Furnace Cooling Technology In Pyrometallurgical Processes (pdf

Transcript

Sohn International Symposium ADVANCED PROCESSING OF METALS AND MATERIALS VOLUME 4 - NEW, IMPROVED AND EXISTING TECHNOLOGIES: NON-FERROUS MATERIALS EXTRACTION AND PROCESSING Edited by F. Kongoli and R.G. Reddy TMS (The Minerals, Metals & Materials Society), 2006 FURNACE COOLING TECHNOLOGY IN PYROMETALLURGICAL PROCESSES Karel Verscheure1, Andrew K. Kyllo2, Andreas Filzwieser3, Bart Blanpain1, Patrick Wollants1 1 Katholieke Universiteit Leuven, Dept. Metallurgy and Materials Engineering Kasteelpark Arenberg 44, B-3001 Leuven, Belgium 2 GK Williams Centre for Extractive Metallurgy, The University of Melbourne, Victoria 3010, Australia 3 METTOP – Metallurgische Optimierungs GmbH, Peter-Tunner-Strasse 19, A-8700 Leoben, Austria Keywords: furnace cooling, vessel integrity, water cooling, review, freeze lining Abstract Reliable furnace cooling technology is a domain of increasing concern to the pyrometallurgical industry as it can significantly increase process intensities, productivity and campaign times of furnaces. Although there are many advantages in using cooling systems, they also impose a variety of problems mainly related to safety, heat losses and sustainability of the operations. The choice of cooling system is hence a matter of trade-offs and differs for every metallurgical application. This paper gives a review of different cooling designs used in the ferrous, non-ferrous and alloying industries. Additionally, the paper systematically reviews aspects of materials selection, manufacturing, installation, water quality and furnace monitoring when using water cooled refractories. Introduction Furnace cooling technology is becoming increasingly important for the metallurgical industry as several trends in metallurgy pose more severe demands on the smelting vessels. The first trend is the evolution towards higher process intensities. This includes both an increase in process temperatures to improve thermodynamics of the reactions and an increase of the bath agitation to improve reaction kinetics and has the advantage that higher production rates can be achieved and smaller smelting vessels with reduced capital and operating cost may be used. A second trend in metallurgical industry is the evolution towards more complex and corrosive metallurgical phases, which is largely driven by the need for new and more efficient processes to produce and recycle materials. The trends towards higher process intensities and more complex metallurgical phases mostly result, in faster degradation of the furnace. Typically the degradation mechanisms of traditional furnace walls can be classified as chemical, thermal, and mechanical stresses. These can appear as a single stress factor, however, combinations of these stresses occur and affect the furnace lining. The most important chemo-thermal wear mechanisms of traditional furnace linings are corrosion by slag, metallic/slag infiltration, metal oxide/carbon bursting, redox reactions, 139 sulphate attack and hydration. An example of this is the acidic, fayalitic slag in the copper industry that can infiltrate into the open pores of the bricks due to its low viscosity. The depth of the infiltration is also a function of wettability of the refractory oxides, surface tension and temperature of the infiltrate, temperature gradient of the brickwork, and size and distribution of the pores. As a result a degeneration (softening) of the brick’s microstructure occurs due to chemical attack [1,2].. Metallic infiltration densifies the brick’s microstructure without any corrosive attack on the brick components but the result is still a loss of flexibility and brick strength. Due to the changed thermo-mechanical properties, any thermal shock leads to crack formation, particularly at the interface between the infiltrated and non infiltrated brick area which could eventually result in a discontinuous wear by spalling. Using an intensive furnace cooling system the temperature gradient is much steeper in general. This means that for a given process any infiltration will solidify closer to the hot face such that the depth of the infiltration is not as deep as normal. Also, if the cooling effect is strong enough to build up an accretion layer or freeze lining in front of the brickwork, chemical attack can be minimized. There are currently a variety of furnace cooling designs around to reduce the wear rate of furnace linings and to form freeze linings. The integration of such cooling designs is a complex process mostly involving seven steps: 1. analysis of the process specifications 2. development of alternatives for the cooling design 3. selection of the cooling design 4. detailed engineering of the selected cooling design 5. manufacturing and installation of the selected cooling design 6. furnace operation and monitoring 7. evaluation and (re)-engineering of the cooling design The goal of this paper is to review some aspects of this process. First, it will consider the theoretical background of water-cooled furnace walls, and the implications of this theory for furnace cooling design. Second, an overview of the most important cooling designs will be presented. Third, some aspects of detailed engineering, manufacturing and installation of cooling designs will be discussed. This includes the selection of materials, manufacturing of cooling channels, design of refractories and aspects of thermal expansion. Fourth, some important considerations of process operation and monitoring will be reviewed. Finally, a global evaluation of the current designs and configurations will be made and some future trends will be addressed in the conclusion. Theoretical Background The campaign time of water-cooled furnaces may be subdivided in two periods, namely a first period in which sufficient refractory is present to protect the furnace wall and a second period where a freeze lining is formed on top of the remaining refractory or directly on the cooling element. The total campaign time of a water-cooled furnace is hence an important function of the stability of the freeze lining and can be dramatically extended by controlling its behavior through adjustment of the cooling design. The theory of freeze linings provides a better insight into the control of freeze linings and may be applied to analyze the process specifications. In the literature [3-5], it is generally assumed that a freeze lining is formed from the thermal balance between heat input from the superheated bath Qbath, heat removal through the cooled 140 furnace wall Qwall and the latent heat of fusion Qfusion at the hot face of the freeze lining. The thermal balance between these three heat fluxes is mathematically expressed by equation (1): Qbath  Q fusion Qwall (1) In one dimension, the heat fluxes Qbath, Qfusion and Qwall are expressed by equations (2) to (4): Qbath hbath ˜ Abath ˜ (Tbath  Tliquidus ) Q fusion Qwall U ˜H f ˜ k fl ˜ Abath ˜ dt fl dt dT dx (2) (3) (4) hotface where: hbath = convection coefficient from the bath to the freeze lining, Abath = total bath/refractory interface, Tbath = bath temperature, Tliquidus = liquidus temperature of the bath material, = thermal conductivity of the freeze lining, kfl T = local temperature in the wall and freeze lining, x = position in the wall starting at the water/cooling interface and pointing to the bath, U = density of the freeze lining, = latent heat of fusion, Hf tfl = thickness of the freeze lining, dT = thermal gradient in the freeze lining evaluated at the hot face of the freeze lining. dx hotface This gradient can be evaluated by solving the one-dimensional heat diffusion equation in the cooling device, the refractory and freeze lining: w § wT · ¨k ¸ wx © wx ¹ Ucp wT wt (5) where cp is the heat capacity of the material. Equations (1) to (5) are valid for transient and steady state situations. In steady state however, the thickness of the freeze lining tfl remains constant in time so that the latent heat of fusion Qfusion is zero. Moreover the temperature distribution in the wall remains steady so that Qwall may be expressed as: Qwall t fl t  refr k fl Abath krefr Abath Tliquidus  Twater t 1  R  cooling  kcooling Abath hwater Abath ¦ air  gaps where: Twater = cooling water temperature, 141 (6) trefr = thickness of the refractory layer, tcooling = thickness of the cooling device, = thermal conductivity of the freeze lining, kfl, = thermal conductivity of the refractory, krefr kcooling = thermal conductivity of the cooling device, hwater = convection coefficient of the cooling water 6Rair-gaps = sum of the thermal resistances over the air gaps between shell and refractory and refractory and freeze lining. If all constants other than the freeze lining thickness are known, the unknown steady state freeze lining thickness can be determined from equations (1), (2) and (6). These mathematical expressions have a number of important implications for furnace cooling design. First, at steady state the specific heat extraction through the furnace wall Qwall/Awall is equal to Qbath/Abath and is solely determined by the convection coefficient from the bath to the sidewall and by the difference in bath and liquidus temperatures (see equations (1), (2) and (6)). Since these parameters are mostly given for the process, the steady state heat extraction through the furnace wall cannot be changed through adjustment of the cooling design. A cooling design thus needs to be designed to extract this steady state heat flux at all times. It can also be deduced from the same equations that increasing the process intensities through an increase of bath temperature or convection, will both increase the heat input from the bath to the wall (Qbath) and the heat extraction through the wall (Qwall). Intensive processes hence require cooling systems that can extract high heat fluxes. Next, it can be deduced from equations (1) (2) and (6) that an increase in the thermal conductivities of the cooling device, the refractory or the freeze lining will lead to an increase in the steady state thickness of the freeze lining. An increase of these thermal conductivities may be achieved through the use of high-conductivity cooling blocks and refractories and/or insertion of high-conductivity rods in the freeze lining; heat removal through the furnace wall will remain unaffected by this change. In contrast, increasing the thermal resistances by allowing more air gaps between the different layers of the wall or using a lower conductivity refractory will decrease the steady state freeze lining thickness but similarly not affect the heat extraction through the furnace wall. A wall design should hence promote good thermal contact between the layers in the wall so that a sufficient freeze lining thickness is assured. Finally, it can be seen from equations (1) to (4) that any disturbance in Qbath or Qwall will be compensated through melting of the freeze lining or solidification of bath material to the freeze lining (Qfusion). For example, a sudden increase of Qbath, due to a sudden increase of bath temperature or the presence of a phase with high heat transfer coefficients (matte or metal phase), will lead to meltdown of the freeze lining (Qfusion < 0) until a new stationary freeze lining equilibrium is reached. Similarly, a sudden decrease of Qbath, due to a sudden decrease of bath temperature, will lead to solidification of extra bath material to the freeze lining (Qfusion > 0). Solidification will also occur if the thermal resistance of the wall decreases, for example, in case of a complete falloff of the freeze lining. The latter is an important worst case scenario for the cooling design as in this situation it needs to extract both heat input from the bath and the latent heat of solidification required to form the freeze lining. The development of technologies that provide technical and economic solutions to these aspects remains an ongoing problem for furnace operators. An overview of the most important designs is presented in the following section. 142 Overview of furnace cooling designs There have been many different designs of cooling systems developed over the last forty years, which can be divided into six general classes [6]: x x x x x x Spray/Shower coolers Plate/Box coolers Stave coolers Internal blocks/rings Panels External jackets. It should be noted here that there are some differences in terminology used for the various classes of coolers. For the purposes of this paper the terminology used by Leggett [6] will be applied. Cooling systems are often compared based on their maximum heat flux capability, with the selection of cooler type based on the equilibrium heat flux in the specific region of the furnace. In doing this, care should be taken in the use of published comparisons. For example, an often quoted comparison [7] gives the maximum heat flux for plate coolers as of the order of 100 kW m-2, while another source from the steel industry states that standard boiler tubes can be used in plate coolers up to a heat flux of just under 2 MW m-2 with the use of copper allowing a maximum heat flux of 5.8 MW m-2 [8]. In reality, the maximum heat flux will depend on the detailed design of the cooling, the refractory selected and the nature of the process, so a global value for any specific cooler type may not be valid. Spray/Shower Cooling Spray or shower cooling is simply the addition of a water spray or curtain down the outer shell of the furnace. It is widely used in the non-ferrous industry, and has the main advantage that it is relatively inexpensive. An additional advantage of this type of cooling is that it is easily installed later on to minimize the refractory consumption. High wear areas like the slag zone can be cooled using spray or shower cooling with a very low investment. To install such a system afterwards, no openings in the steel structure of the furnace are necessary and therefore the strength of the furnace will be as high as before. The main disadvantage of this form of cooling is that it has a very limited effect. The controlling thermal resistance in a furnace wall using shower cooling is the refractory, which is (initially) thick and usually has a low thermal conductivity. Adding external cooling to the outside will have a minimal effect at the hot face, allowing the refractory to wear back until it can maintain an equilibrium thickness. In many cases this thickness is only a few centimeters, even with external cooling, which is generally not able to support the wall above it. An additional drawback of this form of cooling is that the effect decreases in time due to scaling and deposition of dust and microbiological impurities at the surface of the furnace shell. Shower cooling also has problems counteracting hot spot formation over large areas, since film boiling may occur away from the zones of impingement. Moreover this type of cooling has the disadvantage that heat is removed through the furnace shell, which can lead to thermal stresses and considerable deformation of the furnace. This is especially problematic for large furnaces such as iron blast furnaces where misalignments of several hundreds of millimeters may occur at the top of the 143 furnace. Finally, spray cooling is not effective for zones with a downward decreasing section, for example the bosh of the iron blast furnace. Plate/Box Coolers Plate and box coolers are most commonly used above the slag line in the sidewalls and roof of electric arc furnaces, although these coolers are also applied in the bosh and stack regions of iron blast furnaces [9]. Box coolers are simply steel boxes, cooled externally either by water sprays or forced water circulation, while plate coolers are generally made up of an array of tubes with cooling water circulating through them. For higher heat flux applications copper may be used instead of steel. Both plate and box coolers used in arc furnaces will have a pattern of studs or pins protruding from the hot face to aid the bonding of slag to the steel surface. In most cases there is no refractory involved in the construction of these coolers, with the slag build up from material splashed from the furnace forming a protective layer. In contrast, the plate and box coolers applied in iron blast furnaces are always embedded into the refractory brickwork. They can be removed for maintenance if necessary. In arc furnaces, the main problems with this form of cooler are mechanical damage caused during scrap charging and the potential for stress cracking of the plates or pipes, leading to water leakage into the furnace. Leakage is usually avoided by frequent maintenance, but some design factors can reduce the potential for problems. In particular, pipe thickness is critical to give sufficient strength to withstand mechanical damage, while still being thin enough to avoid problems with thermal stress deformation and cracking [8]. In the blast furnace, the main advantages of box and plate coolers are their relatively simple design, the possibility to replace them online and the possibility for visual detection of hot spots on the shell during operation. It is also claimed by some operators that relatively little copper is used, which leads to lower costs when compared to other coolers, like copper staves. The drawbacks of the box and plate coolers are the possibility for thermal stress cracking at the tip leading to water leakage into the furnace, the large number of holes in the furnace shell leading to gas leaks, the loss of operational volume due to the high thickness of the wall and finally the possibility of scaling if an open water circuit or spray is used. Additional disadvantages are that the cooling is non-uniform (see Figure 1) which leads to uneven wear and disruption of the flow of process gases in the furnace. Moreover, it is also often difficult to mount or replace coolers in the tuyere zones of the furnace. In most iron blast furnaces this technology has been abandoned for staves coolers although there are some exceptions [9]. Figure 1. Non-uniform cooling of the iron blast furnace refractory using copper plates [9]. 144 Stave Coolers Stave coolers, generally used in the bosh and stack of iron blast furnaces, are large water-cooled blocks of metal, usually with refractory inserts between them and the hot face. Initially they were constructed of cast iron with cast-in steel pipes inside, but more recently copper has been used to give the stave increased cooling performance and to allow the formation of a freeze layer to provide extra protection. Besides the “conventional” copper stave there are currently a variety of other copper staves available. Staves have been developed with oval in-mould generated cooling channels [10]. The oval channels have the advantage that the total surface of the cooling channels is high and that the thickness of the stave and the amount of copper can be reduced (see Figure 2). There are also staves available with cast-in pipes made of Monel [11]. STAVELETS are another type of staves which have a reduced weight, machining time and a reduced coolingwater flow rate [12]. The cooling channels of STAVELETS are fabricated by welding at the cold side of the staves. Finally, steel staves are being developed to replace the cast iron staves in the low-heat zones of the blast iron furnace. There are several advantages to the use of staves, such as their uniform cooling, the possibility to directly inspect the furnace shell for hot spots and the possibility to do minor repairs to the cooling channels in case of water leaks. The latter can be achieved by introducing flexible hoses into the leaking pipes. The disadvantages of using stave coolers are the difficulty to replace them during operation and the great amount of copper in some of the designs which makes them expensive, however they are generally reusable and the long life attained in many iron blast furnaces reduces the impact of the cost considerably. Stave wear is generally due to three mechanisms; direct attack by hot metal in the bosh, decarburization due to gas contact and thermal stress/fatigue cracking [13]. The effect of these mechanisms can be minimized by careful design of the cooling channels and ensuring good contact between the cooling pipes and the bulk of the stave. Figure 2. Detail of different copper staves: hot rolled and drilled cooling channels (left) and in-mould generated cooling channels (right) Internal Blocks/Rings Internal blocks or rings are commonly used in a variety of non-ferrous furnaces as well as in some iron blast furnaces [14]. They are blocks of copper that are placed within the refractory. These blocks may encircle a cylindrical furnace section as rings, they may be long straight sections or as small, isolated blocks scattered throughout the furnace. Cooling may be either internal or external to the furnace. Internal block coolers have a number of potential problems, particularly if the cooling is internal to the furnace. In this case water leakage can occur as the blocks age, giving the risk of a steam explosion. More commonly, the use of this form of cooling leads to extremely uneven wear, as the cooling is very localized with the low thermal 145 conductivity of the refractories limiting its effectiveness. While the uneven wear can be countered by decreasing the spacing between the blocks, the amount of copper required can make this approach uneconomical. Panels Panels are basically the same as box coolers, but are applied below the slag line, usually in nonferrous applications. They have only limited use at present and only in certain technologies, particularly the zinc slag fuming process and the Vanyukov process. A modification of the panel technology, combining large copper blocks with a layer of refractory, is more common and variations of this are available from major engineering companies. The primary difference between the variations relates to the profiles used on the hot face to maintain the refractory or slag layer (see Figure 3). The major challenge for this type of cooler is in maintaining a protective layer on the metal. This can be particularly difficult in batch operations where thermal cycling can lead to spalling of a low conductivity layer on the metal [15]. The amount of copper involved in these coolers does allow them to withstand very high heat fluxes. Figure 3. Cast EAF Wall Panel [16]. The hot face pockets are designed to retain refractory or slag. External Jackets External jackets are often used as temporary measures in cases where extreme wear has occurred in a relatively local area leading to hot spots on the external wall of the furnace. They have also been used for below bath applications to provide more intense external cooling than can be achieved by spray/shower cooling. This form of cooling has only a limited effect, for the same reasons as has been discussed for spray/shower cooling. An additional drawback of the water jackets compared to spray/shower cooling is that visual detection of hot spots is impossible. Specialty Designs In recent years specially designed cooling systems have been developed to counter the shortcomings of the various existing cooling technologies. These include the modified plate coolers [10-12], variations on block coolers and the Composite Furnace Module Cooling System (see Figure 4), which is a modular design allowing customization for specific furnace regions and has the aim of providing an essentially uniform hot face temperature [17]. This last feature minimizes uneven wear on the surface and can reduce the effects of thermal stresses. The basic principle of the modules is to spread the copper more uniformly throughout the refractory to provide even cooling. The heat flux that can be withstood can be varied by altering the amount of copper present, as well as the refractory material used. This large degree of customization allows the design to use the minimum copper required to provide for the cooling needs of a particular area and also reduces the total heat load on the furnace by not overcooling. 146 Figure 4. The Composite Furnace Module Cooling system Detailed Engineering, Fabrication and Installation of Furnace Coolers Materials for Cooling Design The selection of the metal used in furnace coolers is primarily dependent on the amount of cooling required. In furnaces where the cooling requirements are not particularly high, cast iron or steel is often used. Examples of this are some blast furnace staves where cast iron and steel have been applied, and zinc fuming furnace jackets which are mostly made of steel. Different types of cast iron have been applied (grey and nodular cast iron) because a trade-off between thermal properties (thermal conductivity) and mechanical properties (ductility, fatigue resistance) needs to be made. Steel is used in situations where there is a high strength requirement and extreme cooling is not required such as in some plate coolers and furnace shells used in shower or spray cooling or water jackets. Steel has been tested and used to replace cast iron for staves. For intense cooling copper is the material of choice. It combines a high thermal conductivity with a reasonable melting point and relatively low cost. However, the purity of the copper is important, as even a low level of some impurities can have a significant effect on the thermal conductivity. Copper is also subject to corrosion under some circumstances, particularly in the presence of oxygen at high temperatures. This form of corrosion will occur in some cooler designs, usually where the copper is exposed but is of only limited concern in designs where the copper is fully enclosed in refractory. If new or unique processes are used it is particularly important that the potential for copper corrosion is taken into consideration. When combining copper and steel in piping, careful design considerations should also be made to prevent galvanic corrosion. Some work has looked at the potential for the use of copper with about 4% aluminum, which has been shown to reduce the rate of oxidation, however the reduction in the thermal conductivity by about 40% limits the use of such alloys [18]. Besides pure copper, copper-nickel alloys such as Monel have also been used in some applications as a material to make cast-in pipes [11]. Fabrication techniques Another important factor in the design and construction of furnace coolers is the fabrication technique, particularly with respect to the inclusion of the cooling. This is most important for cooler designs that have the water channels within a block of metal. There are three main techniques for the inclusions of the piping: drill-and-plug, cast around pipes and in-mould generated cooling channels. Two other techniques that have been applied are welding plates to 147 cast or machined coolers and welding pipes or half pipes to the outside of the cooler. Each technique has its advantages and problems. The drill-and-plug technique is relatively simple, involving the drilling of holes directly into the block, with the unused external holes plugged. This gives direct contact of the coolant with the block and so avoids the possibility of air gap formation, which can significantly reduce the heat removal rate. The major drawback of the technique is that the plugs used can leak, particularly if there has been deformation of the cooler, either from mechanical or thermal stresses. There is also the potential for regions of stagnant water to form, leading to the potential for increased corrosion and vapor formation. In addition, it is more difficult to apply in cases where the coolers are curved or intense cooling is required. Cast around pipes are used in both copper and cast iron coolers. This technique involves the casting of the cooling block around the preformed pipes for the coolant. The advantage of this technique is mainly the flexibility in terms of cooling channel geometry. The primary drawback of this technique is the potential for the formation of air gaps at the interface between the pipe and the casting, either during the casting process, or under the effects of thermal stresses during operation. In cases where a combination of steel pipes and a cast iron core is used, there is also the danger for diffusion of carbon into the steel pipe (cementation), which leads to a decrease of its mechanical properties. There are several techniques that may be used to ensure good contact and limited cementation. Plasma spraying of a thin coating of alumina has been used in the production of cast iron blast furnace staves, however this was found not to be fully effective, with an air gap of 0.1-0.3 mm still formed [19]. In copper systems, the use of Monel pipes has been patented, and is reported to give a good bond [11]. Öhler describes a patent concerning “copper in copper” cast [20] where the surface of the copper pipe is coated electrolytically with a nickel layer prior to casting. The use of hot isostatic pressing has also been proposed to remove any air gaps formed during casting, but this may not solve the delamination problem during operation. In-mould generated channels have been used mainly in copper staves. This technique has the advantage that oval cooling channels can be easily fabricated. Oval cooling channels have also been fabricated by drilling. Coolers with external cooling have some added options, including welding plates over cast-in channels and welding pipes or half pipes to the outside of the cooler [21]. Cast-in channels avoid the potential for air gap formation, but require that the finished block be pressure tested to ensure that there will be no coolant leakage. Welding pipes to the outside surface still gives limited contact and is of very limited application; however the use of half pipes gives direct contact of the coolant to the block and can give efficient cooling [12]. Refractory Design The following aspects have to be taken into consideration using highly effective furnace cooling systems: Especially in the non-ferrous metals and ferro-alloy industry, a basic brick lining is required. One of the most important issues is the possibility of brick hydration. Generally, hydration of basic bricks occurs in humid atmospheres, usually at temperatures between 40-120°C. It is characterized by the transformation of periclase into brucite according to the reaction: MgO + H2O = Mg(OH)2 148 (7) This reaction is associated with an increase in the volume of up to 115% [1]. In extreme cases an extensive characteristic crack formation can lead to a sand-like disintegration of the whole brick. The volume increase leads to an uncontrolled brickwork movement and the furnace shell can be affected. Temperatures between 40°-120°C occur in every furnace at a certain point during the heating up procedure, and in cooled systems, including around block coolers, the normal operating temperature of a section of bricks may be in this region. The water source could be mortar or castable, condensation of air humidity or off-gas humidity or water which is coming from outside (for example water leakage from cooling elements). In comparison to furnaces without cooling systems, the water which is already vaporized can condense on cold cooling elements and the surrounding brickwork will be in the critical temperature range much longer. To minimize the risk of hydration the following can be done: - Proper refractory qualities and lining concepts have to be chosen. If possible MgO-free bricks should be used in the critical area concerning hydration, or at least MgO qualities having a CaO to SiO2 ratio lower than one, due to the fact that free CaO increases the risk of hydration. To minimize the amount of water, carbon based – water free – material can be used for ramming mixes and so on. - To guarantee a cooling system without any leakages all cooling devices should be tested under pressure outside the furnace. - The heating up procedure itself has to be adjusted. To avoid hydration all critical areas should be heated up above 400°C very quickly using a temperature increase between 3050 °C/hr. Later on a much more gentle temperature increase is important to obtain a very uniform temperature distribution in the whole furnace. In any case, a very good monitoring during heating up is essential. Expansion considerations In general, expansion problems occur due to differences in thermal expansion coefficient for different materials and the non-uniform temperature distribution. The expansion must be compensated for in the lining design, and therefore a knowledge of the thermal expansion and the reheat changes of the different materials to be used, and of the temperature at each point is imperative. If the thermal expansion is not taken into consideration, or insufficient expansion allowance is made, it can have significant effects on the furnace steel casing. Too high an expansion allowance leads to a leaky lining, joints are not tight and a liquid phase can infiltrate the lining [22]. Cooling elements in sidewalls, for example, influence the temperature distribution in the wall and in the bottom brickwork. Additionally, cooling elements in sidewalls can require a lot of openings in the steel shell or are part of the furnace shell structure itself. In either case this results in an increased weakness of the furnace structure, and this has to be taken into consideration in the expansion allowance calculation. After each heating to high temperature and cooling down cycle, length variation in the bricks remain. This means that a furnace which is heated up a second or third time will not see the same expansion effects as the first time. This effect must be considered at all steps of furnace and/or cooling element design work. As already mentioned, an incorrect expansion allowance calculation or brick hydration can easily destroy a furnace. As soon as the cooling system design is finished and all important operational parameters like water flow rate, heat flux and water temperature are defined, a temperature 149 distribution in the furnace should be calculated. This could be done by using commercial CFD (computational fluid dynamics) software. Furthermore the position and number of heat inlets should be calculated to optimize the heating up procedure [23]. Furnace operation with water cooling Water Quality and Water Temperatures An important aspect of furnace cooling is the quality and operating temperature of the cooling water. A general rule of thumb is that the water quality must be higher when higher water temperatures and/or heat fluxes are reached. This is required because the calcium salts dissolved in water have an inverse solubility product and precipitate when the temperatures are increased [24]. Precipitation of calcium salts leads to scale formation in the pipes and an associated decrease in heat transfer. This may eventually lead to burn-though of the cooling devices. Different types of cooling of metallurgical furnaces can be distinguished [25] of which the three most important ones are (1) cooling with cold technical water, (2) cooling with chemically purified hot water and (3) evaporative cooling. When cooling with cold technical water, the water reaches bulk temperatures up to a maximum of 35°-50°C: once through circuits or return circuits can be applied in this case. The advantage of using cold technical water is that complex multi-stage purifications can be avoided. The main drawback is that enormous water flow rates are needed to keep the water temperature sufficiently low and prevent scaling. Large piping and pumping equipment is needed to cope with these high flow rates. Moreover, the heat stored in the cold water cannot be recovered due to its low temperatures. Using cold technical water in return circuits has the advantage that the amount of dust and dissolved gases in the water can be kept low so that limited scaling and oxidation of the channels occurs. When cooling with chemically purified hot water, the maximum bulk water temperatures vary between 70° and 95 °C and a closed loop circuit is required to maintain the water quality. The advantages of hot cooling water include partial or full recovery of the extracted heat, lower pumping costs compared to cold technical water due to lower water consumption, and reduced burn-through of cooling elements due to reduced scale formation. Hot water cooling also has some drawbacks, such as the increased dependence on heat consumers or heat exchangers and backup systems, a more complex configuration and more complex monitoring equipment. Evaporative cooling has frequently been used in different furnaces, including the iron blast furnace [25], but is nowadays rarely used. Evaporative cooling systems have mostly been operated as pure thermosiphon systems or as thermosiphon systems in combination with circulation pumps. Since the water temperatures exceed the boiling point of water, chemically purified water in a closed loop is required to prevent scaling. In principle, evaporative cooling has the advantage that higher heat transfer coefficients are reached and less water is needed to cool the furnace. This would lead to lower pumping costs, safer operation, the possibility for efficient heat recovery and more sustainable operation. In practice however, evaporative cooling has several disadvantages such as more difficult leak detection, the need for heat consumers near the plant, higher temperatures of the cooling blocks, erosion of copper pipes when high velocities are used and difficult online maintenance due to the danger of burning injuries. Moreover, there are several boiling instabilities that may occur in evaporative systems [26]. The first instability is film boiling (film blanketing), this is the formation of an insulating steam layer at the water/pipe interface when the imposed heat flux exceeds a certain critical value. Secondly, 150 evaporative cooling systems can also exhibit multiple steady state flow rates (Ledinegg instability) and are also prone to oscillatory instability (steam between water slugs can act as a spring). Finally, in closed loop systems with multiple parallel pipes, local variations and oscillations in flow rate may occur. Due to these problems and the associated reduction in design freedom this type of cooling has been largely abandoned. Cooling Element Monitoring The methods used to monitor cooling elements in the non-ferrous and ferro-alloy industry are mainly - optical inspection - water flow rate control - temperature control of cooling water - pressure drop control - temperature control of cooling elements and/or brickwork Above the bath, accretion layers and the thickness of the accretion layer can often be optically inspected. In this case, regular inspection could provide good information about the cooling efficiency. The main drawbacks of this method are that the whole cooled area can normally not be seen and that it only gives qualitative information. The water flow rate, combined with the water temperature increase, gives information about the amount of extracted heat. For an experienced operator the heat loss and the process parameters are indications of the effectiveness of the cooling device. Experienced operators are also necessary to determine unusual changes in the temperature of cooling elements or the surrounding brickwork. Water leakages are usually not detectable using the above mentioned method. Detection of water leaks is therefore often accomplished by measurement of the pressure drop between the water inlet and outlet or by comparing the water flow rates at inlet and outlet. This monitoring is more easily done in a closed water circuit compared to an open circuit. It is also advantageous to subdivide the cooling circuitry in multiple modules, especially in large furnaces, so that leaks can easily be located. Much more attention has to be given to monitor cooling elements which are located under the bath level. Finally, temperatures of the cooling elements and brickwork can be measured to obtain information on the operation of the process and the integrity of the furnace wall. In combination with the data from other measurements and modeling, this information is particularly useful to provide early warnings to furnace operators [27]. Generally it can be said that the better the monitoring of cooling elements the better for the whole process and for the lifetime of the furnace. The efforts that are being undertaken in this domain should therefore be highly encouraged. Conclusion The continuing increase in process intensity in the metallurgical industry will require an expansion in the use of water cooling to retain vessel integrity. There are now a wide variety of cooling designs available, and the integration of the best fit for any given application requires careful consideration at each stage of the integration process. In general, it is best to design the coolers based on the expected maximum heat flux from the furnace. Over-design can lead to 151 considerable capital costs, while under-design can lead to catastrophic failure. The latter can be minimized by the use of cooling systems that keep the water outside of the furnace shell, but failure may also be avoided by ensuring a good understanding of the heat transfer within the furnace, good control over the furnace operation, careful design of the furnace system to allow for thermal expansion and a good knowledge of the thermal properties and stability of the materials used in the furnace construction. Research in the area of furnace cooling technology is fundamental to achieve these goals and may be facilitated through good communication between different furnace building companies, refractory companies, installation companies, customers and universities. Clearly, a further increase of process intensities will be made possible through the use of watercooled freeze linings. This will, however, come at the expense of additional heat losses through the wall [3]. A promising technique to increase process intensities of bath smelting processes while maintaining low heat losses and enhanced safety is the technique of slag engineering [28]. By adjusting the slag towards certain compositions, high liquidus temperatures, high metal recovery, low heat losses and increased safety may be achieved. Further development should also be in the area of higher conductivity refractories, which complement the metal components of the coolers, potentially allowing the use of less expensive and/or more corrosion resistant metals. Additionally, the development of designs which allow the recovery of the energy from the system should be considered to give better overall sustainability of the process. References 1. D. Gregurek, Ch. Majcenovic, “Wear mechanisms of basis brick linings in the non ferrous metals industry – case studies from copper smelting furnaces”, RHI Bulletin, 1 (2003), 17-21. 2. H. Barthel, “Stresses and wear of chrome-magnesite bricks in furnaces used in the copper industry”, (TMS Paper Selection A80-18, 1980). 3. D.G.C. Robertson, S. Kang, “Model studies of heat transfer and flow in slag-cleaning furnaces”, (Paper presented at TMS Annual Meeting, Fluid Flow Phenomena in Minerals Processing, San Diego, California,1999), pp. 157-168. 4. J. Thonstad, S. Rolseth, ‘‘Equilibrium between bath and side ledge. Basic Principles’’ (Paper th presented at the 112 AIME Annual Meeting, The Metallurgical Society of AIME, Atlanta, Georgia, 1983), 415-423. 5. V.R. Voller, “Modeling of ledges and skulls”, (Paper presented at the EPD Congress, Anaheim, California, 4-9 February 1996), 673-681. 6. A.R. Leggett and N.B. Gray, “Development And Application Of A Novel Refractory Cooling System,” (Paper presented at Advances in Refractories for the Metallurgical Industries II, M. Rigaud and C. Allaire, Eds., The Metallurgical Society of CIM, Montreal, Canada, 1996), 519532. 7. J. Merry, J. Sarvinis and N. Voermann, “Designing Modern Furnace Cooling Systems”, JOM, Vol. 52, (2) (2000), 62-64. 152 8. O. Huscher, “Water-cooled vessels in modern high-performance electric arc furnaces”, MPT International, 4 (1996), 36-40. 9. R. Stokman, E. van Stein Callenfels, R; van Laar, “Blast furnace lining and cooling technology – Plate cooler designs – Corus Ijmuiden experiences”, (Paper presented at the AISE, Pittsburg, Pennsylvania, 2003). 10. J.A. Carpenter, “The evolution of Blast Furnace Cooling”, (Paper presented at the AISE, Pittsburgh, Pennsylvania, 1999). 11. E. van Stein Callenfels, R. Van Laar, R. Boonacker, “High performance MTT coolers for demanding furnace applications” (Paper presented at the AISE, Pittsburgh, Pennsylvania, 2000). 12. P. Heinrich, J. Kapischke, F. Reufer, R.G. Helenbrook, “Stavelets – The next logical step in copper stave technology” (Paper presented at the 2001 Ironmaking conference, Baltimore, MD) 141-152. 13. Y. Murai, T. Yamamoto, K. Aoyama and K. Fukamachi, “Design and Evaluation of Blast Furnace Refractories and Cooling System Based on Thermal Load Distribution”, (Paper presented at ISS/AIME Ironmaking Conference, Detroit, vol. 44, 1985), 475-483. 14. R. Stokman, E. van Stein Callenfels, R; van Laar, “Blast furnace lining and cooling technology: experiences at Corus Ijmuiden”, Iron and Steel Technology, November 2004, 21-28 15. K.E. Scholey, I.V. Samasekera and G.G. Richards, “Thermal Stress Failure of Water Cooled Zinc Fuming Furnace Jackets,” Can. Metall. Quart., 33 (4) (1994), 77-84. 16. G. Slaven, A. MacRae and L. Valentas, “The Implementation of UltralifeTM Copper Casting Technology in the EAF”, (Paper presented at AISE, Pittsburg, PA, 2003) 17. A.K. Kyllo, N.B. Gray, D. Papazoglou and B.J. Elliot, “Developing Composite Furnace Module Cooling Systems”, JOM, 52 (2) (2000), 66-67. 18. G. Plascencia, T.A. Utigard and D. Jaramillo, “Copper Cooling Fingers for Furnace Refractories”, JOM, 57 (10) (2005), 44-48. 19. A. Thomas, K. Edwards and S. Webb, “BHP experience in Design, Manufacture and Operation of Staves”, (Paper presented at 4th Intl. Stave Conf., Hamilton, Canada, 1986). 20. Ch. Öhler, “The new „MC – Technology“ from Hundt & Weber”, (Paper presented at the GDMB Copper Committee meeting, Wroclaw, Poland, 29.3 to. 31.3.2006) 21. Y. Hara, T. Takahashi, T. Sakuraya and M. Saito, “Vessel for Molten Metal”, Japanese Patent Application No. 179569, 1990. 22. T. Prietl, H. Antrekowitsch, A. Triessnig, H. Studnicka and A. Filzwieser: “The Evaluation of Refractory Linings Thermo Mechanical Properties”, (Paper presented at the European Metallurgical Conference 2005, September 18-21, Dresden, Germany, Vol.3, GDMB) 10991112 153 23. O. Zach, K. Gamweger, G. Lukesch and A. Filzwieser, “CFD-Modelling in the Non Ferrous Metals Industry”, (Paper presented at the European Metallurgical Conference 2005, September 18-21, Dresden, Germany, Vol.3, GDMB), 1317-1324. 24. J. Gleason, “Establishing peak electric arc furnace cooling water performance”, (Paper presented at the 2000 Electric Furnace Conference Proceedings, Orlando, USA, 2000), 339-348 25. S. Andonyev, Evaporative cooling of metallurgical furnaces, (Mir Publishers Moscow, 1976), 15-35 26. D. Butterworth and G.F. Hewitt, Two-Phase Flow and Heat Transfer, (east Kilbride, Great Britain: Oxford University Press, 1979), 341-373 27. T. Plikas, L. Gunnewiek, T. Gerritsen, Monte Brothers and Alan Karges, “The predictive control of Furnace Tapblock operation using CFD and PCA Modeling”, JOM, October 2005, 3743 28. K. Verscheure, M. Van Camp, B. Blanpain, P. Wollants, P. Hayes, E. Jak, “Zinc fuming processes for treatment of Zinc-containing residues”, (Paper presented at Lead and Zinc 2005 Conference, Kyoto, Japan), 943-960 154