Acknowledgements I would like to thank the legions of grey men wi th whom I have worked and from whom I have learned over the years. Specifically I would like to express my appreciation to James Post, Hung Chen, Herman Tseng, Andrew Jackura, Arthur Chin-Fatt, Michael McCabe and Peter Bergmann who have made significant contributions as the earlier editions developed. I would also like to thank Philip Kerton for providing important revisions for this seventh edition. I am indebted to International Cement Review for their continued interest and support in publishing successive editions of the Handbook.
The author, however, retains proprietary rights over all expressions of ignorance or opinion. Copyright e Philip A Alsop, 2019
Published by Tradeship Publications Ltd Uld King's Head Court, 15 High Street, Dorklng, Surrey, RH41AR, UK. Tel: +44 (0)1306 740363 Fax: +44 (0)1306 740660
Email: [email protected] Website: www.C€mnet.com Managing Editor: Thomas Armstrong Senior Editor: Philip Kerton Editorial: Muriel Bal Design: Duncan Norton Printed in the United Kingdom by Warners (Midlands) Pic, UK. AU 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 prior written consent from Tradeship Publications Ltd. This publication is intended sol~y 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 forthe application of the information.
Preface 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. ]t 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 pre heater kiln technology, which comprises the majority of world production capacity and virtually all kilns installed since 1970
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
an outline of plant assessment and plant valuation
reference to ASTM and EN specifications for cements and for standard methods
a collection of process formulae
a selection of reference data and notes
references to review articles
The Cement Plant Operations Handbook - Seventh Edition
Contents Section A - Process summaries 1. Introduction
1. The basics of cement manufacture - 2. History of cement manufacture - 3. Portland cement in
2. Raw materials
1. Raw materials - 2. Raw mix - 3. Reserves - 4. Crushing - 5. Drying - 6. Pre-blending - 7. Storage and handlins
3. Raw milling and blending
1. Raw milling - 2. Blending - 3. Kiln feed
4. Flames and fuels
1. Chemistry of combustion - 2. Fuels - 3. Physics of combustion - 4. Burner design - 5. Cement kiln burners - 6. Heat transfer - 7. Pollutant formation - 8. Modelling - 9. Fuel storage and firing systems in practice - 10. Insufflation - 11. Alternative and waste fuels
5. Burning and cooling
1. Chemical reactions - 2. Process variants - 3. Kiln burning - 4. Kiln (ontrol- 5. Volatiles in the kiln - 6. Kiln bypass- 7. Kiln start-up and shutdown -8. Kiln refractories- 9. Clinker cooling -10. Kiln mechanical -11. Emergency power -12. Plant control systems
6. Cement milling
1. Storage of clinker and other components - 2. Cement milling - 3. Separators (classifiers) - 4. Ball mill circuit control - 5. Cement storage - 6. Cement dispatch - 7. Distribution - 8. Quality assurance and customer service
7. Quality control
1. Sampling - 2. Chemical analysis - 3. Particle size analysis - 4. Thermal analysis - 5. Microscopy - 6. Virtual Cement and Concrete Testing laboratory (VeCTl) - 7. Calorimetry - 8. Bumability - 9. Grindability - 10. Physical testing-ll. Process control analysis -12. Chromate passivation - 13. Cementquality-14. Setting time - 15. ASTM cement types and specifications - 16. European EN 197 cement specification 17. Composite cements (intergrinds and blends) - 18. Supersulphated cement - 19. Calcium aluminate cement (CAC) - 20. Shrinkage-compensating cements (SCC) - 21. ISO 9001:2000 Quality management system - 22. Concrete problems
1. Maintenance benefits and costs - 2. Failure modes - J. Computerised Maintenllnce Management Systems (CMMS) - 4. Reliability-Centred Maintenance (RCM) - 5. Maintenance cost management - 6. Maintenance organisat ion - 7. Role, planning and control - B. Mobile equipment maintenance - 9. People and indicators
Preface 9. Environment and pollution control
1. Dust coUection ~ 2. Pollution control ~ 3. ISO 14000 - 4. Sustainable development and climate change
10. Hydration of Portland cement
1. Initial stage - 2. Induction or dormant stage ~ 3. Acceleration stage ~ 4. Deceleration stage
11. Plant reporting
1. Definitions - 2. List of reports ~ 3. Inventories and feeders ~ 4. Downtime reporting - 5. Miscellaneous reporting ~ 6. Typical daily production report - 7. Typical process summary data ~ 8. Typical equipment downtime report ~ 9. Plant manning
1. Cost or management at::wunting - 2. Investment justification - 3. Capacity increase by process change ~ 4. Project cost estimation ~ 5. Financial statements
13. Technical and process audits
1. Historical performance ~ 2. Kiln specific fuel consumption ~ 3. Cement mill specific power consumption - 4. Other systems ~ 5. De-bottlenecking ~ 6. Project audit ~ 7. Risk assessment
14. Plant assessment list
1. General ~ 2. Administration and commercial ~ 3. Communication with stakeholders - 4. Quarry - 5. Drying ~ 6. Raw milling ~ 7. Blending ~ 8. Kiln ~ 9. Fuel -10. Clinker - 11. Finish miU- 12. Cement13. Quality control - 14. Packing and dist ribution -15. Emission abatement (dust, NO" S02' etc) - 16. Maintenance -17. Process - 18. Materials analysis - 19. Plant capacity summary - 20. Storage capacity
15. Cement plant construction and valuation
1. New p lant construction - 2. Project management - 3. Cement plant investment costs - 4. Project phases - 5. Plant valuation
Section B - Process calculations and miscellaneous data B1. Power
1. Specific power consumption - 2. Power conservation -3. Three-phase power-4. Motor power output - 5. Peak power tariffs - 6. Power generati150-400
Raw materials or day-to-day fluctuations in material properties, whilst in-plant equipment compensates for shortterm fluctuations that occur wit h an hour-to-hour frequency, or even more rapidly. Pre-blend effectiveness (the ratio of estimated standard deviations [sl for feed and product) is approximately related to the number of layers [N] by: Blending ratio = SIIt. To ensure good blending action, it is preferable that material chemistry varies both above and below its
target set-point on several occasions during the silo residence time. With the availability of real-time, online analysis of mill feed or product, chemistry can be maintained within narrow limits and modern plant designs frequently dispense with kiln feed blending or cut down the size of prehomogenising stockpiles. However, such procedures do not work well when faced with certain origins and frequencies of variability in a major raw miK component, or with corrective
components that are either difficult to extract from storage and accurately dose, or that are difficult to combine within the kiln (Harrisson, 2013). Ratherthan investing in equipment to correct the chemistry of the feed on its way to the kiln, an improved quarry management system to deliver a more consistent supply of limestone provide'S a superior solution.
Both raw meal chemistry and feed rate to the kiln must be consistent to avoid kiln instability and minimise fuel consumption. Short-term feed fluctuations (eg hunting of feed er control) should be monitored, as well as average feed rate. Suspension preheater510se a fraction of kiln feed entrained in exhaust gas. As this fine fraction is usually of a typical composition, kiln feed analysis must be biased to yield thedesired clinker composition. The dust loss, some 5·12 percent of kiln feed, is not usually collected until after it has passed through a raw mill or dryer, so that dust catch is not the same quantity or composition as preheater dust loss. Even if the dust collector catch is returned directly to the kiln, it must still be compensated. Likewise, care must be exercised to minimise chemical disturbance due to dust return, particularly when the raw mill is down and the dust collector catch changes from mill discharge instead of kiln discharge. The least negative option is feeding to the blending silo or to a separate storage tank for subsequent controlled return. If the kiln exhaust passes directly and continuously to dust collection, then the dust may be returned directly to the preheater - sometimes, if low in quantity, re-entrainment can be minimised by insufflation at the hood injection at the upper end of the kiln. Either way, the return rate should be controlled. Kiln feed is monitored by chemical analysis on four- or eight·hourly grab sample'S - not cumulative samples - to determine standard deviation results (see Section 2.6). Analysis is conventional for major oxides with variatIon monitored statIstIcally In terms of CJS or LSF. Kilns, particularly the larger ones, are intolerant of variations in feed chemistry. The general rules for chemistry control at the raw milling stage are similar, no matter how many mix components are available, and are basically simple: • • •
If the chemistry is not at the desired level: change it! If a parameter is below the desired level: increase it to above that level. (Known as 'crossing the target line' - as seen on a control chart - and failure to do this usually re'Sults in 'off·target' batches of raw mea1.) When making a change, pay due regard to previous results and weighfeeder settings. Where possible, use continuous samplers: spot samples can often be misleading and cause erroneous corrections. Residues are very important and their testing and control demands as much care as bulk chemistry checks.
Variations in fuel ash should not be overlooked and ash should be considered as part of the raw mix argillaceous component If tile Quantity fed into the kiln varies, the chemical variability of the dinkerwiH also vary. Plants with high fuel consumption and/or high fuel ash contents should minimise variations by blending the fuel(s>.
Raw milling and blending It should be remembered that standard deviation is not a perfect measure of variation as, simply applied, it does not distinguish between a steady trend and constant fluctuation. Kiln operators need
to respect maximum tolerable li mits upon variability - see Table 3.3. Table 3 3 K,ln feed target v3r1ablht'e5 for vanous
Is an ollernatlW! to tSF or C;' (potenriallficalcium silicore content - see $€clion BS.1), which are
The importance of controlling variations is indicated in Table 3.4. Table 34 likely mfluence of chemIcal vanatlOns upon pro<('ss and quality parameters
Kiln feed Is normally conveyed by bucket elevator to the top of the preheater. Pneumatic conveying wastes energy, and feed de-aeration Is desirable before injection as the entraining air adds to the kiln ID fan load and may reduce capacity. Although about 1.55t raw materials are required to produce It of clinker, kiln feed-to-clinker ratio is typically 1.65-1.75 as weighed, due to the loss of dust with exhaust gas, later collected and returned. The ratio should be periodically reconciled with clinker and cement inventories and with measured dust loss in the preheater exhaust. •
Kiln feed = clinker + Lol + bypass dust + downcomerdust - coal ash where both bypass dust and down comer dust are converted to a loss-free basis.
Flames and fuels
4. Flames and fuels The kiln burning zone is at the heart of the manufacturing process and t he supply of heat energy results from combustion of fuels. Combustio n science involves t hermodyna mics, fluid mechan ics, chemi ca l kinetics and t ran sport processes, and did not emerge until over 100 yea rs ago. In recent yea rs new insights have arisen through adva nces in computer capability, experimental tec hniqu es and asym ptotic methods of appli ed mathematics. Kiln operators can control the following factors when firing fuel in a kiln or calciner.
fuel type fuel handling heat transfer burner momentum excess air emissions.
Other factors that are typically outside immediate operator control can have a dramatic effect on combustion: • • • •
secondary air' momentum tertiary air momentum kiln aerodynamics calciner aerodynamics.
Unfortunately each factor is integral to one system and it is difficult to review one independently of another. This section attempts to address the broad subject of combustion and how it applies to making cement.
Chemistry of combustion
Combustion is a specific group of chemical reactions where a fuel and oxygen react at a sufficientlyhigh temperature to evolve heat and combustion products. Combustion can vary in rate from a very slow decay to an instantaneous explosion: a kiln requires steady heat release at a certain rate. Oxidat ion of industrial hydrocarbon fuels can reasonably be simplified to four basic reactions:
The complete oxidation of carbon C + 0,--+ CO 2 + 394kJ/mOle (94kcaVmole)
The difference in the physical states of the wat er produced in the reaction causes the complication of gross ca lorific value (GCV, or higher heating value) and net calorific value (NeV, or lower heating value) for fuels. Users face the conundrum of paying for gross heat but utilising only net heat.
' Primary a/renters the kiln via the burner pipe, with the fuel. Secondary air is drawn in around the flame. In a cement kiln, this is hot air drown from the clinker cooler via the kiln hood. Tertiary air in a cement kiln system is also hat air from the clinker cooler, in this case supplied to the precalciner by a duct.
The Cement Plant Operations Handbook - Seventh Edition
Table 4.1 Gross and net calorific: value for selected fossil fuels
Oil fuels Oil fuels are produced by refining crude oil or manufactured from coal. They are classified as distillate fuels (such as kerosene ilnd diesel oil), or ilS rcsiduill fuels. The latter come in a range of viscosities and are classified differently in different countries. Typical characteristics are given in Section B6.2. Residual fuels must be heated to become pumpable and reduce viscosity to enable atomisation. The heavier the fuel, the more heat is needed. Owing to their tendency to solidify when cold, handling systems must avoid 'dead legs' as much as possible. Lighter 'white' oils yield a better profit than black
Flames and fuels
fuel oils and refineries increasingly manufacture more of these. The black oils are heavier with different characteristics to the white products, having increasing quantities of asphaltenes. These cracked fuels vary in character, depending on the source of crude and the refining process, and are not necessarily compatible with each other. Under some circumstances, fuel oils form 'gels' in tanks and fuel handling systems with disastrous results. Proposed fuels should always be tested for compatibility with the existing fuel before purchase. Oil atomisation is important because initial drop size determines the size of the cenosphere that is formed and hence the droplets' burn time. Oxygen diffusion is dependent on surface area, but oxygen demand depends on particle mass. Since surface area depends on diameterl and mass on diameterl, larger drops take longer to burn. Most atomisers produce a range of drop diameters, varying from a few micrometres to around 100lOOO[Jm, or even more. A 100[Jm particle burns in about half a second in an industrial flame while a SOO[Jm particle takes about five times as long and a 1000[Jm particle 10 times as long. Since residence time in a flame is typically one second or less, it follows that drops larger than about 200[Jm will not fully burn out before the end of the flame and wi([ either drop on the product, unburnt, or end up in the dust collector. For optimum performance, a range of drop sizes is ideal, fine drops to facilitate ignition and flame establishment and larger drops to maintain a controlled burning rate. However, for the above reasons, the maximum drop size must be below 100-2S0IJm, depending on the exact application. Equally as important as drop size is spray angle. Essentially, mostsprays are conical, with two common types: hollow cone and solid. Hollow cone atomisers are generally preferred, as air mixes most effecti vely with fuel. The small numberof drops in the hollow cone core allows an internal recirculation zone to be established, which assists in maintaining a stable flame front. Most burners vary flow rate to operate over a range of heat liberation rates: atomiser performance must be satisfactory over the entire operating range, since cement plants do not always operate at full load. The drop size of many types of atomiser increases rapidly as the fuel flow-rate is turned down and this can present special problems for plant operation: turndown performance is an important selection criterion for atomisers.
Coals Great care has to be taken handling and burning coal owing to the risk of spontaneous ignition, fire and explosion. As a result, the design and operation of coal-firing systems requires greater specialist knowledge than gas and fuel oil systems. The characteristics of coals vary even more widely than other fuels - from anthracite wi th a high CV, and very low volatile and moisture content, to lignites with moisture and volatile contents of up to cu percent. I ypical properties of some commonly-traded coals are given in Section 66.1. The characterist ics of the coal and its ash have a dramatic effect on plant performance and on maintenance requirements. Relevant properties include:
Volatile content - The higher the volatile content, the more rapidly the coal ignites and burns. Highly-volatile coals (above 3S per cent) tend to present significantly higher explosion risks than those with a volatility below 25 per cent. Coals with volatile contents above 45 per cent require special precautions. Swelling properties - Once the volatiles have been driven off, a coke particle is left. If this is larger than the original particle, it has a more open pore structure and will burn more rapidly than if it shrinks. Moisture content - Coals have two types of moisture: surface moisture and inherent moisture. Generally the higher the inherent water, the greater the coal reactivity and the higher the consequential fire and e~plosion risk. For pu lverised coa l firing, surface moisture has to be removed during grinding. Removal of the inherent water should be minimised, otherwise moisture from t he atmosphere recombines w ith the coal and causes spontaneous heating which can result in fire or explosion. Ash content - Cement manufacture demands consistent levels of ash quantity and composition, and coals have to be selected accordingly. Hardness and abrasion indices - Coal hardness affects coal mill capacity: the harder the coal, the less can be ground and/or the coarser the resulting pulverised coal. The abrasion index is mainly dependent on ash characteristics. Very abrasive coals with high silica ashes cause high wear rates in grinding elements.
The Cement Plant Operations Handbook - Seventh Edition
Petroleum coke Petroleum coke (petcoke), a by-product from oil refining, is the solid residue remaining after extraction
of all valuable liquid and gaseous components from crude oil. The volatile content range is typically 5-15 per cent, depending on the coking process. The main difficulty in burning petcoke is its low reactivity due to this low volatile content. Petcoke has certain advantages, particularly its very high 01 of around 8000kcal/kg (gross), but increasing price in certain market conditions sometimes reduces its attraction. The usually-high sulphur content (3-6 per cent) can also limit usage (Batra et ai, 2005): its use has been banned in some Indian states (Jethmalani, 2017). It exists in four basic forms, delayed ('green'), calcined, fluid coke and f1exicoke. Delayed coke is by far the most common aocl 'green delayed coke' typically has 8-16 per cent volatiles, though higher temperature processing can yield less than one per cent volatiles. It may be 'sponge' or 'shot' and is used - with care - as up to 100 percent of total kiln fueL Petcoke burning usually involves finer grinding than coal and higher excess oxygen to compensate for its low reactivity and achieve complete combustion, which can result in some de-rating of the kiln (Roy, 2001). High·momentum burner design and attention to calciner design can also assist.
Physics of combustion
No chemical reactions can take place until the oxygen in the airis brought into contact with the fuel.. Therefore, aU combustion processes take place in the following stages: mixing . ignition . chemical reaction . dispersal of products Combustion rate depends on the slowest of these stages. In most industrial combustion systems, mixing is slow whilst the other steps are fa st. The rate and completeness of the combustion process is therefore controlled by the rate and completeness of fuel/air mixing. Insufficient fueVair mixing produces unburned CO in flue gases, wasting fu el energy potentiaL Forgood combustion, adequate air must be supplied for complete mixing and the burner must be designed to mix the fuel and air streams effectively and efficiently. Hence, the saying of combustion engineers: ' If it's mixed, it's burnt.'
Fuel/air mixing For most burners, fueVair mixing occurs as a result of jet entrainment. Figure 4.1 shows a free jet issuing from . noule in an ambipnt mPCiium. Friction occurs between the jet boundary and its surroundings, locally accelerating the surrounding fluid to the jet velocity. The accelerated air is then pulled into the jet, expanding it. This process is momentum controlled and continues until the jet velocity is the same as that of its surroundings. The greater the jet momentum, the more surrounding fluid is entrained. A free jet that is able to expand unimpeded can entrain as much of its surrounding medium as it needs to satisfy its entrainment capacity.
fntralnment of~eCit, with monitoring for 0" NO, and perhaps SO,. In addition, there are a number of useful indicators potentially available for estimating conditions inside t he kiln, such as kiln drive power ('kiln amps'), burning zone temperature and exit gas NO, analysis. Aspects of control are discussed below.
Kiln operation is evaluated by: • • • • • • • • •
production rate (tph clinker) operating hours - feed -on (h) involuntary downt ime (h) total fuel rate (tph) proportion of fuel to precakinerjriser (%) specific heat consumption (kcal/kg) secondary air temperature (' C) kiln feed-end temperature (' C) pre heater exhaust gas temperature ( . C)
IDfan suction (mm HP) kiln feed -end 0, (%) downcomer 0, (%) kiln feed-end material
- Nap(%) kiln drive power (kW - often monitored as kiln amps).
There are, of course, numerous other parameters that are logged, both to observe trends that may indicate problems and to provide necessary mean data for process analyses such as heat balances. These other factors include: • • •
primary air flow and burner tip velocity (m/s) specific kiln volume loading(%) speci fic heat loading of burning zone (kcal/h/m' of effective burning zone crosssection area)
cooler air (Nml/h/m' grate area) cooler air (Nm'/kg clinker) cooler (tpd clinker/m' grate area) temperature, pressure and oxygen profile of pre heater (' C, Pa and per cent by volume, respectively).
Burning and cooling
To deduce the reasons for poor performance, the most important aid is comparison data collected during periods when the plant runs well. Such periods are opportunities for staff to fully identify the conditions needed for optimum operation. Modern kiln operation and maintenance should aim for above go per cent run factor (7884h/year), below three per cent lost time per month between planned outages (22h), and cont inuous operations exceeding 100 days (Buzzi, 2003). The best performers surpass 95 per cent run factor year upon year. Excessive heat consumption should be investigated immediately and may indicate incorrect feedrate measurement or feed chemistry, fuel or burner abnormality, insufficient or excess oxygen, air inleakage at kiln seals or preheater ports, low secondary air temperature, or distortion or collapse of preheater splash-plates. Variable or hard-burning mixes should be avoided, as higher temperatures and longer retention times involved in controlling free lime resu lt in large alite and, worse, large belite crystals. Clinker free lime should be as high as needed to avoid the thermal inefficiency of hard burning but safely below the limit for onset of mortar expansion - typic.aUy 0.5-2 per cent. Having reached the target, free lime should, if possible, be maintained within a range of abou t 0.5 per cent. Variation of kiln feed rate or composition makes this control more difficult. Over-burning - a common solution to variable feed chemistry or operator circumspection - wastes fuel, stresses refractories, increases power required for cement milling and reduces cement strength. Sasaki and Ueda (1989) found a 14kcaVkg heat penalty for each per cent reduction in free lime, though ot her references vary. It is important to note that the relationship between free lime and burning zone temperature is far from linear (see Figure 5.4). so that it is easier to overburn clinker - for example, when raw meal c.omposition varies - than to use a softer burning regime tha t risks a potentially large Increase In free lime. Frgure 5 4
Relatronshlp b('twe('t1 clink('J IrE.'€' lime and bllrnmg lon
'. (diff;0.02 per cent and CrPl >0.01 per cent. The use of pure sand and limestone and the low clinker liquid phase require clinker burning at 1600 · c, usually with a mineraliser (uF,). Specific fuel consumption is considerably greater than for grey cement. Quenching clinker with water and trying to recover the waste heat pose challenges to the process and equipment (Pekin, 2000, Clark, 2001 and Schulz, 2003). Various pastel-coloured cements can be made by mixing or, preferably, intergrinding pigments with white cement (Bensted, 1993). Darker colours such as red and brown may be produced from grey cement. Pigments are usually inorganic and should be durable to light and weathering, non-soluble and not reactive with cement. Common pigments, added as 5-10 per cent by weight of cement, are the oxides of iron, manganese, chromium and cobalt, as well as carbon black. Grey cements of equal quality may produce concrete of darker or lighter colour, depending upon the levels of iron and other elements that are present. Masonry cement is used for mortar in masonry construction where good workability and rapid hardening are required. Various inter-ground ingredients may be used, but commonly 20-50 per cent limestone is incorporated together with an air-entraining agent. Masonry cement is typically ground to 5000-6000cm 2Jg and ASTM specification C91 defines three grades: N, S, and M with increasing strength requirements. ASTM C1328 specifies plastic (stucco) cement for plastering applications. There is very little difference between C1328 and C91. ASTM C1329 specifies mortar cement, which is basically masonry cement with a lower air content limit and with additional bond strength requirement.
7.16 European EN 197 cement specification Table 7.10 EN 197 - 1:2011 Cement - Part I: COmpOSition, speCifications and conformity criteria for common cements
Portland slag cement
Portland silica-fume cement
Portland pozzolana cement
Portland burnt shale cement
Portland fly-ash cement
Portland limestone cement
Portland composite cement
Blast fumace cement
18-30 31-50 + 31-50
Siliceom ffy-ash is 10 per cent CoO. Limesrone L has s r.an playa more promin('nt rote in governing appearance. The non-clinker materials are not themselves cementitious but latently hydrau(ic, ie, reacting with lime released during cement hydration to form compounds with cementitious properties. They are widely available and exploited in certain areas but not worldwide. Snellings et al (2012) provide an extensive review of the types of naturally-occurring mineral that exhibit pozzolanic properties, and their evaluation is discussed by Morrical et al (2011). Although generally slower to develop strength than straight Portland cements, the composite cements reduce concrete porosity and, in the presence of moisture, promote self-healing of cracks. The result is cement with reduced heat of hydration, reduced alkali-aggregate reactivity and increased sulphate resistance. Some minerals are ca Icined to produce "artificial pozzola ns", and Norwegian studies have demonstrated scope for alternative chemica! treatment of olivine and serpentine, avoiding the use of fuel. (Justnes, 2009). Although the term 'pozzolan' is often used generically in standards for any active mineral addition, pozzolans should ideally be distinguished from hydraulic materials such as ground granulated slag. These latter react directly with water to form cementitious compounds, while pozzolanic materials, in the presence of moisture, react with calcium hydroxide (either a reaction product or added directly) to form compounds possessing cementing properties. 81astfumace slag, for example, is hydraulic if activated by sodium hydroxide, silicate or calcium sulphate. ASTM C595 requires t hat the manufacturer shall, on demand, state the source and amount of pozzolan added. Under both ASTM and EN standards, a merchant is free to buy pure Portland cement and blend in appropriate minerals to produce composite cements t hat comply with one or more of the specifications. Although C595 refers only to blended cements, a common practice is to add the same pozzolanic materials at the concrete batching plant. Pozzolanic materials thus employed are covered by ASTM C618 (pfa and natural pozzolan) and C989 (blastfurnace slag). EN 15167 covers the use of ground slag as a concrete addition, and EN 13263 covers silica fume. (Note that unlike Europe, in the USA the term 'slag cement' usua lly means 'ground granulated blastfumace slag', not a composite cement!)
Quality control Natul1tl pozzoians are usually soft to grind. ASTM C618 requires that SiCI +AJ.P l+FePl >70 per cent aocl addition rate is 15-40 per cent. A significant safety margin has to be adopted during manufacture to allow for the variability of properties for naturally mined materials. Rapid determination of a hydraulic activity index allows the margin to be reduced. Velasquez et al (2018) indicate how XRD techniques can be modified to determine the amorphous phase in pumiceous materials to provide a more precise
indicator than measurement of acid-insoluble residue or quartz content. Fly ash is classified by ASTM e618 as 'Class F' with sial +A1 20 1+FePI >70 per cent and 'C' if >50 per cent. Some Class C ashes contain sufficient CaO to be appreciably cementitious. Addition rate 1515-40 per cent of total, and a 45~m residue of below 20 per cent is preferred if the ash is to be blended rather than interground. Ash is increasingly liable to contain unburned carbon due to the useoflow-NO. burners in power boiters. For intergrinding or mixing in concrete, ASTM C618 requires that Lol (effectively due to carbon) should be less than six per cent. High carbon levels are undesirable for aesthetic reasons, as the surface colour of concrete is affected, and also fo r technical reasons, as fine carbon interferes with the action of admixtures. Sometimes size separation of ash can improve quality where carbon is mainly in unburned material in the coarsest particles. Other technologies include froth flotation and drying, magnetic separation and electrostatic separation. High·carbon ash may be suitable as a kiln raw material. Several beneficiation processes are in commercial operation. In addition to carbon content, there is also increasing concern over the presence of ammonia from 'slip' during injection to abate NO, emissions. Above SOp pm, such ash may be difficult to use in cement and concrete due to the objectionable odour that arises during concrete production and use, and which can per:;i5t formanyweelu in pooilyv~nlildll'd indoor locations. Limits for safe ammonia conc€fltratlons In ambient air, which vary from one country to another, rarely appear to be exceeded beyond a few hours. But statutory limits tend to drop, and ammonia removal options are on offer. Fly ash is sometimes classed as an 'artificial pozzolan', and its properties are addressed in EN 450, whilst its addition atthe concrete mixer is covered in the UK wi thin 'BS 8500: Concrete- Complementary British Standard to BS EN 206-1'. In contrast to much of the rest of Europe, UK users have preferred classified fly ash, with the coarser particles removed, as this provides greater consistency between deliveries, improved water reduction properties and consequently bener strength performance in mortar and concrete. Typical compositions, with wide variations, are shown in Table 7.14. Table 7 14 Typical comPOSitions
Sulphide S - max
um dehydratlon in a hot mill and are discussed in Section 6.2 and in the paragraph on setting time above. Whilst flash set is irreversible, false set is generally of little practical consequence except in a few central ready-mix concrete plants that use very short mixing times, though it is worth noting that aerated or carbonated cement has a strong tendency to false set, along with strength loss, and setting time is often lengthened. The standard ASTM penetration test methods for early stiffening are C359 for mortar and C451 for cement paste. Although C451 is the optional test under C150, it is widely held that it has little relationship to field periormance of concrete. C359 using mortar gives better correlation, though it is sensitive to water/cement ratio and to mortar temperature. Note also, that false set or flash set may be caused by certain concrete admixtures, particularly water reducers.
Retarded concrete setting - Fresh concrete containing under-sulphated Type V cement and a lignosulphonate wa ter reducer does not set for several days when cold. This is an example of severe cement/admixture incompatibility. The concrete has to be dug out and wasted when this occurs. Slump loss is a normal phenomenon occurring with prolonged mixing. Typically, a 12cm slump will fall to 10cm after 15min and 6cm after 60min. Higher slump losses will occur with porous aggregates, with elevated tempera tures, and with superplasticiser incompatibility, and may be rela ted to: • • • •
insufficient sulphate and high alkali contents in cement accelerated formation of ettringite an excess ot soluble sulphate causing gypsum precipitation inadequate CI' t o control sulphate released into solution.
Mature concrete Low strength of concrete cylinders can result from: • • • • •
high air content incorrect mix, most commonly high water/cement ratio incorrect sampling or moulding impropercuring incorrect capping.
Cement content of hardened concrete can be confirmed (ASTM C1084) by arithmetical comparison of CaO or 5iO l analyses of its constituents.
The Cement Plant Operations Handbook - Seventh Edition
Pop.outs are usually conical, 2-IOcm in diameter at the surface and I -Scm deep. The cause is expansion of aggregate after setting and may be due to: • • • • •
freezing of water in porous aggregate alkali-aggregllte reactivity contamination by burnt lime or dolomite or by broken glass oxidation of sulphide or magnetite in aggregate presence of soft partides such as day lumps, shale, chert and coal.
Alkali-silica reactivity (ASR), is a reaction between alkali in cement and some sensitive aggregates which has ted to the habitual specification of low-alkali cement by engineers in some countries. Reactive aggregates are not ubiquitous and the unnecessary specification of low-alkali cement frequently entails additional production cost and the landfill of process dust. The 0.6 per cent (ASTM C15O) limit may be unnecessary for many aggregates and too high for others. Certain ponolan and slag blends can be effective in reducing the reaction. Cracks - if deep 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. Cracks - plastic shrinkage 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 form -work not pre-saturated.
Cracks - surface 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.
Quality control Attack on hardened concrete by sulphate from external sources often leads to expansio n, cracking
and spalling. ln more advanced stages of attack, concrete may be softened and disintegrated. Sulphates may come from ground- or seawater, soils, road de-icing sa lts and from flue gas emissions, and result in both chemical reactions and physical prnr"S5eS. Incoming sulphate reacts with ca!c:ium hydroxide to fcnn calcium sulphate, which, in tum, reacts with hydrated calcium aluminates to form ettringite in expansive reactions (lea ibid, p. 310). Magnesium sulphate is more aggressive than sodium or calcium sulphate since it also reacts with hydrated calcium silicates teform gypsum and magnesium hydroxide. Long-term concrete exposure studies conducted by PCA indicate that sulphate attack under alternating wet and dry conditions is particularly destructive due to crystallisation pressure. In general, Type V Portland cement, blended cements or high-alumina cements should be used to improve sulphate resistance of concrete. Brown surface discolourations are caused by a very small concentration of Fe2' ions which diffuse from inside concrete through capillary pore solution to the surface and are retained there in the lime efflorescence
produce a dense concrete use cement containing 30-40 per cent slag use cement which has been burned harder to increase C15 formation at the e)(pense of C,5 and free-limp.
A new test method has been developed to test the brown discolouration potential of cements (Haerdtl et aI, 2003). Mapel have proposed a novel blend of alkanolamines that cuts down the level of staining in laboratory tests (Rec.chi et aI, 2013). Surface dusting of floorslabs is caused by: • • • •
high slump concrete premature finishing surface drying premature finishing water adsorbing form work.
Surface scaling is the breaking away of a l -Smm surface layer and may be caused by: • • • •
unsound aggregate freezing and thawing premature finishing excessively fine aggregate (-lSOlJm).
8. Maintenance Maintenance ranges across chem ical, electrica l, mechanical, civil and structural engineering, involving numerous arca ne skills. The activity may be regarded as a production cost, but it is essential if output an d quality are t o be assured: improving maintenance practices offers scope for sign ificant savings. No attempt is made here to add ress the subject in practical deta il. Instead , this chapter presents general concepts, recognising that there are more failure modes than old age and that appropriate analysis of equipment can lead to both greater reli ability and reduced mai ntenance cost. An academ ic paper by Shafeek (2012) provid es a survey of industrial pra cti ce as a background to the study of a Saudi cement plant. The cement industry application of plan ned maintenance is reviewed by Patzke and Krause (1994), cond ition-based ma intenance by Rudd and Wesley (2003) and pla nt engineering by Guilmin (1994). Mea nwhile, useful comments on the cu rrent scene ca n be fou nd in 'Pla nt Services Magazine'.
8.1 Maintenance benefits and costs Maintenance typically represents 15-25 per cent of total manufacturing cost, with best practice said to achieve 10-15 per cent or less. Since the 1970s maintenance departments have had to cut costs whilst increasing plant reliability. This has resulted in smaUerstaff with multiple craft skills, and an increasing use of information and measurement technologies, aided by computer technology. It can be a 'Cinderella' area of manufacturing, where funds disappear into a 'black hole~ with little feel for cost effectiveness or the function 's overall efficiency. Maintenance activity is complex and difficult to get a grip on, and its costs do not often get the same level of attention as energy. However, effective maintenance management is an area for gaining significant competitive advantage. Maintenance obviously costs money, needs people and requires discipline, but there are rewards. Take a 'typical' O.83Mta kiln: the most important things to avoid are kiln stops. Consider two conditions, 'good' being 50 kiln stops/year and 7900h (90 per cent) run, and 'poor' being 140 stops/year and 7300h (83 per cent) run. Costs arise in three categories: power (assumed to be SOkWh/t for steady operation), fuel and lost production. Of course, some stops may not be due to inadequate maintenance: process upsets or poor inventory control can also be involved. Best operating practice maintains kiln operation at a level of 93-95 per cent YoY and, as cost increases when equipment is idle, plants with the highest run factors can register the lowest maintenance costs. Some works, especially those with well-planned control rooms, run stricter discipline for switching off auxiliaries when the kiln stops, classed as 'good' power control. In this situation, good maintenance practice would consume 3kWh/t above the base level, compared to 8kWh/t for poor practice. For a plant with 'poor' power control, both these figures would rise by about 25 per cent. Based upon typica l kiln warm-up schedules, the additiona l fuel consumption in a poor maintenance regime would be about 50kcaljkg, compared to below 20kcaljkg in a good regime. The difference In lost production might amount to 9O,OOOtpa. Adding all these fa ctors, and assuming a sold-out market (and quite moderate profit margin and cost figures), the annual loss due to poor maintenance could easily exceed US$2.5m, without considering knock-on effects upon parameters such as refractory life.
Th e Cement Plant Operations Handbook - Seventh Edition
Other areas can be subjected to similar approximate cost a nalysis, givi ng a 'taster' of the extremely high costs result ing from insufficient or poor maintenance. Even under low demand conditions the cost of unreliability is high. Yanusa-Katingo and Sinha (2013) examine data for a Nigerian cement plant, confirming that there were rewards waitingto be reaped from resources already invested in a condition monitoring system which had yet to yield sufficien tly advance warnin gs of impending failures. As with other areas of plant operation, maintenance can be outsourced to teams run by an equipment supplier, paying a yearly performance-based fee per tonne of product. Syst ems indude those from ABS (Kna benhans, 2011), FLSmidth and several suppliers of equipment for gas sampling and analysis.
8.2 Failure modes Maintenance aims to ensure maKim um a"ailability and effidency of equipment using limited resources of manpower, cost and equipment downtime. 'Ma intenance' is the preservation of equ ipment condition, whil e ' repair' means restoring it to pristine condition; ' patching' is inadequate repair to less than new condition. Historically, cement industry maintenance Involved runningto failure, followed by repairor replacement. In the 1950s the concept of preven ti ve, or operating-time-related maintenance was developed, attempting to predict equipmen t life expe ferrite> belite. C!, is the most soluble major compound and appears to dominate early hydration. Aluminate and, particularly, silicate hydration reactions are extremely complex and many undoubtedly contribute to setting and strength gain of cement Hydration can be approximately divided into four stages (described below and illustrated in Figure
10.1). Figure 10 1 R,lle of heat evolution and structure development of Portland (cmcnt hynri"lt'oo
Rapid formation 01 C-S-HandOi
/ InOOction period i-Icrease i"I Ca>and OH- CDna!I1triltion
