures and NOx formation. FGR is not a viable
method of reducing NOx in a cement kiln burning zone. FGR’s effectiveness relies on cooling the flame
and generating an oxygen deficient (reducing) atmosphere for combustion to reduce NOx formation,
conditions that are not compatible with cement kiln operation. High flame temperature and an oxidizing
atmosphere are process requirements to produce a quality clinker product (PCA, 1999).
Conversion from a direct coal firing system to an indirect firing system is generally necessary before a low
NOx burner can be installed. Based on industry experience, converting a direct coal fired kiln to an
indirect coal fired kiln with a low NOx burner may result in NOx reductions ranging from 0% to about 20%
(Kupper, 1990; Schraemli, 1990). NOx reductions attributable to low NOx burners are usually difficult to
quantify because typically, a number of other process modifications are also made when a kiln coal firing
system is converted to indirect with a low NOx burner. Additionally, rigorous statistical testing of the
“before” and “after” conversion emissions is not normally done (Young and von Seebach, 1990).
Dr. Schraemli (1990) cautioned that incorrect use of multi-channel (low NOx) burners can increase NOx
rather than reduce NOx. He also cautioned that kilns that use blends of fuels tend to have higher NOx
emissions because it is more difficult to achieve the proper flame shape for reduced NOx formation.
Hansen (1985) found that less excess air in the kiln combustion gases results in less NO formation at the
same burning zone temperature. NOx reductions of 15% were reported when operating at 1.6% O2
instead of 2.0% O2 (measured at the exhaust gas exit). In order to operate at these low levels of excess
air, precise process control of the kiln atmosphere must be used (Hansen, 1985).
When cement kilns operate at low levels of kiln exit gas oxygen concentration (e.g., 1.5%), small
variations in the mass flow rate of fuel (dry basis) can create excessive CO emissions because of a
momentary lack of sufficient oxygen for complete combustion. These small variations can result from
limitations on the accuracy of the fuel feeding devices, from changes in the moisture content of the fuel,
and from other transient process conditions. Additionally, the exhaust gases require several seconds to
flow from the main combustion zone to the exit of the kiln. The continuous gas analyzers must sample
the gases, transfer the gases to the analyzers, and the analyzers then provide the analyses. If corrective
action is required, either the fuel rate must be reduced or the induced draft fan flow rate increased. The
response time of such a system may not be sufficient to avoid high CO emissions (e.g., 1.5%). High CO
emissions can result in fires or explosions in the downstream ductwork or air pollution control device.
Secondary combustion zones have proven effective at reducing NOx emissions in cement kilns. In long
kilns, secondary combustion can be partly accomplished by mid-kiln injection of fuel. In such systems,
typically less than 20% of the kiln fuel requirements are burned in the mid-kiln location. NOx emissions
may be reduced because less fuel is burned in the high temperature environment of the burning zone.
Localized reducing conditions may also be produced as the fuel burns, which inhibits NOx formation and
may destroy some of the NOx produced in the burning zone. Conversely, oxygen for mid-kiln combustion
must be drawn through the high temperature environment of the burning zone, which creates a strong
oxidizing environment in the burning zone. Two competing conditions are created: higher oxygen levels
tend to produce higher NOx formation in the burning zone, but because of flame cooling by high excess
air, NOx formation may also be reduced in the burning zone. Therefore, the influence on NOx formation
of mid-kiln injection of fuel can not readily be predicted (Young and von Seebach, 1990). Recent
developments in “mixing air technology” (the subject of another paper today) permit true staged
combustion for NOx control in long kilns with mid-kiln injection of fuel or in a preheater with riser fuel
burning.
The use of fuel in the riser duct of preheater kilns is a NOx reduction technique. As gas temperatures are
maintained at 870 – 980°C (1600 – 1800°F) in the combustion zone of preheater riser ducts and in
calciner vessels, the formation of thermal NOx is negligible and fuel NOx will predominate. However,
NOx emissions will not always be reduced by burning fuel in the riser duct of a preheater kiln (Young and
von Seebach, 1990).
Calciner kilns employ a secondary combustion zone that operates at a much lower temperature than the
burning zone. In a calciner kiln, typically 60% of the fuel is burned in the calciner with the combustion air
coming directly from the clinker cooler. In general, NOx emissions for these kilns are less than from long
wet kilns, long dry kilns, or preheater kiln systems because 60% of the total fuel requirement is burned
under lower temperature conditions where negligible amounts of thermal NOx are formed (Young and von
Seebach, 1990).
Young and von Seebach (1999) also noted that calciner kilns have the potential for staged combustion as
a NOx control technique. All major equipment suppliers offer “low-NOx” calciner designs. Fuel is injected
and burned in a sub-stoichiometric O2 environment to create a strongly reducing atmosphere (relatively
high concentrations of CO). This reducing atmosphere inhibits the formation of fuel NOx and destroys a
portion of the NOx formed in the kiln burning zone. Additional tertiary combustion air is added later to
complete combustion of the fuel.
Several calciner kilns in the U.S. have recently been retrofitted with these “low-NOx” calciners, and
several new kiln lines have been installed using low-NOx calciner technology. Operating experience
indicates a noticeable reduction in NOx per ton of clinker. Industry reports place the NOx reduction
potential of staged combustion with a low-NOx calciner at 30% - 40% when compared to a conventional
calciner kiln system (Bauer, 1990; Dusome, 1994; Bertschinger, 1997).
4.2.2 Secondary NOx Control Techniques
All process control techniques outlined above must be implemented to improve specific fuel consumption
and reduce kiln operating variability before other NOx control techniques are considered. Kiln operation
must be as efficient, stable, and uniform as possible before secondary control techniques are considered.
There are two major classifications of secondary controls: selective non-catalytic reduction (SNCR) and
selective catalytic reduction (SCR). Both techniques are similar in that they use ammonia (or urea) to
chemically react with and destroy NOx contained in flue gases. The operating temperature for SNCR is
typically in the range of 870 to 1040°C (1600 – 1900°F). SCR uses a catalyst that allows the NOx
reduction reactions to occur at lower temperatures, for example, 200 – 425°C (400 – 800°F) (PCA, 1999).
4.2.2.1 Selective Noncatalytic Reduction (SNCR)
Noncatalytic reduction of NO with NH3 was developed by Exxon Research & Engineering Co., USA. The
reagent, typically NH3 or urea, is injected into the kiln system at a location with an appropriate
temperature window of 870 – 1100°C (1600 – 2000°F). Temperature is critical because no catalyst is
used. At higher temperatures, the reagents will form additional NOx; at lower temperatures, the reactions
proceed slowly and substantial amounts of unreacted ammonia will escape.
Other researchers have found that the most effective temperature range for NO decomposition by SNCR
is between about 900 and 1000°C (1650 and 1832°F). Under optimum conditions about 1 mole of NH3 is
required for 1 mole of NO. However, the amount of NH3 is critically dependent on the reaction
temperature. About 0.8 moles of NO is reduced with 1 mole of NH3 at 970°C (1778°F). At higher and
lower temperatures, less NO is reduced per mole of NH3.
With SNCR there is the potential for escape of ammonia, referred to as ammonia slip, in which stack exit
gases contain unreacted ammonia in the amount of 10 ppm or greater. An NH3 excess of more than
100% has little additional effect on NO reduction but ammonia slip increases sharply with increasing
molar ratios of NH3 to NO. In theory, ammonia slip can be minimized but in practice this proves to be
very difficult because temperatures and gas velocities are often inhomogeneous. In pilot plant testing,
molar ratios of 0.5 have produced an ammonia slip of as much as 5% (Scheuer, 1990).
Preheater and calciner kilns operate with kiln gas exit temperatures in a range of 900 – 1100°C (1650 -
2000°F), and SNCR systems have been used on some preheater kilns in Europe. For wet and long dry
kilns, these temperatures exist midway through the kiln. Access to this area is possible only through
ports in the kiln shell as used in mid-kiln firing or with scoops used to return cement kiln dust. Ammonia
must be added continuously in a fixed molar ratio to NOx in order to be effective and to minimize
ammonia slip. Therefore, SNCR is not technically feasible at this time on long wet process or long dry
process kilns.
In summary it can be stated that ammonia slip increases rapidly when the molar ratio of NH3 to NO is
above 1. Unless the molar ratio is held at a low level, the amount of ammonia slip will be significant.
Ammonia slip will cause a detached plume and can increase opacity of the stack gases.
4.2.2.2 Selective Catalytic Reduction (SCR)
SCR uses ammonia, in the presence of a catalyst, to selectively reduce NOx emissions from exhaust
gases. SCR has been used on gas turbines and internal combustion engines in the US (e.g., catalytic
converters on autos). Typically, anhydrous ammonia, usually diluted with air or steam is injected into hot
flue gases. The flue gases are then passed through a catalyst bed where NOx is reduced to N2 gas and
water. A number of different materials have been used as catalysts but titanium dioxide (TiO2) and
vanadium pentoxide (V2O5) are commonly used because they are more resistant to poisoning by SO2.
The optimum temperature for SCR depends on the catalyst but is usually between 300 and 450°C (570 –
840°F) (PCA, 1999).
SCR has been used successfully on “clean” exhaust gases from processes that use either natural gas or
gasoline as a fuel. Attempting to use SCR on exhaust gases from a cement kiln will present a wide range
of problems to overcome. Particulate loading in a kiln gas stream is extraordinarily high. Any kiln using
an ESP for particulate removal may have to install a new fabric filter before installing an SCR process.
The particulate loading from an ESP outage would quickly deactivate a catalyst bed. Kiln exhaust gases
also contain relatively high levels of compounds that can poison or blind a catalyst; such as, alkali salts
(e.g., NaCl, KCl, Na2SO4, K2SO4, etc.) SO2, lime, and many other compounds not found in a “clean”
combustion gas (PCA, 1999).
Exit gas temperatures from dust collectors on wet kilns are in the range of 120 – 175°C (250 – 350°F) and
for long dry kilns are in the range of 175 – 290°C (350 – 550°F). Exit gases for dust collectors in
preheater kilns that use in-line raw mills for grinding and drying raw materials are 95 – 120°C (200 –
250°F). These gases would have to be reheated to employ SCR. The cost of reheating the kiln gas from
these low temperatures would be substantial. The environmental impact and counterproductive nature in
terms of energy required would appear to lend logic to the fact that there are no SCR units currently in
operation on cement kilns. SCR for NOx reduction in cement kilns presents many problems that must be
resolved before it is technically feasible (PCA, 1999).
5.0 Estimated Costs and Cost Effectiveness of Controls
Table 2 presents a preliminary summary of NOx controls that are feasible for cement kilns, the range of
potential NOx reductions from applying these controls, the cost effectiveness of the controls, and effects
on other emissions when using these controls. Table 2 has been reproduced from the Executive
Summary of PCA publication R&D 2227 (1999). Cost effectiveness was estimated by first assuming a
NOx reduction efficiency for each technology and then estimating the annual operating costs. Annual
operating costs included direct and indirect costs. Direct operating costs included fuel, power, labor, and
maintenance. Indirect annual costs included overhead, property tax, insurance, administrative charges
and capital recovery, as prescribed in Chapter 2 – Cost Estimating Methodology, OAQPS, U.S. EPA.
Capital recovery assumed a 10% interest rate and a 15-year equipment life. Estimated cost effectiveness
ranges from $2,100 to $11,500 per ton of NOx removed for those technologies considered feasible for
cement kilns. Costs shown on Table 2 are preliminary, based on industry experience and published
information. Capital costs contained in this report are based on the following (PCA, 1999):
Process Controls – Information from actual installations of various modifications or from engineering
estimates.
Conversion to Indirect Firing with a Low NOx Burner – Information from recent actual installations.
Mid-Kiln Injection of Fuel - Information from recent actual installations.
Low NOx Calciner – Information from a supplier of low NOx calciners with engineering estimates for
installation.
SNCR – Information from limited SNCR installations.

TABLE 2
SUMMARY OF THE FEASIBILITY OF NOx CONTROLS FOR CEMENT KILNS
NOx Reduction Technical NOx Reduction Cost Effectiveness Effect on Other

Technique Feasibility Potential (%) ($/ton NOx removed) Emissions
0 – 30
Process Modifications In full-scale use 3,100 – 8,800 Unknown
[15]*
Low NOx Burners
0 – 20 Unknown
(w/ Indirect Firing) In full-scale use 5,800 – 8,100
[10]*
Mid-Kiln Injection
0 – 30 May increase CO, SO2 or
Of Fuel, Riser Duct In full-scale use 5,100 – 11,500
[15]* hydrocarbon emissions
Firing, Calciners
Have been installed on several full- May increase emissions of
Low NOx Calciner 30 – 40
scale kilns – Offered by several 2,700 – 3,600 CO, SO2, and/or
[30]*
different vendors. hydrocarbons
May increase emissions of
May be applicable only on preheater 15 – 65 +
SNCR 2,100 – 4,700 CO, NH3, and NH4 salts
or calciner kilns – limited data [30]*
(detached plume)
*Values in brackets are the assumed NOx reductions used to calculate the estimated cost effectiveness of each technology.
This table has been reproduced from PCA R&D 2227 copyrighted in 1999.
Croom, J.M., M.S. May, J.C. Phillips, G.L.Young, R.S.MacMann; NOx formation in a cement kiln:
regression analysis. In: Proceedings of the Joint EPRI/EPA Symposium on Stationary
Combustion NOx Control; Vol. 2, Electric Power Research Institute, Palo Alto, CA, 1987.
Gardeik, H. O., H. Rosemann, S. Sprung, and W. Rechenberg; Research Institute of the Cement
Industry, Dusseldorf; Behaviour of nitrogen oxides in rotary kiln plants of the cement
industry; Trans of ZKG No. 10/84 (pp. 499 – 507), Zement-Kalk-Gips No. 12/ 1984.
Gartner, E.M.; Nitrogenous Emissions from Cement Kiln Feeds; Interim Report on Project
HM7140-4330; For presentation at Rule 1112 Ad Hoc Committee Meeting South Coast Air
Quality Management District, El Monte, CA; June 7, 1983.
th
Hansen, E.R., The Use of Carbon Monoxide and Other Gases for Process Control; 27 I.E.E.E.
Cement Industry Technical Conference Proceedings, May 1985
Kupper, D., W. Rother, and G. Unland; Trends on Desulfurization and Denitrification Techniques
in the Cement Industry; In: Proceedings of Emerging Technologies in Resource Recovery
and Emission Reduction in the Cement Industry; Sponsored by the Portland Cement
Association, Dallas, TX; Sept 1990.
Miller, F.M. and A.H. Egelov; Relationship between cement kiln operation and content of NOx in
kiln gases; Zement-Kalk-Gips International; Weisbaden, Germany, 1980.
Nielsen, P.B., and O. L. Jepsen; An Overview of the Formation of SOx and NOx in Various
Pyroprocessig Systems; Presented at: IEEE Cement Industry Technical Conference, XXXII;
Tarpon Springs, FL; May 22-24, 1990.
PCA, Report on NOx Formation and Variability in Portland Cement Kilns Systems - Potential
Control Techniques and Their Feasibility and Cost Effectiveness; by Penta Engineering
Corporation; PCA R&D Serial No. 2227; Skokie, IL; 1999.
Radian Canada, Inc.; Assessment of NOx Emission Control Technologies for Cement and Lime
Kilns; Radian Report No. 714-061-01; Prepared for: Environment Canada, Contract No.
K2035-3-7044; October, 1994
Rother, W.; NOx reduction; The Environmental Yearbook, published by International Cement
Review; Tradeship Publications Ltd.; Surrey, UK; 1997
Schraemli, Dr. W; Experiences with Measures to Reduce SO2 and NOx Emissions form Cement
KiIns In: Proceedings of Emerging Technologies in Resource Recovery and Emission
Reduction in the Cement Industry; Sponsored by the Portland Cement Association, Dallas,
TX; Sept 1990.
Young, G.L., R.S. MacMann, and J. Croom PhD, Technical Report on the Demonstration of the
Feasibility of NOx Emissions Reduction at Riverside Cement Company, Crestmore Plant; In:
Proceedings Before an Administrative Hearing of the South Coast Air Quality Management
District.; El Monte, California, January, 1986.
Young, G.L. and von Seebach, Michael; NOx Abatement in Kiln Plants in the Cement Industry,
Zement-Kalk-Gips, Volume 52, Number 6, 1999.
“STAGED” COMBUSTION FOR LOW-NO X CALCINERS
by:
Brian P. Keefe Robert E. Shenk

General Manager Manager of Pyro Technology
F.L.Smidth Inc. F.L.Smidth Inc.
ABSTRACT
This paper will present NOx reduction theory. From NOx reduction theory, it is hypothesized that an in-line
calciner that combines high temperature combustion and firing under the strongest possible reducing
conditions (without staging air or staging fuel) results both in the simplest calciner configuration and in the
lowest possible NOx and CO emissions. This hypothesis is supported by actual NOx and CO emissions
results from several U.S. plants.
INTRODUCTION
“NOx ” refers to the nitrogen oxides formed during the combustion of fuels in the cement pyroprocess. As
a precursor to ozone formation, NOx has increasingly become an essential topic for both suppliers and
operators of cement plants.
In a modern calciner kiln system, NOx emissions are generated from two sources:
• thermal NOx
• fuel NOx
Thermal NOx forms at temperatures above 1300ºC (2300ºF) and results from the oxidation of molecular
nitrogen (N2) in the kiln’s combustion air to NOx .
Fuel NOx forms at temperatures below 1300ºC (2300ºF) and results from the oxidation of nitrogen
compounds in the calciner’s fuel to NOx .
In the design of modern Low NOx Calciners, it is well known that (1) high temperatures and (2) reducing
conditions are proven methods for suppressing the formation of fuel NOx generated in the calciner and for
destroying thermal NOx generated in the burning zone of the rotary kiln.
“Staged Combustion” is a generic term that is used to describe a well known method for creating reducing
conditions in the calciner resulting from the gradual oxidation of fuel and, therefore, a localized deficiency
of combustion air. There are two methods of Staged Combustion:
• fuel staging
• air staging
In both cases, one or more “reduction zones” are created by the localized incomplete combustion of fuel.
Importantly, due to equally stringent restrictions on CO emissions from U.S. plants, modern Low NOx
Calciners must also not allow such zones of incomplete combustion to result in significant CO emissions
from the lowest preheater stage. Therefore, the best Low NOx Calciners are those they employ both
strong reduction and, separately, strong oxidation zones.
NOX CONTROL TECHNOLOGIES
As mentioned in the introduction, thermal NOx forms at temperatures above 1300ºC and involves the
reaction of nitrogen and oxygen molecules in combustion air. Thermal NOx is produced only in the
burning zone of the kiln. The amount of thermal NOx produced is related to fuel type, burning zone
temperature, and oxygen content. Different fuels produce different flame temperatures. The higher the
flame temperature, the larger the amount of thermal NOx that will be generated. For this reason, oil, coal,
and petroleum coke tend to produce less thermal NOx than natural gas. The rate of reaction for thermal
NOx increases as burning zone temperature increases. Therefore, hard to burn raw mixes will generate
more thermal NOx than easier burning raw mixes. A similar phenomenon occurs with higher kiln loadings
3
(tpd/m ). For a given kiln size, kiln inlet NOx will increase with increased throughput. The rate of reaction
of thermal NOx also increases with increasing oxygen content.
Conversely, fuel NOx is generated primarily in the calciner by the combustion of nitrogen in the fuel. Fuel
nitrogen will do one of three things:
• react with another nitrogen atom to form N2;

• react with oxygen to form NOx ; or
• react with a NOx molecule to form N2 and O2.
The amount of fuel NOx is dependent on a number of factors including fuel type, calciner type, and
calciner temperature.
In a modern pyroprocess system, a number of possibilities exist to reduce the NOx emissions. These
include the following:
• Burnability Improvements
• Low NOx Kiln Burner
• Low NOx Calciner
• SNCR (Selective Non-Catalytic Reduction)
The focus of this paper will be limited to Low NOx Calciners with an emphasis on “staged combustion”.
NOX REDUCTION THEORY:
To understand what is going on in a Low NOx Calciner, it is necessary to know further details of the
combustion process. However, more than 400 known reactions are involved in the NOx reduction and
combustion process, so the technical discussion below will be simplified and limited only to reburning
(burning under reducing conditions) which is the purpose of “staged combustion”.
Reburning in the calciner is caused by a sequence of reactions involving gases, soot and char from coal
pyrolysis, as well as catalytic effects of raw meal and char. There are a series of reactions that occur to
reduce NOx . The following reactions are a simplified version of the mechanisms in the NOx reduction
process:
REACTION 1: CO + NO è CO2, N2
REACTION 2: H2 + NO è NH3 (or N2), H2O
REACTION 3: CHi • + NO è HCN, where “CHi •” is a hydrocarbon radical and i = 1, 2, 3
REACTION 4: NH3 + NO è N2, H2O
REACTION 5: HCN + NO è N2, CO, CO2, H2O
The first and second mechanisms are catalyzed by raw meal and inhibited by oxygen. Therefore, by
introducing fuel and raw meal in the absence of tertiary air, thermal NOx entering the calciner from the
burning zone of the kiln will be destroyed and fuel NOx will be suppressed by the absence of oxygen.
Carbon monoxide, of course, is produced when there is not sufficient oxygen for combustion. Therefore,
bringing the tertiary air in above - or downstream of - the fuel point promotes NOx reduction. The extent
to which thermal NOX is destroyed and fuel NO X is suppressed is dependent upon the strength of the
reduction zone in the lower – or upstream – portion of the calciner where only oxygen from the kiln
exhaust gases exists.
The third mechanism is the reduction of NOx by hydrocarbon radicals formed during the pyrolysis of
calciner fuel. This reaction is favored by high temperature. Therefore, by creating a high temperature, of
o
1050 C or 1100°C, within the reduction zone, it is further possible to destroy thermal NOx and to suppress
fuel NOx . This effect is illustrated by FIGURE 1. Thus, by splitting the raw meal from the second-lowest
stage between (1) the reduction zone and (2) the oxidation zone of the calciner, it is possible to control
the combustion temperature in the reduction zone and, thereby, further reduce NOx emissions.
FIGURE 1: Effect of Reduction Zone Temperature on NOx Reduction
The compounds NH3 and HCN are formed from the degradation of NOx and the pyrolysis of the fuel. As
shown in the fourth and fifth mechanisms, these compounds – in the absence of oxygen - may further
destroy NOx when catalyzed by raw meal to form carbon dioxide and nitrogen gas. Alternatively, these
compounds may actually form NOx in the presence of oxygen.
Finally, the nitrogen present in the char (and/or soot) may either be oxidized to NOx or to N2 as shown in
competing mechanisms 6 and 7.
REACTION 6: Char N + O2 è NO
REACTION 7: Char N + O2 è N2, O2
The char-N (and/or soot-N) to NO conversion depends strongly upon fuel type and combustion
conditions. As a result, the competing mechanisms may vary between 10 to 80% in favor of NO
formation from nitrogen in the char.
CALCINER DEVELOPMENTS AND EXAMPLES OF STAGED COMBUSTION
The Low NOx Calciner is the standard for all new pyroprocess systems. With more stringent emission
regulations, it is unlikely that any new pyro systems will be built without a Low NOx Calciner. Low NOx
Calciners can be of two types: in-line calciners or separate-line calciners. Each has advantages and
disadvantages.
For normal fuels (coal, oil, and gas) an in-line with a minimum calciner residence time of about 3 seconds
is most commonly used. Conversely, for difficult-to-burn fuels such as petroleum coke and anthracite, a
separate-line calciner is usually preferred because its high oxygen atmosphere ensures improved fuel
burnout. However, from NOx reduction theory, it is well known that in-line calciners generally have lower
specific NOx emissions than do separate-line calciners. This is easily seen in FIGURE 2. For this reason,
in-line calciners may also be chosen for difficult-to-burn fuel applications, but the minimum residence time
is increased, as needed, to ensure adequate fuel burnout.
FIGURE 2: Historical Data Depicting NOX Emissions as a Function

of Calciner Type
Generally, both in-line and separate-line calciners are capable of meeting today’s CO/NOx emission
standards for virtually any combination of fuel and feed. The relative success of such calciners, however,
is dependent upon their inherent use of both strong reduction and oxidation zones. Typically, one or
more methods of promoting mixing (e.g. loop duct) to ensure fuel burnout are employed within the
oxidation zone so as to minimize the overall calciner dimensions. By using the proper calciner design
and residence time, it should be possible to maintain a bottom stage CO concentration of less than 100
ppm.
FIGURE 3 (Method 1 of Staged Combustion) presents a typical method of staged combustion which we
will call “gradual combustion”. The concept here is to introduce a weak reduction zone and a weak
oxidation zone over a large retention time. This presupposes that NOx reduction is increased with
increasing retention time.
FIGURE 3 (Method 1 of Staged Combustion): Staged Combustion by the “Gradual Combustion”
Method Employing Weak Reduction and Weak Oxidation Over a Long Retention Time
As shown in Figure 3, a portion of the calciner fuel is diverted to the kiln riser duct, prior to the introduction
of tertiary air. This creates a localized reduction zone which destroys a portion of the thermal NOx
present in the kiln exhaust gas. The purpose of the shallow angle for tertiary air entry is to produce
parallel flows of inert kiln exhaust gas and tertiary air. This increases the effective retention time of the
reduction zone by minimizing the mixing of unburned fuel and tertiary air. The high retention time under
reducing conditions is beneficial for thermal NOX destruction. However, some fuel NO X will necessarily
result because a portion of the calciner fuel is introduced directly into the tertiary air stream. For this
reason, it will often be advantageous in the actual operation of Figure 3 to maximize the calciner fuel split
to the kiln riser. Operation of Figure 3 under such conditions (i.e. stronger reducing conditions) is
consistent with NOx reduction theory for the destruction of thermal NO X and the positive effects of this
approach can readily be seen from FIGURE 4. Unfortunately, the slow (weak) oxidation of the resultant
(strong) reduction zone is prone to coating/buildups under high temperature conditions.
FIGURE 4: Effect of Fuel Split Between the Reduction Zone and the Oxidation
Zone on NOx Emission
As shown in Figure 3, combustion is completed in the loop portion of the tube calciner. Sufficient mixing
and retention time are needed at this location to ensure less than 100 ppm of CO from the bottom stage
exit gas. In fact, a separate mixing chamber at the apex of the loop duct may be needed to ensure such
low CO emissions to destroy the parallel flows of inert kiln gas and tertiary air.
FIGURE 5 (Method 2 of Staged Combustion) presents another variation of staged combustion in an in-
line calciner. Here, both fuel and air are literally staged to create “staged combustion”. Again, the
concept is to introduce weak reduction and weak oxidation zones over a large retention time. Here,
however, a series of weak reduction and oxidation reactions are applied to the kiln exit gas stream.
FIGURE 5 (Method 2 of Staged Combustion): Staged Combustion in the Literal Sense Employing a
Series of up to Three Successively Weak Reduction and Oxidation Zones Over a Large
Retention Time
As shown in Figure 5, fuel and air are alternately introduced at multiple locations along the tube calciner
to create a series of weak reduction and oxidation zones. Again, this concept presupposes that NOx
reduction is increased with increased stages of “staged combustion”. Fuel is first introduced to the kiln’s
feed shelf where a reduction zone destroys a portion of the NOx formed in the burning zone of the kiln. A
drawback of this system is the number of firing points (3) and air inlets (2), both of which must be
controlled via a series of air dampers and fuel splitters. Each successive stage is intended to have a
combined effect and the NOx reduction is assumed to be additive. Again, however, the NOx reduction
theory presented above suggests that the series of weak reduction and oxidation stages are actually
competing between NOx reduction and NOx formation. This is because the series of reduction zones are
weak and relatively undefined.
FIGURE 6 (Method 3 of Staged Combustion) presents a traditional method of staged combustion in a

separate-line calciner. As shown, fuel is introduced to a separate “pot” calciner who’s dimensions are
moderate since the kiln exhaust gases do not pass through it. However, in lieu of a fuel split to the kiln
riser duct, the “pot” calciner must be operated under sub-stoichiometric conditions in order to create a
localized reduction zone for the ‘kiln’s thermal NOx to pass through. This is accomplished by means of a
bypass (quarternary air) duct for tertiary air fed directly to the loop duct. Fuel NOx is suppressed and
thermal NOx from the kiln is destroyed. These two effects are additive.
FIGURE 6 (Method 3 of Staged Combustion): Staged Combustion in a Separate-Line Calciner with
Both Tertiary and Quarternary Air Ducts and Dampers
The primary draw-back of this approach (which applies to all separate-line calciner types) is that some
fuel NO X will necessarily result because a portion of the calciner fuel is introduced directly into the tertiary
air stream. The negative impact of this approach on NO X emissions has already been presented in
Figure 2. This effect is further highlighted by FIGURE 7 which shows 100% of the fuel routed
alternatively between the calciner and kiln riser duct.
To minimize this effect, a quartenary air duct is added, but this has the disadvantage in the number of
control dampers (2) that are needed to control the reduction and oxidation zones. One control damper is
needed to control the kiln/calciner air split and another damper is needed to control the amount of tertiary
air being bypassed to the loop duct. Furthermore, there is some concern about the CO emissions that
would result from such a system given that the retention time after the introduction of quartenary air is
limited. A final drawback to the approach presented in Figure 6 is the inability to control the temperature
of the reduction zone given that we have already seen the strong benefit of elevated temperatures in
Figure 1 for fuel NO X suppression. In Figure 6, this could be achieved by the addition of a meal split
between the “pot” calciner and the loop duct. While such temperature control could have a very positive
effect on NOx emissions as seen in Figure 1, a system with two (2) tertiary air ducts and two (2) meal
splits may pose operational challenges.
FIGURE 7: Effect of Fuel Split Between the Reduction Zone and the
Oxidation Zone on NOx Emissions
FIGURE 8 (Method 4 of Staged Combustion) presents a modern Low NOx Calciner design with neither
fuel nor air staging in an in-line calciner. Such a calciner results in the strongest possible reduction zone
and, separately, the strongest possible oxidation zone. As shown, 100% of the calciner fuel is introduced
to the kiln riser duct. This greatly simplifies the calciner design and it is only logical that addition of all fuel
to the reduction zone will give the lowest possible concentrations of O2 and, therefore, the strongest
reduction zone for minimizing both thermal NOx and fuel NOx . The successful introduction of 100% of the
fuel to the reduction zone demands only a single meal split from the second lowest preheater stage to
control the temperature in the reduction zone. The higher the temperature, the lower the NOx emission
(Figure 1 and Reaction 3). Moreover, NOx destruction is catalyzed by raw meal and inhibited by oxygen
(Reactions 1 and 2).
FIGURE 8 (Method 4 of Staged Combustion): “Staged” Combustion in an In-Line Calciner

Employing the Strongest Possible Reduction Zone and,Separately, the Strongest Possible
Oxidation Zone for Low NOX and CO Emissions
Among the advantages of the calciner design presented in Figure 8 are the use of a single firing point and
a single tertiary air entry without “staging”. Under these conditions, a reduction zone having a residence
time of only 0.2 seconds at 950-1050ºC is more than adequate to achieve the new MACT (Maximum
Achievable Control Technology) Rules that will go into effect as of June 2002 (<1.4 kg/t clinker or 2.8
lb/ST clinker). This is because a very strong transition is created between the rectangular reduction zone
and the cylindrical oxidation zone. The cylindrical calciner design promotes a high circulation rate of
unburned fuel particles at the transition between the kiln riser duct and the conical calciner inlet. This
rectangular-to-cylindrical transition will necessarily promote more mixing (and, therefore, lower CO
emissions) than the laminar flows characteristic of tube calciners. As shown in Figure 8, further mixing is
promoted by (1) the “notch” in the cylindrical portion of the calciner, (2) the transition from the calciner
vessel to the loop duct, and from (3) the loop duct to the bottom stage cyclone.
Such a simple approach to calciner design creates the strongest possible reduction zone for NOx
destruction and the strongest possible oxidation zone for conversion of all CO to CO2. NOx is absolutely
minimized since all of the fuel is introduced where pyrolysis is highest, oxygen content is the lowest,
temperature is highest, and thermal NOx is highest. Under such strong reducing conditions, it has been
demonstrated that the vast majority of NO X reduction takes place with a retention time of only 0.1 seconds
as shown in FIGURE 9.
FIGURE 9: Effect of Retention Time on NOX Reduction
On the basis of NOx reduction theory, it is only logical that the strongest possible reduction zone will yield
the lowest possible NOx emission. This theory is supported by actual emissions results from several U.S.
plants.
CONCLUSION:
“Staged combustion” is a well-known method for reducing NOx emissions from cement plants.
Traditionally, one or more “reduction zones” have been created by the localized incomplete combustion of
fuel. However, as both NOx and CO emissions have become increasingly stringent, such traditional
methods of NOx control via fuel and air staging have been reconsidered. Modern Low NOx Calciner
design combines high temperature combustion and firing under reducing conditions without staging air or
fuel in the calciner. In this way, the operation is kept as simple as possible. Actual emission results
demonstrate that the simplest calciner configuration, with only one fuel and only one tertiary air inlet, will
also provide the lowest possible NOx emission. The low NOx emissions made possible by such calciner
designs are supported by the NOx reduction theory presented in this article. As a consequence, the best
Low NOx and Low CO Calciner designs are those that employ both strong reduction and, separately,
strong oxidation zones.
REFERENCES:
C Jepsen, O.L. and P.B. Nielsen, An Overview of the Formation of SO X and NOX in Various
Pyroprocessing Systems, IEEE Cement Industry Technical Conference, May 1990.
C Jepsen, O.L., Green Cement: Designing Cement Plants of the Future to Meet the Most Stringent
Environmental Codes, International Cement Review, October 1995.
C Thomsen, K., L.S. Jensen and F. Schomburg, FLS ILC-Low NO X Calciner Commissioning and
Operation at Santa Cruz, ZKG International, October 1998.
C Keefe, B.P., Plant Modernization Projects Employ New Calciner Designs, International Cement
Journal, January 1997, pp 40-46.
C Keefe, B. P., O. L. Jepsen and Duc Vu, Unconventional Wisdom: Lonestar’s Unique Approach
to Environmental Challenges, World Cement 1999, pp 68-72.
C Keefe, B. P., O. L. Jepsen, and Tim Menke, Lone Star’s Unique Approach To Environmental
Challenges, International Cement Review, July 2001, pp 45-51.
C Shenk, R. and C. Pepe, Emission Prediction and Process Troubleshooting with Laboratory Tests
and Process Know-How, Profit Producing Technology Seminar, October 2001, F.L.Smidth Inc.
C Miller, S.W., Environmental Considerations for Preheater Design, 2002 IEEE Cement Industry
Technical Conference, May 2002.
LOW NOx ROTARY KILN BURNER TECHNOLOGY :
DESIGN PRINCIPLES & CASE STUDY
By :
Max H. VACCARO
Sales Manager
PILLARD E.G.C.I, Marseilles, France
[email protected]
For presentation at the :

IEEE - IAS/PCA 44th Cement Industry Technical Conference
LOW NOx ROTARY KILN BURNER TECHNOLOGY :
DESIGN PRINCIPLES & CASE STUDY
By :
Max H. VACCARO
Sales Manager, PILLARD E.G.C.I
ABSTRACT only use the technology of air staging but also

In the coming years, numerous cement plants in utilize fuel staging technology using very
USA could be forced to comply with tightening specific fuel injection points which helps to
air emission regulations especially with regard reduce the NOx emissions even further.
to nitrogen oxides (NOx). The United States
Environmental Protection Agency (USEPA) is
currently considering Low NOx rotary kiln HISTORIC BACKGROUND OF COAL FIRING
burners as a cost effective and efficient choice ROTARY KILN BURNERS DEVELOPMENT
for NOx reduction in cement kilns. This paper More than 30 years ago the first multi channel
will discuss the principles behind the design of kiln burner for pulverized coal firing was
low NOx kiln burner and will present data introduced ( see figure 2. ).
gathered from research on an industrially scaled
cement kiln. Finally the results in NOx
emission reduction by means of a low NOx
burner application on a rotary cement kiln in
Europe will be presented.
INTRODUCTION
The operating principle of low NOx burners is a
stepwise combustion which delays the mixing
and input of fuel and air at appropriate stages in
order to achieve a controlled combustion
process. This results in a fuel rich flame core
(Figure 2.)
and lower peak temperatures. Extensive
experience relating to low NOx burners has
In this burner, a channel feeding pulverized coal
been made available since the early 1970’s from
by means of a constant conveying air flow was
boiler furnace engineering. The first generation of
located between two primary air channels for
low NOx burners applied to furnaces were using
swirl air and axial air. The purpose of the interior
the basic principle of air staging (see figure 1).
swirl air component was to expand the
pulverized fuel flow and stabilize the flame by
generating an internal recirculation zone. In this
manner it was possible to control the flame
shape in conjunction with the constricting effect
of the exterior axial air component.
The basic principles of this burner were a low
coal injection velocity in order to reduce wear
and a low conveying air flow in order to keep
additional primary air input (necessary for
adjustment of flame shape) and thus total air
input along the burner, as low as possible.
The design principle of this multi channel burner,
which was generically called the “three channel
burner”, has played a dominant role in the
(Figure 1.) history of rotary kiln firing in the years following
its development.
Nowadays, the new generations of low NOx
burners applied to heavy industrial furnaces not
APPLICATION OF LOW NOX BURNER thus prevent premature mixing of the fuel-/
DESIGN TO ROTARY KILNS primary air mixture with hot secondary air.
In the meantime, Increased environmental
requirements in Europe made it necessary to An early low NOx burner prototype was installed
search for better solutions for reduction of NOx in a German cement plant in 1989 and
emissions while maintaining or improving incorporated the following basic principles for
product quality and kiln operation. Experience reducing NOx formation:
gained from low NOx burner technology in boiler - Proper flame stabilization by generation of a
furnaces (using air staging principle) and also strong inner recirculation zone, promoting
from hundreds of firing systems in rotary kilns, early ignition of the fuel into an oxygen
in conjunction with theoretical studies by the deficient atmosphere.
combustion research department of - Concentration of fuel along the burner axis
Braunschweig university in Germany, finally in order to prevent fuel particles leaving the
concluded with a complete revision of the burner main flame contour and burning at over -
tip geometry (see figure 3). stoechiometric conditions in the hot
secondary air atmosphere.
Axial air - Reduced primary air amount in comparison
Radial air with previous burner designs.
Pulverized coal Approximately 8% to 10% of total
combustion air requirement flows into the
axial, radial, central and pulverized fuel
channels compared to around 15% with a
typical three channel burner.
- Improved burner tip aerodynamics and
adjustable tip position which allows flame
shaping with a minimum of primary air input
Flame stabilizer and maximum outlet velocity as no primary
air valves are used for primary air flow
Extended outer pipe
adjustment.
(Figure 3.) - More even heat distribution in order to
reduce temperature peaks which are
Those revisions were : responsible for higher NOx formation and
- The pulverized fuel duct was relocated to also for increased thermal stress on the
the central part of the burner i.e. inside the refractory lining.
swirl and axial air channels.
- For flame stabilization a kind of bluff - body,
the flame stabilizer, was installed inside the
coal channel. The purpose of the stabilizer,
a concept from boiler burner technology,
was to induce recirculation core eddies in
order to guarantee proper stability of the
flame root.
- The swirl and axial air channels were
optimized by design measures, with respect
to pressure drop, in order to convert the
available pressure of the primary air supply
to a maximum tip outlet velocity so as to
maximize the momentum necessary for
flame control while minimizing the primary
air flow. This burner showed significantly reduced NOx
- The previously designed annular gap for the emissions without affecting process efficiency or
axial air channel was replaced by a number product quality and initiated the path for further
of slots. The purpose was to maintain a development.
perfect concentricity of the axial air channel
and to promote the introduction of
recirculated combustion gases of high CO2 “CEMFLAME” RESEARCH CONSORTIUM
content into the flame root, thus reducing In 1993, a European consortium of cement
the free O2 content. producers and associated research
- An air guide ring was added to correct organizations was formed with the purpose of
outflow direction of axial air component and funding research on a scaled simulation of an
industrial cement kiln. This consortium was
known as “CEMFLAME”. The first research Lastly, it was observed that the ignition point
program was proposed to study the influence of location and temperature was of major
fuels and burner design on NOx generation. importance in NOx generation ; the earlier the
ignition point occurs the lower the NOx
Effect of fuel type : emission level.
During the experimental campaign, fuel type
was observed to have a strong influence on the
flame characteristics. Depending on the CASE STUDY OF REDUCTION IN NOx
pulverized fuel characteristics (type, fineness...), EMISSION WITH A LOW NOX BURNER
a long coal jet could be visible or ignition of the In 1996 the European Community financed a
coal close to the burner was observed. Ignition research program entitled “NOx emissions from
further downstream the furnace generally Cement Manufacture and Evaluation of Various
resulted in higher NOx levels when compared to Possibilities for NOx reduction in the Cement
early ignition at the burner tip. Industry“. A series of tests were performed in a
cement plant, the aim of which was to
Effect of burner design : demonstrate the potential reduction in NOx
The highest NOx emissions were obtained from emissions through the optimization of the main
a mono channel burner design whatever the fuel kiln burner design without affecting clinker
type. Due to high axial momentum promoting quality and production output but without
high entraining of secondary air, the flame was increasing other pollutant emissions.
igniting further downstream in the furnace, and For this purpose, a low NOx burner was tested
at the point of ignition more air was entrained in the SARCHE cement plant in Italy :
into the flame leading to high thermal NOx
formation. SARCHE plant, a 700 tpd LEPOL grate cement
For the case of the generic three channel burner kiln, was originally equipped with an indirect
design, the effects of tangential momentum fired mono channel type burner firing 100%
(radial air), resulted in the following : A slight petroleum coke and having a total primary air
reduction in NOx level was observed due to a input of 10.5% of the total combustion air
part of the coal flow being entrained into the requirement (stoechiometric basis).
oxygen lean condition of the flame centerline.
However, the overall entraining of the secondary A multi channel low NOx burner design was
air was enhanced due to an increased tangential installed to replace the mono channel burner
momentum resulting in higher overall oxygen and was commissioned in March 1996.
concentration at the point of ignition. Both
effects were competing against each other, Commissioning phase :
therefore results in terms of NOx reduction were Due to oversizing of the low NOx burner tip with
not optimum. regard to heat output (24 Gcal/hr instead of 16
Gcal/hr) a higher percentage of total primary air
Conclusions : was found when compared with the original
In view of the above observations, the influence mono channel design (17.5% versus 10.5%).
of the pulverized coal channel location in Notwithstanding the fact of higher primary air
comparison to the primary air channels (axial compared to the original mono channel burner, a
and radial) was proven to be of extreme NOx reduction of 28% at kiln inlet was obtained
importance because of its strong effect on the while keeping low CO and SO2 levels and a
amount of coal being entrained in the local normal kiln shell temperature. However, a
oxygen lean region of the burner. Hence, it was higher specific fuel consumption (830 Kcal/kg
proven that best location for coal channel should clinker versus 805 Kcal/kg clinker) was found,
be in the burner center, then surrounded by both partially due to the high percentage of primary
primary air streams. air and to a change in the burning zone.
Also, it was demonstrated that the air First experimental campaign :

distribution affects the local mixing in the near New settings were made to the burner in order
burner zone and the ignition point. Higher axial to reduce the primary air amount as much as
momentum (like for mono channel burner type) possible while keeping the existing burner tips.
results in high secondary air entraining into the The total primary air input was reduced from
axial air stream. As a result, local mixing 17.5% to 12.5%. This setting was maintained
between the hot and oxygen rich secondary air for some weeks and showed, with a good
flow and the outside of the coal jet is enhanced, repeatability, an average of 45% NOx reduction
thus promoting higher thermal NOx emissions. from original mono channel burner.
The main goal of this experimental campaign performance and clinker quality, this setting was
was the optimization of the burner tip through an adopted as the final one. It is to be noted that
understanding of the effect of axial and radial during this last period, the production was
flow rates in terms of axial momentum and swirl increased to 710tpd (+15 .8%) and specific heat
number. For this purpose it was decided to consumption was reduced to 800 ~ 805 Kcal/kg
install continuous flow measuring devices on the clinker while NOx emissions were reduced by
axial and radial air channels. After installation 35% from those by original mono channel burner
of these flow measuring devices, new tests were design.
performed to investigate the effect of a different
primary air distribution and a higher primary air Mono Channel
Burner
Multi Channels Low Nox burner
(according to Figure 3)
flow on the NOx emission level. This series of Original setting Commissioning First adjustment Revised tip design Last tip design
tests confirmed the fact that the burner tip was Production (tpd) 598 600 600 kiln unstable 710
oversized and justified a revision of the tip Specific heat

(kcal/kg) 805 830 830 kiln unstable 805
design with smaller cross sections. consumption
Total primary air

(%) 10.50% 17.50% 12.50% 11.40% 13.60%
amount
Second experimental campaign : NOx at kiln inlet

ppm 881 631 487 407 577
A new set of burner tips with smaller outlet (10% O2 basis)
sections was installed in June 1996 in order to Equivalent NOx Lbs/ston

emissions clinker
7.19 5.15 3.97 3.32 4.71
further reduce the total primary air input to

11 .4%. At that time a 54% NOx reduction was
Conclusion of the tests :
achieved at the kiln inlet, however, kiln operation
First, a multi channel low NOx kiln burner made
problems arose. The kiln became unstable with
according to Figure 3, by design itself clearly
a reduction in production and a significant
reduces the NOx emission when compared to a
increase in SO2 emissions. An immediate
mono channel burner while keeping low CO and
intervention was necessary in order to permit
SO2 levels.
normal kiln operation without losing production
capacity. A hybrid tip configuration was chosen
Also, the influence of primary air on NOx
using the swirler of the old tip. This tip
generation has been investigated, showing that
geometry was maintained for one month giving
the reduction in primary air positively influences
better operational results, but the SO2
NOx emission. However primary air reduction
emissions were still significant and heat
should be kept within certain limits in order to
consumption of the kiln was not optimal.
maintain a good and stable kiln operation,
Moreover, a buildup on the axial air tip of the
especially in the case of using high sulfur and
burner was observed. A kiln shutdown was
low volatile content pulverized fuel such as
necessary to clean the burner tip in order to
petroleum coke.
avoid flame distortion. The cleaning of the axial
tip resulted in a better kiln operation with a
Lastly, the geometrical feature of the burner is of
slight decrease in SO2 emission but for a
extreme importance in order to avoid plugging
limited time. However, after a few days the SO2
problems that would eventually result in
level became unacceptable because the axial
reduced performances.
air tip was partially plugged again.
Final burner tip design:

SEMI DIRECT FIRED LOW NOX BURNER
Several changes in burner tip design occurred
TECHNOLOGY
between July and September 1996. The best
Generally, low NOx burners are used in
compromise was found by using an increase in
conjunction with an indirect firing system that
the axial outlet cross section while keeping the
allows a reduction in total primary air to the kiln
same number of slots to maintain tip
in comparison to a direct firing system. The
concentricity. This design allowed the use of a
benefits given by an indirect firing system are
higher primary air flow through the axial air
well known but in some cases conversion to
channel and a thus better cooling of the axial air
indirect firing can hardly be justified from an
tip. The primary air input from axial, radial,
economical point of view. Therefore the
central and pulverized fuel channels was
advantages of using of a modern multi channel
increased to 13.6% of the total combustion air.
low NOx burner can be missed.
Operation results were very satisfying with
negligible SO2 emissions and a more stable kiln
However, it is to be noted that installation of a
behavior. A slightly higher NOx emission was
modern low NOx burner technology on a direct
recorded but given the fact that SO2 emission
fired kiln is possible.
had disappeared and that a more stable kiln
behavior positively influenced production
With a few modifications to the coal milling tip by means of an additional channel
circuit it is possible to convert the firing system surrounding the burner. In this way, this
to semi direct i.e. by adding a cyclone on the “additional secondary air” does not enter the
primary air stream coming from the coal mill and main flame which can now fire with a lower
redirecting this primary air to a multi channel primary air amount leading to reduced NOx
low NOx type burner. In this case, the primary emissions.
air to the burner is indeed higher than with an
indirect firing, however the effect of flame
stability, and combustion staging, as CONCLUSION
demonstrated before, may effectively contribute The above paper has discussed the design
to a reduction in NOx emissions. principles of low NOx burners and shown how
NOx emissions can be reduced at the kiln inlet
This type of semi direct firing system was by the means of such burners.
commissioned in 2001 on two long wet kilns at
LOGANSPORT plant in Indiana. The theoretical studies from “CEMFLAME”
NOx data after installation of the low NOx research group have proven that it is possible to
burners are not yet available at this plant. achieve drastic NOx emission reductions by
However kiln stability and clinker quality have burner design. Geometrical feature of the burner
been improved since installation of these low and mostly, the location of coal channel in
NOx burners. According to users, the clinker comparison to the primary air channels are of
reduced iron content is lower and average extreme importance.
clinker free lime have been increased. Finally,
there is no more problem with nose ash ring Additionally, actual testing in SARCHE cement
formation since installation of the new burners. plant has demonstrated that low NOx kiln
burners can, with a good repeatability and over a
Experience gained from the LOGANSPORT long period of time, reduce the NOx emissions
project has lead to a new generation of semi at kiln inlet while maintaining clinker quality,
direct firing systems (see figure 4) which aim to heat consumption and kiln stability, and without
reduce NOx by installing a low NOx kiln burner increasing SO2 and CO emissions.
design and additionally by reducing the primary
air inside the main flame core. Still experience and research in the field of NOx
reduction from the kiln burner standpoint is
progressing. In partnership with ITALCEMENTI a
(1) new campaign of tests will be performed shortly
(5) in Europe with the goal of looking at additional
design features to further reduce NOx from the
(12) current levels obtainable today with a “standard”
(4) (11)
low NOx type burner while maintaining or
(6)
improving clinker quality and production
(8) (9) (10)
performance. The new burner designs are
(A)
(2) (13) subject to several patents :

n US patent 6293208 by :
( B)
R.Guiot/L.Russo/JC.Gauthier/JC.Pillard
(7) n US patent 6315551 by :
( 3)
® P.Salzsieder/G.Endres/R.Guiot
n European patent 0421903 by :
(1) Raw coal silo (8) Sw irl air channel W.Collenbush
(2)
(3)
Hot drying air
Coal mill
(9)
(10)
Axial air channel
Additional secondary air channel
(4) Mill e xhaust fan (11) Kiln bur ner JC.Pillard/JC.Gauthier/J.Pizant
(5) Cyclone (12) Kil n
(6) Mixing injector (13) F resh air inlet damper n European patent 0967434 by :
(7) Prim ary air & conveying ai r fan (14) Tempering air damper P.Salzsieder/G.Endres/R.Guiot
(Figure 4.)
R.Guiot/R.Asdiguian
In this type of semi direct firing system the
primary air going into the main flame is limited
to the amount necessary to provide the required
momentum for flame control. The primary air in
excess (called “additional secondary air”) , i.e.
that which is not necessary for flame control, is
evacuated inside the kiln hood before the burner
"Staged Combustion for NOx Reduction Using High Pressure Air
Injection"
By Eric R. Hansen, Technical Consultant, Cadence Environmental Energy Inc.
Abstract: The effects on emissions by applying high-pressure air injection on rotary kilns to
achieve effective mixing and staged combustion are presented. A technique of using the energy of high-
pressure air jets to achieve cross-sectional mixing in a tertiary combustion zone in rotary kilns resulting in
reduced excess air requirement for optimum combustion is described. An additional result of the mixing
method is the creation of a sub-stoichiometric secondary combustion zone that reduces the emission of
nitrogen oxides.
Introduction: There are numerous regulations providing the incentive to reduce nitrogen oxide
emissions. Plants desiring to expand are faced with the Prevention of Significant Deterioration (PSD)
criteria and Best Available Control Technology (BACT) requirements, and certain plants are being faced
by Federal and State Implementation Plans (FIP and SIP’s) requiring the reduction in NOx emissions due
to concerns for the Regional Transport of Ozone. Additionally, the Regional Haze rule will bring in
requirements for facilities not previously affected requiring Best Available Retrofit Technology (BART).
Thus, the cement and lime industries are faced with continuing pressure to reduce the emissions per unit
of product and are in dialog with the regulating agencies to find reasonable means to accomplish the
objectives of the regulating agencies in a manner that allows continued economical operation of the
facility.
The staging of combustion is a generally accepted means of managing NOx emissions. It shows up as a
prescriptive technology in the “Alternative Control Techniques Document” (EPA ACT Document) and the
“NOx Controls for the Cement Industry” (EPA NOx Report) (references 1 & 2). Additionally, there are
new developments that are now available to enhance the staging of combustion. This is the injection of
high pressure air into the hot process gas to achieve or enhance both the staging of combustion by
managing the availability of oxygen in various zones of the kiln, and by also reducing the amount of
excess oxygen to achieve a given level of completeness of combustion as measured by the CO
concentration. The reduction of excess oxygen required is achieved by effectively mixing the cross
section of the gas in the kiln or duct. This method of mixing also results in limiting the available oxygen in
the primary combustion zone as a result of both reducing the excess air required and providing this
excess air after the primary combustion zone. This technique can be applied in addition to using multiple
fuel firing points on rotary kilns.
Staged Fuel Combustion (SFC): Staged fuel combustion (SFC) as a control technique is well
documented in the regulatory references (1-4). Staged Fuel Combustion is the introduction of a portion of
the fuel in other than the primary, hottest combustion zone of the kiln. By burning a portion of fuel at
reduced temperature, some thermal NOx is avoided and there is possibly some interaction between
localized reducing conditions at the secondary firing point and the NOx from the primary combustion zone
that results in lower overall NOx emissions. The observed NOx reductions may also be a result of the
cooling of the main flame due to the high excess air that results from the fuel substitution. Multiple fuel
firing points are used in precalciner kilns, long kilns firing municipal waste, tires or hazardous waste at
mid-kiln and in preheater kilns that are introducing tires onto the feed shelf. These means of mid-process
firing have received favorable review with respect to NOx reduction in the EPA documents. The EPA
NOx Report states that the addition of waste tires on the feed shelf reduces NOx emissions from
preheater kilns by 30 to 40%. Similarly, long kilns equipped for mid-kiln firing of waste or tires were
reported as achieving significant NOx reductions. From the EPA NOx Report, “In nine tests on dry kilns,
the average reduction in NOx emissions was 33% with a range of 11% to 55%. In three tests on the wet
kilns, the average reduction in NOx emissions 40%. It should be noted that the three kilns that are known
to have CEMS all reported emission reductions of 45% or more.” Figure 1 is a graphic tabulation of NOx
measurement on long kilns with mid-kiln fuel addition showing NOx reduction. The kilns experiencing NOx
reductions by the use of waste derived fuels were primarily motivated by the savings and revenue from
substituting waste for the primary fuel and were not driven by the necessity for NOx management.
In cases where waste-derived fuels are not readily available or there is public resistance to the use of
tires, lump coal could be used to achieve the same mechanism of control through staged combustion
(with some regard for the coal volatility since the hydrocarbons from pyrolysis of the fuel play a role in the
NOx reduction). The full potential of Staged Fuel Combustion for NOx reduction has not been fully
evaluated since the bulk of data available is incidental to the energy recovery activities.
Staged Air Combustion (SAC): Staged Air Combustion is the multipoint introduction of combustion air.
Staged air combustion used in addition to staged fuel firing will allow for the combustion of a portion of the
fuel under sub-stoichiometric conditions to both minimize the formation of NOx from the fuel combustion
and to provide reducing conditions to destroy NOx formed in the high temperature, primary combustion
zone. Staged air combustion is can be achieved in precalciners by splitting the tertiary air so that the
calciner fuel is burned under sub-stoichiometric conditions and the additional air to complete combustion
is added later in the calciner vessel. However, to achieve significant NOx reduction, the reducing
5
atmosphere needs to be created in the temperature zone of 1000 C to 1200 C as indicated by Conroy.
A limitation to the application of staged air combustion at the critical temperatures is the achievement of
adequate mixing. By the use of high-pressure jets, energy can be added to the process gas to achieve
the necessary mixing.
Development of High Pressure Air Injection: The high-pressure air technique to achieve mixing and
staging of the air for combustion was initially developed for long kilns equipped with mid-kiln firing. There
is a high incentive to increase the use of mid-kiln fuel because the mid-kiln fuel has not only fuel value
also generates revenue from disposal fees. The limitation in the amount of mid-kiln fuel substitution is the
loss of flame temperature at the primary burning zone as mid-kiln fuel is increased (the air-through
calciner limitation). This loss of flame temperature is the same phenomena that air-through calciners
experience, limiting the riser fuel-firing rate. Normally, a cement kiln operates with about 10% excess air
to complete combustion. If 15% fuel substitution is done with tires, this results in an increase of excess
air at the main flame to 25% (see Example 1. and Example 2.). The addition of mid-kiln fuel results in
significant cooling and shortening of the main flame due to the higher excess air (which may be the
primary explanation of the NOx reduction observed). Further fuel substitutions, as shown by Example 3,
eventually result in flame temperatures insufficient to maintain stable kiln operation. Generally, kilns that
have higher secondary air temperatures achieve higher fuel substitution rates for this reason. Wet kilns
are typically limited to 15% to 20% substitution and dry kilns to 20% to 25% substitution due to limiting
flame temperatures.
Some kilns needed to increase the kiln exit oxygen levels to control CO while burning fuels mid kiln. In
examining this limiting factor, extensive computational fluid dynamic computer modeling was done on the
atmospheric conditions in the calcining zone of the kiln. In retrospect, it is not surprising that the kiln gas
in the calcining zone of a cement kiln is highly stratified. The gasses at the top of the kiln are hotter, 1300
to 1400 C, while the gasses coming off the calcining bed are only 850 C and have a molecular weight
50% higher than the combustion gasses and as a result have more than twice the density. With the
significant differences in gas density, there is a tendency for the gasses to remain stratified as they pass
through the kiln at over 10 meters per second. Measurements of oxygen levels in the back end of
preheater or precalciner kilns confirm this predicted stratification. When one drops a tire or fuel module
into the calcining zone of the kiln, it falls into this relatively cool pool of carbon dioxide and hot meal and
does not immediately burn but pyrolysises over a period of several minutes. It is not surprising that even
though there is enough oxygen in the kiln to complete combustion, there is some sneakage of carbon
monoxide due to this stratification of gasses. If one could get enough energy into the kiln to do cross
sectional mixing, a reduction in oxygen headroom to get a given CO performance would be possible.
The above considerations led to the development of the injection of high-pressure air into the kiln to
provide the energy to do the mixing. Through computational fluid dynamic computer modeling, geometry
was developed where the energy in an air jet would couple with the kiln gas, creating a rotational
component to the kiln gas and effectively mix the injected air into the kiln gas and mix the kiln gas through
the full kiln cross section. Depending on the pressure of air used, this could be accomplished with as little
as 2% of the kiln gas. Air velocities of 100 to 200 meters per second are injected with a tangential vector.
This velocity compares to the 10 to 15 m/sec kiln gas velocities. There is sufficient energy in the injected
air to impart a rotational component to the kiln gas. The stratification is illustrated in Figure 2. The
amount of kinetic energy in the injected air is about 0.3 kilowatt hour per metric ton of total gas in the kiln.
The amount of air and the pressure are adjusted to specific applications to provide the energy necessary
to achieve adequate mixing and the degree of combustion staging desired.
The injected air can function like tertiary air, which will allow higher mid-kiln fuel substitution rates. Note
the increase in flame temperature calculated in Example 4 over that in Example 3. However, injecting
ambient air at mid-kiln does displace hot air at the null point of the cooler, so generally the air injection is
limited to 10% of the total combustion air. A careful distinction needs to be made between tertiary air and
mixing air. The quantity of mixing air is intended to be about the same as the quantity of the excess air
for combustion. The injected mixing air is principally to introduce energy into the process gas to achieve
cross sectional mixing. The quantity is typically limited to provide only the excess air for combustion so
as not to substantially move the point of heat release of the fuel. The principal function of tertiary air is to
provide combustion air.
The air injection could be placed either uphill or downhill of the mid-kiln fuel introduction point with
favorable results predicted with respect to fuel substitution rates. However, due to the growing int erest in
NOx control, the air injection system was placed uphill of the fuel feeding point in order to limit the oxygen
available in the mid-kiln combustion zone where the temperatures are favorable for NOx destruction.
Example 4 shows the effect on the kiln atmosphere of the injection of 10% of the combustion above the
mid-kiln firing point. Note that the excess air at the main flame is reduced, restoring satisfactory main
flame temperature at the higher fuel substitution rate. Additionally, the zone in the kiln between the mid-
kiln injection point and the air injection becomes slightly reducing, enabling NOx destruction in the
important temperature zone. Figure 3 illustrates the results of using high-pressure air injection on a long
wet kiln. It shows the reduction of SO2 and CO while achieving a 50% reduction in NOx emissions while
doubling the mid-kiln substitution rate. Figure 4 is a frequency distribution of the NOx data showing a
shift in the baseline NOx levels required to make a satisfactory quality clinker.
Application of High Pressure Air Injection on Preheater and Precalciner Kilns: The high-pressure
air injection was initially developed to increase the utilization of waste-derived fuels. However, the
potential to reduce emissions was not overlooked. Preheater kilns are also in need of a retrofit NOx
reduction technique. The staging of fuel combustion by riser firing is reported to have little benefit for
NOx reduction. Nielson and Jepsen report that “preheater kilns firing finely ground fuel into the riser
duct, the NOx content in the exhaust gas may increase on passing through the riser duct. As the NOx
emissions from the kiln may also increase slightly due to an increased excess air rate, the total NOx
6
emissions from the kiln system may increase when starting up riser firing with finely ground fuel.”
However, the firing of coarse fuel such as tires on the feed shelf of preheater kilns is observed to reduce
7
NOx. When tires are not available, lump coal should perform as effectively.
The use of mixing air to create a rotational component and break up of the stratification at the feed end of
a preheater or precalciner kiln may enhance the ability to add secondary fuels on the feed shelf. The
rotation of the kiln gas minimizes stratification at the feed hood, potentially eliminating buildup problems
and allowing increased coarse fuel use. Also, depending on the amount of fuel that continues to burn in
the bed downhill of the air injection point, sub-stoichiometric conditions may also be achieved resulting in
a reducing atmosphere in the critical temperature range for NOx destruction. Feed shelf firing of tires or
lump coal, and high-pressure air injection is a means of retrofitting staged air combustion on preheater
kilns. This type of retrofit may be particularly attractive for preheater kilns with satellite coolers that must
reduce NOx emissions. For these kilns, conversion with tertiary air ducts is particularly costly due to the
requirement to change out the cooler. The addition of lump coal or tires and high-pressure air injection
would have a better cost efficiency per ton of NOx reduced than the addition of tertiary air duct and a
precalciner vessel.
This retrofit is also applicable to the rotary kiln portion of precalciner kilns. For preheater and precalciner
kilns, the amount of air injected is kept to the minimum required to introduce sufficient energy to achieve
mixing. The theoretical fuel penalty for using ambient air instead of preheated air is a result of the
difference in temperature of the air at the null point of the cooler and the ambient air. This hypothetical
fuel penalty is small and can be more than cancelled out if the improved mixing allows operation at lower
excess air levels for the same degree of completeness of combustion.
Lime Kilns - Staged air combustion for NOx control is also applicable to lime kilns. Since lime kilns can
produce a satisfactory product with reducing conditions in the high temperature zone, staged air
combustion can be effectively achieved with the mid-kiln introduction of high-pressure air. The calcining
zones of lime kilns are also highly stratified, and the lime industry uses trefoils to break up the
stratification to improve heat transfer. The idea of introducing energy in the middle of the process to
create a rotation of the kiln gasses will actually prevent the formation of stratification and bring the hot
gasses at the top of the kiln in direct contact with the limestone bed, adding a mechanism of heat transfer
to the kiln. Even for preheater lime kilns, the rotation of the kiln gasses starting several diameters
downhill of the kiln gas exit may serve to minimize operating problems in the preheater riser.
Excess Air Management: The use of high-pressure air to achieve cross sectional mixing of the gasses
in the kiln is a means of reducing the excess air required for a given carbon monoxide level. To take
advantage of reduced excess air requirement, it is important to implement process controls that take into
account the trace gas emissions. Controlling at the minimum excess air level shows significant potential
for NOx management. From the EPA NOx Report, “The use of state-of-the-art continuous emission
monitoring systems and feedback control, excess air can be accurately controlled to maintain a level that
2
promotes optimum combustion and burning conditions in addition to lowering NOx emissions.” A study
8
on the effect of excess air on NOx emissions on a precalciner kiln is documented in a report by Hansen.
Any NOx management technique must be operated and evaluated using emission measurements and
feed back control to be effectively used and accurately evaluated.
Conclusion: Staged air combustion is an available retrofit NOx control technology for long, preheater
and precalciner cement kilns and lime kilns. For cement kilns, this technique is used with staged fuel
addition using lump coal or waste-derived fuels such as tires introduced mid-kiln or on the feed shelf. For
mineral processing kilns that can tolerate reducing conditions in the hot zone, like lime kilns, the air
injection technique can be used alone without the addition of secondary fuel. The air injection technique
reduces NOx by limiting the availability of oxygen in a secondary combustion zone and by reducing the
overall excess oxygen required to achieve optimum combustion by cross section mixing of the kiln gas.
Particularly effective NOx reduction is achieved when the reducing zone can be created in the
temperature range of 1000 C to 1200 C. The air injection technique is a practical means of retrofitting a
kiln to exploit this mechanism of NOx control, and when waste fuels are available, NOx control can be
quite economically implemented.
Example 1
Coal TDF Mixing Air Fan Calculated Value

TPH 7.44 TPH 0.00 % 0.0% 0-10% of Combustion Air Fixed Values
% of BTU 0.0% Kiln Exit Gas
TPH 87.3
Coke °F 772
TPH 0.0 %O2 wet 1.5
Excess Air 8.9%
Combustion Air
TPH 77.7 Dry Raw Material
°F 978 TPH 62.0
A/F 10.4 % Solids 95.5
Flame Temperature Excess Air after SWDF TPH H2O 2.92
Clinker & Ash 4,159 °F Heat Loss 8.9% Excess Air
TPH 39.3 MMBtu/hr 55 CKD
°F 2,650 Oxygen after the Main Flame TPH 2.0
1.98% % Wet
8.9% Excess Air
Example 2

TPH 88.9
Coke °F 742
TPH 0.0 %O2 wet 1.5%
Excess Air 8.9%
Combustion Air
°F 957 TPH 62.0
A/F 10.9 % Solids 95.5
5.05% % Wet
23.4% Excess Air
Example 3

TPH 90.1
Coke °F 721
TPH 0.0 %O2 wet 1.5
Excess Air 8.9%
Combustion Air
°F 942 TPH 62.0
A/F 11.2 % Solids 95.5
7.30% % Wet
34.5% Excess Air
Example 4

TPH 90.1
Coke °F 721
TPH 0.0 %O2 wet 1.5
Excess Air 8.9%
Combustion Air
°F 1,039 TPH 62.0
A/F 10.1 % Solids 95.5
Clinker & Ash 3,597 °F Heat Loss -1.1% Excess Air
5.85% % Wet
24.5% Excess Air Heat Loss Due to Ambiant Air At Mid Kiln in MMBTU/hr -2.2
Heat Loss As A Percentage of Total Heat Input -1.06%
Figure 1
Nox Reduction on Nine Kilns Using Mid-Kiln Firing
NOx Reduction at Cement Kilns

Feeding Whole Tires Using the Cadence Mid-Kiln Technology
1600
1400
1200
1000
PPM
800
600
400
200
0
Plant #6 9% Plant #6 15%
Plant #1 Plant #2 Plant #3 Plant #4 Plant #5 Kiln #1 Plant #5 Kiln #2 Plant #6 23% tires
Tires Tires
Baseline 1359 1372 936 1342 565 513 1053 1053 949
With Mid-Kiln Fuel 883 994 790 600 488 456 755 562 471
Percent Reduction 35.0% 27.6% 15.6% 55.3% 13.6% 11.1% 28.3% 46.6% 50.4%
Baseline With Mid-Kiln Fuel

2600 F
Effective
Gas Rotation
Carbon Mixing of all

Dioxide Kiln Gasses
1550 F
Without mixing air

With mixing air
Figure 2
Illustration of the Stratification of Gasses in the Calcining Zone and the Effect of Mixing air
Emission Reduction Using Mixing Air
700
599
600
500
Emissions in PPM
400
359
313
300
228
200
121
100
10 8 6
0
Carbon Monoxide cor. Sulfur Dioxide ppm Total Hydrocarbons cor. NOx ppm
Mixing Air Off Mixing Air On
Figure 3
Effect of Mixing Air on Carbon Monoxide, Sulfur Dioxide, and Nitrogen Oxide Average Concentration
Histogram of NOx Values
250
200
150
Frequency
100
50
0
50
0
00
50
00
50
00
50
00
0
0
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
10
10
11
11
12
12
13
NOX Concentration in PPM
Without Mixing Air With Mixing Air
Figure 4
Frequency Distribution of NOx Concentrations With and Without Mixing Air
References:
1. “Alternative Control Techniques Document-Control of NOx emissions from Cement
Manufacturing,” U.S. EPA, Office of Air Quality and Standards, February 1993.
2. “NOx Control Technologies for the Cement Industry,” U.S. EPA, Office of Air Quality Planning and
Standards, Ozone Policy and Standards Group, September 19, 2000.
3. “Assessment of NOx Emission Control Technologies for Cement and Lime Kilns,” Environment
Canada, Air Pollution Control Directorate, April 1995.
4. 40 CFR Parts 52 and 98, Federal Implementation Plans to Reduce the Regional Transport of
Ozone; Proposed Rule,” October 21, 1998.
th
5. G.H. Conroy, “Low NOx Pyro-Systems Design and Operation,” 34 Annual IEEE Cement Industry
Technical Conference, Dallas, Texas, May 10-14, 1992, pp303-330.
6. Nielsen, P. B., and O.L. Jepsen. An overview of the Formation of SOx and NOx in Various
Pyroprossing Systems. Presented at the IEEE Cement Industry Technical Conference XXXII, Fl.
May 22-24, 1990.
7. Shumway, D.C. Mitsubishi Cement Corporation’s Cushenbury Plant presented at the IEEE West
Coast Cement Industry Conference. Victorville, CA Oct 1995.
8. Hansen, E.R., “The Use of Carbon Monoxide and Other Gases for Process Control,” IEEE
Transactions on Industry Applications, Vol. IA-22, NO. 2, March/April 1986.
Figures:
1. NOx Reduction on Nine Kilns Using Mid-Kiln Firing.
2. Stratification of Kiln Gasses and Mixing Illustrated
3. Emission Results With and Without Mixing
4. NOx Histogram With and Without Mixing
Examples:
1. Baseline Flame Temperature without Mid-kiln Fuel or Mixing
2. Flame Temperature with 15% Mid-kiln Fuel Substitution
3. Flame Temperature with 25% Mid-kiln Fuel Substitution
4. Flame Temperature with 25% Mid-kiln Fuel Substitution and 10% Mixing Air
General Practices
Working Group
Chair
R. Joseph Pospisil
Hercules Cement Company
Vice Chair
Detlev K. Babel
Humboldt Wedag Inc.
Members
Harley O. Biggs John P. Kline
Mitsubishi Cement Corp. Lafarge (DPC)
Robert K. Crittenden Allan S. Moore

Holcim Inc. Retired (Tilbury Cement Limited)
Ennio de Zorzi Edward S. Papa

Cementi Rossi S.P.A. CEMEX
Max L. Frailey Ted K. Richardson

Capitol Aggregates, Ltd. River Consulting, Inc.
Marc Fuentes Paul J. Riley

RHI Refractories America F.L.Smidth
Anya M. Gill David L. Rydquist

Polysius Corp. Dacotah Cement
D. Randall Jones R. J. Schmidt

Texas Industries Inc. ABB Raymond
Brian P. Keefe Raymond L. Worthington

F.L.Smidth BMH Americas Inc.
COMMISSIONING A 2.2 MT/Y CEMENT PLANT IN MIDLOTHIAN
TEXAS
Marc A. Wurman – Manager of Commissioning, FL Smidth Inc.

William Brown – Project Manager Cement, TXI
Abstract
TXI operates and maintains a cement plant in Midlothian, Texas. The original plant consisted of
four wet process lines with a yearly clinker production of about 1,088,400 mtpy (1,200,000 stpy).
The fuels utilized are natural gas and coal along with waste fuels. With the increasing demand for
cement in the Texas market, TXI decided to expand the plant capacity by adding a large dry
process clinker production line and shut down two of the existing wet kiln lines. Due to strict
environmental restrictions, the plant’s overall clinker production after the capacity increase had to
be achieved while maintaining the new stringent emission permit limits. This was a challenge as
the raw materials contain significant amounts of carbon and pyritic sulfur.
TXI recently commissioned equipment for a new 2,200,000 stpy (6000 stpd) dry clinker
production line, designed and supplied by FL Smidth Inc., that more than doubled the yearly
capacity of the existing facility. This modern clinker production line incorporates all of the latest
technology, in terms of chemical control, emissions control and abatement, and process features
that allows the plant to operate under these strict environmental conditions. This paper will cover
TXI’s recent expansion project and identify the relevant system design issues and operational
results.
Introduction
The TXI, Midlothian plant serves the greater Dallas-Ft-Worth area as well as SE Texas market.
With the rising Texas economy of the 1990’s and expected higher cement demand, TXI decided
to increase the plant capacity and modernize to a dry process system.
Not only were the environmental limitations of the new permit strict for a normal operating cement
plant, but the raw materials did not lend themselves to low emission operation. To accommodate
stringent NOx requirements, a low NOx ILC calciner design was chosen from FL Smidth Inc. Durr
Reeco supplied eleven RTO (Regenerative Thermal Oxidizer) units to eliminate the CO emission
and a wet scrubber was supplied by FLS Miljo to reduce the SOx emissions.
The commissioning of the new clinker production line at the TXI, Midlothian plant was a success.
As with most projects, during the commissioning phase there were a few obstacles to overcome
but the final result is an efficient, reliable plant that is now operating continuously at or above the
designed rated capacity.
The following presents area-by-area descriptions of the process equipment and performance
results to date.
• Stacking and Reclaiming:
Make: FLS Circular Storage

Type: CHO–IH–OF-1388-868 / 102
Mode of stacking: Continuous Chevron
Capacity: Stacking – 1451mtph (1600 stph), Reclaiming – 907 mtph (1000 stph)
FIGURE 1: Stacker / Reclaimer
In order to be able to handle the increased raw material flow from the quarry and eliminate the
older, maintenance intensive raw materials storage and reclaiming equipment, the entire system
was replaced in the expansion project with an FLS Circular Storage system as seen in FIGURE
1. A new, 54,420 mt (60,000 ST) Circular Homogenizing Store (CHO) for the limestone was
erected to store the limestone for all three kilns. The CHO is designed for a stacking rate of 1451
mtph (1600 stph) and a reclaiming rate of 907 mtph (1000 stph).
More tooth racks were added to the harrow to further agitate the pile and improve the material
flow down the pile face. The reclaimer has achieved the rated capacity, but under normal
conditions runs at a much lower rate which is sufficient to sustain the operation of all three kilns.
The limestone is conveyed by belt conveyor from the reclaimer to either a new 454 mt (500 st)
limestone bin serving the new kiln line, or a 181 mt (200 st) bin that supplies the two remaining
wet kilns.
• Raw Grinding System:
Size: FRM 48 / 315

RAR52.5 Separator
Feed System: Center feed with Apron conveyor and seal bin
Capacity: 431mtph (475 stph) at 80% passing 75µ (200 mesh)
Motor Sizes: Table 3170 kW, Classifier 336 kW, Fan 4476 kW
The raw mill system was successfully started with an auxiliary air heater on 20 December 2000
and ran at a reduced capacity due to heat limitations. The mill was started for the first time with
pyroprocessing exhaust gases on 26 January 2001, easily reached and exceeded rated capacity
by 28 January 2001.
The raw material preparation for the new line is accomplished in a vertical roller mill size FRM
48/315, supplied with a 4250 Hp motor, and a high efficiency RAR 52.5 separator. The mill was
designed to grind 431mtph (475 stph) of raw material to 80% passing 75 µm (200 mesh). The mill
itself was oversized in terms of grinding power; however, a large size was chosen to insure that
the mill would be capable of taking in enough hot gas to dry the feed material which contains up
to 15% moisture. A combination of preheater gases, cooler exhaust gases and airheater gases
are utilized for drying in the mill.
In order for the grinding operation to match the drying capacity of the mill, the mill was supplied
with a dual hydraulic roller loading system. This would allow a greater turndown in the mill
grinding capacity as two of the rollers could be operated with a high specific grinding force. This is
required to crush any hard components, while the other two rollers can operate at a lower
pressure to insure that the mill does not waste grinding power.
The mill feed and airlock arrangement was specifically designed to handle the raw materials that
contain 15% moisture and are difficult to handle. As seen in FIGURE 2, a seal bin and center
feed system utilizing an apron conveyor with a spillage conveyor was supplied. The seal bin
feeding method was chosen as it normally improves mechanical reliability, reduces maintenance
and improves feeding performance as compared to a conventional triple flap arrangement. The
mill feed seal is accomplished by an enclosed apron feeder and seal bin arrangement. This
sealing arrangement has proven to provide an effective seal as long as the material level in the
bin is greater than about 4-meters in height. The total leakage across the raw mill was less than
11.0%. The quick response of the apron conveyor enhances smooth stable operation while the
material level in the bin minimizes false air infiltration. The feed material enters the mill via a
central feeding tube in the separator, as shown in FIGURE 3, so that the materials do not stick
and clog up the feed chute.
FIGURE 2:Seal Bin with Apron Conveyor FIGURE 3: Center feed Classifier arrangement
The commissioning of the mill system went well. TABLE 1 depicts the guaranteed levels, the
levels achieved during the performance test and typical operating levels.
TABLE 1: Raw Mill performance
Typical Performance
Parameter Guaranteed
Operating Test
Production – mtph (stph) 431 (475) 476 (525) 443 (488)
Product size-% passing 75 µ (200 mesh) 80 82 91.8
Product moisture - %H2O 1% 0.8 0.66
Guarantees
Lifetime of grinding roller – hours 12000
exceeded
Guarantees
Lifetime of table liners – hours 12000
exceeded
The mill easily achieved its guaranteed production and fineness shortly after the start up. Due to
an unexpected high content of coarse quartz, it was required to grind the raw meal finer than the
guarantees stipulated to improve the burnability. Further optimization of the grinding system was
necessary to achieve the low residue required to improve the raw meal burnability. Two changes
were carried out to reach the goals: The dam ring height was reduced from 2.3% of the table
diameter to 2.08% and the grinding force was increased. These two adjustments improved the
mill specific power consumption significantly. The specific power consumption was improved from
6.2 kWh/mt to 5.16 kWh/mt adjusted to 20% residue on 75µ. Even with the low residue required
the mill has been reliable and able to produce the desired capacity.
• Raw Meal Homogenizing CF Silo and Kiln Feed System:
Size: 25-meter diameter X 53-meters high

Storage Capacity: 27,210 mt (30,000 st)
Mode of operation: Continuous Flow silo with loss of weight kiln feed dosing system
Kiln feed system: Air Lift
A single, 30,000 st CF Homogenizing silo provides both storage

and blending of the kiln feed. The CF silo has seven extraction
points with different orifice sizes as seen in FIGURE 4. The meal
is extracted at different rates around the silo floor to enhance the
blending effect. The silo performance is excellent when the silo
level is maintained at normal operating levels. An official
blending effect test has not been carried out, as kiln feed
homogeneity has been excellent - the standard deviation of CaO
in the kiln feed has been less than 0.11.
The extracted raw meal is collected in a small Loss in Weight

bin, which serves as an on-line calibration bin. During the
emptying periods where no material is supplied to the tank (the
butterfly valve above the tank is closed) the flow rate from the
tank is measured directly as the loss in weight of the tank
content per unit time. The controller differentiates the weight
signal from the load cells and adjusts the position of the flow
control gate accordingly. FIGURE 4: CF Silo
During filling periods this principle cannot be applied because material is being supplied to the
tank (the butterfly valve above the tank is open). During filling periods the position of the flow
control gate is calculated from the requested flow rate based on a mathematical model that
involves the relationship between flow rate, flow gate position and tank content.
The kiln feed is conveyed to the top of the preheater by an airlift system. This system was
chosen for its reliability and minimal maintenance requirements.
The commissioning of the silo system and loss of weight system was successful. There have
been no extraction problems from the silo. The loss of weight system was commissioned and
calibrated in January and the accuracy has been excellent with kiln feed fluctuations less than
1.0%.
• Pyroprocessing System:
4-stage twin string preheater with Low-NOx ILC

Cyclone size for each string : Stage 1 2 x 4.3 m φ
Stage 2 6.0 m φ
Stage 3 6.0 m φ
Stage 4 6.0 m φ
ILC low NOx Calciner: 8.5 m φ x 25.0 m tall
Capacity: 5442 mtpd (6000 stpd)
Alkali Bypass: 40% of kiln gases max
Kiln: 5.0 m φ x 78 m long
Burner: Duoflex
Cooler: CFG / RFT Hybrid clinker cooler with CIS
Commissioning and Optimization of the Pyroprocessing System:

The pyroprocessing system was first commissioned at the end of January 2000. The
commissioning of the plant was successful but there were a few obstacles to overcome. Some of
the milestones reached throughout the commissioning are as follows:
• Fire on in the kiln – 17 January 2001

• First clinker produced – 23 January 2001
• First 24-hour continuous run – 28 January 2001
• Max production obtained – 6497stpd (108%)
Preheater:
The design of the pyroprocessing system and the downstream emission control equipment all
had to take into account the specific raw materials in conjunction with a very strict air permit. With
this in mind, a twin string, four stage preheater with a single Low NOx In-Line Calciner was
chosen and is illustrated in FIGURE 5. The raw materials have a high moisture content, thus
requiring a four stage preheater to supply sufficient heat for drying in the raw mill. The raw
materials available at TXI also have a high carbon and pyritic sulfur content (at times greater than
0.5% for both carbon and pyritic sulfur). This also provides several challenges to the system
design. As a large portion of the carbon combusts in the upper stages of the preheater, the
preheater top stage exit temperature is much higher than normal systems. The spray tower in
each of the downcomer ducts cools the gases for protection of the fan and downstream
equipment. The fact that the carbon is reacting with preheater gases containing only 3% O2
results in a significant amount of CO generation in the upper stages which is not attributed to the
fuel firing in the calciner. TABLE 2 shows a typical temperature, O2 and CO profile of the
preheater. Note that although the bottom stage exit gases contain only 8 PPM CO at 2.4% O2, a
significant increase in the CO concentration to nearly 2500 PPM is seen in the stage II exit gases
due to the combustion of the carbon in the feed at this point.
TABLE 2: Typical preheater profile
0
Location Temperature ( C) O2 (%) CO (PPM)
Kiln inlet 1230 3.4 8
Reduction Zone 1010
Calciner exit 933 2.4 145
Stage 4 average 924 2.4 8
The pyritic sulfur present in the kiln feed also decomposes in the top two stages of the preheater,
releasing heat and resulting in significantly elevated SO2 emissions from the preheater.
The difficult raw materials resulting in the high CO and SO2 emissions were anticipated and
accounted for in the original design of the plant. An innovative feature utilized at TXI is the
sm
patented Cemstar process, where slag is introduced directly into the kiln. Due to the chemical
composition of slag and the energy intensive nature of the steel production process, the material
requires little heat to convert it to cement clinker. The result is reduced fuel consumption and a
lower NOx emission per ton of clinker.
Low NOx ILC Calciner:

The ILC calciner measured 8.5M in diameter, and is in a low NOx configuration as illustrated in
FIGURE 6. All of the fuel is fired in the upper portion of the kiln riser duct, creating a reduction
zone which not only helps to destroy the NOx generated in the kiln, but also prevents formation of
“fuel NOx” from the combustion of the calciner fuel. The temperature in this area is controlled via
a feed split so that it will not be prone to overheating and clogging. The temperature in the
0
reduction zone has been maintained at about 1010 C, further contributing to the NOx reduction.
The NOx guarantees of 309 kg/hr are continuously achieved while maintaining an extremely low
CO emission from the bottom stage.
FIGURE 5: Twin String 4-Stage Low NOx Preheater FIGURE 6: Low NOx ILC Calciner
Clinker Cooler
The cooler has a 7-row CIS section, 13 rows of CFG and 62 rows of RFT. The 13 rows of CFG
alternate with the RFT. Only the stationary rows utilize the air beam technology. The hybrid type
cooler facilitates maintenance by providing more room in the undergrate for inspection and
repairs. This configuration eliminates moveable air beam rows and the need to supply any type of
flexible connections. The layout of the CIS section is illustrated in FIGURE 7 and the benefits are
as follows:
• improves reliability
• reduces the number of flexible connectors
• improves accessibility
• eliminates gaps between grate plates
• reduces amount of wear components
• reduces seal air requirements
FIGURE 7: CIS Section
The cooler operation and reliability has been problem free and there were no stoppages through
out the commissioning due to the cooler. The cooler loss and efficiency were satisfactory at:
• Total Cooler Loss 121 kcal / kg

• VDZ Cooler Loss 103 kcal / kg
• Standard Cooler Loss 97 kcal / kg
• Cooler efficiency 74%
At times the clinker was extremely fine due to periods of high C3S, high silica modulus or excess
sulfur. This fine clinker resulted in a high dust recirculation back to the kiln and the tertiary air
duct. The dust would build up and surge back into the cooler from the tertiary air take off. A
couple of air blasters were installed in the tertiary air take off to minimize surges of hot dust back
into the cooler. Another problem created by the fine clinker size was that a higher than normal
undergrate pressure was realized due to the high dust circulation increasing the load in the
cooler. With an improved chemical composition of the clinker these problems are minimized.
Chemistry:
The chemistry of the kiln feed presented the greatest challenge in obtaining the maximum
production rate of the kiln. The key issue regarding the chemistry was the burnability. The original
burnability testing done depicted an easy burning mix with 0.35% free CaO after 40 minutes in a
0
1427 C furnace. The latest burnability test performed on the raw mix yielded a very hard burning
0
mix with 3.11% free CaO after 40 minutes in a 1427 C furnace. There are a few reasons for the
hard burning mix:
Reasons for poor burnability characteristics:

• High silica modulus
• High coarse fraction of quartz
• Slag addition was lower than expected
• High C3S content
• Chemical variations in the limestone
The silica modulus (SM) was higher than desired due to changes in the limestone chemistry. As
the purity of the limestone increased more sand had to be added to the mix while the alumina and
iron was reduced. The slag contribution could not be increased because the resulting C3A in the
clinker would be too low. With a two component mix (limestone and sand) there was no way
around this problem. The SM would normally run ~2.7 or higher. This is a high SM but normally
not impossible to burn. Due to fluctuations in the limestone composition the silica sand input to
the raw mix had to be increased. This component is difficult to grind in the mill and it has been
proven that the coarse fraction of quartz has a detrimental affect on the burnability. The mix has
been running with nearly 5% of quartz retained on the 45µ sieve. To remedy the silica issues a
few actions were taken:
• Shale was added as a third component to the raw mix

• The silica requirement in the clinker was dropped slightly
• Improve the fineness of the raw mix to reduce the coarse fraction quartz from 5%
down to 3%.
FIGURE 8 illustrates the relationship between the silica modulus and the production rate of the
kiln over time.
Daily production Vs.SR .
270 3.50
3.40
250 3.30
3.20
230
3.10
Pr
od
3.00 SR
uct
ion 210 .
2.90
2.80
190
2.70
2.60
170
2.50
150 2.40
1/28 2/6 2/192/25 3/3 3/9 3/163/223/31 4/8 4/144/204/26 5/7 5/135/195/255/31 6/6 6/166/226/28 7/4 7/107/167/22 8/4 8/108/16
Pro SR
FIGURE 8: Relationship Between SM and Production
The slag addition through the Cemstar system has been lower than expected. Due to the
limitations on the silica content and C3A content of the clinker, the slag addition could not be
increased. The original raw mix expected 5.2% slag addition. The slag addition is normally about
3% - 3.5%. When the slag was increased over 4% the burnability improved significantly. The
addition of the shale component in the raw mix increased the C3A and allowed for more slag
addition, further improving burnability. The slag is very reactive; the higher the slag contribution
through the Cemstar process the easier the mix is to burn.
Occasionally a ring would build up at the kiln feed end due to high temperatures and excess
sulfur. It was determined if the stage 4 material SO3 would exceed 3.3% for an extended period of
time then the ring would develop. This problem was eliminated by maintaining a sufficient amount
of excess air in the kiln (>3.0% O2 ) to avoid reducing conditions. The alkali bypass percent was
adjusted to ensure the stage 4 material SO3 did not exceed the 3.3%.
When the raw meal is reactive and burnability is satisfactory the desired capacity is easily
achieved. To control the excess sulfur the alkali bypass had to run at 37% instead of the
expected bypass of 30%. Another 2.0 kcal / kg can be expected for each percent increase in the
bypass. With this expected increase in heat consumption the achieved value is within the
measuring tolerances. TABLE 3 depicts the performance guarantee test for a 30-day period.
TABLE 3: Pyroprocessing Performance
Pyroprocess Guaranteed Actual

Capacity (mtpd) 5442 5475
Heat consumption (kcal/kg) 726 at 37% bypass 736
Free Lime 1.2 1.0
• Emissions Equipment:
Flue Gas Desulfurization Plant: Wet Scrubber by FL Smidth Miljo

CO and VOC Destruction: 11 Regenerative Thermal Oxidizers by Durr Reeco
Due to the characteristics of the raw materials and the strict air permit, TXI proved to be a pioneer
in implementing new innovative technologies for emissions control of a cement plant. This is the
first cement plant to utilize an RTO system to reduce the CO emissions. A wet Scrubber was
installed to reduce SO2 emissions and to produce gypsum to be utilized in the finish grinding
system.
Both units are located after the main bagfilter fan, and must operate 100% of the time that the kiln
is in operation. The system must handle all of the process waste gas from the pyroprocessing
system, including the preheater, cooler vent and alkali bypass gases. The booster fan for the
system is capable of handling 1,000,000 ACFM of gases, and has a 6500 HP motor. With help
from the suppliers of the new technology the start up of these areas was successful. There was a
substantial learning curve involved for everyone; however, the actual start up of both areas was
not as difficult as anticipated. It has been realized that it is not just the operation of a cement plant
anymore, but the operation of a small chemical plant has been added.
Wet Scrubber system:

The supplied flue gas desulfurization (FGD) system is an open spray tower using limestone fines
and forced oxidation to remove SO2 from the gas stream and produce gypsum as the by-product.
A small stream of sludge containing fine impurities coming from the raw meal fines is purged from
the system and disposed of with the alkali bypass dust.
The FGD system is illustrated in FIGURE 9 and is composed of the following areas:
• Absorber
• Oxidation system
• Absorbent system
• Dewatering system
Absorber: The major SO2 removal occurs in the spray zone where the upward flowing flue gas
contacts the downward spraying alkaline slurry. There are four tiers of sprays but three layers of
tiers are expected to achieve the desired scrubbing. To remove SO2 from the gases, good
contact between the gas and slurry is essential. The recirculation pumps draw slurry from the
tank and pump it to the spray headers where it is distributed over the entire cross sectional area
of the tower. The slurry falls down through the tower and ends up in the tank again. The pH of the
spent slurry is lower because of SO2 absorption. The absorber slurry pH is maintained by adding
absorbent to the slurry. Two stages of mist eliminators are installed on the top of the spray zone
to recover entrained slurry and to minimize water droplets carried over to the RTO.
-2 -2
Oxidation system: The sulphite (SO3 ) generated in the absorber is oxidized to sulphate (SO4 ).
Part of the absorbed SO2 is oxidized naturally by the oxygen in the gases. To ensure complete
oxidation a forced oxidation system is provided. Nearly complete oxidation is required to ensure
an adequate gypsum quality. Two blowers were installed to supply air to the slurry.
Absorbent: The absorbent utilized is raw meal fines supplied from the baghouse catch and stored
in a small silo. These fines are mixed with process water to form slurry of about 30% suspended
solids. The absorbent slurry is stored in the absorbent mix/storage tank.
Dewatering system: Hydroclones are used as the primary dewatering device. The hydroclone
overflow is returned to the absorber to minimize water consumption. Part of the overflow goes to
the thickener tank to purge fine impurities. The hydroclone underflow is dewatered further by the
centrifuges.
Performance of FGD: The scrubbing efficiency of the absorber is excellent and the desired SO2
emissions are comfortably maintained regardless if the mill is running or is down. The main
challenge faced by the system is that the raw meal fines have impurities (VOC’s etc) that act as a
crystal habit modifier (CHM). The shapes of the gypsum crystals become needle like and have
created dewatering problems for the gypsum.
FIGURE 9: Simplified Layout of the Absorber System
An extensive testing program was completed, and the solution is to input a small amount of a
chemical additive to the system to negate the effects of the CHM while maintaining the raw meal
dust as the main absorbent. TABLE 4 summarizes the results of the testing program.
TABLE 4: Summary of Absorber Testing
Effect on
Operational gypsum
Absorbent Effect on dewatering
parameter crystal
structure
Poor hydroclone performance.
Absorber slurry High specific surface area of
Slurry solids Raw meal Thin, needle
density (10 – needle like crystals does not
concentration dust like crystals
20% solids allow for effective filtration in
centrifuge – high gypsum H 2O
Raw meal Thin, needle Poor hydroclone performance.
Absorber pH pH (5.4 – 6.5)
dust like crystals High gypsum H 2O
Potato shaped, Good hydroclone performance.
Change of
Hydrated lime chunky Good filtration in Centrifuge –
absorbent
crystals low gypsum moisture
Cleburne Thin, needle Poor hydroclone performance.
Limestone like crystals High gypsum H 2O
(93% CaCO3)
Effect of Addition of
Raw meal Satisfactory Dewatering performance is
chemical chemical
dust crystals greatly improved
additive additive
Regenerative Thermal Oxidation (RTO) System:
The RTO system uses high temperature, residence time and turbulence to convert CO and
VOC’s to CO2 and H2O. The RTO’s utilize the heat content of the pollutant themselves to
contribute to the purification process thus minimizing fuel costs. The RTO system installed at TXI
consists of 11 individual RTO’s. Each RTO is connected to the stainless steel main inlet and
outlet headers by an individual inlet branch duct and outlet branch duct. Each inlet and outlet duct
is equipped with a pneumatically sealed and operated isolation valve, enabling each RTO to be
isolated from the main system.
The CO destruction efficiency of the RTO’s is excellent, consistently exceeding 90%. The CO
loading into the RTO units was at times double what was expected due to the higher than
anticipated carbon content in the raw materials. However, the destruction efficiency was
adequate to keep the plant emissions in compliance. The destruction efficiency is dependent on
0
the burner chamber temperature. By increasing the burner chamber temperature from 815 C to
0
900 C the CO emissions can be reduced by 10% - 20%. The CO destruction efficiency is also
3
dependent on the specific gas flow rate (m /hr / RTO). Thus with a higher gas flow rate per RTO,
such as when the bleed air damper is open or an RTO goes off line, higher CO emissions result.
The emissions equipment is performing well and all of the FLSmidth Inc. guarantees for
emissions are comfortably maintained in spite of a higher carbon and pyritic sulfur content in the
raw mix. TABLE 5 states the contractual preconditions and the actual for carbon, pyritic sulfur
and Nitrogen in the fuel.
TABLE 5: Preconditions for the Emissions Guarantee
Preconditions Actual
Carbon content in raw mix 0.32% 0.66%
Pyritic sulfur in raw mix 0.42% 0.45%
Nitrogen in fuel 1.39% 1.38%
The TXI emissions guarantees for CO, NOx and SO2 are depicted in the TABLE 6.
TABLE 6: GuaranteesFfor Emissions
Compound Maximum Emission Rate

CO 41.9 kg / hr (0.37 lb/st)
NOx 309 kg / hr (2.71 lb/st)
SO2 150.7 kg / hr (1.33 lb/st)
• Coal Grinding System:

Size: FRM 21 / 42
RAKM 22.5 Separator
Feed System: Weigh feeder and rotary feeder
Capacity: 36 mtph (40 stph) with 15% residue on 75µ, HGI >60
Motor Sizes: Table 373 kW, Classifier 45 kW, Fan 373 kW
A FRM 21/42 coal mill is supplied to grind the coal to 15% residue on 75 µm (200 mesh). The
coal mill was first started on 20 January 2001 during the heat-up of the kiln. Testing time for the
coal mill was limited due to the small fine coal bin (45 Mt) and the fact that the calciner had no
alternative fuel. Some modifications were made to the mill early in the start up to stabilize
operation so it could support the kiln. The dam ring height was increased and a LVT type cone
was installed to increase the mill body velocity.
At this point the coal mill was easily able to support the kiln operation. Further optimization of the
coal mill entailed opening the louvre ring to reduce the system pressure drop to save on fan
power and reduction in the dam ring height to minimize the table power consumption. The louvre
ring velocity was lowered from greater than 80 m/s to about 70 m/s with no rejects. TABLE 7
depicts the coal mill performance.
TABLE 7: Coal Mill Operation after Dam Ring Height Reduction and
Louvre Ring Area Increased
Production rate 35.0 mtph

Residue on 75 µm (200 mesh) 15 %
Mill DP 530 mmH2O
Mill Specific power 10.7 kWh/mt
Mill fan specific power 8.8 kWh/mt
Louver velocity 70.0 m/s
The Hardgrove Index of the raw coal was often less than expected, ranging from 52 to 62. The
best way to handle such variations is to blend the raw coal when storing it.
The coal mill RAKM classifier is performing well with a Rosin-Rammler number greater than 1.0.
There are no coarse coal particles being fired, thereby ensuring better combustion and reduced
CO in the kiln and calciner. There is no residue on 300µ with 1.5% residue on 150µ and 15%
residue on 75µ.
• Finish Grinding System:

Size: 5.0m x 15.5 m UMS shell supported ball mill
N-3500 O-Sepa
Capacity: 145mtph (160 stph)
Motor Sizes: Mill 5595 kW, Classifier 298 kW
The commissioning of the finish grinding system was a success and the mill has proven to be
reliable and able to produce the desired production. TABLE 8 depicts the performance guarantee
results:
TABLE 8: Finish Mill Performance
Guaranteed Obtained
Capacity – mtph 145 148
2
Blaine – cm /gm 3600 3600
The finish mill and O-Sepa Separator system perform well and produce a quality cement. The
desired strengths are realized at a lower than guaranteed blaine value enabling the TXI
2 2
laboratory to lower the blaine control point from 3600 cm /gm to 3400 cm /gm and achieve a
production level of 156 mtph.
The performance of the O-Sepa separator is excellent as expected. The tromp curve yielded
typical high efficiency separator results. TABLE 9 summarizes the separator’s performance
according to the tromp curve.
TABLE 9: Tromp Curve Results
parameter Achieved comment

value
P70 / P30 (ratio of 70% A value of 1 depicts perfect separation. A
recovery to 30% 1.67 conventional separator has a ratio of 2 or
recovery) higher. 1.67 indicates good performance
P50 (cut size) 30 µ Normal
Apparent bypass Apparent bypass is an indication of leakage
of fine material into the tailings. Ideal
5%
separation this number is 0 but anything less
than 10 is considered very good.
E60µ (% recovery of Generally called the coarse end recovery
91% and should be greater than 90%
material of 60µ size)
WACL 127 Weighted average circulating load
RR - number Conventional separators perform with a RR
1.29 number less than one. 1.2 is normally
expected with an O-Sepa
Conclusion:
The plant supplied at TXI Midlothian consists of a complete cement production line from raw
material storage through finish grinding with additional equipment for emissions control. The plant
was successfully commissioned in a reasonable time frame. An integral part of the success was
the cooperation between TXI and FLS personnel. The equipment supplied was designed with the
latest technology and is capable of producing above the guaranteed production rates. The end
result is a reliable, state of the art cement production line, comfortably meeting all FLSmidth
guarantees including fuel consumption, production and emissions on a continuous basis.
SOLVING RAW MATERIAL CHALLENGES
By
Linda M. Hills, Senior Scientist; Vagn Johansen, Senior Principal Scientist; and
F. MacGregor Miller, Senior Principal Scientist
Construction Technology Laboratories, Inc.
Abstract
A comprehensive understanding of the chemical and physical aspects of raw material transformation into
clinker is an important foundation to increasing production, reducing costs, and improving quality at the
cement plant. Clinker formation relies on numerous kiln feed properties and pyroprocessing conditions.
Understanding the influential properties of the feed (chemistry, fineness, uniformity, and mineralogy) can
lead to improvements in its burnability and therefore in the efficiency of plant operations. The paper
discusses clinker formation, its relation to kiln feed properties, and the importance of optimizing
burnability by careful mix control, good mix homogeneity, and tailoring the burning process to the raw mix.
The burnability and kiln feed are discussed in relation to specific fuel consumption.
When the raw materials available make it hard to achieve burnability goals, the use of fluxes and/or
mineralizers may be helpful. Fluxes and mineralizers indirectly affect burnability by promoting clinker
phase formation to occur earlier; their effect on clinker formation and cement quality is briefly described.
Fluoride-containing compounds have proven to be the most effective mineralizers in cement clinkering
reactions. However, an excessive amount of fluoride may delay cement setting time.
Emphasis is placed on the effects of changes in raw material burnability and clinker formation on kiln
operations, finish mill productivity, and the properties of the resulting cement.
Introduction
Plant management is constantly under pressure to maintain or increase production rates and reduce
costs, without sacrificing quality. These three are challenging goals, especially achieving all three
simultaneously. Understanding the process of clinker manufacture - how raw materials are transformed
into clinker - is the key.
This transformation involves both chemical and physical processes. The microstructure of clinker is a
function of kiln feed properties and the pyroprocessing conditions. Knowing what happens where, why,
and how, can lead to optimizing production, costs, and quality. In turn, the clinker microstructure and
composition directly affect how the phases hydrate and contribute to cement performance in fresh and
hardened concrete.
Given its relationship to both efficiency of plant operations and cement performance, the importance of
clinker formation cannot be overestimated.
Clinker Formation
Stated simply, the chemical aspect of clinker formation is the combination of silica with calcium to produce
hydraulic compounds. However, the actual transformation of the raw materials into clinker involves a
combination of chemical and physical processes as the material passes through the kiln system. The
general process of clinker formation is described in Figure 1. The transformation concludes with the
primary clinker phases:
“Alite”: impure tricalcium silicate, generally termed C3S
“Belite”: impure dicalcium silicate, normally termed C2S
“Aluminate”: tricalcium aluminate, C3A
“Ferrite”: nominally tetracalcium aluminoferrite, C4AF
Figure 1. A simplified view of the clinkering reactions in a Portland cement kiln (Hills 2000)
To 700ºC H20
Water is lost from clay minerals. H 20
Dehydrated clay recrystallizes.
Some reactive silica may displace CO2 from CaCO3. Clay particle
Limestone particle
700-900ºC C02 C02

As calcination continues, free lime increases.
Calcination maintains feed temperature at around
850°C. Lower-limed aluminate and ferrite form. C02
900-1150ºC
Reactive silica combines with CaO to begin Free CaO
stages of C2S formation.
1150-1200ºC
When calcination is complete, temperature
increases rapidly.
Small belite crystals form from combination
of silicates and CaO.
Round belite
crystals
1200-1350ºC
Above 1250°C, liquid phase is formed.
Belite and free CaO form alite in the liquid.
Angular alite
crystals
1350-1450ºC
Belite crystals decrease in number, increase in
size.
Alite crystals increase in size and number.
Cooling
Upon cooling, the C3A and C4AF crystallize
from the liquid phase.
Lamellar structure appears in belite crystals.

Figure 2. Clinker silicate phases as observed under
a microscope (polished section with nital etch)
Alite (C3S) normally hexagonal crystals observed in

cross-section, 25-50µm in length.
Belite (C2S) normally 25-40µm rounded crystals with

multidirectional lamellae.
Figure 3. Phases in clinker interstitial

(polished section with nital etch)
Tricalcium aluminate (C3A): Observed in etched

and polished clinker as blue to gray angular crystals
in the interstitial. High alumina ratio will produce
greater amount of aluminate than ferrite.
Ferrite (C4AF) appears more reflective than C3A in

an etched polished surface
Figure 4. Clinker photomicrograph

demonstrates reaction between
belite and free CaO to form alite
(polished section with nital etch)
Free CaO Either properties of the raw

crystals materials, mix homogeneity, or the
pyroprocessing conditions did not
allow the complete transformation
to alite to take place.
Belite crystals
Improving Pyroprocessing Efficiency
The amount of energy needed to form clinker depends on various properties of the raw materials. Using
a burnability equation such as that shown below can identify the specific factors that influence burnability
at a particular plant, and provide the information needed to make appropriate improvements. This
approach can also be used to anticipate changes caused by the introduction of new raw materials.
Replacing raw materials with others because they are less expensive, or to correct the chemical
composition may change how difficult it is to burn the new kiln feed.
Raw Material Properties and Burnability

The raw materials used are designed and proportioned to provide the appropriate amounts of the various
clinker phases. Specific properties of the materials are critical to plant efficiency and energy
consumption, based on their “burnability”.
Burnability is the ease with which the raw materials are transformed into the desired clinker phases, and
is commonly measured by the amount of free (un-reacted) lime remaining in the clinker (low free lime
indicates an easy-to-burn mix). Ordinarily, the burnability is a measure of the ease of formation of alite
from belite and free lime, as formation of the other phases is normally much easier to achieve. The lower
the temperature at which the targeted free lime can be obtained, the better is the burnability of the kiln
feed.
The burnability of a raw mix is determined by its chemical composition, the mineralogy of its component
materials, and its fineness. There are a number of equations that relate the burnability of a feed to the
composition and fineness of the minerals (Fundal 1979, Petersen and Johansen 1979, Christensen 1979,
Ludwig 1973). One example is the burnability equation developed at F. L. Smidth (Theisen 1992),
provided below. The importance of a burnability equation is not to provide an exact value of the free lime
of clinker made from a given raw mixture; it is more important that it gives insight into the effect of raw
material properties on the resulting free lime (clinker burnability) on a relative basis.
CaO1400°C =[ 0.343(LSF-93) + 2.74(SR-2.3)] + [0.83Q45 + 0.10C125 +0.39R45]

Where:
CaO1400°C = is the free lime after burning for 30 minutes at 1400oC
LSF = %CaO/(2.8%SiO2+1.18%Al2O3+0.65%Fe2O3)
SR = % SiO2/(% Al2O3+% Fe2O3)
Q45 = % quartz grains coarser than 45µm
C125 = % calcite grains coarser than 125µm
R45 = % other acid-insoluble minerals, (e.g. feldspar) coarser than 45µm
The first part of the equation (first bracket) represents contribution from the chemical properties of the
raw mix. The LSF (lime saturation factor) represents the CaO of the mix, while the SR (silica ratio or
modulus) is related to the amount of liquid phase at the burning zone temperature. Decreasing the SR is
equivalent to increasing the amount of liquid phase; since the liquid is the transport medium for the
reactants, more liquid can transport more reactants during a given time and therefore, the burnability is
improved. Note the relative magnitude of the coefficients for LSF and SR; the role of the liquid for the
clinker reactions is important, not only for the chemical reactions but also for the formation of clinker
nodules (the physical aspect of clinker formation).
The second part of the burnability equation (second bracket) represents the contribution to the burnability
from the mineralogy and fineness of the raw mix. The percentages of coarse particles represented by
C125, Q45, and R45 are determined by the overall fineness of the raw mix on the one hand, and the
mineralogical properties of the raw materials on the other. Differences in the mineralogy of the mix
components leads to differences in grindability, which in turn will result in variations in the chemical
composition of the different size fractions of the raw mix. For example, since quartz is hard to grind, the
SiO2 content of the coarser fraction of the mix will normally be higher than that of the finer fraction. Note
that the quartz fineness is more significant than fineness of other materials, as indicated by its larger
coefficient in the equation.
The equation provides a “virtual burnability test” to analyze the factors affecting burnability and free lime
of the clinker. There is a relation between the free lime determined by the “virtual burnability test’ and the
free lime of the clinker. Although not identical, it can be assumed that the virtual burnability free lime will
correspond to a certain constant clinker free lime for a given constant kiln operation. Further, for a given
kiln and set of raw materials, if the virtual burnability free lime of the kiln feed changes and the kiln
process parameters remain the same, the clinker free lime will change in a way so the ratio between the
virtual free lime value and the clinker free lime remains the same as before the change. For instance if the
virtual free lime is 3.2 % and the average clinker free lime is 1.4 % for a given steady state of kiln
operation, the ratio is 2.3. This defines a burning index and is a measure for how the kiln in that state of
operation on average burns the kiln feed. Reasons for variations in burnability could be insufficient raw
mix control and/or homogenization of the kiln feed. The resulting clinker will vary in free lime content, and
the normal reaction from kiln operators is to burn harder and eventually operate the kiln on the “hot side”
to avoid large variations in the clinker free lime. One penalty that results is increased specific fuel
consumption.
An example of the relation between the burnability index and the specific fuel consumption is the result of
analysis of data from a 2000 mtpd kiln with planetary coolers, as shown in Figure 5. Based on averages
of kiln feed chemical composition and fineness, the virtual burnability free lime was determined and
together with the average clinker free lime for the same period, the burning index ratio calculated.
3.5
3.4
3.3
3.2
Fuel
Consumption 3.1
(MMbtu/ton)
3.0
2.9
1.5 2 2.5 3 3.5
Burning Index Ratio

(“virtual burnability”/clinker free lime)
Figure 5. Burning Index Ratio (virtual burnability free lime/clinker free lime)
Each point represents the average over a 4-week period. No corrections were made for kiln downtime during the 4-
week periods, which might explain the outliers.
An increasing burnability index means harder burning, resulting in higher specific fuel consumption.
Figure 5 can be used to analyze the effect of kiln feed homogeneity on the specific fuel consumption. As
an example, consider a kiln with average feed composition corresponding to LSF = 98 % and silica ratio,
SR = 2.8 (corresponding to about 66% C3S), with virtual burnability free lime of 3.2% and an average
clinker free lime 1.4%. For a steady state kiln operation and a constant homogeneous kiln feed, the
burning index is 3.2/1.4=2.3. In reality the kiln feed composition will vary between certain limits. In the
first example, assume the variations in LSF and SR are +/- 0.6% and +/- 0.07 respectively. Using the
equation for the virtual burnability test given above, these variations correspond to a variation of the
virtual burnability test free lime between 3.8 and 2.6%. For the steady state kiln operation referred to
above, with burning index = 2.3, this will result in clinker free lime between 1.7 and 1.1%. In the second
example let the kiln feed composition vary +/- 2% in LSF and +/- 0.23 in SR. This corresponds to
variations in the virtual burnability free lime between 4.5 and 1.9%, and for constant kiln operation a
variation of the clinker free lime between 2.0% and 0.8%. As a result of the less homogenous kiln feed in
the last example, the clinker free lime will vary within wider limits. If the operation is changed to harder
burning in order to bring the maximum clinker free lime down to the average value and narrow the
variation in free lime, the burning index will change from the original 2.3 to 4.5/1.4=3.2. From Figure 5
such a change corresponds to an increase of about 6% in specific fuel consumption.
Once again, an understanding of burnability and free lime content is important. There is a relation
between the “virtual burnability” free lime and the free lime of the clinker from the kiln. This burning index
ratio can be used as a process parameter. If the kiln is operated in steady state, the ratio is constant.
However, if the “virtual burnability” deteriorates (i.e. decrease in mix fineness or change in chemistry), and
the kiln is kept in the same steady state, then the clinker free lime will increase. The usual operator
reaction to an increase in clinker free lime is to burn harder. However, if an examination of the kiln feed
were performed to identify the cause for the increase in “virtual burnability” free lime, the corresponding
increase in clinker free lime would be anticipated. The kiln operators would therefore know what to
expect, and would not necessarily constantly operate the kiln “on the hot side”. The examples given
above also emphasize the importance of good raw mix control and kiln feed homogenization vis-à-vis the
specific fuel consumption.
Most often, the mineralogy and fineness change little compared to the chemical composition unless the
raw material sources are changed. Therefore, using the chemical contribution portion of the burnability
equation is helpful. This calculation will give an indication of feed burnability very quickly. An example of
chemical analysis and calculation of its contribution to burnability is below. Note that the trend in
burnability is more easily detected when looking at the calculated chemical contribution instead of the
individual factors of LSF and SR alone.
LSF SR Chemical contribution [0.343(LSF-93) + 2.74(SR-2.3)]

95.2 2.2 0.48
95.6 2.3 0.89
94.9 2.8 2.02
95.9 3.1 3.19
95.8 3.2 3.43
Fluxes and Mineralizers

The use of fluxes and mineralizers can promote clinker phase formation, and therefore improve efficiency
of the pyroprocessing system. By definition, a flux promotes a reaction by increasing the amount of liquid
at a given temperature; for example, Al2O3 and Fe2O3 are fluxes for the formation of alite. A mineralizer
promotes the formation of a particular solid phase through its incorporation in one or more of the solid
phases (normally either alite or belite).
A general view demonstrating the process of clinker formation using a flux or mineralizer, contrasted with
using neither is outlined in Figure 6.
Figure 6. Simplified clinkering reactions in a Portland cement kiln with no additions, with mineralizer, and with flux.
Clay particle Limestone particle
“Normal” clinkering Clinkering with mineralizer Clinkering with flux
Liquid phase formation occurs

earlier (at lower temperature).
Mineralizers promote earlier silicate phase Increased liquid phase allows

formation (some enhance liquid formation, more C3S (alite) to form.
some form intermediate phases).
Without mineralizer or flux, clinker With mineralizers and fluxes, desired clinker phases are formed earlier, and
phases are still forming in kiln. is discharged from kiln. The result: residence time in kiln is reduced, and
less energy is required.
Fluxes and mineralizers have been studied in laboratory experiments and in practice (Flint 1939,
Johansen and Christensen 1979, Surana and Joshi 1990, Erhard 1994, Altun 1999). A series of different
compounds have been used; however, the ones with the most practical application are fluorides.
Fluoride-containing mineralizers have been used since the late 1800’s (Bhatty 1996). Fluoride-containing
compounds such as CaF2, NaF, BaF2, and MgF2, are all effective primarily as mineralizers, although CaF2
has enjoyed the greatest use. Alkali- and alkaline-earth fluorosilicate salts like Na2SiF6 and MgSiF6 can
also achieve similar mineralizing effects (Lea 1971). The effect of 5 to 10% fluorspar on the clinkering of
two cement compositions was reported (Klemm 1976) and as expected, these massive additions resulted
in both a reduction in clinkering temperature and a consequent retardation of cement setting. However,
there was little effect on the 28-day strength of hydrated cements. Although fluoride-containing
compounds have been proven to be the most effective mineralizer in cement clinkering reactions, a
number of other possible fluxes and mineralizers have been studied. These have included gypsum and
sulfate mineral tailings, fluoro-gypsum, phosphorus-containing calcareous tailings in combination with
gypsum/fluorspar and others; all showed some effect either as fluxes or mineralizers but the studies did
not result in practical use.
The benefits of added fluoride in achieving desired free lime levels at lower temperatures are illustrated
by a study in which the same rather hard-burning raw mix was burned with and without the addition of
0.25% fluoride. In this particular case, the fluoride was added as a fluoride salt. The mixes were burned
at 1450°C (2640°F) for 60 minutes. The resulting free lime content of the control clinker was 5.05%, while
that of the clinker to which fluoride was added was 2.31%. A further demonstration of the benefits to
burnability of added fluoride was carried out using the so-called “sandwich” technique. In this technique,
two compacts are prepared. The first is calculated to be an alite/free lime mixture, while the second is
calculated to be an alite/belite mixture after clinkering. The compacts had the compositions shown in the
table below.
Reagent High Lime Compact, % Low Lime Compact, %

Calcium carbonate 78.9 74.3
Quartz (SiO2) 10.73 15.33
Alumina (Al2O3) 3.86 3.86
Ferric Oxide (Fe2O3) 1.93 1.93
Magnesium Hydroxide 3.05 3.05
Sodium Carbonate (Na2CO3) 0.51 0.51
Potassium Carbonate (K2CO3) 1.02 1.02
Alite Alite
Pairs of compacts were made, with and without added fluoride. The
individual compacts were burned at 1450°C for 30 minutes, cooled rapidly Belite Free Lime
to room temperature, and the free lime levels determined. The low lime
compacts had negligible free lime levels (<0.12%), and the high lime <0.12% 13%
compacts had free lime levels around 13%. The high lime and low lime free lime free lime
compact cylinders were placed in intimate mutual contact, and the
cylinders re-burned at 1500°C for one hour. Figure 7 demonstrates this
“sandwich” technique. Alite Alite
Alite
Belite Free
The resulting clinkers were cut, transverse to the joint, etched with Nital Lime
etchant, and examined under reflected light microscopy. The results,
shown in Figures 8 and 9, illustrate the broadening of the region near the
joint where no belite or free lime exists – i.e., where alite formation has
occurred between the high and low lime compacts. The much broader Figure 7. Diagrams
width in the fluoride-doped clinker (about twice as wide) shows how much demonstrate the clinker
faster alite forms in the presence of 0.25% fluoride. “sandwich” before (top) and
after (bottom) re-burning.
Figure 8. Clinker “sandwich” after re-burning Figure 9. Clinker “sandwich” with fluoride after re-burning
With regard to fluxes and mineralizers in general, Bhatty (1996) reports a survey regarding their use in
the cement industry. The fluxes and mineralizers referred to were fluoride-based (primarily fluorspar), and
other non-fluoride-based materials. The latter were industrial by-products or waste-derived materials
primarily used as corrective materials in the mix. They contained iron oxide and alumina and resulted in
decreasing of the silica ratio of the raw mix. This silica ratio reduction is equivalent to an increase of the
clinker melt content, and hence to an improvement of the burnability. The study indicates that cement
plants outside North America use fluxes and mineralizers with better results than do the North American
plants. In most cases improvements in burnability, kiln operation, energy savings, and cement quality
were noted. With the fluoride mineralizer, setting time was reported to increase in some cases.
Christensen (1980) suggested a mechanism for the effect of mineralizers that implies increased setting
time, which is in line with observations with using fluoride compounds.
One must be aware of “side-effects” of using some materials as fluxes or mineralizers. In North America,
the most commonly reported problem when using fluoride-based fluxes/mineralizers is the blockage of
preheaters caused by fines and volatiles. To prevent these blockages, it is recommended to tightly
control kiln temperature and the thermal profile to prevent overheating of the load and generation of
additional volatiles and fines. In other words, the burning zone temperature reduction that has now
become possible because of the use of mineralizers must be implemented.
Fluoride can also be a retarder in cement hydration if present at too high a level. In the example provided
below, a particular plant produced two Type I cements with differing fluoride levels; the fluoride ratio
varied by a factor of four based on the raw materials. The high-fluoride cement had an initial setting time
of 2.4 times longer than the low-fluoride cement. The conduction calorimeter, which measures the
hydration profile of cement pastes at constant temperature, gave a very graphic picture of the difference
in hydration of the two cements, as shown in Fig. 10.
High and Low Fluoride Cements
4.5
4
High Fluoride
3.5
Low Fluoride
3
mV
2.5
1.5
0.5
0
0 10 20 30 40 50
Time, Hours
Figure 10. Calorimeter Results of High and Low Fluoride Cements
The graphs demonstrate the delay in the C3S peak for the cement with high fluoride content. The
maximum hydration rate is delayed from about 9 hours to about 17 hours. It appears from these results
that the amount of fluoride present in the high-fluoride cement was too high for acceptable cement setting
behavior. The behavior of cements with optimum fluoride addition may be considered somewhat
analogous to that of cements hydrated at low temperatures, or to cements hydrated in the presence of
retarding admixtures. Although the early hydration is delayed, the ultimate hydration may indeed lead to
better late strengths (Moir 1982). It is not unusual for cements experiencing slight retardation in setting
time to develop 28-day strengths higher than those of their faster-setting counterparts.
Effects on Operations and Cement Performance

Changes in burnability and/or clinker formation can have important practical impact on what happens in
the kiln, the mill, and the product. If the mix is hard to burn, the operator will be obliged to increase the
burning zone temperature to achieve the desired free lime level. Hard burning will tend to cause low
clinker porosity, large crystals of alite, and often contributes to generation of dust and/or large clinker
balls, instead of good, nodular clinker. It also slows down the resulting cooling process, both because the
maximum temperature is higher, and because the low porosity clinker is more difficult to cool. This can
often result in generation of cement with reduced strength potential and increased water demand. Clinker
temperatures exiting the cooler may increase, further increasing fuel consumption and presenting
handling problems. The high temperature conditions may lead to reductions in clinker alkali and sulfate
level, color variations, and increases in water demand attributable to increased levels of aluminate. The
reduced clinker alkali and sulfate will affect setting time, decrease early strength and increase later-age
strength; and fresh concrete might develop “admixture incompatibility” and changes in its rheological
behavior.
Conclusion
Specific properties of cement raw materials are critical to plant efficiency and energy consumption, based
on their “burnability”. This involves the chemical properties of the raw mix, along with the mineralogy and
fineness of the materials. Determining a “virtual burnability” using these parameters identifies the critical
areas where changes could be made to improve burnability and thereby decrease fuel consumption.
When the quarry and the raw materials available make it hard to achieve burnability goals, the judicious
use of fluxes and/or mineralizers may be a prudent measure. Fluxes and mineralizers promote clinker
phase formation to occur earlier, thereby improving the efficiency of the pyroprocessing system. Fluoride-
containing compounds appear to be the most effective mineralizer in cement clinkering reactions. An
experimental study demonstrated quicker alite formation in the presence of 0.25% fluoride.
Changes in raw materials and/or clinker formation can influence what happens through the kiln, the mill,
all the way through to the cement performance.
This paper demonstrates the importance of optimizing burnability and clinker formation by careful mix
control, good mix homogeneity, and tailoring the burning process to the raw mix. Real energy savings,
potential production increases in kiln and mill, and product enhancement are the rewards for diligently
following these principles.
Acknowledgements
The authors appreciate the contributions of Javed Bhatty, Kenneth Hooker, Alex Mishulovich, Laura
Powers, Rick Stevenson, and Fulvio Tang.
References
Akin Altun, I., “Effect of CaF2 and MgO on Sintering of Cement Clinker”, Cement and Concrete Research
29, 1999, pages 1847-1850.
Bhatty, J. I., Use of Fluxes and Mineralizers in the Cement Industry: A Survey, Portland Cement
Association, R&D Serial No. 2045, 1996.
Christensen, N. H., “Burnability of Cement Raw Mixes at 1400oC. Part I: The Effect of Chemical
Composition; Part II: The Effect of Fineness”, Cement and Concrete Research, Vol. 9, No. 2 and No. 3,
1979
Christensen, N. H., and Johansen, V., “Chemical Effects on Ceramic Reactions”, Journal of the American
Ceramic Society, Vol. 62, No. 5-6, 1979.
Christensen, N. H., and Johansen, V., “Mineralizers and Fluxes in the Clinkering Process. II Kinetics
Effects on Alite Formation”, Proceedings of the 7th International Congress on the Chemistry of Cement,
Vol. 2, Theme I, Paris, 1980, pages I-1—I-5.
Erhard, H.S., Fluxes and Mineralizers in the Clinkering Process – Operational Results with Fluorides,
Portland Cement Association General Technical Committee Meeting, Seattle, 1994, 17 pages.
Flint, E.P., “Mineralizers in Cement”, Rock Products, 1939, pages 40-42.
Fundal, E., “The Burnability of Cement Raw Mixes”, World Cement Technology, July/August 1979.
Hills, L., “Clinker Formation and the Value of Microscopy”, Proceedings of the Twenty Second
International Conference on Cement Microscopy, Montreal, 2000, pages 1-11.
Johansen, V., and Christensen, N. H., “Rate of Formation of C3S in the System CaO-SiO2-Al2O3-MgO
with Addition of CaF2”, Cement and Concrete Research, Vol. 9, 1979, pages 1-6.
Klemm, W. A., and Skalny, J., “Mineralizers and Fluxes in Clinkering Process”, Cement Research
Progress, American Ceramic Society, Columbus, OH, 1976, pages 259-280.
Lea, F. M., The Chemistry of Cement and Concrete, Edward Arnold Ltd. publisher, 3rd Edition, 1971.
Ludwig, U., and Ruckensteiner, G., “Einflüsse auf die Brennbarkeit von Zementrohmehlen“,
Westdeutscher Verlag, Opladen, 1973.
Moir, C. K., “Mineralized High Alite Clinker”, World Cement, December 1982, pages 374-382.
Petersen, I. F., Johansen, V., “Burnability and Clinker Nodule Formation from a Statistical Point of View”,
Cement and Concrete Research, Vol. 9, 1979.
Surana, M.S. and Joshi, S.N., “Use of Mineralizers and Fluxes for Improved Clinkerization and
Conservation of Energy”, Zement-Kalk-Gips, No. 1, 1990, pages 43-47.
Theisen, K., “The Influence of Raw Mix Burnability on the Resulting Cement Clinker”, Proceedings of the
Fourteenth International Conference on Cement Microscopy, 1992, pages 74-88.
The Kosmosdale Expansion Project
By:
Alan Rowley
Cemex
Detlev Babel
Humboldt Wedag, Inc.
Presented at the
th
44 Cement Industry Technical Conference
May 2002 Jacksonville, Florida
Abstract:
Kosmos Cement Company, a partnership between Southdown (now CEMEX, responsible

for operation) and Lonestar Industries (now Heidelberger) decided in 1998 to increase the
clinker production of the Kosmosdale plant from 2500 stpd to 4700 stpd.
To achieve this capacity increase of almost 90%, extensive additions and modifications
had to be made in almost all manufacturing areas. These main areas were the Quarry
stockpile stacker and river load out conveying, a new higher capacity rock barge fleet,
upgrades to the limestone unloading conveying system, limestone storage, raw meal
transport, kiln feed system, pyro-process system, coal grinding, clinker handling, finish
grinding system, cement storage and barge loading system, an addition to the main
electrical substation and a new plant wide control PLC/HMI automation system.
The paper describes the many challenges to execute this unique project in tight and
restricted areas while at the same time maintain ongoing plant operation.
Only a three month general plant shut down was scheduled for demolition and
replacement of the existing pyro-process equipment and tie-ins of the various areas into
the existing systems.
Due to the time constraints this paper concentrates mainly on the pyro-process area.
Introduction:
The plant is located southwest of Louisville, KY with property on the Ohio River. The
plant receives limestone from its quarry operation at Battletown, KY, via daily barge
shipments. Battletown is approximately 40 miles down river from the plant. The plant
also ships product to its terminal network along the Ohio River and to distant points such
as Pittsburgh, PA, Nashville, TN, and Chicago, IL. Bulk shipments are also made via
truck and rail. Cement plant operations on the property date back to 1904. After extensive
study of various capacity increases to the plant, a capacity of 4,700 stpd was finalized
based on the plant’s ultimate raw grinding capacity.
The pyro- process section of the existing plant with an original installed capacity of 1950
stpd, had reached a capacity of 2500 stpd during the mid-nineties through various
improvements including a new ID fan. The system was an Allis Chalmers MIAG type
four-stage preheater, a three station rotary kiln (4.5m dia. X 67 m long) and an Allis
Chalmers two-section inclined/horizontal stroke cooler. This plant was commissioned in
1974.
Requests for proposals were sent out to the major equipment suppliers in February 98 and
after detailed evaluation, a Letter of Intent was issued to KHD Humboldt Wedag, Inc. for
supply of the main pyro-process equipment followed by a firm contract in January 99. A
new Fuller Co. supplied 6000 HP finish mill system originally planned for the Victorville
Plant was transferred to Kosmosdale to support the new clinker capacity.
2
Project Objective:
In major projects, the Project Management team is responsible for the development of a
set of design criteria which is approved by all stake-holders. A portion of the design
objectives developed for Kosmosdale is included below.
• Increase plant clinkering from 2500 stpd to 4700 stpd nominal capacity utilizing
the existing ID fan and main kiln baghouse.
• Increase quarry production, raw material handling and processing, clinker and
gypsum handling, finish grinding, cement transport and cement barge loading
capacity to complement projected clinker production capabilities and sales need.
• Minimize cement manufacturing costs, including raw materials, manpower,
energy consumption, transportation and maintenance costs.
• Minimize impact of construction on customers and operations.
• Consider impacts of kiln downtime due to existing tower and cooler replacement.
• Assume continued Type I/II clinker production.
• Limit specific NOx emission to 3.26 lbs per ton of clinker or less, and CO
emissions of 1.43 lbs per ton of clinker or less to prevent PSD triggers. Take
appropriate measures to limit additional TSP discharge.
• Limit total facility downtime to three month or less.
• Replace the three different current control systems within the plant to the
company PLC/HMI standard, plant wide.
• Where cost justified, make other improvements to the operation that reduce
operating costs in electrical power consumption.
Project scope:
After evaluating all plant sections considering the projected production of 4700 stpd plus
a further increase to 5000 stpd, the Project Team developed the following Scope of
Work for the project:
• Increase the quarry stockpile capacity and barge load out capacity to 1200 stph
• Replace the existing barge fleet with seven new 2400 ston capacity barges to
transport limestone
• Upgrade the river barge unloading system at Kosmosdale to 750 stph from 500
stph including improved transfer points and dust collection
• To reduce operating costs, provide a new limestone stacker and pile withdrawal
feeder to allow direct feeding of the secondary crusher system, thereby bypassing
the MIAG storage reclaimer except in time of wet or cold weather requirements
• Extend the limestone reclaim system to the new pile withdrawal point
• Upgrade the secondary crusher system and provide new electrical room and unit
substation. Automate the control of the secondary crusher system all the way
back to barge unloading system. Allow full control from the central control
room.
3
• Upgrade the crushed limestone /additive measuring and handling system to feed
three raw mill systems. Provide online mix control for limestone and all
additives with a cross-belt analyzer.
The pyro-process area had to be replaced in almost its entirety. This included:
• Innovative kiln feed system including feed weigh bin, impact flow meter and twin
belt-bucket elevators
• New four-stage calciner/low NOx preheater system
• Two integral spray towers for raw mill on/off conditions
• New low NOx and calciner coal burner
• New kiln inlet section and seal
• Replace 46 meters of kiln shell including a new #2 roller station and tire
• New outlet section and seal
• New kiln hood with tertiary duct take-off and tertiary air duct
• New grate clinker cooler with two reciprocating grates, two mechanical drives
and hammer mill type clinker crusher
• New main multi-channel kiln burner
• New high capacity clinker drag conveyor
• New heat exchanger, cooler vent baghouse and cooler vent ID fan
Furthermore, a refurbished RS 963 series coal grinding mill system with indirect firing
(dosing) system, 100 ton ground coal storage, 360 ton raw coal storage bin, CO
monitoring and CO2 fire suppression system was installed.
The balance of the project included the following:

• A new 6000 Hp finish mill with center drive and high capacity air separator/
baghouse system for cement cooling along with new clinker storage bin
• A new gypsum handling system for all three (one new and two existing) finish
mills
• A new 100,000 ton storage dome built on the river side of the highway along with
a 1000 stph belt conveying system and river barge loading system
• A new 15/20 MVA transformer and medium voltage gear along with four new
unit substation and electrical rooms for the new preheater / precalciner, finish
mill, cement storage dome and secondary crusher system
• A new plant wide PLC/HMI control system including significant field control and
electrical room re-wiring.
The simplified Pyroprocess flowsheet is shown on Fig.1.The main components are:

• An Envirocare Conditioning tower integrated into the downcomer.
• A KHD four–stage Calciner/LowNox Preheater with a main cyclone diameter
of 8.2m having a total pressure drop of 65 mbar
• Tertiary duct with dust separation cyclone off the kiln hood
• A KHD two-step, horizontal grate cooler with 91 square meters of grate
cooling area
• An Amerex twelve fan heat exchanger, and Amerex cooler vent baghouse
• TLT cooler vent fan
4
Fig. 1
Schedule:
The initial project schedule predicted an estimated pyro-process completion in May 2000
(18 months) with the finish mill and cement storage dome to follow in June and July
2000.
During the project, two months were lost and the schedule had to be extended. The first
month lost was due to allowing added time for the design-build contractors to fully
develop comprehensive bids. The second month was lost when construction had not
achieved its 60% completion status as precursor to the three month major shut down.
This resulted in a delay to the start of the outage till April 10, 2000.
Lastly, in April 1999, five months into the project, and after a finite element analysis of
the existing kiln shell and stations, a decision was made to replace the number two roller
station including a new tire and 46 meters of kiln shell section during the major three
month outage. The project was well underway and this decision resulted in extending the
outage schedule due to the construction congestion and coordination required. The total
schedule to complete the pyro-process system was 20.5 months.
Fig. 12 of this paper provides an excellent picture of the congested situation around the
pyro-process area.
5
Cost:
The Kosmosdale Expansion project was approved by the Board of Directors of

Southdown and Lone Star Industries at a projected cost of $93,000,000. The new
incremental capacity was expected to result in an installed cost of $125 per ton-year of
new capacity with a total plant cement production capacity of 1.6 million tons per year.
Main project challenges:
Every project has a variety of challenges which have to be dealt with. In our case we
faced the challenge that we had to virtually
construct a new clinker production line in and
around an operating plant. A new line was ruled
out during the initial study due to a lack of space
and the need to reduce emissions from the existing
line as offsets for the new capacity. The plant shut
down period for the necessary tie-ins also had to
be kept to an absolute minimum. This was the
focus early on in the project between the project
team and the pyro-process supplier. By using the
planning and scheduling tool, P3, we determined
that building a new preheater tower adjacent to the
existing tower (Fig.2), and reusing the existing
preheater foundation and structure, the downtime
needed to demolish the existing tower and
reconstruct it to include the riser, new stage four and calciner would consume 12 weeks
of downtime.
The new structure would incorporate part of the new preheater system, the indirect
fired coal system and the kiln feed system. By taking this approach, all these systems
could be built in advance of the major shutdown. Fig. 2
During the shut down period, the existing preheater vessels down to the kiln inlet
chamber would be removed leaving just the steel structure skeleton. The remaining parts
of the new precalciner/preheater system were then installed in the skeleton structure
beginning with the new feed shelf.
Because of the decision to upgrade
the No. 2 tire station, the entire kiln
was also taken down extending the
outage two additional weeks. Of the
original 67 meter length, only 15
meters of the existing kiln shell were
reinstalled.
On the cooler end of the process (Fig.

3), the shutdown schedule dictated Fig. 2
the approach to the new cooler
installation. Working with the pyro-
Fig. 3
6
process vendor, an innovative approach to pre-assembling the cooler was developed that
resulted in this area completing within 11 weeks of starting the shutdown. Without this
cooperative effort between the vendor and the project, the final project scope would have
looked much different than what was constructed. During the shut down period we had to
open up the roof of the building remove the burner platform, demolish the existing cooler
and lift the new cooler in three major sections into the existing cooler pit.
At the same time the existing cooler vent system could be removed and the new heat
exchanger and the baghouse with vent fan could be installed.
A further project challenge was the fact that the construction had to take place not only
around an operating plant but also in extremely tight and restricted work areas (Fig. 4)
Fig. 4
The storage and preassemble area was about 0.5 miles away from the construction site
and with the exception of the clinker cooler each piece of equipment had to be
transported from the storage area over a live rail road track to the work area as needed.
This required a careful logistic and erection sequence planning including numerous pre
shutdown schedule/execution plan reviews culminating in a full project readiness review
meeting.
7
Preheater design:
The standard schematic of a 4-stage Calciner Preheater is shown in figure 5.
Fig. 5 Fig. 6
In order to comply with the previously mentioned requirements (use the existing tower
structure, minimum plant shut down) the preheater had to be split (Fig. 6).
After extensive discussion with the main equipment supplier the solution was developed
to install the double cyclone stage I and the main cyclones stage II and III in the new
tower structure. The inlet chamber, Calciner and stage IV cyclone will be installed during
the plant shut down in the existing tower structure after removal of the existing inlet
chamber and cyclones.
Special design consideration had to be given to the layout routing of the raw meal ducts
and the connection gas duct from stage IV to III.
The actual general arrangement of this unique preheater system is shown in Fig. 7.
The stage I, II and III are installed at the top of the tower which leaves enough space in
the lower part for the kiln feed system and the coal mill with fine coal bin and baghouse.
The plan square area of the new tower is almost the same as the existing tower.
8
Clinker Cooler Design:
Fig. 7
Fig. 8
Fig. 8 shows the actual assembly drawing of the clinker cooler. The cooler has a total
length of 26.5 m and an overall height at the inlet of 7.0 m with an installed grate area
of 91 m2, resulting in a specific loading of 47 mt/m2/day. The design requirements
were to be able to preassemble the lower part of the cooler completely on site during
the pre shut down period including the grates and drives. The design incorporated
split lines to break the cooler into three major sections which could then be lifted into
the existing burner building during the plant shut down. The upper side walls were
installed after the cooler was in its final position.
Erection Phases:
Phase I
This phase was the pre shut down period. During this phase the following tasks had to
be completed:
• Install all new conveying systems up to the tie in points
• Install the new finish mill system
• Erect the new preheater structure adjacent to the existing tower
• Install the coal mill system with fine coal bin and baghouse
• Install the kiln feed system
• Install stages I. II and III of the new preheater
• Preassemble the new clinker cooler adjacent to the existing burner building
9
Fig. 8 shows the progress new preheater erection alongside the existing preheater still in
operation. Fig. 9 shows the new preheater tower almost completed.
Fig. 8 Fig. 9
Fig. 10 shows the preassembled clinker cooler adjacent to the existing burner building.
Fig. 10
10
At the end of phase I the construction site looked like as shown in Fig. 11. The new
preheater structure was completed, stages
I, II and III were installed with all the
connecting duct work. The main crane
was moved into the new position and the
siding of the existing tower was taken off.
The new clinker cooler is assembled next
to the existing burner building and an
additional 450 ton crane is moving in to
handle the removal of the existing cooler
and the installation of the new cooler.
The site was ready for Phase II, the plant
shut down. However there is always
something happening unexpected. As
modifications had to be carried out to the
existing raw meal silo it had to be
emptied. But due to a wrong gate setting
raw meal was conveyed to the existing
preheater instead to the spare silo. About
600 tons of raw mix were dumped into
preheater system and had to be removed.
This clean up effort required 4 days to
correct.
Fig. 11
Phase II
This phase was the plant shut down period. During this phase the following tasks had to
be completed:
• Remove the cyclones, inlet chamber and duct work from the existing preheater
tower
• Install the new inlet chamber, stage IV cyclone and calciner
• Install connecting duct work and raw meal chutes
• Remove the existing cooler and install the new cooler
• Remove the existing cooler vent baghouse, fan and install the new heat exchanger,
• baghouse and vent fan
• Complete all conveying system tie ins.
• Complete the electrical installation and control system
During this phase the work schedule was two-10 hours shift per day, 6 days a week with
the seventh day available to perform critical path schedule work. The manhours to be
expended during this phase would consume fully one half of the plant construction
manhours.
Fig. 12 shows the actual situation on site during the plant shut down.
11
The existing cooler vent system is being removed, the burner building has been opened
up to remove the existing cooler, the kiln has been
cut, and the new cooler is ready to be broken down
into three sections. The whole construction area was
serviced by three main cranes. One was positioned
in the cooler vent area, the second one serviced the
burner building – kiln outlet area and the third crane
was used in the preheater area. Needless to say there
was no room for errors. Each crane movement had
to be monitored very carefully.
The existing preheater tower was stripped of all
equipment (cyclones, duct work etc.). Only the
structural steel skeleton remained (Fig. 13). Some
modifications to the structure were necessary and
some new support platforms had to be installed but
in general the new equipment was designed to fit
into the existing structure.
Although the design has been reviewed several times, we still experienced some
structural interferences especially at the cross over connections from the old tower to the
new tower.
As mentioned before, the cooler

could be broken down after
Fig. 13
preassembly into three major
sections. Fig. 14 shows the first
section weighing 63 tons being
lifted into the existing burner
building. All three sections were
lifted into the final position and
reassembled within three shifts
without any incidence.
(Fig. 15)
Fig. 15
12
Start Up:
Damper closed
Fig. 16
After the five day dry out period, the pyro-process system was started up in the S-
operation (standard start up Fig. 16).During this operation the start up burner, which
becomes later the LowNox burner is used in conjunction with the main kiln burner to
increase the degree of calcinations of the raw mix in the preheater. The tertiary duct
damper is closed as the tertiary air is at this time below 700 °C and all the combustion air
required for the start up burner has to be drawn through the kiln. The raw mix splitter
gate is set to 100% meal distribution to the riser duct, which later on will be called the
LowNox zone.
After one week of operation, the plant experienced a major plug up in the tertiary duct
bend. Due to circumstances which are still unexplained to this day, raw meal entered the
vertical tertiary duct section and the built up was almost up to the transition piece. The
general assumption was that the splitter gate was activated inadvertently.
The cooler operation was unstable and a red river appeared. After discussions the silica
ratio was lowered in the raw mix from 3.1 to 2.9 and later to 2.8 and situation drastically
improved. The red river disappeared. The kiln production was increase to about 2600
stpd and once an acceptable temperature profile in kiln / preheater system had been
13
established (tertiary air app. 700 ˚C, stage 4 inlet app. 860 ˚C), the system was switched
to the R-operation (regular operation).
The R-operation is the regular preheater / calciner operation. The tertiary damper is
gradually opened up and the calciner burner starts up. The splitter gate under the stage III
cyclone is activated to divert raw meal to the calciner section. The feed to the riser duct
section decrease and the fuel to start up burner decreases just to the point to regulate the
stage IV temperature. The plant quickly reached the guaranteed production of 4700 stpd.
The next step was to optimize the system considering the NOx
emission and went into the LowNox operation mode (Fig. 17).
Splitter gate
Fig. 17
The LowNox mode is the operation to reduce the NOx emissions.
The NOx generated in the kiln is reduced by low temperature, fuel rich combustion in the
LowNOx zone (riser duct) of the calciner. The combustion is controlled to produce
sufficient CO to react with the kiln NOx to form N2 and O2 thus reducing the kiln NOx.
The LowNOx burner reduces the NOx produced in the kiln to near zero limited only by
the maximum allowable CO emissions.
At Kosmos we achieved a NOx emission of 2.31 lbs/st which is 70% below the
guaranteed value.
Start
Up
Sche
dule
14
The guaranteed production rate of 4700 stpd was reached about 7 weeks after start up. As
seen in the schedule, the time frame of the S-operation took longer then expected due to
the plug up in the tertiary duct and the final decision to lower the silica ratio in the raw
mix.
The kiln was pushed up to a 5000 stpd production three weeks after the guaranteed
production was reached.
Operating Results:
The above table shows the operating data before the plant expansion, the guaranteed data
and the actually achieved data. In all categories the achieved data exceeded the guaranty.
The only data which is higher is the clinker end temperature, however at the time of the
test the ambient temperature was about 80°F and not 68°F as stipulated in the contract.
Emission results:
15
At Kosmosdale, the new Low NOx precalciner/preheater has achieved a NOx emission of
2.31 lbs per ton clinker (70% of the guaranty max) and a CO emission level of 0.76 lbs
per clinker (52% of the guarantee max.) at a measured production rate of 5136 tons per
day ( 9.3% percent over guaranty).
Conclusion:
The Kosmosdale Expansion project was the largest undertaken by Southdown since
the No. 2 line at Victorville, CA in 1979-1981. The unique preheater structure has
changed not only the skyline of the plant, it has altered the competitive position of the
plant in the market place, one that will endure for many years to come.
• All project objectives were achieved including stringent environmental

emission levels for the modernized line
• A very complex project with many significant challenges were overcome by a
very dedicated project and plant team
• The guarantees were met and surpassed
• The plant is now one of the lowest cost operating facilities in the Cemex US
operations group having reduced their fixed and variable costs by over 40%!
• The project is a success by every measure, having a final installed cost of
$115 per ton of new capacity.
16
ENVIRONMENTAL CONSIDERATIONS FOR PREHEATER DESIGN
By:
Steven W. Miller
F.L.Smidth Inc.
Introduction
Gaseous emissions continue to play an increasingly larger role in the day to day operations of existing
cement plants as well as with the planning of new production lines. As we head into the future, it is
expected that emission limits will continue to become stricter. In the future, cement plants will need to
emit less ST/yr of emissions with higher ST/yr cement productions compared to current and past plant
operations. In addition, through the use of stack CEMs (Continuos Emission Monitors) and more frequent
stack testing, plants will be monitored even more closely to ensure that their emission levels are in
compliance with their permit.
In June 2002 the new MACT (Maximum Achievable Control Technology) rules will go into effect. These
new rules represent the tightest emissions regulations to date and will have an effect on new and
upgraded cement production lines hereafter. The MACT rules do not regulate NOx or SO2 and CO is only
regulated when firing waste fuels; however, each state regulates these emissions for a new cement
production line. It is important to note that states can mandate even lower limits than those given in the
MACT rule and add additional emission requirements. Based on the MACT rule and recent experience,
the emission regulations for major gaseous emissions as of June 2002 will be:
NO X → <1.4 kg/MT clinker (2.8 lb/ST clinker)

SO2 → <0.15-1.5 kg/MT clinker (0.3 – 3.0 lb/ST clinker)
CO → <0.5-1.0 kg/MT clinker (1.0 – 2.0 lb/ST clinker)
*
THC → <50 ppm (dry, corrected to 7% O2)
3
Dioxin/furan → <0.2 ngTEQ/NM dry, corrected to 7% O2 (with <204°C at the dust collector inlet)
* - without waste firing
Note that the ranges given for SO2 and CO reflect how values can change from state to state. In plant
upgrade scenarios, many plants will replace existing production line(s) with a modern, higher capacity
production line. In these cases, the upgraded plant will require lower specific emissions rates in kg/MT
clinker in order to maintain the same MT/yr of emissions as the original plant. This is done to prevent a
PSD (Prevention of Significant Deterioration) review of the plant’s permit, which can cause very long
project delays or even prevent the plant from getting a new permit. In some cases preventing a PSD
review can require emission levels that are much less than would be required for a greenfield plant.
Required NOx levels for upgrades often need to make the 1.4 kg/MT clinker limit regardless of the existing
plant emissions.
As the emission limits get tighter and tighter, it becomes critical to be able to accurately predict gaseous
emissions and incorporate the proper system design to meet permit requirements. A number of
possibilities exist to minimize gaseous emissions. These include systems designs, like the low NOx
calciner, and “tailpipe” designs, like the wet scrubber. Of course, some methods are much simpler and
less expensive than others. In some instances the viability of a new production line can depend on the
extent of added costs required to achieve environmental compliance.
This paper will focus on emissions of NOx , SO2, CO, and THC, and the various methods that can be used
to control these emissions within current limits.
Establishing The Need For Controls
The following are the basic steps that should be followed when planning a new production from a
gaseous emissions standpoint:
Step 1:
Establish the gaseous emission requirements for the new production line. This normally involves
interaction with the appropriate regulatory authority.
Step 2:
Establish the raw materials and fuel(s) to be used.
Step 3:
Perform a full chemical analysis of the raw materials and fuel(s). This includes pyrite and carbon analysis
of the raw mix components.
Step 4:
Based on the individual material chemistries, create the desired raw mix that conforms to the required
cement composition.
Step 5:
Perform a burnability on the raw mix to determine how hard it is to burn. This is achieved by heating a
raw mix sample in a furnace to 1400+°C and testing the free lime content at specific intervals.
Step 6:
Perform an off gas test on the raw mix to determine its gaseous emission potential. An example of an off
gas test is the one performed by F.L. Smidth Inc. at their laboratory. A three gram sample of raw mix is
heated to 1000°C in a furnace in small temperature increments. The atmospheric conditions in the
furnace are kept close to expected conditions in an actual preheater system, with 5% oxygen, 30%
carbon dioxide, and 65% nitrogen. During the test, the exit gas from the furnace is measured for ppm,
THC (measured as propane), ppm CH4, ppm CO, ppm SO2, ppm NO, volume % CO2, and volume % O2.
FIGURE 1: Graph of F.L.Smidth Inc. Off Gas Test Results

The result of this off gas test clearly depicts what is expected to happen with the kiln feed sample in the
preheater. In Figure 1 above, one can see that THC is evolving between 250-550°C, CO is evolving
between 300-700°C, SO2 is evolving between 450-550°, and some NO is given off from 400-700°C. The
general gas emission trends depicted by the laboratory off gas test have been confirmed on numerous
occasions with actual field testing.
From the above graph, the mass emissions of the various pollutants are calculated by summing the area
underneath each curve. Figure 2 below depicts the mass emission result of the graph depicted above.
With this summary table, it can be estimated how much carbon and sulfide sulfur are burning in each
individual stage for a given preheater with a given temperature profile, as well as predict CO, SO2, and
VOC emissions.
THC ppm CO ppm CH4 ppm SO2 ppm NO ppm CO2 % O2 %

Average 0.086 2.492 0.042 0.554 0.675 30.039 5.046
Peak 1.442 26.424 0.576 4.869 4.734 30.739 5.190
mg/kg feed 1.876E+1 3.432E+2 3.314E+0 1.745E+2 1.528E+2 6.502E+7 7.943E+6
THC CO CH4 SO2 NO CO2 O2

100 1.936E-2 5.831E-1 1.273E-1 0.000E+0 0.000E+0 5.667E+6 7.058E+5
150 1.936E-2 5.831E-1 1.780E-1 0.000E+0 0.000E+0 6.799E+6 8.458E+5
200 5.135E-2 5.831E-1 2.137E-1 0.000E+0 0.000E+0 7.806E+6 9.702E+5
250 6.072E-1 5.866E-1 2.448E-1 0.000E+0 0.000E+0 8.781E+6 1.090E+6
300 2.898E+0 2.142E+0 2.619E-1 0.000E+0 0.000E+0 9.766E+6 1.212E+6
350 7.479E+0 2.333E+1 3.437E-1 0.000E+0 1.449E-1 1.080E+7 1.339E+6
400 1.193E+1 7.934E+1 5.615E-1 0.000E+0 2.813E+0 1.191E+7 1.476E+6
450 1.484E+1 1.449E+2 1.104E+0 0.000E+0 5.530E+0 1.313E+7 1.625E+6
500 1.700E+1 2.173E+2 1.821E+0 1.592E+1 1.033E+1 1.448E+7 1.792E+6
550 1.828E+1 2.734E+2 2.547E+0 4.249E+1 2.305E+1 1.603E+7 1.980E+6
600 1.859E+1 3.077E+2 2.994E+0 4.521E+1 4.900E+1 1.780E+7 2.198E+6
650 1.859E+1 3.318E+2 3.160E+0 4.521E+1 7.132E+1 1.990E+7 2.454E+6
700 1.859E+1 3.422E+2 3.283E+0 4.521E+1 8.249E+1 2.242E+7 2.762E+6
750 1.859E+1 3.429E+2 3.314E+0 4.521E+1 8.774E+1 2.551E+7 3.141E+6
800 1.859E+1 3.429E+2 3.314E+0 4.521E+1 8.875E+1 2.929E+7 3.604E+6
850 1.859E+1 3.429E+2 3.314E+0 4.521E+1 8.979E+1 3.386E+7 4.163E+6
900 1.859E+1 3.429E+2 3.314E+0 6.048E+1 9.966E+1 3.976E+7 4.863E+6
950 1.859E+1 3.429E+2 3.314E+0 1.028E+2 1.109E+2 4.699E+7 5.733E+6
1000 1.859E+1 3.429E+2 3.314E+0 1.495E+2 1.325E+2 5.651E+7 6.897E+6
FIGURE 2: Summary Table of Off Gas Test Emissions

Step 7:
Based on all of the lab results, predict the gaseous emissions for the new production line.
Step 8:
Compare the predicted emissions versus the emission requirements. If all predicted emissions are less
than the required emissions, then no additional emission controls are required and the plant can have a
“standard” design. However, if one or more of the emissions are predicted to be above the required
emissions, then an analysis will need to be made to determine which of the control methods as indicated
in this paper will need to be utilized in order to meet the emission requirement.
The following sections cover NOx , SO2, CO, and THC. Included in each section are the parameters that
affect the emission of each pollutant as well as the control techniques that can be used to reduce
emissions.
NOx Emissions
In a modern calciner kiln system NOx emissions are

generated from two sources: thermal NOx and fuel
NOx . Thermal NOx forms at temperatures above
1300ºC and involves the reaction of nitrogen and
oxygen molecules in the combustion air. Thermal
NOx is produced only in the kiln burning zone, as
this is the only place hot enough to achieve this
reaction. The amount of thermal NOx produced is
related to fuel type, burning zone temperature, and
oxygen content. Different fuels tend to produce
different flame temperatures to maintain the desired
burning zone temperature. The higher the flame
temperature, the larger the amount of thermal NOx
that will be generated. Due to their typical flame
temperatures, oil, coal, and petcoke tend to produce
less thermal NOx than natural gas.
The rate of reaction for thermal NOx tends to

increase as the burning zone temperature
increases. Therefore, hard to burn raw mixes which
require a hotter burning zone will tend to generate
more thermal NOx than easier burning raw mixes.
A similar phenomenon occurs with higher kiln
3
loadings (in MTPD/M ). As the production is
pushed in a given size kiln, the backend NOx will
tend to increase.
FIGURE 3: NOx Equilibrium Diagram
The rate of reaction also increases with increasing oxygen content. Running the same kiln with a higher
backend oxygen level will result in a higher amount of thermal NOx being generated as shown in Figure 3.
Fuel NOx is generated primarily in the calciner by the combustion of nitrogen in the fuel. This nitrogen will
do one of three things: 1) react with another nitrogen atom to form N2 2) react with oxygen to form NOx or
3) react with a NOx molecule to form N2 and O2. The amount of fuel NOx is dependent on a number of
factors including fuel type, calciner type, and calciner temperature.
In a modern pyro-processing system, a number of possibilities exist to reduce the NOx emissions. They
- Burnability improvements
- Low NOx kiln burner
- Automatic kiln control
- Low NOx calciner
- SNCR (Selective Non-Catalytic Reduction)
Burnability Improvements
The amount of thermal NOx created in the kiln can be reduced by lowering the burning zone temperature.
This can be achieved by making chemical or physical changes to the raw mix. Chemical changes include
reducing the silica modulus or the lime saturation factor. A lower silica modulus results in a higher
percentage liquid, which promotes heat transfer in the kiln burning zone and improves burnability. A
reduction in the lime saturation factor reduces the potential C3S in the clinker, which results in a lower
burning zone temperature requirement.
Physical changes involve grinding the raw mix finer in the raw mill. With finer grinding, the burnability can
be improved by reducing the number of coarse particles in the raw mix. Coarse particles are normally
defined as silica particles + 45 micron or calcite particles + 125 micron. In many instances, due to raw
material availability, cement quality requirements, and/or raw grinding limitations, burnability
improvements are not always possible.
Low NOx Kiln Burner
The low NOx kiln burner is considered the standard today for modern pyroprocessing systems. Figure 4
depicts an example of the latest low NOx kiln burner design. The benefits of the low NOx burner have
been well documented in the literature and will not be covered in detail in this paper. It can be simplified
by stating that low NOx burners give the best possible capability for controlling kiln NOx emissions.
FIGURE 4: Duoflex™ Burner
Automatic Kiln Control
Automatic Kiln Control systems such as Fuzzy Logic and other expert systems can give a small
reduction in NOx over the long term. This is due to the system automatically controlling burning zone
temperature and backend NOx and O2 to maintain optimum combustion and operating conditions. This
reduces the peaks and valleys in the operation, which can lead to savings of up to 10% for long term NOx
emissions.
Low NOx Calciner
The low NOx calciner is the standard for all new pyroprocessing systems. With the current emission
regulations in place today, it would likely be impossible to build a new pyroprocessing system without
using a low NOx calciner. Low NOx calciners can be of in line or separate line variety. Each has its own
pros and cons. Figures 5 and 6 below show examples of low NOx calciners with in line and separate line
designs.
FIGURE 5: Low NOx ILC Calciner FIGURE 6: SLC-D-NOx™ Calciner
Both calciners use the concept of high temperature staged combustion to suppress the formation of fuel
NOx and reverse the thermal NOx created in the kiln. The low NOx ILC is applicable for most fuel and
feed combinations. The SLC-D-NOX is best suited for difficult to burn fuels such as petcoke and
anthracite due to its combination of having high oxygen and high temperature combustion.
The details of the design and function of low NOx calciners have been covered in other literature and will
not be covered in this paper. However, it can be stated that the low NOx ILC and SLC-D-NOX calciners
are capable of meeting today’s NOx emission standard for virtually any combination of fuel and feed. A
number of these systems are currently in operation today and have proven to have emissions well below
the current benchmark of 1.4 kgNOx /MT clinker.
SNCR
Selective Non-Catalytic Reduction of NOx is achieved by the injection of aqueous ammonia or urea into
the pyroprocessing system. In each case NH2 radicals are created which can enter into a number of
competing radical reactions. The two main simplified reactions are as follows:
(1) NH2 + NO → N2 + H2O

(2) NH2 + O2 → NO + H2O
The ideal is to inject the ammonia or urea in the system where reaction (1) will be the dominant reaction.
This optimum temperature window occurs between 900 and 1000°C. In a modern low NOx calciner
system, this temperature window is located in the riser duct and lower calciner areas. At temperatures
below this range the reaction rate of (1) becomes too slow and above 1000°C reaction (2) starts to
dominate. Two of the drawbacks of SNCR are the potential for a small increase in CO emissions and
ammonia slip. However, the latter item can be minimized by the proper injection of the ammonia or urea
(Figure 7).
FIGURE 7: Ammonia Injection
The cost associated with SNCR includes the handling and injection system for the ammonia water or urea
as well as the cost of the reactant (urea or ammonia water). This cost tends to be lower than for other
technologies such as the wet scrubber and the regenerative thermal oxidizer (RTO). BACT (Best
Available Control Technology) analyses by a couple of plants have found that the cost in dollars per
annual ton of NOx removal is close to the $3500 per annual ton reduced that many states consider to be
an affordable control technology.
To date there are no permanent SNCR installations in the U.S.; however, there are a few SNCR
installations in Europe where NOx emission standards are tougher than in the U.S. The effectiveness of
the low NOx calciner has precluded the necessity for SNCR in the U.S. so far, but as NOx emission
standards continue to get lower and lower it is possible that SNCR could make an appearance in the U.S.
in the future.
To date, there are no “tailpipe” NOx reducing technologies that have been proven to be feasible both
technically and economically in the cement industry. However, as technology continues to develop, the
possibility exists in the future that a “tailpipe” technology such as SCR (Selective Catalytic Reduction)
could be invented that would be viable for the cement industry.
SO2 Emissions
In a modern, dry process cement plant there are two sources of SO2 emissions - the preheater exit gas
and the kiln bypass exit gas. The only source of SO2 emissions from the preheater exit gas is pyritic (iron
sulfide) sulfur in the raw meal. This is due to the fact that the calciner virtually acts as a perfect scrubber
for SO2. All SO2 passing through the calciner is absorbed by the large amount of CaO entrained in the
gas.
Pyritic sulfur combusts to form SO2 in the preheater in the temperature range of 400-600°C. This typically
translates to stage 2 in four and five stage preheaters and stage 3 in a six stage preheater. As the SO2 is
formed in the preheater a certain amount of it will be absorbed onto the CaO and to a lesser extent the
CaCO3 in the raw meal. The presence of CaO at this point in the preheater tower is due to the fact that
the bottom stage cyclones do not have perfect collecting efficiency. A percentage of CaO formed in the
calciner will be carried up the tower with the gas where it can absorb SO2 formed from the pyritic sulfur.
Experience has shown that anywhere from 0-70% of the pyritic sulfur in the raw meal can escape the
preheater as SO2 emissions. The percentage of SO2 formed from pyritic sulfur that escapes depends on
the amount and form of the pyrite in the raw meal, the efficiency of the preheater stages, the number of
preheater stages, and the characteristics of the raw meal.
In a number of pyro systems a kiln bypass is required to either prevent buildups in the riser area or to
remove enough alkali to produce low alkali clinker. Sulfur that enters the system as SO3, as SO4, as
sulfur in the fuel, and as pyritic sulfur which is absorbed in the preheater will all enter into the kiln burning
zone. As the material enters the burning zone a certain percentage of the sulfur will evaporate as SO2
and exit the kiln with the rest of the combustion gases. The percentage of sulfur that evaporates in the
kiln burning zone can range from 25-75% and is a function of the following:
- sulfur to alkali ratio

- raw meal burnability
- excess air
As sulfur enters the burning zone, it will tend to react with alkali to create an alkali sulphate. The more
alkali there is for the sulfur to react with, the less the tendency will be for the sulfur to evaporate into the
kiln exit gas. In cases where there is a large excess of sulfur versus alkali, a higher percentage of sulfur
evaporation is normally experienced.
The harder the burnability of the raw meal,

the higher the burning zone temperature
1200
needs to be to make an acceptable level
of free lime in the clinker. As the
PPM SO 2 in Dedusted Bypass Gases
burnability and burning zone temperature 1000

increase, the percentage of sulfur that
evaporates in the kiln will also increase.
800
The amount of excess air used for the kiln
combustion can have a large influence on
the percentage of sulfur evaporated in the 600
kiln. Figure 8 shows this relationship. It
can be seen from this figure that at a low
excess air percentage the evaporation of 400
sulfur is greatly increased while with high
excess air percentages the sulfur 200
evaporation is greatly reduced. It has
been seen that when the kiln backend
oxygen is maintained at 3% or higher, 0
even larger excess amount of sulfur can 0 1 2 3 4
be handled without buildup problems in
the riser. The downside of operating with % O 2 at Kiln Exit
a higher backend oxygen is that the NOx
emission will increase.
FIGURE 8: Kiln Exit O 2 vs. Bypass SO2 Emission
When a pyro system has a bypass, a percentage of the kiln exit gas is removed and conditioned in the
bypass system. When the bypass gases are removed, the SO2 is still in the gaseous state. As the
bypass gas is quenched with air and further cooled with water or bleed air, a certain amount of the SO2
will be absorbed onto the dust carried in the bypass gas. This dust is caught in the bypass dust collector
and will normally contain between 6-10% SO3. Depending on the percentage of sulfur evaporated in the
kiln and the amount of sulfur absorbed by the bypass dust, the bypass SO2 emissions can normally range
between 0-3.5 lbSO2/STcl. In cases of an extreme excess of sulfur, it is possible to have even higher
bypass SO2 emissions.
In a modern pyroprocessing system, a number of possibilities exist to reduce the SO2 emissions. Each
has its own associated cost and effectiveness in reducing emissions. They include the following:
- Raw material changes

- D-SOx cyclone
- Hydrated lime addition
- Microlime injection in spray tower
- GSA (Gas Suspension Absorber)
- Wet scrubber
Raw Material Changes
As mentioned above, the emission of SO2 is mainly due to the nature of the raw materials and fuel.
Changing the source of minor raw mix additives or using selective quarrying for the limestone are both
ways of reducing the amount of pyritic sulfur in the raw mix. In plants that have a bypass system, a lower
sulfur fuel could be used to reduce the SO2 emissions out of the bypass.
Unfortunately, most plants do not have a lot of flexibility with their raw materials - especially their
limestone source. Most plants with significant pyritic sulfur levels have the pyrite in their limestone
source, and in many instances selective quarrying is not an option. When raw material changes are not
an option, other methods must be used to reduce the SO2 emissions.
D-SOX Cyclone
The D-SOx system is a means of achieving an SO2 emissions reduction in the

range of 25-30%. This level of reduction has been proven with the D-SOx
system in two plants in the U.S. A schematic of the D-SOx system can be seen
in Figure 9. The D-SOx system is designed to take advantage of the free lime
(CaO) that is created in the calciner. A portion of the calciner exit gas is taken
off of the calciner exit duct and goes up to the D-SOx cyclone which separates
most of the entrained dust from the gas. The captured dust is then fed to the
cyclone exit duct where the pyritic sulfur is converted to SO2. This would be
stage 2 for four and five stage preheaters and stage 3 for six stage preheaters.
The free lime in the calciner dust absorbs some of the SO2 to give a 25-30%
reduction. In order to achieve this reduction, a molar ratio of at least 5-6 free
lime to SO2 is required. This translates to approximately 5% of the calciner exit
gas being diverted to the D-SOx system for plants with a significant level of
pyrite.
The exit gas from the D-SOx system is returned to the outlet of the second
stage cyclone. This means that the gas taken to the D-SOx system is
effectively bypassing two to four stages of the preheater tower depending on
the total number of preheater stages. Since the gas moving through the D-SOx
system has an easier path than the gas that goes through the bottom cyclones,
the natural draft of the preheater can be used to pull the gas through the D-SOx
FIGURE 9: D-SOX
Cyclone
system. Therefore, no additional fan is required. In order to control the amount of gas taken through the
D-SOx system, a damper is located in the ducting after the D-SOx cyclone.
Hydrated Lime Addition
One of the simplest methods of reducing SO2 emissions is by injection of hydrated lime - Ca(OH)2 - into
the system. The amount of hydrated lime injected is typically in the range of a 3.0 - 5.0 molar ratio of lime
to sulfur dioxide to be absorbed. Experience has shown that the optimum place to inject the hydrated
lime is in the exit duct of the preheater stage where the pyritic sulfur is being converted to SO2. In four
and five stage preheaters this would be in the stage 2 exit duct, and in a six stage preheater this would be
in the stage 3 exit duct.
Lime addition has also been tried in other places including the spray tower and the raw mill; however, the
SO2 reduction efficiency has not been as good as with the preheater injection. Figure 10 shows the
results of a test that was run at a plant using hydrated lime injection in the preheater and in the spray
tower.
80
70
60
SO2 reduction %
50
Riser Stage 1
40
Conditioning Tower
30
20
10
0
0 1 2 3 4 5
Moles CaO in additive / Moles SO 2 in gas
FIGURE 10: SO 2 Reduction with Hydrated Lime Addition
The hydrated lime is stored in a bin and extracted through a rotary feeder and air sluice into the cyclone
exit duct. An alternate method would be to have the lime bin at ground level and pneumatically convey
the hydrated lime into the cyclone exit duct.
Depending on the level of pyritic sulfur in the raw mix, using hydrated lime injection can reduce the SO2
emissions in the range of 45-70%.
The major operating cost of using hydrated lime injection is the cost of the hydrated lime itself. A typical
cost for hydrated lime is $65-$75/ST.
The capital costs associated with the hydrated lime injection are fairly low. It primarily consists of a
pneumatic transport for the hydrated lime to the feed bin, the bin itself, a rotary feeder, and the chute from
the bin to the cyclone exit duct.
Microlime Injection in Spray Tower
Within the past couple of years, a company has introduced a system to reduce SO2 emissions by adding
microground lime into the spray tower. In order to increase the reactivity of the lime, it is microground to
approximately 5 microns before being added to the spray tower through the spray nozzles in slurry form.
In order to minimize the wear on the nozzles, they are made from a special wear resistant material.
The microground lime is added at a stoichiometric ratio in the range of 3.0-5.0:1. The scrubbing
efficiency of the system depends on how low the gas temperature is reduced in the spray tower (ie. how
much water is added). This is due to the lime’s SO2 scrubbing efficiency increasing as the water amount
in the gas increases, and the fact that a certain minimum amount of water must be injected with a given
amount of lime. The company claims scrubbing efficiencies up to 90% can be achieved in the raw mill
down condition when the preheater gas is cooled to 150°C. Lower scrubbing efficiency is achieved
during the raw mill running condition when the water spray (and therefore lime) in the tower must be
limited to insure the heat in the preheater gas is capable of the necessary drying requirement in the raw
mill. When evaluating SO2 emission requirements and the need for additional SO2 controls, it is important
to factor in the raw mill drying requirements while considering this system.
The additional operating cost associated with the microlime system includes the cost of the microground
lime and the added power and maintenance costs associated with the lime slurry handling equipment.
The microground lime could be created at site with a small slurry ball mill; however, it is likely more
economical to directly purchase the microground lime. The added capital cost for the microlime system
includes the costs for the microground lime handling equipment and the special, wear-resistant spray
nozzles.
 (GSA)
Gas Suspension Absorber
The Gas Suspension Absorber (GSA) is used to reduce the SO2 emissions in kiln bypass systems.
Figure 11 shows the layout of the GSA system. As with a traditional bypass system, the kiln gas is first
cooled in a quench chamber. However, in a traditional bypass this quenching is done by ambient air
while in a GSA system the quenching is done by recycled gas that is in the range of 150°C. The exit gas
from the quench chamber enters the bottom of the GSA, which is similar in appearance to an upside
down spray tower. At the bottom of the GSA where the gas enters, water is sprayed through a single
nozzle concurrently with the gas flow. At this same location bypass dust is also recirculated into the GSA.
Bypass dust has a high free lime content that makes it ideal for the absorption of SO2. This absorption is
catalyzed by the presence of water, so the exit temperature of the GSA is normally kept as low as
possible to give the greatest possible absorption without having condensation problems downstream.
The GSA is designed for a minimum retention time of 3.0 seconds, which gives adequate time for the gas
and suspended dust to interact and achieve the SO2 absorption.
FIGURE 11: Gas Suspension Absorber (GSA)

The GSA exit gas passes through a cyclone where most of the entrained dust is removed. The gas
exiting the cyclone goes to a GSA fan that is operated to maintain a constant volume through the GSA
tower. The dust collected in the cyclone in transported to a small dust bin. This bin has a constant speed
screw conveyor located on the top of the bin and a variable speed screw conveyor located on the bottom
of the bin. The dust extracted from the bottom of the bin is returned to the GSA tower through the screw
conveyor. By returning this dust to the GSA tower, it will have an additional opportunity to interact with
the bypass gas and absorb more SO2.
After the GSA fan, the gas is split between the bypass dust collector and a duct that recycles a part of the
gas back to the quench chamber. This recycled gas is used as the cooling air in the quench chamber
with a bleed air damper that can be opened if needed. By recycling a significant portion of the gas back
to the quench chamber, the bypass dust collector and dust collector fan are approximately half the size of
what they would normally be with a traditional bypass system.
The dust collected in the bypass dust collector is pneumatically transported to join the cyclone catch in
the dust bin. This allows the GSA to operate with the highest possible material recirculation. The screw
conveyor located on the top of the bin keeps a constant head of material in the bin with the material
removed by this screw conveyor being the sole waste product from the system.
The very first GSA system in a cement plant started up in a plant in Norway five years ago and has been
operating since. By creating a high dust circulating load around the GSA, it has been found that the final
SO3 content of the wasted dust can be as high as 25%. This is compared to the bypass dust wasted in a
traditional bypass system that has a typical SO3 content in the range of 6-10%.
Using a GSA in the kiln bypass system gives an SO2 reduction of 50-60% when the bypass dust is used
as the only scrubbing agent. In order to achieve an SO2 reduction up to 90%, lime slurry can also be
sprayed into the GSA. In this scenario the lime slurry would normally be delivered along with the water
spray through the single spray nozzle in the GSA.
Economic evaluations have shown that there is very little price difference between a GSA bypass system
without lime slurry injection and a traditional bypass system. The GSA system has extra costs associated
with having a cyclone, extra fan, and more transport equipment; however, the GSA system costs are
lower for the baghouse and baghouse fan. Due to the costs being similar for a GSA bypass and a
traditional bypass, a GSA system could be installed for systems that do not initially have a requirement for
SO2 reduction. This way the proper equipment will be in place in the future when the emission
requirements for SO2 become stricter.
Wet Scrubber
The wet scrubber gives the highest percentage reduction of SO2 emissions in the main stack. It is able to
achieve a reduction in SO2 emissions of 90-95%. The wet scrubber is a “tailpipe” technology and is
capable of handling gases from both the preheater and the kiln bypass system.
A schematic of a wet scrubber system is shown in Figure 12. It would normally be located downstream of
the main baghouse fan and operate under positive pressure. The scrubber itself consists of a large
pressure vessel with the gas inlet located toward the bottom of the scrubber. Within the scrubber are a
number of banks of spray nozzles that spray limestone slurry into the scrubber. The limestone slurry
reacts with the SO2 in the gas and cools the gases down close to the dewpoint temperature.
FIGURE 12: Wet Scrubber
As the limestone reacts with the SO2, calcium sulfite is formed and precipitates out to the wet bottom of
the scrubber. The bottom of the scrubber is a sump that is maintained at a constant level. As the calcium
sulfite precipitates into the sump a mechanical or air agitator is used to keep the slurry mixed and prevent
settling. In addition, a separate air stream is bubbled into the sump to provide oxygen for the calcium
sulphite to oxidize to calcium sulfate (gypsum). A pump removes the gypsum slurry from the sump for
processing. The slurry first passes through a hydroclone that does a coarse separation of water from the
gypsum. The hydroclone product is then taken through either a centrifuge or a vacuum filter where the
remainder of the gypsum is separated from the slurry. The gypsum removed from the slurry typically
contains in the range of 10-15% moisture and is of such quality that it can be used to replace some of the
purchased gypsum used in the finish grinding system.
After the water and gypsum have been separated, the water is returned to a tank where limestone or raw
mill baghouse dust is added to make the slurry that is introduced in the spray nozzles in the scrubber. In
this manner the scrubber operates without a continuous waste stream. However, a bleed valve is
normally included to remove some water from the system periodically in case there is an undesirable
buildup of materials in the water such as chlorides or alkalis.
Once the gas has passed through the spray nozzles it enters the mist eliminator. As the scrubber is
operating very close to the dewpoint, water and other materials can form a fine mist or aerosol. As the
gas passes through the mist eliminator, a majority of these mists and aerosols are removed. A
continuous water wash maintains a clean surface on the mist eliminator. After the mist eliminator the gas
exits the scrubber and goes to the stack.
As might be expected, the wet scrubber is by far the most expensive option for the reduction of SO2
emissions. Due to its high cost, there are essentially only two scenarios where a wet scrubber would be
used in a cement plant. First would be the case where a grass roots plant is being built and the state or
county regulations are such that a wet scrubber would be the only way that the SO2 emission requirement
could be made. It is essentially a situation of having a wet scrubber or not being allowed to build a plant.
The second condition is when an existing plant is building a new cement line or upgrading one of the
existing lines while possibly decommissioning older existing cement line(s). The way the current
permitting process works in the U.S., there is a big benefit in terms of saved time if the permit for the new
line could be achieved without triggering a PSD review. If the other above-mentioned SO2 reduction
methods do not make it, utilizing a wet scrubber may be the only way to stay within the existing permit
limits for SO2 - especially if there is a substantial increase in clinker capacity. Having to do a PSD review
can result in a long delay in the permitting process, and there is even the possibility that a new permit is
impossible to obtain. In this scenario the wet scrubber is a necessity in order to do the expansion on a
faster schedule or even make the difference whether the expansion can be made at all.
CO Emissions
In today’s modern pyroprocessing system CO emissions are generated from two possible sources:
organic carbon in the raw material and incomplete combustion in the calciner or kiln. The organic carbon
in the raw material can be further broken down into two distinct types of carbon: volatile carbon and fixed
carbon. With the exception of low carbon fly ashes and certain bottom ashes, most carbon found in raw
materials is of the volatile type. Typical raw materials that can contain significant amounts of carbon are
low grade limestone, shale, clay, fly ash, bottom ash, iron ore, and mill scale.
Volatile carbon will normally burn in the temperature range of 300-600°C, which can be found in the
middle stages of the preheater. In this area there is not enough oxygen and temperature for the carbon to
fully combust; therefore, CO is produced. As the combustion of carbon (or partial combustion to CO) is
an exothermic reaction, carbon combustion in the preheater will raise the temperature profile in the
preheater. The higher in the preheater the carbon combusts, the greater the stage 1 exit temperature will
increase. Fixed carbon will normally not combust until it is in or near the calciner where conditions are
more suitable for combustion. Therefore, fixed carbon will not contribute significantly to CO emissions.
Many cement plants today are using fly ash or bottom ash, which are waste products of the power
industry. These materials can contain anywhere from 5% to 25% carbon (and sometime even higher).
The percentage carbon is very dependent upon the combustion practices used by the individual power
generators. Fly ash and bottom ash can typically be used as a raw material additive in small
percentages (up to 12% of the total mix). Since this material is a byproduct of a combustion process, this
type of carbon tends to burn in the lower preheater stages and in the calciner. However, it has been seen
when using high carbon fly ashes and bottom ashes that a portion of the carbon will be volatile and will
burn in the middle preheater stages.
There are a couple of ways to reduce the CO emissions from a new pyroprocessing system. They
- Raw material changes

- Calciner type and retention time
- Feed high carbon raw mix component directly to calciner
- Regenerative thermal oxidizer (RTO)
Of course, regardless of whether any methods for CO reduction are employed, it is assumed that good
combustion practices are used during operation.
Raw Material Changes
In cases where the raw materials contain significant amounts of carbon, raw material adjustments can be
made to reduce the carbon content and ultimately the CO emissions. This is especially true if the carbon
is located in one or two of the minor components. In cases where the carbon is in the limestone, the
options are usually limited. Selective quarrying may be possible to reduce the average carbon content to
a certain extent, but it may not be enough in cases where the limestone contains a large amount of
volatile carbon.
Calciner Type And Retention Time
The CO emission from the calciner can be minimized by using the proper calciner design and residence
time. For normal fuels (coal, oil, and gas) this can be achieved by using an in line low NOx calciner as
shown in Figure 5 with a minimum calciner body residence time of 3.3 seconds. The calciner design
ensures the mixing and residence time needed for ensure good fuel burnout and minimum CO emissions.
For hard to burn fuels such as petcoke and anthracite it is recommended to use a separate line calciner,
as shown in Figure 6. This calciner operates at a high temperature in a high oxygen atmosphere to
ensure good fuel burnout.
By using the proper calciner type and residence time, a modern calciner is able to maintain a bottom
stage CO concentration less than 100 ppm.
Feed High Carbon Raw Mix Component Directly To Calciner
In cases where one of the minor raw mix components is the source of the carbon and CO emissions, a
viable solution is to feed this component directly to the calciner. Due to the high temperature in the
calciner, the carbon in the material will fully combust to give CO2 instead of partially combusting and
giving CO emissions. In cases where the high carbon material is a fly or bottom ash, the material can
normally be fed pneumatically into the calciner as received without any additional processing. In other
cases where the material is coarse (for example, normal ball mill or roller mill feed size), the material
would need to be ground separately in a ball mill before being fed to the calciner.
The additional cost for feeding a component directly to the calciner is small if it does not need to be
separately ground first. From an operating standpoint there is a small increase in fuel consumption and
preheater ID fan power because feeding the component to the calciner gives the same effect as using a
higher ash content coal for fuel. There is a slight increase in fuel consumption and stage I temperature as
the material to gas ratio is decreased within the preheater. The additional capital cost includes the
equipment used to measure and pneumatically convey the material to the calciner. In cases where the
material must be ground separately, there is the additional capital and operating costs associated with
running a small ball mill system. As an alternate to separate grindng, the main raw mill could be designed
slightly larger so that it could batch grind the high carbon component a couple of hours during the week.
Regenerative Thermal Oxidizer (RTO)
The RTO is a “tailpipe” technology that gives the highest possible CO emission reduction. In simple
terms, it is an afterburner in which the gas temperature is increased enough for the CO (and THC) in the
gas to achieve complete combustion. Figure 13 shows a typical RTO unit.
The gas enters the unit on one side where it goes through a ceramic media that preheats the gas. As the
gas flows to the top of the unit, it reaches the combustion area where fuel is fired to increase the gas
temperature to 800-850°C. The temperature is high enough at this level to ensure the combustion of CO
and other volatile carbon molecules (eg. THC). The gas then exits the combustion chamber and passes
back down through the ceramic media where some of the heat is absorbed from the gas before it exits
the RTO unit. The CO and THC reduction efficiency for an RTO is at least 98%.
RTO units are typically designed to handle a specific amount of gas flow. As the gas volume to be
handled increases, a number of same size RTO units are used in parallel to treat the gases. For
example, with the RTO in operation at TXI’s Midlothian, TX plant there is a total of 11 RTO units placed in
parallel.
FIGURE 13: Regenerative Thermal Oxidizer (RTO)
The capital and the operating costs for an RTO are substantial. The installation of an RTO costs millions
of dollars and the operating costs include the fuel fired in the RTO as well as the power consumption of
the fan used to draft the air through the RTO. The fan power consumption is substantial as the pressure
drop across an RTO is >600 mmwg. As with the wet scrubber for SO2, the RTO can be considered as a
last resort to achieve the CO emission requirements.
THC Emissions
Total Hydrocarbons (THC) will be regulated when the new MACT rules go into effect. At the present time
Volatile Organic Carbon (VOC) emissions are currently regulated in cement plants. VOC’S are the same
volatile carbon compounds as Total Hydrocarbons (THC) with the following compounds removed:
1. Methane, ethane, acetone, and inorganic compounds

2. Chlorinated compounds
3. Chlorofluorocarbons
4. Hydrofluorocarbons
5. Parachlorobenzotrifluoride (PCBTF)
6. Cyclic, branched or linear completely methylated siloxanes (VMS)
7. Perfluorocarbon compounds which fall into these classes: cyclic, branched, or linear,
completely fluorinated alkanes; cyclic, branched, or linear, completely fluorinated ethers with
no unsaturation; cyclic, branched, or linear, completely fluorinated tertiary amines with no
unsaturations; and sulfur containing perfluorocarbons with no unsaturations and with sulfur
bonds only to carbon and fluorine
The exact THC compounds not considered VOC’s may vary slightly from state to state.
THC emissions come from volatile carbon compounds found in the raw materials. Low grade limestones,
shales, clays, and mill scales can contain these volatile carbon compounds. These compounds burn at
very low temperatures (100-500°C) in the uppermost stages of the preheater.
As the source of THC emissions is solely the raw mix, control of THC emissions is achieved the same
way as CO emissions from the raw mix is controlled. The raw mix component contributing to the VOC
emission can be removed from the raw mix or fed directly to the calciner. In an extreme case an RTO
could be used.
Semi-Dry System
The semi-dry process can be considered the ultimate environmentally friendly system from the standpoint
of NOx , SO2, CO, and THC emissions. Figure 14 shows the schematic of the semi-dry system. Kiln feed
slurry is fed to the crusher/dryer where the material is flash dried using hot gas at approximately 880°C
from the bottom stage cyclone. The exit temperature from the crusher/dryer is controlled in the range of
150-180°C to ensure that the material exiting the crusher/dryer contains 1% moisture or less. The
material laden gas exiting the crusher/dryer enters a dedusting cyclone that removes most of the
entrained material. This material is then fed directly to the low NOx calciner. Since the material is flash
dried and then fed directly to the calciner, the carbon and pyrite-bearing materials have no chance to
produce CO, VOC, or SO2 emissions. In the calciner, the carbon will burn completely to form only CO2,
and the pyrite will burn to form SO2, which gets fully absorbed by the free lime in the calciner. In addition,
the semi-dry system can operate as either an in line or separate line low NOx to ensure a low NOx
emission. At the Lone Star Greencastle semi-dry plant that started up in 2000 (the first semi-dry system
in North America), emission of NOx , SO2, CO, and THC are all well below permit as well as new MACT
requirements.
FIGURE 14: Semi-Dry System with ILC Calciner
The capital cost of a semi-dry system is below that of a dry system, especially when modernizing an
existing wet plant. The downside is that the semi-dry pyroprocessing system will have a higher fuel and
power consumption than a dry system. In many cases the viability of a semi-dry system hinges on the
environmental requirements. For raw materials that can produce high levels of gaseous emissions, the
semi-dry system can be an attractive alternative to expensive “tailpipe” technologies such as the wet
scrubber and RTO. An economic analysis is required to determine the best solution for each individual
project.
Conclusion
As we head into the future in the cement industry, the environmental requirements will become stricter
and stricter. When evaluating a new cement production line or the upgradation of an existing plant, it is
imperative to incorporate the proper environmental controls into the system design. As the raw materials
are different for each plant, it is not fair to say that all plants must use the exact same environmental
controls. It is important to evaluate each plant on a case by case basis to determine the extent and type
of environmental control(s) required to meet the emission levels in the plant’s permit.
Finish Grinding with Vertical Roller Mills – Operating Data
Chris Oesch, Blaz Jurko

Loesche America, Inc.
Abstract
Vertical roller mills offer significant potential for savings when used for finish
grinding. Two recent installations in Peru and Dominican Republic are being
examined with specific data on cement production, product fineness, specific power
consumption and concrete strength. The vertical mill at Cementos Pacasmayo, S.A.
in Peru was commissioned in December 1999, and the mill at Cementos Nacionales
in Dominican Republic was commissioned in August 2000.
Installed production rates range from 40 stph to over 200 stph per mill. The variety
of sizes makes the vertical roller mill an attractive alternative to ball mills for
additional cement grinding capacity or as stand alone installations for modern cement
producers.
Picture 1
LM 56.2+2 Cement Mill at Cementos Pacasmayo, S.A., Peru (Left)
LM 56.2+2 Cement Mill at Cementos Nacionales, Dominican Republic (Right)
Brief History
Vertical roller mills have been widely used in cement industry for raw material
grinding since the early 1920’s, and are also applied to coal, petroleum coke and
mineral grinding. The major benefit in raw material grinding is significantly lower
specific power consumption and therefore cost savings as compared to the operation
with ball mills. Practically all modern cement plant installations include vertical
roller mills for raw material grinding.
The concept of using vertical roller mills for clinker grinding started in 1980’s, with
the 2+2 technology being patented and introduced in 1992. The 2+2 technology
represents significant breakthrough in the area mostly contested by conventional ball
mills, namely high Blaine product fineness. The 2+2 technology allows for much
finer product with significantly more stable mill operation. Prior to the 2+2
technology, vertical roller mills had problems when attempting to increase the roller
pressure in order to grind a finer end product. The material would evade the grinding
pressure under the rollers, thus creating increased vibration and difficult operating
conditions.
Vertical roller mills have evolved through the years and are currently being used at
various locations worldwide with over 20 cement plants utilizing the 2+2 technology
for grinding slag, OPC or mixed cement. The introduction of slag and pozzolana also
offers an excellent opportunity for vertical roller mill application since the process
directly includes drying, mixing and classification in one grinding unit.
Picture 2
Vertical Roller Mill with 2+2 technology
Vertical Roller Mill Specifications
LM 56.2+2
Mill Weight (incl. classifier, g/box, motor) 796.200 kg/ 1.751.640lbs
Mill Wear Elements Weight 47.500 kg/ 104.500lbs
Mill Table Diameter 5600mm/220inches
Number of Large Rollers 2
Number of Small Rollers 2
Mill Main Drive Motor 4.000kw/ 5.360HP
Typical Mill Block Plan (LxBxH) 17m x 17m x 20m/55.76ft x 55.76ft x
65.6ft
Table 1
Vertical Roller Mill Specifications
Typical Plant Layout
Picture 3
Plant Layout with 2+2 Technology
2+2 Patented Technology
The 2+2 (and 3+3) patented technology allows for much finer grinding by increasing
the stability of the grinding process. The concept was developed in order to reach
higher Blaine values as required by cement manufacturers.
The concept is based on two types of rollers, the large rollers being used for grinding
and the smaller rollers being used for the material bed preparation. The smaller
rollers prepare the grinding bed so that the grinding rollers can apply the increased
hydraulic pressure required to obtain finer grind. When attempting to grind to higher
Blaine values, the material tends to behave very much fluid-like, causing the grinding
bed to disintegrate and the mill to vibrate.
Picture 3
2+2 Technology with small and large rollers
The 2+2 technology allows for production rates of up to 200 stph and the 3+3
technology allows for capacities of over 300 stph in a single mill.
Dynamic Classifier
Another technical advancement that proved very useful for finish grinding with
vertical roller mills was the development of the static/dynamic classifier. It offers
significant improvement as compared with the previous generation of classifiers, also
when classifying high fineness materials.
Picture 5
LSKS High Efficiency Static Dynamic Classifier
Production Data
C. Pacasmayo, Peru C. Nacionales, Dominican R.

Production 170stph to 200stph 200stph
Product Fineness 3600 Bl. to 4800 Bl. 3700 Bl. to 3800 Bl.
Specific Consumption 18.5 to 22 kwh/st 29 to 32.5 kwh/st
(mill drive)
Specific Consumption 25.7 kwh/st
(mill, fan and classifier)
Compressive Strength 1d 1970 PSI 1780 PSI
3d 3585 2970
7d 4325 3840
28 d 5410 4450
Table 2
Production Data
Maintenance Costs
As a rule the maintenance costs of vertical roller mills are significantly lower than
those of ball mills. Comparison data from the past 2-3 years between the vertical
roller mills and the ball mills show a reduction of approximately 30%. The data from
C. Pacasmayo is significant since it represents direct comparison of ball mills that are
currently still in use and the vertical roller mill. In order to save costs, the customer
shifts as much production to their vertical roller mill as possible.
Typical Grinding Costs Ball Mill Vertical Roller Mill Difference

Kwh/st at 3800 Blaine 38.1 kwh/st 25.9 kwh/st 12.2 kwh/st
@ $0.05/kwh $1.91/st $1.29/st $0.61/st
Wear Costs $0.14/st $0.13/st $0.01/st
Total Costs $2.05/st $1.42/st $0.63/st
Based 1,000,000 tpy $2,050,000 $1,420,000 $630,000
Table 3
Typical Grinding Costs
Summary
Vertical roller mills represent an excellent opportunity for cement plants to save on
energy and maintenance costs.
• Cost Savings - The savings when implementing vertical roller mills are
generally more than 10 kwh per ton of ground cement which can result in
hundreds of thousands of dollars per year depending on the energy cost.
Rising energy costs have increased the interest in vertical roller applications
for finish grinding.
• Simple Operation - The flexibility of the vertical roller mill simplifies the
operation of the grinding plant.
• Quality - Quality of cement produced with vertical roller mills is equal or
surpasses the ball mills, notably when considering the early strength data.
CEMENT MILL SYSTEM UPGRADE
By:
Donald A. Longhurst, Sr. Mill Systems Engineer, Krupp Polysius Corp.
ABSTRACT
The following case study examines a range of existing techniques available which have been applied
over the past 15 years at the Arizona Portland Cement plant on three different grinding systems. Four
modifications to these existing systems were completed during this time, and two more modifications are
planned. Planning for future needs has allowed the maximized use of existing equipment, while
increasing both the quality and quantity of finished product to meet the requirements of today and
tomorrow. With possible modifications planned in the future, the finish grinding capacity of these three
systems will have been increased from approximately 74 mtph to 300 mtph through additions to, and the
upgrading of, existing equipment. While the capacity increased by 300%, the specific power requirement
decreased by 20%. This value is for the major pieces of equipment only, based mainly on full use of
installed power
FINISH GRINDING MILL SYSTEMS CASE STUDY
Consumer demands relating to both the quality and quantity of cement have both been on the rise over
the past 20 years. Over the same time period, there have been many improvements and innovations
made in the area of equipment used to produce finish cement from clinker. Fortunately, the advances
made are not on the scale of that made in the computer industry where we have almost reached the
stage of disposability. (Five function calculators and digital wristwatches run by a computer chip are
already there.) That being the case, many cement plants have the possibility to greatly increase their
finish cement grinding capacity while extending the usable life of existing equipment.
In this paper, the air separators discussed are classified as first, second, and third generation. The
second and third generation identified as high efficiency due to the performance level as compared to first
generation air separators. There is also a physical operating difference between the first generation air
separators and the second and third generations. In second and third generation air separators
(identified as high efficiency) air is passed through the separating zone and then through cyclones or a
bag house before re-entering the separating zone. This reduces re-introduction of the product to the
coarse fraction of material. In cases where cyclones are used with second and third generation air
separators, it is usually possible to re-use existing dust collection equipment in place from the original
separator. Where cooling is required, the separating air is discharged from the system, passing through
the separating zone only once. Figures 1-3 pictures are diagrams of three types of separators.
SECOND Separating Air Feed Material
GENERATION
AIR THIRD
FIRST SEPARATOR SECOND GENERATION
GENERATION Feed GENERATION HIGH Separating
Material Air
AIR F`eed
AIR
Material
EFFICIENCY
SEPARATOR SEPARATOR SEPARATOR
Air
Fines
Fines
Separating
Air
Oversize Material
Oversize Material Fines Oversize Material
In some cases, problems with insufficient de-hydration of the gypsum used in most cement products
developed with this arrangement. This was especially true in northern climates where the temperature of
the air may reach sub-freezing levels. In these cases, the problem has been overcome by re-circulating
the filtered air back through the separator in controlled quantities to retain heat in the grinding system.
This type of application was first provided for, although not installed on the D-3 milling system modified in
1988, one of the finish mill conversions discussed in the following examples. Figure 4 air recirculation
diagram.
SINGLE PASS
Recycle Air
SEPARATOR
WITH Exhaust Air
RECYCLE AIR
Fresh
Air Product
Grit
Recycle
The re-circulation of the separating air allows a degree of control over the product temperature, and the
temperature of the separator rejects returned to the mill. Since the de-hydration of the gypsum is a time
at temperature function, it is related not only to the actual temperature of the mill discharge, but also to
the circulation factor at which the system is operated. Higher circulation factors result in a greater
amount of heat loss through convection between the equipment in the system and the surrounding air.
Conversely, lower circulation factors help retain heat in the system. Through engineered system design
and equipment selection, this problem can be reduced. Actual system operation can determine the
lowest temperature for the mill discharge that can be used and still provide sufficient de-hydration of the
gypsum.
Another major piece of equipment that has contributed greatly to capacity increases in grinding circuits is
the High Pressure Grinding Roll. The specific power requirements of a High Pressure Grinding Roll for
grinding to the same fineness as a ball mil is generally in the range of 50% of that required for a ball mill.
Figure 5 High Pressure Grinding Roll.
HIGH PRESSURE GRINDING ROLL

Feed material
Motor Grinding rollers
Planetary gear
Shaft
Bearings
Frame Ground material

Hydraulic system
In many cases where a cool clinker is being ground, the heat for dehydration is obtained from the
inefficient use of power by the ball mill. Since the High Pressure Grinding Roll is more efficient than a ball
mill, there is less heat imparted to the material when passed through a High Pressure Grinding Roll for
dehydration of the gypsum. This further complicates the finish grinding system relative to drying needs.
(Fig 6).
FINISH MILL CAPACITY INCREASES
KWH/T MTPH
50.0 325
275
HPGR & COMBI-
225
40.0
HPGR & HES HES
175
COMBI-
125
HES & MILL
30.0 75
87
89
91
93
95
97
99
01
03
05
19
19
19
19
19
19
19
20
20
20
The enormous increase in grinding capacity achieved at the Arizona Portland Cement plant was
accomplished by modifications to three existing mill systems identified below as D-3, CM-7, and
D-2. Modifications made to these systems are presented in chorological order. This allows a
better illustration of both the implementation of new equipment and the changes in processing
systems as applied to finish cement over the past fifteen years.
D-3 Finish Mill 1988: Upgrade of existing finish mill with High Pressure Grinding Roll and High Efficiency
Separator to Hybrid Grinding System.
This was a typical closed circuit finish mill system (fig 7).
D-3 Mill
Separator
System Prior to De-Dust Vent
Upgrade
55 tph
(3800 cm2/g)
Product
Grit
Recycle
New feed was delivered to the feed end of the mill along with coarse rejects from the air separator.
Discharge from the mill was fed to an internal re-circulating air type separator by a bucket elevator.
Product and coarse rejects from the separator discharged by gravity. A mill exhaust fan was used to
provide for airflow though the mill to remove water vapor from the mill created by a water spray used for
cooling of the material being ground in the mill. To further reduce the cement temperature, the product
from the system was passed through a cement cooler using water as a heat convection medium.
This existing finish mill system was operating in the range of 55 stph with a fineness of 4000 cm2/g
consisting of:
§ One single compartment ball mill 4.41 Meters Diameter X 7.62 Meters long with spiral classifying
liners, Single pinion 2240 kW drive with 1.15 service factor (2575 kW)
§ One 18 ft. 4 in. Diameter Air Separator 450 hp (first generation air separator)
§ One mill exhaust fan with dust collector
§ One cement cooler
Modifications of this system resulted in a capacity increase of more than 100% to 115 stph with a
fineness of 3850 cm2/g. Conversion of the system to a Hybrid Grinding system was accomplished with
the installation of a High Pressure Grinding Roll and a second-generation single pass air separator (fig 8).
D-3 Mill
Conversion
1987
105 mtph
(3800 cm2/g)
The single pass air separator provided sufficient cooling of the product that allowed for operation of the
system without the existing cement cooler for most of the year. Only when the ambient temperature is
above 38 Deg C is the cement cooler required to produce product temperatures of less than 65 Deg C.
Following is a list of the equipment added to complete the conversion:
§ One High Pressure Grinding Roll with 1.7 Meter Diameter X .80 Meter wide rolls and 2 X 672 kW main
drive motors and reducers
§ One single pass High Efficiency Air Separator (second-generation air separator) with a 299 kW variable
speed drive on the separator rotor.
§ One High Efficiency Separator 400 kW exhaust fan, 238000 M3/hr –46 Mbar draft
§ One Static Separator 2.4-Meter Diameter (for product removal from mill vent air)
§ One Static Air Separator 93 kW exhaust fan, 51,000 M3/hr –46 Mbar draft
CM-7 Finish Mill 1995 (Figure 9):
CM-7 Mill
Separator
System Prior to De-Dust Vent
Upgrade
22 tph
(3800 cm2/g)
Product
Grit
Recycle
Upgrade of existing finish mill with High Efficiency Separator. Capacity prior to modification was 22 mtph
with a product fineness of 7.5% retained on 325 mesh, and a product surface area of 3800 cm2/g.
The initial objective was to increase the production rate to 32 mtph, with a reduction to less than 5% of
the product retained on 325 mesh. Secondary objective was to reduce the temperature of the product. It
was foreseen to convert the system to combi-grinding in the future and increase the capacity to 55 mtph.
In anticipation of this, the High Efficiency Separator installed was selected based on the higher capacity.
Prior to the conversion, a water based cement cooler was used to cool the material discharge from the
mill prior to feeding the material to the existing internal air re-circulating air separator. This cooler was
one of the restrictions in the system that limited grinding efficiency. Due to continual blockages in the
cooler, a smooth continuous operation was difficult to maintain. Since the capacity of the system was
eventually to be increased by almost 200 percent, air-cooling of the product by airflow only was felt to be
insufficient. The direction of material flow in the modified circuit was changed so that either a partial, or
total flow of the product could be passed through the existing cooler.
Conversion of CM-7 Mill and future addition Figure 10
CM-7 Mill
Conversion 1995
32 Mtph
(3800 cm2/g)
Feed
Initial Conversion Production Estimates
Existing Mill System High Efficiency Separator Combi-Grind System

With Grinding Aid Addition Modification
Total kW 1729 1934 3183
mtph 22.5 30 60
3800 cm2/g
Cost per mton

(Electrical $0.07/kW) $5.67 $4.52 $3.65
Capacity Increase 0% 33% 167%
Actual capacity achieved after the conversion was 32 mtph at 3800 cm2/g. Product quality also improved
dramatically. Tromp curve analyses indicated a decrease in the separator bypass from 50 percent to less
than 10%. The amount of residue on 325 mesh was reduced from 13% to less than 3%.
Further modification of this system is possible with the addition of a High Pressure Grinding Roll
which has been delayed indefinitely in favor of conversion of a raw grinding mill to finish grinding
(D-2 Mill circuit.
D-3 Finish Mill 1998 (Figure 11):
D-3 Mill Conversion

1998
115 mtph
(3800 cm2/g)
The dramatic improvement in the percent retained on 325 mesh versus surface area with the CM- 7 Ball
Mill circuit supplied the incentive to revisit the D-3 Finish Mill system. A decision was made to replace the
second-generation high efficiency separator with a new third generation high efficiency separator. The
main objective of replacing the existing second-generation air separator on this Hybrid Grinding system
was product quality. The secondary objective was a capacity increase.
The new separator was selected to match the existing fan and dust collector installed with the separator
being replaced. This was possible since there was sufficient capacity in the existing exhaust system to
use an oversized separator. Calculated capacity of the new separator was 20% higher than the
anticipated production level of the grinding system. The new separator was physically smaller than the
one being replaced; very little other than the supports for the separator required modification. Both
objectives were met.
The finished modification resulted in a reduction of the product percent retained on 325 mesh from 9% to
5%. Capacity of the system was increased by approximately 10% to 115 mtph at 3800 cm2/g.
D-2 Finish Mill 2000 (Figure 12 raw mill system):
D-2 Raw Mill

Separator
System Prior to De-Dust
Vent
Upgrade
Product
Grit
Recycle
A new Quadropol Roller Mill designed to meet raw material requirements for the plant well into
the future was commissioned in 1999. The ball mill circuit, which had supplied most of the raw
material requirements over the preceding 20 years, was no longer needed for this purpose.
Conversion of the existing raw mill to finish grinding with a High Efficiency Separator would be
able to augment cement production.
The D-2 Raw Mill Circuit consisted of;
§ One single compartment ball mill 4.73 Meters Diameter X 6.65 Meters long with spiral
classifying liners, Single pinion drive 2240 kW.
§ Two first generation air separators used for drying and separating
§ Auxiliary material feeding and transport equipment, including two mill discharge bucket
elevators.
The approach taken for converting the raw grinding system to finish grinding was fairly straightforward.
§ One new single pass High Efficiency Separator replaced the two first generation air
separators.
§ The existing mill discharge elevators were extended to meet layout demands.
§ The feed system to the mill was modified to supply clinker and additives for finish grinding.
§ A new high airflow feed inlet to the mill was installed
§ Dust collection for the system was modified and added for dry grinding.
§ Existing spiral classifying liners in the mill were not changed.
Capacity of the converted system was expected to produce 55 mtph of finish cement product at 3800
cm2/g. While this system has not yet been optimized, the expected capacity has been exceeded by 5%
(Figure 13)
D-2 Raw Mill

Conversion 2001
55 mtph
(3800 cm2/g)
A further consideration for selection of the auxiliary material transport equipment and the High Efficiency
Separator installed was for future expansion of the system. Plans to modify the existing D-3 Hybrid
grinding system to Combi-grinding are in progress, with the same conversion planned for the D-2 finish
grinding system.
D-3 Finish Mill 200?: The conversion of the existing Hybrid Grinding finish mill system to a Combi-
grinding Finish Mill System will require the placement of a separator between the High Pressure Grinding
Roll and the ball mill. (Figure 14)
STATIC SEPARATOR
Feed Separating air
Fines
Rejects
The purpose of the

conversion is to increase the efficiency of the system, as well as achieve another increase in
production capacity, and product quality. For this grinding stage, a static separator has been
chosen. At this point in the grinding process, the material is comparatively coarse and abrasive
tendencies increased.
With no moving parts, and a fairly coarse separation requirement (1700cm2/g – 2100 cm2/g) the static
separator is a simple and effective choice. The cake produced in the High Pressure Grinding Roll is both
de-agglomerated, and separated within the static separator with the product being directed to the ball mill
circuit, and the grits returned to the High Pressure Grinding Roll. Airflow within the static separator
allows for product size control and to some degree is independent of the flow through system fan. Totals
system airflow though the separator fan allows for variation in the cooling requirement without disrupting
particle sized selection. It is possible in a system of this design to provide too much cooling to allow for
proper dehydration of the gypsum in the ball mill. The quantity of heat per ton of material passed through
the ball mill is decreased substantially when pre-ground material is fed to it. (Figure 15)
D-3 Mill Future 200?

140 mtph (3800 cm2/g)
D-2 Finish Mill 200? : When the D-2 mill was converted from raw grinding to finish grinding with a High
Efficiency Separator addition, it was foreseen that another upgrade of the finish mill system would take
place. The separator added to the system, as well as the auxiliary equipment was sized for a future
addition of a High Pressure Grinding Roll and conversion of the system to Combi-grinding. The initial
conversion added 59 mtph of cement production to the plant; the conversion to Combi-grinding will add
another 95 mtph of cement production to the finish grinding capacity of the plant.
In most cases, the amount of air (and dust) that can be discharged from a plant site is limited. When this
is taken into account, there may be an insufficient amount of air discharged to cool the product to the
desired temperature. In addition to this problem, water use at the plant site may also be limited. In many
areas cooling ponds are not allowed, and in some cases cooling towers may not be possible to install due
to local restrictions. In these cases, closed circuit glycol based cooling systems have been used for
cooling the bearings on mills, fans, separators and other equipment at the plant. It is foreseen that the
water based cement coolers used in the past may also require this type of equipment to cool the water
used and re-circulate it. There are other advantages to this type of system as well as the obvious
ecological ones. Degradation of the cooling capacity of the cement cooler due to calcium and dust build-
up is substantially reduced. This also extends the lifetime of the associated pumps and pipe used in such
a system. (Figure 16)
D-2 Mill Future 200?

154 mtph (3800 cm2/g)
RELATIVE COSTS OF FINISH MILL SYSTEM UPGRADES
The variety of upgrade techniques covered in the three preceding mill systems does not cover all of the
possibilities of course. It is a wide enough spectrum from which some knowledge can be gained
respective to the relative cost of upgrading. Over a time period spanning fifteen (15) years, a numeric
value indicating the cost of the various upgrades would be meaningless without correction for the
inflationary value of the currency referred to. Adding to that confusion is the worldwide sourcing of
materials, and labor, relative to the equipment used in the conversion of the systems examined.
A comparison of the cost of new grinding systems to upgrades for the year the upgrades were performed
was used to develop the graph shown in Figure 17.
R e la tiv e C o s t o f S y s te m U p g ra d e
T o N e w S y s te m
0% 20% 40% 60% 80% 100%
D -3 1 9 8 8 59%
C M -7 1 9 9 5 100%
D -3 1 9 9 8 61%
D -2 2 0 0 1 36%
D -3 2 0 0 ? 116%
D -2 2 0 0 ? 61%
The cost per ton of increased production of finish cement by the system is given as a percentage of the
cost per ton of a new finish grinding system during the time period the upgrade was completed. In the
past few decades, the newer systems used for comparison were for much higher capacities as larger
more efficient mill systems have become available. This also had an effect on the cost per ton of finished
product, lowering the cost per ton for newer higher capacity systems. In order to obtain a comparative
cost relationship, the costs used to develop the graph were also adjusted for the actual capacity
examined in each upgrade. In some cases, the cost of the upgrade was the same or more than a new
system would have cost. In these cases, there were other considerations such as space available, and
improvements that may not have been included in a new system. Another aspect, which also affects this
comparison, is the availability of the higher capacity systems in past years.
The indication is clear that the cost of equipment relative to the total cost of a project has declined over
the past fifteen (15) years. Fig 18
COST OF EQUIPMENT
RELATIVE TO INSTALLED COST
40.0% 42.0% 44.0% 46.0% 48.0% 50.0% 52.0% 54.0%
YEAR
1987 53.0%
1998 49.0%
2002 45.3%
It would appear to be obvious that labor costs have been rising faster than material costs; this is not
necessarily true. Examining this concept in further detail, it must be considered that amount of existing
equipment used compared to new equipment in the resulting system has an even greater effect on the
ratio of equipment cost to installed cost. These differences and the cost of possible maintenance costs
being absorbed into the upgraded system have not been considered in the evaluation. The graph shown
on figure 19 contains figures depicting the equipment involved in the upgrades examined above alongside
a graphic representation of the cost breakdown for civil, electrical, and mechanical expenses.
RELATIVE COST OF UPGRADES
0% 20% 40% 60% 80% 100%
D-3 1988 21% 56% 23%
CM-7 1995 39% 42% 19%
D-3 1998 54% 42% 5%
D-2 2001 38% 51% 11%
D-3 200? 52% 27% 21%
D-2 200? 44% 44% 12%
Civil Mechanical Electrical
It should be noted here again that the ball mill shown for the D-2 Mill 2001 upgrade was an existing raw
mill re-used in place with minor changes. . What is very clear is that where the relative cost of new
equipment is low, there has been a major effort to re-use existing equipment. This results in a more
labor-intensive project increasing the relative costs of the civil and electrical portions of the upgrade and
reducing the relative cost for new equipment.
CONCLUSIONS
There is a great deal of economical potential capacity obtainable though the use of existing equipment.
As illustrated on the graphs, in some cases, the cost of upgrading a mill system can vary considerably,
and each case must be weighed on a different set of scales. In most cases the cost per ton of product
could be expected to be in the range of 40% to 60% of that for a new system. There are numerable other
items which must be considered along with the installed cost, and for each plant the value of one of those
items may have a different priority.
Considering all of the upgrades planned at the Arizona Portland Cement plant, savings in power alone
compared to the original 74 mtph will be almost 3,200 Kilowatts per hour for the new capacity of 300
mtph. With operating hours of 6,500 per year, the power savings over a range of 3cents to 10 cents per
kWh would amount to a savings of between $625,000 and $2,080,000 per year. This is quite and
ecological as well as economical bonus for the refinement of a process, and an achievement to be
commended.
INCREASING CEMENT GRINDING CAPACITY WITH VERTICAL
ROLLER MILL TECHNOLOGY
Gary R Roy
General Manager – Product Management
F.L.Smidth, Inc.
ABSTRACT
With the many large-scale plant expansions made in the US in recent years, there has been a need to
use new and complex burning and emission control technologies to comply with modern emission
regulations. However, in many cases these same plants have chosen to use very traditional technology
for finish grinding – typically a ball mill and separator. With today’s power price issues in the US, and the
large experience base that the rest of the world has with the use of new technology in finish grinding, the
US market is considering the use of modern Vertical Roller Mill technology for cement grinding. This
paper discusses the different options available for upgrading the capacity of an existing plant, provides
comparisons of operating costs, simplicity of operation, product quality considerations, etc., based on the
OK mill project at Phoenix Cement Company (PCC) as the first vertical mill for cement grinding in the US.
INTRODUCTION
When a new cement production facility is to be built, it is relatively simple to make a decision on what type
of technology should be incorporated. The decision will be made based on cost of the installation,
constructability, projected operating and maintenance costs, product quality issues, and customer
preference. For a new installation, all of these variables are fairly well defined and comparisons of one
technology vs another is fairly straightforward.
When an existing facility is investigating how to significantly increase the cement grinding capacity, there
are many different solutions that can be considered. However, it is important to realize that although the
goal of increasing production seems to be very straightforward from a technology standpoint, each retrofit
is a unique situation. The details of the existing facilities can often become the critical factor in making the
justification for one type of system vs another. The justification may then be additionally influenced by the
layout, re-use of existing equipment or other constraints as opposed to technology issues only. What is a
good solution for one installation may be a terrible solution for another.
INCREASING CAPACITY: PREGRINDING OR NEW MILL CIRCUITS
In order to significantly boost the grinding capacity of an existing facility, it requires that additional grinding
power be added. Depending on the situation, this additional grinding power can either take the form of a
pregrinding system attached to the existing grinding circuit, or a completely new grinding installation to
make up the desired additional capacity.
For cement grinding circuits, there are several options available for pregrinding. Depending on the type
of pregrinder and system selected, a 10 to >100% production increase can be expected. Pregrinding
solutions include vertical shaft impactors, hydraulic roller presses, and vertical roller mill pregrinders. Not
only do pregrinders add grinding power to the grinding circuit, but they also provide an opportunity to
increase the overall grinding efficiency of the circuit – so that not only is the capacity increased, but the
overall specific power consumption is reduced compared to the original ball mill circuit. Although many of
published materials are available in the literature regarding these technologies, a quick review should be
included as background for the case study related below.
Vertical Shaft Impactors (VSI’s) have been employed in the cement industry for pregrinding clinker in
order to boost the capacity of existing mill circuits. The VSI reduces the size of the feed by accelerating
the material and projecting it against a stationary surface. The amount of work that is done on the clinker
is directly related to the speed of the rotor, or the impact velocity, and is typically in the range of 2-3
kWh/mt fresh feed. The power consumed by the VSI is roughly equivalent to the power saved in the ball
mill in crushing, therefore little savings in specific grinding power consumption is realized unless the
reduced ball mill feed size eliminates other problems inherent in the original mill. It is typical to have the
VSI in closed circuit with a screening system to prevent any uncrushed particles from being fed directly to
the ball mill.
Literature shows that production increases of 20-30% are possible without changing the separator circuit.
However, each case must be analyzed for the inherent problems with the existing circuit. A more realistic
prediction is a 10-20% production increase for a well operated ball mill. Although the cost of the VSI
itself is relatively attractive, the cost of installing the auxiliary equipment (screens, elevators, etc.) can be
significant, as can the maintenance and down time attributed to wear of the VSI and screens.
Vertical Roller Mill (VRM) Pregrinders have also been used as pregrinders to boost existing mill
production. The vertical pregrinder differs from the vertical roller mill in that only the grinding portion of the
vertical roller mill is employed. There is no airflow or separator attached directly to the machine, as its
purpose is to perform the coarse grinding only. A partial recirculation of the product is made both to
balance the mill operationally and to crush any clinker that passes through the mill uncrushed on the first
pass. With a recirculation rate of 50-70%, the vertical roller mill pregrinder can take up 6-8 kWh/mt feed.
This means that an additional 20-25% more grinding power can be added to an existing ball mill circuit.
In combination with adjustments to the ball charge in the mill, a 30-40% capacity increase can be
attained. For a well operated system, an overall specific grinding power savings of about 10% can be
realized.
With a capacity increase of this magnitude, normally the auxiliary systems in the mill circuit, including the
separator, will need to be evaluated to insure that it can accommodate the increased throughput, else a
larger capital investment will be required. Down time of the original system may be an issue during the
erection period if the separator needs to be replaced. The overall cost of the vertical roller mill pregrinder
is much higher than the VSI due to the much heavier machinery and foundation works, however greater
increases in the capacity can be realized. Also, the wear of the rollers and table is very low, and
maintenance downtime and associated costs are quite low.
Hydraulic Roller Press (HRP) have been used in the cement industry for years, and today there are
hundreds operating worldwide. The major attractiveness to using an HRP is that high pressure grinding
is much more efficient in grinding than a ball mill, therefore a significant power savings can be achieved
which can offset part of the investment and maintenance costs.
Maximizing the work done by the HRP results in a more efficient overall grinding process. In order to
increase the amount of work done by a HRP, it is necessary to recycle a portion of the pressed material
back to the press for further grinding. However, the HRP suffers from operational instability if the amount
of fines recirculated becomes excessive. For these reasons, there are two main arrangements using an
HRP: Pregrinding and Semi-finish Grinding.
In a pregrinding arrangement, the HRP can be erected separately from the ball mill and separator circuit.
This can allow the HRP to be erected while the ball mill is still in operation, reducing the downtime
associated with the conversion. There is no separation of the recycled material back to the press;
therefore the circulation is limited to prevent operational instability. For this reason, the amount of work
that the HRP inputs to the material is only about 4.5 kWh/mt of fresh feed. As the HRP is more effective
in grinding than a ball mill, the result is a decrease in the specific power consumption of the entire mill
circuit by about 10%, depending on product Blaine. Changes can be made to the mill ball charge to take
advantage of the smaller feed size as well. In total, a capacity increase of 25-35% can be achieved by
adding an HRP as a pregrinder.
To maximize the capacity of an existing ball mill, the work done by the press must also be maximized.
This will require a large size press with a high material recirculation rate in order to maximize the specific
power input. As the press operation will become unstable due to the large amount of fines that is
generated, the press must then be put in closed circuit with a grit separator so that the fines in the
material recirculated to the press can be extracted. In this arrangement, the press can do most of the
grinding work, and the ball mill only “finish” grinding the material and flatten the product particle size
distribution. A capacity increase of greater than 100% can be realized in this “semi-finish” grinding
arrangement. The fines separator, elevators, and transport systems will all need to be
upgraded/replaced due to the significantly higher capacity, and therefor the investment cost and
downtime associated with the modification can be quite high. As the wear of the roller press is
proportional to the amount of clinker pressed, there can be some significant maintenance costs
associated with the system.
Milling circuits with a VSI or vertical roller mill pregrinder do not provide a significantly lower grinding
power consumption than a ball mill in closed circuit with high efficiency separator. As a circuit with a
pregrinder has more mechanical equipment, higher maintenance and lower run factors can be expected.
Only the HRP in a semi-finish arrangement significantly lowers the power consumption of the finish
milling circuit. However, this option is more expensive to install and maintain due to the additional
equipment. It is only when the power costs are significant that the HRP can give a positive ROI to offset
the initial capital investment. TABLE 1 gives a short summary of the benefits of each of these
pregrinding options as compared to a standard ball mill system.
TABLE 1: General Comparison of Pregrinding Options
Ball Mill VSI VRM Roll Press Roll Press

(Base) Pregrinder Pregrinder Pregrinder Semi-Finish
System Capacity
100% 110-120% 130-140% 125-135% 135->200%
Potential
Specific Grinding
100% 90-100% 85-92% 75-90% 60-80%
Power
Low High Moderate Moderate High
Maintenance
100 130 115 125 130
Availability 95-99% 80-90% 90-95% 88-95% 80-90%
Ease of
1 2 3 2 4
Operation
New Grinding Circuits can be a more economic solution when considering the options available for
increasing the grinding capacity significantly. Depending on the inefficiencies and physical condition of
the older, existing equipment, actual layout of the plant and requirement for new storage’s, transport and
structures, it may be preferable to install a completely new grinding circuit rather than upgrade the
existing mills. This is particularly attractive when there are several small mills that all need to be
modernized to achieve the desired final capacity. This would have a high cost not only for the additional
equipment, but also in terms of down time and constructability. If some of the existing mills are very small
and inefficient or suffer with high maintenance costs, the incremental cost of increasing the new mill size
can be offset by the operating advantage of shutting down the smallest mills.
CASE STUDY – INCREASING CEMENT GRINDING CAPACITY AT PHOENIX CEMENT COMPANY
Phoenix Cement Company (PCC), located in Clarksdale, AZ, currently operates (2) two-stage preheater
kilns and one long dry kiln, built between the late 1950’s and early 1960’s. The annual output of the plant
is in the range of 600,000 short tons of clinker. In 1997, Phoenix Cement Company began to discuss
plans for increasing the plant capacity to one million short tons per year.
In the initial planning for the expansion, it was considered that the additional clinker capacity would be
gained by adding a small, new kiln line, and maintaining production on the older lines. In order to support
the increased clinker production, it was determined that the overall plant cement grinding capacity should
be increased by about 30%. On site there are four 1500 HP ball mills, in which all of the raw material and
finish grinding take place. As the capacity of both the raw and finish grinding needed to be addressed, it
was originally thought that some synergy could be gained by converting the raw mill to a finish mill, and
install a completely new vertical raw mill. Although this would save the cost of the mill and some
building/civil works, it would involve a lot of transport of materials both to the mill and back out to the silos.
As the project started to develop, the scope was changed to that where a new precalciner kiln would
provide total clinker production, and the older, inefficient kiln lines be shut down. At that time PCC
abandoned the idea of converting the used raw ball mills to finish grinding, and began to pursue the
purchase of an existing 4400 HP Nordberg ball mill for making up the required finish grinding capacity.
There were very brief discussions held at the time to consider the addition of hydraulic roll presses, but it
was determined that the initial capital investments did not warrant its use, and PCC would have enough
grinding capacity with the existing ball mills.
The next challenge came about when the used Nordberg mill ball mill was sold and no longer available on
the market. Naturally, the project then came to consider a completely new ball mill to supplement the
finish grinding capacity. Although more expensive than the used mill, this would be a fairly
straightforward installation comprised of all new equipment.
PCC has a history of installing progressive process solutions in the U.S. cement industry. In 1988, the
first SEPAX flashdryer was installed in the US, operating as not just a classifier, but also a flash dryer for
the raw material. Additionally, Gamma Metrics installed the first on-line analyzer for raw material
proportioning back in 1985. PCC was also one of the first cement companies to explore the use of a
vertical shaft impact (VSI) crusher to boost the cement mill output. However, this VSI was never
commissioned for various reasons.
At the time when PCC was considering the new ball mill circuit, they were introduced to vertical roller mill
technology for cement grinding. For decades, the vertical roller mill (VRM) has been used for grinding of
cement raw materials and coal, and the VRM has become the dominant mill for these applications. The
VRM for cement grinding has been used in other parts of the world for many years; however; it had not
been installed in the North American market to that time. In considering the VRM for cement grinding,
PCC had to not only consider the financial implications of this technology in terms of installed and
operating costs, but also the product quality and comfort of using this technology as a supplement to the
existing finish grinding ball mills.
It is a well known fact that the energy consumption of a vertical roller mill system is significantly lower, 65-
75%, of that required for a comparable ball mill system. See TABLE 2 for a comparison of the proposed
installed power for a new ball mill system and new vertical mill system. As the capacity of the vertical mill
is higher than that proposed for the ball mill option, the table normalizes the power consumption by the
capacity. The operating power saving is expected to be in the range of 10 kWh/mt as compared to a new
ball mill installation.
TABLE 2: Comparison of Installed Power for Ball Mill and Vertical Mill
Ball Mill Vertical Mill
Mill Size UMS 4.6M x 13.0M OK 33-4

Mill Capacity (@ 3900 Blaine) 90.7 mtph 118 mtph
Installed Power
Mill Drive 41.1 kWh/mt 25.3 kWh/mt
Separator Drive 1.6 kWh/mt 1.3 kWh/mt
Fan(s) 4.0 kWh/mt 11.5 kWh/mt
Total 46.7 kWh/mt 38.1 kWh/mt
The vertical mill provides the grinding, transport and separation processes in a single, compact unit that
can perform with the same reliability as seen with roller mills in raw grinding. As a result of the reduced
equipment requirements, the layout of the finish grinding solution is simple and foundation and building
costs are low, especially in temperate environments when an enclosure is not required. The cost
difference between the ball mill and vertical mill system, on an installed basis, can be 10% or less –
depending on the final building requirements for the vertical mill. At Phoenix, the mill is arranged with
only a minimal weather covering. A light, open structure is used, with access for a mobile crane to be
used for maintenance and repair work. FIGURE 1 provides a sketch of the compact arrangement of the
Phoenix installation.
FIGURE 1: Phoenix Ok Mill Arrangement
The quality of vertical mill cement can be different from that produced by a ball mill – and the main
reasons for that is the steeper particle size distribution, and the low retention time which can impact the
dehydration of the gypsum. The vertical mill is quite flexible in it’s operation, and the particle size
distribution can be adjusted though changes in the dam ring, grinding pressure, and airflow rates. A
higher dam ring results in a thicker grinding bed, and a longer retention time on the table. The material
will thus receive more overrollings before leaving the table and therefore a lower circulating load to the
separator. The result is a flatter particle size distribution. A lower grinding force provides a lower energy
input per overrolling, resulting in a higher circulation to the separator and a steeper particle size
distribution. See FIGURE 2 for a graphical depiction of these relationships.
N-value
1.4
1.3 High dam ring

Medium dam ring
1.2 Low dam ring
1.1
78 83 89 100
% Compressive Force
FIGURE 2: Relationship of Product PSD and Operating Parameters
The dehydration of the gypsum can be adjusted by changing the mill exit temperature, which is simply
done by recirculation of the mill exit gases and adding additional heat from an auxiliary furnace if
necessary. Both the setting times and water to cement ratio can be adjusted by controlling the amount of
gypsum dehydration. As gypsum can come from different sources and have different characteristics, the
system is initially set up to operate with a mill exit temperature of 110ºC (230ºF) to insure that the gypsum
dehydration is sufficient, however it can be operated up to 125ºC (257ºF) if required.
The biggest concern regarding the use of the vertical roller mill was the impact on the product quality. As
the market in which PCC operates in is accustomed to receiving cement produced in ball mills, there was
naturally some concern that the product quality would be different than that of the existing ball mill
circuits. The impact of the changing product quality is a very serious marketing concern for any cement
producer. Before PCC would decide on a vertical mill solution, these questions had to be answered.
FLSmidth was able to supply information from many of its vertical mill installations that also operated with
ball mills in order to answer these questions. PCC also went to Cementos Progresso in Guatemala, a
plant that has two OK vertical mills and two ball mills in closed circuit with roll presses, to further
investigate these quality issues. TABLE 3 provides a comparison of the main cement quality data
between the ball mill and OK mill cement at that plant. As can be seen in Table 3, the cement from the
OK mill has a somewhat lower Blaine value and lower sieve residue on the 45 ìm sieve, i.e. a more
narrow particle size distribution. Nevertheless, all other quality figures, including water to normal
consistency, setting times, and compressive strengths are almost identical.
TABLE 3: Comparison f Cement Quality Between
Ok Mill and Ball Mills
Compressive Strength (ASTM)

SettingTime
Blaine Residue MPa
Mill 2
cm /gm %>45µ W/C Initial Final 1 3 7 28
% min min Days Days Days Days
OK 3700 6.4 28.6 130 310 12.3 22.3 26.8 34.2
BM 1 4000 7.5 27.7 130 310 12.8 21.3 25.4 32.0
BM 2 4000 7.6 28.0 130 310 12.7 21.6 25.8 32.9
* PCC cement: 75% clinker, 15% Pozzolan, 5% limestone, 5% gypsum
At the end of all of the evaluations, PCC was confident in the vertical roller mill technology for cement
finish grinding and selected a new, OK mill system to supplement its’ existing cement ball mills.
FINAL FINISH GRINDING SYSTEM DESCRIPTION
The new finish grinding area consists of an OK 33-4 vertical roller cement mill with a 4000 HP (2983 kW)
motor and an OKS-70 classifier with a 200 HP (149 kW) motor that is capable of producing 130 STPH
(118 MTPH) of cement at 3900 Blaine. As mentioned previously, the mill is arranged with only a minimal
weather covering. A light, open structure is used for the mill itself, with access for a mobile crane for
maintenance and repair work. The dust collector is outside the weather protection. The mill exit gas is
vented directly to a baghouse dust collector. The gas then goes through the dust collector fan, after
which part of the gas is recycled back to the mill inlet, while the remainder is vented to atmosphere. There
is little energy wasted on heat generation in the mill as compared to a ball mill, so an auxiliary air heater is
supplied to provide additional heat if required for drying the gypsum. The product collected in the
baghouse is then conveyed to cement product coolers. Two identical cement product coolers (10’0” x
20’8") are provided to cool the cement temperature from 110°C (230°F) to 63°C (145°F). The cool cement
is then conveyed to the existing cement storage silos.
CONCLUSION
Many different options were considered before the final decision to use the OK vertical roller mill for
cement grinding was made by Phoenix Cement Company. With the changing scope of the project, the
options including the use of existing mills, pregrinders, as well as new mill circuits were evaluated, and a
new vertical finish mill was finally selected. Even though there are no running references for vertical
cement mills in the U.S., PCC decided to go forward with this cutting edge technology. After thoroughly
investigating all of the aspects of vertical mill performance, PCC has shown confidence that this solution
will provide them with the most economic operation while producing the high quality product that their
customers demand.
EXISTING MILL CAPACITY DOUBLED
DEVELOPMENTS IN GRINDING PROCESS
by:
IEEE-IAS Cement Industry Committee
Siegfried Strasser, KHD Humboldt Wedag
Summary:
The tube mill is still the most common machine in the cement industry for grinding clinker. Large scale
production upgrades of tube mills are available since 1990, when the Roller Press has been set up for
Semi-Finish-Grinding. The major reasons for upgradings are - production increase, lower energy
consumption and the market demand for higher cement qualities. All three arguments should be fulfilled
for the cement mill conversion, introduced in the following. A pregrinding circuit, consisting of a Roller
Press and a separator has been selected, to more than double the capacity of the existing tube mill. While
this project was one of the first of its nature, a number of problems have been encountered. Finally, a
second upgrading step had to follow the first one, partly replacing major equipment. A new process, even
new machines have been borne, to reach satisfying results. Finally, it has become a new standard for
todays applications. The project introduced, is an example for a number of grinding plants, that still need
the second step of upgrade.
1. Introduction
Even improvements have been made in the grinding department, still more than two third of the electrical
energy consumption in a cement plant is taken by grinding systems. When the Roller Press was
introduced in 1985, simple pre-grinding systems have been used. The energy required for a tube mill has
been cut by 50% and more, when it was substituted by the Roller Press. These results have been spread
quickly to the cement industry. Since than, the grinding technology has undergone substantial changes.
Naturally, it was tried to transfer as much grinding work as possible from the tube mill to the Roller Press.
In 1992, the first two-stage grinding systems have started up. Suppliers and plant operators have been
convinced by the level of production upgrading of existing tube mills.
But the arrangement of the Semi-Finish Grinding, or also known as Combi-Grinding, has become very
complex and high in maintenance. With the first operating results, available from several grinding plants,
an urgent need for improvements was required. With the second generation design, equipment and
process systems are available, that avoid the shortcomings of the original design.
Both development steps are shown by a large scale industrial example.
2. The Original Grinding Plan
The clinker grinding plant under discussion has been built by Asia Cement Corporation in Taiwan. The
plant is located near the city of Hualien, at the east coast of the island. It is a closed circuit tube mill, that
has been started up in 1975, within the erection of a new clinker production line. It consists of a two
compartment tube mill and a cylone air separator, the original ZUB- type. It features today, the so called
second generation design. Type I and II cement are exclusively produced with this mill.
2.1 Equipment specification:
Tube Mill: Dynamic Separator

mill shell diameter 4.4 m housing diameter 4.8 m
mill chamber length 4.5 + 9.5 m separator motor 150 kW
motor (2 sets) each 2000 kW separator fan volume 180,000 m3/h
girth gear / pinion drive fan motor 220 kW
bucket elevator: mill system vent fan:
c-c distance 31 m flow rate 75,000 m3/h
throughput 650 t/h motor 160 kW
motor 80 kW
2.2 Process - operating data:
product: specific energy consumption KWh/t

cement production 105 t/h tube mill 39.1
Blaine fineness 3400 cm2/g separator incl. fan 2.3
residual fineness 325# < 10 % main equipment 42.5
cement temperature < 100 °C complete grinding plant 46.2
2.3 Arrangement, Closed Circuit Tube Mill:
mill bucket
elevator separator
vent duct
separator
ZUB48
feed
bag filter
clinker
hopper
rejects
pre-cyclone
fines
tube mill
mill motor cement

Gear box pump
figure 1: arrangement of grinding plant before conversion

Only one bag filter is used in the mill circuit for venting the tube mill, the separator and the conveying
system. A pre-cyclone is used in the mill vent duct to reduce the dust load of the bag filter.
The tube mill is equipped with a water injection system to keep the mill outlet temperature below 110°C.
Figure 1 shows the arrangement of the grinding plant, figures 2 and 3 give a view of the tube mill and the
dynamic separator, before conversion.
figure 2: tube mill drive, discharge end figure 3: dynamic separator typ ZUB 48 with
cyclones and circulation air fan
3. The Conversion Project, Semi-Finish Grinding
Together with the third clinker production line, built by Asia Cement in Hualien in 1992, it was decided to
increase the production of existing cement mill no.1 with a separate pregrinding circuit. For the raw mill,
with 400 t/h capacity, ordered at the same time, a Roller Press with 2000 kW drive power was required.
For upgrading of the cement mill the identical size was requested. At that time, three smaller Roller
Presses, each with 1000 Kilowatt drive power were already operating, for cement mill no.2 and 3 and for
raw mill no.1, which have been upgraded before 1990.
3.1 Modifications to the Tube Mill Circuit
The reserves of the ZUB-Separator accepted the increase of the air flow and naturally also a higher
production rate. The cement users asked for lower cement temperatures of far below 80°C. A new bag
filter for 260,000 m3/h has been
installed for 100% cooling with
new bag filter
fresh air and collecting the fines
directly from the separator.
Figure 4. The six cyclones and
the circulation fan have been
removed.
To reduce the load of the ZUB-

Separator, the pre-cyclone in
grit separator the mill vent duct to the mill bag
filter has been replaced by a
static grit separator. Approx. 20
t/h product is collected in the
mill filter, which is sent to the
finish product line. Figure4.
Within the conversion project

the pump for the cement
transport has been replaced by
a belt conveyor and a bucket
elevator.
Figure 4: Conversion of the ZUB-Separator and static grit separator
3.1.1 Equipment specification:
Separator bag filter: Static Grit Separator

filter area 2984 m2 diameter 2.6 m
air volume at 90 °C 257,000 m3/h
fan motor 500 kW
3.2 Pregrinding Circuit
The Semi-Finish Grinding Circuit consists mainly of the Roller Press, the Deglomerator, the High
Efficiency Separator and the conveying and dedusting system. To achieve a maximum grinding
capacity, originally it was forseen to add a smaller Roller Press to be operated paralel with the large unit.
Separator rejects have been split to both Roller Presses. Figure 5 shows a top view of the tube mill
existing
building
with
tube mill
figure 6: Building of Cement Mill no.1

figure 5: top view of tube mill and Roller Press new building on the left side
and the Roller Press, as it has been arranged originally. It turned out, that the material flow control has
become extremely difficult, since particle size of the rejects has changed frequently. It was not possible to
stabilise the circuit. The smaller Roller Press has been removed from this circuit lateron. In the following
the operation only with the large Roller Presse is described.
Figure 6 shows the top view of the existing building with the tube mill. The new building beside, has been
set up for the pre-grinding circuit and the new bag filter for the mill-separator.
3.2.1 Equipment Specification
Below table shows the specification for the major equipment of the first grinding stage. It has been
designed for a higher circulation load, with both Roller Presses in operation.
Roller Press: dynamic separator:

Type RP20-170/130 rotor diameter 3.75 m
roller diameter 1.7 m separator motor 500 kW
roller width 1.4 m fan air flow 500,000 m3/h
throughput 720 t/h fan motor 820 kW
motor 2 set each 1000 kW cyclone diameter, 4 set 3.8 m
deglomerator: Bucket Elevator;
rotor diameter 2.6 m throughput 1300 t/h
motor 650 kW Motor, 2 set, each 132 kW
3.2.2 Arrangement of the Semi-Finish Grinding system
Figure 7 shows the arrangement of the grinding circuit with the Roller Press and partly that of the tube
mill. The arrangement and the equipment designed for the grinding circuit was state of the art in 1992.
Roller Press bucket mill bucket

elevator elevator
ZUB separator deglomerator
bag filter
dyn. separator
cyclones
tube mill
Figure 7: arrangement of first grinding stage Roller Press
Description of the arrangement and the process: First Grinding Stage:
The fresh material, clinker and gypsum is extracted from pre-bins by weigh feeders and conveyed via belt
conveyors to the Roller Press (fig. 8,9) feed bin. It is ground by the Roller Press and conveyed
mechanically via belt conveyors and a bucket elevator to the deglomerator. Slabs are deglomerated by
the horizontally arranged rotor and the impact ring arranged around the rotor (fig.12). The rotating speed
is approx. 30 m/s. Via feed chutes the preground material enters the High Efficiency separator. On the top
ring of the rotor cage (fig.12) the material is distributed and separted by the air stream and
the rotation of the rotor. Fines passes the rotor blades (figure 13), the coarse moves downwards between
the rotor and the
guide vanes
(fig.13), arranged
around the rotor.
The required
fineness is
controlled by the
speed of the
separator rotor,
while the air flow is
kept constant. The
coarse material
reaches the pre-
Figure 8: View of Roller Press Figure 9: roller surface, welded,
bin of the Roller
Press via a chute,
together with the new feed, for further grinding. The fines is conveyed via an airslide and a bucket
elevator to the mill hopper.
Second Grinding Stage:
The existing mill hopper for clinker has been used for intermediate storage of the preground product. A
new extraction and feeding system has been installed to handle the fine product. It consists of aeration
equipment at the bottom of the hopper, an airslide with dosing gate and shut off valve and a weigh feeder.
The extracted product is entroduced to the mill inlet. The tube mill has been modified to a single chamber
mill. The coarse ball charge has been replaced by cylpepse. The tube mill circuit has been operated as
before in closed circuit with the ZUB-Separator. The finish product is collected in the new bag filter and
partly in the mill bag filter. With water spray to the mill and 100% fresh air to the ZUB-Separator, the
cement temperature before cement cooler is below 80°C, with clinker temperatures of more than 150°C.
Figure 10: mill inlet, feed by removable air slide Figure 11: mill discharge end
3.2.3 Operating Experience
The clinker, produced at Hualien-Plant is abrasive, difficult to grind and characterised by a high content of
fines. The fine grained clinker caused unstability of the the roller-gap and -pressure and the motor power
of the Roller Press.
The preground product with a Blaine surface of 1800 to 2300 cm2/g, and a temperature of above 120°C,
caused heavy coating of the grinding media, liner plates and discharge slot plates. Even with the use of
grinding aid, the material flow through the mill was restricted by the fine cylpepse charge. Beyond a feed
rate of 190 t/h, the mill feed material was backcharging from the mill inlet.
Figure 12: deglomerator,wear on lining of the impact ring, situation before redesign
With the coating of the grinding media and the high material level in the mill, only less than 80% of the
calculated mill power could be drawn.
After four weeks of operation, heavy wear has been recognised, in the dynamic separator, the
deglomerator and at all transfer chutes. The discharge material of the Roller Press contains a wide
particle size range, with a high portion of fines bilow 90 microns, but also very coarse, sharp edged
particles up to 10 mm can be seen.
Figure 13:
Wear in the dynamic separator
left: distribution ring, to of rotor cage

below, left: rotor cage blades
below, right: guide vanes
This was a short description of problems encountered throughout the startup of the plant. A fast action
was required to solve the wear problem and to stabilise the operation. Besides a number of minor trials
and modifications, following major actions have been taken:
--- redesign of the deglomerator rotor and the lining of the impact ring, use of material with higher wear
resistance
--- redesign of the separator rotor cage and use of hardfaced cage blades
--- improved were material in all areas, as chutes and dedusting pipes
--- replacement of cylpepse by balls with 25 to 15 mm diameter
The sharp edged grains with up to 10 mm size have been in direct contact with the rotating parts of the
deglomerator and the separator. Even with highly wear resistant materials, the life of the wear parts was
not satisfying.
Besides the problem of high wear, the number of equipment and electrical consumers have increased the
maintenance- and spare parts cost. A large number of venting points and a high vent air volume required
extra maintenance.
Even the circuit could be stabilised, it suffered from the availability of the equipment. During a short test
period, the energy consumption has been reduced, but counting on a yearly average, considering the
shut down, it was as high as for the closed circuit tube mill, before conversion.
Besides the operating difficulties, another limitation of the grinding system, but not related to this project is
the drying capacity. Even hot gases are introduced to the dynamic separator, the retention time of the
feed material, before passing the rotor cage is too short for an efficient drying and separation. Sticky
materials caused clogging and unbalance of the deglomerator.
It has to be mentioned at this point, that 6 out of 40 grinding plants, operating in Semi-Finish mode as
described before, have show such an extreme wear- and operating problem. E.g. a grinding plant,
operating in Germany, does not face that wear problems at all.
In the following research, a large difference in wear rates of clinker has been measured and confirmed in
practical operation. The difference is as much as factor five.
Summary of Semi-Finish Grinding, respectively Combi-Grinding
- high wear of dynamic parts of deglomerator and separator

- large number of transfer chutes, subjected to wear
- large number of auxiliary equipment, drives and other electrical consumers
- large quantities of vent air and venting points
- limited application due to low drying capacity
- unstable operation with fine grained clinker
4. The Second Step of the Conversion Project -- Semi-Finish Grinding, Second Generation
During start up of the first

Semi-Finish-Grinding
plants, with the problems 4 4
described before, a static 1 2
cascade-type separator,
optionally with a rotor
cage, was under
research. Three grinding
plants, with similar
problems as shown
before, could be
converted within a shot
time period. Operating
experience was available, 1 – feed
when Asia Cement 2 – air inlet
decided to convert the 3 – coarse
first grinding stage to a 4 – fines,
V-Separator in 1995. air outlet
4.1 The V-Separator
The first approach, to 3 3 1 2

eliminate the short-
comings, was to question Figure 14: V-Separator versus air-separator
the dynamic deglomeration- and separation process.The idea of a cascade type static separator has been
picked up and transferred to applications with the Roller Press.
Air flow simulations and extensive tests from a prototype have given the shape and the specifications to a
new separator in the cement industry. World-wide it is known as the V-Separator (VS).
The principle of the VS can be explained best by comparing it with the conventional static separator, type
LS, known from air swept tube mills. (see fig.14) The LS has the feed material entering together with the
separating air through the riser duct from below. This requires a minimum ratio of air to feed-material for
vertical conveying. Opposed to this, the V-Separator is
fed by gravity from above, while the air is entering the
VS from the side, crossing the material stream. This
makes the feeding rate less dependant from the
separating air. The material loading of the separating
air can therefore be up to three times higher than with
an LS-type airstream separator.
The required product fineness is achieved by the air
speed in the separating channels. The cut point
ranges from 90 microns up to 1.5 mm.
The internals, as step grate and separating channels

have straight and level surfaces, to which a simple but
effective wear protection is applied.
Figure 15: V-Separator of cement mill no.1, inside view
Summary: Features of the V-Separator:
- no moving parts, not even adjustable parts

- maintenance-free operation for years
- wide cut size range, between 0.1 and 1.5 mm
- does four jobs:
- separating of fines
- drying or cooling of feed material
- deglomerating
- mixing of fresh feed and rejects
These features are most welcome, for a stable operation of

The Roller Press.
4.2 Equipment removed

Specification of New Equipment Figure 16: V-Separator discharge end, fresh air inlet
Following modifications have been carried out: (compare arrangement fig.7 and 17)
Installed: Removed:
fresh feed chute to bucket elevator, feeding belt conveyor between weigh feeder and Roller
the V-Separator Press
V-Separator Deglomerator with motor 650 kW
high efficiency separator with motor 500 kW
belt conveyor between b.e. and separator
V-Separator fan 220,000m3/h, motor 315 kW separator fan 500,000m3/h, motor 820 kW
2 cyclones 3.8 m dia with air lock
air slide between VS-cyclones to mill inlet bucket elevator 300 t/h, motor 55 kW
mill hopper bypassed, extraction and dosing
system removed
Above table demonstrates the changes made to the system. A relatively small fan with 315 kW drive
power has replaced equipment with a 1970 kW total motor power. The V-Separator requires only 45% of
the original fan volume. The static pressure drop of the fan is reduced to 65% compared with the fan of
the dynamic separator.
4.3 Arrangement of the V-Separator
The V-Separator fits to the same location that was required for the dynamic separator. More space is
available inside the building, after removal of two large cyclones and ducting.
2 cyclones
feed chute smaller fan
to VS
V-Separator
fresh feed to
V-Separator
Figure 17: V-Separator arrangement
4.4 Operating Experience after conversion with V-Separator:
The clinker feed is introduced to the V-Separator, instead to the Roller Press. This has made major
changes to the operation stability:
- fine and dusty clinker is removed before entering the Roller Press
- fresh feed and V-Separator coarse material are well mixed before the Roller Press
Since the particle size distribution of the product from the V-Separator is different from that of the dynamic
separator, it was a benefit to feed the preground product to the ZUB-Separator instead to the mill inlet.
This has reduced the load of fine in the tube mill. The grinding performance improved.
4.5 Process - operating data:

Blaine fineness 3400 cm2/g Roller Press 11.8
For comparison with the tube mill below table with the process data is shown again.
(2.2 Process - operating data)

Blaine fineness 3400 cm2/g separator incl. fan 2.3
Considering the energy, required for the pur grinding process, the Roller Press is 2.1-times more efficient
compared with the tube mill. The auxiliary equipment for the Two-stage Grinding system consumes more
energy, than that for the tube mill.
Today, process options are available and in operation, which further reduce the number of equipment and
energy consumption. With the major improvements done to the process and machines, Semi-Finish
Grinding went into Second Generation.
During six years operation, no maintenance was required for the V-Separator itself. The feed chute
between bucket elevator and VS needs better wear protection, to avoid maintenance. The fan impeller of
the V-Separator has not been replaced yet, hardfacing is being renewed.
5. Comparison of Grinding Systems
In the following the three stages of the grinding plant are compared with each other. Long term records for
investment- and maintenance cost have been made available by Asia Cement Corporation.
A… Closed Circuit Tube Mill

B… Semi-Finish Grinding, First Generation
C… Semi-Finish Grinding, Second Generation
5.1 Investment Cost
system year mechanical electrical civil total

A 1975 n.a.
B 1991 9.95 1.28 1.24 12.47
C 1995 1.42
( figures indicated are in million US$, the figures have not been corrected to todays value )
5.2 Operating and Maintenance cost
system year Maintenance cost Energy cost sum

A 1975-1992 0.52 2.83 3.34
B 1992-1995 0.99 2.78 3.77
C 1995-2001 0.68 2.36 3.04
( figures indicated are in US$ per ton of cement,

the figures have not been corrected to todays value )
5.3 Maintenance and Spare Parts consumption for Roller Press
The Roller Press type RP20-170/140, in 1992 was the largest machine available. The design grinding
force transferred via the hydraulic cylinders and the rollers to the material bed is 2000 tons. The rollers
are driven by two 1000 kW motors. In 1996 the motors have been increased to 1200 kW each, since
more grinding work ccould be transferred, after the VS-Installation.
Following records have been taken:
operating period: 1992 to 2002

operating hours: 58,000 hours
total cement production: 10,750,000 tons
The wear of the rollers has been recorded during ten years or operation. It is expressed in grams of worn
steel per ton of cement production.
Specifiic wear = 1.4 grams per ton
Replacement parts for the Roller Press from 1992 until 2002:
- roller 3 times, total 5 rollers changed, rebuilding of used roller / tire

- planetary gear 2 times, total 2 gears, motor side gear stage replaced
- motor coupling 5 times, coupling failure due to suppliers manufacturing fault
- hydraulic cylinder 8 set, replaced and refurbished with new seals
- motor 1 replacement due to failure
2 replacements from 1000 to 1200 kW to transfer more grinding energy
5.4. Space requirements
Figure 18 and 19
show the top view of the separator arrangement for the tube mill- and the Roller Press circuit. The bag
filter for the mill separator has been placed in the new building. Semi-Finish Grinding of the first
generation requires four cyclones, and a large fan. The Second generation design with the V-Separator
accepts approx. 30% less space.
Figure 18: Figure 19:

Semi-Finish Grinding, First Generation Semi-Finish Grinding, Second Generation
Deglomerator and V- Separator
High Efficiency separator
6. Outlook
The V-Separator entered the market in 1994. At the same time the VSK-Separator was the final in the
research. The VSK combines the V-Separator with a horizontally arranged rotor cage.
Second conversion of the cement mill in Hualien started in 1995. The developments have continued
durind the past seven years. The VSK-Separator, part of a 120 t/h cement mill, started operation in late
2001 in a German cement plant. The impact of the VSK on the arrangement of the cement mill conversion
is demostrated in figure 21.
6.1 VSK-Separator
air- and material

The VSK-Separator (see figure 20) combines f eed
the static V-Separator and the high efficiency product air
SEPMASTER-separator to one compact unit. outlet inlet
The rotor is arranged horizontally.
The feed material is separated into three hot gas or

different products. The VS-part takes the rough f resh air
work. It receives the coarse feed, it is even
made to receive fresh feed to separate fines midlings
from clinker or other materials. By designing
the air velocity for the VS, more or less coarse
material can be discharged and fed to the rotor
cage, the dynamic part of the VSK.
The circulation load of the static and the coarse

dynamic part can be adjusted separately. outlet Figure 20: VSK-Separator
The fines is passing the rotor with the separating air, the rejects of the rotor cage, the middling, are
withdrawn separately from the very coarse rejects of the V-Separator.
6.2 Proposed Arrangement with VSK-Separator
Semi-Finish- Grinding has

become furthermore attractive.
With the marriage of the V-
Separator and the dynamic
separator to one unit, further
energy savings are achieved.
The pressure drop is reduced by
approx. 10 %. The VSK-
VSK- Separator requires less space.
Separator The wear protected surface area
extension in contact with material-laden air
is reduced, resulting in lower
wear rates and lower overall
investment cost.
The modular design of the VSK
2cyclones, accepts the rotor cage including
fan, ducts
rem oved
housing and drive to be added to
ZUB- the existing V-Separator. (figure
Separator 21, black shaded part) From the
rem oved 21:
Figure VSK-discharge end the duct is
air slide
Semi-Finish Grinding, with VSK-Separator extended connected, leading to the existing
from ZUB bag filter. The air slide feed ing
to VS the ZUB is extended to the VSK.
Following equipment can be taken out of service:
- two cyclones including ducts
- circulation air fan;

the bag filter fan draws fresh air through the VSK
- the ZUB-Separator;
which is the bottleneck, after upgrading, causing high pressure drop, can be removed
Reduced maintenance and spare parts reqirement could be the major reason for a conversion. But the
energy saving of at least 3.0 kWh/t is very attractive.
7. Conclusion
Large scale upgrading of existing cement mills has started in 1992 with the use of large Roller Presses,
building a closed circuit with a deglomerator and a High Efficiency Separator, the so called Semi-Finish
Grinding- or Combi-Grinding System. High operating- and maintenance cost with an increased number of
equipment forced to search for different solutions. With the V-Separator application, this process entered
into the Second Generation design.
Even the technology has been proven since 1994, still Combi-grinding plants have been built using the
first generation design. These grinding plants are being converted now by using V-Separators.
Today, closed to 50 grinding plants are available with a V-Separator in the Roller press circuit - ten are
conversions from first into second generation.
Several lessons have been learned, while going through the upgrading stages of this grinding mill. It will
always be considered, when entering into new developments:
- the simple process and machine applied at the right place, can give higher overall efficiency,
compared with high efficient single units (V-Separator versus High Efficiency Separator and
Deglomerator)
- cylpepse are not useful for high mass throughput of a tube mill; and combined with high feed
temperatures and fine product feed - heavy coating in the mill is the result
- two Roller Presses to operate parallel require its own feeding system and its own separator, with
this arrangement it can serve one tube mill
- variations in wear properties of cement clinker can be as high as factor five
Thanks to the management of Asia Cement, who have taken and shared the risk of a first use of
equipment, and have given a chance to make corrections. The shortcomings of the first generation design
have not only been eliminated, but turned into additional benefits.
J wear is drastically reduced by the V-Separator, the number of transfer chutes is reduced
J no rotating parts and electrical consumers for the VS
J vent air volumes are reduced by 50%, as well as the number of venting points, energy is saved
J the drying capacity is no more the limiting factor of the grinding circuit
J the overall investment- and maintenance cost are reduced
Since 1995, five more V-Separators have been supplied to Asia Cement, four within new grinding plants
for Roller Presses with 2400 to 3000 kW drive power.
Power Generation, Distribution
and Related Products
Working Group
Chair
Mark A. Huber
Lafarge Corporation
Vice Chair
Jean-Paul Morel
Essroc Materials
Members
lb Bentzen-Bilkvist
Retired (Holcim Inc.)
George J. Fergades
Lafarge Corporation
Dean M. Ikemoto
Lafarge Corporation
Gregory J. Miranda
GE Industrial Systems
Jack Stroker
Penta Engineering
Paul Van Teslaar

Siemens Westinghouse Technical Services
OPTIONS FOR THE CEMENT INDUSTRY IN THE
DEREGULATED POWER ERA
By: Gregory J. Miranda, P.E.
Application Engineer
GE Industrial Systems
Abstract: The California energy crisis has brought

electric power deregulation to the forefront for Overview of the Current State of Electric
electricity users in the U.S. Could what happened in Power in the U.S.
California occur in other states? Cement companies
are asking themselves what they can do to prepare The domestic power industry has gone through
for the consequences of electric power deregulation many changes in the past several years and
in the U.S. This paper will give the cement continues to evolve. There are several major trends
manufacturer an overview of the current electric in the U.S. electric power industry today:
power industry and the trends in that industry. The
• Deregulation legislation has slowed in many
paper will then discuss concerns about deregulation
states as a consequence of the California
and outline different options that users might
energy crisis.
consider to address their concerns.
• Electricity generation and transmission
ownership will be concentrated within a few
Introduction large companies.
• There is a current boom in generation
The electric power industry has changed building, but opportunities to expand the
significantly since the Federal Energy Regulatory transmission systems may be limited.
Commission (FERC) started to encourage the • Environmental issues may be game
development of competitive markets for wholesale changers as pressure continues to further
power trading through deregulation of portions of the reduce emissions.
electric power industry. The belief is that • The emergence of an energy trading market
competitive markets will lead to more efficient power and the stability of some of key players may
generation, more technological innovations and be an issue.
eventually to lower retail prices. However, in the • Electric consumers will become a more
meantime there are going to be glitches, as significant force through self-generation and
evidenced by the California crisis and the collapse of load control.
Enron.
California Crisis
Electric power users need to be knowledgeable of
the on going changes in the industry as well as their Most people are familiar with the recent energy crisis
own energy usage to be able to take advantage of in California. Many of the cement plants in that state
the current and future situation. As stated in the have been affected by the rolling blackouts and
paper by Tom Mulholland presented at the 2001 large increases in the cost of electric power.
Vancouver IEEE/PCA CIC Conference, “Customers Knowledgeable electric utility industry insiders
must understand their energy markets in great detail recognize that the California “power shortage”
and must stay current on the constantly shifting moniker is a gross under-simplification of a very
issues. In addition, they must understand their complex set of difficult issues. The real problem is
energy usage patterns, and the implications those one of energy economics. Among the underlying
patterns have on energy buying. Customers must causal factors are:
develop options in order to have flexibility.” • The absence of significant new generation
projects in the state for the last 15 years
The first part of this paper will give an overview of • Environmental constraints that have
the current state of the electric power industry in the severely hampered transmission
U.S. and some of the trends in the industry. Then it reinforcement
will describe options for monitoring energy usage, • Drought in the Northwest and abnormal
evaluating potential energy savings and some of the weather in California
other benefits from energy monitoring systems. • Unanticipated maintenance outages at
Finally, possible options for electric consumers to power plants
best position themselves to operate in the
deregulated power era will be discussed.
• Unforeseen economic boom in the state and generation assets and focus their efforts on
underestimation of the growth of power transmission and distribution. The economic theory
demand is that transmission and distribution are “natural
• A deregulated market structure based on monopolies” in that society will not tolerate multiple
flawed assumptions about market dynamics electric power distribution systems. But electric
• The lack of a market basis to enforce energy can be a market commodity, and the
system reserve requirements business of producing electric power can be
accomplished in a truly competitive market managed
Several developers are performing preliminary by impartial, independent system operators (ISO’s)
screening to help identify grid-related technical who manage regional power markets. Under this
hurdles for proposed new generation sites within theory, the transmission grid is likened to the
California and in adjacent states. Transmission interstate highway system, and requires that all
operators have also been busy determining the final supply-side competitors have equal access to
requirements for integrating new power plants into transmission.
the grid and expediting the process of approving
new generation. Studies have been conducted to In many areas of the US, the wholesale market for
determine how best to reinforce the transmission electric power has been fully deregulated and
grid so as to bring in more power from these remote energy suppliers bid competitively for the opportunity
sources. to sell their product. In some areas, large industrial
and commercial consumers have the option of
The problem in California has been building for choosing their energy supplier, and a few areas are
many years. Therefore, realistic solutions will also experimenting with taking choice to the residential
take time to emerge, but progress is being made retail level.
toward addressing some of the causes.
However, alarmed by California’s energy crisis, with
Deregulation the accompanying blackouts and price spikes, many
states became concerned about what deregulation
The 1978 PURPA act that got the ball rolling on non- could mean for electricity systems across the
utility generation was really the first step toward country. The current power problems facing
deregulation of the electric power industry. California have come about because of poor
Deregulation involves a number of dramatic changes deregulation planning not because of deregulation
in the fundamental structure of the energy supply itself. Never the less, several states that have
industry. approved deregulation plans have delayed or
appear likely to delay implementation of their plans.
Disaggregation of the traditional, vertically integrated Figure 1 shows the status of state deregulation
electric utility was the next step and has progressed efforts as November 2001.
to significant degrees in many areas. This amounts
to forcing electric utilities to divest themselves of
is that transmission and distribution are “natural
monopolies” in that society will not tolerate multiple
electric power distribution systems. But electric
energy can be a market commodity, and the
business of producing electric power can be
accomplished in a truly competitive market managed
by impartial, Independent System Operators (ISO’s)
who manage regional power markets. Under this
theory, the transmission grid is likened to the
interstate highway system, and requires that all
supply-side competitors have equal access to
transmission.
Figure 1: Status of State Electric Industry Restructuring

Activity as of November 2001
Generation turbines. The supply of natural gas, and increasing
costs, may limit its use for the future increased
The concentration of ownership of Investor-Owned
demand for generation. The siting of new generation is
Utility (IOU) generating capacity has increased
also a concern due to the length of permitting for
significantly since the 1992 enactment of Federal
greenfield sites. The uncertainty of deregulation and
Energy Policy Act. Since IOUs currently represent
worries about the capability of the transmission
approximately 75 percent of electric utility generating
systems to wheel the power from the generation sites
capability, generation, sales and revenue in the
to the users or distribution systems may also affect the
U.S., what happens to the IOUs can have a major
rate of construction of new generating capacity.
impact on the future state of generation in the
country. The total number of IOUs owning
There is a trend in U.S. generation away from the
generating capacity has declined from 170 in 1972
central station concept to more reliance on Distributed
to an estimated 143 in 2000. Figure 2 shows how
Generation (DG). With increased system capacity,
the concentration of ownership of Investor-Owned
reliability and quality needs, combined with increasing
Utility generation capacity has changed since 1992.
environmental pressures and technological
Non-utility generation capacity has also increased
advancements has emerged a wide spread recognition
substantially. The portion of electricity generated by
that DG technologies can play a key role in the future.
non-utility generators (NUGs), which include
The emergence of economic technologies for DG has
Independent Power Producers (IPPs), grew from
the potential to radically alter the energy business
about 6 percent of the total generation capacity in
environment over the next decade. Unlike large,
1991 to almost 20 percent in 1999.
central-station power plants, distributed generation
power systems normally range from less than a
100
kilowatt to tens of megawatts in size.
Percent Generation Capacity
10 Largest 20 Largest
80
72.7
DG Technology Typical Range of Sizes
60 55.9
60.3
Combustion Turbines 1 – 30 MW
50.5
Reciprocating Engines 10 kW – 10 MW
40 36.0
39.2 Micro-Turbines <1 – 300 kw
Fuel Cells 1 kw – 20 MW+
Renewable Energy
20
Biomass 5 MW Max.
Wind Turbines 1 kw – 1 MW+
0
1992 1998 2000
Solar 1 kw – 20 MW+
Photovoltaic 1 kw – 1 MW
Figure 2: Concentration of Ownership of IOU Table 1: Distributed Generation
Generating Capacity
Distributed generation comes in many forms using
Between 1978 and 1992, America’s utilities’ capacity various energy sources as indicated in Table 1. Fossil
margins averaged between 25 and 30 percent. fuels are currently the dominant sources of energy for
Since 1992, the capacity margins have declined to DG, as they are for large central-station generation.
less than 15 percent nationwide. There is a current Natural gas and petroleum are the most common fuel
boom in new generation in the U.S. and the decline for mini- and micro-turbines and reciprocating engine
in capacity margin is expected to reverse in 2002. types of DG. However, there is growing use of
The national capacity margin is expected to reach biomass energy sources, such as wood, peat,
18.4 percent in 2004 with electric generating agricultural and municipal waste, etc. Due to
capacity expected to reach 870.0 gigawatts, while environmental concerns, there is also increasing
demand is projected to reach 709.6 gigawatts. interest in what is known as clean distributed power
However, some experts believe that the bubble of including geothermal, solar thermal, photovoltaic and
orders for new generation will burst by sometime in wind generation as well as fuel cells that use
2003. Mainly due to environmental concerns, most hydrogen, propane, natural gas or other fuels to
of the new generation has been driven by gas generate electricity without increasing pollution.
Transmission
Under the old rules, prior to deregulation, generation Eastern Texas Western U.S.
and transmission expansion were planned together to Grid Grid Grid Total
address forecast increases in demand, with the Demand 501,405 53,414 114,830 669,649
overall objective that of “least cost”. The economic (Megawatts)
theorists promoting deregulation initially believed that Supply 582,223 69,769 141,068 793,060
the transmission system would continue to be a (Megawatts)
regulated monopoly service business. Unfortunately, Capacity 13.9 23.4 18.6 15.6
this may have also embodied an assumption that the Margin
transmission system could absorb whatever (Percent)
generation a deregulated supply market would drive Table 2: U.S. Capacity Margins, 2001
along with whatever load the overall economy
demanded. In fact, the industry is currently struggling Deregulation of the electric power industry in the
to address the concern that investment in 1990’s opened the grids to anyone who had electricity
transmission has stagnated, and the former process to sell. New entrepreneurial power brokers and re-
for planning expansion of transmission in conjunction structured utilities now want to sell power to whoever
with expansion in generation resources was a offers the highest price, wherever they may be
casualty of disaggregation. located. The future is uncertain and is an ongoing
concern. Overall use of the transmission system is
The U.S. power network is built in three large grids: growing without significant new construction or
the Eastern, Texas and Western Interconnected upgrades. Investment in transmission systems
Systems as shown in Figure 3. These grids are declined by over 50 percent from 1990 to 1998
connected in a handful of spots by high voltage DC despite the growing demand. Approval of new
lines, but the grids largely operate independently. projects and acquisition of new right-of-ways has
Each grid is composed of a tangle of transmission been difficult. The transmission system has become a
lines operated by various operators, from federal bottleneck to deregulation in the U.S.
power authorities to regulated utilities to large
conglomerates. Various state, regional and federal So far, the absence of investment in transmission has
regulators govern aspects of the grids, deciding how not created any especially intractable problems, but it
much power can enter the grids and flow over each is almost universally agreed that the present situation
set of lines. cannot be allowed to persist. The Federal Energy
Regulatory Commission has proposed formation of
Regional Transmission Operators (RTOs) as a
means of addressing this concern. Among the
expectations of these new entities are an
improvement in the power grid reliability, reductions in
discriminatory transmission practices and increased
investments in transmission infrastructure. It remains
to be seen to what degree RTOs actually emerge and
to what degree they can really attract capital to invest
in transmission.
The other trend that is emerging is that of merchant

transmission. If the transmission grid is analogous to
the free interstate highway system, as some have
proposed, then merchant transmission corresponds
to toll roads. The theory is that investors will
construct transmission assets to be paid for based on
actual utilization, or throughput. So far, one proposal
Figure 3: U.S. Transmission Grid Systems for merchant transmission has been approved by the
FERC, but the project encountered insurmountable
While the nation overall has a 15.6 percent capacity opposition from State regulators based on
margin for 2001, the 3 independent power grids differ environmental concerns.
in their levels of capacity margin. (See Table 2). The
Eastern grid, which is the largest, has a capacity The inherent capability of the transmission system is
margin of only 13.9 percent, whereas the Texas grid governed by the laws of physics – Ohm’s law dictates
has a margin of over 23 percent. how power will distribute through the network, while
heat-transfer considerations establish the physical The effects and costs of poor power quality depend
capabilities of individual branches in that network. A on many factors. For many continuous process
variety of technological innovations have been industries the duration of power unavailability has
proposed for improving the ability to transmit power only secondary impact on cost, and where the initial
through the grid. Many of these involve the instant of interruption may result in the need for costly
application of power electronics to “bend” the rules of repairs or cleanup before production can be restored.
power transmission. But these technologies are Many batch processes industries can withstand
expensive and cost has been an obstacle to momentary interruptions but not prolonged outages.
widespread application.
Cement manufacturing is a combination of both types
Distribution of industry. The cement kiln is a continuous process
and even relatively short interruptions of power can
The electric power distribution systems continue to be
cause problems with the operation of the kiln and
regulated similar to before deregulation. Since the
may affect the quality of the clinker. Longer
distribution systems typically don’t affect the wheeling
interruptions can also cause thermal cycling problems
of wholesale power or connections to large industrial
with the refractory brick inside the kiln. Many of the
electric power users, like cement manufacturers, the
other operations in a cement plant are batch
current status and trends in these systems will not be
processes, like crushing, raw and finish grinding and
discussed in this paper.
packing/shipping. If there is an outage, the lost
production may be made up by running longer shifts.
Power Trading
With the introduction of competition, wholesale power Typical Typical
Spectral Typical Voltage
trading has increased markedly. Power marketers Category Content Duration Magnitude
buy and sell electricity, but unlike most utilities they Transients
typically do not own or operate transmission or Impulsive 5-ns to .1-ms < 50ns to > 1ms
rise
distribution facilities. Currently over 500 companies Oscillatory <5 kHz to 5 µs to 50 ms 0 – 8 p.u.
are classified as power marketers and have filed rate 5 MHz
tariffs with FERC to sell wholesale power in the U.S. Short Duration
Variations
However, actual sales by power marketers are Instantaneous .5 – 30 cycles < 0.1 – 1.8 p.u.
concentrated in approximately 50 companies or less. Momentary 30 cycles – 3s < 0.1 – 1.4 p.u.
The rapid rise and more precipitous fall of Enron in Temporary 3s – 1 min < 0.1 – 1.2 p.u.
Long Duration > 1 min 0 – 1.2 p.u.
late 2001illustrate the volatility of the emerging power Variations
trading market. How this stunning turn of events will Voltage Unbalance Steady State 0.5 – 2%
affect the stability of the power trading market is Wave Form Distortion
th
Harmonics 0 – 100 Steady State 0 – 20%
unknown at this time. harmonic
Interharmonics 0 – 6 kHz Steady State 0 – 2%
Noise Broadband Steady State 0 – 1%
Power Quality Voltage Fluctuations < 25Hz Intermittent 0.1 – 7%
Power Frequency < 10s
Power quality issues are not new to the deregulated Variations
era of electric power. They have always been there,
but with the above-mentioned changes in the Table 3: Categories of Power System
generation, transmission, and selling of wholesale Electromagnetic Phenomena
power there is every indication that that power quality
may become more of a factor when evaluating the Vulnerability of the process to voltage dips or sags
purchase and use of electricity in the future. may include:
There can be different definitions of power quality • Control Error – Loss of control power results
depending on one’s frame of reference. However, in the inability to control the process.
from an electric power user’s perspective power • Contactor Dropout – A voltage dip or sag
quality can be defined as “any power problem can cause a magnetically latched contactor to
manifested in voltage, current, or frequency drop out and stop the motor.
deviations that results in failure or misoperation of • Voltage Flicker – Flicker is the repetitive
customer equipment”. This paper does not go into variation in the intensity of lighting and is
detailed discussion of what constitutes or causes more of a human irritation than a direct cause
power quality issues. Table 3 shows the categories of process interruption. However, it can also
and characteristics of power systems phenomena. be used in a more literal sense to describe a
set of problems in which lighting is
extinguished due to voltage dips.
• Machine Dynamics – Voltage dips or sags A major advancement in PMCS in recent years is the
may limit the ability of a power system to advent of web based monitoring and analysis
distribute power from sources to loads, which systems. A web based monitoring and analysis
could lead to generators not being able to system can be used to analyze energy cost data for
maintain stability. electricity, natural gas, steam, water, or any other
• Stall and Re-acceleration – Motors will stall commodity important to a business. It allows the
if the supply voltage is depressed for a energy user to look at commodity consumption from
prolonged period. For short duration dips or an enterprise level as well as drill down to better
sags the motor may re-accelerate when the understand processes that are major energy
voltage is restored. Re-acceleration involves consumers. Analysis can be done on near real time
higher than normal motor currents that may data accessible from any web browser.
result in further voltage sag problems.
By combining load profiling and load aggregation with
real-time power monitoring, historical data reporting
Options for the Cement Manufacturer and automated alarming and control functions, a
PMCS can help to reduce the costs of day-to-day
operations across the entire enterprise. In a
What can cement manufacturers do to take
deregulated environment, electric consumers can
advantage of the possible savings available with
monitor real-time spot market prices to analyze
deregulation while minimizing the potential negative
energy costs and then avoid high energy costs by
effects of deregulation?
participating in demand response and load
curtailment programs.
Power Monitoring
As mentioned earlier in this paper, in addition to Electrical commodity consumption can be aggregated
understanding their energy markets, electric across multiple facilities to smooth peak usage and
consumers must understand their energy usage then presented to a variety of suppliers to negotiate a
patterns, and the implications those patterns have on lower total cost. To maintain and decrease electrical
energy buying. A power monitoring system is the tool demand charges, load can be monitored and held to
to obtain this understanding but also has the a minimum by shifting non-critical production, (i.e.
capability to provide additional benefits to the user. grinding) to off peak periods. The system provides an
alarm feature to notify plant operations in advance, if
The more a user knows about his electrical system, a peak load is reaching critical levels or if utility
the more reliable it can be made and the more overall curtailment opportunities exist.
efficiency can be obtained. A Power Monitoring and
Control System (PMCS) can capture a lot of The plant can program their electric, natural gas, and
information on the electrical system to enhance water utility rate structures allowing them to
reliability, minimize energy use and provide immediately quantify the impact of shifting load
information to intelligently purchase power in the requirements and economically forecast energy
deregulated market place. consumption for the next production cycle. Energy
commodity efficiencies can be calculated per ton of
A paper presented at the 2000 Salt Lake City product produced allowing the plant to model their
Conference detailed how a cement manufacturer own energy efficiency and benchmark against similar
implemented a PMCS system in an existing plant and plants producing identical product. Variation in
discussed the benefits of the system so this paper will energy consumption can be analyzed to determine
not go into details on the benefits of plant PMCS root cause and translate best practices.
systems. A plant PMCS would provide better
monitoring of the power system, increased system Intelligent load shedding, electrical generation
protection, fault and power quality analysis capability, options, and fuel switching scenarios can be analyzed
powerful preventive and predictive maintenance tools before incurring the risk or cost to implement a
and improved process operation. Intelligent Electronic project. Once implemented, the system can provide
Devices (IEDs) and communications infrastructures closed loop control of selected scenario programs
are constantly decreasing in costs while increasing in without impacting production beyond the input
capability. PMCS systems have been constantly perimeters.
adding features and functions to address the
changing needs of the energy users. And it is now Electricity Rate Structures
much easier to cost justify a PMCS system.
Traditionally, the formula used to calculate the cost of
electric power consumed in an industrial facility
depended on the anticipated magnitude of load, the
voltage at which service was taken, and the number
of supply circuits that the industrial consumer
demanded to achieve his targeted reliability. The
host utility would then apply these physical
considerations against a tariff that had been adopted
by the responsible State regulatory agency to
determine the incremental rate for consumed energy
as well as how such considerations as demand and
power factor would be taken into account.
Recently, these formulas have been modified, and

some of the changes have been dramatic. For
example, some formulas now include a “time of use”
provision. In the simplest case, this means that
power consumed during off-peak hours (typically,
after 8pm but before 6am during the normal work
week, and any time on weekends) costs less than Figure 6: Real Time Pricing (NY ISO)
energy consumed during on-peak hours. The logic
behind this is that in a competitive market, commodity The ability to capture data in real time from a web site
costs should vary with market demand – and when is not difficult to develop. A more difficult challenge for
demand is higher, so should prices. In addition, in an industrial considering the option of real-time
the case of electric power there is the very real pricing is developing the real-time ability to make
consideration that peak demand burdens the energy decisions based on this data. In addition to the matter
delivery infrastructure more than light load, so of how can decisions be made, there is a concern
increasing peak demand can drive the need for about whether consumption can practically be
investment in enhancing that infrastructure in order to trimmed in response to price fluctuations. The cost of
support peak demand. electric power is only one component of the cost of
production of an industrial product, and in reality it
Time-of-use is not the same as real-time pricing. In may not be economically feasible to curtail production
the former case, the rate structure includes two, or at merely because the cost of electric power has
most three, fixed rates and the calculation simply momentarily crossed some threshold.
shifts between these as consumption varies over the
weekly cycle. Real-time pricing carries the concept Another option that may be considered is that of
further by calculating the price for delivered power interruptible rates. In this scenario, the energy
based on the actual market price at the instant of supplier offers a discounted rate in return for an
consumption. The theory is that this pricing structure agreement to curtail consumption upon demand,
encourages the consumer to develop a consumption usually with some provision for advanced warning of
pattern that is responsive to market demand for the need to take this step. Many industrials have
power – that is, to reduce consumption when overall opted for interruptible rates because of the obvious
system peak load drives prices higher, thereby savings, but without realistically anticipating that
helping to dampen spikes in market pricing. Real- curtailment orders would actually be issued. The
time pricing requires two critical elements – the first is experiences in California in early 2001 should be a
a mechanism to inform the consumer of the real-time reminder that “there is no free lunch” and that the
price. In many areas this information is available on possibility of actual interruptions is real.
the Open Access Same-Time Information System
(OASIS) web site maintained by the ISO or grid Distributed Generation
operator. Figure 6 shows regional real-time prices
One of the trends in the electric power industry
captured from the New York ISO web site; this data is
mentioned earlier was the increased significance of
updated at about five minute intervals.
the electric consumer through self-generation. The
option of installing generation within industrial
facilities actually predates formation of the grid as it is
known today. The option for industrial generation
was reinforced by the PURPA act of 1978, and the
more recent aspects of deregulation have increased
the flexibility that generation offers the industrial
concern.
notching, the ability to support voltage as motors start
One concern that must be recognized, however, is on the system, the ability to support fault current
that generation systems have outages. It is not required to achieve selective overcurrent protection,
unusual to see a requirement that generation outages and the ability to reliably detect that the source has
be scheduled on at least an annual basis. been disconnected from the grid in order to avoid
Furthermore, the possibility of forced outages must potential personnel safety issues while back feeding
be recognized. For example, the mean-time- the grid system, just to name a few.
between-failure of smaller combustion turbines is less
than a thousand hours, meaning that one or two One of the more significant activities currently
forced outages must be anticipated each year. underway in IEEE is P1547, a working group within
the Power Engineering Society to develop standards
It is important to carefully understand the Federal, for integrating distributed generation resources into
and especially the State regulations that govern the the overall grid. The major challenge this group
connectability of generation, and the implication of faces is that the technical issues that exist with small,
these regulations on the operating economics of distributed generating units are fundamentally no
proposed generation. For example, in some states, different from those associated with large, central
regulations exist that impose a demand charge on all stations, but the cost of addressing those issues on
power generated and consumed within an industrial small units represents a much large fraction of the
facility if that facility remains connected to the grid total installed cost of the system. Therefore, their
and if the generation displaces power that previously challenge is to find acceptable technical solutions to
had been purchased from a commercial supplier. very old problems that will not impose
disproportionate financial burdens on emerging,
Beyond the question of unit sizing and economics, socially desirable technologies.
options also exist for the form of energy conversion.
Obviously, the conventional solutions include steam Rental Equipment
turbines, combustion turbines, and various forms of
With the rapid and sometimes unexpected
reciprocating engines. The choice between these
consequences of deregulation, such as the power
depends on a variety of non-electrical factors
outages experienced in California, many electric
involving the practicality of true cogeneration,
users are looking for temporary power for various
economics and maintenance requirements. Usually,
reasons. An industrial can rent power generation or
but not always, continuous, or base-load generation
distribution equipment for temporary peak shaving,
scenarios favor steam turbines or combustion
power during plant shutdowns, emergency power
turbines in combined-cycle applications, while
during natural disasters or to provide improved power
reciprocating engines tend to gravitate toward
quality until a permanent solution can be
emergency generation or peak-shaving scenarios.
implemented. Temporary power generating solutions
range from a few kilowatts to over 20 MW, can be set
Alternative Forms of Generation
up quickly and can be rented by the day, week,
The popular press has carried numerous articles month or year. The options may include diesel or gas
about alternative forms of generation. Micro-turbines, reciprocating engine or gas turbine generators.
fuel cells, wind turbines, and photovoltaic cells are all Response to the disaster at the World Trade Center
technologies that are under some form of included several large temporary generation units
development. These technologies currently are installed on an emergency basis.
available only in fairly small unit sizes. Wind turbines
have achieved some degree of respect and have a Distribution equipment such as transformers, low
considerable penetration in some areas of the US, voltage and medium voltage switchgear and even
but in aggregate still represent a very small fraction of prepackaged unit substations can be provided on
the total grid supply. mobile units.
Alternate generation technologies often present Power Quality

technical challenges, especially in the context of As mentioned previously, voltage interruptions, dips
industrial applications. Many of these technologies and sags can cause disruptions of production that
produce electric power either at high frequency, or as can lead to significant costs to the business. A
DC, and the power must be converted to AC for business decision should be made prior to
distribution and utilization. The power converter that implementing a fix to a power quality issue. The cost
sits between the source and the distribution system to apply a voltage dip or sag solution may be
itself then presents limitations and challenges – expensive.
power quality concerns due to harmonics and
preferable if the depth of the voltage
The user should consider the following elements disturbance is severe.
when evaluating a power quality solution:
• Constant Voltage Transformers (CVT) –
1. How vulnerable is the process to various Special transformers that can maintain rated
types of disturbances? output voltage to compensate for voltage
2. What is the net cost of production outages sags and dips of moderate depth.
due to these disturbances?
3. How effective is a particular solution in • Static Series Voltage Regulators (SSVR) –
avoiding these outages? Devices that correct voltage dips and sags
4. How does the cost of the solution compare to due to nearby unbalanced faults or extended-
the savings that can be realized? duration unbalanced loading. Especially
effective for maintaining production continuity
There are several elements of cost associated with during single-phase-to-ground faults on the
voltage interruption that should be recognized and utility supply.
quantified in the economic evaluation: • Superconducting Magnetic Energy
• Cost of Lost Production – In the simplest Storage (SMES) – Devices that store energy
case, this is the incremental margin on in an electromagnetic field within a
product that is not manufactured and supercooled electromagnet and can rapidly
therefore cannot be sold. release that energy upon demand from the
power system. SMES systems are
• Cost of Damaged Product – If the characterized by their ability to supply fairly
interruption causes the quality of the product large amounts of energy over short periods of
to be unacceptable the cost of inferior time, usually seconds or less.
product should be accounted for.
• Uninterruptible Power Supply (UPS) –
• Cost of Maintenance – The cost of reacting Devices that supply critical loads by a storage
to a disturbance. This includes everything battery that is trickle-charged from the power
involved in restoring production, including system. UPS systems are usually designed
diagnosing and correcting the problem, to provide power long enough to permit an
cleanup and repair and environmental costs. orderly shutdown of the process, typically a
few minutes.
• Hidden Costs – This factor may be the most
difficult to quantify but can be significant. If • Battery Energy Storage Systems (BESS) –
the impact of the voltage dip or sag is control Devices that store energy in a conventional
error, it is possible that the impact on product storage battery that is constantly trickle-
may not be readily apparent. Product recall charged from the power system, and can
and/or public relations costs can be return that energy to the AC system when
significant. required. BESS systems are characterized
by their ability to store and supply moderate
Many systems are available for controlling the impact amounts of energy (units up to 20 MW or
of voltage disturbances on the production process. more) over fairly long time periods, often
Listed below are system descriptions of some of the minutes to hours.
more common remedial systems.
• Cogeneration – Generation installed within
• Mechanically Switched Capacitors (MSC) - the industrial system. Unlike emergency
Conventional power factor capacitors that are generation, cogeneration operates
automatically switched on to support voltage continuously so that it is available at the
at a desired level. instant that interruption of the commercial
supply takes place.
• Static VAR Control (SVC) - Devices that
support voltage by supplying vars from
conventional capacitors and that can control Summary
this supply dynamically by electronically
modulating var absorption in shunt reactors The deregulated electric power industry continues to
evolve. The cement manufacture will have to make
• Static Condenser (STATCON) - Electronic many decisions in his consumption and purchase of
devices that supply vars dynamically to power to optimize his business operations. No two
correct for voltage dips. STATCON offers situations will be the same. The cement manufacturer
better response time than SVC, and may be must be knowledgeable about what is happening with
federal and state regulations, how the electric power These can be trying times for the electric power
industry is responding to the changing regulations consumer and knowledge is the key to unlock returns
and the marketplace, what the possible on your energy dollar. Hopefully this paper has
consequences of the changes are to the quality of provided some food for thought as to how to take
power he receives, and how his company consumes advantage of the deregulated power era.
and purchases electric power throughout the
company locations.
References
[1] K. E. Nicholson, R. L. Doughty, L. Mane, G.

Miranda, F. B. Pulaski, “Cost Effective Power
Management Systems”, IEEE Industry
Applications Magazine, March/April 2000.
[2] T. A. Mulholland, “Deregulation, The California
Experience, and Implications for Western
Markets”, IEEE-IAS/PCA Cement Industry
Technical Conference, 2001.
[3] J. Gilbert, “Opportunities for Cement Plants in the
Changing Electricity Industry”, IEEE-IAS/PCA
Cement Industry Technical Conference, 2000.
[4] P. Banak, J. Rebuck, C. Eaves, M. Troia, “The
Installation of an Advanced Power Management
System for Reduced Operating and Maintenance
Expense in a Cement Plant”, IEEE-IAS/PCA
Cement Industry Technical Conference, 2000.
[5] P. Fairley, “A Smarter Power Grid”, Technology
Review, July/August 2001.
[6] R. C. Dugan, T. E. McDermott, G. J. Ball,
“Planning for Distributed Generation”, IEEE
Industry Applications Magazine, March/April 2001.
[7] GE Power Systems Energy Consulting, Applying
Voltage Disturbance Remediation Systems, March
1997.
[8] R. C. Dugan, M. F. McGranaghan, H. W. Beaty,
M. Samotyj, Electrical Power Systems Quality,
New York, NY, McGraw Hill, 1996.
[9] T. J. Overbye, J. D. Weber, “Visualizing the
Electric Grid”, IEEE Spectrum, February 2001.
[10] W. Sweet, E. A. Bretz, “Energy Woes”, IEEE
Spectrum, July 2001.
[11] U. S. Department of Energy – Energy Efficiency
and Renewable Energy Network (EREN) website:
http://www.eren.doe.gov/
[12] U. S. Department of Energy – Energy Information
Administration (EIA) website:
http://www.eia.doe.gov/cneaf/electricity/page/restr
ucture.html
RECOVERY OF INDUSTRIAL HEAT IN THE CEMENT INDUSTRY
BY MEANS OF THE ORC PROCESS

Hilel Legmann
Ormat Industries Ltd
Abstract
The 1,500 kW Organic Rankine Cycle heat recovery system at the Heidelberger Zement AG Plant in
Lengfurt is the first of such systems supplied to the cement industry. This non-polluting power plant
recovers the unused heat available from the grate cooler and generates electricity on a continuous basis
without interfering with the core, clinker production process. Operation of the power plant results in an
annual CO2 emissions reduction of 7,600t equivalent to 29.1% related to the CO2 amount correlated to
the electrical energy used in the plant and 2.63%, related to the total produced CO2.
1. Introduction
Even for an optimized cement burning process, significant heat loss, mainly caused by the heat
of the waste gases, still occurs. The heat balance of a kiln plant reveals that preheater waste
gases and cooler exhaust air account for more than 30% of that heat loss (see Table 1).
TABLE 1
HEAT CONSUMPTION FOR CEMENT CLINKER BURNING
Heat Quantity Referred to Clinker
kJ/kg %
Theoretical heat requirement 1760 54
Waste gas heat (preheater) 755 23
Exhaust air heat (grate cooler) 355 11
Wall heat loss (preheater, kiln, cooler) 335 10
Heat discharged in clinker 65 2
Total 3270* 100
* Note
Data reflects the energy consumption as per 1987. The value for 1998 was 2,905 kJ/kg and the
specific value for Heidelberger Zement is 2,800 kJ/kg. In principle, direct utilization of these waste
heat sources, e.g. for drying of raw material, coal or intergrinding matter, yields maximum
efficiency and may be utilized. However, there are numerous plants where this utilization is either
not possible or is not required, so that the unused heat is indeed lost. The economic order of
magnitude of such losses in a typical kiln line of 2,000 t/d capacity with a 4-stage cyclone
preheater and grate cooler can be assessed as follows: Assuming a preheater waste-gas
temperature of 350°C and grate cooler exhaust-air temperature of 275ºC, approximately
1,100 kJ/kg of clinker of unused heat are lost. When firing coal of a net calorific value of 23,000
kJ/kg, the annual loss to be attributed to unused process heat is approximately US $1.0 – $1.6
million. Such an immense loss of energy should be tackled. Electricity generation is proposed
here as the preferred solution.
2. Electrical Power Generation From Unused Heat Sources
One approach to overcome this economically unsatisfactory situation is to use the waste heat for
the generation of electrical power. In the past few years some cement plant operators have
installed steam waste-heat boilers in the waste-heat piping of their plants and have utilized the
process heat with or without additional external burning to operate a steam generator set.
Reports on these systems have been published.
However, on close examination, the conventional steam technology implies certain drawbacks
with respect to the cement production process:
The use of the relatively low temperature grate-cooler exhaust air, available at continuously
varying temperatures ranging from 330ºF/165ºC to 580ºF/305ºC (see Diagram 1), involves
difficulties with respect to stable steam turbine operation. To overcome this drawback, the
exhaust air temperatures have been raised, in some cases beyond the level required for
clinker burning through additional fuel gas firing. This has increased the fuel consumption in
the plant to unacceptable levels.
As an alternative approach it was suggested to interconnect the steam boilers of grate-cooler

and preheater, whereby the grate-cooler boiler acts as economizer and tertiary evaporator for
the boiler installed for the cyclone preheater. However, not many such systems (if any) have
been installed in the industry.
D ia g r a m 1
T e m p e r a t u r e F lu c t u a t io n in a C lin k e r C o o le r
350
325
300
275
Temperature[C]
250
225
200
175
150
11:33:45
12:41:45
13:49:45
14:57:45
16:05:45
17:13:45
18:21:45
19:29:45
20:37:45
21:45:45
22:53:45
00:01:45
01:09:45
02:17:45
03:25:45
04:33:45
05:41:45
06:49:45
07:57:45
09:05:45
10:13:45
11:21:45
12:29:45
13:37:45
14:45:45
15:53:45
17:01:45
18:09:45
19:17:45
20:25:45
21:33:45
22:41:45
23:49:45
00:57:45
02:05:45
03:13:45
04:21:45
05:29:45
06:37:45
Cooler
Time [ hr Air
] Temperature
1528
The points listed above and additional considerations such as the instability and low efficiency of
the relatively small steam turbines at partial-load operation, caused by the fluctuation of the grate-
cooler air; the necessity to have a dedicated steam plant operator on a shift basis increasing the
personal cost of the plant; the necessity to shut down the complete system in case of failures -
have been the background for a new developed concept.
The heat recovery system operates on the thermodynamic Organic Rankine Cycle (ORC). The
air cooled power module generates electricity from the sensible heat recovered from the clinker
cooler. In order to recover the waste heat from the clinker cooler air, a conventional tube heat
exchanger was installed after the precipitator. The heat is transferred to the OEC by a thermal oil
flow circulating in a closed loop system. The OEC uses the hot thermal oil to evaporate a
hydrocarbon motive fluid (pentane). The pentane vapors are expanded in the turbine, then
condensed in the air cooled condenser and returned in a closed loop to the vaporizer by means
of a high pressure pump (see Diagram 2).
This concept was implemented for the first time in the cement industry at the Heidelberger
Zement AG plant in Lengfurt in Germany in 1998. The one hundred year-old plant is a modern,
well operated and maintained plant having a yearly production capacity of over one million tons of
clinker.
Diagram 2
1511
3. The Thermal Oil - OEC System
3.1 General Description
Following the initial design, it was intended to circulate a thermo oil flow in a closed loop enabling
the recovery of approximately 9.5 MW heat from the clinker cooled air and 10.7 MW from the
preheater gases (total 20.2 MW heat) and feeding it to one air cooled 3.5 MW OEC unit. Taking
into account some budget restrictions, the final design ignored the heat from the preheater gases
and recovered 8.2 MW from the clinker cooler, while feeding a 1.5 MW unit.
Diagram 2 outlines the basic operating principle of the thermal oil/power system for the
generation of electrical power from waste heat in a cement plant:
The heat contained in the wasted clinker cooler is transferred to a thermal oil flow circulating
in a low-pressure loop. The thermal oil heat content is transferred to the OEC motive fluid in
the vaporizer.
The power generation process itself (preheating, vaporizing, expanding, condensing and
pumping of the organic motive fluid) takes place in the OEC unit in a fully automatic operating
mode.
4. Lengfurt – Design Criteria And Operating Experience
Electrical power generation by means of the OEC is a modern, expedient, field-proven option.
The OEC is designed to operate in heat recovery systems transferring the low-temperature heat
sources available in the process, i.e., in the cement industry into electricity. It’s operating fully
automatically and unattended thus lowering significantly the operation and maintenance costs.
The design criteria governing the OEC developments (see Table 2) have been as follows:
1. Matching the design criteria to the specific characteristics of the heat source and to the
environment. Thermal oil (Mobiltherm 594) and the working organic fluid (pentane) were
chosen to enable the system to work within the existing temperatures and pressures.
2. A series of by-passes installed in the hot air flow, thermal oil and pentane loops warrant no
negative interference with the hosting plant.
3. Low speed turbine (3000 rpm) direct connected to the asynchronous generator, and
computerized controls including automatic synchronization and preprogrammed save shut-
down procedure are the key elements for simple operation and maintenance.
4. Pre-fabricated and pre-tested components were the key elements to ensure rapid and low
cost installation no cost overruns.
5. Field proven components are used to ensure maximum system availability.
6. The organic power cycle enables the degree of flexibly required for the continuously changing
heat source conditions in flows and temperatures. (The grate cooler air temperature is
fluctuating between 644°F/340°C and 356°F/180°C causing the thermo oil to fluctuate
between 446°F/230°C and 248°F/120°C. Coping with such fluctuations, the power unit is
generating between 1,500 and 400 kW (see Diagram 3).
Table 2 – Technical Data
Plant Data
Plant production : 1 Mta
Heat recovered from the grate cooler : 8.2 MW
Grate cooler air flow : 150,000 Nm3/hr
temperature in/out : 275/125ºC
Thermal Oil Cycle
flow : 85 t/hr
temperature in/out : 230/85ºC
OEC Data
Rated generation capacity : 1.5 MW
Vaporizer type : Tube and shell type
Materials : TEMA Class “C” BEM
Tubes and shell : Carbon steel
Cooling Side Data
Cooling media : Ambient air
OEC Turbine
Turbine manufacturer : ORMAT
Turbine type : Specially designed impulse type
Rotation shaft speed : 3,015 rpm
Generator
Rated output : 1.5 MW
Rated voltage : 690 Volt/50 Hz/3-phase
Type : Induction
Rotation speed at full load : 3,015 rpm
Number of poles : 2
Motive Fluid
Working fluid type : Hydrocarbon (pentane)
Electricity Generation : 1,300 kW continuous operation, varying according to
inlet conditions (grate cooler air temperature variation).
The generated power is equivalent to about 12% of
the plant’s own electricity demand.
Mode of Operation : Unattended, fully automatic, remote monitored
Additional Effects of the Projects : Reduces pollution of CO2 by 7,600 t/year
Delivery Time : 25 weeks
Erection Time : 3 weeks
System Availability : Above 98%
Recorded (yearly) O&M Cost : Spares (less than) $ 5,000.00
over the last 30 months of Workmanship $10,000.00
operation
TOTAL $ 5,000.00
equivalent to
0.002$/kWh
OEC Generator Output [kW]
1100
1300
1500
100
300
500
700
900
11:33:45
1531
12:42:25
13:51:05
14:59:45
16:08:25
17:17:05
18:25:45
19:34:25
20:43:05
21:51:45
Generator Output
23:00:25
00:09:05
01:17:45
02:26:25
03:35:05
04:43:45
05:52:25
07:01:05
Cooler
08:09:45
Time
09:18:25
10:27:05
hr ]
11:35:45
Diagram 3
12:44:25
Air [Temperature
13:53:05
15:01:45
IN A CLINKER COOLER
16:10:25
17:19:05
18:27:45
19:36:25
20:45:05
21:53:45
23:02:25
TEMPERATURE AND OEC OUTPUT FLUCTUATION
00:11:05
01:19:45
02:28:25
03:37:05
Thermal Oil Temperature
04:45:45
05:54:25
07:03:05
150
200
250
300
350
400
Temperature [°C]
5. Summary
Summarizing the initial 3.5 years of operation, some unique features should be highlighted:
1. The OEC unit has proved that its operation is extremely flexible over a wide range of “off
design conditions”. As a result of changing operating conditions power fluctuations from 400
to 1,400 kW occurs within minutes. The ability of the system to cope and operate in such off-
design ranges without any damage, is unique to the OEC. No other power cycle has proven
similar abilities, suggesting that today, this may be the best available technology.
2. During it’s operation the power module has achieved yearly availabilities over 98% proving
that it operates in the process industry with the same high availability as proven in the
geothermal industry where the system is used to deliver high quality base load power.
3. The operation of the system is fully automatic, with no interference to the clinker production
and without causing any additional work or disturbance to the kiln operators.
4. The operation and maintenance of the system is simple and can be easily handled by the
plant’s own mechanical and electrical staff.
5. During its commercial operation, the system generated some 28 Gwhr without incident or
malfunction, with no major overhaul of components, and no loss of organic fluid.
Analyzing the starting goals of this pilot/commercial project, we conclude that the installation
easily achieved all targeted values. Furthermore, it showed that the ORC power cycle was
successfully used for the conversion into electrical energy of low temperature waste heat
available in the cement plants and could be considered as a viable and feasible alternative to the
steam-generating power plants.
6. Reference List
1. M. Deussner, J. Kümmel, H. Legmann - Waste Heat Recovery in the Cement Industry by

Means of the ORC Process - AFCM, Kuala Lumpur 1988
2. E. Baatz – Generation of Low Temperature Heat via an ORC Power Plant in the Cement
Plant of Lengfurt
3. H. Legmann – ORC at Heidelberger Zement Plant in Lengfurt – One Year of Operation

Experience
2002 IEEE-IAS/PCA
Current Industry Technical Conference
Date: May 8, 2002
Place: Jacksonville, FL
Cogeneration Systems in Material Drying Applications
Presented By:
Dwayne R. Walker, P.E.
Cummins & Barnard, Inc
5405 Data Court Suite 100
Ann Arbor, Michigan 48108
(734) 761-9130
Fax (734) 761-9881
E-Mail: [email protected]
Web: http://www.Cummins-Barnard.com
CUMMINS & BARNARD, INC

CONSULTING ENGINEERS / SINCE 1932
COGENERATION SYSTEMS IN MATERIAL DRYING APPLICATIONS
By:
Dwayne R. Walker, P.E.
Senior Mechanical Engineer
Cummins & Barnard, Inc
ABSTRACT
The potential exists to offset fuel purchases associated with material drying, as well as electric energy
purchases, with the installation of a combustion turbine generator for cogeneration at cement plants, and
other plants where material drying in kilns or furnaces is performed. Cogeneration, also referred to as
combined heat and power, is the simultaneous generation of two forms of useful energy from one
energy/fuel source. A cogeneration concept has been considered due to the coincident electric power
and thermal loads present at cement manufacturing operations. For this application, a natural gas fuel
source would be used to generate both electric energy and heat in the form of turbine exhaust gases.
The purpose of this document is to advise cement manufacturing facilities of the potential of
cogeneration, and present a concept of how cogeneration could be implemented. This document will also
define the pertinent economic drivers and determine how project feasibility is affected. Major
cogeneration plant equipment and systems are described. Feasibility for a test case, which considers
installation of a nominal 13 MW gas turbine, is presented.
INTRODUCTION
A combustion turbine could be installed to generate electricity and to displace a portion of the fuel fired in
drying kilns and furnaces. Typically, ambient temperature air is heated by direct firing of natural gas to
temperatures of approximately 1700 °F. Displacement of fuel required to heat ambient air utilized for this
purpose is accomplished by ducting turbine exhaust to the inlet of the drying kiln or furnace.
In this case, turbine exhaust gases would be heated to drying temperature and conveyed in high
temperature ductwork to the furnace housing. For existing furnace installations, modifications to the
furnace housing to accept the turbine exhaust flow would be required. Existing burners along with their
primary and tempering air system fans and ductwork could remain in place as a backup to the turbine
exhaust system to maintain operations when the turbine would be down for maintenance.
The exhaust gas leaving the turbine at a temperature of approximately 925°F has sufficient oxygen
content to sustain additional natural gas firing in a duct burner as needed to achieve the higher
temperatures required to replicate drying furnace conditions.
TEST CASE ANALYSIS
The test case considers a cement manufacturing facility utilizing two raw grind dryers in their operations.
Ambient temperature air is heated using natural gas furnaces to provide heat for the dryers. Each dryer
requires 40,000 scfm of heated air at 1700°F for drying. The plant electric load averages no less than the
turbine output capabilities, allowing full-load operation of the turbine, and full utilization of energy
generated.
Major Equipment
The combustion turbine-generator considered for the test case analysis has a nominal capacity of 13
MW. The turbine has an exhaust flow of approximately 400,000 lb/hr at 925°F. The turbine should be
typically selected based on the quantity of turbine exhaust gas flow, the object being to provide a good
match for existing drying furnace flow requirements. However, plant electric demand must also be
considered. For the drying application test case considered in this analysis, the drying furnace flow
requirement is 40,000 scfm per dryer. Therefore, the 83,800 scfm (230,700 acfm) of exhaust flow
available from the turbine would coincide well with the requirements of both raw grind dryers. Turbine
exhaust flow would be split to supply both dryers.
The combustion turbine could be housed in a separate building/enclosure if space within existing plant
buildings were unavailable. Distance from the drying application should be kept at a minimum to prevent
long ductwork runs. Proximity to the point of interconnection with the plant electrical system is also a
factor. The unit would be equipped with an inlet air filter and remote air-cooled oil cooler. The inlet air
filter could be equipped with a pulse-type cleaning system if a dusty environment exists. A gas
compressor may be required to furnish nominal 350-psig gas to the turbine if the required pressure is
unavailable from the utility.
The application considered would require a modulating bypass diverter immediately downstream of the
combustion turbine. The diverter would be equipped with a bypass stack to atmosphere with a silencer
such that the process could be bypassed for turbine maintenance or startup/shutdown. The exhaust gas
would be split downstream of this diverter and each stream dedicated to one of the existing drying lines.
Each process gas stream would be equipped with a natural gas fired duct burner controlled by a
temperature signal from the respective process line control system. Corrosion resistant high temperature
ducting will be required to convey turbine exhaust gases to the furnace. Isolation dampers will be required
to modulate exhaust flow and prevent backflow.
Preliminary Feasibility
A benefit and cost analysis was developed to determine preliminary feasibility for test case application
specifically, and for similar applications in general. Due to the large volumes of natural gas consumed by
the gas turbine-generator, the analysis is highly sensitive to this cost. Feasibility is also impacted by the
cost of displaced electric energy at the facility. Capital cost is the other major driver. The following major
notes and assumptions form the basis of this life-cycle analysis:
Economic Notes and Assumptions
• Total project cost = $11.0MM USD

• Cost of electric energy displaced = $0.060/kWh (Year 1 - escalated at 2% annually)
• Cost of natural gas = $5.00 per MMBtu (Year 1 – escalated at 2% annually)
• Cost of turbine maintenance = $0.005/kWh (Year 1 – escalated at 3% annually)
• Cost of standby electric energy - $0.090/kWh (Year 1 - escalated at 2% annually)
Financial Notes and Assumptions
• For simplicity, first year net income was used to determine IRR.
• IRR determined by net savings divided by project cost
Operational Notes and Assumptions
• 8,400 hour per year gas turbine and dryer operation (96% availability)
• Continuous full-load operation of gas turbine
• The waste heat benefit is the elimination of natural gas required to heat 80,000 scfm of
ambient temperature air to turbine exhaust temperature (925°F).
Preliminary project feasibility for the test case based on the above assumptions is summarized below:
Project Cost: $11.0 MM USD

Year 1 Benefits: $9.89 MM
Year 1 Expenses: $8.00 MM
Year 1 Net Operating Income: $1.89 MM
Internal Rate of Return 17.2%
Chart 1 displays internal rate or return vs. natural gas cost for various displaced electric energy costs.
This chart may be used to assess preliminary feasibility based on the assumptions outlined above for a
potential application.
COGENERATION - ELECTRIC GENERATION WITH MATERIAL

DRYING
ACHIEVED IRR VS. NATURAL GAS PRICE & DISPLACED ELECTRIC
ENERGY COST
40%
35%
30%
25%
20% 23.2%
Achieved IRR
21.7%
20.2%
18.7%
15% 17.3%
15.8%
14.3%
10% 12.8%
11.3%
5%
0%
-5%
-10%
-15%
4.00 4.25 4.50 4.75 5.00 5.25 5.50 5.75 6.00
Natural Gas Price ($/MMBtu)
$0.040/KWh $0.045/kWh $0.050/kWh $0.055/kWh
$0.060/kWh $0.065/kWh $0.070/kWh
COGENERATION PRO FORMA FEASIBILITY ANALYSIS
TEST CASE: SOLAR TITAN 130 GAS TURBINE PLANT WITH EXHAUST WASTE HEAT RECOVERY (RAW GRIND DRYERS
PERFORMANCE:
JANUARY FEBRUARY MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER AVERAGE
ELECTRIC DEMAND & GT PERFORMANCE:
CEMENT PLANT PEAK IMPORT POWER (KW) 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000 21,000
CEMENT PLANT AVG POWER DEMAND (KW) 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700 14,700
GAS TURBINE GROSS ELECTRIC OUTPUT (KW) 14,676 14,641 14,170 13,580 13,107 12,678 12,335 12,470 12,886 13,322 13,807 14,376 13,504
GAS TURBINE NET POWER OUTPUT (KW) 14,358 14,323 13,859 13,277 12,811 12,388 12,050 12,183 12,593 13,023 13,502 14,063 13,203
ELECTRIC POWER DISPLACED:
PEAK POWER DISPLACED (KW) 14,358 14,323 13,859 13,277 12,811 12,388 12,050 12,183 12,593 13,023 13,502 14,063 13,203
AVERAGE POWER DISPLACED (KW) 14,358 14,323 13,859 13,277 12,811 12,388 12,050 12,183 12,593 13,023 13,502 14,063 13,203
ELECTRIC EXPORT CAPACITY:
PEAK POWER EXPORTED (KW) 0 0 0 0 0 0 0 0 0 0 0 0 0
AVERAGE POWER EXPORTED (KW) 0 0 0 0 0 0 0 0 0 0 0 0 0
SUPPLEMENTAL ELECTRIC DEMAND:
POWER IMPORTED - PEAK (KW) 0 0 0 0 0 0 0 0 0 0 0 0 0
POWER IMPORTED - AVERAGE (KW) 0 0 0 0 0 0 0 0 0 0 0 0 0
THERMAL DISPLACEMENT:
DRYER AVG HEATING DEMAND (MMBTU/HR) 88.9 88.9 88.9 88.9 88.9 88.9 88.9 88.9 88.9 88.9 88.9 88.9 89
EXHAUST FLOW (LB/HR) 412,200 411,600 404,800 396,000 388,500 381,600 375,900 378,100 385,000 391,900 399,500 407,800 394,408
EXHAUST TEMPERATURE (F) 910 911 915 922 928 934 939 937 931 926 919 913 924
GT EXHAUST RECOVERABLE HEAT (MMBTU/HR) 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78.4 78
FUEL:
GAS TURBINE HEAT RATE (BTU/KWH-HHV) 11,196 11,202 11,292 11,413 11,520 11,631 11,736 11,694 11,572 11,470 11,366 11,364 11,455
FCP HEAT RATE (BTU/KWH-HHV) 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853 5,853
GAS TURBINE NATURAL GAS (MMBTU/HR) 164.3 164.0 160.0 155.0 151.0 147.5 144.8 145.8 149.1 152.8 156.9 163.4 154.5
DRYER NAT GAS FUEL REDUCTION (MMBTU/HR) 78.2 78.1 77.7 77.1 76.6 76.1 75.6 75.8 76.3 76.8 77.4 77.9 77.0
NET INCREASE IN GAS CONSUMED (MMBTU/HR) 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1 86.1
UNIT COST DATA: SERVICE: CONTINUOUS FINANCIAL SUMMARY: $ x 1000

TERM (YEARS): 20
NATURAL GAS $5.00 /MMBTU COMMISSION DATE: 1-Jan-03 INTERNAL RATE OF RETURN (IRR) 17.11%
ELECTRIC SALES PRICE 0.00 MILS/KWH GT DRYER YEAR 1 OPERATING EXPENSES $8,003
WEEKS 50 50
YEAR 1 NET OPERATING INCOME $1,884
ELECTRIC ENERGY 60.00 MILS/KWH GT DRYER
OPERATING HRS 8400 8400
TURBINE MAINTENANCE 5.00 MILS/KWH DOWN HRS 360 360
AVAILABILITY 95.9% 95.9%
TEST CASE: SOLAR TITAN 130 GAS TURBINE PLANT WITH EXHAUST WASTE HEAT RECOVERY (RAW GRIND DRYERS)
PRO FORMA ANALYSIS: ALL VALUES $ X 1000 UNLESS INDICATED OTHERWISE
YEAR 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
ESCALATORS:
CPI 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00% 3.00%
NATURAL GAS 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00%
ELECTRIC ENERGY SALES 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00%
ELECTRIC ENERGY PURCHASES 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00% 2.00%
UNIT PRICING DATA:
ELECTRIC ENERGY PURCHASES:
AGGREGATE ENERGY (MILS/KWH) $60.00 $60.00 $61.20 $62.42 $63.67 $64.95 $66.24 $67.57 $68.92 $70.30 $71.71 $73.14 $74.60 $76.09 $77.62 $79.17
STANDBY ENERGY (MILS/KWH) $90.00 $90.00 $91.80 $93.64 $95.51 $97.42 $99.37 $101.35 $103.38 $105.45 $107.56 $109.71 $111.90 $114.14 $116.42 $118.75
ELECTRIC ENERGY SALES:
ENERGY (MILS/KWH) $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00 $0.00
FUEL PURCHASES:
NATURAL GAS ($/MMBTU) $5.00 $5.00 $5.10 $5.20 $5.31 $5.41 $5.52 $5.63 $5.74 $5.86 $5.98 $6.09 $6.22 $6.34 $6.47 $6.60
PRODUCTION & CONSUMPTION DATA:
ELECTRIC:
CEMENT PLANT ENERGY CONSUMPTION (MWH) 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772 128,772
TOTAL GENERATION (MWH) 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434 113,434
AVAILABLE GENERATION (MWH) 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901
ENERGY DISPLACEMENT 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901 110,901
ENERGY EXPORTED TO UTILITY 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
SUPPLEMENTAL ENERGY IMPORT 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
STANDBY ENERGY 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753 4,753
NATURAL GAS:
GAS TURBINE FUEL CONSUMPTION (MMBTU) 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202 1,298,202
DRYER FUEL DISPLACEMENT (MMBTU) 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546 646,546
REVENUE:
ELECTRIC:
PLANT ELECTRIC ENERGY DISPLACEMENT: $6,654 $6,787 $6,923 $7,061 $7,203 $7,347 $7,494 $7,643 $7,796 $7,952 $8,111 $8,273 $8,439 $8,608 $8,780
EXPORT ENERGY SALES $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
NATURAL GAS:
DRYER FUEL $3,233 $3,297 $3,363 $3,431 $3,499 $3,569 $3,641 $3,713 $3,788 $3,863 $3,941 $4,019 $4,100 $4,182 $4,266
TOTAL REVENUE $170,976 $9,887 $10,085 $10,286 $10,492 $10,702 $10,916 $11,134 $11,357 $11,584 $11,816 $12,052 $12,293 $12,539 $12,790 $13,045
ELECTRIC ENERGY UNIT PRICE ($/KW Average $0.100 $0.087 $0.089 $0.091 $0.092 $0.094 $0.096 $0.098 $0.100 $0.102 $0.104 $0.106 $0.108 $0.111 $0.113 $0.115
EXPENSES:
NATURAL GAS - GAS TURBINE $6,491 $6,621 $6,753 $6,888 $7,026 $7,167 $7,310 $7,456 $7,605 $7,757 $7,913 $8,071 $8,232 $8,397 $8,565
TURBINE MAINTENANCE / OVERHAULS $584 $584 $584 $584 $584 $584 $584 $584 $584 $584 $584 $584 $584 $584 $584
FIXED O&M/LABOR $400 $412 $424 $437 $450 $464 $478 $492 $507 $522 $538 $554 $570 $587 $605
VARIABLE O&M / CONSUMABLES $100 $103 $106 $109 $113 $116 $119 $123 $127 $130 $134 $138 $143 $147 $151
SUPPLEMENTAL & STANDBY ENERGY $428 $436 $445 $454 $463 $472 $482 $491 $501 $511 $521 $532 $543 $553 $564
INSURANCES $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
MANAGEMENT FEES $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
PROPERTY TAXES $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
LAND LEASE $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
TOTAL EXPENSES NPV $68,362 $8,003 $8,156 $8,313 $8,473 $8,636 $8,803 $8,973 $9,147 $9,324 $9,505 $9,690 $9,879 $10,072 $10,269 $10,470
UNIT COST ($/KWH) 15 Yr. Avg $0.083 $0.071 $0.074 $0.075 $0.076 $0.078 $0.079 $0.081 $0.082 $0.084 $0.086 $0.087 $0.089 $0.091 $0.093 $0.094
UNIT COST MINUS FUEL ($/KWH) $0.015 $0.013 $0.014 $0.014 $0.014 $0.014 $0.014 $0.015 $0.015 $0.015 $0.015 $0.016 $0.016 $0.016 $0.017 $0.017
NET OPERATING INCOME $33,265 $1,884 $1,928 $1,973 $2,019 $2,066 $2,113 $2,161 $2,210 $2,260 $2,310 $2,362 $2,414 $2,467 $2,521 $2,576
TEST CASE: SOLAR TITAN 130 GAS TURBINE PLANT WITH EXHAUST WASTE HEAT RECOVERY (RAW GRIND DRYERS)
PRO FORMA ANALYSIS (CONT'D) ALL VALUES $ X 1000 UNLESS INDICATED OTHERWISE
YEAR 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
DEPRECIATION 95% $1,267 $1,098 $952 $825 $715 $619 $537 $465 $403 $349 $303 $263 $228 $197 $171
AMORTIZATION $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
TOTAL NON-CASH CHARGES $1,267 $1,098 $952 $825 $715 $619 $537 $465 $403 $349 $303 $263 $228 $197 $171
EARNINGS BEFORE INTEREST & TAXES (EBIT) $617 $830 $1,022 $1,194 $1,351 $1,494 $1,624 $1,745 $1,857 $1,961 $2,059 $2,152 $2,240 $2,324 $2,405
DEBT SERVICE:
TOTAL PAYMENT $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044 $1,044
INTEREST PAYMENT $739 $712 $683 $651 $617 $579 $539 $494 $446 $393 $336 $274 $207 $133 $53
PRINCIPAL PAYMENTS $305 $332 $361 $393 $427 $465 $506 $550 $598 $651 $708 $770 $838 $911 $991
TAX BASIS -$122 $118 $339 $543 $734 $914 $1,086 $1,251 $1,411 $1,568 $1,723 $1,877 $2,033 $2,191 $2,352
INCOME TAXES:
STATE 6.50% $122 $125 $128 $131 $134 $137 $140 $144 $147 $150 $154 $157 $160 $164 $167
FEDERAL 35.00% ($43) $41 $119 $190 $257 $320 $380 $438 $494 $549 $603 $657 $712 $767 $823
GROSS RECEIPTS $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
CASH BASIS:
NET OPERATING INCOME (NOI) $840 $884 $929 $975 $1,021 $1,069 $1,117 $1,166 $1,216 $1,266 $1,318 $1,370 $1,423 $1,477 $1,532
RESERVE FUNDS $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0 $0
NET BEFORE TAX CASH FLOW $17,601 $840 $884 $929 $975 $1,021 $1,069 $1,117 $1,166 $1,216 $1,266 $1,318 $1,370 $1,423 $1,477 $1,532
LESS TAXES ($8,468) ($80) ($167) ($247) ($321) ($391) ($457) ($521) ($581) ($641) ($699) ($756) ($814) ($872) ($931) ($991)
NET AFTER TAX CASH FLOW $9,134 $760 $717 $682 $654 $630 $612 $597 $585 $575 $567 $561 $556 $551 $546 $541
NET PRESENT VALUE OF NET AFTER TAX $9,134
DEBT SERVICE COVERAGE:

AFTER TAX 1.58 1.73 1.69 1.65 1.63 1.60 1.59 1.57 1.56 1.55 1.54 1.54 1.53 1.53 1.52 1.52
BEFORE TAX 2.12 1.80 1.85 1.89 1.93 1.98 2.02 2.07 2.12 2.16 2.21 2.26 2.31 2.36 2.41 2.47
EQUITY RETURN: 100%

PRE-TAX 38.14% 40.15% 42.20% 44.28% 46.40% 48.55% 50.74% 52.96% 55.22% 57.52% 59.85% 62.22% 64.63% 67.08% 69.57%
AVG 53.30%
POST-TAX 34.52% 32.58% 30.99% 29.69% 28.63% 27.78% 27.10% 26.55% 26.12% 25.77% 25.49% 25.25% 25.03% 24.81% 24.57%
AVG 27.66%
CONCLUSION
The above analysis displays that the potential exists to offset cement plant fuel and electric purchases
with the installation of a combustion turbine generator for cogeneration. Because of the high sensitivity to
the cost of natural gas and displaced electric energy, these items significantly impact feasibility. The
proper combination of gas cost, electric energy cost, and capital cost could allow feasibility of a
cogeneration plant in material drying applications to meet expected financial hurdles.
HIGHER EFFICIENCY – WHAT IS THE REAL COST?
By:
John J. Stroker, Senior Project Manager, Penta Engineering Corporation
Abstract: With the problems that California had with electrical energy last year, it has only emphasized
the need to look at the true cost of high efficiency energy devices used by cement plants since electrical
energy is among the top expenses in producing cement. This paper focuses on low voltage induction
motors and lighting systems. The low voltage induction motors generally account for about 30% of the
electrical energy consumed by a cement plant. The lighting systems generally account for about 3-5% of
the electrical energy consumed by a cement plant but it is a fixed load that does not directly contribute to
production. Life cycle cost analysis is used to determine the true costs of this equipment.
I. INTRODUCTION
True equipment cost is defined as the total of the initial capital expenditures plus the operating costs
which include energy and maintenance. Even though the foremost requirement for a piece of equipment
is its performance in the production chain, nevertheless the justification for selecting one piece over
another should be based upon the initial capital cost plus operating expenses. The following is a review
of two selected areas of electrical equipment to test their cost of ownership.
II. PRESENT WORTH EVALUATION
With equipment requiring the use of electrical energy, there is the problem of weighing initial capital
expenditures against future operating expenses to determine the best return on investment. A method
that can be employed to determine the best return considering both the initial cost and the operating costs
is Life Cycle Costing. A calculation can be made that evaluates the cost of the wasted electrical energy.
This value can be determined by the following formula:
EF = CTn
Where:
EF = Evaluation Factor ($/kW)
C = Average energy cost ($/kWH)
T = Running time per year (Hrs/Yr)
n = Evaluation Life (Yrs.)
Average energy cost is not just the cost per kWH on the billing rate from the utility company. Average
energy cost should include the charge for energy used, fuel adjustments, demand charges, any power
factor penalties, and any other charges associated with the plant electrical service. Average energy cost
is calculated by dividing the total electrical energy cost for a year divided by the kWH’s used during that
year.
This evaluation factor (EF) can be further refined by including the time value of money and an escalation
rate for energy costs. These two factors tend to cancel each other. An equivalent rate can be calculated
using the following formula.
1 + R2
Re = −1
1 + R1
Where:
R1= Energy escalation (%)
R2= Cost of capital (%)
Re= Equivalent rate (%)
Adding in the time value of money and the energy escalation, the evaluation factor would be:
(1 + Re) n − 1
EFPW = CxTx n
R e (1 + R e )
Once the EFPW (Present Worth Evaluation Factor) has been established, it can be used plant-wide for all
types of equipment. It can be used to establish the Present Worth Value of Losses, which can be added
to the initial capital expenditure to obtain a total cost for bid analysis.
100
PWL = EFPW x kW ( − 1)
E
Where:
PWL = Present Worth Value of Losses
EFPW = Present Worth Evaluation Factor
E = Efficiency
kW =Power Consumed
The total cost of a piece of equipment is the sum of its initial cost plus the present worth value of losses
and maintenance costs.
III. EXAMPLES
Low Voltage Motors
Since they are literally all over the plant, low voltage motors are major consumers of electrical energy.
Therefore they should be viewed with regard to efficiency and durability when they are specified. For
example to withstand the environment of a cement plant, the severe duty type of motor should be
specified. An excellent standard to follow for cement plant applications is the IEEE-841 motor design.
Government regulations for motors began when the U.S. put energy efficiency into legislation as part of
wider energy conservation issues. Enacted in 1992, the Energy Policy and Conversation Act (EPAct)
finally became effective in October, 1997. It covers NEMA design A and B, three phase, general purpose
ODP and TEFC motors with base speeds of 1200, 1800, and 3600 RPM. This standard sets the
minimum nominal full-load efficiency allowed in the United States. Most motors in the United States are
now listed as Efficient and Premium Efficient. Following shortly after the October, 1997 effective date for
EPAct, an agreement was reached in Europe where applicable motors were placed in three zones: EFF1
– High Efficiency, EFF2 – Improved Efficiency, EFF3 – Standard Efficiency.
Guaranteed efficiencies should be required with bids and the value ($/kW) assigned to losses for bid
evaluation should be defined in the request for quotation. This way each vendor can put forth his best
offering. This evaluation figure would be the same evaluation factor explained earlier – EFPW. Penalties
to be assessed for failure to meet guaranteed efficiencies should be defined before any order is placed,
and efficiencies are best checked by having random motors of each size and type tested. The penalty
should be based on the same value for $/kW given in bid evaluation.
At this time, it would be well to digress a bit to discuss guaranteed efficiencies versus median efficiencies.
Motor efficiencies for a typical group of motors will tend to follow the typical bell curve. The median
efficiency is the value that 50% of the group is above and 50% is below. The usual guaranteed value is
the one that 95% of the group is above, and 5% is below. Efficiency should be measured at 50%, 75%,
and 100% loadings, with any particular known loading point also requested. The standard by which the
motors are to be tested should be defined. The same motor tested by different standards can produce
different results if the testing methods are not controlled.
Specifications should require IEEE 112, test method B or CSA Std C390, test method 1. These
procedures pin-point stray load losses where the greatest variance can occur from one motor to the next.
They also clearly define the procedure and require testing at full operating temperature or making
corrections for temperature differences. Finally, they require a dynamometer to provide the most
accurate data as a basis for accuracy improvement and stray load loss measurement.
Three different motor applications are examined in the tables below, each with a standard efficiency
motor and a premium efficiency motor, and considering different loading rates and schedules. All motors
are Totally Enclosed Fan Cooled (TEFC), 460 Volts, 1800 RPM, NEMA Design B, 40oC ambient.
MOTORS STUDIED
Motor Efficiency HP Efficiency at Efficiency at Efficiency at Power Power Power
50% Load 75% Load 100% Load Factor at Factor at Factor at
50% Load 75% Load 100% Load
I Efficient 5 87.5% 87.7% 87.5% 70.9% 80.6% 85.9%
I Premium-Eff 5 89.5% 89.7% 89.5% 67.0% 78.0% 84.0%
II Efficient 25 92.3% 92.6% 92.4% 74.3% 80.0% 83.8%
II Premium-Eff 25 93.6% 93.8% 93.6% 76.0% 83.0% 86.0%
III Efficient 100 94.4% 94.7% 94.5% 80.8% 85.3% 85.8%
III Premium-Eff 100 95.4% 95.7% 95.4% 81.0% 86.0% 87.0%
MOTOR COSTS
Motor Efficiency HP Voltage Speed %-Load Annual Initial Cost Total Cost
Hours
I Efficient 5 460/3ph/60 1800 75% 2000 $388 $5,715
I Premium-Eff 5 460/3ph/60 1800 75% 2000 $428 $5,484
I Efficient 5 460/3ph/60 1800 75% 8000 $388 $11,943
I Premium-Eff 5 460/3ph/60 1800 75% 8000 $428 $11,571
II Efficient 25 460/3ph/60 1800 75% 2000 $1,614 $21,300
II Premium-Eff 25 460/3ph/60 1800 75% 2000 $1,659 $20,704
II Efficient 25 460/3ph/60 1800 75% 8000 $1,614 $50,771
II Premium-Eff 25 460/3ph/60 1800 75% 8000 $1,659 $49,795
III Efficient 100 460/3ph/60 1800 75% 2000 $6,697 $74,219
III Premium-Eff 100 460/3ph/60 1800 75% 2000 $7,751 $73,764
III Efficient 100 460/3ph/60 1800 75% 8000 $6,697 $189,463
III Premium-Eff 100 460/3ph/60 1800 75% 8000 $7,751 $187,791
The loading and duty cycles in the above table are typical of cement plant operations, the duty
representing an 8000 hour/year operation (7 days per week, 24 hours per day) and the duty representing
2000 hours/year (5 days per week, 8 hours per day). In all of these examples, the initial cost difference
between the standard efficiency motor and the premium efficiency motor is lost many times over when the
total cost of the motor is considered. As loading and/or duty cycle drops, the advantage of the high
efficiency motor decreases but a cost advantage still remains in these examples. The loading of a motor
as shown is an important part because both efficiency and power factor are affected by motor loading.
Generally as shown in the first table the best efficiency occurs at 75% loading and the best power factor
occurs at 100% loading.
Maintenance is a significant factor in total cost for a motor. Because wasted energy in a motor’s
operation is generated as heat, less efficient motors tend to run hotter. For cooler running motors,
bearing lubrication and motor insulation last longer. Generally, motor bearing life doubles for every 10oC
reduction in operating temperature.
Power factor of the entire plant must be controlled for maximizing the distribution system and minimizing
the distribution losses. For example, a 1000 kVA transformer at .95 lagging power factor is rated for 950
kW of load, but the same transformer at .75 lagging power factor is rated for only 750 kW. However, a
more important reason for controlling the power factor could well be avoiding the penalty clause in the
electrical utility’s rate structure, and this must be done to minimize the average cost per kWH of electrical
energy. A motor should not be over-sized for its load. A motor should be sized to handle its rated load,
its starting load, and any overloading expected during its normal use (flooded screws, etc.). An under-
loaded motor has a detrimental effect on power factor. Over-voltage also has a detrimental effect on
power factor so that it is important that the low voltage substation transformer taps be properly set.
Relation of Efficiency and Power Factor to Voltage Variation
30
20
10 Volts
0 Eff.
-10 1 2 3 4 5 6 7 8 9 P.F.
-20
-30
Percentage Variations from Rated Values
We see here that as the voltage varies from its rated value, the power factor and efficiency also vary but
not in a linear manner.
If a motor cannot be sized to meet the changing load requirements of a particular process such as fans
and pumps where a damper or a control valve is used to control the loading, an adjustable frequency
drive may offer an economical alternative. Besides using only the energy required for the process, most
adjustable frequency drives are of the PWM design and these drives have a power factor of 95% lagging
over the top 50% of their speed range which is better than an induction motor alone. However, the
economic justification for using an adjustable speed drive is more complicated than choosing an energy
efficient motor. The cost difference between a standard starter and a drive is significant. The motor
should be matched with the drive or at least be rated for inverter duty. Care should be taken with lead
lengths to the motor and the harmonics generated into the power system need to be considered.
Lighting Systems
Lighting is an area for reducing energy costs since lighting does not directly contribute to production but is
a consumer of electrical energy. However it still remains true that increased productivity and safety are
the benefits of sufficient lighting. Government regulations governing lighting energy also began in 1992
when the U.S. put energy efficiency into legislation as part of wider energy conservation regulation. This
legislation eliminates the use of some types of incandescent lamps and standard fluorescent fixtures in
commercial and industrial applications. The following chart shows the basic characteristics for various
common light sources.
LIGHT SOURCES
LAMP WATTS INITIAL MEDIAN LUMENS COLOR

LUMENS LIFE PER TEMPERATURE
(HRS) WATT
MEDIUM BASE 100 750 1690 8 3000K
A21
INCANDESCENT
RAPID START 40 3150 18000 79 4200K
FLUORESCENT
LO-PRESSURE 35 4600 18000 131 -
SODIUM
HI-PRESSURE 150 15000 24000 100 2100K
SODIUM
HI-PRESSURE 250 27500 24000 110 2100K
SODIUM
METAL-HALIDE 175 6500 15000 37 3800K
METAL-HALIDE 250 9500 15000 38 3800K
METAL-HALIDE 400 34000 15000 85 3800K
WARM DELUXE 175 6500 24000 37 3300K

MERCURY
WARM DELUXE 250 9500 24000 38 3300K
MERCURY
WARM DELUXE 400 19500 24000 49 3300K
MERCURY
The preceding chart shows that the most in-efficient light source is the incandescent lamp. Its median life
is the shortest and would be possibly the highest maintenance source but it the least expensive initial
cost. The fluorescent lamp is a good replacement for the incandescent source. In the ‘cool-white” color it
is the farthest to the blue side of the color spectrum. If better color rendition is required other colors are
available at a minimal energy cost. The low-pressure sodium lamp has the best efficiency but is a true
monochromatic light source and that is why it shows no color temperature. However it does an excellent
job for security lighting applications especially working with some surveillance camera systems. The
high-pressure sodium lamp has excellent efficiency but is on the yellow side of the color spectrum. The
metal-halide lamp gives an excellent white light source but it is less efficient than the high-pressure
sodium lamp.
The lighting sources that are normally used in cement plants are the following. In un-attended plant
operating areas, the high-pressure sodium vapor lighting would be a common choice. Metal-halide
fixtures are often used in plant areas when better color rendition is desired such as shop and warehouse
areas. Fluorescent fixtures are usually used to light office areas, electrical equipment rooms, central
control rooms, etc. Incandescent lamps would only be used in storage closets or similar infrequently
used applications.
To illustrate the effect of three different light sources and operating times, we will do a tentative design of
a plant maintenance area lit to 50 fc using three different light sources and three different operating times.
Then using the previously described cost evaluation method, the following table shows the total costs for
the chosen lighting systems.
LIGHTING COSTS
Option Wattage Number of Annual Initial Cost Total Cost
of Fixtures Fixtures Hours
I 150 12 2000 $2,220 $4,738

I 150 12 4000 $2,220 $7,256
I 150 12 8000 $2,220 $10,224
II 250 16 2000 $3,296 $8,851
II 250 16 4000 $3,296 $14,406
II 250 16 8000 $3,296 $20,922
III 250 16 2000 $3,296 $8,851
III 250 16 4000 $3,296 $14,406
III 250 16 8000 $3,296 $20,922
The chart shows that least expensive option is No. I, the 150 W high-pressure sodium light fixtures. A
judgment would need to be made as to the affect the color would have on the work done in this area. If
the color is considered important to the work being done in the area, option No. II would be a better
choice. The efficiency of the light source affects the total cost for the light fixture over its life. The annual
hours as shown represent a single-shift operation, a two-shift operation, and a three-shift operation. The
time that a light fixture is lit has a great influence on the energy used and affects the maintenance costs
by affecting the lamp life. The goal is to have the light available when it is needed and to turn it off when
it is not needed. The variety of means for controlling lighting has grown greatly over the past years.
Timers have been used to control outside lighting for years, but they needed to be adjusted with the
seasons in the past. Newer models have a calendar built in to avoid this problem. Photocells are a good
way of controlling outside lighting, but care should be taken with regards to location and particular
attention should be taken in a cement plant installation to avoid dust build-up. Motion sensors can be
used in office and storage areas. PLC’s are a good means to control lighting and there are special
models that are built to mount within the lighting panel and directly control the lighting circuit breakers.
The lighting transformer is usually a dry type, wall-mounted with the secondary neutral grounded. The
primary is usually 480 Volts, but primary and secondary voltage levels could vary by location. If it is
three-phase, it has a secondary usually rated 208Y/120 Volts. It should be a delta-wye winding since
wye-wye or T windings have a high noise level. If it is single-phase, it has a secondary usually rated 240-
120 Volts. It is advisable even if the lighting service level is 480/277 Volts to use a lighting transformer or
down-stream ground-fault protection since it can be most troublesome to take down an entire production
department because a ballast failed to ground in a light fixture and tripped the ground fault relay on the
low-voltage substation main. The efficiency of a lighting transformer is a combination of its impedance,
usually between 2% and 5%, its temperature rise usually 80oC, 115oC, or 150oC rise, and its percentage
loading. A lower impedance means less losses, but it also means higher fault currents available. The
interrupting rating of the lighting breakers should be checked. The other half of the efficiency story is that
all of the heat developed by lack of efficiency of the lighting transformer in the electric room must be
removed by the electric room air conditioner.
IV. SUMMARY
Too often new electrical equipment is chosen on its initial cost alone and that initial price is a small piece
of its true cost. Equipment prices need to be evaluated on their total cost not just their initial price.
V. APPENDIX
Calculations for tables:
Assumed Values for Cost Analysis
C = $0.055/kWh - Average energy cost

R1 = 3% - Energy escalation
R2 = 6% - Cost of capital
N = 15 year - Evaluation life
RE = 2.9%
EFPW = $965 /kWH for 2000 Hrs per Year

EFPW = $1930/kWH for 4000 Hrs per Year
EFPW = $2895/kWH for 8000 Hrs per Year
Initial costs for motors and light fixtures are the estimated costs for the equipment and do not include
costs of installation.
Efficient Motor maintenance at √ HP x 2.2 hours/year ($30/hour)

Premium-efficient Motor maintenance at √ HP x 2 hours/year ($30/hour)
Lighting calculations:
Desired Lighting Level = 50fc
ROOM LENGTH (L) 160 FT.

ROOM WIDTH (W) 80 FT.
2
FLOOR AREA (A) 12800 FT
hrc: 15 FT.
ROOM
hcc: 6 FT.
DIMENSION
hfc: 3.5 FT.
SURFACE CEILING 50 %
REFLECTION WALLS 50 %
FLOOR 25 %
Number of Fixtures = (Footcandles Req’d) x (Floor Area)

C.U. x M.F. x Lumens per Fixture
M.F. = .75 (Maintenance Factor)
C.U. = 5.5 (Coefficient of Utilization)
Number of Fixtures = 12 (Option I)

Number of Fixtures = 16 (Option II)
Number of Fixtures = 16 (Option III)
Lighting maintenance is 1 hour per fixture at rated lamp life ($30/hour)

VI. REFERENCES
1. E. Paul Degarmo, “Engineering Economics”, The Macmillan Co., New York, New York.
2. R.A. Thompson, “Practical Plant Lighting”, IEEE Cement Conference, May, 1977.
3. Frank J. Bartos, “Efficient Motors Can Ease Energy Crunch”, Control Engineering, May, 2001
4. NEMA standard MG 1-1998, Revision 2

2002 IEEE-IAS/PCA
Current Industry Technical Conference
Date: May 7, 2002
Place: Jacksonville, FL
ENERGY EFFICIENCY IMPROVEMENT IN THE CEMENT INDUSTRY
ALLIED PARTNER
OFFICE OF INDUSTRIAL TECHNOLOGY
Presented By:
Brahm Segal
Power Correction Systems, Inc.
1800 South Robertson Blvd. PMB 419
Los Angeles, California 90035
(310)247-4848
Fax: (310)273-7719
E-Mail: [email protected]
Web: http://www.activeharmonicfilters.com
ENERGY EFFICIENCY IMPROVEMENT
IN THE CEMENT INDUSTRY
By:
Brahm Segal
President, CEO
Power Correction Systems, Inc.
ABSTRACT
There are numerous opportunities to improve energy efficiency and reducing the effects of
harmonic distortion in cement manufacturing power distribution systems. The first step is to
choose the most energy efficient equipment, not necessarily choosing the most inexpensive
machinery. The second step is to apply modern technologies to the power distribution systems
for the equipment. Active harmonic filters and power factor correction conditioners automatically
follow changing load conditions, resulting in energy cost reduction. Studies at the Power
Electronics Application center in Knoxville, TN, have shown that harmonics and low power factor
do create voltage and power losses in cables, contactors, and transformers reducing system
capacity and waste energy. Power system modeling, awareness of which technologies to use is
necessary to reduce energy costs. Analysis tools are being developed to assist design engineers
to optimize system efficiency.
INTRODUCTION
This paper identifies the efficiencies of various types of machinery, describes a hypothetical
cement plant case study on power factor correction, discusses harmonics and how to deal with
harmonics, and reveals new lighting technologies.
I. Optimizing Energy Efficiency by Process Technology and Power Factor Correction
A. Equipment Selection
The first step in optimizing energy efficiencies is to select the most efficient equipment. From the
crushing, grinding and conveying systems, to the lighting systems in the manufacturing facility,
proper selection is the key to optimization. The tables below show specific energy consumption
of equipment in the various stages of the cement manufacturing process.
Power Consumption in Various Crushing Systems

Crusher Type Size Reduction KWH/Ton
Gyratory 5-8 .38-.55
Jaw 5-8 .46-1.1
Impact (Fixed Hammer 80 .4-.6
Hammer Mill w/grate (single rotor) 20-30 1-2
Hammer Mill w/grate (double rotor) 30-40 1.3-2.0
Swing Hammer without grate 10-20 .75-.8
KWH/Ton for Raw Material Grinding (150 TPH at 4% Moisture)
Equipment VRM Ball Mill Roller Press
Secondary -- .60 .30
Crusher
Tertiary Crusher -- .40 .40
Grinding Mill 8.00 15.0 9.00
Classifier .70 .70 .70
Disagglomerator -- -- 1.50
Bucket Elevator .20 .80 1.50
Conveyors -- .60 .90
Mill Fans 5.31 2.50 3.20
total 14.20 20.60 17.50
Power Consumption of Blending Equipment

Blender Type KWH/Ton
Mechanical System 2.00 - 2.50
Air Fluidized System 1.00 - 1.50
Gravity System - Inverted Cone 0.25 - 0.50
Gravity System - Multi-outlet 0.10 - 0.13
Power Consumption in Typical Kiln Feed Conveyor Systems

Air Lift Dense Screw Elevator
System Phase Pump
Air Power Consumption (kW) 180.8 104.3 152.0 --
Mechanical Power 11.2 -- 81.9 74.5
Consumption (kW)
Miscellaneous Power 27.9 13.0 6.2 7.4
Consumption (kW)
Total Power Consumption (kW) 219.9 117.3 240.1 81.9
Power Consumption (kWh/t) 1.1 .59 1.2 .41
Typical Fan Drive Capacities (One Million Ton/Year Cement Plant)

Plant Equipment Fan/Blower Total Equipment
Capacity (kW) Capacity (kW)
Crushing 200 2700
Raw Material Grinding 2000 8000
Kiln and Kiln Feed 3100 3800
Cooler 1700 2550
Coal Mill 750 1750
Cement Mill 475 6700
Packing 250 750
Total 8475 28250
Percent of Total Power 30% 100%
Typical Energy Consumption 36 kWh 130 kWh
kWh/t
Energy Savings with 10% 3.6kWh --
Increase in Fan Efficiency
Energy Consumption in Cement Grinding (150 TPH, 3600 Blaine)
Equipment Roller Press with V-Separator Traditional Ball Mill
Grinding Mill 17 33
Separators .6 1.0
Separator Fan 2.9 3.5
Mill Vent Fan -- .5
Auxiliaries 4.5 3.5
Total Power 25 42 (65% extra)
3
Ventilation m /hr m3/hr
Mill Venting -- 55,000
Roller Press Venting 20,000 --
Separator Venting 25,000 300,000
Nuisance Venting 25,000 25,000
Total 70,000 380,000(500% extra)
See Reference 5.
B. Power Factor Correction
Based on a typical 500,000 Tons/Year cement plant, two studies were hypothetically conducted.
One based on kVA demand billing and the other based on kW demand. The hypothetical study
was based on a study of 49 motors (none under 10 HP). All variable speed drive motors were
eliminated from the calculations because a harmonic study is required before correction.
The study was based on an average 80% power factor with a correction to 99% for the 49
motors. With this approach, 1470 kVAR will be added to the system at an estimated cost of
$54,000.00, not installed. Installation may reach $12,000.
In the kW demand report, only the line losses and voltage improvement can be addressed for an
annual saving of $40,835. The kVA report naturally shows a much better payback due to the
reduction in the kVA demand, based on $10.00 per kVA demand per month.
COST AND SAVINGS REPORT (KW Billing only - no kVA charges)

ES-## DESCRIPTION MOTOR PF ANNUAL LINE VOLTAGE P.F.
KW READ COST SAVINGS SAVINGS CHARGE
ES-01 Packaging, 4 x 40 hp 86.4 0.80 $28,787.20 $1,520.16
ES-02 Office, 1 x 200 hp 99.6 0.80 $33,216.00 $1,458.00
ES-03 Crushing, 6 x 20 hp 59.8 0.80 $19,929.60 $890.98
ES-04 Crushing, 4 x 200 hp 405.2 0.80 $135,078.40 $4,989.60
ES-05 Crushing, 4 x 75 hp 152.8 0.80 $50,931.20 $1,848.83
ES-06 Grinding, 8 x 200 hp 810.5 0.80 $270,156.80 $12,830.40
ES-07 Grinding, 6 x 75 hp 225.9 0.80 $75,289.60 $5,706.86
ES-08 Grinding, 10 x 100 hp 504.9 0.80 $168,294.40 $7,767.90
ES-09 Quarry, 4 x 25 hp 49.8 0.80 $16,608.00 $911.25
ES-10 Quarry, 2 x 50 hp 49.8 0.80 $16,608.00 $911.25
TOTAL 2444.7 $814,899.20 $38,835.22 $2,000.00 $0.00
TOTAL SAVINGS = LINE + VOLTAGE +POWER FACTOR = $40,835.22

COST AND SAVINGS REPORT (KVA Billing and kVA charges)
ES-## DESCRIPTION MOTOR PF ANNUAL LINE VOLTAGE P.F.

KW READ COST SAVINGS SAVINGS CHARGE
ES-01 Packaging, 4 x 40 hp 86.4 0.80 $28,787.20 $1,520.16
ES-02 Office, 1 x 200 hp 99.6 0.80 $33,216.00 $1,458.00
ES-03 Crushing, 6 x 20 hp 59.8 0.80 $19,929.60 $890.98
ES-04 Crushing, 4 x 200 hp 405.2 0.80 $135,078.40 $4,989.60
ES-05 Crushing, 4 x 75 hp 152.8 0.80 $50,931.20 $1,848.83
ES-06 Grinding, 8 x 200 hp 810.5 0.80 $270,156.80 $12,830.40
ES-07 Grinding, 6 x 75 hp 225.9 0.80 $75,289.60 $5,706.86
ES-08 Grinding, 10 x 100 hp 504.9 0.80 $168,294.40 $7,767.90
ES-09 Quarry, 4 x 25 hp 49.8 0.80 $16,608.00 $911.25
ES-10 Quarry, 2 x 50 hp 49.8 0.80 $16,608.00 $911.25
TOTAL 2444.7 $814,899.20 $38,835.22 $2,000.00 $72,543.74
TOTAL SAVINGS = LINE + VOLTAGE +POWER FACTOR = $113,378.96
In summary, energy and demand savings would be $113,378 at a cost of $66,000. The return on
investment would be 0.6 years. Energy and demand savings with no power factor kVA penalty
would be $40,835 at a cost of $66,000. The return on investment would be 1.65 years.
The cost of demand and kVA and energy rated vary widely depending on contract agreements.
This would change the economic situation accordingly. Some utilities and federal programs are
now offering incentive programs. Bonneville power has up to 16 cents per kWH incentive for
energy saved. This would improve payback significantly.
II. Harmonics
A. Basic Issues
Major problems resulting from harmonic currents, and low power factor:
1. Over heated: neutral wires, breaker panels, transformers, motors and other equipment.
2. False tripping of Circuit Breakers.
3. Failure of capacitors on DC busses in Variable Speed Drives.
4. Harmonics from local power source.
5. Failure of capacitors used in power factor correction.
6. Component failure caused by resonance with passive filters.
7. Malfunctioning computers and controllers, frozen screens, and lost data.
8. Fires caused by overheated neutral wires.
9. Power grids to deliver up to 50% more kVA than is being paid for, when billing is based
on KW.
These problems are caused by the small conduction angle used by rectifiers in each piece of
equipment, which convert AC to DC. Direct Current (DC) is needed to operate all semiconductor
circuits.
The devices used to convert AC to DC are: switching power supplies, diodes, silicon controlled
rectifiers, insulated gate bipolar transistors, gate turnoff thyristors. These devices are used in all
of the following equipment:
1. Lighting ballasts
2. Variable Speed Drive
3. Uninterruptible Power Supplies (UPS)
4. Computers
5. Instruments
6. Amplifiers
7. Programmable controllers
8. Electroplating
Main Times One

Panel Active
Motor Harmonic
Control Filter
Delta Center
A B
Variable
Times One Speed Motor
Active Drive
Harmonic
Wye Filter
A B
Transformer
/ Panel
A B
Loads Lighting Uninterruptable
Power Supply
Linear Non-Linear Computer
SOLUTIONS:
An engineering study is needed to determine the best choice of harmonic filter, by performance
and cost.
For low power loads, new designs of computer power supplies use power factor corrected
rectifier circuits that provide near unity power factor and low harmonic distortion, with over 95%
efficiency. These are becoming available up to 1kVA commercially. In 1997, only 40% of existing
computer manufacturers choose to spend the extra $25 for this improvement to their power
supply. In 2002 this utilization is civil law in Europe, ruling IEC 1000-2 forces European products
to comply with low distortion and unity power factor. Refer to www.conformity.com Jan 2001
feature article.
A. Passive Filters
On variable speed drives, line reactors and DC link chokes can cut current distortion “THD” from
80% down to 35%. Passive filters may cost 10% less than an Active Filter, but they are tedious to
design, and one must consider upstream transformer impedances to achieve low distortion.
Tuned and detuned passive filters are sensitive to load changes. They must be redesigned or
used in multiple banks switched in and out as loads change. With this method it is possible to get
current distortion down to 8% if load conditions don’t change. Cement plants usually have varying
load conditions on motors.
Computers have been used to control passive filters by quickly switching banks of tuned
capacitors in and out of the circuit to prevent system resonance. Resonance occurs because of
the interaction of the power lines, transformers, and other capacitors with the tuned passive filter.
See Reference #2 for this characteristic. Expensive engineering studies can be reduced in cost
by designing active filters into a solution. It is now possible to modify certain types of variable
speed drives to make shunt connected active filters that are similar in cost to passive filters.
Active filters automatically adjust to changing load conditions without risk of resonance.
Resonance causes up to 50% overvoltage in an electrical system.
Phase shift transformers can reduce harmonic distortion to under 15%. They must be rated for
the full load current and harmonic spectrum; load currents on each phase must be within 15% of
each other to get maximum harmonic attenuation of 70%. Unbalanced loads will limit attenuation
to a minimum of 40%. Varying loads, and system changes require breaker panel rewiring to
balance load current on each phase to improve phase shift transformer filter harmonic
attenuation.
Shunt connected harmonic filters are 30 % to 65% smaller because they only carry the percent of
load current equal to the percentage of harmonic current or power factor, typically 35% to 70%.
Comparatively, phase shift transformers carry 100% of the load.
B. Active Filters
Active Harmonic Filters are used to reduce the electrical harmonic current flowing in a power
circuit neutral and hot lines. The power factor is simultaneously improved to over .90 as the
harmonics are reduced. At both no load and full load, with the correct system design all
harmonics are reduced to below 5% current and 3% voltage distortion (THD).
Harmonics and poor power factor causes energy losses in electrical systems.
Transformers, cables and contactors will dissipate heat energy that can be measured in real
power. This heat increases energy and demand charges and reactive power costs from energy
providers. The range of winding losses is from 1% to 7% in transformers, and ½ to 2% in cables.
Load flow studies by engineering companies can exactly determine these losses when a
complete understanding of harmonic frequencies is utilized in the calculations. Magnetic
engineering is the field of study that covers these issues.
Main Times One

Panel Active
Motor Harmonic
Control Filter
Center
A B
Variable
Speed
Delta Drive Motor
Active Filters are really shunt conditioners. They only carry the harmonic current and have 98%
efficiency for the power they condition. They do this by generating and injecting a waveform 180
degrees reverse to the distortion signal from the rectifiers in the load. Signal processing and
correction signals are generated at the rate of 10KHZ to 15KHZ depending on the spectrum being
cancelled. The inverse current waveform is injected into the power line at the point of current
waveform sensing. The result is that the current is conditioned to become sinusoidal, and is re-
injected into the power line, in-phase with the voltage, thereby reducing distortion current and
voltage, and saving wasted energy in transformers and cables.
The voltage and current resulting in-phase with each other and being sinusoidal, now becomes a
.9 power factor. Current THD and voltage THD are less than 5% at all load levels from 1% to
100%. Voltage distortion is under 3%. These performance measurements are specified in Table
4 of the IEEE 519 – 1992. A clean Harmonic spectrum is maintained even with changing loads.
The filter tracks the spectrum change and maintains low distortion on the power line upstream
and downstream for 100’s of feet from the Active Filter. Software is being developed for optimum
sizing and location of active filters.
The Active Harmonic Filter provides Power Quality improvement in many ways:
1. Reduces energy, demand and reactive power costs from Electrical Power Bills
2. Increases electrical system capacity of existing electrical power systems
3. Improves Manufacturing Output by reducing lost work in process
4. Improves Product Quality through increased precision of control systems and data
5. Improves Productivity through decreased troubleshooting and maintenance
6. Reduces ground to neutral voltage and current and noise
C. Return on Investment:
Investment return can vary from 12 months to 48 months, depending on energy cost, reactive
power cost, demand costs length and gauge of power lines, hours of operation and transformer
ratings. Electrical Utilities are beginning to offer incentives in USA, Canada, Western Europe, to
improve power factor, peak demand, and harmonics.
III. Lighting Technology
A. Wide Spectrum Fluorescent Lights
New light efficiency technology has been proven to reduce energy costs by 45%. Tests of
increased light level at an air base in Hawaii were noticeable without measurements.
Fluorescent lights for outdoor use have become available, which radiate light in the 440
nanometer frequency range. The human eye is twice as sensitive at this wavelength at which the
rods operate. The cones in the eye receive 550 nanometers and are half the sensitivity of the
rods. This results in a situation where the light level can be 45% lower at 440 nanometers, thus
reducing energy cost by 45%. Light meters are calibrated at 550 nanometers. A new measuring
method must be established to allow measurements at 440 nanometers.
Products are available for indoor and outdoor use. High Density Discharge (HID) lamps should
be replaced with this type of light. HID lights take 15 to 20 minutes to ignite the gas. If the AC
power goes off, it takes 10 to 15 minutes to cool the HID lights before the igniting delay. HID
pulse start ballasts speed up the ignition time to 5 to 7 minutes. The new 4,100°F fluorescent
light ignites in less than 2 seconds; it appears to be “instant on”. This is desirable for unoccupied
rooms, especially warehouses or parking lots to save energy when no one is present. HID lamps
of 450 watts have been replaced by 208 watt of these new "wide spectrum lights.” Refer to the
“Fluorescent PowerPoint Presentation” in the index. Performance references are available upon
request.
B. Lighting Controllers
Most facilities, when retrofitted with metal halide or HPS lighting, are assumed to have some of
the highest “efficiency” lighting available, and it is often assumed nothing else can be done.
However, in many instances one ends up with excess lighting, and special electronic controllers
can be installed on lighting branch circuits to control the slight dimming in steady state of these
metal arc systems. The resulting slight voltage drop is more than compensated by a substantial
KW reduction, out of proportion to the percentage light reduction. For example, reducing the
lumens output of metal halide fixtures by 10% could lead to a power decrease in excess of 18%.
Although these devices require engineering to install, one does not have to change out any
fixtures. Additionally, such new equipment could even apply to a new construction project after
completion if it is determined that further refinement in energy consumption is warranted.
Because the lighting controllers do not merely reduce voltage to a lighting circuit by reducing the
voltage waveform, but do so by a more technically sophisticated process of waveform
“reshaping,” the total voltage (RMS voltage) is reduced substantially more than peak voltage. This
means that there is not a one-to-one relationship between energy cost savings and light output. In
fact, depending upon the type, age and condition of the treated fixtures, light output will be
reduced by only one fifth to one third of the level of savings achieved.
Period Of Increased Light Output
Mean Light
Output
Average Life=20,000 Hours=4 Years Extended Life After Treatment
Lights are dimmer today than they were yesterday. And they will be dimmer tomorrow than they
are today. In fact, as lights approach the halfway point in their expected lamp life, they will only
produce about 60% of the light they did when they were brand new.
Lighting controllers set to achieve up to 30% cost savings will initially dim lights about 5% to 10%
from their previous output. However, over time, the average light output from your lights will
actually be greater (5% to 20% greater over the final 50% of the fixture’s actual service life).
With new or nearly new fixtures, using modern lighting controllers set to a modest 30% savings
level, the life of the bulbs are expected to be extended by 40% to 80%. This means that a bulb
with a typical lamp life of 20,000 hours should last from 28,000 to 36,000 hours.
The trend for light output reduction would be extended forward by this same amount of time, as
well. This means that you will have the light output you would normally lose at 10,000 hours still
being produced at 14,000 to 18,000 hours. The lamps will lose output at a much slower rate than
if they were treated by non-electrical wave shaping lighting controllers.
CONCLUSION
Incentives being introduced by federal, state and county programs are improving the return on
investment to get paybacks in the 1 to 5 year range. These programs cover upgrading motors,
variable speed drives, lighting, heating, ventilation, Air conditioning, utilization controllers.
Rebates for upgrades are available for energy and peak demand reduction, as well as renewable
energy.
Productivity, quality, and bottom line net profits can all be improved if a proactive plan is
established and implemented. Regional DOE offices of industrial technology will be introduced to
this information and a wide range of federally accredited education courses will be available
regionally in the coming year. Please contact me through IEEE listings to reach local support
programs in your region. Other pertinent documents are available upon request.
References
1. W.K. Chang, M.W. Grady, M.J. Samotyj “Meeting IEEE-519 Harmonic Voltage and Voltage
Distortion Constraints with an Active Power Line Conditioner” IEEE 94 WM 093 PWRD
2. C.E. McCoy “Static Power Converters, Harmonics, The Power System and You” IEEE-PCIC-
92-44
3. T.M. Gruzs “Uncertainties in Compliance with Harmonic Current Distortion Limits in Electric
Power Systems” IEEE Transactions on Industry Applications Vol. 27 No. 4 July/ August 1991
4. R.L. Almonte, A.W. Ashley “Harmonics at the Utility Industrial Interface: A real World Example”
IEEE Transactions on Industry Applications Vol. 31 No. 6 November/ December 1995
5. S.N. Ghosh, ed. “Energy Conservation and Environmental Control in Cement Industry”
Akademia Books International, 1996.

IEEE Cement 2002 PDF

Uploaded by

Copyright:

Available Formats

IEEE Cement 2002 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

IEEE Cement 2002 PDF

Uploaded by

Copyright:

Available Formats

What is the purpose of the conference?

What is the purpose of the conference?

What are some of the working groups discussed in the document?

What are some of the working groups discussed in the document?

IEEE-IAS/PCA

2002 CEMENT INDUSTRY

May 5-9, 2002

Copyright ©2002 IEEE. Personal use of this material is permitted. However,

Earl Bouse, Jr., Chair Cary O. Cohrs, Chair

Chair Vice Chair

Secretary IEEE Relations/Nominations

Bylaws and History Committee Members

Working Group Chairs

Automation Paper Review

Drives and Related Products Conference Site Selection

Environmental, Maintenance and Safety Paper Selection

General Practices East Coast Subcommittee Chair

Power Generation, Distribution Gulf Coast Subcommittee Chair

Awards and Recognition West Coast Subcommittee Chair

Louis J. DeHayes, Vice Chairman Hans E. Steuch, Vice Chairman

Stuart Pryor, Vice Chairman Mark Morgan, Vice Chairman

Task Committee 3 Task Committee 4

Max L. Frailey, Vice Chairman Jeffrey P. Kemmerer, Vice Chairman

Chairperson Plant Tour

Secretary Special Projects

Renee Williams Chris Horner

Hospitality Spouse Activities

PCA Liaison Advisor

Power Generation, Distribution and Related Products

“Cogeneration Systems in Material Drying Applications”

“Higher Efficiency – What is the Real Cost?”

“Energy Efficiency Improvement in the Cement Industry”

Drives and Related Products

“Spotting Drives for Large Motors”

“Selecting the Right Preheater Fan and Drive”

“Alternative Drive Solutions for the Cement Industry”

“Solving Raw Material Challenges”

“The Kosmosdale Expansion Project”

“Environmental Considerations for Preheater Design”

“Cement Mill System Upgrade”

“Increasing Cement Grinding Capacity with Vertical Roller Mill Technology”

“Existing Mill Capacity Doubled Developments in Grinding Process”

Environmental, Maintenance and Safety

“Staged” Combustion for Low-NOx Calciners

“A Real World Innovative Concept for Plant Information Integration”

“Total Integrated Device Operation Network (TIDON)”

Leroy C. Cockrell, Jr. Andrew Robinson

Craig Gotro Jay Shah

Andrzej Janik Gregory Witte

For Presentation at the

RECOGNITION OF THE PROBLEM:

Initial Scope and Plan:

We also recognized that the 6 phases of most typical projects were:

TIME FOR A CHANGE:

Revised Scope and Plan:

TEMPLATES AND PHILOSOPHIES:

SIMULATION AND I/O CHECKOUT:

Documentation of the electrical systems is important for construction, but it is also

Urs Herzog, Holcim Group Support Ltd. Switzerland

Warehouse Personnel Client