35

,:::
............
./:!
:.
J
,.,,
...
6
,:.
.:~,
/ ,$
</, ,,
MILL TESTING,
OPTIMISATION AND
TARGETING
D BAIRD
MILL TESTING, OPTTMISATION AND TARGETING
OVERVIEW
This paper is divided into three sections which relate to the TESTING,
OPTIMISATION and performance TARGETING of cement mills.
SECTION A : BALL MILL TESTING
The various methods used for testing ball mills are described, especially mill axial
samplingtesting. Whilst the paper concentrates primarily on cement mills, these
test methods can be equally applied to raw mills. Hence some of the examples
shown include raw mills.
SECTION B: CEMENT MILL OPTIMISATION
This section provides general guidelines on how to optimise the internal
cordigurationof cement millsto suit the typeshnge of cements produced, These
notes includerecent feedback on several different types of mill internal liners and
diaphragms.
SECTION C : CEMENT MILL PERFORMANCE TARGETING
This section examines the various methods by which the target petiormance for
open/closed circuit cement mills can be established using a knowledge of the
following:-
G
Cement grindability data
G
The results from mill inspections and axial sampling tests.
The results of the BCI cement millbenchmarkingexercise are also reviewed in this
section.
SECTION A
BALL MILL TESTING
SECTION A
1.
2.
3.
4.
5.
6.
BALL MILL TESTING
CONTENTS
INTRODUCTION
MONITORING MILL PERFORMANCE
2.1 Mill Throughput Tests
2.2 Power Drawn
2.3 MU Product QudIty
2.4 Feed Grindability
2.5 Temperature
2.6 Air Flow
AXIAL SAMPLING TESTS
3.1 Full Test Procedure for an Aid Test
3.2 The Quick Test Method
AXXAL TEST FOR A CEMENT MILL
PRACTICAL EXAMPLE
4.1 Mill Performance Before and After Tuning
4.2 Background to Testwork
4.3 Mill Inspection
4.4 Axial Sampling Test Results
4.5 Gas Circuit Tests
4.6 Power Drawn
4.7 Medium/Long Term Optimisation
MILL INSPECTION AND MAINTENANCE
5.1 Lining Plates
5.2 Diaphragms
5.3 Meda
5.4 Voidage Filling
5.5 The Importance of Regular Mill Maintenance
and the use of Axial Tests
FINE MEDIA IN CEMENT MILLS
APPENDIX IA
APPENDIX HA
APPENDIX IIIA
APPENDIX IVA
APPENDIX VA
APPENDIX VIA
APPENDIX VHA
Cement Mill Faults
The Fineness of Samples taken immediately prior to the First Chamber
Diaphragm
Essential Preconditions for use of Finer Media Gradings in Chamber
Two of Cement Mills
tial Sampling Curves for Cement Mill - Second Chambers - Sieve
Residues
Example of How the Volume Loading within a Mill can be tiected by
the Accumulation of Unground Material (Nbs)
Mass Balance on a Raw Milling Circuit
Example of the Quick Test method for PucialSampling Tests
1. INTRODUCTION
The efficiencyof grinding depends upon a number of factors, and a variation of one or more of.
these causes deterioration of mill performance. If this goes unchecked very inefficient grinding
occurs resulting in a very poor quality mill product.
CarefiJ routine observation of mill residues and power used for the grinding process will show
when efficiency begins to fall off and whether a thorough check on performance is necessary.
2. MONITORING MILL PERFORMANCE
In order to monitor a mills perfimrmnce, the following data is required:-
G
Mill Throughput
G
Power Drawn
G
Mill Product Quality
G
Feed Grindability
G
Mill Temperature/Product Temperature
G
Mill Air Flow/Cooling
Much of the above information should normally be recorded as part of the Works routine
procedures. When routiie data is unobtainableor is suspect then the following tests and checks
may be carried out.
2.1 Mill Throu~hDut Tests
2.1.1 Weigh Feedem
All too ofien weigh fkeders can give a misleadmgpicture of a mills throughput. Direct readings
of the mill throughput from a weigh feeder or totaliser are subject to possible errors in the
calibration of the feeder. Regular checks on the calibration of feeders in accordance with
manufacturers recommended procedures can reduce the degree of error.
A simple method of checking the accuracy of a weigh feeder is by measuring the weight of
material over a known length of belt under steady feed conditions, knowing the belt speed
enablesthe throughput to be estimated. Sufficient length of belt must be sampled and account
taken of any cyclic variation in feed rate if this method is to be accurate.
/
I I I I , 1 I
\
I I I I I I 1 I I
u
L
!
_
-
Chain tests can be used to check the accuracy of weigh feeders. After zeroing the scale, the
weigher belt is loaded up with a set of chains which are calibrated to cover the weighing range
of the scales. The scale indicator and recorder can be checked from a knowledge of the chain
loading and belt speed.
2.1.2 Salt Tests
Another method for determiningthroughput is the salt t~ where salt is used as a tracer through
the mill. Under steady conditionsa constant amount of salt is added to the mill feed and the rise
in chloride level of the millproduct is obsemed. Atypical procedure for a Salt Test on a cement
mill would be as follows:-
G
G
G
G
G
G
Sample the mill feed and finished cement for approximately 30 minutes before starting
the test in order to establish control conditions.
When the mill is running steadily add accurately measured equal quantities of salt at
regular intervals for a period of two hours.
The addition rate of rate should be
approximately 0.5/0of the mill output per minute.
Sample the mill product at regular intervals for up to 3 hours.
Anaiysethe samplesfor chlorideusing the chromate direct titration method. Determine
the purity of the salt added and determinethe chloride content of the clinker and gypsum
feed.
Plot a-graph of% Cl against time and note the steady average value to which the
chloride level rises. (M) as shown in F@ure 1.
Determine the output by the following mass balance where:-
X
=
Chloride entering mill in kghnin
Y
=
Cement Mill Output
z= 0/0Chloride in the cement prior to salt addition
M= YoChloride in the cement after salt addition ,=
From which y maybe calculated:
~ _ x. (loo - M.(NaCl/Cl)
M- 100.Z
This test method is more complicated than belt sectioning/weigh offs and is also prone to
sampling and analytical errors. Hence this test is not commo@y used.
,
2.1.3 Cement Weigh Off
In this test the cement mills product is diverted into an empty clean silo where it can be
separately packed off and weighed. For valid results, the test must be run for sufficiently long
time, i.e. at least 24 hours. Errors will arise if the silo used cannot be effectively emptied out
before and afier the test due to build up.
2.1.4 Clinker Drop Tests and Volume Measurements
In cases where space allows for the collection of feed belt material, a drop test maybe carried
out by diverting the material through some form of by-pass into a pre-weighed dumper. By
collecting the fed material over a known period the mill throughput can be estimated. ,
Another method which is not particularly accurate but which can be used to give a rough guide
to mill output, is the method of measuring the fall in level of clinker in a feed hopper, whilst the
mill is running with a steady feed.
Samples are taken during the test to determine the clinker bulk density and the S03 level in the
clinker and fished cement. An SOSmass balance then enables the gypsum addition rate to be
calculated whilst the clinker throughput is estimated from the bulk density and fdl in volume in
the hopper. Errors arise in this method from level measurements and differences in the degree
of compaction and segregation effects which may alter the bulk density of clinker in the, hopper
from that measured on the feed belt.
2.2 Power Drawn
The most usefid method of checking the power drawn by a mill is by taking routine readings
from an integrating kWh meter. Such readings are vexy often taken on a weekly basis. Spot
checks can be made by timing a number of revolutions of the disc of the kWh meter and applying
the appropriate correction factor for the meter.
If neither of these tests can be carried out, then an estimate of the power drawn can be made
form ammeter readings. A knowledge of the voltage and power factor enables the power drawn
to be deduced though such estiniates are often subject to large errors. From records of the mill
throughput and the power drawn, the power consumption in kWh/tonne is calculated.
Gross Power =43 V.I Cos#
where V = Voltage, I = Current, Cos ~= power factor
4
Gross power is the power to the mill motor, Motor and gearbox losses are normally between
5 and 10%dependingupon the drive systemused. Hence the nett power absorbed by the charge
in the mill is typically 90 to 95/0of the gross power absorbed. See notes in section 3.1.
2.3 Mill Product (Mudity
When referringto a millsoutput, reference should also be made to those quality aspects which
can affect the output. It is normal to check a cement mills product for suxfacearea and sieve
residues at 90 and 45 microns. A record of S03 content is important as the form of sulphate
addition, whether it be gypsum or anhydnte can have a significant effect on mill outputs by
altering the grindability of the feed. Factors which influence the grindability characteristics of
clinker are summarised below:-
2.3.1 Clinker Grindability - Effect of Clinker Chemistry Variations
The following factors have been shown to have a detrimental effect on clinker grindability and
hence cement mill output.
G
A lower LSF clinker ie a higher proportion of CJ in place of CJ.
G
Lower free lime
G
High liquid phase
G
Lower gypsum addition
G
Higher clinker SOJ
With any source of clinker supply there will be natural variations in the clinker chemistry and
hence grindability. If a mill automatic control circuit works correctly then it should enable
..additional output to be achieved during periods when grinding lower grindability clinker.
Figure 2 shows the relationshipbetween clinker chemistry and mill only power consumption for
two closed circuit cement mills grinding OPC and SRC clinker when grinding to a constant
surface area of 306m2/kg. This illustrates the effect of LSF, C3S and C2S on clinker
grindability.
Hence, variations in the quality of clinker produced can have a significant effect on the
petiorrnance of any cement milliig circuit. This aspect needs to be considered when evaluating
the performance of any mill circuit.
2.4 Feed Grindabilitv
Changes in the grindability of the clinker can tiect mill performance and so it is advisable to
cany out grindability tests on the clinker at regular intervals. When carrying out axial tests on
a mill, as well be described in greater detail later on in this paper, it is recommended that
approximately 50kg of average clinker sample is taken for a grindability test to be carried out.
5
56
Cn
54
d?
52
50
22
18
94
93
- Relationship between Specfic Power Consumption and Cement
Chemistrv for two Closed Circuit Cement Mills
c@
No.2 Mill (S.R.C)
-... No.1 Mill (OPC)
-..
.*.-
......<
%.-
%.
G..*
-..*
%..
-..
*-----
-...
%.-
%.-
.
-.#---
.*.*-
-*.--
*-..
.4
...----
.*-*-
.**.. .-
-- No 1
..-.--
.
...---a-
-----------
-- . . . .
%*.-
%..m
------ No.1 Mill (OPC)
..----
-=----
--....
.*.-
{
G-...*
--..*-
%...-
LSF
No.2 Mill (S.R.C)
,
90
32. 5
33
33. 5
The result of this test enables the actual mill performance to be compared with the theoretical
performance and is usefid in showing how efficiently the mill and individual chambers are
performing highlightingareas of the mill where the performance can be improved by alterations
to the mill charge etc.
Gypsum addition has a significant impact upon the grindability characteristics of cement.
Additivessuch as limestonealso afkcts the cement grindability characteristics and the following
rules of thumb can be applied.
+ 10?0
+ 1??0
2.5
Limestone addition reduces the B.C.I. grindability figure by approximately 2.5%
Gypsumadditioncan equate to+ 5% mill output u + 12m2Agsurface area at the same
output.
Temperature
Problems arise when hot clinker is fd to a millor the mills cooling system, i.e. induced draught
or water injection do not fimction properly. Thermocouples can be used to monitor both feed
and product temperatures and the latter can be used to control a water injection control loop.
2.6 Air Flow
For adequate ventilatio~ the quantity of air through an open circuit mill should be 2-3 volume
changes of air per minute. Here the term volume refers to the free volume above the charge in
the mill and estimates are made using a standard temperature of 11OdegC.
Whh more recent and larger closed circuit mill installations sufficient ventilation for up to 7 air
changes per minute is oilen provided. However, at this high level of airflow, there is a risk that
the cement cmied out of the mill to the filter plant is much coarser than the finished product.
Thus mill airflows are usually reduced to a level below 7 air changes per minute.
There are a number of difficulties involved in making measurements of air flow through a mill.
Measurements taken around the ducting leading to the dust filtering plant can be meaningless
if the mill has poor seals as the resultant air flow figures are more likely to indicate inleak rather
than ventilation air flow.
Measurements recently taken at one UK Works indicate that whilst the air flow through the
cement mill filtering plants was adequate only 200/0of this air flow was actually being drawn
through the mill owing to poor mill outlet seals. Pitot measurements in this region suffer from
problems of blocking Pitot tubes due to dust and humidity. To measure the quantity of air
actuallyflowingthrough the mill, an anemometer can be used at the mill inlet with the feed rate
off the mill. Draught irdcators maybe provided at the mill inlet to give some rough idea of the
quantity of cooling air through the mill. The differential pressure across a cement mill being
typically40-60 mmwg (for open circuit mill with moderate cooling airflow) and 75-100 mmwg
(large closed circuit mills with high mill ventilation).
7
.
:,
.,
,,
Samplingp(obe
Vent pipe
\
$
Hood
l!iR@-
13usttrap Drying
Ein
Mill
Inlet pipe
&Thermometer
~ I?otometer
: !$Zodetector -lJ-
1
N20bottle
~IGURE 3 NzO Tracer Method
Another method of assessing the air flow through the mill is to use nitrous oxide into the mill
inlet as a tracer. The concentration of N20 in the exit air is measured using an infrared detector.
FQure 3 shows the arrangement for N20 tracer testing on a mill. A similar technique has been
applied, in a limited number of cases, using Carbon Monoxide as the tracer gas.
If either the routine performance data or any of the above tests show that there has been a
deterioration in the perfiorrnanceof a particular mill, then it is advisable to carry out a more
detailed examination into the internal state of the mill as well as an axial test.
3. AXIAL SAMPLING TESTS
An axial sample testis means of determining how well a mill is grinding along its length. Such
a test can highlightarm withinthe millwhere the grinding is not being carried out as efficiently
as it should be. When coupled with the results for the grindability test it is then possible to
compare the overall performance of the mill as well as individual mill chambers with the
theoretical performance predicted.
To ensure that correct conclusions are drawn from any axial sampling test, it is essential that
representative samples are taken throughout the mill othenvise results can ofien be misleading.
The full test procedure for an axial sampling testis as follows.
3.1 Full Test Procedure for an Axial Test --
G
Sample the mill fd and product under steady conditions for approximately 40 minutes
prior to stopping the mill. Record the mill output and power consumption
In the case of a closed circuit milling system. take samples of the mill and separator
circuit before crash stopping in order to construct a mass balance and to estimate the
separator efficiencies. The mass balance can be used to estimate the mill recirculating
load and allowany rejects weighing device to be cross-checked. Mass balances should
be based on 45 micron residues for cement mill circuits and 90 micron residues for raw
mill circuits. Appendix VIA shows a typical mass balance/separator efficiency
calculation for a raw mill circuit.
G
Stop both the mill and the f~ simultaneously. Ifthe fd is stopped before the mill then
the residual material within the mill will be ground finer than normal and this will make
the overall mill efficiency appear higher than it actually is.
G
After allowing sufficient time for cooling, enter the mill and take axial samples. Divide
up the mill internally into sampling points typically 0.5-1 m apart. Samples should also
be taken at the diaphragms. At each point on the axis, an average sample should be
taken of the material along a line at right angles to the axis of the mill. The material
shouldbe taken from points a fewinches below the ball charge and not from the surface.
In the case of a three chambered mill, take larger samples in the first and second
chambers than in the third chamber.
9
CLINKER GRADE
mm
SURFACE AREA
B.S. SIEVE
mz I kg
.$
+19
0 2
+
y
8
- 19 +95
0 3
~3
~.
-
-9 5
+4.8
0.6
0 + 16
38
+7 -4 8
+2-4
1 1
m
-7 + 14 -2.4
+1 2
183
-i4
+ 25 -1 2
+0 6
3s46
-25
+ 52 -0-6
+0 3
6.42
-52
+100 -03
+0 .15 17 8
-1oo
-(j .15 Measured directly
(Lea Nurse)
PIGW 4 surface Area of Different ClinkerGrades
Ensure that allsamples taken are filly representative ofmatetial tittin the mill. Forexample
if nibs of unground material are present, sample these and do not simple discard them. Take
larger sample quantities if large quantities of nibs are present.
G
Allowthe samplesto cool before measuringthe surfice area in the case of a cement mill,
ardor sieveresidues in the case of a raw mill. The coarser samples whose sutiace area
cannot be measured directly must be graded and their surface areas calculated from
Figure 4. For a cement mill, a check should also be made on sieve residue of samples
throughout the mill: of particular importance are the residues at 45 and 300P m, and at
2.36mm sieve sizes. (see Appendix IVA)
G
Measure the height above the charge and calculate the VOvolume loading from Figure
5. From this the weight of me&a in each chamber can be calculated using a value for the
average media bulk density, if none is available the use a media density of
approximately:-
4.3 t/m3for Chamber 1
4.5 t/m3for Chamber 2
Ifa third chambered mill is tested the media density figures are 4.3,4.4 and 4.5t./m3for
chamber 1,2and 3 respectively. Figure 5. is an approximation to the formula shown in
Figure 6. Normally a computer spreadsheet approach should be used to provide
consistent calculation method with the accurate date for each mill.
The height above the charge is best measured with the filling slightly run down,
otherwise a false highvalue for volume load will be obtained. In addition large quantities
of unground nibs can cause the charge to bulk hence giving a false impression of the
actual volume loading in each chamber (see example in Appendix VA)
G
Using the power equation, calculate the power absorbed by each mill chamber and for
..
the mill overall. Compare thk with the figures obtained from the kwh meters.
Nett
where
Note:-
kW = 0.2846 D. A.W.N.
D = Mill diameter inside the lining in metres
A = 1.073-J where J is the fractional volume loading
W = Weight of media in tonnes
N = Mill speed in rev/rein
nett kW = Gross kW x n where n = 0.95-0.9 depending on losses in
motor/gearbox etc.
11
0-s0 0-55
FIGURE 5 Ball
Mill -
0-60
Volume
O-65
Ratio
Loading
0-70 0-75 0-80 0-85
FIGURE 6
VOLUME LOAD FORMULA
H = Height above charge (m)
D = Mill internal diameter inside lining (m)
% Volume Load = ~
[ 1
(0.25.Cos(2m(WD - 0.5))) - ((HID - 0.5). (S2)))
,,
The factor of net/gross power is very important since it can give an indication of what is
happening inside a ball mill. Some typical factors found with different mills areas follows:
Net/Gross Factor Mill Tme
0.94-0.95 Modem efficientmill with central drive, efficient gearbox
and drive motor. Good material flow through mill
without hold up and charge expansion.
0.93-0.90 Ghth gear driven mill of older design with higher drive
system losses
0.88-0.89 Check to see whether mill could be running with
minimum powder hold up causing charge to pull higher
kW.
Generally, if the net/gross ratio exceeds 0.95-0.96, then it is likely that the volume
loading estimate is affectedby charge expansion due to hold up of material or excessive
quantities of nibs etc. This condition is very common in the following cases.
Cement mills using Fine (25-15rnm) media in chamber 2.
First chambers of cementkaw mills with coarsehrd feed materials, worn ball
charge and liners giving poor crushing action.
Inadequate ventilatiotilocked diaphragms causing excessive material hold up
within the mill.
G
Plot axial graphs as follows:-
(a) Cement Mill
Surface area versus position along mill axis
Residue versus position along mill axis
(b)
Raw Mill
.
Residue versus position along mill axis
Ifthe volume loadiigs vaxysignikantly then it can be usefid to plot cement surface area
versus the cumulative power drawn along the axis of the mill. However, current BCTC
guidelines put more emphasis on the shape of the cement.lmeal residue curve changes
along the mill axis. AppendixDAgives details of the recommended sieve residues which
should be examined when inteqweting axial sampling test results.
The axial samplingcurves should show a steady rise in cement surface area and an even
reduction in cement (or raw meal) residue along the axis of the mill chambers.
Graph contains any flat sections or sections where the rate of surface production is low
as indicated by a shallow slope of the graph, then this indicates areas of the mill where
the efficiency is low due to:-
Incorrect ball size
Insufficient charge
Blocked diaphragms
The cement/raw meal residue versus position along mill axis curve is usefil for showing the
following:-
Whether or not the material leaving say, chamber 1, is sufficiently fine for fine
grinding in chambers 2/3. See the guideline figures for cementhaw mills given
in Appendix 11A.
Whether or not there is any accumulation of oversized materkd (ie nibs or
spitzers) in the chambers due to coarse feed size, inadequate first chamber
crushing, classification of nibs by classi&ing liners in chamber 2 etc.
Whether or not the secondhhkd chamber media is sufficiently fine to allow a
steady reduction in 45 micron residu-ein a cement mill secondhhird chamber.
G
Calculate the gross power drawn by each chamber of the mill and work out the gross
kWh/tonne figure for each chamber. Compare these results with the guidelines given in
Section B.
G
Measure the media grading in chamber by the following means:-
G
First Chamber
Remove media from the charge, record its size and number of balls in each size
range 90,80,70, 60mm etc. Try to measure as many balls as possible with a
minimum around 50. Take 2-3 measurements in the first chamber.
This will only give a rough guide to the media grading, but can help to decide if
the charge is say, too worn and in need of regrading.
G
Second Chamber
When taking the axial samples it is often convenient to take samples of the media
at the same time. When taking every other sample in a mill with a classi&ing
15
lining the sample and media can be dug out of the mill together. Load the
samples into, say, cement bags and remove these from the mill.
Sieve off the material, sample and record the total weight of the media. Count
the number of balls present and estimate the average ball weight. Plot a graph
of average ball weight against position in the chamber. If the classifying lining
is working correctly then obviously there should be a gradual reduction in
average ball weight along the length of the chamber.
G
Cany out a sii grading of the fd clinker. This is important when determining the size
of media to be added to the millslist chamber. Usually the maximum feed size dictates
the maximumball sii that shouldbe added to the first chamber of the mill, for example,
for 19mmcIinkerthe typical maximummedia size is approximately 90mm. If the clinker
proves particularly hard to grind then media up to 100mrn size may even be necessary
in the extreme case. Normally one varies the proportions of 90mm media in the charge
to suit different types of cement milling mndhions and different mill configurations.
Please refer to the media grading guidelines given in Section B.
3.2 The (hick Test Method
tial sampling tests are only of any use if the results are correctly interpreted and acted upon
with a short/medium/long term action plan formulated.
To have any impact, the results need to be reported as soon as possible after the tests have been
carried out.
To speed up this process, BCTC use a Quick Test method which fmtures
Taking the minimum number of essential samples.
Carries out the minimum analysis based primarily on the key sieve residues.
The Quick Test procedure is not rigidlyfixedand very ofkq the number of samples taken may
vary to suit milling condhions. for example, usually 3 samples would be taken at the inlet
middle and before the diaphragm in chamber 1.
If there is evidence of nibs accumulation prior to the intermediate diaphragm then it maybe
usefid to take an extra sample at O.5m prior to the diaphragm. This shows whether or not the
coarse material at the diaphragm is common throughout the chamber or is due to build up.
Another usefid type of test which can be carried out is the Quick Test applied to a mill in
which the mill output has been pushed prior to the crash stop. This method is particularly
useiid when testing raw millsto determine what the limiting factor on performance is. Care has
to be taken not to push the millf~ rate too Ibr if the circuit is unstable. However, some useiid
Mormation can be shown up by the test. A recent test on a raw mill showed that as the f=d rate
16
was increased, the material leaving chamber 1 coarsened up and nibs started to buildup at the
diaphragms. These nibs then started to blind the diaphragm slots and this resulted in coarse
material spilling into chamber 2 via the central ventilation grid.
Solution - Coarsen up chamber 1, charge by adding extra 90mrn media and
reducing 60/70mm media quantity.
,. Reduce size of central vent grid to permit higher volume load and
to cope with charge expansion.
4. AXIAL SAMPLING TEST FOR A CEMENT MILL
Practical ExamDle
The following example demonstrates how the results of axial sampling tests can be interpreted,
acted upon and improved performance obtained.
4.1 Mill Performance Before and After Tuning
Mill Absorbed Outr.)ut B.C. Conventional
Case Power (kW\ tQl mill efflciencv 0/0 M. S.P.F.
Before 2714 65.3 138 1.021
After 2830 70.7 144 1.061
The BC conventional mill efficiency and M. S.P.F. (mill surface production factor) are both
measures of the mill efficiency compared with the BC cement grindability test estimate. These
factors are described in section C.
4.2 Background to Testwork
The closed circuit mill is a USA mill equipped with a large conventional separator and large
elevator transport capacity. This design has been proven to give a good efficiency when used
in conjunction with grinding aid.(see sections B and C)
Representatives samples of clinker and gypsum were taken over a period before and after the
test. These were used for cement gnndability testing at BCTC. The test results showed an
increase in cement grindabilityfrom 125/0previouslyto the current level of 129/0. This increase
was due to higher C# levels in the clinker and alone would have accounted for 3Aloss of mill
output .
17
The plant urgently needed to maximise the cement milling capacity to cope with high seasonal
demand. This prevented stoppages for major mill changes and the aims of the testwork were as
follows:-
Maximise output by means of minimum changes to ball chargeiintemals without
major internal replacement.
Obtain output gain NOW!
Identi@ mediudlong term plan to optimise mill petiormance.
The mill was equipped with a 3356Kw motor but was limited to drawing 2714Kw by volume
loading constraints. There were problems with the design of the stepped chute feeder. This
regularlybackspilledclinkerlgypsumat the mill inlet and thus limited chamber 1 volume loadlng.
4.3 Mill Inspection
Details of the mill internal inspection report are shown in Table 1. The key points areas
follows:-
a)
b)
c)
d)
e)
9
d
The mill produces high surface area cements at 391m2/Kg (3800 cm2/gBlaine)
The mill critical speed is very high at 80.79?! compared with a typical value for
a modem mill of 76/0. The discharge side of the intermediate diaphragm
featured a915mm diameter outlet. Thk limited the volume loading to around
35%under static load conditions.
The mill was equipped with wave type liners in chamber 1 compared with our
normal recommended liners i.e. step liner or Lorain bar lift liner. However, it
would be unwise to change the liner to these types as the lifting action could
prove too severe due to the high critical speed:
The interm~late diaphragmwas badly blocked by small broken media with only
2.09V0effective slot area. Cleaning of the slots could increase this to 5.27V0
which is acceptable.
The first chamber charge contained worn undersized media which needed
regrading to avoid it breaking up and blocking the slots.
Material levels in chamber 1 showed the chamber to be running very light of
material.
The first chamber was charged to 37.3V0volume load but this included some
expansion of the charge which is allowed for by using a lower media density of
4.097 = t/m3 in chamber 1.

.
.-m _=-
MILL INSPECTION
, *TIONS- M~~l sampling teaL
13r7/94
$ b?&RE&:- E%AMPLE
MILL NO.:- Cemant MN No.2
NOTE5- Before charging,the mill drew 3640 HP or 2714 kw.
Approxtmate)y10-12 stons of I?Wd@was addad to Chambar 2 after the teat.
Mediiedded wee W4tol.5inchapprox
MILL MOTOR. - 33S6KW MILL SPEED (RPM)= 17.42 M53wrWt
MILL MOTOR:. 4530 HP MILL SPEED (RPM)= f7,44 Mlnakeed for CbfNMN
MEANINTERNAL DIAMETER= 3.622 m CRITICAL SPEED = 21.69 rpm
PERCENT CRITICAL=
0h6rnber 1
60.79 %
-==x-
EffSCtii LSWthm . $. Wmerl@a cftdal mill length.
30.39 %
Internal tiameter m
s Total chambarhsngthm= lt.090
Internal volume m3 = 3602
Liner type.- Wavelinerwith nomine16U-65 mmprofiiencwwrn tounder(%3mm.
culditikrndlinera- Liners inthefirat 7rmwareworn togrowadprdite.
WPhfwmu=mk 1OuM
-.
Diaphwn ty*.- Modem IWffergratesinatalled on~iiftertype~ support~.
Circumferential dote with self clearing prcfle which ara partiallyworn.
The outer segments surfaces were @ttad
The d@ragm has slottedgratea on its dwharge *.
Ixaphm e- Tha outer ssqments are badly bkxked by srnall/bmkan made end the slda ~
denibbad duringthe inapectlonHwew,tha methcd IX cleaning these alota
allowsthetrap machato drop into the charge and reblockthe d@ragm,
There is a signikant quantityof small rnada mthe first chamber Wctr
will tend to braak up and reblock the mtennadii diaphiagm.
Dk@wagm SW a&e8 and area:-
Slct lengths:- 12 eegmerb
WTER-Twtn ekXe mm 1!50 125 65 60 140 M6 156
115 1s0 165 fro 476 17!3 185
Sloteizesmm= 6 6
PeftwY &%w M
& Jdia.
6 0
Cond!tii of StOtS=
Perc entage blookage=
Maxtmumslot area without Mookage =
-i@Ot SfWt With_=
0.3377 m2
0.01s9 m2
165
1%
6
INNER.Twirl emsmm lm 115 ?30 140 150
E 115 105 190 160 I@
Single elota mm 145 150 It%
S&tsizesmm= 6 e 6
condition of elota = PwtkMyt&d Lw Lmedia
Percentage MOckage=
Mexlmum slot area without bbckage = 0.2570 m2
ERecuve slot area with bbckaga = 0.2165 m2
Total slot length m=
Diaphm@rneffectii cross sectional ar6a (w@ 11.262 WEIGHTEO MEAN SLOT SIZE= 6.(H) mm
Dmphragmelc4 areesaaparagnt c#CSA= 5.27 % If ail slots are clear.
2.CQ % Wkh aiota bbckage.
Maximum slot aim= 6mm Minimum slot eizw %Inm
Racammanded slot size= 6 mm
DISCHARGE OF INTERMEDIATE DIAPHRAGM IN CHAMBER 2:-
lNNER-Slot sizes mm = 6 6
waaIa L.
e 6
Conditionof slots=
OUTER-Slot SiZeS mm = 6 6
Conditionofslota= &e4Y46%bto&dbY bmkdeldrJiR
l~~mvcentralwgrtd
-.
Diameter Qf central vent grid= 670 mm Meeheb= 25
Areaofcerrtral wwgridm2= 0.353
Minimumheiiht abow charge if media is aama level as wnt grid (m)=
Voktmebadat whiimedi aissameteve ieewmtgnd%= XK) lb
Medh gradirrg-chambe$ 1
-
Sectoc- outlet
.
Oe#eke Nod
mm balls
z
100 0
m o
so o
66 3
2
:
:
: 3
w 3
56 4
50 0
0
z 2
35
30 :
-z ;
T@al
Nod x)
.
rim SALL 62.0
SUE MM
Intd
No.of
batle
o
0
3
12
0
7
6
2
73.5
OVERAU
CHAMSER 1 MEAN BALL SUE=
2.230 HID= 0.56s
67.8 MM
corrMkmdrr@te:- Char9econtaina ebniflcant cwentiiy of fine media and naada wmening.
The average ball size should be Increased to S0 Guidelines.
Sas DAWN Cabuwrxrs Wihirskateathetao mechergeexgemk mhaaomurud.
tab
-
totat=
46mmmperaegmed
emm
3120mmpereagment
450QPaf-
6.m nun
6.00 mm
TABLE 1
-
Mullnlet
...
3
Diameter d inlet insole Iinmg (m) =
Diamaterof inletcone beforeendwall Wars (m)=
0.8
0.6
~
Minimum hewght abowJchaqe if mediaiasame ISWIIaa inletlnmnbn (m)- 2.296 WD= o.m6
volumebad*whichrnedii iaeemelmml eainletcone%= s.m %
-al M
.
Rmaida= +20mmabowcherge (25Uoftddchafnb@
Middle =%me level (50% C4chamber)
Fsllingside=Eabmated-120 mmE&owcharga(25%dc hamb@)
(XmraIl swags matwial Might= -25 mmbelowchew.
chamber 2
-aaa
Ethxtiw Lengthm = 7.72 ~ *M till w:-
@41.81
internaldiarnatarm = S636 Aw!rags-mirw3.82m; rnmF3.865m.llnas=2 by3W mm.
Inbsmal vohmmm3 = 8922
Mnertypw- wmlypawith eomm WvspldBa.
Lineris Ingoodcondti.
MWm w- Urcumfafeniiel aids d Mersnt design tointsnnediste d@Msagrn
Thareiea soliiaecbrm whhnoabtain c41.69mdwwtar.
Thereiss gapbaksan the solidcenterandthe inrmarsegments.
~ -
Theouterrowof segmentsarebadlyblockedbybmlrenfundemkadmsdk
TtraarSaare badiywzfn OapecWylnttra innerSOgmOntO.
~e~y Siza Uld am-
12aqmenls
OUTER- H mm 75 mmsldw 28
123mms~ 22
SMsimemms 16 16 18 17 Ie
18
cendiidel@s= PslUly:&Jed byLn rLs.
Percentage biockage=
?ilaximumslc?ar eawithoutbbckage= O.mm m2
Effac4kalatarea with bkmkage= 0.1814 m2
S&t le n git w. 12 Segmmis
INNER- el@emm 60mmalda= 2 75mmekHa=
85mm slots= 2 120mmebts=
70mmalOb@ 1 126mmalota=
ac4@zsemm= 15 14 12 12
14 14 :: 13 3s
30
2
48
1s
cendiidakAa=
PadwaybY
bdummneli midia. -
Percentageblocke@=
Maxh-numslata reewithoutbbckege= o.6m9 M2
Effactkask4 areawittrbkckega= 0.3643 m2
Diaphragma?hxUVScme eectimsl area (m2J= 11.557
DiephragmaJ@areaaaaparcantc4CSA= 14,82 % mardmumifallatdesre-. - -
CX@wa9mslc4arae eesrnmmtafCSA= 6.45 %aOtualwkh blOckadakAS
Maxtmui SW size= 18 mm MWmumakxaiza= 12 mm
RammmW@eld- 8mm
Chemk20utletd@hragm antralvef4gftd
okmeterd-Ventgnd= @ii CeMar-nd@lceble
MedagmMgamk 2
.. -------- . -
6a@$x- Iniet MirtrSa
.
Ball Size Nod Nod
mm belle
.
m o 0
65 0 0
m o
75 : 0
m o 0
66 0 0
z : ;
50 2
45 ;
40 1: e
: ; 2
25 0 23
20 0 4
Totei
Nod 27 im
bstb
.-
MEAN 6ALL 38.7 30.2
S&X MM
OVemllrnaanbaneize= 20S4 mm
conditiond madiw- Themedksgmdin ginctudeaewnewwaizedmedi i d@m40mr nsl ze.
Li ghtc oat i ngon rnedWtinin@a butdiephmgmwe moracosted.
~w
23.8
tMrSltc+lsmbUrnakdal ls!ml=-12mm bek4vcharge.
l%odbek~hout
-==-==============
Beltspeed(fLWn)= Weight d Ssmple(lba)=
6ett speed(rn/min)= 4H Weight C4ample= n:
Feadtonnagatph= 73.19 Bettsectionlength(m)= 25
Fssd_st@r* 80.70 Rlehta= 8t%30cln@
Feedsisegmdlng
-===-.= ==S==
Size(ii)
2
1
0.75
0.s
0.375
Idnue 0%%
AUTHOR-D. BAIRO VEmiolt-1 DAIE-le/7m4
- Wekwm)
-
50.80
25.40 1.:
19.05 0.4
1;7CJ 0.6
o.e
6.35 1.1
20.1
waigMlb
tabP 6060mmpsra8gmawt
19. 27mm
,- --
TABLE 1 - CONTINUED
h) There was no evidence of nibs leaving chamber 1 so the first chamber was doing
adequate fd siie reduction despite its worn charge. Hence to maximise the mill output,
it is necessaryto maximisethe charge in chamber 2 as far as possible. Hence some 10,9
tonnes of fine me&a was added to chamber 2 to give 37.3XO volume ioad. To permit this
higher volume loading in chamber 2, a steel retaining ring was fitted to chamber 2 side
of the intermediate diaphragm. This reduced the vent grid diameter from 915 to 675mrn.
This would permit the volume loading to be raised fi.u-therto 38.9?40using 25-20mm
media once this media was available
i)
j)
k)
Chamber 2 contained some media of 38rnm size and above. To maximise the
he grindingpdormance of the charge, it is necessary to use a finer top up size
of around 25mm.
The inlet step chute feeder designwas modified to permit higher volume loading
at a fiture date if required.
The axial sampling test results (see section 4.3) and the mill inspection showed
that there was no need to add extra media to chamber 1 in the short term. Once
chamber 2 charge is maximized the mill still had spare motor capacity to permit
higher volume loadings in chamber 1 and 2.
Chamber 2 has a worn Magotteaux classi&ing liner. Unfortunately the
classi&ing Iiner design tends to reduce the mill internal diameter and fhrther
limits the maximum power that can be drawn. The mill is effectively
overmotored for the size of its shell.
4.4 Axial Samding Test Results
The attached data (Table 2) and axial curves (Figures 7 & 8) show the following:-
a) The size reduction in chamber 1 is good with acceptable residues in the sample
taken prior to the intermediate diaphragm.
b) The 45 micron sieveresidue reduction in chamber 2 is only average and the mill
would benefit flom a higher proportion of finer media of 25-20mm size. Please
note that USA cement mill ball charges do not tend to be as fine as UK ball
charges using the BCTC guidelines for fine media. Section B gives fbrther
details. The USA mills operate ve~ efficiently with high circulating loads and
it is known that very fine media (15-17mm size) tends to increase hold up within
the ndl.
Once the mill chamber 2 charge was optimised, fi.u-theraxial sampling tests were
necessary to decide on what action is required concerning chamber 1 charge
grading.
21
24
20
16
12
8
4
0
Fiqure 7 - AXIAL SAMPLING TEST CURVE - Example
Percent Retained on 2.36mm, 1.18mm and 300um Sieves
~
k v v

3 6 9
14 19 24 29 34
Distance from Feed Spout (ft.)
D 2360pm
+ 1180pm
o 300pm
1
70
60
50
40
30
20
10
Figure 8
- AXIAL SAMPLING TEST CURVE - Example
Percent Retained on 45pm and 90Bm Sieves
.

u
3 6 9 14 19 24 29 34
Distance from Feed Spout (ft.)
/
~ 90pm + 45pm
/
TABLE 2
AXIAL SAMPLING TEST - SIEVE RESIDUE RESULTS ,
0/0Cumulative residue at sieve size:-
Chamher 2.36mm 1.18mm 300mm 90mm 45mm
1 InletAl 12.9 17.8 24.8 46.5 55.5
InletB1 2.2 4.3 13.6 46.7 52.1
OutletCl 0.9 2.6 11.5 39.7 58.5
2 InletM 0.8 1.9 9.9 38.7 58.1
Met B2 0.1 0.2 3.7 32.1 56.0
InletC2 o 0.1 1.2 25 55.3
InletD2 0.1 0.5 1.2 21.5 54.1
OutletE2 0.3 0.4 1.3 20.0 53.7
Comparison ofmatial fnenessleavingChamber1
0/0Cumulative Residue
SieveSize SampleC1
I
BCTCGuideline
2.36rnrn 0.9 1
1.18mm 2.6 6
3ooum 11.5 20
encematerialleavingChamber1issuillciently fmetobehandledbyfmemediainChamber 2.
5 Gas Circuit Tests
Measurements of the mill airflow plus inieaking air levels showed that these were below the
optimum. The poor design of the stepped chute feeder and high inleakingair over the mill discharge
resulted in low airflow through the mill. Hence these areas needed to be improved by redesigning
the step chute feeder and sealing up the mill discharge hood, bag filter doors etc.
4.6 Power Drawn
Table 3 shows the mill power drawn estimates for both chambers, before adding the extra 10.9
tonnes media to chamber 2.
The mill first chamber was drawing approximately 12.33 kWh/tome cement at 65.3tph output.
Op_g cbber 2 charge allowed the output to be increased to 70.7tph which is equivalent to
11.39 kWMtonne cement. Both figures are higher than the recommended 9-11 kWhAome for
chamber 1. This data plus the axial samplingtest data showed that chamber 1 had adequate charge
to handle higher outputs. The mill was lacking power for fine grinding in chamber 2,
24 .
TABLE 3
MILL VOLUME LOAD AND POWER DRAWN
CHAMBER NO.
I
1
I
2 I TOTAL/AVE
I
Effective Length 3.37 7.66 11.03
r
Percent of total length 30.55 69.45
Diam. inside lining (m) 3.7900 3.8360
Mean Diameter (m) 3.82
Mill Speed (rPm) 17.44
Criticid Speed (rPm) 21.59
Percent Critical Speed 80.79
VOLUME LOAD
ESTIMATION
Ave Ht above charge (m) 2.2767 2.3925
0.601 0.624
0/0Volume load 37.27 34.41 35.28
Chamber Volume (m3) 38.02 88.53 126.55
Media Density (t/m3) 4097 4.288
Media Volume (m3) 1417 3046 44.63
Media Weight (t) 5805 130.63 188.68
Chamber Free Volume (m3) 23.85 58.06 81.91
MILL POWER
CONSUMPTION
Chamber nett KW (DAWN)
I
765
I
1813
I 2578
Total gross KW (Meter)
!
2714
i I
Ratio nett.lgross power :
0.950
I I I
Estimated power losses (%) 5.039
Gross KW per chamber
I
805
I
1909
I
2714
MILL OUTPUT I I I
Cement tph 65.3
Gross kWh/tonne per chamber 12.33 29.23 41.56
Cement M2/Ktz 391
I
25
FIG.9 BLUE CIRCLE INDUSTRIES - PLYMSTOCKWORKS
TOOL FOR CLEANING DIAPHRAGM SLOTS
Insert end into slots and level out
blockage material.
Dont allow nibs etc. to fall back onto
charge - put downsheets /
collect any material falling
bags to
out of slots.
SECTION A - A
I
4 b 65mm bar
I G welded tp
~6mm plate
m
plate
80mm
J
A
*90mm~
+230mrn+30mm-
---575mm G
30mm
4.7 Medium/Lonz Term OWimisation
4.7.1 Mill Diaphragms
a)
b)
c)
d)
Due to the breakage of small media in chamber 1 and the lack of any nibs trap,
there is a high tendency for diaphragm slot blockages. Solving this problem is
critical if the mills are to be optimised and if the circulating loads are increased.
Hence a correctly designed nibs trap is required with the following features:-
Locate aerated nibs trap in airslide close to elevator discharge.
Use double slide gates to discharge nibshroken media.
screen out nibshoken rneda and return cement to elevator using a small
vibratory screen beneath nibs trap.
Despite the diaphragm blockage problems, the material levels in both chambers
were low. Hence material flow through the mill in is not a major problem at
present. It could become a problemif the slots are not kept clean at higher outputs.
The outlet diaphragm design needs to be improved using a modem Pfeiffer style
diaphragm with 8mm circumferential slots. The replacement diaphragm should
have a cental ventilation grid of no more than 675mm diameter to allow up to
38.9%0volume load (speci& 4070 volume load in any enquiry). The present
diaphragm has oversized slots of 12-18mm size.
Use BCI diaphragm denibbing tool for cleaning diaphragms as per Figure 9.
4.7.2 Liner Plates
a) When liner plates need replacing in chamber 1- use a simple wave type liner.
Note:- this recommendation is not the normal liner recommendation (see section
B for details) but is necessary here due to the untypically high mill critical speed.
b) Chamber 2- the classifying liner reduces mill internal diameter. When this liner
requires replacement - consider a simple wave or ripple profile liner to increase the
mill internal dkrneter and maximise ball charge.
4.7.3 Media
a) Increase volume loading to 38.9?? in chamber 2 using 25-50 mm media top up.
Use typical USA mill me&a grading as per section B guidelines.
27
b) Top up charge in chamber 2 with 25mm instead of 38-50mm media in the fhture.
c) Regrade chamber 1 and review grading after retesting mill with extra charge in
chamber 2. See Section B for chamber 1 media grading guidelines.
5. MILL INSPECTION AND MAINTENANCE
During an axial test or as part of a programme of routine mill maintenance, it is usual to carry out
an examination of the mill internals. Regular inspection of mill internals is an essential part of
maintaining optimum performance. A carefid inspection will help to support the findings and
conclusions of an axial sampling test. The two processes should always be considered together.
Special attention should always be paid to the following points during an inspection:-
5.1 Lhinp Plates
Examine the plates for any $gns of wear, cmting and breakage. Normally one expects a reasonably
long life from lining plates and it is important to keep an eye out for any unexpected wear or
breakages so that suppliers quahty can be checked.
5.2 DiaDhrapms
Examine the diaphragm for any breakages, wear and blockage. If the diaphragm shows signs of
blockage thenit is important to determine what has caused the blockage as this can tiect what
action needs to be taken, for example, the presence of nibs could indicate the absence of sufficient
quantities of larger size media.
5.3 Media
,,
Inspect the mdla for wear and breakages. From the results of the axial test ball grading, check the
-g at the s@ed grading if this exists and from this determine what sized media should
be added or whether or not the charge should be re-graded. Note any differences between the
media levels in each chamber since too great a step up in level can cause hold ups along the mill
unless a lifter type diaphragm is used. Note any coating of the media due to poor mill ventilation
or moisture.
....
5.4 Voida~e Filling
During an axial test check whether the feed material fills the voids of the balls. Over-filling may
indicate diaphragm blockages and a restriction to flow whilst under-filling could be causing
excessive ball wear and heat generation.
28
Ifa mill has been brought down for examination due to a specific fault, for example, its output has
fidlenor nibs are present in the product, then there area number of possible explanations for this.
AppendixI lists some common cement millfaults together with their possible causes and remedies.
5.5 The Importance of Remdar Mill Maintenance and the Use of Axial Tests
Figure 10illustratesthe importance of regularly maintaining the correct level of charge in a mill by
indicating what happens when the charge is allowed to run down in a d~ raw mill, over a period
of time. It can be seen that as the power drawn by the mill has fallen due to wear on the charge,
the tonnage has fallen and the kWh/tonne have risen.
Approximately f40,000 per annum could be saved on power costs by restoring the mill to its
previous performance. In additionto power savings there would have been additional benefits due
to increased raw meal availability.
F@re 9 demonstrates the point that maintaininga mill at its optimum performance requires routing
recharging of the media with periodic regrading to remove tramp or undersized charge. Changes
in milling systems take place gradually over periods of years. Initially, lack of attention to
maintenancewill save repair costs but then gradually starts to increase operating costs as the plant
efficiencytails off. The plant may then be faced with the problem of having high revenue cost for
mill internals tier a period of say, 5-7 years time when Iinerskiiaphragmsneed a major replacement
programme. Cement/raw mill performance can be maintained closer to the optimum by having
short/mediurn/long term action plans with planned replacement schedules. Guidelines for the
selection of mill internals (liners, diaphragms, ball gradings) are reviewed in Section B and these
guidelines should be consulted before ordering replacement mill internals.
By regularlymonitoringthe millsperformance and by carrying out axial tests from time to time, it
should be possible to determine the optimum performance from a mill. In addition to providing
information on how efficiently the grinding process is being carried out within a given mill, axial
tests also enable an insightto be giveninto the effect of other process changes which can affect the
millsperformance. For example, the effects of feed pre-crushing, changes in gypsum addition rate
and feeding cooler clinker to cement millscan all be investigated more thoroughly by means of axial
tests.
There is a tendency to onlyconsider carrying out an axial test and other mill tests when something
has gone wrong and a millis not performingas well as it should. However, it is equally important
to cany out axial tests when a mill is performing well so that we can establish why it is performing
well. By carrying out axial tests on a regular basis it is possible to build up a record of mill
operating data, thereby enabling factors such as optimum charge grading to be determined.
29
.
,
,
,
,
,
.
I
I
,
t
1
0
8
.
0
m
*
N
w

o
n o
I
E
c

+
T
.
P
H
.
4
-
0
/
0
g
o
p
m
r
e
s
.
6. FINE MEDIA IN CEMENT MILLS
In order to maximise the effectiveness of fine grinding by the use of fine media (ie 25-15mm size)
in cement mills, a checklkt of essential preconditions is attached (as Appendix IIIB). Special
attention should be paid to these points when testing mills with fine charges. All too ofien, fine
media is not as effective as it might be, due to the reasons:-
G
Poor first chamber performance resultingin nibs/coarse material entering second chambers.
G
Lack of adequate mill cooling and ventilation leading to coating.
G
Excessive water injection plus (ii) leading to media coating.
G
D]aphragmblockage by nibshramp metalhoken media causing overfilling of the chamber
and cushioning of the fine charge.
31
1.
2.
3.
4.
5.
6.
7.
8.
APPENDIX 1A
CEMENT MILL FAULTS
Output Decreases
Cement Too Coarse
Cement Too Fine
Mill Fills
Mill Empties
Choked Diaphragms
Clinker Nibs in Cement
Formation of Coating
APPENDIX IA
CEMENT MILL FAULTS
1. OUTPUT DECREASES
POSSIBLE CAUSIJ
REMEDY
a. Increase in clinker hardness and size
Examine raw fed composition and kiln
burning conditions.
b. Mill charge too worn
Dump charge, re-grade and makeup new
charge.
c. Mill volume loadlng too low Measure each chamber and add correct
tonnage of new bodies.
d, Too smali bodies used for makeup Charge should be dumped and re-graded,
or (less satisfactory) use largest bodies for
make up.
e. Diaphragm slots partially blocked Inspect Diaphragms clean slots. Further
action as in Section 6
f. Fractured diaphragm Replace section(s).
+
2. CEMENT TOO COARSE
POSSIBLE CAUSE REMEDY
a. Increase in clinker hardness and size Examine raw fd composition and kiln
burning conditions.
b. Grinding media too large Make up with smaller media e.g. 50/50 of
the two smallest sizes.
c. Mill volume loading too low Measure each chamber and add correct
tomage of new bodies.
d. Diaphragm slots too worn Replace as soon as possible (with steel
diaphragms, weld bar into worst slots as a
temporafy measure).
e. Coating forming Reduce mill temperature. Use dryer feed.
Increase mill ventilation.
1
3. CEMENT TOO FINE
POSSIBLE CAUSE REMEDY
a. Output decreased See Section 1.
b. Last chamber media too small Dump and re-charge or make up with large
media.
c. Diaphragm slots practically blocked Locate source of blockage by sound and
examine diaphragm. Further action see
Section 6.
...
.
4. MILL FILLS
I
POSSIBLE CAUSE
I
REMEDY
a. Harder feed clinker Check rawfeed composition and kiln ~
burning conditions. Increase 1st chamber
piece weight.
b. Too much feed Run without feed until chamber sound is
normal. Then use slightly reduced feed. If
mill fills again consider (a), (c) and (d).
c. Worn gnridingmedia in 1st chamber
I
See Section lb.
d. Choked diaphragm(s) See Section 6.
e. Coating forming
I
See Section 8.
5. MILL EMPTIES
POSSIBLE CAUSE REMEDY
a. Insufficient feed Increase feed slowly.
b. First diaphragm slots worn See Section 2d.
c. Decrease in clinker hardness and size Check raw feed composition and kiln
.,
burning conditions decrease 1st chamber
piece weight.
d. Media in last chamber too large Dump charge and refill with charge of
smaller piece weight.
Note: I.famillfillsseverely, cleiuingisMIicultduetotheexcessivetemperature risethatresultsfromrunrihgtith
reducedventilationandfeed. Alittlegrindingaidpouredintoeachchamberthroughthemaindoorintum
(startingwiththelastchamber) accelerates theclearing.
,.
2
.-...,,
6. CHOKED DIAPHRAGMS
POSSIBLE CAUSE REMEDY
a. Steel diaphragm slots closed by Burn/grind out slots to original width,
metal flow replace diaphragm soonest.
b. Tramp metal in feed Greater care and tidiness when scrap metal
is handled and repairs are carried out in
proximity to clinker store and clinker
handling systems.
c. Grinding body pieces formed in Examine cast iron and very hard grinding
mill rneda for casting faults and excessive
brittleness. Inspect mill for very large
bodies mixed with very small bodies (if
found, dump charge).
d. Over worn grinding media Dump charge and make up with new.
e. Clinker nibs See Section 7.
f Coating forming in slots See Section 8.
7. CLINKER Nl_BSIN CEMENT
POSSIBLE CAUSE REMEDY
a. 1st chamber diaphragm slots too
-.
Replace soonest.
worn
b. 1st chamber diaphragm has open Repair as necessary.
crack/k ill fitting or loose
c. Exceptionally hard and/or size Examine raw meal quality and burning
feed control (if permanent, increase ball size in
1st chamber).
d. Maximum ball size in 2nd Increase piece weight by adding 70mm
chamber too small balls.
3
8. FORMATION OF COATING
,
POSSIBLE CAUSE REMEDY
a. Excessive moisture in feed Mix dry feed with moist feed to reduce
moisture content.
b. Inadequate ventilation Increase by opening fan damper ,or identifj
source of in-leaks after mill and reduce.
Inspect mill to find possible blockage
within mill. Differential pressure across
mill should be around 40-60mmwg.
c. Too high milling temperature Reduce clinker temperature, increase mill
cooling and ventilation.
d. Grinding body charge size Use grinding aid (including internal water
generally too large ~ cooling) if charge is otherwise petiorrning
satisfactorily i.e. ,when grinding RHC
(rapid hardening cement) in and OPC
(Open Circuit Mill).
I
,.
4
APPENDIX HA
THE FINENESS OF SAMPLES TAKEN IMMEDIATELY PRIOR
TO THE FIRST CHAMBER DIAPHRAGM
1. Samdin~ Method
Itis important to use common sense when taking samples within a mill. If visual inspection of the
material shows it to be relativelyfine, then small(200 gm) sampleswill suffice for carrying out sieve
analysis. However, if large quantities of nibs are present, it is advisable to take larger samples, i.e.
around 0.5-1 kg in weight.
2. Samrde Analysis
Coarse samples should be
recommended range for raw
Sieve Size
76 mm
50 mm
25 mm
12 mm
6mm
Microns
2,360
1,180
300
90
75
45
)
)
)
)
graded through the convenient sieve sizes available on site. A
mill and cement mill testing are as follows:-
Notes
Choose coarse sieve sizes to suit typical feed size of stone or
clinker to the mill. Of particular importance for raw mill
assessment.
Important for anaiysingsamples prior to the intermediate diaphragm in
raw/cement mills.
Important for analysingsamples prior to the intermediate diaphragm in
raw/cement mills.
Important for analysingsamples prior to the intermediate diaphragm in
raw/cement mills.
)
Use for mass balance calculations around raw mill circuits.
)
Important for analysing samples ex cement mill second chambers. Do
not normallyuse for raw mdl testing. Use for mass balance calculation
around cement mill circuits.
1
3. Guideline values for material samded before intermediate diaRhra~m ie. leaving
chamber 1.
G
Cement Mills
.,. ..
See Appendix III section (1) for details
G
Raw Mills
Check for any accumulation of material in the plus 12* size since residues can increase
if there is a buildup of coarse material at the diaphragm due to poor crushing action etc.
The 300 micron residue is also a usefbl gl.lideto ~-- --I--- --A--.-..:.
111 S1QlldlllUGl ~Gllu1lllaluu 1=.
Cumulative % Residue
on 300 Micron Sieve
~
33 (A)
43
03)
72 (c)
Efficient
underrun
first chamber handling fine stone - mill
- capable of slightly higher oytput.
Typical mill installation - efficient first chamber
handling reasonable size stone.
Mill with first chamber over-filled due to coarse
feed_size and inadequate ball size. Extra 90mm or
even 100mm media required to improve crushing
action.
..
As can be seem in the case of this design of a mill a more realistic target residue would be around
40% retained on 300 micron i.e. twice the residue considered acceptable for a cement mill.
2
APPENDIX HL4
ESSENTIAL PRECONDITIONS FOR THE UTILISATION OF SMALLER MEDIA
SIZES IN SECOND AND THIRD CHAMBERS OF CEMENT MTLLS
1.
2.
3.
4.
5.
6.
7.
8.
The first chamber charge should be in good condition and
fineness of samples prior to the intermediate diaphragm.
Sieve Size O/OCumulative Residue
2.36 mm l%
1.18mm 6%
200 urn 20%
When determiningthis fineness, it is necessary to take large
produce the following typical
samples and not to ignore any
unground clinker nibs present. If nibs are present, the first chamber media grading should
be checked and adjustedwith extra 90/80mm instead of 70rnmmedia. Any media less than
55mrn should be removed.
The first chamber should have an efficient reverse step/lifter type lining.
The diaphragmslots must be clear and even with no excessive gaps which can allow coarse
material to enter the second chamber from chamber 1.
The outlet diaphragm must be clear of blockages, etc, to avoid chamber overllling and
hence cushioning of the fine charge.
Check the second chamber axial sampling cume for any signs of nibs/coarse material from
the first chamber which can inhibitthe fine grinding characteristics. Figure IIIA shows and
example of two rni.11% one with an efficientand one with inefficient first chamber. It is noted
how the second chamber of the latter mill has to carry out some of the crushing action
which should have been carried out in the first chamber.
The mill must be adequately vented, i.e. with typically three air changes per minute (or
0.25kg air per kg cement) for open circuit mills or typically five changes (or 0.4kg air per
kg cement) for closed circuit mills. These airtlows refer to air flow actually entering the mill
inlet which is oflen lower than at the filter due to inleaking air.
Avoid high millingtemperatures above 120C, which can cause coating in the second (and
third) chambers. Coating can adversely affect the pefionnance of media especially when
using finer charges.
Excessive water injection to control otherwise high milling temperatures should also be
avoided wherever possible in order to prevent hard coating on media.
50
40
30
20
10
0
FIGURE lIIA
!
------------- ------
------
--
,--. O-
--
---
-*
+ 45 micron
-*-
G9 MILL B
G99... -.-..=* -mm- m-------- ------
b----
---0 ----
G
+ 2.36 mryc.
Ge* .*
G9*
***
B
/
Nibstransfenedftom
Ge.
G*-
..*
. chamberone of Mill B
-.
--
A
0-
0---- 0-----------
n I
FIGURE A - AXAIL SAMPLING
TEST -
SIEVE RESIDUES FOR
CHAMBER TWO
ILL A
I ) Efficient first chamber
charge and lining.
ii) Fine ( 25- 15mm ) media
in chamber two.
J!!llLu
i)
ii)
iii)
I
I
20
40
6i
80
0
% CHAMBER LENGTH
100
inefficient first chamber
nibs prior to
intermediate diaphragm
pooriiningand charge.
Conventional charge
(60 - 17mm ) in chamber two.
Poor classifying iining gives
reverse classification.
II
APPENDIX IVA
AXIAL SAMPLING CURVES FOR CEMENT MILL
SECOND CHAMBERS - SIEVE RESIDUES
General
In additionto examinhg the surface area versus nett kW drawdchamber length relationship for the
second chamber of a cement mill, it is also usefid to examine the sieve residue relationship.
This is best examined by using a 45 micron sieve or even finer sieves such as 32 and 25 micron
sieves if these are available.
Very oflem the ti~ sampling curves for second chambers show a steady reduced in 90 and 300
micron residue throughout the chamber. This gives the impressionof efficient grinding taking place.
However, cement is predominantlyfiner than 45/32 micron and to assess the fine grading efficient,
these sieve residues should also be examined.
Figure IIIA shows axial sampling curves for two closed circuit cement mills. Mill Awas an
efficient mill using fine media (25-1 5mm) in its second chamber. The material leaving the first
chamber of this mill was fine and within the guidelines shown in Appendix III.
Mill B used a ftily conventionalmedia gradhg consisting of 60mm to 17mm media. The mill also
had a form of class@ingliningwhich did not fi.mctionwell and resulting in reverse classification of
the media.
Comparing the 45um residue shows that the material at the inlet to the chamber was of similar
fineness i.e. around 50% cumulative residue.
However, the chamber outlet samples were as follows:-
Mill Outlet Residue Reduction in Residue
~ 45 micron Across Chamber
A 46% 4%
B 28% 22%
This shows the superior grinding characteristics of the finer mdla used in conjunction with an
efficient first chamber charge in Mill A.
To improve Mill Bto give a similar performance would have involved costly replacement of the
liner/charge. It was therefore decided to reduce the average replacement media size from 60mm
to 25mm size and replace some of the coarser 60mm media by the available 25-17mm media.
APPENDIX VA
EXAMPLE OF HOW THE VOLUME LOADING WITHIN A MILL CAN BE
AFFECTED BY THE ACCUMULATION OF UNGROUND MATERIAL (NIBS)
G
Background
The following results were obtained on a 4200 KW raw mill, which was designed to
produce 220 tph raw meal. The mill was crash stopped afier running at 75% of its design
load during commissioning. It was decided to cany out an axial sampling test on the mill
before raisingthe charge to 90% load. The millwas crash stopped and it was noted that the
first chamber was very fill of unground stone. Later on the mill was run out and the height
above charge measurements were rechecked.
G
ADDiirent First Chamber Volume Loadinp Followinp Crash StOD
Average height above charge (H)
=
Inside lining diameter of chamber (D) =
=
Apparent volume load
=
G
Volume Loading After Mill was Run Out
Average height above charge (H)
=
=
Volume load
=
Tomage of media loaded into mill
=
Chamber internal volume
=
Volume of media
=
Media Density
=
3.07m
4.42m
3.07
= 0.695
4.42
25.5%
3.275m
0.74
20%
88 tonnes
92.1m3
3251 x 0.02= 18.42m3
88.0
= 4.78 tlm3
18.42
G
Amarent Media Densitv
Due to the presence of nibs occupying approximately 25.5- 20.0= 5.5% of the ,
internal volume of the mill, the apparent media density is less than that shown
above.
i.e. Volume of media plus nibs
=
92.1 X0.255= 23.4@
88.0
Apparent media density
= = 3.75t/m3
23.49
G
Conclusions
Hence, unlessthe volume loading of the chamber had been checked with the
mill run out it would have appeared that the chamber contained more media
than it actually did, i.e:-
92.1 X0.255X 4.78
=
112.3 to~es
Another means of cross-checkingthe actual media loaded is to calculate the
KW drawn for each chamber using the power formula and cross-checking
these results against the figures from the mill kwh meter.
2.
APPENDIX VIA
MASS BALANCE ON A RAW MILLING CEMENT
General
The example chosen to demonstrate mass balance calculations on a milling circuit is that of a
double rotator mill. It was felt that this represents one of the more complex closed circuit
millingsystems availablewhichcouldbest illustrate some of the techniquesJproblems involved.
Generally speaking, most closed circuit cementlraw milling systems should be simpler to
evaluate than the example shown.
Mill Circuit Samples
Rejects from static separator
Fines from cyclone atler static separator
Rejects from Wedag separator (feed to
chamber 2)
Feed to Wedag separator
Mill central discharge
Fines from Wedag separator
.. Finished raw meal
Mass Balance Calculations
G Mill Throu~hDut
Mill feed rate
Feed moisture
DWtonnage
G Finished Raw Meal
0/0Cumulative retained
on 90 micron sieve
66.00
9.24
76.22
52.90
49.50
2.70
6.30
=
42.8 tph
(from totaliser readings
..
prior to test)
=
3.1%0
=
42.8 X0.969= 41.5 tph
Dry tonnage
=
41.5 tph
The finished raw meal consists of (a)
(b)
:. (41.5 -x)
Mass balance on 90 micron residue
(2.7) + (41.5 -X) 9.24
x
Hence, fines ex cyclone
fines ex cyclone
fines ex Wedag separator
=
tph of iines ex Wedag separator
=
tph fines ex cyclone
=
41.5 (6.30)
=
18.7 tph
=
41.50 -18.7 = 22.8 tph
G
Reiects Feed to the Seuarator
The fd to the separator consists of mill discharge material together with rejects from
the static separator.
Basis 1 tph of separator feed
let y
=
tph of material ex mill
central discharge
Mass balance on 90 micron residue
y (49.5)+ (l-y) 66.0
=
52.90 X 1.0
Y
=
0.79
i.e. 79% of the feed to the separator consists of material from the mill central
discharge.
G
SeDarator Performance
(a)
(b)
(c)
2
o/ORecirculating Load
76.22 -2.70 = s 15
Circulating factor C
=
76.22 -52.90
Recirculating load
=
C-1 = 2.15 i.e. 215%
Fines Efficiency
100 (100 - 2.7)
= = 65.6?40
315- (2.70 + 2.15 X 76.22)
Coarse Efficiency
100 (2. 15 76.22) = 984%
=
2.7 + 2.15 X 76.22
The coarse (or rejection) efficiencyis good. The fines efficiency is on the low side due to the
fine cut point of the separator (2-3?40 plus 90 micron as opposed to 5.0%+90 micron target -
see comments in conclusions section). The recirculatingload estimate does not agree with that
indicated by the rejects weigher (1090A). However, the two cannot be directly compared.
G
Mass Balance Over SeDarator
Fines ex separator
=
Let rejects
=
Balance on 90 micron residues
(R+ 18.7) 52.9
=
R= 40.3 tph
Hence, separator feed
=
G
Reiects ex static SeDarator
These form 21% of feed to separator =
Hence, mill central discharge
=
G
Indicated Versus Actual Recirculating Load
(a) Indicated
Mill rejects weigher
=
!- Mill feed dxy
=
Hence, indicated recirculating load =
0)
Estimated
Calculated rejects ex Wedag =
Mill feed
=
Estimated recirculating load =
18.7 tph
R tph
R (76.22)+ 18.7 (2.7)
40.3 + 18.7= 59.0 tph
59.0 X0.21 = 12.4 tph
59- 12.4= 46.6 tph
45.1 tph
41.5 tph
45.1
x 100 = 109?/0
41.5
40.3 tph
41.5 tph
40.3
= 97%
41.5
These are within the emors of accuracy that can be expected.
3
( Figures in brackets are 90 micron sieve residues)
b
Fresh Feed _
41.5 tph
42.8 tph
Separator
4
I
T
I
(6&.0%)
\
12.4 tph
11
n
(76:22%)
40.3 tph
Fines
Cyclone
18.7 tph
(9.24%)
22.8 tph
(6.30%)
Finished raw
FIG. VIA -
MASS BALANCE ON RAW MILL CIRCUIT
meal 41.5 tph
Mass Balance and Conclusions
Figure VIA gives the completed Mass Balance.
The above mass balance must be treated cautiously for the following reasons. These illustrate
some of the problems with calculating circuit mass balances.
1. The estimated finesfromthe cycloneis 22.8 tph or 55% of the finished raw meal. This
is most probably an over-estimate. Normally, one would expect only 30XOof the
product to come from this cyclone.
2. The mass balances are carriedout usiig 90 micron residues. Considering the very fine
product produced bythe Wedag separator it maybe advisable to use a finer sieve size
for constructingthe mass balance. It is best to carry out fbrther sieve gradings on say
45 or 75 urn residuesto see whichresults relate best to the measured rejects flow rate.
3. In the example showq the Wedag separator was producing a finer product in order to
compensate for the coarser product from the cyclone after the static separator. The
target residue was 5V0on 90 micron. This proved to be usefid information. Atler the
te~ the static separator was inspected internally and it was found that the suspended
bobbeneath the central cone of the separator was misaligned. This was subsequently
put right and finer residues resulted.
4. To cany out a fill assessment of the separator performance the efficiency figures
should be estimated over the fill range of particle size results. Reference to a single
residue efficiency will only give a very rough guide to separator performance. When
comparing results from the same mill, reference should be made to the level of
recirculatingload, i.e. an apparentlypoor fines efficiencyresult may simply be a feature
of operation with highrecirculatingloads rather than any fault with the separator-itself.
Construct TROMP curves for the separator product using fi.dlpsd analysis.
5. When two fines or rejects streams are mixed (as in the above example) if the analysis
of the materials before and after mixing is similar then it becomes very difficult to
estimate their respective tonnages. If sieve residues are no help - try tracer techniques
or check for any chemical variation.
6. To improve accuracy - several samples should be taken horn the circuit and analysed.
APPENDIX VHA
EXAMPLE OF THE OUICK TEST METHOD
FOR AXIAL SAMPLING TESTS
1. MILL INSPECTION REPORT
Tables VII (i) and VII (ii) and Figure VIIA sudse the essential data from the mill
inspection concerning media gradings, liners, diaphragms, etc.
1.1 Observations
The following points were noted during the test and inspection.
G
G
G
G
1.2
There is an excessive quantity of unground nibs at the intermediate diaphragms.
The first chamber lining has a relatively shallow step of only 45mm compared with a
recommended 70mm. This gives a poor crushing action. There is an identical mill at
this Works whichhas a conventional 70mrn reverse step lining. The latter mill has no
evidence of nibs before the diaphragm. Both mills have a good coarse charge in
Chamber 1 which is more than adequate for the fine clinker feed size.
The estimated gross mill kW is too high using the DAWN formul~ i.e. 2390 kW as
opposed to an actual 2200 kW absorbed. This over-estimate is most probably due to
the presence of large quantitiesof nibs in Chamber 1which causes the charge to bulk
and give a fhlseover-estimate of the volume load. It is best to re-check volume loads
after the mill has been run out.
The intermediate diaphragm has gaps between its adjacent segments which allow
coarse material to enter Chamber 2. The outlet diaphragm slots are badly blocked and
the second chamber should be tipped and regraded to remove undersized and broken
me&a (i.e. -12mm size). The effective area of the outlet diaphragm is low and fiuther
blockages could cause hold-up problems. All slots should be cleaned out and all gaps
between segments should be sealed up. New diaphragm segments are necessary to
replace all the badly worn and burned outslots which have jagged profiles.
Test Method
The number of samples taken was nine in total (four in chamber 1 and five in ;hamber 2).
TableV5(i) shows the sievegradings analysed. Note that greatest importance is attached to
sieve residues at 2.36 urn, 300 urn and 45 um. Surface area results in chamber 2 could also
be carried out. The mill was crash stopped in the early hours and allowed to cool for 1-2
hours using modest mill cooling airflow. All samples were analysed and the results, plus
conclusions, drawn up (axial sampling cumes, media gradings, belt weigh off results) within
the same day.
Samples of the separator f- fines and rejects were also analysed. These confirmed that the
separator efficiency was reasonably good and was typical of medium efficiency separation
using a Wedag type separator.
1
TABLE VII (i
) : MILL INSPECTION REPORT
CHAMBER
- .1 2
Chamber Lc nglh(m) 3.76 8.484
El%c tivediamekr(m) 3.63 3.658
Averagehcighl above chwgc (m) 2.354 2.358
/o %hrmc load 31.38 31.85
Chamber volume (m) 38.91 89.16
Media volume (m) i2.21 28.40
Media density assumed (Vm) 4.3 4.5
Estimafc~ weight m~la(1) 52.50 127.80
Chamber ficc volume (m) 26.70 60.76
Wctt kW drawn fA6.48 1576.04
3ross kW drawn 695.14 1694.67
rotrrl 8rosa kW drawn (2389.81 - too high)
4ctual grow kW drawn
Factor formcdirr density, eta to
~chicve richra[ gross kW
3cpth of ccmcnt above charge Material below chmga Material Gt acme Icvel
hnrling 0( mcdirJlinin@
.incr type Bolkcc Rcvcrai Step Ckasifyitg
Mpth atcp on Chamber I lining NtA
Mill speed -15.7 rpm
Weighted avcrcgc internal dinrnclcr (m) = 3.649m
X0critical -71 .lOA
CHAMBER
Orrtkt dianhramq 1
2
mm slot size (ran8e) Gvcraga 7 (range 4-8)
9 (range 8-9)
Maximum clot ma with sII slots ckar (m) cffkdivc 0.564 0.67912
Actual slot ara (m? cfktivc 0.34304 0.41264
D@rhragm C.!LA. (m~ 10.349 10.509
G/0 slot am maximum 5.45 6.46
G/. slot ma actual 3.31 3.93
~cdia Distribution
Average
Inkt Ball !%m
Size (m) Weight % Sample pt. (mm)
Mcdii sin distribution by either random 90 23.2 31.0
rampling ((X) ot ball weighthize CIl) 85 25.0 : 27.5
80 !3.9 3 27.5
75 13.5 4 19.5
70 10.9 s 18.0
65 10.0 outlet
60 2.9
-60 0.6
x
3vcrall wcra8c ball aim in chamber (mm) 77 .21.5
h pqmliOrr +30mm media in CD NIA 3.78
~mpchcs of tramp matcnal or broken media in chmgc Negligible
~ proportion of undcrsi?xd (@mm) mcdii in CI 0.6Va NIA
II
1
2
3
4
1
2
3
4
5
\
TABLE VII (u) . AXIAL SA
..
.
MPLING TEST SAMPLES$
CUMULATIVE h RETAINED
o
+ 10mm +5mm +2.36mm +1.18mm +300/A +90/A +45/z
2.96 8.17 14.62 20.76 34.63 50.16
3.18 5.26 9.15 15.56 35.40 49.17
1 3.09 .3.53 4.77 8.81 27.32 ,49.75
5.23 5.44 5.77 7.85 24.25 45.17
.
0.74 1.84 16.66 41.10 67.02
.
0.46 0.62 6.68 31.43 59.65
2
.
0,32 0.38 3.07 20.20 54.67
. .
0.29 0.31 1.31 12.57 43.48
2.58 2.60 3,48 13.22 57.34
1
!
!
I
l
l
I
I
1
1
1
1
1
1
1
1
1
1
I
I
1
[
I
I
1
.
o
0
m
N
o
0
3
m
v
m
m
a
%
SECTION B
CEMENT MILL OPTIMISATION
SECTION B
GENERAL GUIDELINES TO OPTIMISE
PERFORMANCE OF OPEN AND CLOSED CIRCUIT
1.
2.
3.
4.
5.
6.
7.
8.
9.
1 0 .
11.
CEMENT MILLS
CONTENTS
INTRODUCTION
MILL TESTING/ROUTINE INSPECTION
MILL CONFIGURATION -CHAMBERLENGTHS
MILLLINERPLATE DESIGN
4.1 First Chamber Liner Design
4.2 Second Chamber Liner Design
MEDIA GRADINGS
5.1 Chamber 1 Media
.5.2 Chamber 2 Media
DIAPHWGM DESIGN
MILL VENTILATION
MILL POWER DRAWN, MILL CRITICAL SPEED AND MEDIA LOAD
GUIDELINES FOR MILL FIRST CHAMBER POWER
CONSUMPTION
CLOSED CIRCUIT MILL OPTIMISATION
THE PURCHASING OF NEW CEMENT MILL INTERNALS
APPENDIX lB: Recent experiences with cement mill internals
APPENDIX IIB: Notes on the performance of USA cement mills
.... .
1. INTRODUCTION
This paper is intended as a set of generid guidelines to assist plant personnel in optimizing the
performance of open and closed circuit cement mills. Very oilen, when efforts are made to
optimise a millsperformance, certain compromises may have to be made in the following areas:-
G
G
G
Me&a grading - available stocks of media plus the wish to avoid excessively high media
replacement costs often results in compromise charges being used.
Existing plant design limitations - older mills often have inadequate ventilation/filter
plants, etc. Whhout major redesign/capital expenditure it is often not possible to achieve
current ventilation standards.
Factors such as veryhighclinker fd temperatures and inadequate mill cooling will often
limit the extend to-which, say, fine media can be used in secondhhird chambers.
Hence, common sense has to be applied when mill optimisation is carried out and all too often
compromises are necessary. It is important to be aware of the factors which limit a mills
petiormance so that on-going improvements can be made in key areas. Routine mill testing and
inspection is essential to any programme of mill optimisation. When the opportunity arises to
repla% diaphragmdiners, etc., it is essentialto make sure that outdated designs are not repeated
and that the latest/optimum designs are applied.
2.
G
G
G
G
3.
MILL TESTING/ROUTINE INSPECTION
Axial samplingtests - make use of the quick test method which concentrates on getting
the maximum amount of essential information out of the minimum amount of
sampling/analysis. (See 3.2 in section A).
A minimum of one axial sampling test per mill per year is essential. Preferably 2-3 tests
shouldbe carried out per annum on each mill. there is no point carrying out tests if firm
conclusions are not drawn and if recommendations are not acted upon.
Make greater use of cameras to obtain a photographic record, of mill inspections.
Photographs are usually more convincing than words.
Routine mill inspections shouldbe carried out on a monthlybasis in order to check for any
obvious failures or problems plus charging levels as a preventative measure. This will
ensure that mill internals are maintained.
MILL CONFIG~TION - CHAMBER LENGTHS
Typical chamber lengths for open/closed circuit cement mills are as follows:-
/o OF OVERALL MILL EFFECTIVE LENGTH
MILL CLOSED
Chamber 1 30 28-34
Chamber 2 20 72-66
Chamber 3 50 Not applicable
(plus 4 if fitted)
I 100% I 100%
The above is a general guide only. Certain ding systems, such as closed circuit mills for white
cement, ofien use much shorter first chamber lengths (eg Ras Al Khaimah - first chamber = 14Yo).
Similarly, the use of pre-crushing equipment such as Roll Presses often makes it necessary to
adjust chamber conjurations/charges etc. Chamber lengths are often a compromise, especially
if the mill is used for more than one cement type, eg RHC and OPC. First chamber lengths have
to be suitablefor the worst case, ie highest throughput at the lowest surface area and the hardest
grindability.
2
The followinggeneralisedguidelinescan alsobe appliedwhen selectingthe optimum first chamber
length:-
QisJ3 % of total effective lemzth
High critical speed (75%)
Stepliner, Coarse Ball charge 28?40
Average surface area (350 mz)
Low critical speed (70-72%)
Low surface area (280-300m2/kg) 34%
Reference is drawn to section 9 of this paper in which the power input requirements of chamber
1 are reviewed.
3
4. MILL LINER PLATE DESIGN
4.1 First Chamber Liner Design
G
The standard recommended liner design for Chamber 1 is the reverse step liner with a
ripple profilewhich is availablefrom several manufacturers, e.g. Magotteau~ Bradley and
Foster, etc. See Figure 1 for details of this liner design.
G
G
G
G
When this liner wears, the step distance is reduced and there is less separation of the
charge. Some older liners were often provided with a relatively shallow step of only 40-
50mm. This is insufficient and the recommended step sizes areas follows:- ,,
Minimum = 60mrn
Maximum = 80mm
Typical = 70mm
The optiium step size is a fi.mctionof mill speed. With slower speed mills, it is desirable
I
to use the maximum step size in order to compensate for the loss of lifting action.
Reverse step liners are suitable for ball mills with critical speeds in the range 69-77Y0.
However, be carefi,dof using this liner design if the mill has a very high critical speed of
80-81%. In the later case a simplewave type liner is suitable as referred to in the example
given in section ~ part 4.
Lorain liners (see Appendix IB) are usefid for USA style cement mills featuring high
circulating loads with grinding aid usage. Thk liner gives high lift characteristics which
can help to overcome the cushioning effects of very high material throughputs. There
agai~ avoid using this liner if the mill critical speed is very high at 80-810/o.
If a reverse step liner is installed to replace a smooth type liner - remember to take into
account any limitationsto the maximum mill loading/power drawn (due to any reduction
in mill internal diameter).
A recent development in liner design is the DUOLIFT liner available from Magotteaux
(see Figure 1.1). This liner is claimed to give more effective lifling/crushing action than
reverse step liners. However, following installations in the UK and overseas, we have
serious reservations concerning this liner design (see Appendix IB).
Under no circumstances use a smoothtype liner such as the Voest-Alpine grooved liner
design, which gives little lifting action and tends to result in high levels of nibs. This
design has recentlybeen altered to incorporate activatorplates, although there is no data
to show whether its petiormance has been improved.
4
FIGURE 1: BOLTED REVERSE STEP LINER FOR CHAMBER ONE
N OF
ROTATION
i
NOTE - REVERSED STEP LINING PLATE DIMENSIONS ARE TYPICAL ONLY
FIGURE 1.1: DUOLIFT LINER
L,,,
/.
/ I
I w
b
\\
m
o
v
4.2 Second Chamber Liner Desire
G
Classifying liners - several designs are available and these give varying results. Some
liiers (e.g. UBE type) can even give complete reverse classification! C1assif@g linings
came into vogue when open circuit mills were converted Ilom say 3 to 2 chambers using
conventional (60- 17mm) media in Chamber 2.
G
It has been found that classiig liningsoften class@well on conventional mediL but less
well on finer sized media (i.e. 25- 15mm). Hence, with the current tendance to use finer
reed% there seems littlejustificationinusing classifying liners. A simple ripple type liner,
such as the FLS DRAGPEB, is usually adequate with finer sized media.
G
A disadvantage of classi@ingliners is that they can allow clinker nibs to be classified as
small media, causing build-up of material and blockages of second chamber outlet
diaphragms.
G
Another disadvantage of classi~ng linings is their greater overall depth and the loss of
effective mill diameter/power drawn which results. In addition, where mill shell water
cooking is used, the thickness of this liner limits heat transfer. Potential 10SSof mill
diameter is particularly significant on smaller diameter mills.
G
In cases where the mill is not restricted by volume load considerations (ie, the fill motor
load can be drawn at around 28-32V0volume load) classifying linings can be an effective
lining with a long effective life.
G
For USA cement mills, the Lorain liner can prove effkctive in chamber 2 since it allows
some lifl characteristics to handle the high material throughput. The Lorain liner also
does not cause the same loss of internal diameter as a classifying lining. This is an
essential feature for mills operating with up to 40/0volume loading in chamber 2.
..
G
See Appendix IB for fi.uther details of current chamber 2 liner designs
7
5. MEDIA GRADINGS
5.1 Chamber 1 Media
Typical recommended first chamber media gradings are within the following range of fine and
coarse gradings. All media should be H*d with matching liner/diaphragm materials.
I
WEIGHT VOIN CHAMBER
Charge Type: Fine Coarse.
Ball Size
mm
90 23 36.5
80 32 29.5
70 20 24.0
60 25 10.0
I
100%0
I
100%0
,,
The choice of media gradings for Chamber 1 will depend upon several factors, ie:-
G
Size and hardness of clinker feed to mill, together with the length of the first chamber.
G
Type of liner - if a poor first chamber liner is fitted, it may be necessary to use a very
coarse gradingto compensate for the lack of any lifting action. With the new DUOLIFT
liner design there has been a tendency to use a relatively fine first chamber charge. This
is not yet proven as validwithinBCI. As a general rule, the coarser grading shoivn above
is probably more applicable within BCI UK plants.
.,.
G
As a general rule, the Coarse grading is more suitable for
(a) Low critical speed (68-72Yo)
(b) Worn step liners
(c) Coarse clinker (50mm plus)
(d) Hard clinker
(e) Short first chamber (28Yo)with significant production of lower SSA cement ie,
300m2/kg.
The Fine grading may prove more suitable for mills which have well sized first
chambers, grindingfine clinker fkedto higher cement surface areas (370-440 m2/kg). The
above are rough guidelines and have to be applied with common sense. Hence, if a
8
cement mill has to produce a wide range of cement types o~ say, 300 to 440 m2/kg S.S.~
some compromiseswith respect to grrdng may have to be made. Hence a coarser ,charge
maybe needed to suit grinding at 300 m2/kginstead of the optimum finer media grading
required for 440 m2/kg. Hence the tendency is towards the coarser grading in many cases.
.
G
As a general guide, the first chamber charge should be adjusted to give the following
typical fineness of material sampled immediately before the intermediate diaphragm.
o/OCumulative Residue
Sieve Size
2.36rnm 170
300 urn 20%
When taking these samples, it is essential to take large samples (1 kg) and sieve all the
material at 2.36mm in order to detect any nibs of unground material (see Mill Testing
paper). If nibs are present coarsen the charge and check effectiveness of liner design.
5.2 Chamber 2 Media
G
Current practice favours the use of liner media (i.e. 25-15mm size) in second/third
chambers instead of conventional 60-17mm media gradings since fine media is more
efilcient for fine grinding. Under the correct milling conditions, fine media has been
proven to give increases in output of up to 25V0,subject to the initial condition of the
media grading as well as the mill itself
G
Under the followingadverse conditions, the effectivenessof fine media can be limited by:-
...
Inadequate mill ventilation
Poor cooling within the mill
.- Excessivelyhighmillingtemperatures and/or high water injection rates leading to
coating problems
Chamber overfilling problems due to blocked outlet diaphragms, etc., causing
excessive material hold-up
Excessive coarse material entering ex Chamber 1 due to factors such as poor
diaphragms, liners and/or media in Chamber 1
Under these conditions, fine media will tend to become coated and its fine grinding
petiormance becomes limited due to cushioning effkcts, etc.
In the Mill Testing Paper (Appendix 111A)a list of essential pre-conditions fpr the optimum use
of fine media has been drawn up. These preconditions should be examined for individual mills
where it is proposed to use finemedia.
,,
G
As a general guideline, typical fine chamber charges for two chambered open and closed
circuit mills for UK style cement mills are as follows:-
WEIGHT VOIN CHAMBER
Ball Size
mm Open Closed
=-E-E
I
100% ,1 1(.)Ovo
I
,,
This applies for closed circuit millswith 150-200% circulating load. Whh three chambered open
circuit millsthe media is distributed with typically, 30-25mm media in Chamber 2 and 20-15mm
in Chamber 3.
The above gradings are only typical and in practice may be finer/coarser according to
circumstances. In most cases the grading actuallyused in a mill will be a compromise which takes
into account the following factors:-
The quantity/cost of rejecting +30rnm media horn the charge. Obviously the finer the
charge is made, the greater the cost of replacing ovetsized media becomes.
Adverse milling conditions such as coating may make very fine media less effective due
to coating and cushioning. Under these circumstances a slightly coarser or conventional
charge may be used. Common sense has to be used when deciding upon charge fineness.
I.fa xnilIalready has problems with material transportatio~ then finer media may worsen
these problems. Attention to diaphragms and possible use of grinding aids can help to
overcome some transportation problems. ~ ,,
G
USA cement mill - due to the highcirculatingloads (350-400%) u~ the second chamber
media gradings do not tend to be as fine as shown above. A problem with fine media is
its tendency to expand and incr~e the apparent volume loading. Thk makes diaphragm
design more critical and can limit the maximum power drawn by the mill.
10
Ball Size
(mm)
38
32
25
22
19
Coarse Grading
Weight YO
17
18
29
27
9
Fine Grading
Weight %
o
15
30
30
25
II 1 I
1~
100%
1Oovo
Note that the proportion of 19rnrnmedia is significantlyless than the typical UK mill guideline
gradings. h additio~ very fineme&a of 15mrnsize is not normally recommended. Although the
high usage of grinding aids can help to offset some of the problems associated with using fine
media, we have not yet tended to use very fine media gradings in the USA mills. Some notes
concerning the performance of USA cement mills are contained in Appendix IIB.
11
6.
G
G
G
DIAPHRAGM DESIGN
All diaphragms used should be of the double diaphragm/lifter/flow control type, i.e. the
level of material in the chambers is controlled by a positive conveying action. An
adjustable form of flow control is highly desirable to give variable control of material
levels within the chambers.
All diaphragmsegments should be constructed of wear resistanthard wearing materials.
Avoid use of steels which can spread thereby causing peening over of the diaphragm
slots leadhg to closures and blockages.
Optimum slot sizes and slot areas depend upon the mill lengthkliameter ratio as well as
the circuit design (i.e. open or closed).
Typical guidelines are as follows:-
Ot)en Circuit (Typical 4 to 5.5 L/D Ratio)
Chamber Outlet Diaphragm ~ z
Slot size (mm) 67
Slot area (%) 4-5 5-6
Closed Circuit (Typical under 3.5 L/D Ratio)
Chamber Outlet Diaphragm 1 z
Slot size (mm) 68
Slot area (?!) 5-6 6-8
Note the followk
G There shouldbe a steady increase in slot size and area on progression through the
mill. This is to avoid excessive material holdup due to, say, the build-up of nibs
in diaphragms. It ensures that any coarse material ex Chamber 1 will not cause
blockages in subsequent chambers.
G When former open circuit mills of 4-5.5 L/D ratio are closed circuited, it is
essentialto maxiniiseslot areas to avoid material transportation problems through
the mill. If slot areas are not maximised, this can otlen impose a limit on achieving
the optimum mill recirculating load as well as the minimum cement residues (i.e.
narrowest particle size distribution).
G
In the above tables the quoted areas are for the slots only, they do not include the
area of the central ventilation grid. This point must be made clear to suppliers
when orderingkpecifing new diaphragms.
12
G
Slots should preferably be circumferential and the slot profile should not be of a type
whichencourages blockages by tramp metal/clinkernibs, eg, provide a self cleaning slot
profile as per the Pfeiffer and Magotteaux Opticontrol diaphragms (see Appendix IB)
7. MILL VENTILATION
s
Recommended mill ventilation airflows are as follows:-
Ckcuit Open Closed
Air changeshnin 3 5
kg aidkg cement 0.25 0.4
G
When designingventilationsystems, filters should be seized with adequate allowance for
inleakingair over the ventilation circuit. Filter plants should be sized for at ieast 5 (open
circuit) and 7 (closed circuit) air changes per minute. All air volumes are related to
changing the free volume above the charge. Air volumes are based upon a standard air
temperature of110C.
G
The millinlet, outlet and diaphragmdesignshould not impose a high restriction to airflow
through the mill. Ensure that the following design features are incorporated:-
G
Use a stepped chute mill inlet with an inlet trunnion scroll as per Figure 2. Avoid
the use of drum type feeders which restrict airflow.
G
Maximise the hood discharge area in order to reduce air velocities to allow dust
-drop out prior to the filter. Keep ductwork runs to the mill ventilation system as
short as possible with the minimum number of bends and horizontal runs. Flt
adequate sealing before airslides, etc., at the discharge end. Wkh closed circuit
mills, design the filter system such that filter dust is returned to the separator.
This prevents the ventilationairtlowbeingcut back to match final cement fineness.
In open circuit millingsystemsthis filter dust obviously cannot be re-classified and
will thus affect the finished cement surface area.
G
Use double diaphragms with central ventilation grids. On smaller diameter mills
(e.g. 895 kW or less) ensure that the area of the ventilation grid is not excessive,
thereby restricting volume loading. When central ventilation grids are used it is
essential to make sure that diaphragm slots are kept clean and-are adequately
sized. Ifno~ there is a risk of coarse material overfilling Chamber 1 and flooding
through the central grid into Chamber 2.
G
Low airflows through cement mills often result from the following:-
G
High restriction/blockages of ductwork to filter plants
G
High irdeaking air over mill hood, filter plant and ductwork
G
Blinded filter bags/poor cleaning, etc.
13
-
. .
.
.
. .
.
.
.

.
G

.
\
.

.
.
1
I
.
.
.
q
G
Measurement of millventilation airflows can be earned out by means of relatively simple
test work as follows:-
-
Measure all sources (where possible) of mill inlet airflow using a hot wire
anemometer.
Measure the mill exit cement temperature plus the exit air and ambient air
temperature.
Measure the exit airflow to the filter using a Pilot tube.
Carry out a heatimass balance over the mill exit airflow. If there is a large
differencebetweenthe temperature of the exit cement and the air to the filter, then
this will indicatea highlevelof inleakingair which can be calculated from a simple
heat and mass balance.
Compare the measured mill inlet airflow with that estimated using heatimass
balance techniques over the mill exit. An average of the two results will give an
approximate indicator of the actual airflow through the mill itself
Calculate the air changes per minute 170mthe measured volume loadings/free
volume above charge data.
Using this method on a UK 746 kW cement mill gave the following results:-
Measured mill iniet airflow
=
18.7 kg/rein
Estimated mill exit airflow (by heat balance) = 25.7 kglmin
Average airilow
=
22.2 kglrnin
Filter exit airflow
=
52.7 kglmin
Mill inlet airflow as a percentage of the
filter exit airflow
=
42. 1?40, i.e.
Inleaking air
=
57.9% of total filter exit airflow!
15
8. MILL POWER DRAWN. MILL CRITICAL SPEED AND MEDIA LOAD
,,
The power drawn by a mill can be estimated using various formulae such as the DAWN formula.
In order to optimise mill performance, in particular the lifiing/cmshing action in Chamber 1, the
following conditions are preferred.
High mill critical speed - typically 75-77%
Adequate reverse step liner fitted in Chamber 1
No chamber overilllingor excessivenibs which reduce the effective media density
Efficient charge in all chambers
Under these conditions, the overall power drawn per tonne media in the mill is typically 14-16
gross kW/tonne media
.
w - DOnot confbse this with the guidelines for power drawn per tonne cement as outlined
in section 9 below)
For older mills which have:-
Low critical speed i.e. typically 68-70?40
Poor chamber 1 liner which allows slippage of the charge
Chamber overfillinghnefficient media, etc.
The gross kW/tonne media maybe only 10-12 gross kW/tonne media.
Hence, it is worthwhile checking the critical speed and gross kW drawn per tonne media loaded
for each mill. It maybe possible to compensate for the low critical speed of some older mills by
optimizing media gradingdliner designs, etc, as outlined within these guidelines.
16
9.
The optimum configuration of a mill will depend upon the types and range of cements to re-
produced. For example, a mill producing high surface area (380-400 m2/kg) cements will tend
to require less power to be absorbed in its first chamber than a mill which is optimised for
producing coarse cements (280-300 m2/kg). However, if the same mill has to be used for both
cement types, a compromiseis needed. Hence the chargein the first chamber cannot be reduced
too far or the mill will tend to overfill and block when producing higher outputs at the lower
cement surface areas.
The recommended guideline for first chamber gross power consumption is as follows:-
9-11 gross kWh/tonne cement
Current BCTC mill optimisation philosophy is as follows:-
Put sufficient media/power into chamber 1 to achieve adequate crushing of
material leaving the chamber.
Avoid exceeding the above guideline first chamber power consumption unless
having to compensate for less efficient mill design ie low critical speed or worn
chamber 1 liners.
Maximise fine grinding charge in chamber 2.
Note, if the mill has to produce a wide range of cement surface areas then it is necessary to
compromise ie,
G
~~ed OPC 2238kW mill
65tph at 300 m2/kg
Chamber 1 power
=
670 kW
Chamber 1 gross power consumption = ~ = 10.3 kWh/tonne cement
65
G
R.H.C.
Same tonnage media/power in chamber 1 = 40tph at 400 m2/kg
Chamber 1 gross power consumption = ~ = 16.8 kWh/tonne cement
40
17
Hence, when producing RHC cement, the first chamber power is too high. However, if the media
is reduced in chamber 1 (or chamber 1 is shortened), there will be insufficient power available to
handle 65tph OPC at 300 m2/kg. Hence a decision has to be made on whether the mill
performance is to be optimised for Bagged OPC of R.H,C production.
Possible compromise for Bagged OPC
Chamber 1
,
9 kWh/tonne at 65tph requires 585 kW
Chamber 2 1653 kW available
25.4 kWh/tonne on Bagged OPC
41.3 kWh/tome on RHC
Any such compromise must be based upon a detailed working knowledge of the mill from
inspectionskixkdsamplingtests etc. It is known that certain milling systems can perform well on
only 8 kWh/tonne cement in chamber 1 whilst others may have to handle clinker and extenders
needing around 12 kWh/tonne.
come to a sensible compromise.
Hence, common sense and factual testwork data is required to
,,
18
10. CLOSED CIRCUIT MILL OPTIMISATION
Testwork can be carried out on closed circuit milling systems in order to establish the optimum
mill recirculating load. The aim of the tests is to minimise the overall circuit ldHw/tonne and
maximise output. In additio~ quality parameters such as cement residues are also related to
recirculating load levels and the effects of changes in residue also has to be taken into account.
Test work aimed at optimizing closed circuit milling systems is often a lengthy procedure and
results are not always conclusive. In order to cany out these tests, fill particle size analysis
results of the separator feed, fines and rejects streams should be carried out using analysers such
as the SEDIGRAPH or C.I.L.A.S. type. These allow the parameters such as separator by-pass,
Rosin-Ramler slope, etc., to be determined. When trying to establish the minimum power
consumptionfor the millingsystem the effects of varying recirculating load upon ancillary power
consumption must also be taken into account. In some cases, there maybe physical limitations
which restrict the optimum recirculating load from being achieved such as diaphragm slot
size/are~ elevator capacity limitations, etc. ,
Ideally, fi,dlparticle size analysis equipment should be available at the Works to allow routine
testing of cement psd as well as the separator performance parameters. This provides usetld data
which can be logged along with the normal routine quality/petiormance data such as:-
G
outputs
G
kW absorbed, mill and ancillaries
G
Cement surface areas, strengths, etc
G
Cement residues
G
Clinker chemistry e.g. CZS,CqSand S03
G
Cement grindability
Computer based mill modelling techniques are used within BCC Technical Services in order to
simulate mill operation. These techniques can prove very usefid for predicting the theoretical
gains to be achieved by:-
G
Optimizing the recirculating load for existing milling systems
G
Optirnising the media gradings
G
Conversion from either open circuit or closed circuitlconventional separator
operation to high efficiency separator operation
MN modelling techniques are most usefid when used in comection with normal mill/separator
testing procedures. They can help to quanti@ the benefits to be gained through plant
improvement/modifications. bother use of the model is that of mill conversion tender
evaluation. The mill model has been used to cross-check suppliers claims for increased output
by conversion to high efficiency separation operation at Cauldon, Dunbar, El Melon, etc.
19
11. THE PURCHASING OF NEW CEMENT MILL INTERNALS
Use of the various mill efficiencytargeting methods can help to formulate any cases for changing
mill internals in the short and long term. It is recommended that the current mill internals be
compared with BCI guidelines as contained in the Cement Optimisation paper. This, paper is
updated at regular intervalsto reflect current user experiences with new and existing designs, etc.
Avoid buying mill internals based upon suppliers claims.
Ml equipment suppliers have a vested
interest in sellihgworks their equipment which may not be the qptimum for a given duty. There
have been several bad examplesof new millinternals being ordered which have not produced any
improvement in millingefficiency. In certain caseq internals have been ordered which contravene
the BCI guidelines and have resulted in a reduction in milling efficiency.
Generally speaking, any renewal of mill internals should give an improvement in milling
performance as worn linersldiaphragms, etc. are replaced. Unfor@nately, suppliers often
recommend changes to
G
Mill Chamber lengths
G
Media gradings
G
Lining thicknesses
which can prove detrimental to the efficiency of a mill. Whilst a short term benefit may be
achieved by the new internals, there is a risk that the long term efficiency is adversely affected.
Hence the message is Consult the BCI guidelines and speak to BCTC Technical Centre
engineers.
These guidelines are generalised guidelines only and have to be applied with common sense in
order to arrive at cost effective solutions to optimizing mills.
..
20
APPENDIX Ill
RECENT EXPERIENCES WITH CEMENT MTLL INTERNALS
1. ALTERNATIVE LINER PLATE DESIGNS FOR CHAMBER 1
1.1 Magotteaux Duolift Liner
1.2 Voest Alpine
1.3 Lorain Liners
1.4 Step Liners
2. CHAMBER 2 LINERS
2.1 Magotteaux Classi~ng Linings
2.2 Lorain Liners
2.3 Christian Pfeiffer Classi@g Liners
3. DIAPHR4GM DESIGNS
3.1 FLS Combidan Diaphragms
3.2 Magotteaux Opticontrol Diaphragm
3.3 Pfeiffer Third Generation Diaphragm
.. 3.4 Magotteaux Airfeel Diaphragm
1. ALTERNATIVE LINER PLATE DESIGNS FOR CHAMBER 1
1.1 Ma~otteaux Duolift Liner
There have been several installations within the BC Group. This liner design has been installed
at Weardale, Cauldon, Cemento Melon and Aberthaw Works with varying results. The main
conclusions drawn are summarised below and the liner design is shown in Figure 1.1.
G
G
G
G
G
1.2
No proven advantage over normal (i.e. 60-80mm) step type liners, apart from the
maintenance advantage of having a beltless design (less spillage, bolt breakages, etc).
The DuoMIliner was removed from Weardale No. 3 cement mill following a decrease in
millpefiormance and less stable mill operation. Other factors (coarser ball charge, etc.)
also contributed to this problem.
Magotteaux subsequently fitted a step liner replacement.
There is some evidenceto suggest that the Duolifi is not suitable for mills of high critical
speed (75/0plus) and can cause higher power losses.
The Duolift liner in No. 19 raw mill at Cemento Melon was seen to give no advantage
over FLS flat place liners. WMst the Duolift lining helped to reduce the level of nibs
before the intermediate diaphragm, the lining (plus chamber 2 classifyhg liner) caused a
reduction in millinternal diameter. There was no evidence of any improvement in output
over the similar No. 18 raw mill with its older FLS internals.
AberthawWorks No. 1-3mills - improvements to the mill ventilation and the installation
of Duolii linershave generallyincreasedproduction. However, the first chamber absorbs
a relativelyhighpower (16 kWh/t) to achieve a satisfactory fineness. This could limit the
maximum mill potential output.
Voest A]Dine - Activator plate Lining
This is a development of the VA grooved liner plates which were unsuitable due to high charge
slippage/excessive nibs problems. Test work on the liners used at Fujairah cement showed that
the liner gave poor Iifiing action. The design has been modified to include raised sections or
activators to increase lift.
.
Since experience with this design in cement mills is limited, the following experiences with raw
milling have been included:-
This design has been installedinthe limestone mill at Rawang Works and the raw mill at Kanthan
Works.
At Rawang Works, whilst improved power consumption is claimed, any improvement in
petiormance must be judged in the light of the poor state of the liners which were replaced. An
1
area of concern is the reduced mill power tid whether or not this represents a real power saving.
The inference is that the design could still allow charge slippage and therebylimit the maximum
usefid power drawn by a mill.
The MFL liners installedin Kanthan Works raw mill gave reasonable performance, but tended to
be prone to build-ups when handlingwet materials, When wet rriaterialscoats up the gap ,between
adjacent lifters, the Iiner becomes inefficient., For contractual reasons these liners ,were replaced
by conventional s when the mill was uprated by Roll Press addition in 1993-4.
The Activator plate design offered by Christian Pfeiffer is also available for cement mills.
However, the raised step on these plates is only 40mm and the lining does not appear to promote
good crushingaction in Chamber 1. Hence, this design cannot be recommended especially when
used with mills of low critical speed (68-720/0).Figure 3 shows de@ls of the MFL Activator plate
designs used in Malaysia.
1.3 Lorain Liners
This bar type lifter is common in the USA but less common in the UK. The liner seems to
achieve good charge separation/liRhg action in the mill. Hence, we would not advocate
replacement of this liner design by newer designs. The high lift characteristics may well be the
most appropriate lining type for mills such as those commonly found in the ,US~ i.e. with:
high volume loadings (38-40%)
,
,
high recirculating loads (300%)
1
See Figure 4 for details of the Lorain liner.
,,
.,
Beware of usiig a high lifl liner design, such as the Lorain, with high critical speed mills, i.e. 78-
80V0critical speed.
1.4 SteD Liners
,.
Avoid the use of the Magotteaux low step liners (step 35-45mm) such as installed in Malaysia
(Kanthan No. 3 and Rawang No. 5 cement mills). These give poor crushing actio~ even when
used with a coarse first chamber ball grading.
Some Magotteaux step liners have medium step distances, i.e, 50-60mm which may work well
when new and in millswith higher critical speeds (75/0plus). However, when the step wems to
around 40mm it is less effective. Hence, the prefemed step distance is 60-80mm.
,,
Avoid using a very deep step distance (80mm) if the mill Mo critical speed is 76?40or above.
2
FIGURE 3 COMPARISON BETWEEN
ACTIVATING LINER PLATES
F-l
3$9
I
1-
C)RIGINAL-tiFL ACTIVATING LINER
i
,
___.. -
.
.
FIGURE 4 LORAIN LINER PLATES FOR
CHAMBER 1 AND CHAMBER 2
,..
CHAIIBER 1
cHAllBER 2
1.4.1 Beltless versus Bolted Step Liners
Followinginstallationof a bokless first chamber step liner in Plymstock No. 2 cement mill, BCC
were concerned over the efficiency of thk design compared with the conventional bolted liner
design.
Comparative tests were carried out in Magotteauxs pilot plant and these showed the following:-
For volume loads below 32%, the beltless liner absorbs more power than the
bolted liner.
For volume loads of 29%, the liftingangle (see sketch) is 4% higher for the bolted
liner.
Overall the difference in absorbed power is 1.7% higher for the beltless liner.
However, Magotteaux concluded that in operation, there is little actual difference between the
two designs. BCTC view is as follows:- -
The beltless is less efficient in
design.
.
The beltless design tends to
accumulated in chamber 1.
actual operation when compared with the bolted
produce a tailing of oversized nibs which can
-The beltless liner should not be used with slower speed mills (ie, 68-72%critical
speed) especiallyif such mills have a small first chamber and have to grind lower
S.S.A. cements or hard/oversized clinker.
At Cemento Melon, No. 16 cement mill is fitted with the Magotteaux beltless liner design. The
mill speed was increased from 70 to 78.6/0,following gearbox repairs. This allowed the mill
output to be increased from 23 to 24tph at a similar power consumption. The higher speed
allowed chamber 1 to be shortened and at this higher speed the beltless liner gave reasonable
crushingaction. Whilst the beltless step liner is considered to be inferior to the bolted design its
performance is acceptable at higher critical speeds (76-78%).
Figure 5 shows the comparative design of Magotteaux Bolted and Beltless lineqs.
5
o z
L
L
l
2
j
n o z z i
2. CHAMBER 2 LINERS
2.1 Mwzotteaux Classifying Linings
Mention has already been made of how classiijing linings can limit output on mills which are
volume loading limited. This is due to the 10SSof internal diameter and reduction in critical speed.
Magotteaux are aware of BCTC views on this and have developed a thinner classi&ing lining
which was recently fitted to the new larger mill shells of Cookstown 1 and 2 mills. The liner
appears to give good classi$ing action Cookstown but was also reported as giving some reverse
classificationelsewhere. Figure 6.1 shows a comparison between classifying liners and Dragpeb
liners. Figure 6.2 shows the Cookstown Lher design.
2.2 Lorain Liners
These liners are suitablefor second chambers of USA mills where high material throughputs and
high volume loadings require a more aggressive action to the charge.
2.3 Christian Pfeiffer Classifvinp Liners
The designof cktssi&ingliner plate f~tures a very deep (190mm) wedge liner plate coupled with
a flat plate with grooved profile and activator section at one edge. It is fitted to Ravena No.4
cement mill.
Whilst the liner appears to give reasonable ball classification this design cannot be recommended
since it limits the mill internal diameter and runs the risk of classif@g nibs to the outlet of
chamber 2. In the USA mill application, great care has to be taken to avoid nibs and maintain
even 6mm gaps in the intermediate diaphragm to avoid nibs.
7
E E
w
g
>
-
4
\
a
5
3
n
v
5
3 m
:
A
&
m
.
-
I
A
L
-
o
m
*
0
S
3. DIAPHRAGM DESIGNS
3.1 FLS Combidan Diaohra~m
This diaphragm design features large armour castings which have gaps between them to allow
material flow. The diaphragm slots consist of a screen with 6 or 8mm pefiorations.
This designis installedinthe new cement millsat Aberthaw and Kanthan Works. Problems arose
with the Abefihaw mill which are associated with the lack of any means of regulating the flow
control through the diaphragm. The diaphragm was modified by cutting away four of the nine
lifter plates and converting these to an adjustable plate.
This modification proved reasonably successful and allowed greater material retention in
Chamber 1. See Figure 7.1 and 7.2 for details.
However, the Combidandiaphragmis not seen to be as good as either the Pfeiffer or Magotteaux
Opticontrol design. hy fbture Combidan installations must be provided with the flow control
modifications as per the Aberthaw mill.
The Combidan diaphragmalso relies on a good first chamber liner and ball charge to avoid nibs.
Clinker nibs can quickly block the perforated screens and there is little self-cleaning action since
the media is not normally in contact with these screens. The vent grid slot sizes tend to be the
same size as the screens which makes them prone to blinding.
The FLS outlet diaphragm design used at Kanthan and Aberthaw Works also had a tendency to
strip material out of the second chamber. Hence, it was necessmy to block the outer row of slots
to improve material retention.
Generally we note that when using the combidan diaphragm design, FLS/Fuller tend to use a
comparativelylong first chamber in order to avoid potential nibs blockage problems. This often
risks achievingoptimum efficiency from the mill especially when grinding higher S.S.A. cements
as witnessed at Aberthaw.
3.2 Mmzotteaux opticontrol DiaDhra~m
The Slegten diaphragm has been refined to use circutierential slots in a pattern similar to the
Pfeiffer third generation diaphragm. This improves the self-cleaning action of the slots. A
diaphragm of this design was installed in Cemento Melon No. 21 cement mill in 1992 and
appeared to perform well after modifications were made to reduce the size of the central
ventilation grid to allowfor charge expansion. Some problems have arisen with second chamber
fine media back spillage. Care has to be taken when selecting diaphragms for high volume
loadings or when using fine media which can expand. See Figures 8.1 and 8.2 for details of the
diaphragm slots whilst Figure 9 gives details of the adjustable scoops used for materials flow
control.
During 1994, there were problems with the Cemento Melon No. 21 diaphragm support structure.
Severe cractig of the Ih.me segments has occurred and this has been phrtially blamed upon small
meda backspillage from chamber 2. We, therefore, have reservations about recommending ttis
10
..
FIGURE 7.1- FLS COMBIDAN DIAPHRAGM
..
G
9
.
8
Centrepiece
1
Scf eeni f g pbte
G
Residue space
~ 1 /~
Liftem
I
I
%=. .
G* . ~%. . GX.
G
# 6-. .. .: -. ..- -.
>
..- .
R
. . . . .
. .
0 IF - Ull
.
.
FIGURE 7.2 MODIFICATION TO COMBIDAN
DIAPHRAGM AT ABERTHAW WORKS
MODIFICATION - REPLACE 4 OUT OF 9 LIFTERS [FIXEDI
BY MOVEABLE LIFTERS
.,.
I
I
[/////////
I I I
i
/
INVERTED PLAN
.
FIGURE 8 . 1 - MAGOTTEAUX - OPTICONTROL DIAPHRAGM
.
.
.
L ~-
.

&_ -

=
.,
E

.
_

..-
~
.

3k=

-
._
. ~
-

SELF
CLEANING
SLOT
PROFILE
FIGURE 8.2
(
A) PROBLEM - MATERIAL FLOW THROUGH VENT GRID
SOLUTION
>
,-
B) CLOSURE PLATE - INSIDE VENT GRID
.. ,
FIGURE 9- FLOW CONTROL SCOOPS
Thescoopscaneasilybe adjusted by
turningthemontheiraxes,
PositionA will not allow the scoop to
pick up material,thusthe diaphragm
fillsup with material,
Position B shows the scoop partially
open,
If set at positionC the scoop isfully
opened and will pick up the maxi-
mum amount of material. If several
scoops are in this position, the level
of material in the diaphragm will
decrease,
The differentsettingsof the scoopswill
not affectthe mill output:
Q scoop output= Susetd area X Hactive height
= constant
1
A
m
I!!#l
U
.-, ,*
H
.<,
I
diaphragmfor larger mills (ie above 4.OMdiameter) especially thnse with fine media and charge
expansion characteristics.
,!
,
Magotteaux have modified their design and their latest Opticontrol diaphragm design will be
installed in the new Kanthan CM5 and Pasir Gudang cement mills. Before ordering ,my new
Opticontrol diaphragms it is important to check that the latest improvements are included.
Before ordering this diaphragm always check the required volume loadings and whether or not
the mill has a tendency for charge expansion. For example, using the DAWN spreadsheet, we can
estimate the followingmaximumventilation@d sizeswould be n@ed for Cauldon No. 4 cement
mill with its Magotteaux intermediate and outlet diaphragms.
Mill Chamber 2 Height Above Maximum Vent
kW Volume Load Charge Grid Diameter
/0 (m) (mm)
,,
1418 38 1.726 552
1406 37 1.748 596
1394 36 1.771 642
1368 34.15 1.815 - 749
(Present) (Actual)
Allowances for charge expansion when using fine media (25-15mm) in chamber 2.
Typical
=
2-4% higher v~lume load
Worst cases
=
7% higher volume load
3.3 Pfeiffer Third Generation Dia~hragm
This diaphragmhas given good results in USA and Malaysia. The flow control is easy to adjust.,
There is only one area of concern and that relates to the diaphragm segments. As with earlier
Pfeiffer diaphragms, where the segments meet there are areas where any movement of the
segments can allow gaps to occur. Whhout any backing plates behind these points, the gaps can
allow coarser materialshibs to enter chamber 2.
Ailer discussions between BCTC Chiistian Pfeiffer have a~eed to modiijr their diaphragm!
Details of the diaphragndmodifications are shown in Figures 10.1 and 10.2.
Whh this modificatio~ the Pfeiffer design is believed to be very suitable for closed circuit milk,
especially when operating with high circulating loads. Care has to b,etaken with open circuit
mil.1$since many of the modem lifter diaphragm designs can tend to strip material out of the first
chamber.
,
16
FIGURE 10.1- PFEIFFER INTERMEDIATE DIAPHRAGM
Self-Cleaning Feature
The front side of the slotted
plates have continuous
concentric grooves which
coincide with the slots. Balls
are guided by these grooves.
Material which would
otherwise plug the slots is
forced through. This self-
cleaning effect assures the free
area for material flow.
I
I
I
I
I
I
1
I
1
8
I
1
#
i
t
I
I
1
1
I
1,
I
1
I
I
t
I
1
I
FIGURE 1092 PFIEFFER DIAPHRAGM MODIFICATION
SKETCH OF
ADDITIONAL
DIAPHRAGM SUBSTRUCTURE
BETWEEN RINGS OF
TO CLOSE UP GAP
SLOTTED PLATES
Please note that the earlier first and second generation Pfeiffer diaphragms tended to be less
suitable due to the lack of flow control which tended to result in underfilled chambers. When
ordering replacement Pfeiffer diaphragms ensure that:-
(a) Design is third generation type.
(b) Specifi maximum volume loading required in both chambers. The design is
suitable for volume loadings up to 40/0.
3.4 Ma~otteaux Airfeel Dia~hragm
A recent Magotteaux development,the firfeel diaphragmhas yet to be evaluated, but is probably
not suitablefor open circuit applications. We have concerns over the flow control action of this
design, which is mill ventilation rate dependent.
19
APPENDIX ITB
NOTES ON THE PERFORMANCE OF USA CEMENT MILLS
We are canying out ongoing surveys of the USA plants which have shown the following:-
s
When testing the cement millsgenerallygive good efficiencyfigures when compared with
the BC grindability test.
G
whilst there is no singIecommon design basis and there are many different combinations
of mill internals, separators, etc., some aspects of the mills are similar i.e.:-
Common use of grinding aids.
Mills operate at high circulating loads, i.e. 300V0.
Conventional separators are of large size in order to cope with high circulating
load.
Separator efficiency is good considering ageldesign.
Use of high liil liners such as the Lorain type.
G
As an example, at Tulsa Works, the millswere tested and were found to be operating with
a mill efficiency of around 125Abased upon the Blue Circle grindability test data. This
efficiencylevel was the same as our target value for closed circuit mills. However, it was
apparent that the mill internal conditions were very poor with poor linings in Chamber 1
and excessive levels of nibs. The mills were producing around 34.5 stph (31.3 tph) of
Type 1cement at 3500 Blaine and grinding aids were used. Hence, although it was to be
expected that the mill pefiormance could be improved, it was difficult to predict to what
level.
Themills were subsequently modified by installation of the following:-
Pfeiffer intermediate diaphragms with high transport rates
Lorain lifters (simple bar lifier type desi~ non classi~ng)
Ball charge
Separate mill ventilation installed plus minor circuit modifications
Mer optimisation, the mills achieve 36-38 stph output, which implies a mill efficiency of 130-
138XO.
The mill output was higher than expected due to the practice of running over the rated motor load
(these millsare typically run at 1240 kW compared with nominal mill motor ratingof1119 kw)
with high media volume loads.
NOTE:
When ordering new diaphragms for USA mills, etc., ,wherehigh volume loadings are used (36-
I
~
40??) - always speci@this to the diaphragmsupplier so that their design can be suitably modified.
The Pfieffer diaphragm has various sizes of inner diaphragm segments available to suit different
volume loadings.
Hence, a combination of the following factors:-
,,
Use of grinding aids.
Mill designfatures as summarisedabove appears to permit an increase in the BC
target mill efficiency to around 135% for closed circuit mills.
1
,
,.
2
SECTION C
CEMENT MILL PERFORMANCE TARGETING
SECTION C
CEMENT MILL PERFORMANCE TARGETING
CONTENTS
1. TARGET PERFORMANCE FOR CEMENT MILLS
1.1 Mill Power Drawn
1.2 Mill Running Time
1.3 Mill Charge Levels
2. CEMENT MILLING TARGET OUTPUT AND CEMENT RESIDUE
FOR A GIVEN SURFACE AREA
2.1
2.2
2.3
2.4
2.5
2.6
2.7
Target Performance for an Open Circuit Cement Mill - Example
Target Performances for a Closed Circuit Cement Mill
Open to Close Circuit Mill Conversion
Mill Surface Production Factor Method for Targeting Closed Circuit
Cement Milling Performances
Mill Net to Gross Power Ratio
Mill SurfaceProduction Factor - Target Values for different closed circuit
milling systems
Summa~ of Closed Circuit Mill Pefiormance Targeting
3. HIGH EFFICIENCY SEPARATOR CONVERSION EXAMPLE
3.1 Results Obtained from Conversion
3.2 Effect of Changes on Mill Output
4. BENCHMARKING CEMENT MILL PERFORMANCE
4.1 Methods Used for Benchmarking Cement Mill Performance
4.2 The 1994 Benchmarking Results for Cement Milling
4.3 How To Improve Cement MNing Performance
4.4 How to Improve Benchmarking Methodology
4.5 Conclusions
1. TARGET PERFORMANCE FOR CEMENT MILLS
It is important to set a target performance level for any milling system in order to identifi its
potential performance. This paper sets out some guidelines on how this can be achieved. The
factors influencing the overall annual average performance of a mill are considered individually.
1.1 Mill Power Drawn
With adequate maintenance, it shouldbe possibleto maintaina mill at its fill absorbed power over
long periods. However, in practice, there can be limitations to achieving this, such as:-
Peak demandperiods resulting in mills having to run continuously without maintenance,
e.g. charge replenishment, diaphragm cleaning, liner replacement etc.
Resource limitations, e.g. lack of Iabour, limitations on capital/revenue expenditure on
replacement mill internals etc.
Other plant limitations, e.g. mill gearbox, product transportation, mill circuit design
limitationsetc. Acombinationof these factors could mean that the installed motor power
of any mill may not be filly utilised.
For example, there is no merit in charging a mill to draw till power if (a) the additional
millthroughput cannot be handled by, say, the cement transportation system or if (b) the
period between major gearbox repairs is reduced to unacceptable levels. (Note the above
are real limitations currently applying to mills within the Group).
For target purposes, the following figures can be considered.
Target annual average power =97% of installed motor power
Typical Best annual average = 100?40installed power
Many mills,especiallyin the USA consistently run above the design motor load! Check on type
of motor, type and design of gearbox, motor cooling, start up loading etc before giving any
consideration to this method of operation. The long term effects upon motor/gearbox life etc
have to be considered.
The maximum power input to a ball mill will depend upon the efficiency of the motor and
gearbox.
A typical breakdown of these is as follows:-
Mill motor
Mill motor efficiency
Maximum mill motor input power
Gearbox efllciency
Net power to mill
Overall power losses
Net/gross power ratio
~
ie. 6/0power losses,
=
2238kW
=
95?40
,
=
2356kW
=
99??
=
2238 X0.99
=
2216kW
=
2356-2216 = 140kW
,2216
=0.94
=
2356
-.
Hence, the maximum input power to the mill gearbox and the mill motor efficiency must be
checked before decidingthe maximum input power. At the same time, the mechanical condition
of the gearbox must be considered
1.2 Mill Running Time
G This is subject to considerations such as maximum demand and the other plant equipment
which has to be run. Maximumdemand considerations may typically make it desirable to
stop the mill for, say, 4-5 hours per day for periods of time.
G
Atypical target annual average mill running time figure is 85% assuming no major M.D.
limitations.
,.
G Best annual average mill running times are targeted at 90Y0.
G
Do not confhse the difference between peak and average running time, i.e. during peak
demand periods, many millsmay exceed 90% ruining time for several weeks. However,
even when there is consistent high demand all year round, an annual average figure of
90% is a high target to achieve.
2
1.3 Mill Chmwe Levels
G
As the charge level in a mill is increased, the power drawn per additional tonne of media
loaded reduces since the centerline of the media gets closer to the centerline of the mill.
G
In addition to the limitations of mill motor and gearbox sizing, there may be certain
internal limitations which limit the maximum charge level, i.e.
Size and design of mill inletioutlet trunnions.
Lining plate type/design and lining thickness.
Type and design of diaphragms.
Do diaphragmshave central ventilation grids or solid centres? A large central ventilation
grid may limit the maximumvolume loading - since above a certain level, the media and
material maybe above the level of the vent grid. In this condition, there is a risk of coarse
material entering chamber2 via the vent grid instead of being held back by the diaphragm
slots.
Certain types of smooth lining plates allow slippage of the charge (e.g. Voest-Alpine
grooved liner plates) which can reduce the maximum usefi.dpower drawn antior result
in high volume loads to draw fill motor power!
Relativelythick liner plates such as classivlng linings, can result in a loss of mill internal
diameter. This in turn can result in higher volume loadings to draw fill motor power.
Check the design of any mill trunnion liners. A scroll should be fitted to help transport
material (and any media thrown back) into the mill.
Expansion of the mill charge caused by accumulation of unground material or by high
throughputifine media can also limit maximumcharging levels. Always check the volume.
loading with the mill run out and compare this figure with the levels taken afier a crash
stop (see example in i4ppendix VA of Mill Testing paper A).
G
Hence, the above factors need to be taken into account before considering the maximum
volume loading/power drawn for any mill. Always be aware of mills which have been
over-motored such as the 1600 HP (1194 kW) mills which have 1200 HP (895 kw)
shells and a usefid power drawn limit of around 1000 kW. Just because the mill motor
plaque gives a figure of X kW installed power, this does not inevitably mean that 100%
of X kW can be usefully absorbed during operation! If there are any limitations which
limit the maximumpower drawn (e.g. liner/diaphragm type etc) - then consider ways and
means of overcoming these by, say, the use of alternative mill internals. In the case of
Cookstown Works two 1600HP cement mills, the mill shells were replaced by larger
diameter shells in order to allow the maximum motor power to be used.
3
2. CEMENT MILLING
TARGET OUTPUT AND CEMENT RESIDUE FOR A GIVEN SURFACE AREA
G
The cement particle size distribution for a given surface area can be characterised by the
RosinRamler slope (taken from the slope of the psd plot on Rosin Ramler graph paper).
Efficient operation of a cement milling system is usually indicated by a steep R-R slope,
and lower cement 45 micron residue for a given surface area.
Table 1 shows how the cement residue and sutiace area can be used to predict the R-R
slope. Table I also shows typical cement residues and R-R slopes for open and closed
circuit cement mills.
This comparison can give an indication of the relative efficiency of a milling system.
Please note that the comparison applies to neat (i.e. clinker plus gypsum only) cements.
When @ding extended (e.g. limestonefilledor pozzolanic) cements, the relationship will
not be the same as shown in Table 1.
Ifa cement mill produces a range of cement types, examine the SSA/R-R slope data for
each cement type. This can help identifi those milling systems with (a) reasonably
efficient performance at low surface areas, (b) poor efficiency at high surface area
cements. This is typical of the 6000 HP cement millswith FLS-CV separators which have
poor performance at high surface areas (380-400 m2/kg plus) due to poor separator
performance.
G
Cany out a cement gnndability test to estimate the theoretical performance of a mill.
The BCI gnndability test provides ve~ usefid data which needs carefil interpretation
before it can be used to target a mills petiormance. However, the results are open to a
wide degree of interpretation and care has to be exercised when applying the data to
different milling circuits.
The following example has been chosen to demonstrate how an open circuit mill
performance can be assessed and targets established.
5
2.1 Tar~et Performance for an O~en Circuit Cement Mill - Examde
Using the BCI cement grindability data, together with the actual mill operating data, the
performance of an open circuit mill can be established thus:-
Mill output
=
Mill power drawn
=
Cement sufiace area =
Cement 45 micron residue =
Rosin Ramler slope
=
BCI cement grindability =
From grindability curve -
Revs/lb at 380 m2/kg =
12.3 tph
640 kW gross
380 m2/kg
14% retained
0.97
109% (at 300 mz/kg)
105 revdlb
Hence, Theoretical Gross power consumption
Revs/lb x 0.381
=
kwhh
0.9
=
44.5 kWh/t (Gross)
Note: In this method the power losses for the mill motor and gearbox are
assumed to be 10O/O. Hence a netigross factor of 0.9 is used above.
Actual mill Gross power consumption
640
=
=52.0 kWh/t (Gross)
12.3
Mill efficiency
100 x Theoretical gross kWh/t
=
Actual gross kWh/t
100 x 44.5
=
52
..
This compares with the following Target open circuit milling eficiency:-
0/0 Mill Efflciencv
Open circuit mill without grinding aid
(See notes in section 2.2 for fhtiher details)
Hence, the performance of this mill at 85.6?40efficiency
115%0
is very poor!
Using a target figureof115?40 milling efficiency for this open circuit mill,
Target Gross kWh/t (mill only)
If the millis charged
output would be:-
44.5
= 38.7 kwhh
1.15
to its design motor load of 709 kW, the potential maximum
709
= 18.3 tDh (Maximum)
38.7
For OPC cement, an efficient open circuit mill could achieve 7% residue at 380
m3/kgsurface area, i.e. half of the current cement residue.
Target residue = 7% at 380 m2/kgat 1.1 R-R slope
Action reauired
G
Carry out fill detailed inspection of mill internals.
G
Cany out axial sampling tests and use datahesults to optimise mill media
grading.
G
Check mill ventilation and mill cooling systems.
G
Draw up action plan and establish improvement programme, i.e.
modernisationof mill liners, diaphragms, feed system, installation of new
mill ventilation bag filters etc. Use BCTC cement mill optimisation
guidelines and per Section B.
G
Repair mill shell and overhaul
motor load to be drawn.
mill gearbox in order to allow the fill mill
7
2.1.1 Limitations when targeting Open Circuit Cement Mill Performance
The BCI Target open circuit mill efficiency is not a fixed parameter for the following reasons:-
G
The use of grinding aid can increase the target efficiency figure by 5-1O%.
G
It is possible to achieve higher efficiency levels in certain mills with the following
characteristics: -
Under
Clinker feed of small size and relatively soft to crush
Mill optimised for fine grinding with the minimum power consumption required
for first chamber crushing.
,
these circumstances, open circuit mill efficiencies of 120%have been achieved.
G
The target efficiency is dependant upon the mill design and the finished cement surface
area. Hence, the production of higher surface area cements usually results in a lower
target efficiency figure for the following reasons:-
More cushioningof charge, coating problems due to finer material within the mill.
Remember the shape of the grindability curve and the reasons why the slope of
the cuwes is less at higher surface areas!
Reduced output and limitations to mill cooling airflow/water injectioncan cause
problems with milltemperature control at higher surface areas. This worsens the
potential coating kndcushioning problems.
The target mill efficiency figure could thus be modified to allow for higher surface area
...
cements as follows:-
-
115 300
110 325,
105 350
100 375
8
The best open circuit mill data for BCC UK works shows a maximum milling efficiency of
120?40for 350 m2/kg. Hence, in the example given in Section 2.1, whilst the mill could achieve
a target efficiencyof 1150A,it would be safer to assume a figure of 10041giving the following:-
Target Gross kWh/t (mill only)
_44J=

44.5
1.00
Mill output at the design 709 kW
709
= = 15.9 tph
44.5
This is still 29.3% higher than the current mill output of 12.3 tph.
2.2 Target Performance for a Closed Circuit Cement Mill
Here agai~ a simplistic approach which can be used is to take
milling efficiencies.
the following standard target
SYN!an
Target AMill Efficiency
Closed Circuit Mill
- no gfinding aid 125%
Closed Circuit Mill
- with grinding aid 135%
Urdiortumtely,there are limitationswith this approach and these can be shown up by the following
cases: -
G
Closed circuit mills with older separators or undersized separators - these mills tend to
suffer a deterioration in millingefficiency at higher surface areas. The problem can often
be compounded by mill coolig and coating problems. Examples of this are the 6000 HP
FLS cement mills with poor C.V. type separators and with poor mill cooling and
ventilation facilities.
G
U.S.A mills or modem mills designedwith large efficient separators, high mill circulating
loads, etc. Under these circumstances the target mill efficiency will tend to increase with
increasingcement surface area. Allowancecan be made for this within the formula for net
power consumptio~ i.e. using the FLS formula with the grindability test data for net
power consumption:-
Open circuit Net power consumption = 34.0 kWh/tonne
Cement sufiace area
=
345 m21kg
Equivalent power consumption for closed circuit milling with a high efficiency separator
=
Where factor
=
=
Hence, closed circuit power consumption
34.0 x factor
1.2- (0.001 x m2/kg)
1.2- (0.001x 345)
0.855
0.855 X 34.0
29.1 kWh/tonne net
10
2.3 O~en to Closed Circuit Mill Conversion
The above formula is useful in so far as it demonstrates the benefits of Open to Closed circuit
conversion. The formula shows that greater mill power ,consumption savings are achievable at
higher cement surface areas. The above formula demonstrates the known facts that:-
G
G
Open to Closed circuit mill conversion will not produce significant power consumption
savings at low surface areas (unless the conversion also improves milling efficiency by
optimizing the mill internals or coolinghentilation aspects).
Open to Closed circuit mill conversion is most beneficial when producing high sufiace
area cements. Part of the benefit can arise from the BktineBonus in which the same 28
day strengths can be achieved at a lower surface area due to the lower cement residues.
Care has to be taken to ensure that any loss of early strengths does not cause problems
or limit the Blaine Bonus.
AnyMill power savingsmust be offset by higher ancilkq power consumption costs which
can often result in higher overall milling energy costs, i.e.
Ancillarv kWh/tonne cement
Conventional separator
Mechanical conveying of
cement 4-5

High Efficiency separator

with pneumatic conveying 8-10
The claims for outputipower consumption benefits must always be tested. Appendix I,
shows an example in which a 16.2V0gain in output was claimed from converting a mill
with a mediumefficiencyseparator to a high efficiency type. However, when the benefits
of the conversion are examined more closely, it was found that the separator conversion
gave only 3.6?40 of the overall output benefit and the major gains were due to optimizing
the mill charge and media grading!
11
o m
l
m
T
I
I
.
.
W
J

*
\
m
0 % @
I
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
I
A
o
-
l
!
!
!
!
!
!
!
!
!
!
!
!
o c
o
o C
9
o 0 N
o

m 0 C
9
0
I
I
1
I
l
=
I
I
I
o m
3 ?
j
n
I
I
m o
I
I
2.4 Mill Surface Production Factor Method for Tarzetin~ Closed
Circuit Cement Milling Performance
BCTC currently use a different approach to targeting closed circuit cement mills with efficient
separators. This method uses two approaches, as follows:-
G
Fromthe cement gtindalility cuwe, estimatethe current millingefficiency using the simple
approach shown in the open circuit millexample. Compare the milling efficiency with the
above targets of 125% (no grinding aid) and 135A(with grinding aid). In thk method,
the mill power losses are always fixed at 10?4o, giving a net/gross power ratio of 0.9.
,s
Estimate the Mill Surface Production Factor (M.S.P.F.) for the milling system. The
M. S.P.F. is a measure of how efficient a closed mill is when compared with the BCI
gnndability cuwe.
Whh reference to Figure 1, the BCI grindability curve, it can be seen that the curve is usually a
straight lineUPto approximately200 m2/kg. Extendingthe same slope of this line up to the target
cement surface area would give a graph of net power consumption versus surface area. If the
ACTUAL net power consumption of the mill fits on this extended line, then the Mill Surface
production factor is 1.0.
G
When the perfiorrnanceof several U.S.A. cement mills was compared, the following was
noted
A plot of actual net mill only kWh/tonne versus surface area on the grindability
curve gave many points above the line for U.S.A. cement mills.
.
The slope of the line obtained through the actual operating data was typically 5-
10% steeper than the linear part of the gnndability curve, i.e.
M. S.P.F. = 1.05 tol.lo
The millsexaminedall used grindingaids and operated with high circulating loads

with large efficient separators.

G
Examination of some conventional UK closed circuit milling installations with
I@h/mediumefficiencyseparators, low circulatingloads(150Yo) and no grinding aid gave
actual operating points some 5- 10%0 below the slope of the linear part of the grindability
cuxve, i.e.
.
M. S.P.F. = 0.90-0.95
13
2.5 Mill Net to Gross Power Ratio
In this method, the ACTUAL net power of the millis used rather,than the fixed factor of 0.9 used
for B.C.I. millingefficiency. The actual net to gross power ratio will depend upon the type of mill
drive system and is typically as follows:-
Mlll Drive ~
Net/Gross Power Ratio
Efficient Central Drive Mill 0.94-0.95
Efficient Girth Gear Drive Mill 0.93
Inefficient/older Gh-thGear
Drive Mill
0;90
For fimther details please refer to the additional notes in Section 3.1 of part A (Mill Testing)
2.6 Mill Surface Production Factor Tawet Values for
Different Closed Circuit Millin~ Svstems
From the above study, the following Target Mill Surface Production factors have been
established.
MILL SYSTEM TARGET MILL SURFACE
PRODUCTION FACTOR
Conventional Closed Circuit Mdling System
Circulating load 100- 200%
No grinding aid 0.9-0.95
Conventional Separator
Medium/High Efficiency Separator
No grinding aid 0.95-1.05
Circulating load 100- 200%
MediudHigh Efficiency Large Separator
High circulating load (300 - 400%)
..
Grinding aid 1.05-1.10
Mill internals optimised for above method of
operation with high material transpofiation rate
BCTC have developed computer spreadsheets for use with the above mill targeting methods.
Care has to be taken when establishing targets and fin-ther methods are also available to assist in
this process, e.g. the B.C.I. closed circuit mill and the mill media models.
14
2.7 Summarv of Closed Circuit lt4ill Performance Tar~eting
Hence, in summary, BCI recommend that ~ methods for targeting mill performance levels
should be used for closed circuit milling systems, i,e.
G
BCI Milling Efficiency
s
Mill Surface Production Factor.
The target efficiency should then be established by comparing both sets of results and making a
balanced judgement.
All targets for-milling performance must be realistic and achievable. Hence, the targeting data
cannot be considered alone and supporting data from axial sampling tests and mill inspections is
also essential. Avoid setting too high a target performance if the evidence from axial sampling
tests and inspections does not reveal any obvious areas of inefficiency. Bear in mind that mill
internals have finite lives before wear/tear affect their efficiency. There is no point setting the
theoretical maximum efficiency target now if it is not planned to replace the mill internals for
several years.
15
..
3. HIGH ElWIClENCl SEPARATOR
CONVERSION EXAMPLE
3.1 Results Obtained From Conversion
Before
Afler
0/0Increase
in @tDUt
Output (tph) 126.3
146.8
16.2%
Surface Area 391
382
M2~g
KW mill motor 3900
4230 ~
KwH/t
30.9 28.8
Separator
bypass %
20 10
Rejects surface
area m21kg
115 86
I
Rosin-Rammler
slope 1.15 1.17
.
16
3.2 Effect Of Changes On Mill Outrmt
@
Increasing mill charge
Uprate pneumatic conveying system
Repair mill gearbox
Mill motor load increases
Predicted output
Output gain
s
Immove mill efllciency
By use of coarser media
Mill efficiency
Output gain
3900-4230kW
= 126.3 X = =
137 tph
3900
=
8.5V0
n Chamber 1 and finer media in Chamber 2.
123- 128?40
=
4.1%0
TOTAL DUE TO MILLING
IMPROVEMENTS = 12.6A
A,
G
HiizhEfflciencv Separator
Reduced surface area of cement (Blaine Bonus) for same 28-day strength.
Predicted output = 126.3 ~5
(381)
= 130.8 tph
I
OUTPUT GAIN = 3.6%
I
17
4 BENCHMARKING CEMENT MILL PERFORMANCE
Blue Circle Tectilcal Centre carry out an annual review of performance of the cement mills
operating within the Blue Circle group and its associated companies. For this exercise the Blue
Circle conventional mill efficiency method is used together with a simple formula for the
grindability curve. The results of this exercise and the conclusions drawn are presented in this
section.
4.1 Methods Used For Benchmarking Cement Mill Performance
,The first attempt at benchmarking of cement mills (i.e. primary benchmarking) that was carried
out was a simple comparison of specific power usage per tonne of cement for the whole milling
system. This approach could give a fake impressionas to the milling efficiency as no account was
made of the following factors:-
G
Type of milling system (open/closed)
G
Cement grindability
G
Target surface area
G
Use of grinding aid
In order to make a more realistic comparison between individual mills, a more comprehensive
method of secondq benchmarkingwas used to assess the 1994 performance data. This method
takes into account the following factors.
4.1.1 Mill Absorbed and Installed Power
To achieve m@mum output from the mill, it is essential to rpn the mill motor at its maximum
absorbed power. Practically, this is not always possible but for the purpose of benchmarking this
should be assumed to set the production target.
It should be noted that certain B.C. Cement Mills (e.g. Plymstock CM2, Cauldon CM4) cannot
draw the maximum motor load due to their internal design i.e classifying liners reduce internal
diameter.
Other B.C. cement mills (e.g. Tul~ AberthawCM1O)tend to run with an absorbed power above
the installed power. Remember that the motor input power may be 5-6% higher than the rated
output power.
4.1.2 Surface Area and Grindability
Millingsystems throughout the group have both significantlydifferent feeds, both in terms of feed
mixture and grindability, and the cement from these systems has a wide variation in product
specific surface area. To take account of these variations and to allow an appropriate target to
be set, B.C. cement grindability tests are earned out on the clinker, gypsum and additives used.
BCTC have built up historical data concerningindividual works ,cement grindabilitydata. Recent
data is lacking for some plants. Hence the historical data has been used with the appropriate
correction factors for any changes in S03 levels, additive dosage rates etc.
18
Prediction of the theoretical power consumption is based upon the grindability curve (see typical
example in Figure 1) which relates the specific surface area of the cement to Revs/lb of the test
mill. The latter is directly related to the theoretical mill power consumption by the formula
Theoretical
Mill only
Gross ,kh/tonne
(Revs/lb at product SSA) x 0.3815

0.9
The mill only power consumption is compared with the theoretical kwhhonne to give the B.C.
conventional milling efficiencywhich is defined as follows:
Conventional B.C.
Milling Efficiency YO =
~~. x Theoretical Kwh/t
Actual kwhlt
4.1.3 TvDe of Millinp Svstem and the target B.C. Millinp Efflciencv
The BCI grindability test is an approximation of the open circuit milling situation, although the
actual specific power is expected to be less than that predicted in the laboratory. Similarly, the
specific power for a closed circuit mill is also expected to be less than for open circuit. In
addition, the use of third generation separator will yield lower specific power than for first or
second generation separators. The specific power value obtained from the grindability curve can
thus be converted to the target kwhhonne for open or closed circuit mills by using the following
factors.
0/0 B.C. Milling Efficiency
Without ~
MlllinRSvstem Grinding Aid Grinding Aid -
Open Circuit 115
Closed Circuit with
125
conventional separator 120 130
Closed Circuit with
high etliciency separator 125 135
This assumes that using a conventional separator or a high efficiency separator increases the
potential millingefficiencyof an open circuit mill by around 5 and 10% respectively; while use of
a grinding aid in all three millingsystemsincreasesthe potential milling efficiency by around 10O/O.
If the VOgrindability figure is known, the BC grindability curve can be related to a simple cuwe
fit. The conventional mill efficiency (E) can then be determined using the following formula:-
E
[
GxS
1
K
11.5 - [0.044 (s/100)3]
19
Where K = actual energy consumption mill only (kWh/tonne)
G= BCI grindability (??)
,
s = Lea and Nurse Specific $.uface area (m2/kg)
4.1.4 Overall Targeting of the Cement Mill Output
The above factors are combined together as follows:-
Installed
Actual Output x Motor (Power) x Target
(tph)
(P)
Efficiency (Yo)
Target Mill
Output (tph) =
Absorbed motor kw x Actual % mill efficiency (E)
Hence, to be a top petiormer a cement mill must be kept up to charge to draw its rated power
and must have the optimumconfiguration of mill internals with adequate ventilation and cooling.
4.2, The 1994 Benchmarking Results For Cement Milling
The results for open and closed circuit cement rnllls are show in Figures 2 and,3 respectively
(secondary benchmarking data). In overall energy consumption terms, the cement milling
kwh/tonne figures for 1992-94 are shown in Figure 4. This data is primary benchmarking data
and does not distinguish between types of mill, surface area targets etc.
4.2.1 Interpretation of Results
The performance data is expressed as ?40 of target based upon the above parameters. Caution
has to be exercised when interpreting results and a detailed knowledge of individual mills is
necessary before leaping to conclusions. Some conclusions which can be drawn are as follows:-
,
The closed circuit cement mills 1,2 and 3 at Ravena Works have good milling efficiencies
but are overrnotored. Hence the maximum mill usefhl power is taken as 4000 HP and
not the installed 4,500 HP.
G
G
G
The Ewekoro/Shagamu millingefficienciesare too high dueto the high surface area from
precipitator dust addition.
As mills NOS. 1, 2 and 3 at Rawang were converted in the latter half of 1994 the target
outputs were calculated assuming a high efficiency separator was in operation.
The pozzolanic cement ground at Cemento Melon, Chile, tends to reduce mill internal
coating problems and milling efilciency levels tend to be high.
20
:
;
;
~
:
;
:
:
:
:
:
:
i
:
.
1
:
.
.
.
.
::,m
,..
.
.
.
.
.
.
.
.,
.,..,
.
.
.
.
.
.
:
~
:
:
:
:
:
:
:
:
:
:
:
-
I
d

=
=
=
-
-
-
t
,
,.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
I
.
.
.
.
.
.
.
.
.
.
,
,
.
.
,
.
.
,
.
.
,
,
.
,
I
.
.
.
.
.
.
.
.
.
.
.
,
,
,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
,
.
,
.
.
,
,
,
,
,
.
,
.
,
,
,
,
.
.
I
.
.
.
.
.

4
.
.
.
.
.
,
,
.
.
,
.
,
.
,
.
.
-
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
,
.
,
.
.
.
,
.
,
.
.
.
.
.
.
,
,
,
o z
,
.r
y
;
:
:
,
:
@
.
I
.
:
.
:
p
:
:
.
:
.
:
.
:
.
:
:
:
.
:
.
:
.
:
m
i
i
E
l
.
;

.
.
,
.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
,

.
.

.
.

.
.
.
.
.
.

.
.
.
.
.
.

.
l
.

.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.....
.

,
.
.
.
.
.

~
,
.
,
,
,
,
,

.
.
.
.
.

.
.
,

.
.
,

.
,
.
.
.
.
.
.
.
.
.
t
;
,
.
,
,
.
,
.
.
~
.
.
.
.
.
.
.
t
.
.
.
.
.
t
.
,
,
:
.
.
.
.
.
.
.
.
t

,
I
,
.
r
I
.
.
.
,
.
.
.
.
.
.
.
.
.
;
.
.
:
.
,
.
.
,
,

.
.
.
.
..
.
.
.
.
.
..
.
.
.
.
.
,
,
1
.
-
.
.
.
.
.
.
;
.
.
.
t
.
.
.
.
.
,
.
. .
.
.
,
.
.

.
.
.
.
.
.
I [
,
.
I I
1 [
.

.
.

,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..
.
.
.
.
.
.
.
.
.
.
.
.
.
,
.
-
.
.
,
.
,
.
.
~
.

.
.
[
.
.
.
.

.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
~

.,
.
.
.
.
.
..
,.,
.,,.
.
.
.
.
.
.
.
l
.
.
:
.
.

.
,
.
.
,
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
~

.
.
.
.
.
.
.
.
.
:
.
.

.
.

.
.
.
.
.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
.
L
.
.
.
.
.
.
.
.
.
.
.
L
.
4
I
I
,
,
,
1
v
/
/
/
/
/
/
/
/
/
.
.
.
.
.
.
.
.
.
-
.
.
.
.
.
-

..
.
..
.
.
.
.
..
.
.
..
.
.
.
.
..
E
1
I
.
,
,
,
,
{
.
,
,
.
,
,
,
,
,
1
/
/
/
/
/
/
/
/
/
.
.
.
,
.
.
a
-
e
.:-U
=
T
=
.
=
.

.
-

-
.-
.,e
..-
=
,=
,
,=
...:.:..
[.
.
.
.
t 1
<
-

/
/
/
/
/
/
,
,

.
.
,
=
_
_
.
.
.
,
,
,
,
,
,
.
,
,
.
.
,
.
,
-
-
,
.
T
r
r
m
l
(
Z
/
/
/
/
/
/
/

t
.

.
.
.

,
,
,
r
r
r
m
.
,

.
.
=
-
.
I
I
T
r
m
As the absorbed mill power
powers were assumed:-
G
G
G
I
G
G
G
No. 21 mill
No. 16 mill
appears higher than the actual at Melon the following absorbed
4230 kW
890 kW
No. 10, 14 and 15 mills 700 kW
Note that the age/mechanical condition of several cement mills hasresulted in the mill
power being limiteddue to shell cracking problems. This applies to Cemento Melon 10,
14 and 15 mills.
The use of classif@g liners in certain cement mills prevents these mills from drawing the
maximum motor power. Examples are:-
Ravena Mills
Cauldon CM4
Plymstock CM2
.TheBenchmarking exercise is based upon data for the ,maincement produced. In cases
such as Aberthaw CM1O, it is not possible to fhlly optimise the mill for Bulk OPC when
it also has to produce Bag OPC and RHC. This results in a compromise to milling
efficiency i.e. chamber 1 has to have sufficient power to cope with 90 tph of Bag OPC.
This results in power wastage when grinding Bulk OPC as 78 tph.
Hence some compromise of milling efficiency is unavoidable where a single fill has to
produce several cement types with a wide range of outputs.
The r~orted grindabilityfrom Aberthaw of 127% corresponds to a neat clinker/gypsum
cement. In practice as around 4/0limestone is added agrindtillityof117/0 was assumed.
The reported grindabdityfromDunbar is regarded as being slightly higher than typical and
for the calculations a lower value of 139% was assumed.
The reported grindabilities for Rawang and Kanthan were determined in Malaysia.
Previous results showed that BCTC obtained lower values and consequently the current -
reported values were reduced accordingly.
The swingmill at Tulsa was not included in this exercise as it is also used to supplement
raw fked production.
Technical survey data has shown that the efficiency of the SVeardale cement mills is
amongst the highest for open circuit mills. However, when the rpill installed power figures
are examined, these are generally just below the installed power.
The Westbury cement mills tend to use more power than the installed motor power (not
exceedng the motor input power liinitations). Hence the milling efficiency compares well
on an installed power basis.
24
G
Detailed testwork on the Weardale and Westbury open circuit mills do not support their
relative rankhg. Whilst the Westbury mills tend to run at higher power consumption
than Weardale, the Westbwy millingefficiencies are generally lower than Weardales.
G
Hope and Northfleet works have 6000 HP closed circuit cement mills. It is known that
the efficiency of these mills are good when producing low surface area cements.
However, millingefficiencylevelstail off when the millshave to produce high surface area
cements due to the very poor pefiorrnanceof the CV. separators and mill cooling/coating
problems.
4.3 How Can We Act On These Results To Imwove Cement Millhw Performance
To illustrate how the benchmarkingresults can be used to establish targets, an example of a 1491
kW closed circuit cement mill is given below:-
4.3.1 Current Performance
Mill output
=
Absorbed Power
=
Cement Surface area
.
BC Cement grindability
=
BC Conventional mill efficiency
=
37 tph
1368 kW
320 m2/kg
122%
109VO
4.3.2 Potential Mill Performance
Target mill efficiencyfor closed circuit mill with high efficiency separator without grinding aid =
125V0
Potential output at 1491 kW
and 125/0eflciency
Actual output as% of target
.37 XM2XE
1368 109
=
46.2 tph
= 100 x x-
= 80.lVO
46.2
25
FIGURE 5
CM Axial Test
Residue (%)
80
60
40
20
SSA (cm/g)
let Chamber - central 2nd Chamber
Dia
\!/
/l\
.
.
.
.
!
-,
.
I
o
ABCD
Av.
400
80
60
40
20
0
ABCDEF
2.36mm 1.18mm 300 micron 90 micron 45 micron SSA
+- * +
media size (mm)
let Chamber
67.5
ABCD
Bntra
Dia
240
220
200
180
- SSA (cm/g)
I SSA Av.mediasize (mm) I
160
140
120
%d Chamber
40.3
ABCDEF
220
200
180
160
140
120
4.3.3 Mill Inspection and Axial Sampling Test Results
Figure 5showsthe fill tidsampling test results and media grading. Theconclusions drawn
from this data plus the mill internal inspection showed the following:-
G
The mill first chamber is too long at 34%0. This should be shortened to 28%0of the total
length. This will allow more fine media to be added in chamber 2 (LONG TERM).
G
The ixiterrnediatediaphragms have large diameter central ventilation grids. Modi~ these
by fittinginternal retainingrings to allowthe volume loads to be increased - note that mill
currently only 91.8?40of its rated motor power. (SHORT TERM).
c
Replace the classifyingliner in chamber 2 by a Dragpeb liner to permit higher ball charge.
The classifyingIinerreduces the internal diameter and this also limits the maximum power
drawn (LONG TERM).
G
Regrade chamber media to BCTC recommended grading (SHORT TERM).
G
Regrade chamber 2 media and remove media of 60- 30mm size. Replace this by 25-
15mmmedia (SHORT TERM).
Hence, the benchmarkingresults indicate the potential to make a significant improvement to this
millsperformance. Once the targets are established- detailedtestwork and inspections are needed
to identifi what we must do to improve the mill. Please note that the improvement programme
shown above identifies short and long term proposals. Obviouslythere is no sense in replacing
existingIinersif these are in good condition. However, when the mill requires major repair, use
this opportunity to move the diaphragm/change liner designs etc. Hence the optimisation of any
mill has to be an economic compromise.
27
4.4 How We Can ImDrove Our Benchmarking Methodolon
4.4.1 Targeting Methods
The BC conventional mill efficiency method was
simplicity and greater flexibility when dealing with
,
chosen for the 1994 data review due to its
many types of milling system.
However, this simplemethod does have several limitations,in particular, the method tends to set:-
i) a too high a target performance may be set when producing low surflacearea cements
(280-300 m/kg)
ii) a too low a target pefiormance maybe set in mills producing a mediumhigh surface area
cement (350-380 m2/kg) if grinding aids are used and large separators are operated
resulting in high circulating loads.
BCTC have therefore recently developed the Mill Surface Production Factor (M. S.P.F) method
for targeting cement mills. This method is usefid for targeting cement mills in U. S.A. which
feature:-
Use of grinding aid
High mill circulating loads
Medium/Mgh cement surface areas
Large conventional separators.
Using the BC conventional mill efficiency, figures of 140-150% efficiency can be achieved which
are much higher than the target figures shown in section 1.3. The M. S.P.F. method is more
appropriate for targeting mills in USA and Chile but this method also has its limitations.
In the fiture, we may use a combination of these two methods for targeting purposes. In
additio~ the current technical suweys being carried out on U.K./overseas works will help BCTC
to establish a more accurate data base concerning individual mills, their performance levels, the
factors affecting their efficiency as well as their performance targets.
4.4.2 Accuracy of the Reported Data
The 1994Benchmarking exercise has shown up several anomalies in the data presented. Further
plant surveys have revealed that some of the data reported by individual works is misleading due
to factors such as:
G
Inaccurate assessment of individualmill outputs due to inaccurate weigh feeders, booking
methods.
G Mill only power consumption figures are inaccurate due to lack of accurate metering
andlor inclusion of some ancillary plant.
28
G
Additives such as Limestone precipitator dust causing major reduction in cement
grindability.
G
Artificiallyhigh cement surfaceareas dtie to dust addition and subsequent hydration of the
cement which results in an artificially high cement surface area (if the artificially high
performance of the Ewekoro and Shagamu cement mills).
Hence, the results obtained from the Benchmarking exercise need to be read with some caution
at present. These results can onlybe as accurate as the data reported. BCTC will amend obvious
data errors as far as possible but we do urge each works to report the data as accurately as
possible.
4.4.3 Targeting According to Cement Type
Please note the comments contained in section 2.2 paragraph (vii) concerning the efficiency
compromises which have to be made when a single mill has to produce several types of cement.
The Benchmarkingdata is based upon the ~ product cement only. In time it may be possible
to extend the evaluation to examine individual mills producing several types of cement. This
could show up how much efficiency is lost by the need to compromise when producing several
different cements.
..
29
4.5,
G
G
G
G
G
G
G
Conclusions
The benchmarking exercise provides a valuable tool for comparing the performance of
cement mills within the BC Group.
Having identified those mills which need improving, we have to set realistic targets for
their petiormance.
The target performance should be based upon the following:-
Detailed axial sampling test data.
Cement grindability data.
A working knowledge of any mill plus ancillary plant limitations e.g. FK pump,
bucket elevator, separator capacity.
The target performance should be judged using the above data, common sense as well as
the estimates of-
9
-
B.C. conventional mill efficiency
Mill surface production factor. 1
Reference should be made to the latest Cement Mill Optimisation guidelines which are
updated regularly by BCTC.
BCTC are continually updating our know-ledge base on the performance of new mill
internal designs as well as the methods for targeting mill performance.
Please consult BCTC engineers prior to mill refurbishment programmed so that the
opportunity for improvement is used. Please do not simply repeat old internal
designs/original mill configurations if these do not give the optimum performance.
30