Professional Documents
Culture Documents
I- FA E.
CEMENT I_ A NI 1-
OPERATIONS HANDBOOK
1
Second Edition July 1998
TERNATIONAL
Revie w w
ACKNOWLEDGMENTS
I would like to thank the legions of grey men with whom 1 have worked and from whom 1 have learned over the years. I acknowledge with thanks the help of Mr James Post with whom I worked on the first edition. For their help with this revision, I wish specifically to thank Dr Hung Chen, Mr James Korthals, Mr Mark Mays and Mr Herman Tseng who have reviewed all or part of the manuscript and have made a number of valuable suggestions and corrections. I am particularly grateful to Mr Andrew Jackura for writing the chapter on Reliability and Maintenance and for substantial contributions to other sections. Southdown, Inc has kindly encouraged the revision of this book. The Company has not, however, reviewed the text and is in no way responsible for factual errors or contentious suggestions. As before, the author retains proprietary rights over all expressions of ignorance and opinion. Finally, I am indebted to Mr David Hargreaves of international Cement Review for his continued interest and support in the re-issue of this small work.
PREFACE
The first edition of this handbook appeared to find some use within the Industry and we are encouraged to revise and expand the material. As previously noted, while there are a number of excellent books covering plant design, process engineering, and cement chemistry, there is a fairly sparse literature addressing cement plant operations and little to serve the hapless novitiate. Again for brevity, the objective has been constrained, and whole areas of operations technology and management have been omitted as being inappropriate to address in so limited a compass. It is also appreciated that regulations, specifications, and even operating practices are not universal, and our observations should be discounted accordingly. The scope attempted comprehends:
A consideration only of cyclone preheater kiln technology which comprises more than 80% of world production and virtually all kilns installed since 1970. The use only of metric units. A review of major plant sub-systems with a proposed list of data which should be available to plant and corporate management, and some suggestions regarding problem areas and possible solutions. A summary of cement types and concrete problems. A collection of process formulae. A selection of reference data and notes. An outline of plant assessment and plant valuation. References to review articles and a limited bibliography. Addresses of pertinent organizations.
The assessment of cement plant equipment and operations involves numerous terms and numbers, many of which are prone to varying definitions. We would like to offer the following comments and suggestions towards standardization: Cyclone preheater kiln - There is much confusion in terminology regarding air-suspension preheater kilns with and without secondary firing or precalcination. The term preheater does not distinguish air-suspension from grate preheaters, and preheater is frequently used in distinction from precalciner. Dry process, of course, also refers to long dry kilns and fluidized beds. We would, therefore, advocate "cyclone preheater" and "cyclone preheater with precakiner". Plant capacity - Annual capacity can relate to, various assumptions for kiln operation and for cement intergrinding. A reasonable standard is the designed, or established, daily clinker production assuming 85% annual run factor and 5% cement additives:
Annual cement capacity = Clinker tlday x 365 x 0.85 / 0.95
Some elaboration is still required for the plant which produces 3,000t/day of clinker and 6,000t/day of cement, but it is better understood as a 980,000t/year plant than 1,860,000t/year. Kiln run-factor - Various definitions have been encountered including fire-on time, and running-time exclusive of planned shut-downs. Feed-on time is suggested to be the most significant parameter and should be expressed as a percentage of 8760 hours per year.
This is a divisive subject. Uncomfortably for those of us who have always numbered from the top, it must be asserted that, with the proliferation of cyclone preheaters of other than 4 stages, numbering from the bottom allows more meaningful correlation from kiln to kiln.
-
OBITER DICTUM
Joseph Aspdin of Leeds patented Portland cement in 1824 for a ground mixture of calcined limestone and clay. The designation "Portland cement" was originally due to a resemblance in colour and character to limestone found on the Isle of Portland off the south coast of England. The recognition that high temperature clinkering (C3S formation) greatly improved performance is attributed to Isaac Johnson in 1845. Joseph Ransome patented the rotary kiln in England in 1885 and the air-suspension cyclone preheater was patented by VogelJorgensen in Czechoslovakia in 1932. "In the old days their monstrous red-hot bodies revolved with a cosmic roar and howl, belching hellish flames...and over all the acrid stench of cement." Fyodor Gladkov; Cement (1925)
m
wri
rn
Cement Plant
Operations Handbook v
Publisher: David Hargreaves, International Cement Review Layout & Design: Mary Flack, Paul Benewith Copyright Philip Alsop All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical or otherwise, without the prior permission of Tradeship Publications Ltd. This publication is intended solely for use by professional personnel who are competent to evaluate the significance and limitations of the information provided herein, and who will accept total responsibility for the application of the information. Printed by Bishops Ltd, Portsmouth, United Kingdom Front cover: The CBR cement works in Lixhe, Belgium Photographer: Mr Serge Brison
CONTENTS
Section A Process Summaries
1
3
13
21
5 FINISH MILLING
1 Clinker Storage 2 Finish Milling 3 Separators 4 Cement Storage 5 Cement dispatch
61
1 M7-1WI.X
6 POLLUTION CONTROL
1 Dust Collection 2 Pollution control
75
7 QUALITY CONTROL
1 Chemical Analysis 2 Sampling 3 Physical Testing 4 Process Control 5 Microscopy 6 Cement Strength 7 Cement Types & Specifications 8 Cement Intergrinds 9 ISO 9000 10 Concrete Problems 11 Domestic Water Treatment
85
8 MAINTENANCE
1 Failure Modes 2 Computerized Maintenance Management Systems 3 Reliability Centered Maintenance 4 Cost Management 5 Organization 6 Role, Planning & Control 7 Mobile Equipment Maintenance
107
9 ACCOUNTING
Cost Accounting 2 Investment Justifications 3 Project Cost Estimation 4 Profit & Loss Statements
115
121 125
135 136
139
3 Conveying
1 Comparative power consumption 2 Fuller Kinyon Pump 3 Bucket Elevator 4 Belt Conveyor 5 Screw Conveyor 6 Air-slide 7 Drag Conveyor 8 Tube Belt Conveyor 9 Sandwich Belt Conveyor 10 Pneumatic Capsule Conveyor 11 Water pump
147
4 Milling
1 Sieve Sizes 2 Circulating Load 3 Separator Efficiency 4 Tromp curve 5 Critical Speed 6 Charge Loading 7 Grace Factor 8 Mill Power
151
Handbook ix
1N O
157
6 Fuels Data
1 Coal & Petroleum coke 2 Fuel Oil: 3 Natural Gas.
167
7 Materials Data
1 Bulk Densities 2 Specific Gravities & Grindabilities 3 Coefficients of Linear Expansion.
171
8 Conversion Tables
1 Length: 2 Area: 3 Volume 4 Density 5 Pressure 6 Energy 7 Weight 8 Miscellaneous
172
9 Miscellaneous Data
173
x Cement Plant
Operations Handbook
1 Geometrical & Trigonometrical Formulae 2 Greek Alphabet 3 Atmospheric Pressure 4 pH & Normality 5 Laboratory Reagents 6 Sea Water Composition 7 Elemental Abundance 8 Hardness of Materials 9 World Cement Production 10 Ship Capacities
10 Statistics
1 Standard deviation 2 F Test 3 X2 Test 4 Linear regression
179
183 185
193
14 Plant Assessment Data List 15 Cement Plant Valuation & Construction Cost References
1 Review papers 2 Books 3 Addresses of pertinent organizations
index
217
1 INTRODUCTION
Cement is "a substance applied to the surface of solid bodies to make them cohere firmly" or, more specifically, "a powdered substance which, made plastic with water, is used in a soft and pasty state (which hardens on drying) to bind together bricks, stones, etc in building" (SOED). Portland cement is a calcined material comprising lime and silicates which is mixed with sand and stone and, upon hydration, forms a plastic material which sets and hardens to a rock-like material, concrete. Confusion between cement and concrete is endemic among the uninitiated.
Portland cement is manufactured in a series of processes which may be represented as shown:
Clay Silica Iron
1,14 4
Crusher
Limestone quarry
I
Limestone stockpile
Gypsum
Additives
Clinker
Cement sibs
4114
Shipping
Limestone (calcium carbonate) and other materials containing appropriate proportions of calcium, silicon, aluminum, and iron oxides are crushed and milled to a fine flour-like raw meal. This is heated in a kiln firstly to dissociate calcium carbonate to calcium oxide with the evolution of carbon dioxide, and then to react calcium oxide with the other components to form calcium silicates and aluminates which partially fuse at material burning temperatures up to 1450C. The reaction products leave the kiln as a black nodular material, clinker. The clinker is finally interground with a small proportion of gypsum (to control the rate of hydration) and the fine product is cement.
r-
2 RAW MATERIALS
2.1 Raw Materials
The composition of portland cement varies from plant to plant due both to cement specifications and to the mineralogy of available materials. In general, however, an eutectic mix is sought which minimizes the heat input required for clinkering and the total cost of raw materials, while producing a cement of acceptable performance. An approximate analysis for raw mix on ignited basis, or for clinker, is:
CaO 5102 Al203 Fe203 MgO Mn203 T102 503 K20 Na20 65-68% 21-23% 5- 7% 2-4% 1-5% 0.1-3% 0.1-1% 0.1 -2 % 0.1 -1 % 0.1 -0.5%
Note that, with a substantial proportion of the raw mix being CaCO3, heating either in a kiln or in a laboratory furnace evolves some 35% by weight as CO2; this results in a requirement of approximately 1.5t of raw materials to produce 1t of cement, and also requires that analytical data be clearly distinguished between "raw" and "ignited" basis. Cement mixes vary from "cement rock", a single component which, as mined, contains appropriate proportions of the required minerals, to 4 or 5 component mixes comprising one or two grades of limestone, a shale or clay, and one or more additives to augment Si02, Al203 or Fe203 levels. Kiln feed typically contains 78-80% CaCO3 so that limestone can only fall close to this level to the extent that it also contains the other ingredients. It is essential to have sufficient flux (Al, Fe, Mg, F) to promote fusion in the kiln, but MgO should not exceed 4-5% or the cement may be expansive. Excess alkalis (K, Na) affect both kiln operation (build-ups) and product quality (alkali-aggregate reactivity). Excess S causes kiln build-ups and limits the addition of gypsum which may
result in setting problems. Conventional wisdom suggests that the stoichiomatric ratio of alkalis to sulfur should be kept between 0.8-1.2 (see, however, Sec. B5.14). Excess Cl causes serious problems for preheater operation. Materials, as mined, therefore, are typically proportioned: Limestone (CaO) Shale or clay (510 2, Al 203 & Fe2O3) Additives (5i02, Al203 or Fe 2O3) 85% 13% <1% each
Normally, cement plants are located on limestone deposits and shale or clay is sufficiently ubiquitous for most plants to mine this locally. Additives are usually brought in, albeit in small quantities. Mining plans are developed according to the geology of the materials. If the limestone is not homogeneous, it may be necessary to blend rock from different areas in order to maximize recovery, and it may also be necessary to mine selectively in order to avoid low grade material or problems such as alkalis. Mining and hauling are commonly monitored by:
Blasting Stripping ratio Loading Hauling tonnes/kg explosive volume waste removed/volume used rock tonnes/hour of equipment tonnes/hour per truck
production records are most conveniently kept in dry tonnes but moisture levels of mined, hauled, and crushed rock must be considered.
All
Apart from chemistry, grindability is also a factor in selecting raw materials. In particular, silica additives containing large-grain quartz are very difficult to grind and can result in hard burning and high fuel consumption. If quartz silica is employed it should, preferably, have a natural grain size of less than 301ntt. In recent years, cement kilns have been increasingly employed to utilize industrial by-products (eg mill scale) and to dispose of industrial waste materials (eg water treatment sludge) in return for disposal fees. Such materials include:
Ca agents:
Si agents: Fe agents:
tin slag converter flue dust mill scale blast furnace slag stone working residues
s
LL
F agents:
The use of waste materials for cement manuhas led to incorporation of a much wider range of trace elements and their effects are reviewed by Bhatty (PCA Bulletin RD109T).
such as P, Ti, Cr and Mn. facture
Apart from raw materials, gypsum and fuel are required for cement manufacture together with various pozzolanic materials (both natural and by-product) if interground cements are produced.
2.2 Reserves
A knowledge of limestone and, to a lesser extent, shale reserves is necessary, particularly when justifying investment to increase plant capacity. Reserves are classed according to the detail in which they have been explored: Class A or proven reserves: Extensive drilling has confirmed quantity, mineralogy, variation, mining and legal availability. Class B or probable reserves: Sufficient drilling to allow presumption of quality and availability. Class C or indicated reserves: Widely spaced drilling gives extent and some knowledge of quality. Class D or inferred reserves: Initial exploration and consideration of geology allow general assumption.
Handbook 5
2.3 Crushing
Primary crushers should be capable of accepting shot rock with the minimum of wastage or of preliminary size reduction. Typically feed should be less than 120cm and, either the feed hopper should be protected by an appropriate grizzly, or a hydraulic breaker may be installed to reduce oversized rock. Commonly there are primary, secondary and, occasionally, tertiary crushers in series. Most crushers are operated in open circuit though, frequently, they are also preceded by a screen or grizzly to bypass fine material direct to product. Crushed rock should ideally be -12mm for feed to ball type raw mills and 30-80mm for roller mills or less than 5% of table diameter. Location of the crusher may be either at the quarry or the plant and is largely a function of haulage vs conveying costs (Heur; WC; 11/97, pg 34). Mobile crushers are common in aggregate quarries but rare for cement (RP; 9/1994, pg 31). Hopper shapes and feed mechanisms are important considerations where material flow may be a problem. In particular, the handling, crushing, conveying, and storage of wet clay materials prior to drying are prone to difficulties (Reed; World Cement; 2/1993, pg 21). Crushers may operate by compression or by impact. Compression machines comprise single and double jaw crushers and gyratory crushers (Figure 1).
Double Toggle Jaw Crusher has a simple compression motion with jaw
angle of 15-20. It is effective for hard abrasive materials with Iow (<5%) moisture and gives a reduction ratio of about 6:1. Single Toggle Jaw Crusher also has a measure of vertical jaw motion which adds attrition to compression. Sticky materials can be handled but wear rate is increased. Gyratory Crusher operates by pressure between a gyrating cone and a stationary or spring-loaded crushing ring. Hard, abrasive materials can be handled with reduction ratios of about 5:1. Compression type crushers normally produce a cubic product with a low proportion of fines; being choke fed, plugging will result with feed moisture in excess of about 5%. Wear is low but reduction ratios are usually less than 8:1.
6 Cement Plant
Operations Handbook
ir
Impact machines may be either hammer mills or impact crushers. These are usually preferred for limestones with quartz contents of less than 10%. Wear is greater than for compression crushing and there is a larger proportion of fines, but moisture contents up to about 12% can be accepted and reduction ratios up to 50:1 are possible.
2.4 Drying
The handling characteristics of materials relative to moisture content vary widely. In general, drying can be performed in the raw mill from an aggregate of up to 15% moisture but pre-drying may be necessary for certain materials to facilitate their handling. Drying is commonly effected either in combined crusher-dryers, in rotary dryers which can either use dedicated hot gas generators or waste heat from kiln exhaust, or in autogenous mills. In each case, the air flows will involve dust collection while the high humidity of the used gas usually favors electrostatic precipitators over bag houses.
;e n ;)e
y
g Ts te
n
Monitoring includes: Production rate, tonnes/hour Operating hours Involuntary downtime Feed moisture, % Product moisture, % Heat input (for dedicated heat generators) kcal/kg Heat consumption is most conveniently calculated on equivalent clinker basis so that it may be compared arithmetically with kiln heat to assess total process heat. If raw materials are partially predried, they may cause serious handling problems. Steam evolution from the hot material causes build-up and can plug dust collectors. Also the dry, fine fraction is liable to flush if held in intermediate storage. There are advantages in both handling and heat consumption if raw materials can be handled up to mill feed and dried in the raw mill.
y
is
2.5 Preblending
If the limestone and clay/shale are both of high consistency, or if their differences in form would result in segregation, there may be justification for handling each separately up to raw mill feed and proportioning them with the mill feeders. More commonly, material variation can be mitigated, and buffer storage more economically provided, by a mixed preblend pile, either longitudinal or circular. The limestone (and clay/shale if premixed) are metered and fed simultaneously to a feed conveyor. There are two major stacking systems: Chevron stacking in layers along a single axis with the feed conveyor sweeping backwards and forwards along the length of the pile (Figure 2.1). Winrow stacking in longitudinal strips side be side and then in successive layers; this avoids the segregation which characterizes chevron stacking (Petersen; WC Bulk Materials Handling Review; 1994, pg 30) but otherwise there is little difference in blending efficiency. The winrow system requires a more complex and expensive stacking belt arrangement (Figure 2.2).
-
it
)r Le Le
Chemical analysis of the material, especially if premixed, as fed to storage is essential. This may be either by continuous sampling followed by
Figure 2
Stacing conveyor
Stacing conveyor
Rake
Reclaim boom
Discharge conveyor
Discharge conveyor
periodic conventional analysis, or by neutron activation analysis (Tschudin et al; ZKG 5/1994; pg E141) which can continuously analyze the material flow falling through a chute or, better, traveling on a belt conveyor. The latter method has the obvious advantage of rapidity and of avoiding a very difficult sampling problem. Preblend effectiveness (the ratio of estimated standard deviations [escl] for feed and product) is roughly related to the number of layers IN] by:
Blending ratio = esd(feed)/esd(produco = V N/2
Typically a pile is built of 100-400 layers yielding a blending ratio of 3-6 for raw data and 6-15 if the variations within each reclaimed slice (which should be eliminated by normal handling) are ignored (Labahn). Recovery to mill feed is either by end or side reclaim: End Reclaim various systems to scrape an entire end face with a transverse scraper at floor level moving material to a discharge conveyor (Figure 2.3). Side Reclaim a boom mounted scraper working end to end along the side of the pile. This gives less effective blending and recovery flow rate is not constant (Figure 2.4). Two piles are normally operated with one being built while the other is recovered. Length to width ratio should be at least 5:1. The preblend pile is usually the principal buffer storage between quarry /crushing and raw milling; it should be minimally sized to maintain mill feed during the longest anticipated interruption in the supply of new rock which may be week-end shut-down of the quarry, or the time required for maintenance of crushing equipment. Total capacity of 7-10 days consumption is normal. Preblend piles are usually covered, both to prevent rainfall on the fine and, therefore, absorbent material, and to contain fugitive dust. Circular preblends are of significantly lower construction cost than longitudinal beds, but preclude subsequent expansion. Preblending should be monitored for compositional variation of feed and discharge to determine a blending ratio.
TM
z 2
Figure 3
Classifier Feed
Roller
Louvre Table
Hot gas ex preheater
Rejects trap
Drive
Feed Mill
Product
classifier; fines pass out with the air current while coarse material falls back onto the feed table. Material drying occurs in air suspension between table and classifier. Circulating load is typically 800%. Smooth running depends upon the stability of the material bed; occasional vibration may be due to abnormal feed, particularly dry or fine material. External circulation is frequently employed to optimize mill efficiency (ZKG; 9/1992,pg E249). The mill is started with the rollers lifted away from the table. When lowered to their grinding position, actual metal to metal contact should be prevented by limit switches and consistent feed. Material which is not carried upwards by the air stream falls from the table to a rejects trap, but every effort should be made to exclude tramp metal which can damage the grinding surfaces. Loesche mills are defined by grinding table diameter (dM), number of grinding rollers, and roller size; eg L.M41A50 is 4.1M in diameter with 4 roller modules of size 50. Primary roller mill controls are: s s s mill differential pressure which controls circulating load (usually 500-750mm WG) inlet gas temperature (normally up to 500C) inlet gas flow
1 z
It should be noted that the material cycle time of a roller mill is usually less than a minute against several minutes for a ball mill. Thus, control response should be appropriately faster. Optimum fineness of kiln feed must be determined empirically but, typically, raw materials should be ground to finer than 10% + 170# (881.) or 15% + 200# (74u) and adjustment is made on the separator. A narrow particle size distribution is optimal because fines tend to increase dust loss by entrainment in exhaust gas, while coarse particles are harder to react in the kiln and result in high free-lime, and/or excessive fuel consumption. Specific power consumption depends upon material hardness and mill efficiency. For ball mills, the range is approximately 8kWh/t (mill drive
only) for soft, chalky limestone to 25kWh/t for hard materials. System power consumption is typically 30% lower in roller mills, and combined power consumption may be 10-30% lower if a roll press or other machine is used for pre-grinding. Mill product is monitored either by continuous, on-line analysis or by laboratory analysis of hourly grab or composite samples. Computer programs are employed to effect feed corrections in order to maintain the desired optimum composition of long-term (say 6-hour) average. Automatic feed correction will probably be ineffective if any of the feed silos have been contaminated. Raw mills are monitored by: Production rate, tonnes/hour Operating hours Involuntary downtime hours kWh/tonne (mill motor) % of Connected Power (relative to mill motor rating) Product fineness, -170# Product moisture, % Limestone, kg/Eqt Clay/shale, kg/Eqt Additives, kg/Eqt (Note: Equivalent tonnes (Eqt) are theoretical tonnes of cement produced from given quantities of raw materials; see Plant Reporting, Section 11.2) Additional operating parameters required periodically include: Circulating load, % Ball usage, g/t (for ball mills) Chemical analysis of +170# fraction The coarse (+170#) fraction may be lime-rich or, more likely, silica-rich relative to the total sample but should show constant bias. Fundal (WC; 7/1979, pg 195 & FLS Review #22) suggests that quartz particles larger than 45p and calcite larger than 125p will not react completely under normal burning conditions; both fractions should, ideally, be limited to not more than 1%. Incompletely reacted silica results in belite (C2S)
dusters and unreacted calcite results in free-lime, both of which prevent alite (C3S) formation inhibit cement strength gain. Separators are covered in more detail under finish milling (Section 5.3). Roller mills require high efficiency separators to avoid material bed problems and high pressure drop associated with the accumulation of excessive fines in the mill circuit.
3.2 Blending
There are various blending silo designs. The two major types involve turbulence (in which the material is tumbled about by the injection of high volume air through air-pads on the silo floor) and controlled flow (where sequenced light aeration of segments of air-pads causes layers of material in the silo to blend by differential rates of descent within the silo). Controlled flow silos may have multiple discharge chutes, or an inverted cone over a center discharge within which the meal is fluidized. Compressor power consumption is approximately:
Turbulent mixing (air-merge) Controlled flow, inverted cone multi outlet (Bartholomew; ICR; 9/1995, pg 66)
Turbulent mixing can be operated batch-wise or continuously. The former involves either a filling cycle corrected progressively to average the target mix, or a sequence of filling, mixing, sampling and analyzing, correcting, remixing, and then feeding to kiln. Continuous blending involves simultaneous feeding of the silo, overflow to a second silo and discharge to kiln feed. Modern blending silos are generally of continuous, controlled flow type with each silo having capacity of more than 24 hours kiln feed and yielding a blending ratio of 3-5. Apart from power savings, the effective capacity of a CF silo is some 20% greater due to the higher bulk density of meal which is not heavily aerated. Retrofit modifications are available for converting air-merge systems to controlled flow for approximately $0.5 million per silo.
Blending silos should be monitored by: Blending ratio (esdfeed /esd product)
Compressor kWh/tonne throughput
Blending silos are prone to internal build-up of dead material, particularly if feed material is wet or if aeration is defective, and periodic (1-2 years) internal inspections are necessary. As raw meal is liable to solidify if left inactive (during a kiln shut-down for example), blending silos should be emptied or recirculated when not in use.
may also be carbonate titration. If oxides are measured, statistics should use C3S or LSE Kiln feed should typically have an estimated standard deviation for hourly grab samples of less than 0.2% CaCO3 or 3% C3S. (Note 1.0% CaCO3 is equivalent to 10-15% C3S.) It should be born in mind that standard deviation is not a perfect measure of variation as, simply applied, it does not distinguish between a steady trend and constant fluctuation. One might also beware of the dubious practice of eliminating extreme values from a set before calculating standard deviation: a major equipment supplier once tried to perpetrate this upon us to prove compliance with specification of his blending system. Kiln feed is normally conveyed by bucket elevator to the top of the preheater to minimize power consumption. If this conveying is effected pneumatically, de-aeration is desirable before injection as the entraining air otherwise adds to the kiln ID fan load and may reduce kiln capacity. Kiln feed to clinker ratio is typically 1.65 - 1.75 and, after determination, should be periodically reconciled with clinker and cement inventories and with measured dust loss in the preheater exhaust.
Kiln feed = Clinker + Lol + Bypass dust + Downcomer dust
-
Coal ash
where both bypass dust and downcomer dust are converted to ignited basis.
Cyclone preheater kilns have developed rapidly since the 1950s and have been virtually the only type of cement kiln installed over the past 25 years. The first units were 4-stage preheaters. Single string preheaters are limited to about 4500t/day (with up to 10MO cyclones) and larger kilns now have two- and even three- strings together with precalciners (secondary combustion vessels between kiln and preheater) allowing unit capacities in excess of 10,000t/ day. Heat recovery has also been improved, where heat is not required for drying raw materials, by using 5-and 6-stages of cyclones, and redesign of cyclone vessels has allowed pressure drop to be reduced without loss of efficiency (Hose & Bauer; ICR; 9/1993, pg 55). Exit gas temperatures, static pressures, and specific fuel consumptions for modern precalciner kilns are typically:
6-stage 5-stage 4-stage 280 310 350 450mm 1-120 400mm 350mm 710kcal/kg (NCV) 725kcal/kg 750kcal/kg
_Ad
Temperatures are 20-30 lower without precalciners and older systems are usually 20-30 higher than the above. Early 4-stage cyclone preheater kilns commonly have pressure drops of 700-800mm (higher if ID fans have been upgraded without modifying cyclones and ducts) and specific fuel consumptions of 850-900kcal/kg (Figure 4).
Figure 4
350deg C 750mm WG
r=93%
Feed
n=80%
WOdegC
700deg C 300mm WG
n=70%
11\
67GclegC
820deg C 150mm WG
000deg C 50mm WG
CYCLONE PREHEATER TYPICAL TEMPERATURE & PRESSURE PROFILE AND CYCLONE EFFICIENCIES
In cyclone preheater kilns without precalciners, the feed is 20-40% calcined at the kiln inlet. Riser firing increases this, and addition of a caldner vessel allows up to 90% calcination before the meal enters the kiln. Precalcination should not exceed 95% as completion of the endothermic reaction would allow a dangerous material temperature rise before entry to the kiln with probable build-up and plugging. The major cyclone preheater configurations are shown in Figure 5. Other terms frequently encountered include: NSP (New Suspension Preheater) - Precalciner technology which was developed in Japan in the early 1970s. MFC (Mitsubishi Fluidized-bed Calciner) RSP (Reinforced Suspension Preheater) - Onoda design of precalciner vessel. AT (Air through) - Precalciner or riser firing using combustion air drawn through the kiln. AS (Air separate) - Precalciner using tertiary air.
Figure 5
Calciner
Tertiary
SLC - Precalciner fed by tertiary only. Can be single, 4' double, or triple string preheater
Other kiln performance factors include: Primary air tip velocity, M/ sec Specific kiln volume loading, % Gas velocity in burning zone, M/sec Specific heat loading of burning zone, kcal/h per M 2 of effective burning zone cross-section area. Cooler air, NM3 /h per M2 grate area Cooler+primary air, NM3 per kg clinker Temperature, pressure and oxygen profile of preheater Excessive heat consumption should be investigated immediately and may be indicative of incorrect feed-rate measurement or feed chemistry, burner abnormality, insufficient or excess oxygen, air inleakage at kiln seals or preheater ports, low temperature of secondary air, and distortion or collapse of preheater splash-plates (Saxena et al; WC; 2/1996, pg 44). Clinker free-lime should be as high as possible to avoid the problems of hard burning, but safely below the onset of mortar expansion; typically between 0.5% and 2%. Having established the target, free lime should, if possible, be maintained within a range of about 0.5%. Variation of kiln feed rate or composition makes this control more difficult. It should be appreciated that overburning - a common solution to variable kiln feed chemistry or operator circumspection - wastes fuel, stresses refractories, increases the power required for cement milling, and reduces cement strength. Sasaki & Ueda (ICR; 8/1989, pg 55) found a l4kcal/kg heat penalty for each 0.1% reduction in free lime. A convenient supplement for free-lime measurement is the more rapid determination of litre-weight. This involves screening a sample of clinker from the cooler discharge to approximately +5/-12mm and weighing a standard 1 litre volume. Litre-weight is typically 1100-1300g/L (varying inversely with free-lime) but the target range should be determined with a minimum equivalent to the established free-lime upper limit. Secondary air temperature should be as high as possible in order to recover the maximum heat; usually 800-1000C. Maximizing secondary air temperature involves optimizing clinker bed depth and air volumes injected to the first cooler compartments. Note that secondary air temperature is difficult to measure unless there is a hotgas take off from the hood for tertiary or coal mill air; an unprotected
z
$20
0 0 1-1
thermocouple in the hood is liable to erroneous and misleading measurement due to radiation. Precalciner kilns maximize the heat input to the calciner and, typically, 60% of fuel is fed to the calciner while 40% is burned in the kiln. This serves to minimize the size of the rotary kiln and its heat loading; it does not reduce specific fuel consumption. It has been widely found that preheater kilns without precalciner vessels can still benefit from feeding 10-20% of total fuel to the kiln riser. Kiln operation is noticeably more stable and brick life is extended. This is also a useful means of consuming low grade fuels or waste materials. The limit to fuel injection at the riser depends upon its size and consequent gas retention time, and upon fuelair mixing characteristics; overfuelling results in preheater operating problems, an increase in exit gas temperature, and CO in the exhaust. The vortex finders (dip tubes) of lower stage cyclones were for many years prone to collapse and, frequently, were not replaced. A new segmented design in high-temperature alloy is now available for original installations and retrofits (WC; 10/1994, pg 39). However, the effectiveness of the vortex finders should be carefully assessed by review of preheater temperature and pressure profile and of specific fuel efficiency both before and after the tubes are removed or fall out; in many cases there is scant justification for re-installation. For kilns with grate coolers, the burner tip should be in the plane of the kiln nose (hot) or slightly inside the kiln providing it does not suffer damage from falling clinker. The burner should normally be concentric with, and on the axis of, the kiln. Some operators prefer to hold the burner horizontal and even tilted into the load. Such orientation may result in reducing conditions and must be adopted with caution. It should be appreciated that both burner position and tip velocity are intimately related to hood aerodynamics and can not be considered in isolation. Kiln rings are sections of heavy coating, usually in the burning zone, though sometimes also near the back of the kiln, which can grow to restrict both gas and material flow and eventually force shut-down. Conversely, ring collapse causes a flush of unburned material. Ring formation in the burning zone is commonly attributed to operational fluctuations though a low coal ash-fusion temperature or high mix liquid phase will increase the risk (Bhatty; Proc ICS; 1981, pg 110). Early
detection is possible with a shell scanner and rapid reaction is essential. Such ring growth may be countered by varying kiln speed or by small movements (10cm) of the burner in and out. Rings at the back of the kiln are usually associated with the volatiles cycle, particularly excessive sulfur at the kiln inlet. It is evident, though of little help, that rings are structurally more stable in small diameter kilns. Recurrence merits an investigation of cause(s) (Hamilton; ICR; 12/1997, pg 53). Certain plants have raw materials which contain significant proportions of hydrocarbons (kerogens), typically up to 3%, or may wish to dispose of oil contaminated soils. If fed conventionally to the top of the preheater, the hydrocarbons will tend to distill at intermediate temperatures and exit with the flue gas if they do not explode in the EP (Ryzhik; WC; 11/1992, pg 22). To prevent the resulting pollution, and to make use of the heat potential, kerogen containing materials should be injected at higher temperatures; usually to a 1- or 2-stage preheater if the hydrocarbons are present in the limestone. The high temperature exhaust may then be used for drying or for power co-generation (Onissi & Munakata; ZKG; 1/1993, pg E7). If the hydrocarbons occur in a minor constituent, this component may be ground separately and fed to the kiln riser. Petroleum coke, or the residual carbon in fly ash used as raw material, being involatile, can be added conventionally with kiln feed and yield useful heat (Borgholm; ZKG; 6/1992, pg 141). Some fly ash, however, contains high and variable carbon (1-30%) and, unless preblended, can seriously destabilize kiln operation.
w
2
szo 0 0
discrete control alogrithms, proprietary input/output (i/o) modules that interface loop controllers with field devices (eg pressure transmitters, damper operators), and proprietary software running on standard PCs for the MMI. Almost all DCS vendors (eg Honeywell, Rosemount, Bailey) design redundancy into the SCADA system and to multiloop controllers which yields very high reliability. DCSs also come with high level programming software which automatically takes care of common programming tasks and greatly facilitates system configuration and maintenance. All major PLC suppliers (AllenBradley, Siemens, GE/Fanuc) offer controllers which interface with DCSs, and a common architecture for DCSs employed in cement plants uses integrated multiloop controllers for analog control and PLCs for discrete control; with some 80% of cement plant control loops being discrete, this uses DCS controllers only for the few analog loops which require them while using the less expensive PLCs for discrete control. Such interfaced PLCs continue to be favored for discrete control due to speed, ease of programming, and reliability. Open Distributed Control Systems comprise SCADA software running on standard PCs, proprietary software running on proprietary PLCs for performing analog or discrete control alogrithms, proprietary i/o modules interfacing PLCs with field devices, and proprietary software running on standard PCs for the MMI. While a standard PC is used for both MMI and SCADA tasks, compatible software from a single vendor is used. The primary advantage of the PC system is the ease and economy of upgrading speed and memory. However, while hardware costs are lower than with proprietary DCSs, programming costs are usually higher because automated high level programming software is not yet available. Also, to obtain the same level of redundancy, additional PCs must be incorporated. The present cost savings for a PC system may be 10-15% less than for DCS (Feeley; Control Magazine; 11/1997, pg 40). The current trend is for DCS and PC systems to converge. Both DCS and PLC vendors are moving away from proprietary hardware and software to more open systems while it is increasingly common to find control systems running on PC platforms with software performing
multiloop control functions as well as MMI and SCADA (Walker; WC; 3/1996, pg 68). Various expert systems (Linkman, Fuzzy Logic, etc) can be overlaid to the computerized control platform and can give dramatic improvements in kiln stability, fuel efficiency, clinker quality and, consequently, in production rate. However, since process response to controlled variables changes overtime, such systems require constant attention. Adaptive programs are now being developed.
z
cits
Control is effected by adjustments to kiln speed, fuel rate and ID fan speed. Whether normal operation is manual or automated, most kilns are still liable to upset periods due to ring building, coating loss, etc and, while every effort should in any case be made to minimize such perturbations, stability is a prerequirement for effective computer control. Kiln feed and speed are usually controlled with a fixed linear relationship and unilateral variation of kiln speed should, at most, be used only as a short-term expedient (eg to control a kiln flush). It has been asserted that for many kilns speed should be kept constant in the upper range of feed rates (Clark; WC; 3/1994, pg 43). Kiln speed should be such that volumetric loading is within the range 10-13% (Section B5.10). Typically cyclone preheater kilns rotate at 22.5rpm (50-70cm/sec circumpherential speed) and have material retention times of 20-40mins. Precalciner kilns rotate at 3.5-4.5rpm (80-100cm/sec) for similar retention times. Retention in the preheater is 20-40secs. It has been asserted by Scheubel (ZKG; 12/1989, pg E314) that CaO, upon calcination, is highly reactive but that this reactivity decreases rapidly so that slow heating between 900-1300C can result in increased heat of formation of cement compounds. Keeping the same kiln retention time with increasing degree of calcination of the material
entering the kiln results in extending this transition and there is evidence that the introduction of short, 2-pier, kilns for precalciner systems has led to the reduction of material residence time before entering the burning zone from some 15 minutes to 6 minutes with resulting improvement in clinker mineralogy and grindability. Kilns are frequently operated to the limit of the ID fan. In this case, low oxygen must be corrected by reducing both fuel and feed. Precalciner kilns burn fuel at the kiln hood using combustion air mainly drawn from the hot end of the (grate) cooler, and in the calciner using combustion air drawn from either the hood or the mid- section of the cooler via a tertiary duct. Most precalciner kilns have dampers in the tertiary duct, and some have dampers in the riser, to control relative air flows to the two burners in order to maintain the desired fuel split. Frequently these dampers fail and it is then essential to adjust the fuel flows to the actual air flows. This is effected by maintaining oxygen at the kiln feed-end at, say, 2%. The gas probe at the kiln feed-end should project inside the kiln to avoid the effect of false air inleakage at the kiln seal; this is a difficult location for gas sampling and an adequate probe is essential (ICR; 6/1995; pg 51). CO should, and NOx may, also be measured at the kiln entrance. The oxygen level required at the kiln inlet will depend upon kiln stability and combustion efficiency; with a good flame, 1-2% 02 should result in less than 100ppm CO while an unstable flame may yield in excess of 0.1% CO with 3% 02. In a cyclone preheater kiln without riser firing, the downcomer oxygen analyzer serves both as back-up to the kiln inlet unit and to monitor air inleakage across the tower; an increase in 02 of more than 2-3% suggests excessive inleakage. In a precalciner kiln, an additional gas analyzer may be installed in the exhaust duct from the bottom cyclone and, again, this should be operated at as low an oxygen level as is consistent with less than 100ppm of CO. Useful information on kiln operation can be obtained from frequent (2-hourly) analysis of clinker for 503, and periodic (8-hourly) sampling of the underflow from the bottom cyclone stage(s) for SO3 and alkali determination. Normal SO3 levels (typically 0.6% in clinker and 2-3% in underflow) should be determined and maintained. In precalciner kilns, retention time and heat loading are particularly low and alkalis (K,Na)
tend to pass through to clinker while sulfur is volatilized and builds a cycle at the back of the kiln exacerbated by the deficiency of alkalis. If the kiln is burned too hot, this cycle increases excessively until build-up or cyclone plugging occurs. This is matched by an abnormally low 303 content in the clinker. Eventually, if the kiln is allowed to cool, this sulfur is released and transient high clinker SO3 results. Such variation in clinker SO3 will also give rise to varying grindability in the finish mill. It cannot be overemphasized that kiln stability, fuel efficiency, finish grinding power consumption, and cement quality all depend greatly upon the provision of a kiln feed and fuel with minimal variation both of chemistry and feed rate. Healthy suspicion should be nurtured towards both instrument signals and manually reported data. Particular areas for mistrust are: s False instrument signals of which pressure sensors and gas sampling probes are particularly liable to failure. s Short term variations masked by electronically damped signals. Feeder variations especially when the material is either sticky or fine and dry. Chemical variations hidden by faulty analytical methods, statistical mistreatment, or outright fraud. Variations in kiln behavior always have a cause; any variations which cannot be explained by observed feed deviation or known operational disturbance should alert to the possibility of faulty data. For management scrutiny it is useful to have either a "read-only" CRT which can be interrogated without interfering with kiln operation, or a strip chart recorder which minimally shows kiln feed, kiln speed, and kiln drive amps. Automated kiln control seems, unfortunately, to have reduced operators' habits of looking in the kiln and inspecting the clinker produced. Modern kiln and cooler camera systems, however, are excellent tools (Prokopy; RP-C; 5/1996, pg 38) for observing flame shape and position of the load in the kiln (dark interface of unburned material), "snowmen" (build-up on grates below the hood), "red rivers" and excessive blow-through in the cooler. The appearance of clinker can also be instructive; preferably black with surface glitter, dense but not dead burned, absence of excessive fines, dark grey cores. Brown cores are
m
cto
0 0
usually due to reducing conditions in the kiln but can also be due to the decreased permeability of clinker resulting from high belite and sulfate concentrations which inhibit oxidation of ferrous (Fe 2+) iron to ferric (Fe3+) during cooling. This in turn is due to chemical variation of kiln feed and to low volatilization of sulfur in the kiln (Scrivener & Taylor; ZKG; 1/1995, pg 34). Other causes have also been proposed (Jakobsen; WC; 8/1993, pg 32). Brown clinker is associated with increased heat consumption, reduced grindability, cement strength loss, and rapid setting. Certain alarms on the kiln system are critical. Apart from normal mechanical alarms and the routine monitoring of kiln shell for refractory failure, the potential for explosion requires particular care. Gas analysis is conventional at the feed end of the kiln, at the downcomer, and at the dust collector entrance. CO above 1% should cause alarm, and above 2% should cause fuel, and EP if so equipped, to shut off. Flame detection is particularly vital during warm up of the kiln and fuel should be shut off by interlock if the flame is lost. When the kiln is up to temperature it is common to deactivate the flame detector but it should be impossible to start a kiln without this protection. The light-up of kilns is potentially dangerous as there is insufficient temperature in the system to ensure continuous combustion. Unburned fuel accumulates rapidly in the kiln and, if then ignited, will probably explode. It is important that ignition be achieved as soon as the fuel is injected and, if the flame fails during warm-up, the kiln should be purged with 5 times the volume of kiln, preheater, ducting, and dust collector (probably some 3-5 minutes) before reignition is attempted. A simple and reliable ignition system has been described by Davies (ICR; 9/1996, pg 77).
follows agreement by production and maintenance management that all work is completed, that all tools and materials have been removed, and that all doors are closed. Work may, with discretion, continue in the cooler during warm-up but no workers should remain in
Figure 6
the cooler at the time of ignition. Commonly, warm-up from cold takes 24 hours from ignition to feed-on, but may be increased if extensive refractory work requires curing. A typical chart is shown (Figure 6) indicating the desired rate of increase in back-end temperature (this may also be set out in terms of fuel rate), the kiln turning program, the introduction of feed (usually 50% of full rate), and the increase of fuel, speed and feed to normal operation which should take another 8 hours from feed-on. For PC kilns, fuel is supplied to the calciner at the same time as, or soon after, feed-on. ID fan should be operated to approximately 10% 02 at the back of the kiln to feed-on whereupon the normal 02 target is adopted. For coal fired kilns, warm-up almost invariably employs gas or oil with switch-over to coal at the time of feed-on. If the coal mill uses hot gas from the cooler, there may be a delay before heat is available from the clinker. Before and during warm-up, equipment checks should be performed to ensure that each unit is ready to operate when required. Warm-up from shorter stops where the kiln is still hot, say stops of less than 24 hours, are conventionally accelerated to half the shutdown time.
Suggested inching is as follows: 0 - 2 hours continuous 2 - 4 hours 1/4 turn every 15 minutes 1/4 turn every hour 4 -12 hours If the shut down is for less than a day and does not involve entering the kiln or preheater, then heat should be retained either by stopping the ID
34
Cement Plant
Operations Handbook
fan immediately and shutting the preheater dampers after 2 hours, or (if there are no dampers) shutting down the fan after 2 hours.
feed end
8 months 70-85% alumina Basic, dolomite, or spinet 6-10 months 70% alumina 21 months 40% alumina 21-37 months
Kilns with precalcination average significantly longer brick life. A detailed historical record of refractory replacement and review thereof are important to minimize cost and service interruption. Typically, brick from the kiln nose to the back of the high-alumina brick section should be replaced if found to be 10cms or less in thickness, but such a rule-of-thumb is subject to much variation depending upon operating considerations. A useful practice is to drill through the brick every meter whenever the kiln is down and coating has been stripped (wider spacing and lesser frequency is adequate in the low alumina brick area). Such drilling requires discretion to locate the shell and to identify irregular circumpherential wear. Non-intrusive instruments to measure brick thickness are also available (eg Hoganas Linometer). The extent of coating should be observed whenever the kiln is entered and, roughly, basic brick should extend back to the top of the coated zone. Changes in fuels, feed, or burning conditions will affect the location of the burning zone. Coating location and refractory condition are usually monitored during operation with a shell scanner (eg Wulff; ZKG; 11/1993, pg E300). Kiln shells should also be inspected visually, particularly under tires where small hot spots may be concealed from the shell scanner. Warm areas of shell can be controlled by use of a fixed fan array or of movable fans which can be directed at the area. "Red spots", when the kiln shell reaches incandescence, should always be a cause for alarm and should not be allowed to persist for any length of time. If the hot
spot is a dull red and is in the burning zone it may be possible to re-coat the area and continue operation. Specifically, a small sharp hot spot, relating to the loss of one or two bricks, occurring in the burning zone can be "repaired" by stopping the kiln for 2-5 minutes under the load with an air lance cooling the spot. However, response must be rapid and the long-term problems caused by warping of the shell should always be born in mind. Red spots on surfaces other than the kiln may be temporarily secured by building a steel box on the outside to cover the hot area and filling the box with castable refractory; the box should be cut off and permanent repairs effected during the next kiln shutdown. There is an extensive literature on kiln brick types and performance (eg Scheubel; ZKG; 1/1994, pg E22: Wright; WC; 12/1994, pg 2). Brick usage averages 850g /t of clinker produced for cyclone preheater kilns and 500g/ t for precalciner kilns (Scheubel & Nactwey; ZKG; 10/1997, pg 572). With considerable variation, installed brick thickness is related to internal kiln diameter:
<4.2M diameter 4.2-4.7M 4.7-5.2M >5.2M 180mm 200mm 225mm 250mm
The two major bricking techniques are the epoxy method and the "ring-jack" method (Mosci; Brick Installation in Rotary Kilns; RefrAmerica 1995). Both have their place; the ring-jack is usually faster for long installations but does not allow turning of the kiln which may be important if other maintenance is to be performed on the shell, drive, or seals. Typically, installation after clean-out is at the rate of 0.5M/hour. In addition, monolithics, which comprise castable and plastic refractories, have various uses from the rapid gunning of large areas or
complex shapes to the molding of burner pipes and distorted kiln nose rings (Fraser; Proc IKA; Toronto; 1992). Castables are concretes with refractory aggregate and a high-temperature resistant (high Al203) hydraulic binder. Castables may be "heavy" or "lightweight insulating" and are classed: s standard (>2.5% CaO) s low cement (1.0-2.5% CaO) s ultra-low cement (<1.0% CaO) They are mixed with water, molded or gunned in place, and allowed to cure for 24 hours before heat is applied. The low cement types are very sensitive to water content and water addition must be strictly minimized; this results in poor workability and requires intense vibration for installation. Plastics have the consistency of modelling clay and are usually rammed into place though they can be gunned if so formulated. They are classed: s standard s air-setting s chemically bonded (usually phosphate bonded) Plastics have to be heated at specified rates to cure which can allow immediate kiln warm-up to begin. Castable and plastic refractories require steel anchors to hold them in place. The design and array of these anchors is critical to allow necessary movement of refractory against steel, and expansion joints are also required. The choice of refractory is dictated by the required service temperature, the potential for chemical attack, and the abrasion to which it will be subjected. Although provision is made for expansion when installing refractories, retaining rings attached to the shell have been found essential to resist the thrust resulting from rotation and inclination of the kiln. One should be installed approximately 1M in from the nose casting and others as desired further up the kiln but not within 5M of a kiln tire. Conventional rings are rolled plates 40-50mm high and slightly less than the width of one brick positioned such that the uphill brick is in
Cement Plant
Operations Handbook 37
tight contact, while the ring itself is covered by a slightly raised row of brick. Connection to the shell should be by plug welding (Refratechnik Manual). The low profile of the ring results in massive loading of the contact surface with the uphill brick with consequent risk of failure. Gortan et al (ZKG; 10/1994, pg E281) have developed a wedge-shaped retaining ring comprising a honeycomb of special alloy steel filled with castable refractory. Although specially shaped bricks must be laid on the rings, the system is claimed to produce a dramatic improvement in refractory life in difficult cases.
4.6 Insufflation
Insufflation is the injection of dust into the kiln flame. The dust may be
either metered into the primary air (using a venturi), or it may be separately conveyed pneumatically and injected adjacent to the main burner. The effects of insufflation are to: s increase the luminosity, and hence the heat transfer, of oil and gas flames. s reduce flame temperature and, hence, thermal NOx production s return fine dust to the kiln with maximum chance of incorporation into clinker rather than re-entrainment with the exhaust gas. s increase volatilization of alkalis in the dust to facilitate production of low alkali clinker. s allow production of small quantities of special clinkers by adjustment of mix design without transitioning the whole blending and kiln feed systems. Alternatively, if oil or gas are temporarily substituted for coal as kiln fuel, the effect of coal ash upon clinker chemistry can be maintained by injection of an appropriate mix. s facilitate the addition of small quantities of hazardous waste which must be fed directly to the high temperature zone of the kiln. It should be recognized that material entering the kiln at the hood uses high-grade heat for preheating which is less efficient than conventional kiln feeding, though this may be offset by better heat transfer. With coal as fuel, however, the emissivity may well be reduced and fuel-air mixing may deteriorate. Insufflation should not normally contribute more than 5% relative to clinker weight or flame temperature will be excessively reduced.
38 Cement Plant
Operations Handbook
c
z
szci
0 0
The criteria for bypass operation are complex depending upon input
concentrations, stoichiornetric balance between alkalis and sulfur, intrinsic volatility, kiln retention (the heat loading and retention time is shorter in precalciner kilns), and upon cement specification (Farag & Kamel; ZKG; 10/1994, pg 586). von Seebach (Frac ICS; 1983, pg 31) gives the average effectiveness of a 10% bypass as a reduction of 45-65% for Cl, 20% for SO3, 10% for K20 and 4% for Na20. Automated control of the bypass has been proposed based upon sampling and analysis of the hot meal (Triebel et al; ZKG; 10/1994, pg E272). Chloride circulation should be limited to 5-10g/Kg clinker depending upon the degree of precalcination (Farag & Abbas; ZKG; 1/1995, pg 22) Much cement must now meet low-alkali specification (Na20+0.658K20 < 0.6%) while excessive SO3 in the clinker inhibits C3S formation. Typically, however, kiln problems may be expected if total alkalis or SO3 exceed 1%, or chloride exceeds 0.015% relative to clinker weight.
It should be noted that natural 802 scrubbing in cyclone preheaters is very efficient. SO2 found in the stack originates from pyrites or organic sulfur in the raw materials, not from sulfate or from fuel sulfur.
usually priced in terms of gross heat (heat available assuming water in combustion product is condensed to recover latent heat of vaporization). In practice, only the net heat is employed (assumes that water in combustion gas is released as vapor). The difference varies with fuel:
Gross kcal/kg 5500-7100 Coal Oil, #6 10200 Natural Gas (kcal/M3) 6200 It Net kcal/kg 5400 -7000 9700 5600 Diff 2% 5% 10%
should also be noted that the gas flame is of lowest emissivity and, requiring more combustion air per unit of heat, is the least efficient. Kiln production typically increases by 2-3% when gas is replaced by coal. On the other hand, gas is the cheapest and easiest fuel to handle and is conventionally billed after use rather than requiring advance purchase and inventory. Assuming 850kcal/ kg clinker and 2% excess 02:
Flame Temperature 2250C 2350C 2400C Combustion gas NM3/million. Cal 1.23 1.31 1.45 Total exhaust gas NM3 /tonne clinker 1360 1420 1550
Coal Oil
Gas
Coal, much more than oil or gas, is liable to compositional variation. The nature of production and handling by major suppliers should minimize short-term fluctuation while long-term variation can be compensated by analysis and normal kiln control procedures. If, however, the supply is from small-scale or multiple suppliers, adequate blending must be effected prior to use. Stockpiling of coal requires vigilance as spontaneous combustion is common, particularly with wet or pyrites-containing coal. Smoldering coal should be dug out, the site spread with limestone dust, and the coal then compressed. If long-term storage is necessary, the pile should be compacted and sealed with coal tar emulsion. Thermocouples embedded 1-2M below the surface allow monitoring for combustion. Coal is usually dried, ground so that the residue on 200# (750 is not more than 0.5 x % volatiles, and injected with carrying air at a pressure of 120-150g/ cm2 and tip velocity of 60-80M/sec. A more precise determination of optimum fineness according to coal type has been described by Seidel (ZKG; 1 /1995, pg 18).
Oil may require pre-heating to reduce viscosity and is injected with a nozzle pressure of about 20kg/cm 2 except for pressure atomized systems which employ pressure to 100kg/cm 2 . Droplet size should be less than 2001 to permit burn-out while still retained within the flame. Gas is usually received at 10-70kg/cm 2. Primary air is not essential and the gas is injected as axial, or a mixture of axial and swirl, flow at 3-10kg/cm2 and a tip velocity of 300-400M/sec (injection will normally be limited by sonic velocity - 430M/sec for methane at 0C). Gas requires turbulent diffusion and its heat flux tends to be released more slowly than with oil or coal; peak heat release is usually about 20M into the kiln against 5-10M for oil. This results in slower response to control changes which makes for more difficult control of the kiln. It should also be noted that, with a higher ignition temperature than oil or coal, natural gas cannot be reliably reignited off hot kiln lining. Recently, a novel "precessing jet" burner has been introduced which both increases the luminosity of the gas flame and reduces NOx emissions (Manias & Nathan; WC; 3/1993, pg 3). The technology has since been extended to mixed gas-coal flames (Vidergar & Rapson; Proc IEEE; 4/1997, pg 345). Combustion comprises Mixing Ignition Combustion reactions Product dispersion
In most industrial combustion processes, mixing is the slowest and, therefore, rate-determining step. Hence the risk of excessively reducing primary air. It is also apparent that optimum combustion depends upon the aerodynamics of the kiln hood or calciner vessel and that burner tip momentum must be related to secondary air momentum, a technology known as "flame control" (Mullinger; WC; 6/1993, pg 20). Burner design, therefore, involves more than fuel rate alone. In recent years the cost of fuel, which, for most plants, is the largest single cost factor, has stimulated a search for low cost alternatives. Petroleum coke has certain advantages, particularly its very high heat content, but increasing price in some markets has reduced its attraction. The usually high sulfur content (3-6%) also limits rate of addition.
It should be noted that there are two main types of pet-coke: "delayed" and "fluid". The preponderant type comes from the delayed batch process in which feedstock is heated under vacuum to about 500C; the residue, "green delayed coke" has typically 8-16% volatiles while calcining at about 1700 yields less than 1% volatiles. Delayed coke may be "sponge" or "shot", can be milled with coal, and is commonly used to 60% of total fuel (to 100% has been claimed). Fluid coke consists of small spherical particles resulting from a continuous coking process at about 650. Volatiles are typically 5-10% and the coke is too hard for conventional milling. Fluid coke is injected unmilled at 10-20% of total fuel (1CR; 10/1993, pg 55). Numerous other by-product and waste fuels have been used and many command disposal fees. Progressively, however, source reduction is diminishing the supply of easily handled liquid solvents and waste oils, and the available materials are, increasingly, solids, aqueous sludges, or scheduled hazardous materials involving onerous regulation. With such materials both consistency and possible contaminants must be monitored. Tires are potentially attractive though whole tires are more problematic to handle while shredding or pyrolisis eliminates much of the cost benefit. Also, fuels added discontinuously, such as whole tires or containerized waste, derate the kiln as sufficient oxygen must be maintained to support the peaks of combustion.
Liquid waste fuels: tar distillation residues used oil petrochemical waste paint waste
Solid waste fuels:
chemical wastes waste solvents wax suspensions asphalt slurry oil sludge rubber residues used tires battery cases wood waste rice chaff nut shells sewage sludge
pyrolysis gas
paper waste pulp sludge petroleum coke plastic residues domestic refuse refuse derived fuel oil-bearing earths
Gaseous waste:
landfill gas
Kilns employing alternative fuels have detailed specifications to prevent operating or environmental problems, and each shipment is
Ir"
sampled and checked before unloading to ensure compliance. Typically, heat value should not be less than 4500kcal/kg, chloride is limited to 1% and most plants decline fuels with PCB content exceeding 50pprn. Waste fuels may be burned in the kiln, the riser, or the precalciner.
Roller mills can dry coal of up to 10% moisture beyond which the mill is derated according to manufacturers design data. Similarly mills are normally designed for 55 Hardgrove index and harder coals (lower HGI) will result in derating. Finally, a 10% fall in capacity between maintenance is assumed and allowed for in sizing a coal mill. If the mill and fan have separate drives, capacity can often be increased by raising the table speed (eg from 900rpm to 1200rpm); the fan must be subjected to separate engineering evaluation. Mill inlet temperature should not exceed 315C and coal should not be dried to below 1% surface moisture. Mill discharge temperature is limited to 65C for indirect systems and 80C for direct. Carrying air velocity must be maintained above 20M/sec to avoid dust settlement (Recommended Guidelines for Coal System Safety; PCA; May 1983). Vendors specify a minimum airflow, typically 1-1.5kg (0.8-1.2NM3) air per kg coal, which must be maintained even when the drying requirement is negligible. This airflow is required to ensure coal entrainment. Fires are usually the result of rags or wood lodging within the mill and may be detected by an increasing discharge air temperature unrelated to increased inlet temperature or reduced feed rate. Fires in direct firing mills are extinguished by adding feed to act as a heat sink and lowering the mill inlet temperature. Mills in indirect firing systems conventionally employ CO monitoring to detect combustion (thermocouples are too slow to respond); extinction is effected either by water injection or, better, by CO2 or N2 with CO2 predominant. Hot air for drying coal can come from cooler exhaust (normal air) or preheater exhaust (low oxygen). The inlet temperature to the mill is controlled to maintain the outlet temperature as described above and dedusted in a cyclone. Tempering to about 370C is effected by bleeding in cold air between cooler and cyclone. If preheater exhaust is used it will typically be at 300-350C with 5% 02 and 6% moisture; the temperature and moisture must be considered in the system design. Although most mill fires occur on start-up or just after shut down, the low oxygen atmosphere does reduce the risk of a mill fire during normal operation.
cci
z
0 0
Figure 7
Kiln
Cooler
To Precalciner
Coal mill
In the United States all coal mill systems are designed in accordance with NFPA Standard 85F which requires the equipment to withstand pressures of 2.5 times the absolute working pressure; this equates to 3.5kg/cm2 (50psi). Direct coal firing involves a single pipe burner through which the mill carrying air together with entrained coal are injected at a tip velocity of approximately 80M/sec. Tip velocity must always be substantially greater than the flame propagation velocity which may be up to 25M/sec. The pipe is usually narrowed near the tip to generate turbulence and improve fuel-air mixing. For indirect firing, three channel burners of various designs are employed (Figure 7). The outer annulus is for axial or jet air, next is an annulus for coal, and inside is a second air annulus with swirl vanes at the tip. To achieve adequate momentum for flame recirculation with the reduced primary air of indirect systems, jet air may be up to 440M/sec and swirl air 160M/sec (KFID Pyro-Jet burner). The center may be used for a second fuel. Coal firing almost inevitably involves a normal operating condition where pulverized coal is in contact with air before reaching the burner; this is a condition that should, theoretically, be avoided due to its inherent risk of fire and explosion. The design of a coal firing system is, therefore, critically important and should consider numerous factors including the following: Minimum ignition temperature of air/fuel mix; coal ignition temperatures range 200-750C though the ignition temperature of volatile products may be lower. Minimum explosive concentration of fuel in air is about 40g/M 3 (notehaduspirenothmgus) s Maximum permissible oxygen concentration to prevent ignition is 12%. The entire pulverized coal system must be designed to contain 3.5kg/cm2 (NFPA 85F) or with explosion relief as appropriate (NFPA 68). Coal dust explosive tendency increases with volatiles content and with fineness, and decreases with water content and with inert dust diluent (eg limestone). Coal dust is liable to spontaneous ignition which increases in risk with thickness of dust layer and with the presence of easily oxidized
w
2
0 0
contaminants such as pyrites (more than 2%). A smoldering fire can become explosive if disturbed. Coal obviously can be handled safely but pulverized coal should always be considered as a potential explosive. Safety considerations are reviewed in a publication, Coal Fires and Explosions: Prevention, Detection & Control; EPRI Research Project 1883-1; Final Report, May 1986.
48 Cement Plant
Operations Handbook
is exhausted taking with it more than 100kcal/kg clinker. Such coolers were typically rated at 35-40t/M 2 /day and operated with clinker bed depths of 200-400mm. Air beam designs are now available from all cooler manufacturers which, with individual variation, provide for the direct ducting of cooling air to hollow grate support beams whence the air is directed into the clinker bed more efficiently than is possible from the former undergrate plenum and grate array. This type of cooler allows 1.5-2NM3 /kg to be used with a rating of 45-55t/M 2 /day and bed depth of 800mm. The grates used with air beams, "controlled flow grates" (CFG) are of a higher resistance design with air passing horizontally through slots into recesses in the grate surface; this largely eliminates fall through of clinker fines and renders the air flow less dependent upon bed resistance. Similar grate designs are also available for the non-recouperating zone but with lower resistance as required by plenum aeration; these are referred to as "reduced fall through" (RFT) plates (Roy & Brugan; ICR; 2/1995, pg 47). Numerous refinements include a pendulum frame for the moving grates which is claimed to minimize drive maintenance, a static inclined arrangement of the first 6-9 air beams which serves to distribute more evenly the clinker entering the cooler, and a hydraulically operated heat shield at the end of the secondary (or tertiary) air recouperation zone which reduces high-grade heat loss to exhaust (von Wedel; ICR; 4/1995, pg 75). The static grate section has been found particularly effective and, typically, costs $0.5 - 1.0 million installed. As kiln discharge rate and clinker size can vary with kiln operation, the pressure drop across the clinker bed on the cooler will also vary. Undergrate air pressure is maintained constant by interlocking pressure (usually of the second compartment) with first grate drive speed. The second drive is controlled at 10% faster than the first if of equal width. An interesting theoretical analysis of grate cooler operation has been made by Moles & Bernstein (ZKG; 3/1995, pg 125). As large clinker nodules and blocks of coating cannot effectively be cooled, grate coolers incorporate clinker breakers. Size reduction is more effective if performed at a mid point in the cooler using a roll crusher. Older coolers, however, normally employ hammer breakers which, due to the difficulties of maintenance at high temperature, are
m
20
0 0
more often located at the discharge. Visual inspection of the cooler is important; in particular persistent "red rivers" indicate a problem with grate geometry or air distribution and "blow through" indicates excessive air flow to a particular compartment. Saint-Paul & Slot (ICR; 8/1995, pg 61) describe a scanner for continuous monitoring of cooler operation. Acoustic horns have been suggested as a means to improve cooler heat recovery (WC; 1/1995, pg 26). Clinker coolers are monitored by: Secondary air temperature, C Tertiary air temperature, C Discharge air temperature, C Discharge air volume or fan amps Clinker discharge temperature, C For process analysis, a detailed record is also required of individual cooler compartment under-grate air pressures and of grate drive speed and power. Broken grates can allow excessive leakage of hot clinker to the under-grate compartment with risk of major damage. Thermocouples or level indicators placed below the grate drive permit an alarm for clinker filling up within a compartment. Clinker production is not usually measured directly and is calculated from kiln feed with confirmation from cement shipment. However, it is helpful to have a point in the clinker transfer system from which clinker production can be loaded to a truck to allow weight checks. Other coolers are occasionally encountered which avoid the need for separate dust collection: Rotary coolers are simple rotating drums which lift the clinker to fall through the incoming combustion air stream effecting heat exchange. These coolers are limited to small kilns. Planetary coolers comprise a ring of tubes attached to the kiln shell and turning with the kiln, which serve as multiple rotary coolers. These coolers can cause mechanical problems on the kiln and tend
to complicate flame aerodynamics; they are, therefore, limited to special applications. Also it is difficult to balance clinker flow to the cooler tubes. Rotary and planetary coolers are less effective than grate coolers, being limited in cooling air to that which can be consumed in the kiln for combustion. However, with efficient internals, clinker can be cooled to about 180C (Steinbiss; ZKG; 8/1992, pg E210). Replacement of planetary coolers by grate coolers is not uncommon (Cohrs; ICR; 9/1995, page 40). G Coolers (from Claudius Peters) are effective in producing low temperature clinker in combination with a planetary cooler, a Iow-air-volume grate cooler, or to accommodate kiln capacity increase beyond the grate cooler capacity. Clinker, after the primary cooler and clinker breaker, is allowed to settle slowly in a bed with heat exchange to air cooled tubes (Harder et al; ICR; 9/1994, pg 56). The cooler is built from standard modules with the length determined by capacity and the height determined by required temperature drop. These units tend to require little maintenance. However, fine clinker may impact retention time and level control requires care due to the long lead-time from discharge gate adjustment.
-
co c
z
R:4
0 0
stresses on the refractory lining which is also exposed to thermal stresses. The lining absorbs these stresses through minute amounts of deflection within the individual refractory units (bricks) and through relative motion between refractory units. The material strength of the individual refractory unit, together with the strength of the joints between units, enables the kiln lining to deform to some extent without failure; however, excessive stresses will lead to lining failure. Though it is extremely difficult to quantify refractory lining stresses, it is easy to conclude that any condition which exacerbates stresses to the kiln shell will increase risk to the kiln lining and should be avoided. Proper alignment of the kiln support rollers and maintenance of tire pad clearances are the two primary precautions for minimizing stress. Roller shafts must be on the same slope as the kiln though they do deflect due to the kiln vertical load by as much as 100 - 50011. All shafts on a given pier must be parallel to avoid generation of unnecessary thrust bearing loads. Shafts should be slightly skewed relative to the kiln shell's theoretical axis at each pier to generate some thrust on the kiln tire that pushes it, and consequently the kiln, uphill. On a properly skewed set of bearings, rollers ride downhill against their thrust bearings. Slight changes in skew can relieve the rollers' downhill thrust. Although all kiln thrust rolls are designed to support the entire kiln's downward load, only certain types, such as those with hydraulic actuators, are designed to operate this way continuously. When all rollers are properly skewed, and under stable process conditions, each support roller imparts a small uphill thrust to the riding rings and, thus, to the kiln shell so that the kiln will contact its thrust rolls only intermittently during each revolution. Evidence of excessive support roller thrusting can be detected by temperature comparisons of the thrust bearings and from bearing wear rates. Unfortunately, most kilns do not have bearing metal thermocouples so that temperature measurement must be indirect. The two most common methods are to monitor the temperature of the bearing housing where the thrust bearing ("button") is mounted, or to measure the temperature of the roller shaft thrust shoulder using an infrared pyrometer directed through the bearing housing handhole. Another excellent way to check for excessive thrusting is carefully to rub the fingers across the surface of each roller noting if the surface feels smooth or rough while wiping in the uphill or downhill direction. This is often referred
to as "checking for fishscales" since the roller surface will feel rough in one direction and smooth in the other. Kiln seals are required at the inlet and discharge ends to exclude false air. The feed end seal must protect against 4 - 5cm WG differential pressure compared to 1cm or less for the discharge seal. False air at the discharge seal displaces hot (900-1000C) secondary air so is detrimental to fuel efficiency. False air at the feed end not only reduces capacity and efficiency but, introducing cold air in the middle of any volatile cycle, tends to cause serious build-up and interfere with both material and gas flow. There are numerous designs of kiln seal and most work reasonably well if they are properly maintained. The most common replacement seal design utilizes flexible sheet metal plates arrayed around a steel ring mounted to the firing hood. The plates contact the rotating kiln shell cowl (Geiger; WC; 12/1995, pg lb). Even within this type there are many variations in design. At the discharge end, the most common cause of poor sealing is overheating of the sheet metal plates. Many kiln designers have found that it is important to protect the internal surfaces of the seal plates from exposure to radiant heat. This is usually achieved by appropriate provision for radiant heat shielding and cooling air flow. Apart from overheating, the most common cause of seal failure is material loss resulting from stationary and rotating component contact. Since wear between these components is unavoidable, regular inspection and opportune maintenance is essential. It is also important that the rotating component runout be held within the seal's capability.
cc;
20
rZ
0 0
about half the strength relative to ambient temperature so that it is essential that shell temperatures be monitored continuously, recorded and alarmed. Infrared imaging systems are used increasingly for this purpose and also to provide information on refractory condition, coating thickness, and tire creep. Creep should never be zero and may typically be up to about 2cm per revolution. Any shell temperatures in excess of 350C should be monitored closely and if corrective action is deemed necessary to establish or modify the coating it should be commenced before the shell temperature reaches 500C. Continued operation at or above this temperature will generally result in permanent shell deformation or crack initiation. Normally there are two thicknesses of shell at each tire; the thicker plate directly under the tire is known as the tire course and the thinner plates uphill and downhill are known as flanking plate. The plate between tire sections is even thinner than the flanking plate. The most common location for shell cracking is at the transition between the flanking plate and the thin shell plate that spans between piers. Failures generally occur at the toe of the weld joint on the thin plate side. It is often acceptable simply to mark the extent of these cracks while continuing to operate until an opportune shutdown. The crack should then be removed by burning or gouging from the outside a groove with sides making an angle of about 60. Magnetic particle or dye penetrant inspection should be used to determine that the entire crack has been removed. Welding should be performed with appropriate filler metals laid down in straight beads. Each bead should overlap the one below it by about half the bead width to provide heat for relief of weld shrinkage stresses in the underlying bead. Beads should be laid in straight parallel lines; weaving of the beads should not be allowed as this technique usually results in poor quality and high residual stresses. The layers of weld bead should be built up until the toe of each weld bead lies on the line connecting the lips of the groove; ie the weld should form a tapered surface from the thicker plate to the thinner. There is no need to stress-relieve after welding but the shell plates must be maintained above about 5C during the entire process. The common practice of drilling the ends of cracks to stop their propagation is usually ineffective unless the hole is calculated for the particular situation and is of the order of 2 inch diameter. At the first opportunity, the repair should be accessed from the inside and the full length again burned or gouged to remove the root pass and reweld as described for the initial repair.
54 Cement Plant
Operations Handbook
Apart from overheating, the most common cause of shell cracking is probably fatigue due to excessive alternating stress generated from kiln rotation with high tire pad clearance. Though tire creep should be logged daily for each tire, creep is not the best indication of tire pad clearance. This should be measured with a device commonly known as the Obourg Pen Tester (Chapman; Recommended Procedures for Mechanical Analysis of Rotary Kilns; Fuller Company 1985, Fig 10) which yields a trace of the relative motion between tire and shell during several revolutions. More importantly, regular ovality measurements should be taken at each tire to verify the actual shell deflection. Generally, when percentage ovality exceeds a value:
Shell diameter in metres / 10
c
2
corrective action should be considered both to limit the negative impact on refractory life and also to prevent excessive shell stress. On tires adjacent to the kiln gear, it is common to find normal or even low ovality while tire pad clearance is high. When tire pad clearance on these tires reaches a level which is associated with excessive ovality on the other tires, corrective action should be instigated. Although the gear is able to hold the kiln shell with minimal ovality, it is not designed to handle the stress caused by this situation. Tire creep varies with shell temperature and the continuous measurement offered by some shell scanners is valuable. Some creep is essential at all times and it should not normally exceed about 2cm per revolution; correction is effected by shimming of tire pads. Tire thrusting or excessive thrust loading on a tire is indicated by hard contact between a kiln tire and its retaining mechanism. The most frequent cause of excessive tire thrust loading is a slope difference between support rollers and the kiln axis through the tire. Other, less common, causes are conical wear on tire or roller and excessive tire pad clearance. Axial thrusting of a kiln tire should be corrected before significant metal is removed from the tire sides because the resulting undercut is difficult to remedy. It is generally ineffective to attempt to relieve tire thrust by skewing support rollers; a thrusting tire requires one or more corrective actions: the support rollers returned to design slope, the rollers and /or tire reground to cylindrical, or the tire pad clearance reduced.
$20
0 0
Cl
Kiln drives generally utilize girth gears and pinions designed to give
over 20 years of continuous service if lubrication and alignment are maintained. Lubrication should be in accordance with the gear supplier's
recommendations for viscosity at operating temperature. An infrared pyrometer is used to measure the gear and pinion tooth flank temperatures and at least three measurements across the gear face should be recorded monthly. A properly aligned gear set will vary no more than 5C across the gear mesh. Alignment corrective action should be considered when temperature variation exceeds 10C. The tip to root clearance between gear and pinion should be measured routinely and after any refractory failure which may have resulted in shell damage. This measurement should be made on the uphill and downhill end of a pinion tooth every 60 of kiln rotation to ensure that clearance and gear runout are adequate for continued operation. Any changes in support roller position should be considered for their effect on gear alignment and it is, therefore, normal to move tire axes on all piers except the drive pier. Over 80% of the horsepower generated by the kiln drive motor is, by design, intended to elevate the kiln charge. However, excessive kiln misalignment will greatly increase the base load power required. Motor designs provide for short term loading of up to about 2.5 times the motor rated current and torque to overcome inertia and static friction for starting. Precalciner technology and the desire to minimize retention time of material between calcination and sintering have resulted in a trend to increased kiln rotation speed. As older kilns are upgraded it is common for drive speeds to be increased and this is usually accomplished in one of three ways: s gear reducer ratio changes s weakening of d.c. motor field strength s utilizing higher than rated motor frequency on a.c. induction motors with variable frequency drives Before deciding to speed up a kiln, it should be verified that the drive motor has adequate reserve capacity to accommodate increased torque demands from process upsets. If a motor is run close to its current limit, load changes resulting from operating conditions may cause uncontrolled changes in kiln speed which will seriously exacerbate the problems of kiln control. Thus, a marginal drive motor should be replaced before attempting to increase kiln speed. During normal operation the motor should draw approximately 60% of capacity. This provides additional torque to overcome thermal bowing, uneven coating, and kiln misalignment
Kiln alignment. Most kiln designers limit bearing pressure between support rollers shaft and bearing to around 35kg/cm 2 though some designs allow for higher pressures if the bearings are self-aligning. Allowable bearing pressures are determined by the bearing materials used, either brass or babbitt, so that the length of the bearing journal must be selected to keep these pressures within design limits. The lubricant film generated by rotation of the support roller journal is usually not over 125 thick so that anything which impacts this film thickness is of consequence. It is essential to inspect the shaft surface routinely during shut downs and remachine before excessive circumferential scoring may cause penetration of the oil film, temperature elevation, and bearing failure. Kiln designs make generous allowances for the weight of the kiln charge, refractories, and accretions. To operate reliably the support roller journal surface and bearing clearances must be adequate as must the viscosity, supply, and cleanliness of the bearing lubricant. Variations in coating pattern resulting in nonuniform shell temperature distributions can cause temporary changes in the shell's theoretical axis of rotation. These temporary, process induced, bends in the shell axis combined with other permanent shell deformations can cause excessive bearing pressures and subsequent bearing failures. The lowest risk operating state then, is the one that allows the kiln to accommodate as much transient shell misalignments as possible. To achieve this state the kiln shell's theoretical axis must be measured while the kiln is operating. This is actually done somewhat indirectly by measuring the position of the riding rings, assuming the rings and shell are perfectly round, and calculating the location of the kiln axis at each tire. Then imaginary straight lines are drawn between each tire axis to determine if the tires are high, low, left, or right of a theoretical straight line representing the kiln shell's axis. If the kiln shell is relatively straight, the location of each tire's measured axis should lie within about 3mm left or right of the theoretical kiln axis when viewed from above the kiln. The vertical locations of the tire axes relative to the theoretical straight line kiln axis may vary significantly from pier to pier among kilns. Generally, on a three support kiln, the middle support rollers bear the highest load and often must be set low relative to the theoretical kiln axis in order to avoid excessive bearing pressures. The most effective means to make the final alignment adjustments is to
utilize an ovality gauge to measure the shell deflections at each tire. The support rollers should then be adjusted to give equal deflection between left and right rollers on a given pier as well as between different piers. This is not, however, accomplished by obtaining the same ovality percentage at each pier because ovality is significantly influenced by the tire pad clearance. Bearing temperatures also are indicators of the support roller loading and, consequently, should be monitored closely. Tire ovality should not, by design, exceed 0.2%. Tires which are too thin give rise to excessive ovality and this can only be remedied by tire replacement. Kiln support rollers are designed to bear the weight of the kiln as well as some of the downhill thrust acting along the axis of the inclined kiln. The support rollers on each pier should have their axes aligned parallel to the theoretical axis of the kiln between each pier and slightly cut to impart an upward thrust to the tire on each pier. This will result in the support rollers lightly touching against their thrust bearings. The sum of all support roller thrusting should keep the kiln from contacting its thrust rollers continuously when internal coating is normal and uniformly distributed. Note that longitudinal shell expansion at operating temperature is approximately 20cm and it should be confirmed that the tires are centered on the rollers when hot. Corrosion of the kiln shell is not normally a serious problem unless high levels of sulfur or chloride are present. Corrosion is reviewed by Kotter & Bartha (Proc. Refratechnik Symposium, 1986, pg 104).
Alternatively, at very least, there should be a gasoline engine to drive the kiln at low speed. Certain manual procedures are then essential such as withdrawing the burner pipe from the kiln hood and opening doors at the top of the preheater if there are no automatic vents. The more reliable the main power supply and the more infrequent the use of the emergency system, the more important is maintenance of the standby equipment and rehearsal of the procedure. Serious distortion of the kiln shell will result if it is not turned within 15-30 minutes of a crash stop.
0 0
5 FINISH MILLING
5.1 Clinker Storage
Clinker as discharged from the cooler is liable to be, at least intermittently, at high temperature. A convenient transport system avoiding intermediate transfer is a steel deep-pan conveyor which can elevate the clinker at up to 45 to the top of the clinker storage. In combination with cement storage, there should be adequate clinker capacity to maintain cement shipments during kiln maintenance shut-downs and, if the market is cyclical, to bridge low and high shipping periods. While total clinker capacity should be equivalent to at least 14 days of kiln production, there must also be separate storage for different types of clinker (if produced) and for high free-lime clinker. Clinker storage serves also to blend the clinker and a silo should not be filled and discharged to milling at the same time unless the silo has multiple discharge points to avoid short-circuiting. It is inadvisable habitually to run clinker silos down to the steel cone as abrasion will eventually cause structural failure. Loading of hot (>150C) clinker to concrete silos should also be avoided as spalling and delamination will progressively weaken the silo walls. High free-lime clinker must be blended into mill feed with circumspection to ensure that the cement is not expansive. A maximum composite free lime for the cement should be established and used as a control parameter. There are obvious economies of scale for clinker storage and single silos up to 150,000t capacity have been constructed (Heine; ICR; 6/1991, pg 48). The desirability of large capacity storage buildings has resulted in a range of structures from conventional sheds and steel or concrete silos, to domes of various shapes either clad steel structures or concrete shells sprayed on inflatable forms (Tavares; WC; 3/1995, pg 67). Outside stockpiling of clinker is not usually worthwhile if subject to rainfall; it is also prohibited in some locations. Partial hydration of clinker before grinding seriously reduces strength and increases setting time. Recovery from the stockpile should not contribute more then 1.0%
to cement Loi and can seldom exceed 10% of clinker fed to the mill while cement produced is of higher unit cost due to added handling costs for recovered clinker, and to milling to increased fineness to avoid strength loss. If clinker must be put outside, it is advisable first to screen out fines (-4#) to minimize fugitive dust and hydration.
62 Cement Plant
Operations Handbook
complete, transition to melt formation should be as rapid as possible to minimize growth of belite and CaO crystals. This transition is delayed by a long, lazy flame which may be due to poor mixing, coarse coal, or insufficient burner tip momentum (Weil-trauch; Influences of Burning Process on Clinker Grindability; Polysius Corporation Technical Seminar, October 1997). Higher clinker SO3 gives harder grinding and higher free lime gives easier grinding (increasing FL by 0.2% reduces finish mill power requirement by ca 1%). Any reduction in power which can be obtained by varying mix design and burning practice must be considered in the context of overall manufacturing cost and cement quality. Use of high-chrome charge and liners is essential to avoid frequent sorting and recharging. However, tramp metal may occasionally require separation and, eventually, the smallest balls will wear to the point where they begin to plug the discharge screen. Ball usage should be of the order of 30-50g/tonne cement. Mill first compartments typically have balls sized 60-90mm and second compartments 20-50mm (Fuhrmann; ZKG; 9/1991, pg E199). First compartments have lifting liners to promote impact. Second compartments are usually fitted with classifying liners if a range of ball sizes is employed; it has been observed, however, that classifying liners are ineffective unless feed rate and grindability are constant. Also, too high a circulating load will tend to cause reverse classification. While coarse grinding benefits from a range of ball sizes, greater efficiency of fine grinding may result from single sized small (ca 20-25mm) balls. Note that discharge screen slots must be about 3mm wider than diaphragm slots. Air flow is typically 1-2M/sec referred to the open mill cross section above the ball charge and separator vent air for cooling is some 2-3NM3 /kg of mill throughput. Ball mills normally operate at about 75% of critical speed and 25-35% volumetric charge loading. Circulating load is typically 100-200% for raw milling, 200-300% for finish mills with mechanical separators, and 150-200% with high efficiency separators. Specific power consumption usually decreases with lower charge loading but so does production rate. Since other circuit power is essentially fixed, the design charge loading and rated mill power should normally be maintained.
A useful test is to shut the mill down on load and turn the fan off immediately. On entering the mill, it should be observed that the top layer of balls in the first compartment half floats in feed; exposed balls require more smaller balls to fill voids, or the blanking of an outer annulus of the diaphragm, or the closing of scoops on an adjustable diaphragm (Gudat & Albers; ZKG; 3/1992, pg 73). Also, samples should be taken every half meter through both compartments for screening. Any divergence from progressive size reduction indicates flow or charge problems. If negligible particles in the second compartment are greater than 3mm, then the ball size and/or loading in the first compartment may be reduced. Diaphragm and discharge screen should be inspected for wear, holes, and plugging. The energy efficiency of ball mills is very low, particularly for coarse grinding. In recent years, various circuits have been introduced incorporating hammer mills, roll-presses, and roller mills for pregrinding ahead of the ball mill. An impact crusher in closed circuit can yield product with 100% -3mm; this approximates to the size of material passing from first to second compartments of a 2-compartment mill so that the large ball charge can be replaced by ca 20-25mm balls with a potential mill capacity increase of 20% and a combined system specific power reduction of 5-10% (Stoiber et al; ZKG; 3/1994, pg E94). Roller mills employ a mixture of compression and shearing which reduces power consumption, but the rate of wear for rollers and table in clinker grinding has been a problem. The use of specially heat-treated high-alloy chromium steel is now claimed to give 20,000 hours before resurfacing (Ellerbrock; ZKG; 4/1994, pg E104) and of late, roller mills have been installed as stand-alone cement mills with an energy savings of some 40% relative to ball mills (Shimoide; WC; 9/1996, pg 68). Roll presses have been used successfully for pregrinding in a variety of circuits and they have also been used as stand-alone cement mills (Feige; ZKG; 11/1993, pg E305). Many early roll presses suffered from roll surface and bearing failures but, progressively, operating pressures have been reduced, roll sizes increased, and metallurgy improved to achieve satisfactory performance. The product of a roll press is a compacted cake which is either fed in open circuit to a ball mill or must be deglomerated before classifying. It was observed that the particle size and shape of clinker from a roll press resulted in high wear rates on
64 Cement Plant
Operations Handbook
conventional classifiers. A V-separa tor is now available which both deglomerates and classifies in a static configuration of stepped plates down which the material cascades through a cross flow of air (Strasser et al; ZKG; 3/1997, pg 140). Various roll press circuits are shown in Figure 8. FCB have addressed the reconciliation of efficient compression grinding with operating simplicity, and have developed the Horomill. This is a cylindrical mill shell, rotating above critical speed, with a single idler roller and internal fittings to control the flow of material (ICR; 2/1995, pg 29). The roller is in free rotation but is hydraulically pressed against the shell. The mill may operate either in closed circuit with a separator or be used for pregrinding in a ball mill circuit. The Horomill is approximately one third the length of the equivalent ball mill arid, when standing alone, power savings of 30-50% are claimed. Specific power consumption for clinker/gypsum grinding in a ball mill should be approximately: 3000cm2 /g 3200 3400 3600 3800 4000 24.4kWh/t 26.8 29.4 32.0 34.7 37.5
Combined power can be reduced by some 20% with pre-grinding. Abnormally high power consumption may be due to mill inefficiency, but is as likely to be caused by overturned clinker. Laboratory grinding techniques to determine standard grindability indices (kWh /t) are ubiquitous. Theisen (WC; 8/1993, pg 17) has developed an empirical formula to correlate grindability to alite (C3S) and belle (C2S) crystal sizes determined by microscopy: kWh/ t (@ 3500cm2 /g Blaine) = 33.4 x 0.32C3SN + 0.27C2S - 13.2TEA where C3SN = mean alite size, = % belite C2S = % grinding aid TEA
Figure 8
00
Roll press Bell mill
PREGRIND CIRCUIT
Separator Feed
Product
00
Roll press Ball mill
HYBRID CIRCUIT
COMBINED CIRCUIT
The effect of changing fineness on mill production rate and specific power consumption is approximated by: (Blaine 1 / Blaine 2) 1 .5 = (t/h 2 / t/h 1) = (kWh/t 1 / kWh/t 2) where Blaine = cm 2 /g Mill discharge temperature should be kept below about 110C though partial dehydration of gypsum in the mill, up to about 60%, increases cement strength without excessive false set (Jorgensen; ZICG; 10/1988, pg 497). Water sprays may be used in the mill to control temperature but this will, inevitably, sacrifice some cement strength; cool clinker or air sweep of the mill or separator are to be preferred. High mill temperatures also exacerbate material agglomeration and coating of balls and liners, significantly increasing specific power consumption. Grinding aids (usually ethylene glycol based or triethanolamine based) may be added to the ball mill to reduce such electrostatic agglomeration (Sumner; ICR; 11/1993, pg 72). The optimum addition rate should be determined which balances enhanced grinding against power savings to minimize cost. Grinding aids also serve to reduce pack set which is a cement handling problem. More recently, composite grinding aids have been developed which yield significant strength increase in both mortar and concrete; it has been observed that response depends upon clinker chemistry, most notably when ferrite (C4AF) is in the range 712% (Slim et al; ICR; 6/1997, pg 46). Ball mill monitoring should include: Production rate, tonnes/hour Operating hours Involuntary downtime hours kWh/tonne mill drive Connected power, % of mill motor rating Cement temperature, C Grinding aid usage, grams/tonne Ball usage, grams/tonne Circulating load, % Additional monitoring will be required if pregrinding is effected; specifically the combined kWh/tonne of both units will be of importance. Particle size may be determined by a number of techniques but is usually monitored by air permeability through compressed powder (Blaine
z
s
Cement Plant
Operations Handbook 67
method). Particle size distribution is represented by the Tromp curve (Section B4.4). The ratio Blaine:-325# should be minimized; typical is 4042 for mills with mechanical separators and 36-38 for high efficiency separators. It is considered that the particle size fraction below 311 contributes most to 1-day strength though it also increases water demand, while 3-2511 governs 28-day strength. Note that hydration only penetrates 3-411 in 28 days. Particles larger than 2511 make a negligible contribution to strength. Typical particle size distribution for a cement of 3600cm 2 /g Blaine is:
Below 18011 150 100 80 60 40 % 100 99.9 99.3 98.5 96.0 88.2 Below 3011 20 15 10 8 6 % 81.6 64.8 53.4 40.4 34.6 28.4 Below 511 4 3 % 24.9 21.2 17.2 12.8 7.1 1.9
2
1 0.5
Fine cement with a narrow particle size range (as is possible with high efficiency separators) gives high mortar strengths but, it should be noted, may also give rise to high water demand which will yield low concrete strength. Concrete strength at fixed workability, not fixed water, is ultimately critical, and standard mortar strengths are significant for process control and specification only - not directly for concrete product design. This should be considered in retrofitting a high efficiency separator to an existing mill; replacement is usually justified by increase in production due to reduced Blaine requirement for a given mortar strength and the benefits do not always carry through to concrete (Detwiler; ZKG; 7/1995, pg 384 & 9/1995, pg 486).
5.3 Separators
Several types of separator are employed in mill circuits and there are numerous variations of each type: Grit separators (Figure 9.1) are used to dedust mill air-sweep. They have no moving parts and effect separation by the cyclonic air flow induced by guide vanes. Radial setting of the vanes gives minimal, and near tangential gives maximum, dust separation. Grit separators should be adjusted so that dust collector catch has the same 325# as product; the dust collected can then be conveyed directly to product cement.
Figure 9
Gas discharge with fine dust entrained to dust collection and fan Gas discharge with fine dust entrained to dust collection and fan
Rejects
Drive
Feed
1'
.7"
\11
Main fan Control valve Auxiliary fan Dispersion plate Return vanes
I Drive
Gas discharge with fine dust entrained to dust collection and fan Rotor
Mechanical separators (Figure 9.2) are the traditional classifiers of mill product. The material is fed onto a rotating dispersion plate whence it is spun off into a rising air stream. Coarse particles either fall directly from the dispersion plate or are rejected between the auxiliary fan blades and the control valve. Fine dust remains entrained through the main fan and becomes detrained as the gas flows downwards with both decreasing velocity and diversion through the return vanes. Operating adjustments are the number of auxiliary blades, the clearance between auxiliary blades and control valve, and the radial position of the main fan blades. High efficiency cyclone separators were introduced to improve on the mechanical separator's low efficiency in fines recovery (Onuma 8z Ito; ZKG; 9/1994, pg 535). A simplified process flow for the 0-SEPA (Figure 9.3) is as follows. Material is fed onto a rotating dispersion plate whence it is dispersed into the classifiying air stream which is sucked from tangential inlet ducts through fixed guide vanes. Separator loading is up to about 2.5kg feed/M 3 air flow A horizontal vortex is formed by the rotor which classifies particles between centrifugal force and the inward air flow. The fine fraction exits upwards with the air exhaust for subsequent dust collection while the coarse fraction falls and is discharged from the bottom of the vessel. Fineness is increased by increasing rotor speed. Rotary classifiers, used integrally with a roller mill (Figure 9.4), involve the upward flow of dust-entrained air with decreasing velocity to a static or rotary classifier; coarse particles fall back to the mill while fines leave with the air exhaust for external dedusting. Possible mill circuits are legion, from a single open-circuit mill to a combination of mills and separators (Onuma & Ito; WC; 9/1994, pg 80). The situation is rendered particularly interesting when increased performance is required from existing equipment and new units are combined with old. The only generalization which can be made is an obvious caution that balance and control are essential.
With cement storage typically costing $250-400 per tonne and representing 10% of total plant capital cost, there is strong pressure to minimize storage capacity and to minimize the number of silos for a given capacity. Dagnan (Proc IEEE Cement Industry Conference; Salt Lake City, May 1986, pg 131) suggests a guideline of 30 days production and details the factors involved. He specifically addresses the economics of small volume special products. Cement should, if possible, be below 60C when conveyed to silos. The storage of cement at greater than 80C causes dehydration of gypsum and reaction of the released water with fine cement particles resulting in loss of early strength and promotion of false set in the cement as well as build-up in the silos. If high temperature is an intractable problem, the effects can be mitigated by reducing gypsum addition and replacing up to half the gypsum by anhydrite (Reid; WC; 4/1997, pg 104). After the mill, cement can be cooled using a dedicated cement cooler (Kochmann & Ranze; ZKG; 10/1997, pg 556) or by water jacketing a pneumatic conveying line. Build-up and the dubious performance of many silo level detectors may lead to erroneous inventories. Graham (ICR; 2/1997, pg 51) describes a low-frequency acoustic technology for flow enhancement and silo cleaning. Manual measurement should be made at a point 2/3 of the radius out from the center and measurements should be reconciled with metered filling and discharge of silos to detect anomalies. 5.5 Cement Dispatch In general, shipments of cement range from almost wholly bulk in developed markets to almost wholly sack in less developed countries. The economics of distribution are complex and have important effects upon plant locations and capacities. It has usually been accepted that a plant can effectively serve a radius of about 300kM using road transport. Betty & Scott (ICR; 7/1993, pg 36) give formulae for estimating unit bulk shipment cost in the United States:
23-25tonne road truck = $10/t/day + [$0.0276/t x kM(one 2000t rail shipment = $0.53/t + [$0.0212/t x kM] 25,000t sea shipment = $1.90/t + [$0.00075/t x kM]
r-
way)]
Thus, a 320kM (200 mile) haul would cost approximately $19/t by road and $7.50/t by rail. This, of course, may not reflect door-to-door cost
71
where road haulage normally offers an advantage. Road transport also is usually more flexible than rail for small, irregular, or short-notice shipments and is more likely to offer direct delivery to point of use. By comparison, 6000kM trans-Atlantic sea freight would cost about $6.50 (subject to the cyclical variation of shipping costs). It is obvious, therefore, that coastal areas are very exposed to low cost imports unless protected by tariffs. And certain routes can be served by ships on backhaul at even lower rates. Similarly it can be seen that plants with deep water docks have almost unlimited market potential. Polansky (ICR; 12/1995; pg 47) reviews the operation and economics of bulk carriers of greater than 25,000DWT. Economic bulk distribution requires minimal demurrage costs for filling and discharging bulk vessels. Such costs, if incurred, should periodically be compared to the cost of upgrading loading/unloading equipment. Sacking adds approximately $4-6/t to production cost and can, of course, increase shipment costs due to the additional handling involved. Sacks are most efficiently handled either on pallets or shrink-wrapped (WC; 7/1992, pg 42). Cement sacks are typically 94Lb, 50kg or 40kg net weight. Anyone who has seen 40kg girls handling 50kg cement sacks in Asia must be bemused by the need felt by the European Union, at considerable expense to producers, to limit sack weight to 25kg as a health and safety issue (Reitmeyer; WC Bulk Materials Handling Review; 1993, page 32). Sack weight standard deviation should be determined in order to minimize overweight. ASTM C150 Sec 11.3 specifies that packages more than 2% underweight may be rejected, and that if 50 random sacks average less than the stated weight, the whole shipment may be rejected. Thus, with an average of the stated weight and a standard deviation of less than 1%, fewer than 1 in 20 sacks could be rejected. Any margin of designed overweight is at management discretion, but a conservative target would be the stated net weight plus the mean bag weight (200-250g for 40/50kg bags) plus 1 x standard deviation. The major cost element in sacking is the 'craft paper and this has historically been a commodity volatile in cost and, occasionally, in short supply. If a plant has a significant sack market, therefore, some effort is
Handbook
justified in purchasing and stockpiling paper and in optimizing sack configuration against breakage (Grundy; ICR; 7/1993, pg 50). In-plant sack breakage of less than 0.3% should be maintained and this can be assisted by careful attention to sack handling and load-out conveyors. All sacks should be date-marked, albeit in code. This gives a means of tracing the cement's history in the event of subsequent complaint. It is not unknown for complaints to be received of hydrated or low strength cement which may be years old. Records allowing the tracing of questioned shipments to silo and production are also advisable. Bulk loading of road or rail trucks is normally about 400t/h, and of ships and barges considerably faster. 25T road trucks can be loaded at 6-minute intervals from a single loading spout with appropriate equipment and procedures. The filling spout is telescopic and has a conically tapered nozzle to seal against the tank opening. The filler comprises a central cement delivery pipe surrounded by an outer suction tube which vents at least 1000M 3 /h and is ducted to a dust collector. Load weight is determined either by tare and load weights of the truck (preferably on dedicated truck scales under each loading point), or by loading from a hopper on load cells which is itself weighed before and after loading and which can be programmed for a specific delivery (Brix & Pauer; WC; 3/1995, pg 106). To prevent spillage, the loading spout should be fitted with a level sensor and cut-off switch. The design of modern road tankers is reviewed in ICR; 1/1995, pg 40. A third mode of shipment is the "big bag", a 0.5-2 tonne flexible bulk container comprising a woven polypropylene outer ply and a polyethylene liner. The bags may be single or multiple use and can be handled by slings or forklifts. There tends to be little, if any, cost advantage over sacks, and to date the use of big bags has been limited to special situations such as oil well cement (Wild; ZKG; 5/1994, pg E148) or as temporary "silos" for small batch plants.
-n
r-
1-1
shut-down due either to kiln interlock or to electrical failure, and this must be acceptable under emission regulation. Debell & Yong (WC; 6/1995, pg 22) describe the potential for upgrading existing EPs.
Bag filters comprise filters of either woven fabric (which employ bag
shaking or reverse air flow for cleaning) or needle felts (which are cleaned by reverse air pulse). Air to cloth ratio (M 3:M2/h) should be 30-50 for woven fabrics and 100-140 for needle felts. Pressure drop is typically 150-250 mmH2O and efficiency 99.95%. The normal fiber is
M O
6 POLLUTION CONTROL
polyester which can operate up to 150C, while polyamide can be used to 230 and glass fiber to 280. Reverse pulse air should be dried and of 6-8kg/cm 2 pressure, and bag tension approximately 0.4kg/cm of bag circumference. Bag house pressure drop can be reduced and, thus, capacity increased by use of acoustic horns (WC; 1/1995, pg 26). While the risk of explosion with unburned fuel or reducing conditions is obvious for electrostatic precipitators, it should also be recognized that, unless anti-static bags are employed, a similar risk attaches to bag filters. Conditioning of kiln exhaust gas is necessary before dust collection. The exhaust gas from many kilns is used for drying raw materials and this process serves both to cool the gas and to raise its humidity before dust collection. The gas temperature should be below 170C for both baghouses and EPs and, for the latter, a moisture content of ca 15% (v/v) is required in exhaust gas. It is important, however, that the gas should not be allowed to pass through the dew point before release,
especially if significant S or Cl is present. A conditioning tower is usually provided for periods when raw mill or dryer are bypassed to allow gas cooling by water spray. The alternative is to add tempering air but
this significantly increases total gas flow and fan power consumption, and may lead to derating of the kiln while bypassing the raw mill or dryer. The operation of conditioning towers is notoriously problemati-
cal due to the dirty atmosphere in which the water sprays must function and to the large turn-down necessary for control; various spray systems are claimed to meet these needs (Berube; WC; 3/1996, pg 77). 6.2 Pollution Control
Environmental regulation is, of course, very much a matter of national and local ordinance. However, certain generalizations can be made about air, water, solid, and noise pollution. Cement plants are primarily concerned with air emissions. Water discharge should not be a concern beyond handling normal domestic waste and storm water run-off with its potential for leaching from stockpiles and spillage. The temperature of discharged cooling water may also be subject to control.
mon in basic brick, has largely been eliminated due to its alleged toxicity and all used refractory may now be incorporated into kiln feed (after crushing and grinding) or sent to landfill. Note that if the limestone quarry also produces aggregates, recovery through the crushing system must be carefully managed as basic brick contamination of aggregate can be catastrophic due to expansive hydration. The only other likely solid waste will result if kiln dust is discharged to relieve a volatile cycle in the kiln. Hitherto, landfilling such dust has involved minimal expense, but increasing regulation, particularly in the United States (EPA Regulatory Determination on Kiln Dust issued January 1995), may eventually encourage processes to recycle rather than landfill. Passamaquoddy Technology (ICR; 3/1991; pg 45) and Fuller Company (WC; 3/1993; pg 31) offer processes for dust recovery. Material spillage can be collected by either fixed or mobile heavy-duty vacuum systems (ICR; 12/1991; pg 49) and, preferably, returned to the process. Noise originates primarily from fans/compressors and screw/ drag-chain conveyors. Noise affecting plant workers can be controlled by ear protection. Noise at the property boundary is best considered at the design stage but can be mitigated by enclosure, insulation and, for fans, sound attenuators (Fuchs; ZKG; 7/1993, pg E185). Although most fan problems occur with short exhausts, some tall chimneys have been found to resonate unless silenced. Quarry blasting is a specific problem which may involve charge design and timing to minimize disturbance to neighbors.
t1
11 I 101
emissions, detached plumes resulting from hydrocarbons or ammonium compounds can present a problem which is not solved by conventional dust collection. Considerable understanding of detached plumes has been acquired by certain plants and some information has been published (Guemez-Garcia & Ganatra; WC; 12/1994, pg 50). In the United Kingdom, new dry-process plants should meet the following standards: Particulates 50mg /M3 (Clinker coolers - 100mg/M 3) SO2 750mg/M3 1200mg /M3 NOx (HMIP Chief Inspector's Guidance to Inspectors; Process Guidance Note IPR 3/1) While emission regulations stipulate particulate levels, actual measurement is a protracted procedure involving isokinetic sample collection. Automated systems are available for continuous opacity monitoring which measure light attenuation across the stack. Some regulators also recognize visual estimations such as the Ringelmann Smoke Chart. The Ringelmann chart is a series of cards with increasingly dense cross hatching representing opacity from 0-100%; where applicable, emissions should not normally exceed 10% opacity (between Ringelmann 0-1).
Card 0 1 2 3 4 5 All white 0% opacity 20% 40% 60% 80% All black
100%
Dust suppression can be important in both quarry and plant where dry materials are handled and where unpaved surfaces are used by mobile equipment. Various engineered and chemical-spray systems are reviewed by Carter; RP; 5/1995, pg 19. CO is formed by the incomplete combustion of carbonaceous materials. Oxidation to CO2 takes place in the presence of excess oxygen at temperatures above about 680C. CO in stack emissions is usually attributed to overall deficiency of oxygen in the burning zone or to poor fuel/air mixing. However, Sadowsky et al (ZKG; 5/1997, pg 272) have found that many cement raw materials contain 1.4-6g organic carbon
per kg clinker, that these oxidize below 680C, and that 10-20 % of the oxidation results in CO irrespective of the level of excess 02 (Note that 2g C/kg clinker with 15% conversion yields about 250ppm CO at 5% excess oxygen). Obviously this is too low a temperature for post oxidation to CO2 and suggests that some CO observed at the stack may not indicate combustion problems and may not be easily rectifiable. NOx is formed during fuel combustion by oxidation of nitrogen compounds in the fuel (fuel NOx) and of the nitrogen from combustion air (thermal N0x). Thermal NO increases with flame temperature above 1200C, with retention time, and with increasing free oxygen. NO produced in the kiln is related to: s flame temperature s flame shape s type of fuel s excess 02 s gas retention time in burning zone s load temperature s load retention time in burning zone A short, hot flame, excess combustion air, and high secondary air temperature all raise NO, while fuel moisture, dust insufflation, or a high dust entrainment in secondary air reduce NO formation. Also, a hard burning mix will lead to increased NO and, certainly, overburning is to be avoided. In most precalciner vessels, combustion occurs at less than 1200C so that mainly fuel NO is produced from secondary firing. NO may be oxidized to NO2 at lower temperatures but normally NO comprises more than 90% of NOx emissions. NOx is, however, conventionally expressed as NO2 for regulatory purposes. NO produced by secondary firing depends on: s fuel N content s excess 02 s flame temperature
-0
a
0
2
0
2
Cement Plant
Operations Handbook
79
Some reduction of thermal NO produced in the kiln can occur in the precalciner at temperatures 800-1100C by the reaction: [N] + NO N7 + [0] Gross NOx emissions are typically in the range 500-1500ppm and, being so closely related to kiln combustion conditions, measurement of NOx provides a useful means of monitoring kiln operation. Recently, pressure has been applied by regulators to reduce NOx emissions and various techniques have been developed: LOW NOx burners which are designed to recirculate low-oxygen, used combustion gas through the burner (Mullinger; WC; 12/1994, pg 27). Stepped combustion gives NOx reduction of 30-50%. In this process fuel is burned under reducing conditions in the kiln riser at 850-950C with subsequent addition of tertiary air for re-oxidation of CO (Deussner; WC; 12/1995, pg 52). Catalytic reduction of NO using ammonia or urea has also been successful (Steuch et al; ZKG; 1/1996, pg 1) and Polysius have developed activated carbon filter technology (Polvitec) to scrub NOx, SO2, NH3 and other pollutants from kiln exhaust (Rose & Brentrup; ZKG; 4/1995, pg 204). Haspel et al (ICR; 1/1991, pg 30) working with NOx as a control parameter for kiln operation, have discovered that, although the NOx produced is mainly thermal, it is a good indication of burning zone condition only with burners which provide stable flames with good recirculation. With non-recirculatory, low primary air burners, there are interferences with its correlation to burning zone temperature. Specifically, non-robust burners can yield CO of more than 1000ppm (0.1%) with 2-3% 02 at kiln inlet and this can totally confuse logical kiln control responses. Petroleum coke combustion, too, is particularly sensitive to secondary air temperature. Thus, additional factors affecting NOx at kiln exhaust include: s s s kiln atmosphere (NOx is degraded by CO > 3000ppm) alkali cycle increases rapidly with burning zone temperature and with reducing conditions secondary air temperature
SO2 is produced in the kiln both by oxidation of fuel S and by decomposition of sulfates. SO2 thus produced is almost totally scrubbed by K2O, Na2O and CaO in the cyclone preheater. The lower volatility of the alkali sulfates leads to their predominantly exiting with clinker (unless relieved by a gas bypass) while CaSO4 will largely revolatilize in the burning zone and results in a sulfur cycle building up at the back of the kiln and the lower preheater cyclones. In extreme cases, this cycle will cause accretion and blockage problems unless relieved by a gas (or meal) bypass. This is exacerbated by the hard burning required for low alkali clinker and frequently leads to strict sulfur limits on feed and fuel. Sulfides and organic sulfur in raw materials, however, oxidize in the preheater and largely exit with exhaust gas. SO2 can theoretically oxidize to SO3 at low temperature but, in practice, more than 99% of gaseous sulfur will be SO2. With SO2 emissions being increasingly regulated (the US Clean Air Act mandates limits of 2.5Lb/million BTU or 4.5kg/million kcal by 1995, and 1.2Lb/million BTU or 2.2kg/million kcal by 2000), the only solution if such raw materials cannot be avoided is to scrub the exhaust gas. There is an extensive literature of NOx and SO2 chemistry and mitigation (eg Nielsen & Jepsen; FL Smidth Review #96) Unfortunately, regulators have not yet extended their jurisdiction to natural phenomena such as the 1991 eruption of Mount Pinatubo which is estimated to have injected into the atmosphere SO2 equal to between 10 and 100 times the present worldwide annual production (Economist; 21 Nov 1992, pg 97). Gas analysis for process control and emission monitoring involves a wide range of proprietary instruments. For occasional measurement of most gases of interest, absorption tubes with a small syringe pump provide a simple, accurate, and low cost method (MSA, National Draeger, Sensidyne). emissions are of particular concern where alternative fuels are burned. It is generally accepted that dioxins and furans will not be released in significant quantity if burned under the following conditions:
Dioxin
-a
c
0
73
Minimum temperature Minimum retention time Minimum excess oxygen - liquid fuels - solid fuels Maximum CO (@ 11% 02)
(Krogbeumker: ICR; 5/1994, pg 43)
Cement kilns can easily meet these requirements with adequate process control and well-designed burners. The emissions associated with alternative raw materials and fuels are reviewed by Krogbeumker (as above) and Uchikawa (WC; 12/1994, pg 58). Although it does not help in meeting regulatory limits, the level of risk associated with dioxins appears based more on political expediency than scientific evidence (Rigo; Solid Waste Technologies; 1-2/1995, pg 36). It may be noted that there are 75 chlorinated dibenzo-p-dioxins and 135 chlorinated dibenzofurans, all with different toxicities. Many regulators recognize a scheme of "International Toxic Equivalents" (ITEC). Following are the basic structures:
Toxic heavy metals of common concern, in decreasing order of volatility, are Hg, Tl, Cd, Se, Sn, Zn, Pb, Ag, Cr, Be, Ni, Ba, As, V. They are ubiquitous in trace quantities, and the manufacture of cement from natural minerals is usually of negligible consequence as most metals are retained in the clinker (Sprung & Rechenberg; ZKG; 7/1994, pg F183). The introduction of alternative raw materials and fuels requires monitoring both of inputs and of distribution between cement, kiln dust (if discharged), and gaseous exhaust (ICR; 5/1992, pg 71: von Seebach & Tompkins; RP; 4/1991, pg 31). Regulation varies with jurisdiction but has, unfortunately, frequently lost sight of the value of the cement kiln to destroy organic wastes and
to encapsulate waste metals in concrete. Also, the legal position has become confused by consideration of waste burning kilns as incinerators, by considering the products of processes incorporating hazardous waste materials as themselves automatically hazardous, and by assumptions that any measurable toxic metal is dangerous even when lower than its natural elemental occurrence in the earth.
0
1
7 QUALITY CONTROL
The plant laboratory is responsible for: s Analysis of raw materials and fuel. s Mix design which should reflect individual raw material costs, production costs (principally drying, grirtdability, and burnability), and product quality s Preblending of raw materials and blending of kiln feed to minimize chemical variation into the kiln. s Process control which involves raw materials moisture, raw meal fineness, clinker free lime, degree of calcination and volatiles contents of hot meal at the kiln inlet, proportioning of clinker and gypsum, and fineness of milled cement. s Chemical analysis and physical testing of cement to confirm and certify compliance with specification. s Miscellaneous functions such as water treatment and laboratory measurements associated with process engineering and pollution control.
should beware of the computer generated calibration with a zero slope and an intercept equal to the previously unvarying result. It should be noted that x-ray fluorescence is not normally a primary method of analysis and must be calibrated from standard samples though standardless XRF is useful for approximate analysis of materials for which no calibration has been established (Moller; WC; 9/1995, pg 75). Also, even more than for wet analysis, the actual sample analyzed is of the order of micrograms and requires considerable faith in the sample reduction procedures which may start with several hundred tons of process material. Raw material sample size should reflect material particle size and homogeneity. The sample should be crushed and split repeatedly to obtain the analytical sample. This is why the recently applied neutron activation system which analyzes a total process stream is instinctively preferable.
Neutron activation using an isotope source of neutrons results in characteristic secondary gamma radiation. Gamma radiation being of much higher energy than x-radiation (0.5-5mev vs 1-50kev) is less subject to attenuation so that bulk material (10-100cm thick stream) can be analyzed where only the sample surface is sensitive to XRF. The gamma spectrum is resolved by energy dispersion allowing rapid, on-line analysis of process streams without sample preparation. If layered materials are to be measured, as on a raw mill feed belt, it is advisable to pass the mixture over a transfer point to homogenize the stream. Accuracy and precision now approach that of XRF and sampling error is drastically reduced (Bourgon et al; WC; 5/1995, pg 46). X-radiation only penetrates a few microns into the sample surface. Some minerals are prone to render the surface nonrepresentative; examples in raw materials are quartz which may be present in large particles and platy micreous minerals. Such materials may require fusion to avoid analytical bias (Philips offers the Soled Perl'X system). Pedersen & Wismann (WC; 5/1997, pg 34) review XRF sample preparation and make an interesting point in estimating relative contributions to total analytical error as: XRF measurement 10-15% Sample preparation 25-30% Sample collection 55-65%
86 Cement Plant
Operations Handbook
Each element in the sample attenuates the x-radiation of other elements to varying degrees so that interelement corrections are necessary and calibrations are not linear. While wide matrix calibration programs are available, most calibrations used for process control apply only to a narrow range of composition. Discretion should be exercised to avoid the use of inappropriate calibrations; for example, the use of a kiln feed calibration to analyze bypass dust. It should also be remembered that most analytical methods measure elemental concentration; thus Ca is measured but, by convention, CaO is reported. X-ray powder diffraction does offer the ability to identify and quantify chemical compounds such as CaO, CaCO3, CaSO4, CaSO4. 1/2H20, CaSO4.2H20, or C3S. Powder diffraction offers relatively low precision but is useful for comparative studies of kiln chemistry (Beilmann; ICR; 8/1993, pg 46), and is increasingly used for rapid determination of free lime in clinker (Bonvin et al; WC; 8/1994, pg 11). Indeed, to meet the requirement of rapid analysis and adjustment for efficient control of a continuous operation, increasing use is being made of automatic systems for sampling, sample conveying and preparation, and for chemical and physical measurement (Buhrer & Vohren, WC; 11/1994, pg 26). In the absence of x-ray fluorescence, kiln feed must be designed and monitored largely from the acid/base titration of carbonate. For accuracy and precision, this depends upon meticulous manual analysis under conditions of mind-numbing routine: possible but unlikely. The routine wet analysis of all major oxides (Si, Ca, Al, Fe, Mg, K, Na, S) in composites of individual raw materials, kiln feed, clinker, and cement is becoming less common. While various classical (wet) analytical methods may be employed, specifications usually give details of methods for each component which should be used as reference test methods in case of dispute. These methods, in turn, should be used for analysis of standard samples (eg US National Bureau of Standards standard reference materials, "SRMs"). It should be recognized that the maximum acceptable difference between the average of duplicate analytical results and the SRM certificate values (ASTM 0114, Table 1) are:
1im<
O 2 O
23
0.05%
These limits are important in themselves and indicate that no oxide, except alkalis, should be reported with greater precision than 1 place of decimals, though raw data for statistical calculation should retain 2 places. Thus, if a laboratory reports an oxide to 2 or more places, or a cement compound to more than the nearest whole number, this would suggest that they do not understand precision and may well not understand much analytical chemistry either. The laboratory should possess a flame photometer for alkali determination and a bomb calorimeter for solid fuel. An automatic sulfur analyzer (for fuel) may be useful, and clinker microscopy should at least be available as a service if the plant laboratory is not itself equipped.
7.2 Sampling
Sampling involves both the collection of a portion from a large quantity of material and its reduction to a small fraction which is representative of the whole. This is as important as analysis and is often much more difficult, particularly where large particle (rock) size and heterogeneity require that large samples are taken and reduced systematically by successive crushing and splitting to the quantity actually analyzed. The ASTM procedure for collecting aggregate samples is described in D75, and the reduction process in C702. The standards for coal are D2234 and D2013 respectively. Theoretically, the required sample size is estimated using standard formulae and predetermined factors for heterogeneity, particle size, and required precision (Pitard; Heterogeneity and Sampling, Vol 1; CRC Press, 1989). There are also various simpler guidelines for sample size such as: 1 Between 0.2% and 2% of total flow depending upon heterogeneity and particle size (Labahn).
Minimum Sample Max Particle Size 10mm 1 Okg 25 20 30 60 80 40 100 50 120 60 150 75 175 90 (ASTM D75; Standard Practice for Sampling Aggregates)
.0
C
For most practical purposes, much smaller samples than this may be used. For raw materials, some 50x the weight of the largest particle (rock) may suffice while at each splitting stage, the largest particle should not exceed 2% of the total fraction weight.
0 z
0
s The optimal use of analysis and storage facilities to blend the kiln feed to minimal variation. s The management of cement inventories For convenience, process control usually employs composite terms such as Bogue compounds, lime saturation factor, silica ratio, etc (defined in Section B5.1) rather than individual oxides for monitoring deviations and calculating statistical variation.
7.5 Microscopy
Clinker microscopy is long established, but became a widespread tool for process control and problem solving with the introduction of the Ono powder-mount method in the 1970s. There are mixed opinions of the Ono method as a routine procedure for kiln monitoring and cement strength prediction, but the technique, together with thin sections and reflected light microscopy, constitute a valuable resource to a plant laboratory. The Ono powder method, in gross simplification, involves crushing clinker and screening to -200#/+325# (74-44g). 1.715 refractive index immersion oil is used to mount the sample for microscopic examination using transmitted polarized light. Measurement is made of alite (C3S) size, alite birefringence, belite (C2S) size, and belite color: Alite size is a measure of kiln temperature rise between ca 1200-1450C where C2S is combining with CaO to form C3S. Rapid heating is desirable and is indicated by alite of 15-20p., while slow heating produces alite of 60g or above.
Alite birefringence, the difference between the refractive indices of blue and red light, is related to maximum kiln temperature. High temperature is desirable and is indicated by birefringence of 0.008-0.010 against 0.002 for a cool burning zone. Belite size reflects retention time in the burning zone above 1400C.
Maximum retention is preferred and is indicated by average crystal length of 25-40g while short retention yields 5-10g. Belite color relates to the rate of initial cooling to below 1000C
(ie cooling in the kiln before entering the cooler). Rapid cooling is desirable and results in clear belite crystals while slow cooled clinker gives yellow to amber color. Ono proceeded to correlate these four measured parameters with 28-day mortar compressive strength. Parallel techniques are the inspections of polished sections prepared by embedding the same 74-44t powder fraction in epoxy resin, cutting the cast, and polishing the surface for examination under reflected light. mounting whole nodules of clinker in epoxy resin, cutting and polishing, also for reflected light examination. A typical photomicrograph is shown below. cutting a thin section of an epoxy resin mount for examination using transmitted light.
0 a
2
1-1
0
Alite (C35)
These techniques allow much additional data to be collected but are not as simple or rapid to perform as Ono's method. Direct estimates of cement compounds can be made (as opposed to Bogue calculations), free lime can be estimated, crystalline MgO can be identified, and certain phenomena such as belite nesting can be observed (Chromy; ZKG; 12/1992, pg E325). The latter correlates to coarse silica particles in kiln feed and has an adverse effect upon 28-day and later strengths. A review of the Ono method has been made by Campbell in a 1992 PCA report, "A Summary of Ono's Method for Cement Quality Control with Emphasis on Belite Color". Dr. Campbell has also published a standard text "Microscopical Examination and Interpretation of Portland Cement and Clinker", available from PCA.
strength may use differing cement contents, water-cement ratios, and mixing and molding methods; all of which can lead to wide disparity in quoted strengths from identical cement samples.
ASTM C150 SPECIFICATION FOR PORTLAND CEMENT Type I Type II Type III Type IV Type V A LA General purpose Moderate heat/moderate sulfate High early strength Low heat of hydration (rare) High sulfate resistance Air-entraining Low alkali
resistance
Associated ASTM test methods & specifications: C109 C114 C151 C191 C204 C266 C359 C451
C465 C563
Compressive strength Chemical analysis Autoclave expansion Vicat setting time Fineness by air permeability Gilmore setting time Early stiffening/false set/flash set (Mortar method) Early stiffening/false set/flash set (Paste method) Processing additions
Optimum SO3
C917
Chemical Limits (%): I Si02, min Al203, max Fe203, max MgO, max 503, max (C3A<8) (C3A>8) Lol, max Insol res, max C3A, max C4AF+2C3A
(Optional)
III
IV
0.60
58 0.60
0.60
0.60
0.60
12
min Fineness, min, cm 2/g 2800 Autoclave exp, max, % 0.80 Strength, min, MPa 1 day 12.4 3 day 24.1 12.4 10.3 7 day 17.2 19.3 28 day (note A cements have lower strength requirements) Gilmore IST/FST, min - 60/600 minutes all types Vicat, IST/FST, min - 45/375 minutes
12 22 16 2800 0.80
12 22 16 2800 0.80
12 22 16 2800 0.80
12
2800 0.80
6.9 17.2
Typical cement compound concentrations for the five cement types are:
C3S C2S C3A C4F
Type
I II III IV V
55 51 56 28 48
19 24 19 49 30
10 6 10 4 4
7 11 7 12 10
Additions are limited to water and calcium sulfate (and air entraining agent). At the option of the manufacturer, process additions may be
used which meet C465. C465 requires comparative testing with/without additives; difference shall not exceed: 1 Water requirement for normal consistency shall not increase by more than 1%. 2 Setting time (Vical initial) shall not increase/decrease by more than the lessor of 11-I or 50%. 3 Autoclave expansion shall not increase by more than 0.10. 4 Strength shall not decrease by more than 5%.
ASTM C595 SPECIFICATION FOR BLENDED HYDRAULIC CEMENTS
Cs C
11
Pozzolan or slag used for intergrinding must satisfy requirements for pozzolanic activity; primarily that a standard intergrind shall have not less than 75% of the 28- day strength of the control cement mortar.
Types covered are: Type IS Type IP Type P Type S Type I(PM) Type l(SM) Portland blast-furnace slag cement containing 25-70% slag Portland pozzolan cement containing 15-40% pozzolan Portland pozzolan cement containing 15-40% pozzolan but with slower strength development than IP. Slag cement containing at least 70% slag normally used with hydrated lime for making masonry mortar Pozzolan modified portland cement containing less than 15% pozzolan Slag modified portland cement containing less than 25% slag
z
0
All of these types can be specified for moderate sulfate resistance (MS), air entrainment (A), or for moderate heat of hydration (MFI).
Chemical Limits (%):
MgO, max S as S03, max Sulfide 5, max IR, max Lol, max Water-soluble alkali, max 3.0 2.0 1.0 3.0 4.0 2.0 1.0 4.0 0.03 5.0
I(5M),I(S)
l(PM),RIP
5.0 4.0
Physical Limits:
Autoclave expansion, max, % contraction, max, % Vicat IST, min, minutes FST, max, hours
I(SM),IS, l(PM),IP
0.50 0.20 45 7
IS(MS) IP(MS)
0.50 0.20 45 7
S
0.50 0.20 45 7
P
0.50 0.20 45 7
4.1 10.3
10.3 20.7
In 1992, ASTM approved an alternative, performance based, specification for blended cements, C1157. Adoption has been slow, but it is anticipated that both this specification and a similar performance based specification for portland cements will eventually prevail in place of the present mix of compositional and performance criteria. European Standards are issued by the European Committee for Standardization. EN 196 covers sampling and both physical and chemical testing of cement. DD ENV 197-1:1995 (draft) covers composition, specifications and conformity criteria. Rejection is permitted under ASTM C150 if any requirements are not met. BS 12 and ENV 197 employ "autocontrol" by the producer which generates statistical test data on cement as shipped; a finite though small incidence of defects is permitted.
Although C595 refers only to blended cements, a common practice is to add the same pozzolanic materials at the concrete batch plant. Pozzolanic materials employed thus are covered by ASTM C618 (fly ash and natural pozzolans) and C989 (blast-furnace slag).
Natural pozzolans are mainly volcanic materials but include some diatomaceous earths. They are soft to grind. ASTM C618 requires that Si02+Al203+Fe203 > 70% and addition rate is 1540%. Fly ash is obtained in large quantities from coal-fired power generation.
The ash is principally an aluminosilicate material in the form of small hollow glass spheres; this results in very low bulk density (ca 0.8t/M 3) which should be considered in transportation. ASTM C618 classifies fly ash as "F" with Si02+Al203+Fe203 >70% and "C" if >50%. Some Class C fly ashes contain sufficient CaO to be themselves appreciably cementitious. Addition rate is 15-40% of total. Alternatively to intergrinding, much fly ash is used as a concrete admixture. Ash is increasingly liable to contain unburned carbon due to the use of low-NOx burners in power boilers. For intergrinding or mixing in concrete, ASTM C618 requires that loss on ignition (carbon) should be less than 6%. High carbon ash may be suitable as a kiln raw material. Typical compositions, with wide variations, are:
Cement
CaO % Mg0 % 5102 % Al203 % Fe203 % Bulk density, t/M 3 65.0 1.5 20.0 6.0 2.5 1.5
Pozzolan
5.0 2.0 60.0 15.0 5.0 1.6
Fly Ash
5.0
2.0
quenched in order to fracture the glass structure and expose active surface. The activity of slag increases with the ratio CaO+MgO/Si02+Al203 and with glass content; the two, however, appear to be inversely related. Addition rate is typically 25-70%.
Masonry Cement is used for mortar in masonry construction where
interground ingredients may be used but commonly 20-50% limestone is incorporated together with an air-entraining agent. ASTM specification C91 defines three types of masonry cement: N, 5, and M with increasing strength requirements. Masonry cement is typically ground to 5000-6000cm2/g
ISO 9000 Contains only guidelines for selection and use of other standards.
ISO 9001 Applies to facilities which design, produce, install and service products and services. ISO 9002 Applies to facilities which provide goods or services to customer to standard specifications. ISO 9003 Applies only to final inspection and test procedures. ISO 9004 Also contains guidelines. ISO 9002 is the appropriate standard for cement manufacture. The standard comprises 19 sections (a 20th applies only to ISO 9001).
1 Management Responsibility
Covers policy commitment, designation of responsibilities and provision of resources.
2 Quality System
Quality Manual and Standard Operating Procedures. 3 Contract Review Establishes customer requirements and ability to meet them.
6 Purchasing
Documented qualification of subcontractors and suppliers, though not necessarily to ISO 9000.
9 Process Control
Documented procedures and equipment maintenance.
18 Training
Establish and maintain training appropriate to individual skills and program 19 Servicing Not usually a contractual obligation for cement.
20 Statistical Techniques
Document where used for process control or product testing.
which yields a concrete with: Slump (ASTM C143) Density 28-day strength (ASTM C39)
Virtually all concrete mixes involve admixtures. These fall into three groups which are self explanatory: s Air entraining agents s Accelerators s Water reducing and set retarding agents also known as plasticizers and super-plasticizers Air entraining agents may be added either by intergrinding in cement or directly to the concrete mix. Desired air content is usually 1-8% by volume, the higher contents being employed where the concrete is subject to freeze-thaw cycles. Increasing air content, however, yields lower strength. It is a curiosity of the cement industry that the product is tested and certified either as a paste of cement and water or as a mortar containing standard sand. Such testing can give only a limited indication of performance in concrete. Specifically, cement strength tests ignore both water demand, which varies significantly with fineness, and possible interactions with the wide range of concrete additives. Optimum SO3 (C563) is the level which maximizes mortar strength at 1-day while concrete containing water reducer usually requires a higher optimum SO3 (Tagnit-Hamou & Aitcin; WC; 8/1993, pg 38). It is also known that high alkali cement requires less air entraining additive so that alkali variation in cement can cause strength variation in concrete. In general, strength decreases with increased water content and the minimum water allowing satisfactory workability should be used. This is a common source of problems as concrete workers find the addition of excess water a convenience. Potential problems with concrete are numerous and their investigation and solution is a specialist task. However, as cement is frequently, and usually incorrectly, blamed for defects, some common areas of contention may be summarized.
ri 0
z
0
Slump loss is a normal phenomenon which takes place with prolonged mixing. Typically, a 125mm slump will fall to 100mm after 15 minutes and 50rnin after 60 minutes. Higher slump losses will occur with porous aggregates, with elevated temperatures, and with superplasticizer incompatibility. Surface cracking may be due to: rapid or premature surface drying of finished concrete excessive working or premature floating which causes a high cement-content surface high slump (excessive water content) dusting on of dry cement to hasten drying Plastic shrinkage cracks are random, relatively deep cracks which form away from the edge of the slab while still plastic. They are caused by: rapid drying of surface low bleeding characteristics (mitigate with chemical admixture or by adjusting fine aggregate gradation) sub base, aggregate, or formwork not pre-saturated Deeper cracks can be due to: high slump concrete with consequent high shrinkage poor aggregate gradation rusting of re-bar too close to surface structural settlement from lack of footings restrained concrete or inadequate relief joints alkali-aggregate reactivity freeze/thaw of non-air-entrained concrete Cracks, when stable, can be repaired by injecting epoxy. Surface scaling is the breaking away of a 1-5mm surface layer and may be caused by: unsound aggregate freezing premature finishing excessive fine aggregate (-100#) Surface dusting of floor slabs is caused by: high slump concrete premature finishing
surface drying water adsorbing formwork Yield Concrete is supplied by volume (M 3 or yd 3) but is batched by weight. The unit weight (density) of concrete is sensitive to air content, and high air mixes may lose air (and volume) due to contamination (eg oil) or to excessive mixing. Commonly, however, yield questions result from incorrect measurement of volume due to uneven ground or sagging forms.
1-1 in<
z
0
Efflorescence is usually due to passage of water through cracks carrying unreacted lime to the surface where it reacts with CO2 to precipitate CaCO3. It is prevented by ensuring that the concrete is watertight. Occasionally, other water-soluble salts such as alkali sulfates can be involved.
improper storage allowing partial hydration. Packset is a phenomenon of electrostatic charging of cement particles which, upon storage and compaction (particularly in bulk distribution vessels), results in resistance to flow. Unlike hydration, however, once packset is broken, the cement flows freely. The electrostatic charging is prevented by addition of cement grinding aids. A standard method for the determination of packset is available from WR Grace & Co, Construction Products Division [62 Whittemore Ave, Cambridge, MA 02140, USA; fax (617) 497 4515]. False set is the result of gypsum dehydration due to overheating of cement (CaSO4.2H20 s CaSO4. 1/2H20). Upon mixing with water, crystallization of reformed gypsum causes stiffening of the mix within a few minutes. Usually this is only a problem with very rapid mixing systems as false set can be broken up after a few minutes and no subsequent problems result. Hot cement is a common complaint. It must be admitted that cement over about 60C can be unpleasant to handle, but, as it constitutes only 10-15% of the concrete mix, its heat contribution is seldom critical. Heat generated by hydration in mass pours, of course, is a different issue and involves the type of cement employed. Alkali silica reactivity (ASR) or alkali-aggregate reactivity, which leads to the habitual specification of low alkali cement by engineers in some countries, is of concern to, but outside the control of, cement producers. While alkali-aggregate reactivity is a serious problem, reactive aggregates are not ubiquitous and the unnecessary specification of low alkali cement frequently entails additional production cost and the disposal to landfill of process dust. Mather (eg Proc PCA-GTC, 9/1982) has argued that cement should be specified with more discretion, as the 0.6% (ASTM C150) limit may be unnecessary for many aggregates while being still too high for some. Furthermore, the known effect of certain pozzolan and slag intergrinds in reducing alkali reactivity should be taken into account.
-
Delayed Ettringite formation (DEF) has recently achieved some notoriety. Ettringite, 3CaO.Al203.3CaSO4.31H20, forms early in the hydration of cement leading to setting and hardening of cement. It appears
104
that under certain circumstances, principally curing temperatures above about 70C, ettringite decomposes and subsequently reforms in the presence of water to cause expansion and cracking of concrete. The search for causes has become somewhat emotional but, to date, there is no evidence that this is the result of anything but incorrect concrete practices. In summary, it is believed that while cement quality may occasionally be the cause of concrete problems, this is much less likely than other causes. The most common cement problem is the shipping of an incorrect cement type or contamination during transportation.
Cement Plant
FAI LU RERAT
Wearout I period I
Thus, it was accepted that inappropriate time-scheduled maintenance may increase risk of failure by reintroducing infant mortality to stable systems. This led to the idea of condition-based maintenance which depends upon recognition that many failures give advance warning through such parameters as temperature or vibration. The optimum maintenance program should, therefore, identify and adapt to the failure modes for each equipment item or component. Since the probability of failure is nearly constant over most (ca85%) of the equipment life, systems often appear to suffer random failure. Condition-based and operating-time-based maintenance can only be partly effective in controling these failures and unacceptable residual failure rate must be addressed by modifications to system operation and/or design. Maintenance represents, typically, 15% of total manufacturing cost and, since the 1970s, maintenance departments in the cement industry have shared the need for increasing cost discipline along with increased plant reliability. This has tended to result in smaller staffs with multiple craft skills and an increasing use of information and measurement technologies.
Operations Handbook
labor and overtime tracking safety record-keeping cost history for maintenance of each item of equipment
Essential to the CMMS is an examination of the logic underlying every preventive maintenance task to determine whether it is cost effective, whether the task is in fact effective to prevent the subject failure, and to ensure that all known failure modes are addressed with appropriate preventive practices. Implementation of a CMMS without this scrutiny will only perpetuate previous system defects with greater efficiency. It is desirable that the CMMS interfaces with the company's accounting systems to avoid the inefficiency and potential for inconsistency of multiple data entry.
be relaxed for oversized ancillary equipment. The same priorities should be employed by the maintenance department in allocation of the scarce resources at their disposal (manpower, expense, and down-time), and these may vary from time to time depending upon equipment and inventory situations. Reliability centered maintenance was developed in the aerospace industry and its effectiveness is attested by the very low failure rates achieved there. The concept also underlies the various quality management systems. (Ireson & Coombs; Handbook of Reliability Engineering and Management; McGraw Hill, 1988) It has generally been accepted in the cement industry that the greatest run-factor and maintenance cost efficiencies are achieved by running kilns to refractory failure subject, of course, to analysis of such on-line equipment monitoring as is possible (eg ID fan vibration). Planned maintenance is conventional on all major equipment except the kiln, and the specific job list should incorporate feedback on deficiencies from both production and quality departments. For kilns, planning does occur but most work is floated until inevitable failure. Shut-down tasks should include: Lubrication, filter cleaning, etc. s Attention to problem areas identified by operators, eg frequent alarms, vibration, etc. s Inspection and measurement of wear parts From previous inspections, or at fixed periods, part replacement or service should be performed
Maintenance labor costs are essentially fixed though overtime and outside contractors may constitute a significant variable element. Studies have shown as little as 28% of an 8-hour shift may be spent actually working on equipment and, more than with other jobs in a plant where a routine is established, productivity requires efficient scheduling and supervision. Equipment downtime is another major maintenance resource which should be used as efficiently as possible. For kilns, which are run to failure, the best use must be made of the unavoidable downtime and it is essential to plan in advance for all routine checks and for accumulated non-critical repairs. This may also involve contingency planning for possible tasks that are not confirmed until access is obtained. Once the required job list has been established and the critical path determined the work should be prioritized and scheduled to minimize the outage. Ancillary equipment which can be made available for scheduled preventive maintenance can accommodate other considerations; for example, maintenance on mills may be best performed during off-peak power tariff periods, and maintenance on crushers should be done when the quarry is not operating. A final consideration is that scheduled maintenance should keep as closely as possible to the schedule so that the production department may have confidence that the equipment will be available when they need it; loss of maintenance credibility is a major factor in operators' refusal to shut down equipment for routine attention and a downward spiral into breakdown maintenance. Inputs to the CMMS come from work orders which should: s identify the equipment by code which determines the cost center for charging maintenance costs and s describe the problem to be investigated or the repair to be made An important function of plant management is to determine operating strategy with respect to maintenance. Some flexibility is appropriate to respond to periods of high market demand when non-essential maintenance can be deferred and to periods of low market demand when the time may be available for deferred maintenance or equipment modification but, frequently, costs are under pressure. Plant management should define what is required from maintenance, their responsibility and authority, and the indices and metrics by which maintenance performance will be assessed.
z rn z
Prioritization of maintenance tasks is essential in order to allow timely inspection and repair of equipment before failure. Resources will never be considered adequate so that priorities must be set by plant manager, maintenance manager, and the heads of operating departments. The bases for priority are largely self evident and include; the kiln must be kept in operation to avoid production loss which cannot be made up; ancillary equipment can often be allowed to shut down without permanent loss; redundancy reduces priority equipment failures which impact the ability to load out product and satisfy the customer safety items must be either corrected or temporarily neutralized (eg by locking out or roping off) environmental or regulatory items may sometimes be deferred if a plan is communicated to, and accepted by, the regulator Priorities and their justifications should be communicated to shift supervision so that they are not lost in the heat of new crises. Ultimately, manning levels with or without outside resources must be sufficient to cope with all tasks deemed necessary. A list of deferrable, non-critical jobs is required to achieve efficient employment of a work force with a varying load of priority work.
;NY N31N IV
inadequate prioritization of scheduled work It is essential to identify and eliminate the root causes, otherwise the department will begin to organize for perennial breakdowns and will become entirely reactive. Inevitable results will be either excessive manning levels or excessive overtime, and the expansion of parts inventories to cover any eventuality. With carrying costs running 10-25% of inventory value depending upon taxation, depreciation, and the cost of capital, a significant expense can be generated which may well be ignored by plant management if it does not impact plant operating earnings. Risk analysis can be effective to establish optimum stock levels but an analysis of equipment failure and of procurement cycles must be integrated with the analysis. In this context it should be noted that considerable savings in inventory cost may be achieved if suppliers can agree to carry stock in the plant and charge only upon use. There may also be potential for selling back to a supplier stock which is no longer required.
with conventional cement plant maintenance. Traditionally, therefore, mobile equipment is under the control of the quarry department. With increasing equipment size, the number of units is frequently inadequate to allow efficient management of a dedicated shop and workforce so that many plants have resorted to outside shops for all but the routine checks and adjustments which are, in any case, conducted by operators. Frequently such shops belong to the equipment dealer who can offer efficiencies of familiarity, technical support, parts inventory, and the ability to apply varying levels of resource according to need. In-house mobile equipment facilities are now largely associated with remote locations or an abnormal fleet composition. However, whether or not maintenance is conducted in-house, a system must be established to track the maintenance and service history of each piece of mobile equipment and this may be either the quarry department's production record keeping program or the maintenance department's CMMS. Operating hours are logged at each routine inspection so that preventive maintenance activities can be initiated when appropriate. Mobile equipment is subject to an increasing risk of wear related to operating hours while the condition of the machinery can be closely monitored through analysis of the lubricating and hydraulic fluids which come into contact with the various components. Oil monitoring has become a sophisticated tool employing spectrodiernical analysis and particle concentration and size distribution which, both from absolute levels and from trending, allow identification of components subject to abnormal wear or nearing failure. The optimum frequency of oil analysis is related to service conditions and the characterization of failure types to be managed.
9 ACCOUNTING
9.1 Cost Accounting
A detailed cost accounting system is essential to identify and control manufacturing costs and compare actual cost with budget. This should encompass each item of moving equipment, purchased materials, and a breakdown of manufacturing over-head costs. Detailed costs should then be summarized monthly as unit costs (per tonne cement) under various heads, eg:
Cost by Process Area: Quarrying & crushing On-site raw materials to raw milling Raw milling & additives Blending Burning & cooling Finish milling & additives Bulk handling & loadout Manufacturing overhead Sub-total Bag premium Cost by Natural Expense: Operating salaries Operating & service labor Supplies, rentals, services Purchased raw materials Fuel Power Mobile equipment Maintenance, including labor Grinding media Refractories
This allows rapid scrutiny of cost trends and abnormalities which can be further investigated, if necessary, in the detailed accounts. There is sometimes conflict between the requirements of financial and of management accounting; the former preferring uniform monthly allocation of periodic refractory and major maintenance costs which can hide abnormal actual costs. If necessary, separate reports should be prepared which provide operators with more relevant information.
An operating budget should be prepared before the start of each financial year. Management determines projected production by month and by cement type; the limitation or "principal budget factor" is usually either market or equipment capacity. Each department estimates its own detailed costs, and the combined estimate should be reviewed in the light of previous actual costs and anticipated process and cost changes. Few manufacturers operate in sufficiently stable environments
Cement
OMMfiT1))
to avoid significant variances. Flexible budgeting allows provision for different levels of production. Comparison with budget will indicate any variances. Variances will be due either to change in quantity or change in cost/price and, whether favorable or adverse, must be identified, explained, and, if necessary, corrected. While budget reconciliation is a valuable operations management tool, an accurate budget is also essential for cash flow management. Fixed costs apply only to fixed labor, fixed power charge, lease/rental costs, depreciation and amortization, insurance, and taxes (other than income); virtually all other manufacturing costs are variable.
Variable costs are fuel, power, materials & services, and purchased raw
materials. Excluding depreciation, typically 65% of manufacturing costs are variable.
Overhead absorption is usually on a per tonne basis. Unless a flexible budget is employed, differences between projected and actual production must be reconciled retrospectively. Cost by type of cement should be estimated as accurately as possible in order to establish the net margin for each type. This requires identification and quantification of all associated cost elements including transition, product transfer, and storage costs. Special cements involving only finish milling and additives can be quite easily estimated, special clinkers which involve intangibles such as kiln refractory life and mechanical wear and tear can not. Costing of spare parts and purchased materials inventories can employ
various methods (FIFO, LIFO, standard cost, etc); with modern data processing, weighted average costing is probably the most satisfactory.
available for identifying duplications. While automatic order generation is desirable, any items costing more than some predetermined minimum amount (say, $1000) should be directed for management scrutiny prior to order.
Major spare parts should be kept under review. Slow wearing parts
such as mill diaphragms should be monitored and ordered an appropriate time ahead of anticipated failure. Parts subject to catastrophic failure such as large motors, kiln rollers, mill girth gears, and fan impeller shafts require judgement to balance their high inventory cost against potential production loss. If possible, equipment should be standardized so that one spare will cover multiple units or the part can be taken from a less critical unit (eg from finish mill to raw mill); a record should be kept of which gears have been turned and how many hours they have operated; and contingency plans should be made for emergency repair of large shafts and rollers. Reciprocal agreements between plants with similar equipment may also be possible but advanced planning is desirable.
An annual capital budget should be prepared covering items and projects exceeding a predetermined minimum amount (say, $50,000) which are capitalized and depreciated over a number of years rather than expansed to current operations. Major maintenance items such as mill linings which are replaced at periods greater than 1 year should be included. Normally, however, large projects (exceeding, say $500,000) must be considered on an individual basis.
Net present value reduces future costs and revenues to present value
using an assumed "discount rate".
Internal rate of return takes the sum of future benefits, after tax, and
- 0, 1 M n M [ j)_M
calculates the interest at which the total equals present project or acquisitiOn cost. Many companies have a minimum or "hurdle rate" for investments to be considered and this must, at least, exceed the after tax cost of money. The margin of excess should vary with the perceived risk involved. Pay-back is a crude measure comprising project cost divided by anticipated annual revenues or savings. This ignores discount rate but is useful for periods of less than 3 or 4 years.
and insolvent at the same time. Cash flow is variously defined and is determined from the "Statement of Cash Flow": Cash inflow = New borrowing + Funds from new equity + Operating income + Sale of assets Cash outflow = Dividends paid + Tax paid + Purchase of assets Net cash flow = Cash inflow - Cash outflow = Increase in working capital
10 MISCELLANEOUS
-
The following points are among many which require constant or periodic consideration: Stockpile inventories are often calculated from production and consumption figures. At least monthly, all piles should be surveyed and their capacity calculated from standard bulk density assumptions. For large, disorderly stockpiles, flyovers are particularly valuable; aerial digital imaging is now accurate to 1M horizontally and 15cm vertically. Silos for most materials are prone to developing dead material which is both wasteful of capacity and misleading as to inventory. Silos and hoppers may be designed for: Mass flow where material at every point is in motion during discharge. This requires smooth walls with steep surfaces and no abrupt transitions. Funnel flow where material from the top surface only discharges through a vertical channel above the outlet while surrounding material remains static. This occurs in squat silos and hoppers with insufficiently steep walls. It should be noted that the angle of repose of free material is usually inappropriate for surface slope design (Reed; WC; 2/1993, pg 21). Periodically, all silos should be examined and cleaned, either manually from inside or using the various remotely controlled mechanical techniques now available (ICR; 6/1991, pg 56). Certain clinker silo designs leave substantial quantities of dead material in the bottom; such silos, upon construction, are better pre-filled and run down to refusal with limestone rather than high value clinker. Weigh-feeders should be calibrated regularly and cross-checked against both inventories and indicated feed-rates at other stages of the process. Impact flow meters are particularly liable to instability and error if located in a moving air flow.
Exception reporting should be developed so that any abnormalities of
process rate, efficiency, down-time, spare-part consumption, and cost are quickly identified.
Cement Plant
Operations
Handbook
121
Equipment numbering should allow easy identification and location. A typical system comprises area (eg kiln = 4), equipment type (eg bucket elevator = BE) and a serial number; thus 4BE2. All maintenance and cost records should refer to equipment number. Safety lockout of electrical equipment for maintenance is vital and standard procedures must be documented and enforced (Morgan; P&Q; 7/1992, pg 32). Every employee must have a personal lock which is applied to the breaker of individual units (eg pumps), or to the lock-box of multiple systems (eg mills), to prevent inadvertent or premature re-energizing. Operators' data logging is important even when data is recorded automatically as it helps to bring the operator's attention to process changes. Either the operator or the shift supervisor should also keep a narrative log of equipment shut-downs (with detailed explanation), alarm faults, requirements for maintenance or process investigation, process inventory situation, personnel accidents, and any other matters that require recording or reporting. The shift supervisor is, in effect, the plant manager outside of office hours and should be given the training and the authority to match the responsibilities. Among the more important duties are to supervise operations, to decide on shut-down of major equipment, to set priorities for shift maintenance, and to decide when call-out is required, either for management or for additional maintenance or service people. Process alarms should be carefully designed. Certain events are critical and require immediate attention, but if aural or visual alarms are excessive, important alarms may be missed. Critical alarms should be designed so that cancellation is impossible before the problem is corrected. Frequently, equipment failure causes interlock shut down of other upstream equipment; it is useful to devise alarms that direct attention immediately to the unit which caused the failure. Interlocks are not uncommonly jumpered (either by hard wiring or by programming) to allow maintenance or to cope with some temporary abnormality. Such jumpering must be recorded and active jumpers frequently reviewed to prevent inadvertent permanence.
11 PLANT REPORTING
11.1 List of Reports
Minimal reports to plant management for monitoring operations include: Daily: Production Report with production, downtime, and inventory by area (milling, burning, etc). Production Report. Quarry Report with production figures and details of drilling, blasting, loading and hauling. Process Summary with operating data and efficiencies for each area. Downtime Report with total downtime and detailed breakdown for each area. Shipping Report with total cement shipped broken down by type, by bulk vs sack, by conveyance (road, rail, etc.), and by destination market. Paper sack inventory reconciliation and sack breakage. Quality Summary with raw material, process and product analyses, and statistical variation. Mobile Equipment Report with availability, fuel consumption, and details of major downtime. Manufacturing Cost Summary with total unit cost and detailed breakdown by area, by individual equipment and by grouping (power, fuel, labor, etc). Inventory Schedule valuing product, process, fuel and warehouse inventories. Order Status itemizing deliveries which have been rescheduled or are overdue. Manpower report comparing actual numbers with establishment by department, and including overtime, hiring and terminations. Safety Report detailing all accidents and total days worked (to month-end) since last lost-time accident. Projects Report covering planning, ordering, progress, and budget of capital projects managed by plant staff.
Monthly:
Equivalent tonnes facilitate the compilation of materials and process cost contributions to the unit (tonne) of cement produced. An equivalent tonne of cement is usually assumed to be 950kg clinker and 50kg gypsum. Then, for example, if there is 80% limestone in the raw mix and a kiln feed: clinker factor of 1.6: 1 Eqt Limestone = 950 x 1.6 x 0.8 = 1216kg
Also, if raw mill operation costs $3.00/ tonne of kiln feed ground: Unit cost of raw milling = $3.00 x (950x1.6)/1000 = $4.56 The system requires adjustment when pozzolanic or masonry cements are produced which differ significantly from 95% clinker.
period. It is suggested that the main reason only need be recorded though the other tasks should be noted in order to explain anomalous totals. For example, if a kiln is shut down for four weeks due to high cement inventory, and major rebricking and maintenance are completed at the same time, the ascription should be voluntary downtime, but the subsequent low annual downtime must not be incorrectly construed. Any kiln voluntary downtime should be entirely due to excess clinker/cement inventory due, in turn, to depressed shipment. By design, the kiln should be the limit on plant production and it should never be slowed or shut down either for lack of kiln feed, or for lack of clinker and cement space. If this does occur, then either maintenance or the capacity of secondary equipment is inadequate and should be reviewed. Likewise, there should be substantial voluntary downtime or spare capacity on all equipment other than the kiln.
Plant Operation: Primary crusher Crusher-dryer Raw mill Kiln Finish mill
Raw Milling Production, t, Raw Cement Eq Operating time, % Downtime, h, Total Voluntary Mill, kWh/t % of rated kW Fineness, % -170# Meal moisture, % Material usage, kg/Eqt, Limestone Shale Silica Iron ore Ball usage, g/t
Burning & Cooling Production, t clinker t cement Eq average, daily t clinker Operating time, % Downtime, H, total voluntary Production rate, t/H Specific heat, kcal/kg clinker Coal, t/H Precalciner fuel, % Kiln exhaust gas, C Preheater exhaust gas, C Average freelime, % Feed end, % Lol SO3 Coal, % moisture ash kcal/kg (net, adb)
Finish Milling Production, t, total Type 1 .... Operating time, % Downtime, total voluntary Mill, % of rated kW kWh/t, total Type 1, etc Gypsum usage, % Grinding aid usage, g/t Ball usage, g/t
Quality Summary
Kiln feed, Type 1, C35 C3A SiR Clinker, Type 1, C35 Free CaO Cement, Type 1, C3S, % C3A, % 503, % Blaine, cm2/g +325#, % Initial set, hrs:mins Strength, 1 day, kg/cm' 3 day 7 day 28 day
1 320 Z 1 m -co 0 I 2
Ey
Raw Mill: Voluntary Sub-total Involuntary Feed system Conveying FK pumps/lines Separator Liners/screens Balling Mill mechanical Mill electrical Power off Inspection Misc maintenance Sub-total Total No raw materials Blend silos full
Kiln: Voluntary Inventory control Misc operating Feed system SP plug Kiln mechanical Kiln electrical Power off Control system Dust collector Refractories Cooler mech Cooler elec Fuel system ID fan Scheduled Sub-total Total
(Note kiln downtime is counted when feed is off)
Finish Mill: Voluntary No clinker No gypsum Cement silos full Sub-total Involuntary - Feed system Conveying FK pumps/lines Dust collector Separator Liners/screens Balling Mill mechanical Mill electrical Power off Inspection Misc maintenance Sub-total Total
Complete project management-from concept development to execution A wide range of state-of-the-art equipment and systems Service to our customers 24 hours a day, insuring smooth start-up, trouble-free maintenance and continued efficiency
:01
Put Fuller's technology, products and services to work for you. Call us today.
wi
IN FULLER.
A member of the FL Smidth-Fuller Engineering Group
Fuller Company
2040 Avenue C Bethlehem, PA 18017-2188 (610) 264-6011 FAX (610) 264-6170 www.fullerco.com
0
V3
Process formulae quoted here have, where possible, been selected or modified to be simple rather than precise. They are largely intended as "sanity checks" and, for the purposes of process engineering or design, more precise formulae are usually necessary. Note that metric tonnes are represented by "t" and short Tons (2000Lb) by "T".
1 POWER
1.1 Specific Power Consumption
Power consumption rose significantly with the introduction of dry process kilns and has continued to rise with conversion to coal, increased fineness of cement, and with the demands of environmental protection. Typically, power consumption is presently 11O-120kWh/t cement which may be broken down:
Quarrying & preblend Raw milling Blending Burning & cooling Finish milling Conveying, packing & loading Total 5% 24 6 22 38 5 6kWhit 28 7 25 44 6 116kWhit
For most large fan applications, frequency controlled variable speed drives offer significant power savings over damper or inlet vane controlled fixed speed ac drives albeit at higher capital cost. A technical and cost comparison is made by Bosche (WC; 6/1993, pg 2).
Energy efficiency should obviously be built in to original plant design. For existing plants, most remedies involve significant capital cost which must be justified by anticipated power savings. (Shenoy & Chacko; ICR; 10/1997, pg 60)
1.5 Co-Generation
The generation of electric power from kiln waste heat is not usually considered viable. Heat not required for drying raw materials is better conserved using 5-or 6-stage preheaters and high efficiency clinker coolers. Certain exceptions may apply, however, when power tariffs are particularly high or where co-generation is subsidized or mandatory (Onissi & Munakata; ZKG; 1/1993, pg
3;1
z
rpm2 rpmi
p2 1)1
TM) 2
kW2
rpm2\3 rpmi
Qi
( rpmi
kWi
For a fan with fixed speed and fixed resistance, the static pressure and kW increase in proportion to gas density. For a fan with fixed speed and fixed resistance, the effects of changing impeller diameter (D) are:
Q2 _ Q/
P2
( D2 ) 2D2
kW2
D2)5
i Di
pi
Di
klAri
Di
Thus, flow can be increased by increasing either speed or diameter, though increasing the diameter involves less power. Increasing either raises tip speed which develops more pressure and, often, more noise. Flow adjustment during operation is effected either by varying speed or by dampering. Speed adjustment is the more efficient but must be offset against the higher cost of variable speed motors. Inlet dampers are more efficient than outlet dampers. System static pressure varies with the square of the flow.
For fans operating at significant elevation, designed actual flow and fan drive power increase inversely with atmospheric pressure (Section B9.3) assuming no change in system back pressure. If system geometry stays constant, drive power increases inversely with the cube of atmospheric pressure. Leakage and friction losses should be minimized. The fan should be designed to convert gas velocity into static pressure by gradually increasing the cross sectional area of the path. Rapid slowing gives rise to "shock loss". Leakage from tip back to inlet is reduced by use of closed-sided impellers. Fan curves are for one speed, impeller diameter, and gas density only. Fans are usually operated to the left of the curve where pressure and kW do not change greatly with changes in flow. Curve shape depends upon impeller blade shape Radial blades show the least variation of static pressure with flow. Efficiency is 55-60%, increasing by about 5% if the impeller is closed-sided, and decreasing by 5% with dust entrainment. Backward curved blades provide the lowest static pressure for a given rotor size. They are, however, the most efficient, typically 75-80%, again slightly lower with dust entrainment. Backward curved fans are now standard for kiln id. Forward curved fans develop the maximum static pressure for a given rotor size
RADIAL VANES
Duct design should minimize pressure loss, and allowance should be made for the additional fan power required to move entrained dust (Wilson; RP; 5/1991, pg 67).
7>11 z
z
Cement Plant
Operations
Handbook 141
539.5
K values for conventional pitot tubes are effectively 1, and for S-type tubes 0.8-0.9 (see manufacturer's calibration). K values for orifice/venturi with constriction diameter d and pipe diameter D: d/D Orifice K Venturi K 0.731 0.7 0.6 0.683 0.5 0.658 1.012 0.993 0.4 0.646 0.640 0.984 0.3 (Note: venturi recovery cone of 15 assumed) Determination of air flow in ducts using a pitot tube involves measurement at several points across a section of the duct to allow for non-uniform flow and is described in detail for regulatory practice in the United States by EPA 40 CFR (7-1-91 Edition) Part 60, Appendix A. Measurement should be, if possible, at least 8 diameters down-stream and 2 diameters up-stream from any bend or irregularity. Measurements are traversed across two diameters at right angles, with 12 points per traverse for diameters greater than 0.6M, and 8 points for diameters 0.3-0.6M. Closer proximity to disturbances is compensated by additional traverse points. This procedure cannot be used when the flow is cyclonic and no point should be less than 2.5cm from the wall. The arithmetic mean of all velocity measurements is used to calculate flow. For approximate process as opposed to regulatory measurements, fewer points may be used.
Insertion distances should be: 12 points/traverse 0.021 D 0.644 D 0.067 0.750 0.118 0.823 0.177 0.882 0.250 0.933 0.356 0.979 8 points/traverse 0.032 D 0.677 0 0.105 0.806 0.194 0.895 0.323 0.968 6 points/traverse 0.044 D 0.704 D 0.147 0.853 0.296 0.956 4 points/traverse 0.067 D 0.750 D 0.250 0.933
PITOT TRAVERSE FOR GAS FLOW MEASUREMENT IN CIRCULAR AND RECTANGULAR DUCTS LARGER THAN 0.6M DIAMETER
Rectangular duct of section length L and width W yields an equivalent diameter: D = 2L.W/(L+W) and this diameter is used to determine the number of measurement points (12/traverse above 0.6M, 9 below). The section is then divided into that number of equal (square) areas and measurement made at the center of each area.
20 30 40 50 60 70 80
H2O content of saturated air: mg/L = 3.84x10-6xT4 + 2.93x10-5xT3 + 0.014xT2 + 0.29T + 4.98 where T = temperature, C
Cement Plant
Operations Handbook
145
hard facing, the impeller vanes can also be protected by guide vanes located near the hub which act to deflect particulates away from the main vanes (Krift; ZKG; 9/1994, pg E235). Fan housings together with cyclones and pneumatic conveying lines can be protected by lining either with fused cast basalt or, for high temperatures, materials made by fusing alumina, zirconia and silica (ICR; 7/1991, pg 52). Pneumatic conveying lines are prone to failure at elbows. It has long been standard to embed the elbow in a concrete-filled box. However, as the pipe wears, increased turbulence accelerates and extends the wear. T-bends have been found to simplify construction of pneumatic lines, reduce wear, and do not increase pressure drop but should not be used close to the pump where velocity is low.
3 CONVEYING
3.1 Comparative Power Consumption for Lift
kWh/t/100M 1.10 0.59 1.20 0.41
It must be emphasized, however, that the sizing of pneumatic conveying systems is complex, depending principally upon line length and configuration.
0Mt IMW
3.6 Airslide
Air slides are inclined at 4-10 with aeration compartment lengths of up to 5M. Air pressure is typically 40cm H20, air volume is 2M 3 /M2 /min and power consumption approximately 0.01kW/t/10M (Labahn). Dust collection is required to maintain upper compartment under negative pressure.
of fugitive dust, together with ability to climb steeper gradients and turn tighter corners. There are variations of design (Bahke; ZKG; 3/1992, pg 121: Hashimoto; WC Bulk Materials Handling Review; 1994, pg 40), some of which involve special belting with cables embedded in reinforced edges. Installation cost and power consumption are comparable to troughed belts but maintenance is less. Tube conveyors up to 650rnm diameter have been built carrying 2500t/H of material up to 400rrun in size.
711
400#
37g
Feed (F)
separation of fines. Bypass is the proportion of fines which are rejected; a perfect separator would have 0% bypass and an efficient separator should have Iess than 20%.
100
Operations Handbook
My M3 = Mi/MD Gv M3 = 1tr2LVL/100
G kg/min/M3 = MR(CL+100)/6000Gs where r = Mill inside radius, M L = Mill inside length, M VL = Charge loading, % MR = Mill production rate, kg/h R = Mill retention time, min MD = Mill discharge bulk density, kg /M3 CL = Mill circulating load, %
An analysis of 79 cement mills gave an optimum Grace factor of 263 and an acceptable range of 180-300. Void fill should be close to 100% (Welch; The Grace Factor A New Tool for Cement Process Control Engineers; ASTM; 1984). Retention time is determined using fluorescein (7g / 10t/h of combined mill feed and circulating load see Mardulier & Wightman; RP; 8/1971). Samples taken at discharge are shaken with water, filtered, and the solutions observed under UV light to determine peak concentration. If XRF is available which is equipped for Mg analysis, MgO can be used as tracer; addition should be such as to give sample concentrations of greater than 2Oppm, more if the milled material contains significant Zn.
The approximate power consumption of a mill can also be calculated from: kW = 7.5Gw4r) where Gw = ball charge, t D = mill internal diameter, M Traditionally, ball mills have been driven by single girth gear & pinion drives which are satisfactory up to about 5000kW. Larger mills are driven by girth gears with double pinions, central drives with epicyclic gears, or ring motors (ICR; 8/1994; pg 57).
30 40 50 60 70 80
Steel density is assumed 7.8g.cm 2. Bulk density of a mixed ball charge may be taken as 4550kg/M 3.
154
:slim -ma ]
A = +451.t residue after acid wash identified by microscopy as non- quartz acid insoluble Note Q, C & A are expressed as % of total raw mix sample
(DuToit; WC; 3/1997, pg 77)
Heat out: Exothermic cryst. of dehydrated clay 10 Exothermic formation of cement cmpds. 100 Cooling clinker 1400-20 360 Cooling CO2 900-20 120 Cooling and condensing steam 450-20 20 610 Net theoretical heat required to form 1Kg clinker, Q = 420kcal Heat of formation for specific mix designs can be calculated: (Q = 4.11A1 + 6.48Mg + 7.646Ca - 5.116Si - 0.59Fe)
Heat out: Theoretical heat of clinker burning Moisture evaporation (k9KF x % x 5.40) Kiln exhaust (_NM 3 x D x SH x C) Entrained dust (,_kg x 0.22 x C) Bypass gas (_NM 3 x D x SH x C) Bypass dust (_kg x SH x C) Cooler exhaust (NM 3 x 1,793 x 0.237 x C) Clinker (1kg x 0.20 x C) Radiation, preheater kiln cooler Total kcal/kg clinker out Note:
1 All weights and gas volumes are relative to 1kg clinker. 2 Temperatures (C) are relative to 20C (or ambient). 3 Kiln feed, kiln dust, and clinker specific heats can be calculated from standard values for components and temperatures, confirmed empirically, or typical values may be used (kcal/kg/C): 20 0.21 0.19 20-500 20-700 20-900 20-1100 20-1400 0.265 0.259 0.249 0.262 0.236 0.242 0.229 0.220
4 Fuel may be on "as received" or dried basis but weight and kcal/kg should be consistent, and "net" heat content should be used. 5 Clinker cooling air + primary air + false air + net combustion gas (ie 0.5 H2O) + calcination CO2 + moisture = kiln exhaust + cooler exhaust. 6 Evaporation is kiln feed x % moisture x latent heat; if coal is direct fired, its moisture must also be considered. 7 Exhaust gas density (D) and specific heat (SH) can be calculated from its components (Section 2.4) or approximately, from the typical gas composition given in Section 5.13.
,
20
20-200
20-400
20-600
20-800
20-1000
air
CO2
N2
8 Bypass gas and dust analyses are required for acceptable estimates of
gas density and specific heats. 9 Radiation losses can be determined by integrating measurements of surface temperature and emissivity for sub-areas. Alternatively, approximations such as those employed above are frequently used.
6 M/sec 10 9.5 13 24 18 28 25 20
1.2 2.09
Note that volumetric loading depends upon estimated retention time and upon the degree of precalcination of material at the kiln inlet. Alternatively, the specific kiln capacity is expressed in relation to the effective kiln volume as kg clinker produced/M 3 /hour with typical ranges: SP without riser firing 80- 90kg/M3/h ILC or SP with riser firing 100-130kg/M3 /h SLC 120-150kg/M3 /h Very large SLC kilns are now being designed with over 180kg/M 3 /h.
with indirect firing, and specific fuel consumption of 850kcal/kg clinker. Then coal consumption is 0.115kg/kg clinker and combustion gases produced are: C 76.0% H 5.0% S 2.0% 0.115 x 0.76 x 22.4/12 0.115 x 0.05 x 22.4/2 0.115 x 0.02 x 22.4/32 = 0.1631NM3 CO2 = 0.0644NM3 H2O = 0.0016NM3 S02
z
20
Total 02 required for combustion = 0.115 ((0.76x32/12)+(0.05x16/2)+(0.02x32/32) = 0.2814kg less 02 from coal: 0.115kg x 5% = 0.0054kg -> 0.2591kg x 22.4/32 = 0.1930NM 3 Then equivalent N2 from air = 0.1930x0.79/0.21 = 0.7260NM 3 CO2 from calcination of raw meal to yield 1kg clinker (assuming kiln feed LoI of 35%): [(1/0.65) - x 22.4/44 = 0.274NM3 CO2 H2O from kiln feed (assuming kiln feed:clinker factor of 1.7 and 0.5% H2O): 1kg x 1.7 x 0.005 x 22.4/18 = 0.0106NM 3 H2O Then exhaust gas with no excess air, is: CO2 0.4371NM3 = 35.3% 0.0750 H2O = 6.0 0.0016 = 0.129 (1290ppm) SO2 = 58.6 0.7260 N2 Total gas = 1.24 NM3/kg clinker or 1.52 kg/kg clinker Thus, total combustion air required (primary, secondary, tertiary, and false air) can be approximately estimated from: NM3/kg clinker = kcal/kg / 1000 x 1.122 where kcal/kg = net specific fuel consumption kg air/kg clinker = NM 3 /kg x 1.494 Total kiln exhaust gas can be approximately estimated from: NM3/kg clinker = [(kcal/kg / 1000 x 1.122) + 0.274][1 + (%02 x 4.76/10W
where %02 = free oxygen content This exhaust gas, with various levels of excess air, has the following density (wet basis), specific heat, and dew point:
Density g/L SH cal/g/C Dew point C
No excess air 1.487 2% 02 1.469 5% 02 1.441 10% 02 1.395 Note excess air assumed dry
38 36 33 26
Typical gas volumes for combustion products (with no excess air) vary with fuel: NM3 /kg clinker - Coal = kcal/kg / 1000 x 1.122 Oil = kcal/kg / 1000 x 1.185 Natural gas = kcal/kg / 1000 x 1.322 where kcal/kg = net specific fuel consumption In each case add 0.285NM3 /kg clinker for calcination and moisture
M Pt, C 776 801 772 1069 884 B Pt, C 15005 1413 1689 Eutectics have even lower melting points.
1450 891
851
896D
Volatilization in burning zone and condensation in preheater may be represented as shown (Norbom; Proc IEEE; Miami 1973). The external cycles through dust collector are not here considered; if dust is not wasted, then virtually all "e" is returned to the kiln.
PREHEATER
Conventional factors defining the circulation and typical values for SP kilns without bypass are:
SO3 K20 Na20 0.67 0.53 0.55 0.88 0.92 0.80 0.40 01/0.4 0.10 3.3/1.4 1.40 3.20 (=1 without bypass) CI 0.99 0.97 0.05 45.0
Primary volatility, El = b/d Secondary volatility, E2 = b/d V = e/d Valve, Circulation factor, K = b/a Residue, R = da
(Ritzmann; ZKG; 8/1971, page 338)
Note primary volatility applies to raw feed and secondary volatility to recirculated material. SO3 volatility increases with stoichiometric excess over alkalis. Excess sulfur will exacerbate the cycle concentration at the back of the kiln and can be effectively countered by adding alkali. Excess alkali will largely pass to clinker. Unless low alkali product is required and a bypass is operated, it is suggested that sulfur be added only to achieve stoichiometric balance in the material entering the kiln (bottom cyclone discharge), not in the overall system, (Mortensen et al; ZKG; 2/1998, pg 84).
6 FUELS
6.1 Typical Coal Analyses (% as recd/mineral-matter-free)
H2O
Lignite Sub-bituminous Bituminous, HVoI LVoI Anthracite 45 25 5 5 3 Vol 25 35 45 20 5 Fixed C 30 40 50 75 92
TI
m
3700 5400 7200 8500 7500
kcal/kg(Gross)
Ash is normally 5-15% of air-dried weight. Ash composition varies widely according to mineral contamination. Sulfur varies 0.2-7% and comprises mainly organic sulfur and pyrites with traces of sulfate. Pyritic sulfur can to some extent be removed by standard washing techniques. Proximate analysis: H2O + Vol + Fixed C* + Ash = 100% Ultimate analysis: C+H+N+S+ 0* + Ash = 100% (Vol = % volatiles) (* not normally determined; obtained by difference) Gross Heat, kcal/kg = 80.8C + 22.45S + 339.4H - 35.90 Net Heat, kcal/kg = 80.8C + 22.45S + 287(H-0/8) - 6W where W is H2O content, % Gross - Net = 51.5H* where H* is total % H2 including H2O (Gross Heat is the theoretical heat of combustion which assumes that water produced is condensed. In practice, water is usually released as vapor so that only Net Heat is recovered.)
Typical Petroleum Cokes H2O Vol Fixed C Green delayed 8% 11% 82% Fluid 0 5 86
SG Crude oils Gasolines Kerosines Gas oils Diesel oils Lubricating oils Heavy fuel oils 0.80-0.97 0.70-0.79 0.78-0.84 0.82-0.90 0.82-0.92 0.85-0.95 0.92-0.99
10,200
N2
CO2 kg/M 3 Net kcal/M 3
57.6H25
Most natural gas is free from sulfur but, if it occurs, is usually removed before delivery. Liquified gases yield the following: Methane 1 litre liquid Propane Butane
7 MATERIALS
7.1 Bulk Densities of Materials for Silo Storage (kg/M 3)
Aggregate,fine coarse 1500 1600 1500 Flyash Iron ore Limestone Raw meal Sand Shale/clay 550 2700 1400 1250 1600 1000
Cement Clinker 1360 Coal, bituminous, bulk 850 450 Coal, pulverized
8.2 Gypsum rock Iron ore 10.2 Limestone 11.5 Sandstone Silica sand 12.2 Blast furnace slag (Duda; Cement Data Book 1, 3rd Ed, pg 129/130)
1VI 31
8 CONVERSION TABLES
1 Length 1 mm 25.4 mm 1M 1.61 kM = 40 thou = 1 inch = 3.28 ft = 1 mile 2 Area 1 M2 1 acre 1 HA = 10.76 ft2 = 4840 yd2 = 2.47 acres
3 Volume 1 M3 28.32 L 3.785 L 1 ft3 4.546 L = 35.31 ft3 = 1 ft3 = 1 gal (US) = 7.48 gal (US) = 1 gal (Imp)
4 Density = 62.4 Lb/ft3 1 g/cm3 1 gal(US) water = 8.345 Lb = 22.4 L 1 gmole gas = 359 scf 1 Lbmole gas 1 cm3 mercury = 13.6 g
5 Pressure 1 Atmos = 1.034 kg/cm2 =14.7psi = 1.013 bar = 101.8 kPA = 76.0 cm Hg = 1033 cm H2O = 10.13 N/cm 2 = 101.4m H2O
6 Energy 1 kW = 860 kcal = 3413 BTU = 1.34 HP = 3.60 MJ = 1.80 BTU/Lb = 3600 BTU/5T
1 kcal/kg
1 kPA
8 Miscellaneous = 0.001 grain/scf 2.29 mg/NM 3 1 BBL(oil) = 42.0 gal(US) 1 BBL(cement) = 376 Lb
9 MISCELLANEOUS DATA
9.1 Geometrical & Trigonometrical Formulae:
3.14159 26535 89793 23846 26433 83279 50288
Cone, surface area = Itr(l+r) where I = slant height volume = 1/3irr 2 h Segment of circle Area = nr 2 (0/360) - Ch/2 where r = radius of circle (I)= angle subtended by chord at center C = length of chord h = perpendicular distance from chord to center
Trigonometrical Tables
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 sin 0.0000 0.0175 0.0349 0.0523 0.0698 0.0872 0.1045 0.1219 0.1392 0.1564 0.1736 0.1908 0.2079 0.2250 0.2419 0.2588 0.2756 0.2924 0.3090 0.3256 0.3420 0.3584 0.3746 0.3907 0.4067 0.4226 0.4384 0.4540 0.4695 0.4848 cos 1.0000 0.9998 0.9994 0.9986 0.9976 0.9962 0.9945 0.9925 0.9903 0.9877 0.9848 0.9816 0.9781 0.9744 0.9703 0.9659 0.9613 0.9563 0.9511 0.9455 0.9397 0.9336 0.9272 0.9205 0.9135 0.9063 0.8988 0.8910 0.8829 0.8746 tan 0.0000 0.0175 0.0349 0.0524 0.0699 0.0875 0.1051 0.1228 0_1405 0.1584 0.1763 0.1944 0.2126 0.2309 0.2493 0-2679 0.2867 0.3057 0.3249 0.3443 0.3640 0.3839 0.4040 0.4245 0.4452 0.4663 0.4877 0.5095 0.5317 0.5543 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 sin 0.5000 0.5150 0.5299 0.5446 0.5592 0.5736 0.5878 0.6018 0.6157 0.6293 0.6428 0.6561 0.6691 0.6820 0.6947 0.7071 0.7193 0.7314 0.7431 0.7547 0.7660 0.7771 0.7880 0.7986 0.8090 0.8192 0.8290 0.8387 0.8480 0.8572 cos 0.8660 0.8572 0.8480 0.8387 0.8290 0.8192 0.8090 tan 0.5774 0.6009 0.6249 0.6494 0.6745 0.7002 0.7265 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 sin 0.8660 0.8746 0.8829 0.8910 0.8988 0.9063 0.9135 0.9205 0.9272 0.9336 0.9397 0.9455 0.9511 0.9563 0.9613 0.9659 0.9703 0.9744 0.9781 0.9816 0.9848 0.9877 0.9903 0.9925 0.9945 0.9962 0.9976 0.9986 0.9994 0.9998 cos 0_5000 0.4848 0.4695 0.4540 0.4384 0.4226 0.4067 0.3907 0.3746 0.3584 0.3420 0.3256 0.3090 0.2924 0.2756 0.2588 02419 0.2250 0.2079 0.1908 tan 1.7321 1.8040 1.8807 1.9626 2.0503 2.1445 2.2460 2.3559 2.4751 2.6051 2.7475 2.9042 3.0777 3.2709 3.4874 3.7321 4.0108 4.3315 4.7046 5.1446
0.7986 0.7536 0.7880 0.7813 0.7771 0.8019 0 7660 0.7547 0.7431 0.7314 0.7193 0.7071 0.6947 0.6820 0.6691 0.6561 0.8391 0.8693 0.9004 0.9325 0.9657 1.0000 1.0355 1.0724 1.1106 1.1504
0.6428 1.1918 0.6293 1.2349 0.6157 1.2799 0.6018 1.3270 0.5878 1.3764 0.5736 1.4281 0.5592 1_4826 0.5446 1.5399 0.5299 1.6003 0.5150 1.6643
0.1736 5.6713 0.1564 6.3138 0.1392 7.1154 0-1219 8.1443 0.1045 9.514 0.0872 11.43 0.0698 14.30 0.0523 19.08 0.0349 28.64 0.0175 57.29
TT
Yu Ff E kv 52 1..)
mu
(m)
tau (t) upsilon (u) phi (f) chi (x) psi (y) omega (w)
1
15 25 16 6 17 37 9
i3 NV11 33I
b)
Metal hardness is usually determined by: A known load is applied to a surface using a spring-loaded hammer and a standard impression point; the depth of the impression is then measured. Range is approximately 60 for aluminum to 6-700 for hard alloy steels. Rockwell Two loads, one large and one small, are applied by compression to the surface; the difference in impression depths is measured. Values are approximately 1/10th those for Brinell. Both methods employ standard impression points which may be steel balls for soft materials but should be diamond for hard surfaces.
Brinell
(Lafarge)
1380 Mt/Y World total (ICR, Global Cement Report, 2nd Ed)
Mean - 242kg
During 1997, S0million t of new capacity was commissioned with an estimated 7-8million t closed down. Of the new capacity, 13% was from greenfield plants 47% was from new lines at existing plants 40% was from plant upgrades
23 tonnes (25T) Road bulk truck (U.S.) Rail bulk car, small (U.S.) 68 tonnes (75T) 91 tonnes (100T) large (U.S.)
n13N1/11 33
10 STATISTICS
Accuracy is the closeness between the observed value and the "true
value"; inaccuracy is usually due to systematic errors. Precision is the closeness between successive experimental values; this is the subject of statistics. Random errors give rise to Gaussian (normal) distribution. Standard deviation, a = "NI E(x-R) 2 /(n-1) where x is the experimental value r( is the arithmetic mean n is the number of determinations The number of "degrees of freedom", c, = n-1 since two values are required to define a line and only values in excess of two provide statistical information. Where n is finite, an approximation only of standard deviation is obtained, the "estimated standard deviation", s. As n increases, it becomes closer to the true standard deviation, a. For Gaussian distribution, 68.3% of values will lie between la and -1a, and 95.4% between 2a and -20. To convert s to % confidence limits, it is multiplied by a "t" factor. This gives the probability of a single determination falling within the indicated range.
GAUSSIAN DISTRIBUTION CURVE
-30-20 10 x 10 20 30
There are, of course, other theoretical distributions such as Poisson, binomial, and rectangular, but most process and analytical measurements can be handled by Gaussian statistics. In employing statistical analysis to quantify material and process variation, it is important that the input data should be grab samples or spot
measurements not composites or averaged signals. Input data period should also reflect actual cycle timing. For absolute measurement of variation, as opposed to trends, determination of method standard deviation (am) must be determined and subtracted from observed standard deviation (co) to give net standard deviation of the variable (r): ao2 am2 It follows that the precision of any test method must be significantly greater than that of the process variation being controlled.
F
test allows estimation of whether two values of s are from the same population. If calculated F = sA 2 /sB2 (where sA > sB) is equal to, or less than, the corresponding F value given in tables for a specified probability, then identity can be deduced. X2 test (chi squared)determines fit of data to theoretical distribution. For present purposes this will be confirmation of normal distribution about a mean. Data are collected into sub- groups and the percentage of observed values falling within each sub-group (0) is compared to theoretical probabilities (E), eg:
sub-group range
observed (0)
Then X2 = E (O-E) 2 /E is compared to tables to determine probable fit for the given degrees of freedom (for 6 groups, (0=6-1=5). Then X 2 should normally be less than 11 and greater than 1; greater than 11 indicates poorer than 5% probability of fit while less than 1 suggests too good a fit for natural data. Other numbers of sub-groups, other probability levels, and distribution other than Gaussian can be used. Comprehensive X2 tables are then required.
Linear regression is the best fit of xy data to a line: y = MX C where: m = s xy /sx2 c = y - sxy .x/s2 sxy = E(x-x)(y-y)/n Covariance is a measure of the inter-dependence of two variables; independence gives a correlation (r) of 0, and perfect dependence gives 1 or -1. r = sxy /sx.sy
"t" Tables
0 1 2 3 4 5 6 7 8 9 10 20 30 40 120 = 80% 3.078 1.886 1.638 1.533 1.476 1.440 1.415 1.397 1.383 1.372 1.325 1.310 1.303 1.289 1.282 90% 6.314 2.920 2.353 2.132 2.015 1.943 1.895 1.860 1.833 1.812 1.725 1.697 1.684 1.658 1.645 95% 12.706 4.303 3.182 2.776 2.571 2.447 2.365 2.306 2.262 2.228 2.086 2.042 2.021 1.980 1.960 98% 31.821 6.965 4.541 3.747 3.365 3.143 2.998 2.896 2.821 2.764 2.528 2.457 2.423 2358 2.326 99% 63.657 9.925 5.841 4.604 4.032 3.707 3.499 3.355 3.250 3.169 2.845 2.750 2.704 2.617 2.576 99.9% 636.62 31.598 12.941 8.610 6.859 5.959 5.405 5.041 4.781 4.587 3.850 3.646 3.551 3.373 3.291
Normal Curve - Percentage of Values Within (tx6) of Mean t % (two tails) % (one tail)
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.5 2.0 3.0 4.0 3.98 7.93 11.79 15.54 19.15 22.58 25.80 28.81 31.59 34.13 43.32 47.73 49.87 50.00 7.96 15.86 23.58 31.08 38.30 45.16 51.60 57.62 63.18 68.27 86.64 95.45 99.74 100.00
"F" Tables
(90% probability) 2 sA/sB 1 2 18.5 19.0 3 10.13 9.55 4 7.71 6.94 5.99 5.14 6 5.32 4.46 8 10 4.96 4.10 20 4.35 3.49 40 4.08 3.23 3.84 3.00 (98% probability) 2 1 SA/SB 2 98.5 99.0 3 34.1 30.8 4 18.0 21.2 6 13.74 10.92 8 11.26 8.65 10 10.04 7.56 20 8.10 5.85 40 7.31 5.18 6.63 4.61 3 19.2 9.28 6.59 4.76 4.07 3.71 3.10 2.84 2.60 4 19.2 9.12 6.39 4.53 3.84 3.48 2.87 2.61 2.37 6 19.3 8.94 6.16 4.28 3.58 3.22 2.60 2.34 2.10 8 19.4 8.85 6.04 4.15 3.44 3.07 2.45 2.18 1.94 10 19.4 8.79 5.96 4.06 3.35 2.98 2.35 2.08 1.83 24 19.5 8.64 5.77 3.84 3.12 2.74 2.08 1.79 1.52
"X2" Tables
tp
99% 0.02 0.11 0.30 0.55 0.87 1.24 1.65 2.09 2.56 8.26 22.16 37.48 70.06
95% 0.10 0.35 0.71 1.15 1.64 2.17 2.73 3.33 3.94 10.85 25.51 43.19 77.93
1 2 3 4 5 6 7 8 9 10 20 40 60 100
10% 2.71 4.61 6.25 7.78 9.24 10.64 12.02 13.36 14.68 15.99 28.41 51.81 74.40 118.50
5% 3.84 5.99 7.81 9.49 11.07 12.59 14.07 15.51 16.92 18.31 31.41 55.76 79.08 124.34
1% 6.63 9.21 11.34 13.28 15.09 16.81 18.48 20.09 21.67 23.21 37.57 63.69 88.38 135.81
The two standard explanations of cement setting and strength development are Le Chatelier's interlocking crystal theory and Michaelis's colloidal gel theory (Benstead; WC; 8/1991, pg 21: Chen & Odler; Cem and Conc Res; Vol 22; 1992, pg 1130). Aluminate and, particularly, silicate hydration reactions are extremely complex, and many undoubtedly contribute to the setting and strength gain of cement. The subject is described in some detail by Lea (The Chemistry of Cement and Concrete). C3A is the most soluble of the major cement compounds and appears to dominate early hydration. C3A reacts with sulfate in solution to give ettringite (C3A.3CaSO4.32H20), the interlocking crystal growth of which may contribute to setting. Ettringite later converts to monosuiphoaluminate (C3A.CaSO4.12H20). In the absence of sulfate, C3A hydrates to C4A.14H20 and C2A.8H20 which crystallize to cause flash set. This explains the importance of gypsum as set retarder. However, it has been observed that setting is largely independent of C3A concentration and it is now believed that both setting and strength development are largely caused by hydration of C3S to tobermorite (C3S2.4H20), a gel of variable composition. Hydration of cement typically involves combined water of about 22% relative to clinker weight. The relative contributions to strength development are shown in the following diagram:
Compressive strength development of individual Cement Compounds
Pozzolanic activity comes from the reaction of soluble Si02 from the pozzolan with CaO in solution. As free CaO will always be present in solutions in contact with hydrated cement, pozzolanic reactions provide "self-curing" of cracks in pozzolanic concrete.
saA
0
rol
rn
0
'73
Specific fuel (kcal/kg) 900-1000 1200-1500 900-1200 800- 900 800- 900 700- 850
Length:Diameter
32 - 38 32 - 38
2,000
2,000
11,000
It has also been observed that the grindability of clinker differs significantly with kiln type. Relative power consumptions for clinker types are:
Lepol kiln Cyclone preheater kiln Long wet kiln Long dry kiln 100 (softest)
107
To enhance evaporation of water by increasing surface area for heat exchange, to facilitate the handling of feed as it transitions from slurry through plastic material, and to detrain dust from kiln exhaust gas, chain systems are hung within the kiln shell (Figure 10). A typical system would comprise one to two diameters of bare shell followed by one diameter of curtain chains as a dust curtain (curtain chains are lengths of chain about 75% the diameter of the kiln and attached at one end only in successive circles around the circumference of the shell). Next come some five diameters of spiral curtain chain to break up and convey the drying (plastic) feed down the kiln (Figure 10). While curtain chains are
easier to manage, garland chains have been claimed to give better efficiency (garland chains are attached to the shell at both ends; the attachments should be 90 apart in a spiral down the axis of the kiln and the chain should hang slightly below the center line). Usually there is a second section of bare shell near the down-hill end of the chains to reduce circumpherential imbalance in gas temperature and material conveying. Duda recommends chain design parameters of 12% of daily clinker production for total chain loading, and surface area of 6-8.5M 2/M3 of chain section volume; de Beus (ICR; 12/1997, pg 41) suggests 15% and 610M2 /M3 respectively. Chain consumption is about 100g/ t clinker. The feed material leaves the chain section at 5-10% moisture and proceeds to the preheating, calcining, and burning zones of the kiln. Total material retention time in a long wet kiln is approximately 3 hours and gas is discharged at 150-200C. Dust loss with exhaust gas should ideally be 8-10% but is often much higher as kiln production is increased with resulting increase in gas velocity. Return of dust to the slurry system is inadvisable as it frequently causes agglomeration and sedimentation. Up to 5% relative to clinker weight can be returned by insufflation into the kiln flame; beyond this quantity, flame cooling becomes unacceptable. Alternatives are separate slurrying in a vortex mixer and parallel injection with the main feed, and return using scoops which inject the dust slightly downhill from the chain section or into a bare section of kiln near the downhill end of the chains. The basic causes of high dust loss, however, are gas velocity and chain design and condition.
0
I
z m
In both types of long kiln, chain fires can have catastrophic results and are at risk if the chain temperature exceeds about 1200C under oxidizing conditions; corrective action should be taken if chain temperature reaches 1000C.
Figure 10
GARLAND CHAIN
Gas duct Nodulizer 7-1: 1 Grate
CURTAIN CHAIN
Atz
I Kiln Cooler
Air injectors
Traditional shaft kilns were basically holes in the ground using mixtures of unground feed roughly mixed with solid fuel and burned in batches with natural draft. The lack of feed homogeneity together with non-uniform ventilation gave rise to widely varying temperature and oxidizing conditions so that quality was low and erratic. Considerable advances have been made and Rajbhandari (WC; 1/1995, pg 65) describes Spohn's black meal process as one of the most advanced shaft kiln technologies presently available (Figure 10). Practical unit capacities are 20-200t/d. Raw materials and solid fuel are ground together and nodulized (black meal process). Alternatively, but less effectively, raw mix and fuel can be ground separately and then blended and nodulized (white meal process). As with the Lepol kiln, stable nodules are important and usually require both a clay component and a solid fuel with less than 16% volatile content. The fuel may be coal, charcoal, coke, or petroleum coke. The kiln shaft is filled with the prepared mix and air is blown into the shaft at and near the base. The material is in turn heated, calcined, and burned at progressively higher temperatures as it moves down the shaft countercurrent to the combustion air. Near the base the clinker, with fuel already consumed, is rapidly cooled by the injected air and is discharged through a gate. Production is continuous with new feed added at the top to balance discharge. The process is, therefore, basically similar to that of rotary kilns. The principal difference is in uniformity; the rotary kiln ensures that the material is constantly agitated and that all material is subject to the same retention time and heat transfer. In the shaft kiln, however, there is a definite thermal gradient with the core material reaching a maximum temperature of ca 1450C, some 200 higher than material at the walls. Differential melting of the material tends to increase air flow velocity at the walls which reinforces heat loss though the walls to exacerbate the difference. More sophisticated shaft kilns can compensate for this with increased peripheral fuel concentration and reduced wall heat loss. Retention time above 1250C is typically 30 minutes. Increasing the air flow through the bed both increases production rate and clinker quality; the necessary air injection pressures (1000-2500mm WG) require an efficient air lock on the clinker discharge; either a triple gate or a controlled choke flow (seal leg). In practice, the seal leg is too dependent upon clinker bulk density and porosity to be effective and
Operations Handbook
the triple gate is preferred. In recent years there has also been a trend to increased diameter and reduced height to an aspect ratio of 2.5-3.0 but the diameter is usually limited to about 2.4M as increasing diameter makes uniform air distribution more difficult to achieve. Modern shaft kiln designs can be fully instrumented and PLC controlled.
$20 -0
0
rn
Large particle quartz in raw materials should be avoided if at all possible as grinding to reactive size, say <3011, by simple raw milling will result in excessive power consumption, in overgrinding the softer components, and in increasing cyclone preheater dust loss. Laboratory
grindability tests should be made for each raw material (Kannewurf, Grindability Standard; PCA Technical Report MP-8; 1956). Laboratory burnability tests can also be helpful (Maas & Kupper; ZKG; 4/1994, pg E27) as, to a lesser extent, can empirical formulae such as given in Section 5.3.
Operations Handbook
A heat balance should be constructed for the preheater, kiln, and cooler and the "heat out" values compared to standard values such as those given in B5.6 adjusted for the above factors. Significant abnormal heat loss is almost invariably associated with preheater exhaust and/or cooler exhaust. High preheater exhaust heat loss may be due to: False air from kiln seals and preheater access points s Ineffective cyclone preheater splash plates s Missing or distorted cyclone dip tubes Defective mix design, homogeneity, feed particle size, and feed rate consistency Hard burning to low free lime Defective coal particle size and feed rate consistency Poor kiln burner design and calciner retention It should be noted that, if heat is required in kiln exhaust for drying of raw materials, improving heat exchange in the preheater may be counterproductive; there is still, however, a benefit in optimizing burning conditions and mix design. If cooler exhaust or clinker heat loss is excessive, this will correlate to low secondary air temperature and poor cooler operation. Excessive kiln radiation may indicate poor kiln coating which may suggest an unstable flame. It has been shown (Roseman et al; ZKG; 12/1987, pg E280) that the effective heat loss varies with the source of loss due to the grade of heat involved. For a 4-stage cyclone preheater with a 2-stage grate cooler: I kcallkg heat loss from
Preheater exhaust must be compensated by 0.87kcal/kg in additional fuel energy Radiation from 4th (top) stage 3rd stage 2nd stage 1st (bottom) stage Kiln radiation loss Feed material reactions (eg overburning) Cooler radiation, vent, and clinker 0.22 0.44 0.76 1.18 1.18 1.18 1.46
321
This suggests that heat losses from the cooler are the most serious as heat lost from secondary or tertiary air must be compensated by addi-
Plant
Date
14.1. General
1 Property map 2 Land title 3 Surrounding land use 4 Plant lay-out 5 Plant age 6 Space & scope for expansion 7 Foundation conditions 8 Truck patterns 9 Process flow 10 Rainfall 11 Temperature high/low 12 Paving 13 Housekeeping 14 Insurance 15 Power supply 16 Power tariff on/off peak 17 Emergency generator
N3W SHSV 1 NV 1
18 Water source & consumption 19 Water treatment 20 Availability of waste materials 21 Cement market & seasonal variation
14.2. ADMINISTRATION
1 Manufacturing costs by area & natural expense 2 Manufacturing cost by type 3 Cement price by type 4 Cement price by market 5 Data processing 6 Employee numbers 7 Organization chart 8 Employee housing/transport 9 Safety records 10 Purchase procedure, local/import
14.3. QUARRY
1 Materials type 2 LS blasting, loading, hauling 3 LS crushing 4 LS reserves 5 Clay mining, transporting
14.4. DRYING
1 Raw materials moisture 2 Drying method
14.6. BLENDING
1 Silo type, dimensions, capacities 2 Blending ratio 3 kWhit 4 Discharge conveying
N3W 5 3 SS VIN
14.7. KILN
1 Kiln feed metering 2 Kiln type & dimensions 3 Kiln capacity, rated/actual 4 Run factor & downtime analysis 5 Voluntary downtime 6 Kiln seals 7 Kiln shell 8 Kiln tires 9 Kiln cover 10 Precalciner, tertiary 11 Preheater splash plates 12 Preheater build-up 13 10 fan build-up 14 Bypass 15 Cooler type & dimensions 16 Cooler condition 17 Cooler fans 18 Clinker breaker 19 kcal/kg clinker 20 Auxiliary drive
14.8. FUEL
1 Fuel type(s) 2 Start-up/back-up fuel 3 Firing system 4 Coal moisture 5 Coal mill type 6 Waste fuels
14.9. Clinker
1 Clinker conveying 2 Clinker types 3 Storage capacity 4 Outside storage 5 Silo tops, condition
N3IN 53SV IN
8 Cement conveying
14.11. CEMENT
1 Cement types 2 Silo capacity 3 Silo tops, condition
14.15. Maintenance
1 PM system 2 Shop facilities 3 Local machining 4 Local motor rewinding 5 Major maintenance planned/required
14.16. Process
1 Existing reports 2 Process engineering 3 Redundancy 4 Recorders, logging 5 Vibration monitors 6 Emission limits, dust 7 Emission limits, SO2 etc
N31/11 53V IN
14.18. Permits
1 Operating permits 2 Emission permits
t Storage capacity of limestone available to raw mill Storage capacity of clay available to raw mill Storage capacity of blending/kiln feed silos Storage capacity of clinker (excluding outside) Storage capacity of cement
Days usage
rn
m
ma
Financial Value is the historical cost of the assets less depreciation and
depletion and is the easiest valuation to obtain. As a going concern, asset value can be derived from the standard financial equation Assets = Liabilities + Equity or: CA + FA + OA = CL + LD + E where CA = current assets (cash equivalents & accounts receivable) FA = fixed assets (original cost less depreciation and depletion) OA = other assets (non producing assets & intangible assets) CL = current liabilities (payables and current portion of LD) LD = long term debt (maturity greater than one year) E = equity or: Valuation = FA + OA + (CA-CL) = LD + E where (CA-CL) = net working capital For most purposes, excluding financial reporting and return-on-investment calculations, the financial value of an asset is of little use. This is especially true during periods of rising prices when historical cost misrepresents the current cost of the assets being valued. Fair market value of a plant can be determined using several methods.
c
0
Briefly, the fair market value of an asset represents the amount a willing buyer would pay to a willing seller, neither under compulsion to participate dn the transaction, and where both have full knowledge of the facts and circumstances regarding the transaction. In practice, of course, this situation rarely exists. Discounted Cash Flow values an asset based upon the net present value of all the after tax cash flows the asset produces over a period of time, ideally twenty years or more, including the residual value. Valuation = NPV (R, (CFyri, CFyr2..(CFyrx + RV))) where R = discount rate,% The calculation of the cash flow resulting from the asset requires subjective assumptions. This is especially true with the cash flows (revenues) due to the uncertainty of projecting volumes and prices into the future. The effect of this uncertainty can be minimized by preparing values based on "most likely" and "worst case" projections. In valuing an asset for investment purposes, the discount rate will equal the investor's after-tax cost of capital.
Replacement Value is the current cost to replace the asset adjusted for depreciation to recognize wear-and-tear and obsolescence. Depreciation is based on replacement cost less salvage value divided by useful life.
In practice it is desirable to employ all or several of the above valuation methods and apply a sanity test upon completion. Several projects and acquisitions in recent years doubtless resulted from sophisticated economic analysis but clearly failed the sanity test. As a going concern, plant valuation can be based upon: 1) Replacement value applying depreciation for wear-and-tear and obsolescence to current new replacement cost. For asset value pur-
poses (as opposed to tax write-off which may be accelerated), depreciation schedules are typically straight-line over: Mobile equipment 3-10 years Machinery & equipment 10-20 years Buildings 20-40 years 2) Net earnings (after interest, amortization, depreciation, and tax) adjusted for the average market P/E ratio of comparable companies and the average acquisition premium of comparable deals. Valuation = Net earnings x P/E + premium P/E ratios and acquisition premiums vary considerably with time and market area. Typical ratio would be 12 and premium 33%. Simple share-price/earnings-per-share ratios are considered crude parameters due to distortions of earnings by items such as depreciation. An alternative measure has been proposed (Economist; 18 Feb 1995, pg 80) which takes the ratio of enterprise value (debt + market value of equity - estimated market value of non-relevant assets) to cash flow (annual cash flow from core business before interest, taxes, and depreciation). This EV /CF ratio, though more complex to determine, does give a more realistic measure than P /E.
m
rn
1-1
If the subject plant is part of a multi-plant company, determination or confirmation of appropriate profit center figures is necessary as cost allocation or transfer pricing practices may cause distortion. A particular problem is encountered when attempting to value operations or perform cost/benefit analyses in countries subject to subsidies and price controls, or in large companies where transfer pricing may be suspect. "Shadow pricing" is an attempt to impute free market prices. Foreign exchange risk is nearly always significant in developing countries as local currency finance, if available, is prohibitively expensive. Devaluation will, almost inevitably, be followed by direct or indirect control of product prices while debt service, spare parts and imported supplies, and, usually, fuel is payable in hard currency. Hedging is, of course, possible but expensive and a number of companies in South East Asia have suffered from this exposure during 1997. Inherent in the calculation of future earnings is a market projection over
the life of the investment. Such long term projections are never easy but are particularly fraught when considering developing markets. A review of long term forecasting methods is given by Joos (ICR; 8/1992, pg 29).
There is, of course, considerable variation due to specification, location, degree of local fabrication, labor and materials costs, and cost of capital. Conroy (Proc IEEE, 5/1998, pg 211) discusses the cost of new and upgraded production lines in the U.S. There is also considerable variation in schedule for construction. Conventionally, process and preliminary engineering design, and permitting are completed before procurement and contract bidding begin; however, where expedition is essential the risks of overlapping engineering with construction may be accepted. Site preparation is also frequently performed by plant forces or separately contracted. Then assuming permitting does not impact the critical path, a typical schedule would be: Cumulative 0 months Start detailed engineering 2 Issue tender documents 5 Contracts for equipment and construction 7 Ground breaking 12 Complete civil work 12 Major equipment delivered Complete mechanical erection 18 20 Complete electrical and instrumentation 21 Begin commissioning Commercial operations 24
REFERENCES
JOURNALS CITED
BSH = Bulk Solids Handling ICR = International Cement Review P&Q = Pit & Quarry RP = Rock Products WC = World Cement ZKG = Zement-KaLk-Gips
REVIEW PAPERS
General
Cement Plant Modernization; Dumas; ICR; 5/1991, pg 54. New Developments by Krupp Polysius; Ritzmann; ZKG; 4/1993, pg 173. Plant Engineering; Guilmin; ZKG; 5/1994, pg 131. Planned Maintenance; Patzke & Krause; ZKG; 5/1994, pg 135. Power Reduction; Fujimoto; WC; 7/1993, pg 25. Use of Heavy-Duty Fans; ICR; 1/1995; pg 53. Future Technology; Part 1 WC; 9/1994, pg 66: Part 2 -WC; 10/1994, pg 49: Part 3 WC; 11/1994, pg 47: Part 4 -WC; 12/1994, pg 58. Materials, Milling & Conveying Crushing; McCarter; ICR; 2/1996, pg 51. Ensuring Reliable Flow from Hoppers, Silos, Bins & Bunkers; Reed; WC; 2/1993, pg 21 Bed Blending Homogenization; Petersen; WC Bulk Materials Handling Review; 1994, pg 30. Grinding Energy & Plant Throughput Improvement; Marchal; WC; 9/1994, pg 84. Comminution Technology and Energy Management; Ellerbrock & Mathiak; ZKG; 11/1994, pg E296. High Pressure Grinding Rolls; Ellerbrock; ZKG; 4/1994, pg E104. Loesche Mill for Cement Grinding; Brundiek; ZKG; 4/1994, pg 179. Separators in Grinding Circuits; Onuma & Ito; ZKG; 9/1994, pg 535.
High Efficiency Separators; Detwiler; ZKG; 7/1995, pg 384 & 9/1995, pg 486. Pneumatic Conveying; Hilgraf; ZKG; 3/1993, pg E84 & 5/1993, pg E145.
Dust Collection
Electrostatic Precipitators in Cement Plants; Werner; ICR; 8/1991, pg 61. ESPs for the Future; ICR; 5/1995, pg 29. Troubleshooting a Dust Collector; Davis; WC; 10/1994, pg 14. Use of Acoustic Horns in Cement Production; Andersch & Kramer; WC; 1/1995, pg 26.
Cement Terminals
Terminal Solutions; Haschke; ICR; 5/1995, pg 36.
BOOKS
Chapman; Recommended Procedures for Mechanical Analysis of Rotary Kilns; Fuller Company (1985). Croom; Filter Dust Collectors Design and Application; McGraw-Hill, Inc. (1994). Duda; Cement Data Book, Volume 1 Process Engineering, 3rd Ed (1985). Volume 2 Electrical Engineering, Automation, Storage, Transportation, Dispatch (1984). Volume 3 Raw Materials (1988). Bauverlag GmbH, D-65173 Wiesbaden, Germany.
Cement Plant
Fax: 49 6123 700 122 Johnson; ISO 9000 Meeting the New International Standards, (1993). McGraw-Hill, New York. Kosmatka & Panarese; Design & Control of Concrete Mixes, 13th Ed (1988); Portland Cement Association. Labahn/Kohlhaas; Cement Engineers' Handbook, 4th English Ed (1983). Bauverlag GmbH. Lea; The Chemistry of Cement and Concrete, 3rd Ed (1970). Edward Arnold, London. Taylor; Cement Chemistry (1990). Academic Press, London. Peray; The Rotary Cement Kiln, 2nd Ed (1986). Chemical Publishing Co, New York. Peray; Cement Manufacturer's Handbook (1979). Chemical Publishing Co, New York. Perry & Green; Chemical Engineers' Handbook, 6th Ed (1984). McGraw-Hill, New York. Blasters' Handbook. (formerly published by El_ duPont de Nemours & Co) International Society of Explosives Engineers, 29100 Aurora Rd, Cleveland, OH 44139, USA. Fax: (216) 349 3788 Buffalo Forge Fan Engineering Handbook, 8th Ed (1983). Available from Buffalo Forge sales offices. Buffalo (USA) Office Fax: (716) 847 5208 Cembureau; World Cement Directory. World Statistical Review Cement Standards of the World. (see below)
Three I Publications Ltd, 1-5-16 Uchikanda Chiyoda-ku, Tokyo 101, Japan. Fax: 81 3 3291 3764 International Cement Review; The Global Cement Review. Tradeship Publications, Old Kings Head Court, 15 High Street, Dorking, Surrey RH4 lAR, U.K. Fax: 44 1306 740660 North American Combustion Handbook, 3rd Ed, Vol 1 (1986) & Vol 2 (1997). North American Manufacturing Co, Handbook Dept, 4455 E 71st Street, Cleveland, OH 44105, USA. Fax: (216) 271 7852
ADDRESSES OF ORGANIZATIONS
American Petroleum Institute, 211 N Evray, Suite 1700, Dallas, TX 75201, USA. Fax: (214) 748 7962 Order desk 1220L Street NW, Washington, DC 20005. Tel: (202) 682 8375 Fax: (202) 962 4776 American Society for Testing & Materials (ASTM), 1918 Race Street, Philadelphia, PA 19103-1187, USA. Tel: (215) 299 5400 Fax: (215) 977 9679 American Society for Quality Control; (ISO) Registrar Accreditation Board, 611 East Wisconsin Avenue, Milwaukee, WI 53201, USA. Tel: (414) 272 8575 Fax: (414) 272 1734 British Standards Institution (BSI), 389 Chiswick High Road, London W4 4AL, England. Tel: 44 181 996 7111 Fax: 44 181 996 7048 Cimeurope S.a.r.l., (Publications subsidiary of Cembureau The European Cement Association), 55 rue d'Arlon, B-1040 Brussels, Belgium. Tel: 32 2 234 1011 Fax: 32 2 230 4720 European Committee for Standardization (CEN),
5 33N311 3d311
36 rue de Stassart, B-1050 Brussels, Belgium. Tel: 32 2 519 6811 Fax: 32 2 519 6819 Institute of Electrical and Electronics Engineers (IEEE), Cement Industry Committee, 445 Hoes Lane, PO Box 1331, Piscataway, NJ 08855, USA. Tel: (732) 981 006. Fax: (732) 562 5445 Intercem (Organizers of international conferences on trading, transportation and handling of cement) Congrex House, 169 Stafford Road, Wallington, Surrey SM6 9BT, England. Tel: 44 181 669 5222 Fax: 44 181 669 9926 International Kiln Association, 15 Woodgate Road, Mile End, Coleford, Gloucestershire GL16 7QF, England. Tel: 44 1594 810426 Fax: 44 1594 810426 International Organization for Standardization (ISO), Case Postale 56, CH-1121 Geneva, Switzerland. Society for Maintenance & Reliability Professionals, 401 N Michigan Ave, Chicago, IL 60611, USA Tel: (312) 321 5190. Fax: (312) 527 6658 US Portland Cement Association, 5420 Old Orchard Road, Skokie, IL 60077, USA. Tel: (708) 966 6200. Fax: (708) 966 9781
Operations Handbook
C)
1 -4
co
l'a 0 CD ia to ln
01
to 4. F1 M to b CO
a, 17 -4 Cc,
CO
0:.
ro b -.. 0 (0 2.11
CI
Ps)
CO
It
'-' co s
C.
= ...,.
co
EiCD
0)
_.
.31
21 2 ,.-' 13I
,.
0,
a)
63 A
gt) cb
*
Z.
" C N co Oa
c' (0 ul
4, co CO
n1 CD
el
CD
Ls? 0
ip IM -4
a .....
cn P ........, - in 0 --... CO
co
0)
co
-.. 0 .o.
Ca
co co
CO
0 0c.r,
(D
0 CD t a) ''' 4, co
CD
1. -
--s Ca
0 13 al
CD
co rn fo * -4 !'" MA O T. 0 N A cn eo 7;, ,, -4 co c* b _
Z co
IV
Q DC
J., u.,
Di 0 -4
1 f' MI 4.
-4
th
11 r.,
-I
(
NA0
(0
1,3
s.^ -4
Pa _.
ro Si A - 0 PO 8 0 "I . Co 0 at 13 A
c* Z "
:--r
A c., 3 3
In
EL I
...
,-1.
O
(0
-0 '4 N '
c'" 9 a .C.
-4
a. .
131
..4
TC- 0 a rn -4
(0
Cr' >
-4 > 4,
0 ks
a 4' Cn
4D
N ci
ar
C 6
0:1
0-
at
Ca ,
6
ba
,A
(13 '='
0, 03 0 o. :1". 0. ' C.
,.,
ul N o,
i' 1Xl
cn
0 00
vDC
o
E
0, . >..... 0 4 j j
. N
h.,
_..
N m cp 61) (.0
CO''
co
0,
th
-Ti
Ca
Fo
CD M
00
in
Co n
Co
A ca
r.a.
-4
0. o
OP CO "I ch (0
TS 1:1
3 0
-4 0") c,, co
b cn
W Cn 0 0 5 co
co
F.
0 s-l th g
A
x , .
T.2 1
N 33I 0, K.:m 4, !-,, pc co 10 s,, co 2 Z , b 0 X a. zo .., C0 ;.,, 7:. -L , eri 0 a 2 w t (0 = crs E.,
co --
INDEX
Page Abrasion resistance 145 Acoustic horns 50, 76 Addresses, organizations 214 Air distribution 142 Alarms 122 Alkali, S and Cl limits 39 Alkali-silica reactivity 104 Alkali:sulfur ratio 4 Audit, process 193 Atmospheric pressure 174 Bag filters 75 Bibliography 212 Blended cements 95 Blending 17 Blending ratio 11 Brinell hardness 176 Brown clinker 31 Budget, capital 107 Budget, operating 115 Bulk densities 171 Burnability factor 157 Cash flow 131 Cement big-bags 73 Cement bulk loading 73 Cement capacities, new 177 Cement comparative testing 89 Cement compounds & ratios 157 Cement dispatch 71 Cement hydration 183 Cement intergrinds 96 Cement milling 61 Cement packing 72 Cement particle size 67 Cement production, world 176 Cement storage 70 Cement strength 92 Cement types & specs 92 Chemical analysis 85 Page Chi squared tables 182 Chi squared test 180 Circulating load 151 Clinker cooling 48 Clinker grindability 62 Clinker heat of formation 158 Clinker storage 61 Coal data 167 Coal firing 44 46 Coal firing systems Concrete problems 100 Conversion tables 172 Conveying, belt 148 Conveying, bucket elevator 148 Conveying, drag chain 148 147 Conveying, FK pump Conveying, pneumatic capsule 149 149 Conveying, sandwich belt 148 Conveying, screw 148 Conveying, tube belt Conveying power 147 Cooler efficiency 162 Corrosion, kiln 58 Cost accounting 115 181 Covariance Co-generation 137 Crushing 6 Cyclones 75 Definitions Preface Delayed ettringite formation 104 Depreciation 207 Detached plumes 78 Dew point 145 Dioxins 81 126 Downtime reporting Drying, raw material 8 Drying, raw mill 13 Dust collection 75
Dust suppression 78 Earth, elemental abundance 175 Electricity, 3-phase 137 Electrostatic precipitator 75 Emergency power 58 Equipment numbering 122 F tables 182 F test 180 False air 144 False set 67,104 Fan build-up 142 Fan laws 139 Flux, kiln feed 3 Fly ash 97 Free lime 25 40 Fuels Fuels data 167 Gas conditioning 76 Gas data 142 168 Gas fuels data Geometrical formulae 173 Greek alphabet 174 Grindabilities 171 Grinding aids 67 Hardness 176 Horomill 65 Insufflation 38 Interlocks, jumpered 122 Investment justification 117 ISO 9000 98 27 Kerogens Kiln, Lepol 188 Kiln, long dry 187 Kiln, vertical shaft 188 Kiln, long wet 186 Kiln alignment 57 Kiln burning 23 Kiln burning temperature 158 38 Kiln bypass
157 Kiln coating tendency 29 Kiln control 162 Kiln drive power 55 Kiln drives 162 Kiln exhaust gas 18 Kiln feed 19 Kiln feed:clinker ratio Kiln gas analysis 30 160 Kiln gas velocities 159 Kiln heat balance 161 Kiln loading, volume 51 Kiln mechanical 35 Kiln refractories 161 Kiln retention time 26 Kiln rings & build-ups 53 Kiln seals Kiln shell. 53 161 Kiln slope 194 Kiln specific fuel 29 Kiln speed 32 Kiln start-up & shut-down Kiln types 185 175 Laboratory reagents 171 Linear expansion coefts. 181 Linear regression 107 Maintenance 97 Masonry cement 82 Metals, toxic 90 Microscopy 153 Mill, Grace factor 63 Mill air flow 155 Mill ball size 154 Mill ball weights 152 Mill charge loading 155 Mill charge wear 152 Mill critical speed Mill diaphragm 64,155 154 Mill power 153 Mill retention time
Mill specific power 65,196 Mobile equip. maintenance 113 Normal curve area 179 NOx 79 Oil fuels data 168 Oilwell cement 92 Opacity 78 Operators' log 122 Orifices 143 Packset 104 Periodic table End Petroleum coke 42, 167 pH & normality 174 Physical testing 89 Pitots 143 Plant assessment data list 197 Plant construction cost 208 Plant reporting 125 Plant valuation 205 Pollution control 75 Power conservation 136 Power consumption 136 Power generation 137 Power tariffs 127 Preblending 9 Preheater cleaning 40 Preheater configurations 22 Process flow Profit & loss 119 Project cost estimating 118 Quality control 85 Quarry operations 4 Raw material sampling 88 Raw materials 3 Raw milling 13 References 211 Reliability centered main. 109
Report, daily production 128 Report, downtime summary 132 Report, process summary 129 Reserves 5 Review papers 211 Rockwell hardness 176 Roll press 64 Roller mill, cement 64 Roller mill, raw 13 Safety lock-out 122 Sea water composition 175 151 Separator efficiency Separator types 68 Shift supervisor 122 Ship capacities 175 Sieves 151 121 Silo maintenance 97 Slag SO2 81 145 Spray cooling 145 Stack draft 179 Standard deviation 179 Statistics Stockpile inventories 121 Sulfur cycle 30 "t" tables 181 Trigonometric tables 173 143 Venturis Volatile cycles 38,164 116 Warehouse Waste fuels 43 Waste raw materials 4 149 Water pump Water treatment 105 121 Weigh Feeders X-ray diffraction 87 85 X-ray fluorescence
*
Operations Handbook
Tradeship Publications Lt Old Kings Head Con 15 High Strel Dorking, Surre RH4 1A United Kingdoi Tel: +44 (0) 1306 740363/3E Fax: +44 (0) 1306 7406E E-mail: tradeshipdial.pipex.cm Website: www.cemnet.co .0