Boletín Refractarios RHI
Boletín Refractarios RHI
Boletín Refractarios RHI
Chrome Corundum
New Thrust Lock System for Applications in Glass Novel Gas Purging System
Rotary Cement Kilns Melting Furnaces for Copper Anode Furnaces
RHI Bulletin >2 > 2013
The Journal of Refractory Innovations
The products, processes, technologies, or tradenames in the
RHI Bulletin may be the subject of intellectual property rights
held by RHI AG or other companies.
2<
RHI worldwide
Refractories Simply RHI Attends UNITECR 2013
Explained in the New RHI Canada >> The 13th Biennial Worldwide Congress on Refractories,
Animation Video UNITECR13, was held from September 1013, 2013, in Victoria,
British Columbia.
Austria >> A new RHI video, titled More than 600 international participants were present, including
What are Refractory Products? pro- refractory producers, raw material suppliers, customers from vari-
vides a short animated explanation of ous industry segments, and academia. Considered the most impor-
the importance, application, and produc- tant refractory conference worldwide, it provides the opportunity
tion of refractories. Available in both to discuss the latest trends in the sector of high-temperature mate-
English and German on the RHI website rials.
(www.rhi-ag.com), it is designed to be RHI participated in various sessions, presenting topics including
informative for nonexperts and high- zirconia characterization methods, the challenges for Western mag-
lights the indispensibility of refractories nesia-carbon brick manufacturers, and investigation of thermal
for everyday life. shock in lower slide gate plates. Additional papers focused on dif-
ferent destructive and nondistructive methods to characterize mag-
nesia and magnesia-chromite bricks, improved physical properties
of alumina-silica bricks using sol impregnation, and the perfor-
New Transport and Storage mance advantages provided by hybrid spinel technology in basic
Brochures Available bricks for cement rotary kiln applications. The latter topic is also
included in the first paper of this Bulletin (page 10).
Austria >> Following frequent customer
requests, RHI created new bilingual
(German/English and English/Russian)
brochures concerning transport and stor- INTERSTOP Awarded Delivery of Stopper
age issues, with additional languages
planned. Containing state of the art three- Control System for the New Baosteel Steel
dimensional illustrations, they highlight Plant in Zhanjiang
the main packaging types, labels, and
storage advice. Furthermore, an insight China >> Baosteel Group Corporation is one of the largest steel
into the undesirable hydration process is producers in the world, with an installed capacity totalling 45 mil-
provided, along with possible on-site lion tonnes per annum. Currently, Baosteel is building a completely
tests. new basic oxygen furnace plant in Zhanjiang (China), as part of a
Due to different customer requirements, strategy to concentrate production units and operate process tech-
the new brochures are available in Steel nology that meets market requirements.
and Industrial versions, both of which are The steel plant will operate three BOFs with a nominal capacity
available on request from RHI Sales and of 350 tonnes. As highly sophisticated steel grades (e.g., silicon
Technical Marketing personnel. steels) are scheduled to be produced, converter slide gates for slag
retainment are foreseen. Furthermore, one ladle furnace and two
RH degassers are planned.
In the initial phase, with startup scheduled for the fourth quarter
TUNFLOWThe RHI Tundish of 2015, 2 x 2 strand slab casters from Siemens VAI Metals Tech-
Flow Modifier nologies will be installed with an annual steel output of 6 million
tonnes. In the next stage, the existing 2 x 1 strand Danieli continu-
Since its market launch in 2010, RHIs ous casting machines (CCMs) from Baosteel Luojing will be relo-
impact pot and flow modifier for the cated and revamped to a 1 x 2 strand CCM at Zhanjiang. After all
tundish has successfully entered 19 coun- three CCMs are operational, the total output will reach 8.75 million
tries all over the world. Experienced tonnes per annum in 2016.
researchers and the use of CFD and water Baosteel invited international bids for the entire stopper
modelling simulations are the basis for control system to be installed initial on the 2 x 2 strand slab cast-
RHIs customer-orientated approach with ers. The scope of equipment offered by RHI included stopper rods,
this product, enabling steel plants to make electrically actuated stopper drives, the newest generation of VUHZ
the best of their existing conditions, such mould level sensors, an advanced argon management system
as tundish geometry, by using the opti- including data storage, and the process control system. A team
mum tailored flow modifier concept. from INTERSTOP in Switzerland and Shanghai worked jointly on
Under the brand name TUNFLOW, RHI the bid, convincing Baosteel of the technology leadership in this
sold over 4000 pieces in 2012 and expects field of flow control. As a result, on October 18, 2013, Stopinc AG
total sales of more than 6000 pieces in (Switzerland) was informed it had been awarded the contract.
2013, with an ongoing trend for further Delivery of the equipment is scheduled for the third quarter of
growth in the next years. 2014.
>
> 33
RHI worldwide
RHI Clasil Receives IRMA TMS Annual Meeting in February 2014
Award for Excellence USA >> The 143rd TMS Annual Meeting and Exhibition will take
place at the San Diego Convention Center, California. More than
India >> At the Indian Refractory Mak-
4000 of the worlds business leaders, engineers, scientists, and
ers Associations (IRMA) 52nd annual
other professionals in the materials field are expected at this event,
general meeting, held on August 17,
which is scheduled from February 1620, 2014, and provides the
2013, the IRMA Awards for Excellence in
opportunity for an outstanding exchange of technical knowledge.
20122013 were announced. RHI Clasil
RHI will be presenting technical papers during the prestigious
Ltd., (India) received the accolade Over-
EPD Symposium in honour of David G.C. RobertsonCelebrating
all Excellence in Export Performance for
the Megascaleincluding a paper titled Modelling and Compari-
20122013 in the large scale sector. This
son of Refractory Corrosion at RHIs Technology Center. RHI will
is the second consecutive year RHI Clasil
also participate in the 5th International Symposium on High-Tem-
has been honoured with this award.
perature Metallurgical Processing. RHI personnel from numerous
The IRMA was established in 1958 and
departments look forward to seeing you at booth No. 229 to dis-
is the national organization representing
cuss any point of interest relating to the presented topics and other
companies designing, engineering, and
refractory-related areas.
installing refractories in India. Currently,
more that 70 refractory manufacturing
companies are members of the IRMA.
New Casting Sequence Record for RHI at
Outokumpu Stainless Alabama
RHI Commissions Three USA >> Outokumpu Stainless USA (Calvert, Alabama) is a green-
field steel plant that produced its first heat on November 16, 2012.
BOF Taphole Changing Units Currently, RHI has a supply contract for 100% of the isostatically
at Tata Steel Jamshedpur pressed continuous casting refractory requirements.
On September 12, 2013, the first 8 heat sequence was success
India >> Tata Steel is one of the largest fully cast with a RHI submerged entry nozzle (SEN), surpassing the
steel producers in India with an installed previous 6 heats. This achievement was recognized and celebrated
capacity of 10 million tonnes per annum. by the Outokumpu casting personnel, along with RHI.
The plant has three BOF steel melting The stainless steel grade was 304, in a mould width of 1260 mm,
shops designated LD#1 (2 x 165 tonnes), with a total casting time of 499 minutes. The SEN (DELTEK
LD#2 (3 x 165 tonnes), and LD#3 (2 x 165 HX10Z98WX08), along with the ladle shroud (DELTEK A100) and
tonnes). Whereas LD#1 and LD#2 have stopper rod (DELTEK A100D99N) were produced at the RHI Say-
been operating for a long time, LD#3 brook plant (Ohio, USA).
was only commissioned in 20112012. An enormous amount of effort from the RHI and Outokumpu
In LD#1 and LD#3, the two vessel BOF team members contributed to this achievement, and additional
operation requires a high-efficiency tap- improvements are still possible.
hole exchange mechanism to ensure
maximum productivity is achieved. RHI
was approached by Tata Steel to assist
in this regard and after a detailed techni- 56th International Colloquium on Refractories
cal discussion it was finalized that RHIs
Taphole Breakout Device (TBD) along
2013
with ISOJET C type tapholes would be Germany >> The 56th International Colloquium on Refractories,
the best possible solution to achieve the held in Aachen from September 2526, 2013, provided an impor-
desired taphole exchange time. As a tant opportunity to present current research topics as well as
result, RHI received an order for three develop and recruit the next generation of refractory engineers and
TBD machines and the ISOJET C taphole scientists. Diverse new ideas, with the focus on industrial applica-
requirement for both shops. tions, were introduced by European and international industries
RHI worked closely with Tata Steel and and research institutes. RHI presented papers regarding basic
within a period of 12 months all four refractory wear in the nonferrous metals industry; an investigation
BOFs in LD#1 and LD#3 were converted and verification of ultrasonic and resonant frequency measurement
to the ISOJET C taphole system and the methods for fired refractory products; and the importance, organi-
three TBD machines were successfully zation, and evaluation of interlaboratory tests.
commissioned. The taphole exchange Once again RHI had a booth in the exhibition area, providing a
time frame as well as the taphole life meeting place for creative brainstorming and knowledge exchange.
achieved so far is in line with the cus- Since a considerable number of delegates came from outside
tomers expectations. Europe, this networking platform had an international character.
4<
4 <
Nucor Steel Texas Awards South America Hosts the Copper 2013
Contract for INTERSTOP International Conference
Automatic Mould Level Chile >> The 8th Copper/Cobre 2013 conference took place from
ANKERHARTH Mixes er is an essential and often crucial step in order to achieve the
requested performance level. Therefore, for each grade all the
Austria >> In the last 50 years, ANKER- quality-relevant processing steps, from the right storage conditions
HARTH has become the No. 1 hearth ram- through to installation, are now described in tailor-made operating
ming mix for electric arc furnaces on the instructions. These descriptions are as comprehensive as neces-
global market. To celebrate this success sary, while being as simple as possible. Additionally, pictograms
story, all ANKERHARTH big bags from the are included in all instructions, which highlight the most important
Breitenau and Hochfilzen plants will have characteristics and provide hints for proper mix application. In the
anniversary stickers in the coming future these pictograms should also enable users with reading dif-
months. Further information regarding the ficulties to understand the main content of the instructions and
ANKERHARTH product range can be ensure correct utilization. As part of the project, the operating
found in the RHI Bulletin > 1 > 2013Steel instructions for nonbasic and basic mixes have also been
Edition. standardized.
>5
Total dedication throughout the year
RHI personnel provide a range
of refractory solutions
Seasons greetings
from staff at the
Technology Center Leoben
www.rhi-ag.com
ANKRAL Q-Series
Pure innovation
setting the pace
Excellence
www.rhi-ag.com
in Refractories
Editorial Contents
A decade ago we launched the RHI Bulletin and over 200 articles 10 Hybrid Spinel Technology Provides
have been published describing innovative products and services Performance Advances for Basic Cement
from RHI as well as broader refractory-related topics. Available in
Rotary Kiln Bricks
both printed and digital format, it regularly communicates how
We lead the Industry. Everywhere. Anytime. 13 New Sol-Bonded Product Members
Gunning Experiences and Shotcasting
This edition contains a wide range of papers, principally focusing Applications
on the cement, glass, and nonferrous metal industries. The first
17 RHI Thrust Lock System for Cement
paper describes how hybrid spinel technology in the ANKRAL
Rotary Kilns
Q-series incorporates brick flexibility into magnesia bricks, which is
not only essential for lining mechanically critical areas but is highly 20 Safety in the Mining IndustryA Value
advantageous in the case of alkali salt infiltration. This is followed by 25 Low Carbon Economy Roadmap for the
a description of the new sol-bonded monolithics that have been
Ceramic Industry
developed for shotcasting and shotcreting applications, enabling
rapid and efficient installation of large amounts of material. Addition- 27 Chrome Corundum: An Alternative to
ally, the extended production programme to facilitate global access Isostatically Pressed Products
to the sol-bonded products is highlighted. A newly designed retain- 31 Postmortem Studies of Standard Silica,
ing system for the outlet zone in rotary kilns is the subject of the next No-Lime Silica, and Fused Cast AZS from
paper, illustrating how the RHI Thrust Lock System will significantly
Oxy-Fuel Fired Soda-Lime Glass Melting
improve refractory service life in this highly stressed kiln section.
Furnaces
Advances described for the glass industry include an economic 36 Microstructural Analysis of Magnesia
alternative to isostatically pressed chrome-containing material for Bricks Operating Under Altered Conditions
areas in glass furnaces that require high corrosion resistance. The in the Regenerator Condensation Zone of
bonded alumina-chrome developed by RHI provides an additional Glass Melting Furnaces
benefit of superior thermal shock resistance. Two further papers
describe postmortem studies that demonstrate the STELLA GNL
40 RHI US Ltd Provides Value-Added Services
no-lime silica brand is advantageous for crown applications in oxy- to Environmental, Energy, and Chemical
fuel fired glass furnaces and microstructural analysis of magnesia Sectors in the USA
bricks operating under altered conditions in the regenerator conden- 42 American Petroleum Institute Refractory
sation zone. Installation Certification Programme
An innovative gas purging system for stationary and tilting copper 44 Determining Resistance to Abrasion at
anode furnaces is introduced in the nonferrous section, as well as Ambient TemperatureImproving
thermal shock resistant alumina-chromia products for the copper Comparability Between Laboratories
industry, and FEA modelling of a lead rotary furnace that resulted in 50 New Innovative Gas Purging System for
a modified lining design and improved furnace lifetime. Stationary and Tilting Copper Anode
Furnaces
Additional papers in this edition describe the full service capability
now offered by RHI US Ltd., the American Petroleum Institute refrac- 54 Thermal Shock Resistant Alumina-Chromia
tory installation programme that has been developed to improve Products for the Copper Industry
refractory reliability and service life, and an approach to increase the 59 Thermomechanical Finite Element
reproducibility of material abrasion testing results. Furthermore, an
Simulation of Heat-Up Processes in a
article regarding safety in the mining industry provides an insight
Lead Rotary Furnace
into this critical issue and the low carbon economy roadmap for the
ceramic industry is summarized in a paper that also includes the first
PRE results providing an average European carbon footprint for a
range of refractory products. Subscription Service
In closing I extend my gratitude to all the authors who kindly contrib-
and Contributions
uted to this edition and the editorial team for their dedication to this
We encourage you, our customers and inter-
publication, many of whom have been involved from its inception. ested readers, to relay your comments, feed-
back, and suggestions to improve the publica-
tion quality using the contact details below.
Yours sincerely Furthermore, to receive the RHI Bulletin free of
charge please e-mail or fax your details to the
Bernd Buchberger Subscription Service using the form on the
Corporate Research and Development back page.
RHI AG
E-mail: [email protected]
Phone: +43 (0) 502 13-5323
Fax: +43 (0) 502 13-5237
>9
RHI Bulletin > 2 > 2013, pp. 1012
Introduction spinel brick, while types B and C were based on hybrid spi-
nels and different sintered magnesia grades (Table I). All the
Cement rotary kiln refractories are exposed to severe
brick types were fired at the same temperature.
stresses originating from thermal, mechanical, and chemi-
cal loads. In particular, kiln rotation places basic refractory
materials under critical stresses because magnesia-based Dynamic Youngs Modulus
bricks are comparatively brittle. Therefore, special additives Initially the dynamic Youngs modulus of all three materials
are used, so-called flexibilizing additives, to reduce the brit- was determined to characterize the thermomechanical
tleness in order to fulfil the high thermomechanical require- behaviour. This investigation was performed by measuring
ments in cement rotary kilns. the ultrasonic velocity as a sample was heated from room
temperature to 1400 C and as it cooled back to room tem-
In addition to the thermomechanical stresses present in perature [1].
rotary kilns, chemical load must also be considered, espe-
cially in view of the steadily increasing use of alternative The Youngs modulus or modulus of elasticity describes the
fuels. The infiltration of volatile components like K2O, SO3 , amount of stress required to create a defined, reversible
and Cl leads to two effects: Corrosion of the ceramic brick deformation in a material. It is also a characteristic of the
bonding and thereby degeneration of the brick structure, amount of stored elastic energy at this defined deformation
and a second no less critical effect, densification of the brick (i.e., a high Youngs modulus indicates a high amount of
structure by alkali salts that results in loss of thermome- stored elastic energy). The higher this value, the higher the
chanical flexibility.
brittleness and the lower the thermal shock resistance of
the material. When the dynamic Youngs modulus of a pure
While in the past chrome ore was used as a standard flexi-
magnesia brick sample without flexiblizer was determined,
bilizing concept, nowadays spinel group (AB2X4) minerals
it showed a very brittle behaviour that slightly reduced with
are state of the art. These minerals have a lower thermal
increasing temperature and then increased back to the orig-
expansion than magnesia, and the greater the thermal mis-
inal value as the sample cooled to room temperature (Fig-
match compared to magnesia, the more effective the reduc-
ure 1). In comparison, the Youngs modulus was signifi-
tion in brittleness. The most commonly used minerals are
cantly lower when MA spinel flexibilizer was present. Fur-
magnesia-alumina spinel, MgAl2O4 (MA spinel); iron-alu-
thermore, the temperature-dependent change of the
mina spinel, FeAl2O4 (hercynite); and manganese-alumina
Youngs modulus was significantly different for the magne-
spinel, MnAl2O4 (galaxite). The main focus of the research
sia MA spinel sample compared to the pure magnesia
performed was to develop, through the systematic use of
different spinels, high-performance bricks for cement rotary
kilns able to withstand high mechanical loads in combina-
A B C
tion with chemical stresses. An additional aim was to
reduce the loss of flexibility once the brick structure is den- MgO 88.0 87.2 86.1
sified by alkali salts. CaO 0.7 0.7 1.6
SiO2 0.3 0.3 0.8
Spinel-Containing Magnesia Bricks Al2O3 10.5 9.1 9.0
Two different brick types based on sintered synthetic high- Fe2O3 0.5 2.7 2.5
purity magnesia (MgO > 98 wt.%) and one type based on
sintered natural magnesia (MgO > 97 wt.%) were produced Table I. Chemical composition (wt.%) of the magnesia-based
on a pilot scale for the analyses. Type A was a magnesia brick types containing different spinel concepts.
10 <
RHI Bulletin > 2 > 2013
refractory (see Figure 1). At room temperature the Youngs load/displacement curve during stable crack propagation
modulus was low, slightly increasing as the sample was and indicates a materials ability to absorb thermomechani-
heated up. Upon cooling the Youngs modulus significantly cal stress. To provide a more detailed understanding of the
increased until it reached a maximum at approximately different brick types thermomechanical properties, wedge
10001100 C. With further cooling the Youngs modulus splitting tests were performed at 1100 C, the temperature at
returned to the original value at ambient temperature. The which the most brittle behaviour was determined from the
increase during cooling down was caused by the thermal dynamic Youngs modulus. Brick type B showed a signifi-
misfit between magnesia and flexibilizer. cantly higher Gf (area under the graph) than brick type A
(Figure 3). The maximum vertical force occurred at a higher
The type B material containing hybrid spinel showed a displacement compared to the brick type A and after passing
lower overall dynamic Youngs modulus compared to the the maximum, the vertical force remained at a higher level,
conventional magnesia MA spinel brick sample (type A) which indicates a higher resistance against crack propaga-
(Figure 2), as did type C. In addition, not only was the tion. Brick type C showed similar results to brick type B.
Youngs modulus at a lower level during the entire heat-up
In summary, both brick types containing hybrid spinels
procedure, which would lead to a lower risk of spalling at
showed a significantly lower dynamic Youngs modulus and
the hot face during heat-up in a cement rotary kiln, the
a very high resistance against crack propagation compared
increase during cooling was also much less. Since the brick
to the magnesia MA spinel brick. These thermomechanical
type B did not show such a significant increase in brittle-
properties are advantageous in high mechanically loaded
ness during cooling down, the danger of spalling at the hot
rotary kilns.
face during kiln cooling would be considerably reduced.
Thermomechanical Properties of
Wedge Splitting Test
Infiltrated Bricks
The wedge splitting test enables the mechanical fracture
In order to examine the thermomechanical behaviour under
behaviour of refractory materials to be examined and can
chemical load, further laboratory investigations were per-
be performed at temperatures up to 1500 C [2]. The spe-
formed. Crucibles were prepared from the three different
cific fracture energy, Gf , is determined by evaluating the
brick types. The crucibles were filled with KHSO4, heated to
950 C, and held for 24 hours. The dynamic Youngs modulus
800 of the crucible material was measured ultrasonically before
and after the materials had been subject to chemical loading.
700 Pure magnesia brickbrittle
Dynamic Youngs modulus [GPa]
600
As expected the noninfiltrated crucibles with hybrid spinel
technology showed a lower Youngs modulus compared to
500 noninfiltrated magnesia MA material (Table II). However, the
400
Type mass (%) E0 (%) E1 (%)
300
A 5.7 100 118
200 B 5 65 97
100 C 5.9 64 95
Magnesia brick with flexibilizerlow brittleness
0
Table II. Mass increase (mass) and relative Youngs modulus of
0 500 1000 1500
the three brick types before and after chemical loading. E0 is the
Temperature [C] relative Youngs modulus prior to chemical loading compared to
a noninfiltrated magnesia MA spinel sample and E1 denotes the
Figure 1. Dynamic Youngs modulus of brittle (pure magnesia) relative Youngs modulus after chemical loading compared to a
and flexible (magnesia MA spinel) brick samples. noninfiltrated magnesia MA spinel sample.
800 900
n A n A
700 n B n B
600
Youngs modulus [GPa]
600
Vertical force [N]
500
400
300
300
200
100
0 0
0 500 1000 1500 0 1 2 3 4 5 6 7 8 9 10
Temperature [C] Displacement [mm]
Figure 2. Dynamic Youngs modulus of magnesia MA spinel (A) Figure 3. Wedge splitting test performed on magnesia MA spinel
and magnesia hybrid spinel (B) brick samples. and magnesia hybrid spinel brick samples at 1100 C.
> 11
RHI Bulletin > 2 > 2013
heavily infiltrated crucibles with hybrid spinel technology Established several years ago, the first member of the
also had a lower Youngs modulus than the nonchemically Q-series, ANKRAL QF, is the top grade in the series. It has
loaded magnesia MA spinel material (Figure 4). This result demonstrated outstanding performance in critical kiln areas,
justifies the expectation that magnesia hybrid spinel bricks for example tyre sections subjected to alkali salt attack [3].
with outstanding thermomechanical properties provide In many cases the lifetime of previous installations has
excellent service lifetimes in high chemically loaded cement been increased significantly, in some cases even more than
rotary kilns. Since magnesia hybrid spinel bricks strongly doubled. The experience gained from in service results with
infiltrated with alkali salts are not as brittle as noninfiltrated ANKRAL QF was the starting point for further product
standard magnesia MA spinel bricks, the sensitivity of den- developments. For example, the hybrid spinel technology
sified bricks containing the hybrid spinel technology to ther- was adapted for use with RHIs own sintered magnesia
mal shock and thereby spalling at the hot face is signifi- resources. The result was ANKRAL Q2, a product with an
cantly reduced. unparalleled price/performance ratio. This RHI brand also
has a long list of successful references [4]. As installation
The Q-SeriesA Complete Product Range results have shown, ANKRAL Q2 is able to successfully
compete with high cost competitor products and also offers
Based on research results and experience gained from suc-
solutions for highly loaded transition zones in alternative
cessful installations and brick performances, the product
fuel fired kilns.
range containing the hybrid spinel technology has been
extended and a comprehensive new product line is avail
ANKRAL Q1 has been operational in various kilns since the
ablethe Q-series.
beginning of 2013 and supplements the high-grade Q-series
product range, also focusing on highly loaded transition
zones in alternative fuel fired kilns.
125
n A
115 n B Additionally, the Q-series offers a solution for the sintering
n C zone. With ANKRAL QE RHI has developed a hybrid spinel
Relative Youngs modulus [%]
References
[1] Triessnig, A., Studnicka, H. and Prietl, T. Determination of Thermo-Mechanical Properties at the RHI Refractories Technology Center Leoben.
RHI Bulletin. 2006, No.1, 3337.
[2] Tschegg, E., Fendt, K., Manhart, C. and Harmuth, H. Fracture Properties of Refractory Materials Under Uniaxial and Biaxial Loading. RHI Bulletin.
2010, No. 1, 4047.
[3] Jrg, S., Gelbmann, G., Krischanitz, R. and Fritsch, P. ANKRAL QFA New Brick to Combat Tyre Stresses. RHI Bulletin. 2010, No. 1, 812.
[4] Jrg, S., Gelbmann, G. and Krischanitz, R. ANKRAL Q2An Innovative Solution for Transition Zones. RHI Bulletin. 2012, No. 2, 811.
Authors
Gerald Gelbmann, RHI AG, Technology Center, Leoben, Austria.
Roland Krischanitz, RHI AG, Industrial Division, Vienna, Austria.
Susanne Jrg, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Roland Krischanitz, [email protected]
12 <
RHI Bulletin > 2 > 2013, pp. 1316
> 13
RHI Bulletin > 2 > 2013
Brand Al2O3 SiO2 Fe2O3 SiC CCS (N/mm2) LTE ATL SL Raw materials Additional information
(wt.%) (wt.%) (wt.%) (wt.%) 110 C 1000 C 1000 C (C) (months)
(%)
COMPAC SOL F53G-6 50.0 47.0 0.7 - 45 45 0.55 1570 18 Dense fire clay, low iron
COMPAC SOL M64G-6 61.0 34.0 0.7 - 55 65 0.55 1670 18 Mullite
COMPAC SOL S64G-6 63.0 36.0 0.5 - 55 60 0.55 1670 18 Andalusite
COMPAC SOL FS99-3 1.5 98.0 0.2 - 20 30 0.00 1600 18 Fused silica
COMPAC SHOT FS99-3 0.2 99.5 0.1 - 12 25 0.00 1650 18 Fused silica Shotcreting, hot repair
CARSIT SHOT F30SB-6 39.0 30.0 0.6 30 63 75 0.55 1550 18 Dense fire clay, low iron/SiC Shotcreting
CARSIT SHOT M10SB-6 58.0 28.0 0.9 10 55 65 0.55 1650 18 Mullite/SiC Shotcreting
CARSIT SOL F10G-6 45.0 42.0 0.7 10 45 50 0.55 1550 18 Dense fire clay, low iron/SiC
CARSIT SOL F30G-6 38.0 31.0 0.7 30 45 45 0.55 1550 18 Dense fire clay, low iron/SiC
CARSIT SOL M10G-6 57.0 30.0 0.7 10 55 65 0.55 1650 18 Mullite/SiC
CARSIT SOL M30G-6 44.0 20.0 0.7 30 65 75 0.55 1580 18 Mullite/SiC
CARSIT SOL S10G-6 60.0 23.0 0.7 10 55 55 0.55 1650 18 Andalusite/SiC
CARSIT SOL S30G-6 44.0 22.0 0.7 30 55 55 0.55 1580 18 Andalusite/SiC
Table I. Product range of sol-bonded gunning and shotcreting mixes for the cement industry. Abbreviations include cold crushing
strength (CCS), linear thermal expansion (LTE), application temperature limit (ATL), and shelf life (SL).
Figure 1. Cooler roof in a German cement plant 1 year after Figure 3. Cooler roof section in Figure 1 after 2 years in opera-
installation. The darker orange field in the background was lined tion. The orange coloured cement-bonded material in the back-
with a competitors conventional cement-bonded gunning mate- ground showed heavy wear.
rial. The other areas were lined with CARSIT SOL F10G-6.
Figure 2. Reverse view of the cooler roof section in Figure 1. Figure 4. Cooler roof section in Figure 2 after 2 years. The
Both gunned materials looked very similar in terms of wear after anchoring was completely exposed in the orange coloured
1 year and the performance was considered equal by the cus- cement-bonded material in the foreground and material loss had
tomer at this time. occurred due to significant crack formation.
14 <
RHI Bulletin > 2 > 2013
easily. Shotcasting is a wet gunning method using adapted product. In fact two products were created containing silicon
vibration castables. It involves mixing the monolithic with carbide, CARSIT SHOT M10SB-6 and CARSIT SHOT F30SB-6.
the required amount of liquid in a paddle or compulsory The acronym SB in the brand name indicates sol-bonded and
mixer, pumping the wet material via double piston pumps to distinguishes these products from the conventional cement-
the application area, and finally gunning with the use of a bonded brands.
small amount of compressed air and an accelerating liquid.
Initially the cooperation with Moeller resulted in the installa-
For shotcasting applications, the well-known RHI sol-bonded tion of more than 100 tonnes of CARSIT SHOT M10SB-6 and
vibration castables were slightly adapted to improve their CARSIT SHOT F30SB-6 for a Swiss cement plant repair (i.e.,
pumping propertiesmainly to reduce friction in the pipes cooler, kiln hood, nose zone, and inlet chamber area) in spring
and hoses. In addition, it was necessary to develop a com- 2013 (Figure 7). A few weeks later a further cooler repair was
pletely new accelerating liquid, since the sodium silicate carried out in a German cement plant (Figure 8) by Moeller.
solution (DIKASIL K) commonly used for cement-bonded
materials does not work with the cement-free bonding sys- Sol NewsRecent Developments
tem. As it is very effective, only a few litres of the newly
In the last years, the focus for sol-bonded castables was
developed DIMASIL SB-LIQUID accelerator are needed per
mostly on Europe, with a more recent extension of the
tonne of shotcasting mix.
demand worldwide. Therefore, whilst virtually all sol-bonded
castables were originally produced in Urmitz (Germany),
To fulfil the very tight development schedule, ensure optimal
RHIs largest alumina monolithic production plant, now a
workability, and to profit from the experience of an installa-
selection of sol-bonded products are also available from RHI
tion company, RHI worked closely together with Moeller
production sites in India, Canada as well as Mexico, facilitat-
Feuerfesttechnik GmbH & Co. KG (Germany). Moeller was
ing access to these sophisticated products for customers
very interested in this project because it sees great potential
globally.
in the sol-bonding features, particularly the short heat-up
period. The trials included tests with different pumps, a real-
North and South America
istic gunning environment, including overhead gunning and
anchoring, as well as sample preparation to examine the After the first excellent results with sol-bonded materials in
physical properties after installation (Figures 5 and 6). North and South America delivered from Europe, production
of selected castables from the sol-bonded product range was
Development during the 4 trial days at Lemgo (Moeller main extended to the RHI plants in Tlalnepantla (Mexico) and Burl-
yard) and the RHI Urmitz plant is a very good example of the ington (Canada). In addition, products for torcreting applica-
successful collaboration between a refractory producer and tions are in preparation (Figures 9 and 10) and will be avail
an experienced installation company to create a new able by the end of 2013.
Figure 5. Machinery and setup for a shotcreting trial with the Figure 7. Kiln hood (roof) gunned with CARSIT SHOT M10SB-6
German installation company Moeller Feuerfesttechnik in 2013 at a Swiss cement plant in spring 2013.
at their main yard in Lemgo.
Figure 6. Shotcreting trial at Lemgo showing overhead gunned Figure 8. Shotcreting application of CARSIT SHOT M10SB-6 in a
panels. The lower left section generated a gunned sample for cooler area at a German cement plant in spring 2013.
further testing.
> 15
RHI Bulletin > 2 > 2013
Figure 9. First sol-bonded gunning trial at the RHI Tlalnepantla Figure 10. Gunning equipment at the RHI Tlalnepantla plant
plant (Mexico) in April 2013. (Mexico), including an Allentown pressure vessel gunning
machine.
References
[1] Blajs, M., von der Heyde, R., Fritsch, P. and Krischanitz, R. COMPAC SOLThe New Generation of Easy, Safe, and Fast Heat-Up No Cement
Castables. RHI Bulletin. 2010, No. 1, 1317.
[2] Fritsch, P., von der Heyde, R. and Krischanitz, R. COMPAC SOLOperational Experiences With the Easy, Fast Heat-Up No Cement Castable.
RHI Bulletin. 2011, No. 2, 4245.
[3] von der Heyde, R., Krischanitz, R., Hall, D. and Zingraf, E. COMPAC SOLThe Success Story Continues With Gunning Mixes and New Product
Developments. RHI Bulletin. 2012, No. 2, 1216.
[4] Schtz, J., Maranitsch, A. and Blajs, M. New Oxycarbide Refractory Products Demonstrate Outstanding PropertiesFirst Practical Results.
RHI Bulletin. 2012, No 1, 1619.
[5] Hochegger, M., Blajs, M., Nonnen, B. and Zottler, P. First Practical Results With COMPAC SOL A100S-15A High-End Sol-Bonded Castable
Designed for EAF Deltas. RHI Bulletin. 2013, No. 1, 1419.
[6] von der Heyde, R., Taha, H. and Sadek, W. The Keys to Improvement. World Cement. 2013, 44, No. 8, 7376.
Authors
Ren von der Heyde, RHI AG, Industrial Division, Mlheim-Krlich, Germany.
Roland Krischanitz, RHI AG, Industrial Division, Vienna, Austria.
Milo Blajs, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Roland Krischanitz, [email protected]
16 <
RHI Bulletin > 2 > 2013, pp. 1719
Figure 1. Destroyed brick lining at the retaining ring of a rotary kiln. Figure 3. Veitscher-Magotteaux system (four steps).
Figure 2. Rectangular retaining ring design. Figure 4. Veitscher-Magotteaux system showing a 40 mm brick
override (arrows) in the direction of the kiln outlet.
> 17
RHI Bulletin > 2 > 2013
RHI Thrust Lock System brand termed ANKRAL TLS was created for this special
application. Moreover, due to the raw material basis com-
The challenging target was to design a new system that
prising the purest sintered synthetic magnesia and a special
would provide additional benefits compared to already
spinel concept, ANKRAL TLS offers excellent resistance
existing concepts as well as the possibility to upgrade cur-
against corrosion and clinker melt infiltration. Depending
rent retaining designs. Instead of the common brick and
on the installation position and working conditions, high-
rectangular retaining ring design, a special skew brick set is
alumina bricks or SiC bricks can also be used.
used in the RHI Thrust Lock System (Figure 5). Due to the
skew brick inclination (>> 20), the contact face between
Continuing the basic RHI Thrust Lock System idea, a slightly
retaining elements and brick lining is much higher and modified skew brick set (Figure 7) can be installed mid-kiln,
therefore the resulting contact pressure is considerably directly in front of an outlet segment, or additionally as part
reduced. In addition, the axial lining thrust is diverted into of an existing Veitsch-Magotteaux system.
radial and circumferential forces and consequently mechan-
ical stress in the bricks is further reduced.
(c)
Figure 5. New RHI Thrust Lock System installed at a retaining Figure 6. Three-dimensional FEM models of the (a) Veitsch-
ring, comprising special skew brick sets (cream and brown) and Magotteaux system (one step 20), (b) common brick and rectan-
an expansion joint (white). gular retaining ring design, and (c) RHI Thrust Lock System
(skew brick 40). Dark green indicates tensile stresses and light
green to dark red depicts the increasing compressive stresses.
18 <
RHI Bulletin > 2 > 2013
Figure 7. New RHI Thrust Lock System installed mid-kiln. The first trials are planned for the second half of 2013. Tech-
nical support, investigation of current wear mechanisms,
process analysis, and installation supervision will be pro-
Conclusion vided by RHIs experts.
After decades of struggling to cope with lining problems in
the outlet zone, RHI is able to provide a promising concept A patent application for the new RHI Thrust Lock System
applicable for all rotary kilns. has been filed.
References
[1] Marschall, H.U. and Wiry, A. Computer-Aided Thermo-Mechanical Stress Modelling of Different Rotary Kiln Outlet Designs. RHI Bulletin. 2004,
No. 2, 1216.
[2] Landthaler, A. Drehrohrofen. German Patent DE 2643412 C3, 1987.
Authors
Andreas Wiry, RHI AG, Industrial Division, Vienna, Austria.
Hans Ulrich Marschall, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Andreas Wiry, [email protected]
> 19
RHI Bulletin > 2 > 2013, pp. 2024
Thomas Drnek
The data regarding fatalities does not indicate such a strin- In summary, the significant improvements realized in South
gent development for all three companies. While the Africa are due to the Chamber of Mines of South Africa being
improvement at Anglo American shows a very significant very active in the area of safety, with the mining industry acting
two-thirds reduction over 5 years, decreasing from close to 45 as the most important supporting factor. In recent years multi-
to around 15 fatalities, the values for Rio Tinto peaked in 2008 ple aims have been defined, very positive progress has been
due to a helicopter crash in which 10 exploration crew died. achieved, and results are reported on a regular basis [2,3].
2.0 50
n BHP Billiton n BHP Billiton
1.8 45
Lost time injuries/2 x 105 working hours
1.2 30
1.0 25
0.8 20
0.6 15
0.4 10
0.2 5
0.0 0
2006 2007 2008 2009 2010 2011 2005 2006 2007 2008 2009 2010 2011
Year Year
Figure 1. Annual injury frequency rates for three large mining Figure 2. Annual fatalities in three large mining companies [1].
companies [1].
20 <
RHI Bulletin > 2 > 2013
Table I. MSHA coal mining industry daily fatality report on July 5, 2012, for the USA. Abbreviations include underground (UG) and
surface (S) mining [5].
0.35 80
0.30 70
Fatalities/1 x 106 worked hours
60
0.25
Absolute number
50
0.20
40
0.15
30
0.10
20
0.05 10
0.00 0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2001 2003 2005 2007 2009 2011
Year Year
Figure 3. Annual mining-related fatality rates in South Africa [2]. Figure 4. MSHA annual mining-related fatalities in the USA [4].
> 21
RHI Bulletin > 2 > 2013
conditions in mines is a primary task for all miners? Having The aim of CORESafety is to reduce fatalities to zero and
identified these two factors, it appears the underlying causes decrease the injury rate by 50% within 5 years.
are both related to the following education issues:
In addition to these initiatives, the US Assistant Secretary
>> Miners not having enough experience. of Labor for Mine Safety and Health, Mr. Joe Main, called
>> Mining engineers being insufficiently educated. for the creation of incentives so operators improve safety
practices as well as stronger civil and criminal penalties
The reason for an overall decrease in experience amongst to act as a powerful deterrent against mine operators
US miners is that the average age has increased from 35 knowingly or persistently putting the lives of miners at
years old in 1980 to 47 in 2008 [6]. As a result of retirement, risk [10].
the experience gained by senior miners is gradually being
lost, and safe behavioural practices are no longer being Safety in Europe
transferred from experienced miners to inexperienced
When the situation in Europe was examined, it showed
young miners. In addition, after only six months new miners
that the number of accidents and fatalities in the mining
can now take an exam and upon successful completion are
sector is very low compared to other industries (Figure 5).
considered sufficiently experienced to work on their own
Unfortunately, these are the only overall European mining
anywhere in a mine.
statistics available; however, they do reveal the severity of
Regarding the second issue, due to high salaries offered by the accidents that have occurred in the mining sector,
the industry to graduates with a bachelors degree as well since the percentage of fatalities is fivefold higher than
as high tuition costs at US universities, fewer young engi- that of injuries [11].
neers are willing to study for masters and doctoral degrees.
This has resulted in excessive aging of university staff, in Sweden and Austria
addition to mining authorities and research institutions fac- To provide a more in-depth assessment of the situation,
ing a shortage of postgraduate employees. Sweden and Austria were examined individually, since
they both have a vital mining industry and very good sta-
CONSOL EnergyAbsolute ZERO Value System tistics are available [12,13]. In both countries the lost day
A positive example in the USA of a mining company that injury frequency rate (LTIFR) (i.e., number of occupational
has been very effective in creating a safe working environ- injuries per 1 x 106 working hours) shows a general decline
ment is CONSOL Energy [7], where 97% of the more than for both countries (Figures 6), namely in the last decade
7500 employees worked in 2007 without a single accident the rate of 25 injuries per 1 million worked hours
occurring. The company lists three principal factors that are decreased to 7.7 and 12 in Sweden and Austria, respec-
required to achieve such figures [6]: tively. Furthermore, the LTIFR in Austria reduced signifi-
cantly from 70 injuries per 1 million worked hours in 1970
>> Strict compliance with all safety regulations. Although, to approximately 25 in 2000. The reason for this trend was
these provisions alone are generally not sufficient and better equipment, stricter legislation, good education, and
effective. a general improvement in public health. The number of
>> Application of all available technologies even beyond the fatalities is in the range of 0 to 1 per year, with Austria
mandatory legal requirements, including mine gas moni- having 1 in 2010 and 2011, and Sweden having 1 in 2002,
toring and wireless communication systems. 2007, and 2009. In both countries there were zero fatalities
>> A corporate culture where the top priority is safe, in the other years.
accident-free behaviour in the workplace. Employees
assume a central role in this culture and are responsible Mining Education in Austria
for their own safety and health. The employees are not In part, the relatively low number of injuries and fatalities
only authorized, but explicitly required to immediately in Austria can be attributed to the training and education
correct any and all unsafe working conditions and if opportunities available for mining personnel in this coun-
necessary stop the production process in order to rectify try, which are divided into:
any defects. Regular training sessions and formal evalua-
tions of the safety behaviour also contribute towards >> Miner.
creating an active zero accident culture among all >> Foreman.
employees [7]. >> Mine manager.
22 <
RHI Bulletin > 2 > 2013
After the basic course has been completed, one or more spe- University level education is subdivided into bachelors and
cial courses can be taken that focus on different topics (e.g., masters degree programmes. In addition, doctoral studies
open pit, underground, and mineral processing). The special can be performed. Most students in Austria graduate with
courses comprise the same number of lessons as the basic a masters degree. A bachelor programme (7 semesters) is
course [15]. In Autumn 2013, the Berg- und Httenschule HTL offered at the University of Leoben under the title: Mineral
in Leoben also started to offer a 5 year course focused on raw Resources Engineering. After completing this B.Sc., a mas-
materials for pupils typically aged 1519 years old [16]. ters programme (3 semesters) can be undertaken.
Construction
Manufacturing
Education
0 5 10 15 20 25 30
Figure 5. Percentage of serious and fatal accidents at work in Europe, according to economic activity (total cases: 3942999, fatalities:
4898) [11].
30 80
70
No. of injuries/1 x 106 working hours
25
60
20
50
15 40
30
10
20
5
10
0 0
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 1970 1973 1987 1996 2000 2002 2004 2006 2008 2010 2011
(a) Year
(b) Year
Figure 6. Annual lost day injury frequency rates in the (a) Swedish [12] and (b) Austrian mining industry [13].
> 23
RHI Bulletin > 2 > 2013
This is more focused and there are two areas available: Summary and Recommendations
Mining and Tunnelling, and Raw Materials Engineering [17].
The analysis of mining safety data revealed there are no
comprehensive global statistics available. While the Interna-
Safety in China tional Labour Organization (ILO), a United Nations agency,
Currently, no statistical data is available regarding mining provides some information, the data regarding injuries and
safety in China. The only information is from media reports fatalities is incomplete. Large mining companies have very
(e.g., Xinhua) and the Government website [18]. The official good data and show very positive progress with their aims
number of fatalities reported in coal mines for 2010 was to achieve zero injuries in the workplace. In contrast, analy-
1920 (2436 in 2009); however, it is assumed that the number sis of the available government data revealed a mixed situ-
of unreported cases is considerably higher, since a Chinese ation, with some governments providing access to very
miner is 100 times more likely to die in an accident than a good data and others that do not have or want to publish
US miner [19]. such statistics. However, in general from the large mining
company and specific country statistics there appears to be
The Chinese government is committed to improving the sit- a positive trend in the area of safety. One key factor influ-
uation and a safety tax was imposed in 2012. This equates encing safety is good training and education for all person-
to ~ US$4.8/tonne of coal and funds are planned to be used nel working in the mining industrystarting with miners
to improve the safety of facilities, increase safety education, and ending at executive management level.
set up health damage prevention programmes, build under-
ground shelters, and monitor equipment that is a major In conclusion, the mining community must pay even more
danger source. The results of these measures have not attention to safety issues, and also promote more detailed
been published to date, but the impact could be enormous training and education opportunities for all mining person-
because of the huge revenue available from such a taxation nel. Additionally, the International Organizing Committee of
scheme. the World Mining Congresses needs to place far more
attention on safety issues. Therefore, it was very positive to
see several sessions at the 23rd World Mining Congress
(Montreal, Canada), in August 2013, were focused entirely
on safety.
References
[1] Annual reports and Sustainability reports for Rio Tinto, BHP, and Anglo American (20072011).
[2] http://www.bullion.org.za/documents/ar_2011-small.pdf
[3] http://www.bullion.org.za/documents/AR_2012_web_Final.pdf
[4] www.msha.gov
[5] http://www.msha.gov/stats/charts/coaldaily.asp
[6] Brune, J. Occupational Safety and Health in US American Mines Problems and Potential Solutions. World of Mining - Surface and Underground.
2012, 64, No. 6, 382389.
[7] http://www.consolenergy.com/about-us/corporate-responsibility/safety.aspx
[8] www.nma.org
[9] http://www.coresafety.org/index.html
[10] http://www.dol.gov/ocia/congressionaltestimony/20100520_JMain.htm
[11]http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Health_and_safety_at_work_statistics
[12] http://www.svemin.se/MediaBinaryLoader.axd?MediaArchive_FileID=f4416fc2-1c6a-4276-bf1f-1cfc1245459f&FileName=%C3%85rsrapport+201
1engelsk.pdf
[13]http://www.bmwfj.gv.at/EnergieUndBergbau/SicherheitImBergbau/Documents/Sp%C3%B6rker,%20Unfallstatistik.pdf
[14] http://www.htl-leoben.at/files/spezialkursprogramm2013_1_1.pdf
[15] http://www.htl-leoben.at/index.php?article_id=69
[16] http://www.htl-leoben.at/index.php?article_id=162
[17]http://starter.unileoben.ac.at/fileadmin/shares/starter/docs/Studienbrosch%C3%BCre_13_Web.pdf
[18] www.gov.cn
[19] http://www.bbc.co.uk/news/business-11497070
Author
Thomas Drnek, RHI AG, Raw Materials Division, Breitenau, Austria.
Corresponding author: Thomas Drnek, [email protected]
24 <
RHI Bulletin > 2 > 2013, pp. 2526
> 25
RHI Bulletin > 2 > 2013
Furthermore, the write-off and lost production costs would The study showed the product carbon footprint directly cor-
account for an additional 40 billion and it would result in related with the energy intensity, which mainly depended
absolutely unprofitable running costs. on the kiln load. It also revealed that environmental require-
ments, such as mandatory post process waste gas combus-
Therefore, while the refractory industry will significantly tion, can increase the energy intensity by up to 30%.
contribute to resource and energy efficiency, it is not feasi-
ble to the extent outlined in the EU 2050 roadmap. Since The European refractory industry is committed to continu-
refractories will continue to be indispensible for down- ing this exercise since it provides the possibility to reveal
stream users in the future and enable resource and energy emission-related trends. In addition to increasing the num-
efficient production in other sectors, a regulatory framework ber of product groups covered by the analysis, future aims
that enhances European competitiveness, supports sustain- include the use of national rather than European conversion
able growth, and avoids the risk of carbon and job leakage factors.
must be considered the primary target.
References
[1] http://ec.europa.eu/clima/policies/roadmap/
[2] http://www.roadmap2050.eu/attachments/files/Volume1_fullreport_PressPack.pdf
[3] A Roadmap for Moving to a Competitive Low Carbon Economy by 2050. Communication From the Commission to the European Parliament,
the Council, the European Economic and Social Committee and the Committee of the Regions. COM(2011) 112 final, Ed.; EU-Commission,
Brussels, 2011.
[4] http://www.cerameunie.eu/en/news/european-ceramic-industry-launches-2050-roadmap-paving-the-way-to-a-better-future
[5] Paving the Way to 2050. The Ceramic Industry Roadmap. Cerame-Unie, Brussels, November 2012.
[6] http://ec.europa.eu/europe2020
[7] http://ec.europa.eu/europe2020/europe-2020-in-a-nutshell/flagship-initiatives/index_en.htm
[8] http://www.pre.eu/
[9] PRE Product Carbon Footprint Report. European Refractories Producers Federation, Brussels, September 2013.
Authors
Franz Maier, RHI AG, Technology Center, Leoben, Austria.
Clare McFarlane, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Franz Maier, [email protected]
26 <
RHI Bulletin > 2 > 2013, pp. 2730
2250
2200
Temperature [C]
2150
2100
2050
2045 5 C
2000
1950
0 20 40 60 80 100 0.9 0.8 0.9 3.6
Al2O3 Cr2O3
Composition [mol.%] Notch depth [cm]
Figure 1. Phase diagram showing the continuous miscibility of Figure 2. Static plate corrosion test results. RHIs SUPRAL RK50S
Al2O3-Cr2O3 [1]. and SUPRAL RK70 chrome corundum grades (containing 53 wt.%
and 72.5 wt.% Cr2O3 , respectively) were compared to isostatically
pressed chrome and fused cast AZS material for 144 hours at
1500 C in soda-lime glass.
> 27
RHI Bulletin > 2 > 2013
cast AZS with 40% ZrO2 , whereas the difference between More recently, chrome-bearing material with > 30 wt.%
the different chrome-containing materials was minor: The chromium oxide has been installed in soda-lime glass fur-
uniaxially pressed chrome corundum grades were nearly as naces. The material is only used in zones with high wear
resistant to corrosion as the isostatically pressed product. rates, for example the weir wall in container glass furnaces
(Figure 3), doghouse corners, and the throat. Since 2004,
At several positions in glass furnaces, good thermal shock RHI has also supplied uniaxially pressed chrome corundum
resistance is essential. This is a weak point of isostatically grades with a chrome content of 3070 wt.% for about 40
pressed high chrome material. Thermal shock resistance is soda-lime glass furnaces worldwide. A closer look at the
tested by heating up cylindrical samples to 950 C and then weir walls has shown that chrome corundum greatly
quenching them in water. Isostatically pressed products extends their lifetime, when compared to the use of fused
only survive a few such cycles, 2 or 3 in general. cast AZS.
Due to good structural flexibility, chrome corundum mate- Weir walls significantly influence the glass melt flow profile,
rial exhibits a much better thermal shock resistance. The supporting homogenization. At the same time the wall
addition of small amounts of zirconia-based phases further retains impurity particles, which could otherwise lead to
improves this characteristic. These grains have a different glass defects. Thus, chrome corundum weir walls contribute
thermal expansion behaviour compared to chrome corun- to longer production periods with high quality glass.
dum and this creates a zone of microcracks in their vicinity.
The microcracks can stop the growth of larger cracks, which Batch material is fed into the furnace through the dog-
would otherwise destroy the material. This effect is called house. Due to the abrasive action of unmelted batch as well
microcrack reinforcement. Results of a thermal shock test as thermal shock, the doghouse corner blocks are especially
comparing various chrome-containing refractory grades are subjected to significant corrosion. When completely worn,
listed in Table I. The excellent thermal shock resistance of the doghouse corners must be extensively repaired.
chrome corundum enables it to be installed even as hot Chrome corundum doghouse corners (Figure 4) fulfil the
repair overcoating tiles behind corroded soldier blocks on
container and float glass tanks.
Practical Experiences
Today, chrome-bearing refractory grades are used when
excellent corrosion resistance is required. For example, fur-
naces for the production of C-glass are entirely lined with
chrome corundum materials, permitting a furnace lifetime
of 4 or more years. Since 1980, RHI has also supplied SUPRAL RK50S
chrome corundum grades to insulation fibre producers and
over the last 12 years about 70 furnaces worldwide have
been equipped entirely or in major sections with these
materials.
SUPRAL RK30S
Table I. Thermal shock resistance (TSR) of selected chrome- Figure 4. Doghouse corner in a container glass furnace compris-
bearing refractory grades. ing two chrome-bearing refractory grades, SUPRAL RK30S and
SUPRAL RK50S.
28 <
RHI Bulletin > 2 > 2013
n REFEL 1240
n DURITAL RK50S
10000 n DURITAL RK30S
Electrical resistivity [ohm/cm]
n Chrome 95
n Soda-lime glass
1000
100
10
1
400 600 800 1000 1200 1400 1600
Temperature [C]
Figure 5. Electrical resistivity of DURITAL chrome corundum Figure 7. Wear of AZS overcoating tiles (white) in the foreground
grades compared to iso-chrome, fused cast AZS, and soda-lime and chrome corundum overcoating tiles (black) in the back-
glass melt. ground, both after 2 years application.
(a) (b)
Figure 6. (a) significantly corroded AZS soldier blocks and (b) installation of chrome corundum overcoating tiles providing hot repair.
> 29
RHI Bulletin > 2 > 2013
25
Installation of the overcoating tiles
20
Cr2O3 [ppm]
15
10
0
6 months 9 months
Summary
Chrome corundum materials from RHI exhibit excellent
characteristics for different applications in the glass
production process. By using fused raw materials in various
DURITAL and SUPRAL brands (i.e., RK30S and RK50S), the
properties are improved further because the fusion process
ensures homogeneous mixed crystals with a well-defined
Al2O3 /Cr2O3 ratio.
Reference
[1] Bunting, E.N. Phase Equilibrium in the System Cr2O3 -Al2O3. Bur. Standard. J. Research. 1931, 6, 947949.
Authors
Markus Dietrich, RHI AG, Industrial Division, Wiesbaden, Germany.
Stefan Postrach, RHI AG, Industrial Division, Wiesbaden, Germany.
Corresponding author: Markus Dietrich, [email protected]
30 <
RHI Bulletin > 2 > 2013, pp. 3135
1600
Crs
0
NC2S3
Standard Silica
1400
0 0 NC3S6
1500
110 100
Silica bricks are commonly used by the glass industry. The
Trd
N2CS3 900
good corrosion resistance at temperatures of > 1470 C, a NS2
NS 870
800
> 31
RHI Bulletin > 2 > 2013
The brick from near the throat had a lower Na2O content at
the hot face (0.4 wt.%) in comparison to the centre and cold
face (1.3 wt.%). This was a result of alkali infiltration from (a) (b)
the hot face towards the cold face. Due to dissolution of
wollastonite, the CaO content at the hot face was low
Figure 2. Silica crown bricks after 14 years application in an oxy-
(0.7 wt.%). Whilst melt containing SiO2 , CaO, and Na2O had fuel fired container glass melting furnace. Brick from (a) near the
dripped down into the glass bath, part of this melt would throat and (b) near the doghouse.
also have migrated towards the cold face, resulting in the
CaO content being even higher in the centre (3.5 wt.%) than
at the cold face (2.5 wt.%) where the CaO content was com-
parable to that of an unused standard silica brick (2.5 wt.%).
Enrichment of CaO in the brick centre has also been
reported in the literature [12]. The main mineral phase in all
samples was cristobalite, which indicates the brick tempera-
ture had become higher than 1470 C during service. The
brick was more porous (24 vol.%) at the hot face than in the
centre (16 vol.%), as a consequence of melt dripping out at
the bricks hot face. Scanning electron microscopy of the
brick hot face clearly revealed sharp edges caused by the
melt dripping off and silica grains surrounded by melt in
the remaining areas (Figure 3).
In the case of the brick from near the doghouse, the Na2O 500 m
content at the hot face was relatively high (2.5 wt.%). Since
the temperature of this brick was lower than the brick from Figure 3. Scanning electron micrograph of the hot face of the sil-
near the throat, the Na2O had not infiltrated further in the ica brick located near the throat.
Table I. Chemical composition, phase analysis, and apparent porosity of the silica crown bricks from near the throat and doghouse
after 14 years service in an oxy-fuel fired container glass melting furnace. Determination of the oxides by X-ray fluorescence analysis
after ignition at 1050 C (1) and on the original sample (2).
32 <
RHI Bulletin > 2 > 2013
direction of the cold face. Consequently the Na2O content in Sample 1 represented no-lime silica with 1 wt.% Na2O, cor-
the centre and at the cold face was low (0.7 wt% and 0.3 responding to an approximate situation that can occur at
wt.%, respectively). At the hot face the CaO level was low the brick hot face during service (see Figure 4). Sample 2
(1.0 wt.%) because the newly formed SiO2 , CaO, and Na2O was standard silica also containing 1 wt.% Na2O, providing
containing melt had dripped down into the glass bath. It a direct comparison to Sample 1. Sample 3 was standard
was evident that enrichment of CaO in the direction of the silica containing 2 wt.% Na2O, which is a realistic composi-
cold face had not occurred, since the CaO content at the tion for the hot face of a standard silica brick during ser-
centre and cold face (2.4 wt.% and 2.3 wt.%, respectively) vice (see Figure 4).
corresponded to CaO levels in an unused silica brick. Due
to the lower operating temperature in this region, the main The thermodynamic calculations were performed over an
mineral phase in the brick was tridymite. appropriate temperature range and the different phase
amounts (wt.%) were plotted against the specific tempera-
No-Lime Silica ture (Figure 6). In the case of Sample 1 (STELLA GNL)
there was hardly any CaO, and therefore only SiO2 solid
As described, the corrosion process affecting standard sil-
phases (trydimite up to 1470 C and cristobalite above
ica bricks in oxy-fuel fired furnaces is initiated by alkali
1470 C) and melts with two different compositions were
vapour causing dissolution of wollastonite in the bonding
formed. The wollastonite (CaOSiO2 ) that was present in
matrix [13]. The resulting melt then plays an important role
standard silica dissolved at 1310 C (Sample 2) or 1215 C
in the further corrosion of SiO2 . Since a modified bonding
(Sample 3).
system without wollastonite eliminates this weak point in
silica bricks, a silica brick without limeSTELLA GNLwas
In Figure 6 the exact amount of each melt is indicated as a
developed by RHI.
percentage on the curves. For example, at 1400 C Sample 1
(STELLA GNL) had 6.7 wt.% of Melt 1 and 3.1 wt.% of Melt
In a field trial, standard silica and STELLA GNL bricks were
2, a total melt amount of 9.8 wt.%. However, at the same
installed in the superstructure of an oxy-fuel fired container
temperature (1400 C) the total amount of melt in Sample 2
glass furnace. After 3 months the Na2O diffusion was
(standard silica with 1.0 wt.% Na2O) was 16.0 wt.%, and in
measured in these samples. The results showed that Na2O
Sample 3 (standard silica with 2.0 wt.% Na2O) it was even
diffusion into the standard silica was significantly higher
higher at 24.2 wt.%. At 1500 C, all the samples had a
than into STELLA GNL (Figure 4).
higher amount of melt than at 1400 C, but the STELLA GNL
2.5
n Standard silica
n STELLA GNL
2.0
Na2O in silica [wt.%]
1.5
1.0
0.5
0.0
0 10 20 30 40
Distance from hot face [mm] 500 m
Figure 4. Na2O diffusion into standard silica and STELLA GNL Figure 5. Light microscopy of a cut section through the STELLA
bricks after a 3 month field trial in the superstructure of an oxy- GNL brick installed for 3 months in the superstructure of an oxy-
fuel fired container glass furnace. fuel fired container glass furnace. No significant attack at the hot
face was visible.
> 33
RHI Bulletin > 2 > 2013
6.1
90
6.7 Fused Cast AZS
8.0 When oxy-fuel technology was first introduced, fused cast
85 AZS was used for crown and superstructure applications
because standard silica materials showed increased corro-
80 sion under the operating conditions. However, a technical
SiO2(S4) disadvantage of fused cast AZS is that exudation of the
75 SiO2(S5) glassy phase occurs during operation and is greater in the
case of oxy-fuel firing. This can be clearly observed from
70 the results of a trial that compared fused cast AZS and
1000 1100 1200 1300 1400 1500 1600 STELLA GNL in an oxy-fuel tableware glass furnace. After 3
months in the superstructure the fused cast AZS showed
(a) Temperature [C]
significant exdudation at the hot face while STELLA GNL
Sample 1 Total melt amount had only a slightly wetted surface (Figure 7). Additionally,
No-lime silica STELLA GNL 1400 C: 9.8 wt.% there was higher sodium infiltration into the fused cast AZS
containing 1 wt.% Na2O 1500 C: 13.4 wt.% than into the STELLA GNL (Figure 8). Whilst the risk of exu-
dation can be avoided by using fused cast alumina instead
of fused cast AZS [14,15], STELLA GNL is an economic
100
alternative to fused cast alumina.
Na2Ca3SiO15(S)
95
6.9 Melt 1
Phase distribution [wt.%]
9.2
90 CaSiO3(S) CaSiO3(S2)
13.2
85 16.0
19.3
80
SiO2(S4)
75
SiO2(S5)
70
1000 1100 1200 1300 1400 1500 1600
(b) Temperature [C]
Sample 2 Total melt amount
Standard silica 1400 C: 16.0 wt.%
containing 1 wt.% Na2O 1500 C: 19.3 wt.%
Figure 7. (a) fused cast AZS and (b) STELLA GNL after a 3 month
field trial in the superstructure of an oxy-fuel tableware glass fur-
100 nace.
Melt 2
4.7
95
Na2Ca3SiO15(S)
Phase distribution [wt.%]
8.2
8
90 10.5 Melt 1
n AZS32
7 n STELLA GNL
CaSiO3(S2)
85
Na2O in samples [wt.%]
17.4 6
CaSiO3(S)
80 20.4 5
24.2 4
75 SiO2(S4)
27.7 3
SiO2(S5)
70
2
1000 1100 1200 1300 1400 1500 1600
(c) Temperature [C] 1
Figure 6. Results of the thermodynamic calculations performed Figure 8. Na2O diffusion into fused cast AZS and STELLA GNL
for Samples 13, defined in Table II, using FactSage (Version after a 3 month field trial in the superstructure of an oxy-fuel
6.3). tableware glass furnace.
34 <
RHI Bulletin > 2 > 2013
Conclusion
Corrosion resistant refractories for crown and superstruc-
ture applications in oxy-fuel fired soda-lime glass melting
furnaces have undergone continuous improvement. The
first generation materialstandard silicawas replaced by
a second generationfused cast products. However, the
new generationno-lime silicacombines the advantages
of standard silica such as the low weight, no exudation, and
economic price with high corrosion resistance.
References
[1] Kobayashi, H. Advances in Oxy-Fuel Fired Glass Melting Technology. Presented at XX International Congress on Glass, Kyoto, Japan. Sept.,
26Oct., 1, 2004.
[2] Gridley, M. Philosophy, Design and Performance of Oxy-Fuel Furnaces. 57th Conference on Glass Problems: Ceram. Eng. Sci. Proc. 1997, 18,
Issue 1, 114.
[3] Ibbotson, A., Hunter, J., Brown, J.T. and McMahon, A.K. The Development of a Large Oxy-Fuel Glass Melter and an Assessment of its
Performance and Potential. Presented at European Seminar on Improved Technologies for the Rational Use of Energy in the Glass Industry,
Wiesbaden, Germany, Feb., 46, 1992.
[4] Faber, A.J. Corrosion of Refractories Under Oxy-Fuel Combustion. Glass Production Technology International. 1997, 6166.
[5] LeBlanc, J. Impact of Silica Attack on Soda Lime Oxy-fuel Furnaces. Presented at Ceramic Industry Manufacturing Conference and Exposition,
Pittsburgh, USA, Oct., 11, 1995.
[6] Godard, H.T., Kotacska, L.H., Wosinski, J.F., Winder, S.M., Gupta, A., Selkregg, K.R. and Gould, S. Refractory Corrosion Behaviour Under Air-Fuel
and Oxy-Fuel Environments. 57th Conference on Glass Problems: Ceram. Eng. Sci. Proc. 1997, 18, Issue 1, 180207.
[7] Duvierre, G., Zanoli, A., Boussant-Roux, Y. and Nelson, M. Selection of Optimum Refractories for the Superstructure of Oxy-Fuel Glass Melting
Furnaces. 57th Conference on Glass Problems: Ceram. Eng. Sci. Proc. 1997, 18, Issue 1, 146163.
[8] RHI Product Information: STELLA GNLA New Silica Brick Without Lime Bonding. Glasstec Exhibition 2006, Dsseldorf, Germany.
[9] Mulch, C., Weichert, T. and Postrach, S. Neuer keramisch gebundener Werkstoff fr den Oberbau von oxy fuel beheizten Glaswannen auf SiO2-
Basis. Presented at Httentechnische Vereinigung der Deutschen Glasindustrie e.V., Meeting, Niederdollendorf, Germany, March 2006.
[10] van Limpt, J.A.C. Modeling of Evaporation Processes in Glass Melting Furnaces, Ph.D., Thesis, Eindhoven University of Technology, Netherlands,
2007.
[11] Morey, G.W. and Bowen, N.L. High SiO2 Corner of System Na2O-CaO-SiO2. J. Soc. Glass Technol. 1925, 9, 232233.
[12] Faber, A.J. and Verheijen, O.S. Refractory Corrosion Under Oxy-Fuel Firing Conditions. 57th Conference on Glass Problems: Ceram. Eng. Sci.
Proc. 1997, 18, Issue 1, 109119.
[13] Beerkens, R., Verheijen, O., van Limpt, H. and Smits, B. Evaporation of Alkali Compounds and Silica Corrosion in Glass Furnaces. Presented at
HVG-NCNG Colloquium on Operation and Modelling of Glass Melting Furnaces, Aachen, Germany, Nov. 8, 2005.
[14] Monofrax Alumina Fused Cast Crowns. Monofrax Report No. 1, 1997.
[15] Winder, S.M., Selkregg, K.R. and Gupta, A. Update on Selection of Refractories for Oxy-Fuel Glass-Melting Service. Ceram. Eng. Sci. Proc., 1999,
20, No. 1, 81105.
Authors
Rongxing Bei, RHI AG, Industrial Division, Wiesbaden, Germany.
Klaus Santowski, RHI AG, Technology Center, Leoben, Austria.
Christian Majcenovic, RHI AG, Technology Center, Leoben, Austria.
Susanne Jrg, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Rongxing Bei, [email protected]
> 35
RHI Bulletin > 2 > 2013, pp. 3639
Introduction
Flue gases in oil or gas fired soda-lime glass melting
furnaces consist mainly of fossil fuel combustion prod-
ucts such as SOx (in the case of oil firing), CO, CO2 , HCl,
NOx , O2 , and N2. Additional components in the flue gas
can be SO2 , stemming from the refining agent, and
NaOH, mainly originating from the glass melt [1]. Under
oxidizing operating conditions, volatiles like sodium and
SO3 condense in the so-called condensation zone of
the glass melting tank regenerator, resulting in the final
formation of Na2SO4. This condensation zone occurs
in the temperature range of approximately 7001100 C
(Figure 1). In the case of oxidizing operating conditions
and excess SO3 , aggressive free sulphate is formed
besides sodium sulphate. This resulted in the develop-
ment of forsteritic-bonded magnesia-zircon bricks [2],
which provide the highest corrosion resistance against
sulphate attack.
36 <
RHI Bulletin > 2 > 2013
The effect of reducing conditions on the waste gas in the microscopic investigations were carried out on polished
condensation zone is that SO3 is not stable and as a result sections by optical light microscopy and using a JEOL 6400
sodium sulphate cannot form in the condensation zone. As scanning electron microscope equipped with an energy
a consequence, aggressive sodium remains in the flue gas dispersive spectrometry (EDS) analysis system to provide
[1]. Since sodium has a high reaction potential with SiO2- chemical microanalyses.
rich brick materials, pure, low-iron magnesia bricks are an
appropriate lining recommendation for the regenerator Results and Discussion
condensation zone operating under reducing conditions. In
Magnesia-Zircon BricksOxidizing Atmosphere
sodium silicate glass melting tanks, particularly in the case
and High Sulphate Load
of gas firing where there is a very high load from NaOH in
the flue gas, pure magnesia bricks show the best perfor- Forsteritic-bonded magnesia zircon brick brands like
mance. RUBINAL EZ and RUBINAL VZ are a suitable lining material
for regenerators operating with an oxidizing atmosphere
In summary, depending on the furnace atmosphere, suit where there is an excess of sulphate in the flue gas. One
able and well-proven refractory lining recommendations important microscopic feature of this brick type is that ini-
can be made. However, in cases where during operation tial magnesia and zircon brick components completely
the atmosphere within the regenerator heating phase react during the firing process to form magnesia, forsterite,
changes drastically from the usual oxidizing to reducing and zirconia. Well-selected raw materials that form a dense
conditions and vice versa, serious chemothermal attack on microstructure with homogeneously distributed compo-
the previously well-performing refractory lining can arise. nents, a strong forsteritic bonding structure in the brick,
In extreme cases, the regenerator lining lifetime can and the development of a forsteritic belt protecting the
decrease drastically. magnesia provide the microstructural basis for a satisfac-
tory lining life. Due to its acidic character, forsterite does
The following examples provide a comprehensive view of not react with sulphate. This has been confirmed in several
the microstructural changes that occurred when bricks postmortem studies of samples, not only from the glass
were subjected to a drastically changed atmosphere in the industry, where magnesia with Ca2SiO4 in the interstitial
phase is corroded by sulphate to form merwinite, monticel-
regenerator. The samples were investigated microscopi-
lite, and finally the stable reaction product forsterite [1,6,7].
cally and compared to unused material.
Figures 2 and 3 provide impressive microstructural exam-
ples of a forsteritic-bonded magnesia-zircon checker brick
Investigation Procedure that achieved a lining lifetime of 17 years in the condensa-
To generate results with the highest possible quality and tion zone despite very high sulphate load. Even in the cen-
accuracy levels, sample investigations were carried out tre the SO3 load was 6.77 wt.% with a molar alkali/SO3 ratio
according to international standard procedures. For exam- of 0.32. At the immediate surface there was magnesia cor-
ple calibration was performed using internationally certi- rosion due to sulphate attack (see Figure 2) involving the
fied standards. Na was analysed by inductively coupled formation of Na-Mg-sulphate. However, the forsteritic
plasma optical emission spectrometry (ICP-OES) (ISO bonding matrix was stable and almost unaffected. At a
26845), which covers the range from parts per billion to depth of 7 mm there was Na-Mg-sulphate infiltration and
high wt.%. Sulphur determination (DIN 51085) using LECO densification of the open pores but no bonding structure
CS 200 covered a range of 030 wt.% sulphur in a refrac- corrosion and only very minor corrosion of the forsterite
tory sample with an accuracy of 0.1 wt.%. The protected magnesia grains (see Figure 3).
4
2
demonstrated with a postmortem sample that had become
affected when the atmosphere was severely changed by the
operating conditions. In this extreme example a 6.8 wt.%
1 2
SO3 content with a molar Na2O/SO3 ratio of 0.31 was found
even in the centre of some bricks. Generally, preferential
corrosion of the Ca2SiO4 interstitial phase as well as MgO
2
can occur. Sulphate attack in combination with a certain
2 1
amount of sodium and MgO results in the formation of Ca-
3 sulphate, Ca-Na-sulphate, Ca-Na-Mg-sulphate, and forsterite
(Figure 6). Consequently, there is also corrosion of the MgO
bonding structure. As a result of MgO participating in the
sulphate attack, the formation of idiomorphic, cubic shaped
500 m
3 100 m
Figure 2. Magnesia-zircon checker brick after 17 years in service. Figure 3. Microstructural detail 7 mm from the surface of a mag-
At the near surface, up to a maximum of 5 mm from the hot nesia-zircon checker brick after 17 years in service. No corrosion
face, corrosion of magnesia (2) due to the formation of a sul- of the forsteritic bonding matrix (1). The forsteritic belt (arrows)
phate phase (1) is visible. The forsteritic matrix shows very high protecting the magnesia grains (2) and Na-Mg-sulphate in the
resistance against sulphate attack (3) and pores (4) are indicated. pores (3) are indicated.
> 37
RHI Bulletin > 2 > 2013
Magnesia-Zircon BricksReducing Conditions demonstrated with a postmortem sample that had become
and Sodium Attack affected when the atmosphere was severely changed by the
operating conditions. In this extreme example a 6.8 wt.%
The following example is a magnesia-zircon brick sample
SO3 content with a molar Na2O/SO3 ratio of 0.31 was found
that was exposed during a laboratory trial to a reducing
even in the centre of some bricks. Generally, preferential
atmosphere containing an excess of sodium compared to
corrosion of the Ca2SiO4 interstitial phase as well as MgO
sulphur. The temperature was cycled from 800 C to 1100 C
can occur. Sulphate attack in combination with a certain
and back to 800 C. Firing during the 30 minute heating
amount of sodium and MgO results in the formation of Ca-
phase was under reducing conditions (35% CO) whilst the
sulphate, Ca-Na-sulphate, Ca-Na-Mg-sulphate, and forsterite
30 minute cooling period was under oxidizing conditions.
(Figure 6). Consequently, there is also corrosion of the MgO
A total of 96 cycles were performed. The forsterite reacted
bonding structure. As a result of MgO participating in the
with the supplied sodium to form Na-Mg-silicate (Figure 4).
sulphate attack, the formation of idiomorphic, cubic shaped
This resulted in a corroded, softened bonding structure,
MgOpericlasecan also occur (Figure 7). This can be
especially near the surface. Under strongly reducing condi- explained by the slight thermal changes and related MgO
tions the forsteritic bonding structure in magnesia-zircon saturation limit changes in the salt melt. Higher thermal
bricks is corroded by sodium attack, which increases with load results in an increased ability for MgO to dissolve in
the exposure time and begins at the brick surface. In the melt. When this MgO saturated sulphate melt cools,
extreme cases even ZrO2 can be corroded to form Na-zirco- MgO precipitates, forming idiomorphic, cubic crystals
nate or Na-Zr-silicate. within the sulphate.
2
1
2
3 1
100 m 50 m
Figure 4. Magnesia-zircon brick sample that had undergone a Figure 6. Pure magnesia brick, approximately 1.5 mm from the
laboratory corrosion test where the heating and cooling phases hot face brick, after sulphate attack caused by an oxidizing
were under alternating reducing and oxidizing conditions, respec- atmosphere. Some residual MgO bonding bridges remain.
tively. 96 cycles were performed between 800 C and 1100 C. MgO (1), Na-Mg-(K)-Ca-sulphate (2), Ca-sulphate (3) are indicat-
Mainly sodium attack occurred. Newly formed Na-Mg-silicate ed. SEM-EDX analysis of point 2 detected 11.6 wt.% Na2O, 18.7
(1) minimal Na-sulphur phases (2), corroded forsteritic matrix wt.% MgO, 66.0 wt.% SO3, 1.2 wt.% K2O, and 2.5 wt.% CaO.
(3) with ZrO2 (white), MgO (4), and pores (black) are indicated.
38 <
RHI Bulletin > 2 > 2013
Conclusion
There are well-proven solutions to line the regenerator con-
2 densation zone in glass melting furnaces that are selected
according to the operating conditions such as the fuel type
and whether the waste gas atmosphere is oxidizing or
reducing. The microstructure of the various refractories is
1 optimized for the special operational requirements in this
critical lining zone. However, when during the heating
phase the usual atmosphere in the regenerator is drastically
changed, the flue gas also significantly changes its corro-
sive potential. Atmosphere changes can occur for instance
1 1
in the case of operating problems or due to NOx reduction
measures taken during the life cycle of an existing regener-
ator lining. Therefore, a possible impact on the existing
glass tank regenerator lining and its performance must be
50 m considered. In extreme cases drastically lowered lining per-
formance can occur.
Figure 7. Pure magnesia brick after sulphate attack caused by an
oxidizing atmosphere. Precipitation of idiomorphic, cubic shaped
MgO (1) from the sulphate melt (2) is visible.
References
[1] Heilemann, G., Schmalenbach, B., Weichert, T., Postrach S., Lynker, A. and Gelbmann, G. New Solutions for Checkers Working Under Oxidizing
and Reducing Conditions. 67th Conference on Glass Problems: Ceram. Eng. Sci. Proc. 2008, 28, Issue 1, 183194.
[2] Weichert, T. and Schmalenbach, B. Use and Further Development of Magnesia-Zircon Bricks in the Glass Industry. 55th Conference on Glass
Problems: Ceram. Eng. Sci. Proc. 1995, 16, Issue 2, 6873.
[3] Kido, N. Recent Problems in Glass Melting from the View Point of Refractory and its Solutions. Advanced Materials Research. 2008, 3940,
593600.
[4] Schmalenbach, B. and Riepl, K. Impact of Reducing Atmosphere on the Corrosion of Refractories in Regenerators of Glass Melting Tanks. Proc.
UNITECR 01, Cancun, Mexico, Nov., 48, 2001; pp. 11351141.
[5] Hughes, D.E. Melting Glass With Reduced NOx Emissions. United States Patent 4,328,020, 1982.
[6] Bartha, P. and Sdje, J. Degradation of Refractories in Cement Rotary Kilns Fired With Waste Fuels. CN Refractories. 2001, 5, 6271.
[7] Jrg, S., Majcenovic, C., Liprandi, F., Mura, R. and Nievoll, J. Outstanding Performance of ANKRAL ZE at the Caravate Plant, Colacem S.p.A., Italy.
RHI Bulletin. 2007, No. 2, 3134.
Authors
Christian Majcenovic, RHI AG, Technology Center, Leoben, Austria.
Gerald Gelbmann, RHI AG, Technology Center, Leoben, Austria.
Rongxing Bei, RHI AG, Industrial Division, Wiesbaden, Germany.
Corresponding author: Christian Majcenovic, [email protected]
> 39
RHI Bulletin > 2 > 2013, pp. 4041
Engineering Material
n Basic engineering n Production
n Detailed engineering n Packaging
nE ngineering plus n Transportation
(modelling and simulation) n Warehousing
40 <
RHI Bulletin > 2 > 2013
the job start, which enabled daily cost monitoring and com-
parison of the actual performance against budgeted
resources.
Customer Satisfaction
The feedback from Chemtrade was very positive, with state-
ments including RHI can work on this site anytime, as they
demonstrate safe, professional work ethics, and are forth-
coming with immediate creative solutions. An additional
Figure 2. Sulphur recovery unit. comment from Chemetics was RHI has shown us a new
avenue that we will certainly use for our future projects.
accounts for equipment rental and third-party purchases,
while Controlling established RHI US Ltd., business/contrac- RHI will promote its Modular Solution System to the global
tor licenses and created workers insurance cover. New EEC, and cement and lime markets through RHI Canada,
roles and responsibilities were added to current positions in RHI US Ltd., and Marvo operations, with the objective of
Canada regarding estimates, customer sales representa- securing large capital projects, multi-year maintenance con-
tives, and accounts receivable. RHIs tailored Project Track- tracts, and providing all key components for a value-added
ing System (PTS) was uploaded with project details prior to service.
Authors
Joseph Plunkett, RHI Canada Inc., Industrial Division, Sarnia, Canada.
Friedrich Schweighofer, RHI Canada Inc., Industrial Division, Burlington, Canada.
Corresponding author: Joseph Plunkett, [email protected]
> 41
RHI Bulletin > 2 > 2013, pp. 4243
Figure 1. Fluid catalytic cracking unit: A typical application area Figure 2. Qualification test according to API 936: Gunning a test
where API certified personnel are required. panel.
42 <
RHI Bulletin > 2 > 2013
References
[1] http://www.techstreet.com/products/1587645
[2] http://www.api.org/certification-programs/individual-certification-program-icp
Authors
Joseph Plunkett, RHI Canada Inc., Industrial Division, Sarnia, Canada.
Dieter Pirkner, RHI AG, Industrial Division, Wiesbaden, Germany.
Corresponding author: Joseph Plunkett, [email protected]
> 43
RHI Bulletin > 2 > 2013, pp. 4449
Introduction ASTM C704M - 09 in 2009. In this update the blast gun and
venturi system dimensions were defined in more detail and
Refractory resistance against abrasion is a key issue for
a new clause was added that describes known factors that
many industrial furnace applications where particle-loaded
affect the results. The findings in this clause are based on
gas jets are expected during the process. For example, they
a ruggedness test performed using float glass samples.
occur in furnaces for the petrochemical industry, waste
incineration, coal-fired power plants, and in steel reheating
pusher furnaces.
Pressure gauge
44 <
RHI Bulletin > 2 > 2013
Statistically significant effects were reported if the inside the designated abrasion value of 9.3 cm3 0.3 cm3 for this
diameter of the inlet air nozzle and the air pressure were material, the standard air pressure is altered, which is con-
not as defined in the standard. In the latest revision ASTM trary to the established standards where the pressure is fixed
C704M - 12, published in 2012, precision of the apparatus at 4.5 bar. Using the new method, the apparatus setup details
setup was further defined. Supplementary requirements for are of secondary importance. A number of tests were per-
highly abrasion-resistant materials were also added. The formed to examine various aspects of this approach in detail,
major task of this clause is to more precisely describe the including a round robin test that clearly demonstrated cali-
testing conditions that are attained, for example by mount- brating the system on the basis of a defined abrasion value
ing additional pressure gauges. It is evident from the inclu- provides reliable data to examine abrasive wear of refractory
sions that the ASTM Committee C08 is heading toward a materials.
more rigid definition of the apparatus dimensions, espe-
cially the particle acceleration system. Preliminary Examinations
Prior to the round robin test, investigations were undertaken
Main Factors Influencing Abrasion to understand how variable air pressures and different
During the Test amounts of abrasive material affect the results. In addition,
The refractory material volume loss is dependent principally an alternative material (i.e., B4C) was used for the nozzle
on the impact energy that is discharged when highly accel- tube.
erated SiC particles hit the sample surface. It is proportional
to both the particle velocity and weight (i.e., grain size). Air Pressure Influence
Therefore, only narrow SiC grain distributions are suitable In this preliminary test performed at two laboratories, boro-
for the test. It is also necessary to examine the grain size silicate glass samples from a European producer with a bulk
distribution within one batch if disintegration effects are density of 2.22 g/cm3 were abraded at predefined air pres-
obvious. In addition, the particle surface plays a role. The sures between 3 and 6 bar. All the other parameters were
particles should have angular, jagged edged surfaces and kept constant and in accordance with the procedure
therefore only be used once. The general grain shape can described in ASTM C704M - 09.
be checked by measuring the bulk density.
The results show there is a linear ascending correlation
Many factors affect the particle velocity during acceleration. between the volume loss and increasing air pressure over
Of major importance is the standard air pressure provided the measured range (Figure 2). However, the measurements
to the venturi system and both standards specify 4.5 bar. from the two laboratories have a different slope, indicating
However, particle acceleration inside the blast gun is not a slightly different setup of the testing devices. For example,
only dependent on the pressure but also on the chamber if an abrasion value of 7.4 cm3 is required, the air pressure
dimensions inside the blast gun and the specific dimen- must be adjusted to approximately 5.50 bar at Laboratory
sions of the venturi nozzle. The latter is described in detail A and 4.56 bar at Laboratory B.
in ASTM C704M - 09. However, for as long as ASTM C704
has existed, a simple but essential problem has been dis- Influence of the Abrasive Material Amount
cussed among those applying the standard outside the US: and Nozzle Tube Material
The specified blast gun is not available in Europe. More
In the subsequent evaluation, borosilicate float glass samples
over, since 2009 the ASTM has stated that the described
from the European producer detailed above and standard float
blast gun is the only one permitted.
glass samples as recommended by ASTM C704M - 09 were
abraded with different amounts of SiC (i.e., 8001200 g).
After the particles have been accelerated in the venturi
chamber, they pass through a glass tube with strictly
defined dimensions (i.e., length and diameter). Due to vari
able surface roughness, the specific material used to manu-
facture the tube can cause different friction effects and may 10.0
decelerate the SiC grains to a greater or lesser degree dur- n Laboratory A
9.0 n Laboratory B
ing their passage through the tube. Since 2007, these and
additional factors known to affect the results have been
8.0
listed in ASTM C704, but without any quantification. y = 2.1255x - 2.2325
Abrasion [cm3]
R2 = 0.99439
7.0
In summary, the evolving standards have defined the test
conditions and the apparatus setup more and more strictly 6.0
y = 1.5113x - 0.8143
in order to get better interlaboratory reproducibility. How- R2 = 0.98207
5.0
ever, as a result it has become more difficult to assure the
apparatus achieves the predefined conditions and the 4.0
method has become more error-prone. Therefore, this
paper presents an alternative solution that is much easier to 3.0
perform and leads to highly reliable results. It proposes 3 4 5 6
deregulating the apparatus setup specifications but defining Air pressure [bar]
a preliminary step and adjusting the gas pressure to
achieve defined abrasion of a float glass test specimen,
Figure 2. Linear relationship between borosilicate glass plate vol-
described as the calibration standard in ASTM C704. This
ume loss (cm3) and the air pressure (bar) used to accelerate the
material, broadly used as a reference standard, is particu- abrasive particles. The results were derived in two independent
larly suitable because it is highly homogenous. To obtain laboratories.
> 45
RHI Bulletin > 2 > 2013
Two sets of tests were performed, one with conventional deeper pits with a smaller eroded area at the surface, namely
glass nozzle tubes and a second with abrasion resistant B4C the blast jet appeared to be more focused. However, the
nozzle tubes. Each test was repeated twice with the same slope of all the tests was very similar. Therefore, it can be
tube. All other parameters were kept constant and in accord- concluded that the nature of the blast jet is of minor impor-
ance with the procedure described in ASTM C704M - 09, for tance for reproducibility of the method, if the air pressure or
example the predefined air pressure was 4.5 bar. All tests the amount of abrasive material is the adjustable parameter.
were performed using the same abrasion tester.
Interlaboratory StudyRound Robin Test
In all the tests performed, the abrasion loss showed a strict
As previously described, the performance and reproducibility
linear correlation with the amount of SiC used over the
of the standard test methods according to ASTM C704 or ISO
measured range (Figure 3). Therefore, the amount of SiC
16282 are highly dependent on the particle size and velocity
can also be used to standardize the abrasion tester. For this
before they hit the sample surface. Therefore, it is very diffi-
approach, the mass of SiC has to be adjusted until the mea
cult to reproduce the precise test conditions defined in the
sured abrasion value is equivalent to the standard value of
standards. As the preliminary test results described above
9.3 cm3 0.3 cm3, as defined in ASTM C704M - 09 for the clearly show, the abrasion loss is significantly and systemati-
reference float glass plate. cally controlled by the air pressure, which generates the SiC
particle acceleration in the venturi system. The round robin
The linear correlation between the amount of SiC used and test was performed to examine the interlaboratory reproduci-
the abrasion level indicates that materials with a low abra- bility of the test when performed according to ASTM C704M
sion resistance (> 20 cm) can be tested using a decreased - 09 and how the reproducibility is affected if prior to testing
amount of SiC (e.g., 500 g instead of 1000 g). This would the air pressure is adjusted until the abrasion value of a
reduce the possibility of blasting through the sample. Fur- standard ASTM float glass sample is 9.3 cm3 0.3 cm3.
thermore, abrasion values obtained using a different amount
of SiC can be used to determine the equivalent values for
Round Robin Test Setup
1000 g SiC.
In the round robin test, 6 different shaped and monolithic
In addition, Figure 3 clearly demonstrates that the actual refractories were examined in 10 laboratories. The materials
abrasion values measured are influenced by the nozzle tube were selected with the aim of covering a wide range of
material used. For example, the B4C tubes typically led to abrasion values (i.e., 530 cm3). Table I details the refractory
higher abrasion loss than the conventional glass tubes. How- materials tested.
ever, it was also evident that B4C tubes increased the repro-
ducibility of individual measurements since the results from
Sample Bulk density Refractory type Abrasion
repeat tests had the same linear regression. In contrast, the
(g/cm3) (cm3)
results obtained using glass tubes showed variations of up
A 2.15 Fireclay FC40 brick 30
to 2% for the same test conditions. As a result of the stan
dard glass tube erosion after a single test, this nozzle type B 3.00 Alumina-chromia ACr80/5 brick 15
must be changed after every test whilst the B4C nozzles CD 3.15 Magnesia-chromite MCr50 brick 7
have the advantage that they can be used approximately E 3.00 Sintered magnesia based brick 15
100 times because the erosion rate is very low. F 2.20 Conventional refractory concrete 8
G 2.80 Ultra low cement castable 5
When the erosion pits were examined after the test it was
evident that the blast jet slightly differed according to the Table I. Overview of the refractory materials examined in the
tube type. For example, the B4C tubes typically produced interlaboratory test, including the approximate abrasion value.
9.0
tation technique (IET) and ultrasonic measurements were
also performed in three dimensions. The results of these
8.0
tests enabled the dynamic Youngs modulus to be calcu-
7.0 lated and compared (Figure 4). Determining this material
6.0 property was important because the abrasion resistance is
highly dependent on the homogeneity of the test samples.
5.0
This is exemplified by the series of fireclay brick FC40 sam-
4.0 ples (i.e., refractory sample A) that were examined in one
800 900 1000 1100 1200
laboratory and show that the abrasion resistance is strongly
SiC [g] dependent on the sonic velocity (Figure 5). A scatter in the
sonic velocity of 10% for any refractory sample set was
Figure 3. Linear relationship between the volume loss (cm3) and
amount of SiC abrasive (g) accelerated using a constant air pres-
regarded as admissible for the round robin test.
sure of 4.5 bar. Standard float glass (A/continuous lines) and boro-
silicate glass (E/dashed lines) samples were tested using standard A preliminary result at this stage in the investigation was
glass nozzle tubes (S/blue lines) and B4C nozzle tubes (N/red lines). that determining the sonic velocity is a powerful tool to
All the conditions were measured twice without changing the
tube. The repeat measurements determined with B4C tubes were evaluate the abrasion resistance of a single refractory mate-
very close; therefore, only a single regression line is visible. rial type, for example material A.
46 <
RHI Bulletin > 2 > 2013
For the round robin test, every laboratory was instructed to air pressure was adjusted to attain an abrasion value of
follow the same procedure. Initially, a test run was performed 9.3 cm3 0.3 cm3 with ASTM standard glass samples.
with the standard ASTM float glass plates using the test con-
ditions defined in ASTM C704M - 09 (i.e., at a fixed air pres- In summary, 10 laboratories examined 6 different refractory
sure of 4.5 bar). The measurements were repeated multiple materials under standard conditions and with the adjusted
times. Since the samples were highly homogeneous, the air pressure. Typically, three glass plates were necessary to
results obtained were particularly useful to examine the inter- adjust the air pressure to within the tolerance range. In addi-
laboratory reproducibility. Duplicate refractory samples (see tion, the standard glass plates were tested at least 4 times.
Table I) were then examined under the standard test condi- The results were evaluated with the PROLab Plus (QuoData
tions. Subsequently, a second series of duplicate tests was GmbH, Germany) software, which is in accordance with ISO
performed on standard ASTM float glass plates and the six 5725-6:1994. The results and precision statistics are detailed
different refractory materials. However, prior to the tests the in Table II.
Table II. Results of the round robin test conducted at 10 laboratories on the 6 refractory samples detailed in Table I and standard float
glass samples. The abrasion tests were conducted according to ASTM C704M - 09 or using the modified test procedure where the air
pressure was adjusted so a standard float glass sample had an abrasion value of 9.3 cm3 0.3 cm3. Abbreviations include standard
deviation of reproducibility (SR), standard deviation of repeatability (Sr), reproducibility interval (R), and repeatability interval (r).
90 3300
Median
Dynamic Youngs modulus [MPa]
80 25%75% 3200
70 Range without
Sonic velocity [m/s]
outlier
3100
60 Outlier
50 3000
40
2900
30
2800
20
10 2700
A B C D E F G 22 24 26 28 30 32 34 36 38
Refractory type Abrasion [cm3]
Figure 4. Box plot of the dynamic Youngs modulus determined Figure 5. Approximate linear relationship between the abrasion
using the impulse excitation technique. The results are shown loss and sonic velocity for six different fireclay brick FC40 sam-
for the six different refractory materials detailed in Table I, name- ples (i.e., refractory type A) within the range defined.
ly fireclay FC40 brick (A), alumina-chromia ACr80/5 brick (B),
magnesia-chromite MCr50 brick (C and D), sintered magnesia
based brick (E), conventional refractory concrete (F), and an ultra
low cement castable (G).
> 47
RHI Bulletin > 2 > 2013
Standard Glass Samples The refractory material abrasion results from the round
robin test are summarized in Table II. Whilst the data
The round robin abrasion results for the standard ASTM
shows that calibrating the equipment by adjusting the air
float glass material are shown in Figure 6. The mean
pressure only slightly enhances the repeatability (Sr)
empirical value was 9.294 and the tolerance limits were
within laboratories (i.e., 7.54% compared to 8.19%), it is
fixed at 8.620 and 9.992 cm3, which corresponds to a
clearly evident that this procedure significantly decreases
Zu-score of < 2.00. With a single exception, only the abra-
the standard deviation between laboratories (SR). Across
sion values determined after calibrating the equipment by
all the tested samples, the average relative standard devia-
adjusting the air pressure fell within the statistical tolerance
tion between laboratories was more than halved from
limits. The scatter of results in a single laboratory is indi-
cated by the boxes. Laboratory 5 showed the widest scat- 31.48% to 14.81%. Furthermore, the results clearly show
ter. On further examination, it transpired that this labora- that the abrasion testers are capable of producing reliable
tory had specific problems with fluctuations in the air pres- and consistent results.
sure supply that were solved by installing a pressure ves-
sel. In summary, when the standard glass material abrasion If only the results of the standard glass plates are consid-
results from all the laboratories were combined it was clear ered, the enhancement achieved by adjusting the air pres-
that calibrating the equipment by adjusting the air pressure sure is profound. While the standard deviation for all the
significantly reduces the scatter of results (Figure 7). sample types between laboratories was improved by a fac-
tor of 2, the highly homogenous glass plates really dem-
onstrate the improvement, namely the relative SR was
0 12 decreased almost 20 fold from 28.25% to 1.72%. This
clearly indicates that the gas adjustment method can pro-
11 duce very precise results and that the heterogeneity
between refractory samples is the limiting factor.
Tolerance threshold (9.992 cm3)
10
Average
Abrasion [cm3]
0
L1 L2 L3 L4 L5 L6 L2 L1 L3 L10 L5 L9 L7 L8 L9 L7
Conclusions
st st st ad ad ad ad ad ad ad st ad ad ad st st The round robin test was performed using a broad range
Laboratory
of shaped and unshaped refractory materials that had
diverse blast abrasion resistances (i.e., 530 cm3). The
Figure 6. Round robin abrasion results for standard ASTM float
glass material. The tests were performed at 10 laboratories major outcome of this work is that to provide interlabora-
according to ASTM C704M - 09 with a fixed air pressure (st) or tory comparability it is better to adjust the air pressure of
using the modified test procedure where the air pressure was the venturi system and calibrate the equipment with
adjusted so a standard float glass sample had an abrasion value
of 9.3 cm3 0.3 cm3 (ad). The triangles indicate measured
respect to the abrasion rate of a standard float glass sam-
results. ple rather than define rigid regulations for the apparatus
setup as recommended by ASTM C704M - 12. When the
air pressure was adjusted in this manner the overall stan
dard deviation between laboratories was reduced by a fac-
0 tor of 2 when refractory materials were tested (i.e., SR
Median
25%75% decreased from 31.48% to 14.81%) and by a factor of 20 if
12
Range without outlier only the results of the standard glass samples are consid-
Outlier
ered (i.e., SR decreases from 28.25% to 1.72%). The poorer
Abrasion value [cm3]
11
precision regarding the refractory materials can be attri
10 buted to the heterogeneity inherent in the samples. Sonic
velocity measurements prior to the abrasion test can help
9
identify outlier samples. The SR can also be positively
influenced by multiple measurements due to enhanced
8
repeatability.
7
Over the measured range, the refractory material abrasion
0 loss showed a linear correlation with the air pressure and
Standard Adjusted the amount of abrasive used. Therefore, these parameters
can be adjusted over a broad range without a significant
Figure 7. Box plot of all the standard glass sample measure- decrease in the precision of the method. However, it is
ments performed at the 10 laboratories according to ASTM important to note that the air pressure supply should be
C704M - 09 with a fixed air pressure (standard) or using the checked carefully. Temporary pressure fluctuations caused
modified test procedure where the air pressure was adjusted
so a standard float glass sample had an abrasion value of by an insufficient air supply will cause poor repeatability
9.3 cm3 0.3 cm3 (adjusted). and deliver in most cases lower values.
48 <
RHI Bulletin > 2 > 2013
Authors
Olaf Krause, University of Applied Sciences Koblenz, Hhr-Grenzhausen, Germany.
Gerhard Urbanek, RHI AG, Technology Center, Leoben, Austria.
Hartmut Krber, Deutsches Institut fr Feuerfest und Keramik GmbH, Hhr-Grenzhausen, Germany.
Corresponding author: Gerhard Urbanek, [email protected]
> 49
RHI Bulletin > 2 > 2013, pp. 5053
Klaus Gamweger
50 <
RHI Bulletin > 2 > 2013
longitudinally down the mid section of the furnace bottom patented solution to this problem is a slidable, air-cooled pip-
towards the endwalls and through the steel shell to the ing system (Figure 2).
outside.
The angled pipe connection from the plug is inserted into a
The main problem with implementing such a gas purging protection pipe, which provides an air-tight seal and is
system in a tilting or stationary furnace is the difference affixed in the permanent lining. The space between the pro-
between the thermal expansion of the wear lining and the tection pipe and the surrounding brickwork is filled with ram-
permanent lining, which becomes more and more critical the ming mix. Since the angled pipe from the plug can slide
larger the furnace. The bricks in the wear lining must be very within the larger protection pipe, this compensates for any
tightly connected to each other to prevent metal penetration plug movement caused by thermal expansion (Figure 3). The
in between the bricks. Therefore, the refractory layout in the nitrogen supply is realized with a flexible metal hose. Addi-
furnace is designed to provide significantly higher thermal tionally, the plug bottom is cooled to minimize the risk of
expansion in the wear lining than in the permanent lining copper leakage into the pipe system as a result of metal pen-
during heat-up and operation. Usually the bottom lining etration between the bricks in this area. For this purpose, the
moves from the centre of the furnace bottom towards the purging plug is equipped with a double-layered steel bottom.
endwalls, where the expansion is partially compensated. Compressed air is connected to the plug using a second flex-
ible metal hose and is blown between the two steel plates,
The purging plug and well block are installed in the wear lin- cooling the lower plug area and surrounding region. The
ing whilst the system piping is located below in the perma- cooling effect has been visualized in computational fluid
nent lining. Since the plug piping crosses the boundary layer dynamics (CFD) simulations, as shown in Figure 4. In situa-
between the wear and permanent lining, high tensions are tions where the piping is long and the cooling air line could
introduced that can cause damage to the welding or even the overheat, an additional compressed air line is incorporated
pipes. In this case, a nitrogen leak would occur at this spot, within the protection pipe to facilitate cooling. Due to the
subsequently resulting in the plug becoming infiltrated with overall cooling effect, the temperature underneath the plug
copper. Therefore, the relative movement of the two lining falls below a critical value so if metal infiltration occurs, the
layers has to be compensated to prevent pipe damage. The copper freezes and a molten metal leak is avoided.
Double-layered
steel bottom
Protection pipe
Compressed air
Thermocouple
Nitrogen supply
Figure 2. Conical purging plug with well block, supporting brick, Figure 3. Purging plug gas piping system.
and gas connections.
Temperature [C]
30 600 1200
Figure 4. CFD simulation. Air-cooling of the plug bottom, surrounding region, and pipe.
> 51
RHI Bulletin > 2 > 2013
Online thermocouples monitor temperatures at two posi- melting rate is an inherent result of the metal movement.
tions in the system. The first measures the temperature However, the impact on oxidation and reduction processes
inside the purging plug, which indicates plug wear, whilst requires more detailed examination. Especially in tilting
the second records the bottom plate temperature, monitor- anode furnaces, with a high number of tuyeres in the side-
ing the air-cooling efficiency. wall, an interfering effect from the plugs has to be avoided.
Purging Plug Exchange Figure 5 shows the difference in flow velocity with and with-
out the use of a gas purging system during reduction in a
The lifetime of purging plugs used for this application is
tilting anode furnace. The red areas indicate the most rapid
between 18 months and 3 years. In some cases the furnace
lining has a longer durability so plug replacement may be melt flow, with the highest reaction kinetics occurring in
necessary. Due to the modular setup it is possible to replace these regions. In these areas the most material exchange
a worn out or blocked plug without renewing a large area occurs, the poling effect is very high, and a shorter reduc-
of the furnace hearth; only about one square metre has to tion time is necessary to get these areas to react with the
be removed and rebuilt. Furthermore, the piping remains tuyere gas. Therefore, these zones have to be maximized by
untouched in the permanent lining. This can be easily exe- appropriately positioning the purging plugs and adjusting
cuted during periodic repair to the tuyere zone. the gas flow rate.
Velocity [m/s]
0.0 0.5 1.0
(a)
(b)
Figure 5. CFD simulation of a tilting anode furnace (a) with and (b) without purging plugs.
52 <
RHI Bulletin > 2 > 2013
Conclusion
The new cooled, slidable piping technology enables gas
purging systems to also be installed in stationary and tilting
furnaces. The system can withstand all expansion-induced
tensions in the furnace lining and has proven to be a safe
and long-lasting solution. It enables all the benefits of inert
gas purging in anode furnaces to be realized and is provid-
ing process improvements and energy savings in multiple
applications worldwide.
References
[1] Rigby, A.J. and Lanyi, M.D. Porous Plugs in Molten Copper Production and Refining. Smelting Operations and Advances: Proceedings of the
Copper 99-Cobre99 International Conference, Phoenix, Arizona, Oct., 1013, 1999; pp. 505516.
[2] Lee, S.S., Kim, B.S. and Choi, S.R. Application of the Porous Plug System in the Anode Furnace at Onsan Smelter. Proceedings of the Yazawa
International Symposium on Metallurgical and Materials Processing Principles and Technologies. Vol. 2: High-Temperature Metal Production, San
Diego, California, March 36, 2003; pp. 447458.
[3] Acuna, C. and Sherrington, M. Efficiency of Porous Plugs in Fire Refining of Crude Copper. Proceedings of the Yazawa International Symposium
on Metallurgical and Materials Processing Principles and Technologies. Vol. 2: High-Temperature Metal Production, San Diego, California, March
36, 2003; pp. 265279.
[4] Gamweger, K. Evaluation of the Efficiency of a Gas Purging System in a Copper Anode Furnace. Proceedings of the Sohn International
Symposium on Advanced Processing of Metals and Materials. International Symposium on Sulphide Smelting, Vol. 8, San Diego, California,
Aug., 2731, 2006; pp.499508.
[5] Zach, O., Gamweger, K., Lukesch, G. and Filzwieser, A. CFD-Modelling in the Non Ferrous Metals Industry. Proceedings of EMC 2005, European
Metallurgical Conference, Dresden, Germany, Sept., 1821, 2005; pp. 13171324.
Author
Klaus Gamweger, RHI AG, Industrial Division, Vienna, Austria.
Corresponding author: Klaus Gamweger, [email protected]
> 53
RHI Bulletin > 2 > 2013, pp. 5458
Sonja Breyner, Klaus Santowski, Thomas Prietl, Alfred Spanring and Anna Franzkowiak
54 <
RHI Bulletin > 2 > 2013
ideal composition was found when the Cr2O3 content was provide fresh material and prevent saturation of the slag
between 1030 wt.% and the zirconia component exceeded with refractory material oxides.
10 wt.%. In addition, > 20 wt.% fused alumina-chromia grain
was added to maintain the corrosion resistance at an appro- After testing, the samples were cut in half and macroscopi-
priate level. cally analysed. Cross sections of the samples can be seen in
Figure 2. The wear and infiltration results measured from
Physical Testing the cross-sectional profiles are presented in Figure 3 and
Table III.
In order to evaluate the material properties of a thermal
shock resistant (SR) alumina-chromia product, the test
results were compared to a standard alumina-chromia
grade based on the same Cr2O3 content.
Sample BD AP CCS
(g/cm3) (vol.%) (MPa)
Standard 3.50 15.0 160
SR 3.47 15.0 140
Table II. Comparison of the SR and standard grade hot physical Figure 1. Rotary furnace test (a) in operation and (b) gas burner.
properties. Abbreviations include refractoriness under load (RuL)
and hot modulus of rupture (HMoR).
> 55
2 cm 2 cm
(a) (b)
Figure 2. Cross section through the (a) standard and (b) SR grade samples after the rotary furnace test.
(a) (b)
Figure 3. Wear (red section) and infiltration (dashed line) graphic measurements of the (a) standard and (b) SR grade samples after
the rotary furnace test.
56 <
RHI Bulletin > 2 > 2013
Figure 4. Cross section through a SR alumina-chromia grade Table IV. Chemical analysis of the SR alumina-chromia grade
after 117 heats in the area surrounding the tuyeres. Samples for after 117 heats in the area surrounding the tuyeres. The tests
chemical analysis (AD) and microscopic polished sections (17) were performed according to DIN 51081 and DIN 51085 on ignit-
are indicated. ed samples (1050 C) by X-ray fluorescence analysis and on orig-
inal samples using C-S elemental analysis, respectively.
100 m 1000 m
(a) (b)
Figure 5. Microstructure of the SR alumina-chromia grade after 117 heats in the area surrounding the tuyeres. (a) Immediate hot face:
Reaction zone with partial slag coating (R) and a thin Fe-Ni-Cu-Zn-oxide rich seam (arrows) around a Cr-corundum grain (1).
(b) 10 mm from the hot face: Crack running parallel to the brick hot face filled with slag and Cu-sulphide (red lines).
5
2
3 1
4 3
2
1 50 m 500 m
(a) (b)
Figure 6. Microstructure of the SR alumina-chromia grade after 117 heats in the area surrounding the tuyeres. (a) 50 mm from the
hot face: Corroded Zr-mullite visible, corundum (1), corroded Zr-oxide (2) with Zr-silicate at rims (3), Cu-sulphide (4), and Cr-corun-
dum in the matrix (5). (b) 80 mm from the hot face: Brick joint (red rectangle) free from infiltration, Cr-corundum (1), Zr-mullite (2),
and Ca-Mg-sulphate (3).
> 57
RHI Bulletin > 2 > 2013
2 1
3
5
1 2
1000 m 10 m
(a) (b)
Figure 7. Microstructure of the SR alumina-chromia grade after 117 heats in the area surrounding the tuyeres. (a) and (b) cold face:
Cr-corundum (1), Zr-mullite (2), Na-Mg-Al-Fe-Cr-sulphate (3), glassy phase (4), and pore (5).
It can be concluded that the alumina-chromia brick infiltra- Summary and Conclusions
tion was comparably lower than in the sintered magnesia-
Alumina-chromia grades underwent further development to
chromite grade. Nevertheless, effects of high chemical and
combine high acidic slag resistance with improved thermal
thermal load on the microstructure (e.g., recrystallization
shock resistance. Whilst the materials were found to exhibit
and mullite degeneration) could be found as far down as
a decrease in bulk density as well as hot modulus of rupture
the middle section of the postmortem sample. In addition,
compared to the standard products, laboratory investiga-
traces of sulphate were detected at the brick cold face. How-
tions demonstrated both improved corrosion and thermal
ever, the average residual thickness of several postmortem
shock resistance of these hydration-resistant materials. To
samples was relatively high compared to the sintered mag-
evaluate this brick type in a highly oxidizing atmosphere
nesia-chromite grade.
combined with significant temperature changes and acidic
slag, a 30% Cr2O3 alumina-chromia grade, RESISTAL
Chemical Analysis
RK30SR, was installed in the tuyere zone of a copper con-
Four sampling positions for the chemical analysis were verter. The field trial results, including detailed chemical
chosen to generate comprehensive data through the post- and mineralogical investigations of postmortem samples,
mortem sample (see Figure 4). The results are provided confirmed that the thermal shock resistant alumina-chromia
in Table IV. It is evident that directly at the brick hot face products can be successfully applied in areas experiencing
there was a loss of Al2O3 (~ 13 wt.%), Cr2O3 (~ 6 wt.%), high thermal stress combined with chemical attack from
and ZrO2 (~ 3 wt.%) when compared to the middle section acidic melts and sulphur. Therefore, these new products will
(125150 mm). In addition, there was a significant supply also be suitable for a number of other application areas in
of Fe2O3 to the hot face (~ 6 wt.%) from the slag. At the cold primary copper smelting, converting, and refining.
face 0.47 wt.% SO3 was detected. This represents a lower
value than usually observed with magnesia-chromite bricks
due to lower penetration into the alumina-chromia brick
microstructure.
References
[1] Emblem, H.G., Davies, T.J., Harabi, A. and Tsantzalou, V. Alumina-Chrome Refractories. Proceedings of UNITECR 91, Aachen, Germany, Sept.,
2326, 1991; pp. 415421.
[2] Fraser, T. M. Development and Application of Chrome-Alumina Refractories. Application of Refractories: Ceramic Engineering and Science
Proceedings, Vol. 9, Issue1/2, 1988; pp. 110.
[3] Davenport, W.G., King, M., Schlesinger, M. and Biswas, A.K. Extractive Metallurgy of Copper. 4th Edition; Elsevier Science Ltd: Oxford, 2002.
[4] Routschka, G. (Ed) Pocket Manual Refractory Materials. 2nd Edition; Vulkan-Verlag: Essen, 2004.
Authors
Sonja Breyner, RHI AG, Technology Center, Leoben, Austria.
Klaus Santowski, RHI AG, Technology Center, Leoben, Austria.
Thomas Prietl, RHI AG, Industrial Division, Vienna, Austria.
Alfred Spanring, RHI AG, Industrial Division, Vienna, Austria.
Anna Franzkowiak, RHI AG, Industrial Division, Vienna, Austria.
Corresponding authors: S
onja Breyner, [email protected]
Alfred Spanring, [email protected]
58 <
RHI Bulletin > 2 > 2013, pp. 5962
(a) (b)
Figure 2. (a) thermomechanical FEA model and (b) furnace represented by this model.
> 59
RHI Bulletin > 2 > 2013
As discussed in a previous issue of the RHI Bulletin [1], the From the formula it is apparent that the von Mises stress
thermomechanical behaviour of refractory materials is very is independent of the mathematical sign of the principal
complex and includes temperature dependencies, plasticity, stresses, implying that tensile stresses (positive sign) and
visco-plasticity, and time-dependent effects like relaxation. compressive stresses (negative sign) are treated the same
The target of the simulation discussed here was to find the way. Based on the theory behind the von Mises stress this
cause of high stresses and to minimize them. Therefore, a is valid for plastic materials like steel. However, brittle
temperature-dependent linear elastic approach for the materials like ceramics, concretes, and refractories are sig-
description of the material behaviour was considered suffi- nificantly more sensitive to tensile stresses than to com-
cient and was chosen knowing that such an assumption pressive stresses, since the tensile strength of refractory
would result in the model predicting much higher stresses materials is typically one-tenth of the compressive
than those expected in reality. However, independent of the strength. Hence the von Mises stress is not appropriate to
absolute stress levels, minimizing them would reduce the describe the failure behaviour of refractory materials,
risk of further failures, even if it is not possible to prove instead the maximum principal stress, the largest tensile
whether failure would be expected. stress, is typically used in a first approach to evaluate the
risk of crack formation.
In the first set of models, the influence of the following
heat-up procedures was investigated: The von Mises stresses in the steel shell after the 72 hour
holding time at maximum temperature are shown in Figure 3.
>> Slow heat-up in 48.5 hours to 1200 C (heating rate The highest stresses in the steel shell were observed in
~ 22 C/hour). the bolts, represented by the red areas. In order to compare
>> Rapid heat-up in 33 hours to 1200 C (heating rate the stresses caused by the three different heat-up proce-
~ 40 C/hour). dures, the average stresses in a bolt were plotted over time
>> Rapid heat-up procedure with holding times used for (Figure 4). As discussed in the previous section, the abso-
refractory monolithics, with a total heat-up time of 58 lute values of the stresses should be considered too high.
hours (final heating rate ~ 40 C/hour). Therefore, the average von Mises stresses in the bolts are
presented relative to the stress level at the end of the slow
All the heat-up procedures were followed by a 72 hour heat-up procedure modelling time.
dwelling time, at a constant temperature of 1200 C.
As shown in Figure 4, prior to the heat-up procedure the
bolts in the model were pretensioned at about 15% of the
Initial Determined Stresses in the Steel Shell
final stresses. At the end of each heat-up procedure (58
The von Mises stress is a stress calculated according to the hour time point on the graph) the stresses were at approxi-
formula below (1) [2], and represents the load on plastic mately 80% of the final values, and continued to increase
deformable materials. It was created to provide a scalar for the next 50 hours after the end of the heat-up proce-
value that can be used to compare the three-dimensional dures. Assuming that the stresses at the end of the heat-up
stress in a structure with the tensile strength of for example were below but close to the tensile strength of the bolts, the
steel. further stress increases provide an explanation for the
observed bolt failure several hours after the end of the heat-
1 up procedure. When the stresses are compared after 120
sV = ( s1 s2 ) 2 +( s2 s3 ) 2 +( s3 s1 ) 2 (1) hours, it is evident that they were almost the same for all
2 procedures. In summary, the models described above were
able to explain why failure occurred several hours after the
In this formula the index V indicates the von Mises stress end of the furnace heat-up and that the rapid heat-up proce-
and the indices 1, 2, and 3 are the three principal stresses. dure was not responsible for the observed damage.
1400 140
Relative averaged von Mises stress in a bolt [%]
1200 120
Hot face temperature [C]
1000 100
800 80
Temperature slow heat-up
600 Temperature rapid heat-up 60
Temperature rapid heat-up
400 with holds 40
Stress slow heat-up
Stress rapid heat-up
200 Stress rapid heat-up 20
with holds
0 0
0 24 48 72 96 120 144
Time [hours]
Figure 3. Von Mises stresses in the steel shell after the 72 hours Figure 4. Development of the relative von Mises stresses (solid
holding time at 1200 C, namely the end of the modelled time lines) in a bolt over time during and after the different heat-up
frame. Relative high stresses are indicated in red. procedures (dashed lines).
60 <
RHI Bulletin > 2 > 2013
Stresses in the Refractory Brick Lining stress-free lining. Furthermore, the unique slag washing
procedure used in this type of furnace results in a very spe-
The observations then raised the question of whether it
cific situation regarding the furnace and refractory lining
would be possible to shorten the heat-up time in order to
design: Not only must it be able to withstand the stresses
minimize furnace downtime and energy consumption.
caused by thermal expansion of the refractory material dur-
However, prior to discussing this point it is necessary to
ing the slag washing procedure at 11001250 C, but a tight
highlight that the model assumed a linear elastic behav-
lining is required after cooling down to operating tempera-
iour of the refractory materials, neglecting relaxation.
tures in the range of 800900 C.
Since the refractory brick relaxation is in part significantly
shorter than the modelled time frame, the comparative
In order to investigate the stresses in the steel shell and
stresses and interpretation of the modelling results are refractory bricks that are expected to occur in these two
valid. temperatures ranges, an additional FEA was performed to
investigate a slow heat-up to 850 C. This simulation indi-
The main reason for a slow furnace heat-up procedure is cated that the stresses after a heat-up to 850 C are about
the requirement to avoid thermomechanical overload of 50% of those generated by the heat-up to 1200 C, high-
the refractory lining bricks. This can result in damage to lighting the challenge regarding furnace design and brick
the brick structure, a reduced furnace lifetime, and in the grade selection for such a procedure.
worst-case scenario in brick spalling during heat-up. In
order to examine the potential risk of the three heat-up
Lining Design to Minimize Potential Failure
procedures causing brick damage, the maximum principal
stresses were modelled for each scenario (Figure 5). To address the final question of how to minimize the risk of
further failures, four different lining designs were
The stresses in Figure 5 are shown at time points during
the three different heat-up procedures when the highest
maximum principal stresses were observed. Those
stresses observed with the rapid heat-up procedure were
about 30% higher than with the slow heat-up schedule.
Therefore, the slower procedure decreases the risk of brick
damage compared to rapid heat-up. The stresses occur-
ring during the procedure with holding times, which
addresses the requirement for drying times in the case of
water-containing refractory monolithics, were in the same
region as those observed with the rapid heat-up without
holding times, indicating that the heat-up rate is the domi-
nant factor creating these stresses.
(a)
> 61
RHI Bulletin > 2 > 2013
compared using FEA. One was the standard refractory lin- The modelling showed that while introduction of expansion
ing modelled in the initial analysis that did not include any allowances between the bricks within specific rings (orange
expansion allowances in the endwalls. Two additional mod- line) resulted in minor stress reduction, including them in
els enabled different expansion allowance possibilities in between the rings resulted in a significant decrease in the
the endwall to be investigated: The first involved including stresses (green line). Based on the experience that the fur-
expansion allowances in between the bricks of specific nace failed a number of hours after the end of the heat-up
rings at the endwall while the second introduced expansion procedure at a stress level exceeding 80%, the reduction of
allowances between the specific rings. For reference, a the stresses in the design with the expansion allowances in
fourth model without any expansion allowances in all areas between the rings to approximately 60% at the end of heat-
of the furnace lining was examined. Development of the rel- up and less than 80% multiple hours later infers a signifi-
ative averaged von Mises stresses in a bolt over time for cant reduction of failure risk with this lining design. A more
the four different lining designs is shown in Figure 6. The detailed analysis of stresses in the different lining designs
plotted stresses are relative to the stresses in the standard confirmed compressive stresses in all directions at the lin-
design (blue line) at the end of the modelling period. The ing hot face, indicating that all expansion allowances were
red line shows the stress development that would be closed.
expected without any expansion allowances.
Summary
180 The thermomechanical FEA of a lead short rotary furnace
Relative averaged von Mises stress in a bolt [%]
References
[1] Marschall, H.U. Finite Element Simulation of RefractoriesThe Challenge. RHI Bulletin. 2012, No. 2, 4144.
[2] Beitz, W. and Kttner, K.-H.(Eds). Dubbel: Taschenbuch fr den Maschinenbau; Springer Verlag: Berlin, 1990.
Authors
Hans Ulrich Marschall, RHI AG, Technology Center, Leoben, Austria.
Daniel Kreuzer, RHI AG, Industrial Division, Vienna, Austria.
Alfred Spanring, RHI AG, Industrial Division, Vienna, Austria.
Corresponding authors: H
ans Ulrich Marschall, [email protected]
Daniel Kreuzer, [email protected]
Alfred Spanring, [email protected]
62 <
RHI Bulletin
The Journal of Refractory Innovations
To facilitate efficient distribution of the RHI Bulletin, amendments or e-mail the required alterations. In
we would be grateful if all recipients could check addition, if you would like to register and receive the
their delivery details are correct and select the pre- RHI Bulletin on a biannual basis, please complete
ferred format. Please use the form below to fax any the form below and return it to us by fax or e-mail.
Title
First name
Surname
Company
Department
Street / no.
Postbox
City (town)
Area code
Country
Phone
Fax
Phone: +43 (0) 502 13-5323, Fax: +43 (0) 502 13-5237, E-mail: [email protected]
RHI Bulletin >2>2013
The Journal of Refractory Innovations