Heat transfer engineering an international journal, tập 32, số 3 4, 2011

Heat Transfer Engineering, 32(3–4):189–196, 2011 Copyright C Taylor and Francis Group, LLC ISSN: 0145-7632 print / 1521-0537 online DOI: 10.1080/01457632.2010.503108 editorial Heat Exchanger Fouling: Mitigation and Cleaning Strategies 1,2 ¨ H MULLER-STEINHAGEN, M R MALAYERI,2 and A P WATKINSON3 Institute of Technical Thermodynamics, German Aerospace Centre (DLR), Stuttgart, Germany Institute of Thermodynamics and Thermal Engineering, University of Stuttgart, Stuttgart, Germany Department of Chemical & Biological Engineering, University of British Columbia, Vancouver, Canada Heat exchangers are the workhorse of most chemical, petrochemical, food-processing, and power-generating processes The global heat exchanger market is estimated to top a total of $12.7 billion by 2012, with an increase of 3–5% per annum [1] Despite this very positive market outlook, manufacturers are under increasing pressure to produce heat exchangers that are more efficient in terms of heat recovery and use of material, while at the same time being faced with fluids that are increasingly difficult to process One major problem directly related to these requirements is the deposition of unwanted materials on the heat transfer surfaces, which occurs in the majority of heat exchangers [2] Conservative studies estimated that heat exchanger fouling leads to additional costs in the order of 0.25% of the gross domestic product (GDP) of industrialized countries, and that it is responsible for 2.5% of the total equivalent anthropogenic emissions of carbon dioxide [2, 3] Therefore, efficient mitigation and cleaning methods must be available to safeguard the operation of heat exchangers Two basic approaches are possible to combat heat exchanger fouling, namely, mitigation (including on-line cleaning) and off-line cleaning techniques The general criteria for the selection of any of these strategies are: • Extent of required cleanliness • Mitigation and cleaning costs • Time intervals between cleaning cycles • Dominant fouling mechanism • Severity of fouling • Type of heat exchanger Heat exchanger fouling may effectively be mitigated at the design stage of the heat exchanger through the following steps: Address correspondence to Dr Mohammad Reza Malayeri, Institute of Thermodynamics and Thermal Engineering, University of Stuttgart, Pfaffenwaldring 6, D-70550, Stuttgart, Germany E-mail: m.malayeri@itw.uni-stuttgart.de In the following, various mitigation and cleaning techniques are discussed and areas for further developments are identified MITIGATION APPROACHES Figure lists the main methodologies for the mitigation of fouling in industrial heat exchangers It is understood that mitigation techniques are not limited to those given in Figure 1, which have been selected because they are widely used and known to be successful in a number of applications The general preference is to mitigate fouling firstly through proper design of heat exchangers, then by on-line mitigation techniques In reality, a combination of these methods may be necessary to combat fouling Mitigation of Heat Exchanger Fouling by Design • Selection of a suitable heat exchanger type and geometry • Omission of operating conditions that promote fouling • Optimum design with adequate velocities in the heat exchanger and that avoids hot spots, bypass flow, or dead zones ã Design for easy cleaning 189 ă H MULLER-STEINHAGEN ET AL 190 Fouling mitigation techniques Design of heat exchangers Feed dilution On-line Change of operating conditions Feed filtration Chemical inhibitors Thermal shock Shear stress increase Projectiles Reverse flow or pulsation Mechanical Inserts Wire brushes Physical Scrappers Surface coating Ultrasound Electric, magnetic Gas rumbling Figure Various fouling mitigation methodologies Măuller-Steinhagen [4, 5] extensively discussed how these various options can be implemented in the design stage of heat exchangers to reduce fouling Compact heat exchangers, such as plate-and-frame heat exchangers, spiral flow heat exchangers, and fin-tube heat exchangers, have been found to experience reduced fouling in many (but certainly not all) applications due to increased level of turbulence, reduced surface temperatures, and homogeneous flow distribution [5] While this has only been known for convective heat transfer to liquids, Esawy et al [6] recently showed that the buildup of deposits during pool boiling of CaSO4 solutions can be substantially reduced by the presence of fins on the tube outside (Figure 2) Scraped-surface heat exchangers [7] where rotating installations continuously keep the pipe internal surfaces free from deposits have been used in industry for many years Their investment, operation, and maintenance cost, as well as the complex geometry and maintenance, limit this technique to applications where very severe fouling occurs Fluidized-bed heat exchangers are a very effective technology to reduce or even eliminate scale formation in many types of applications Particles of different materials and shapes are transported upward through the vertical heat exchanger tubes together with the fouling liquid [8] They are then separated from the liquid and returned in an external downcomer In addition to having a slightly abrasive effect on the heat transfer surface, the particles will also improve the tube-side heat transfer coefficients Typical applications include, e.g., desalination, processing of aqueous solutions, and processing of hydrocarbons heat transfer engineering On-Line Mitigation of Heat Exchanger Fouling The purpose of on-line mitigation is to keep the heat transfer surfaces in an acceptable state of cleanliness in order to maintain high operating efficiency and plant availability On-line mitigation includes many different methodologies that can be divided, in order of their applicability, into (i) changing operating conditions, (ii) chemical, (iii) mechanical, and (iv) physical approaches A breakdown of these techniques has already been presented in Figure 1, and this subsection briefly outlines each Figure Comparison of fouling resistances for smooth and finned tubes at a heat flux of 200 kW/m2 and a CaSO4 concentration of 1.6 g/L [6] vol 32 nos 3–4 2011 ă H MULLER-STEINHAGEN ET AL individual technique and its limitations More details are provided in ref [5] Filtration and/or washing Removal of contaminants can substantially reduce fouling For example, removal of materials such as sodium, sulfur, or vanadium from fuels prior to combustion and contaminant removal from combustion gases are two approaches to mitigate gas-side fouling Water washing has helped to overcome some of the fouling problems experienced with crude oils and with residual oils in marine applications by removing sodium and sediment Feed dilution or blending Refineries are increasingly becoming more complex as heavier crude oils need to be processed Accordingly, more severe fouling is expected, particularly in the refinery preheat heat exchanger trains Among various mitigation techniques, diluting feed by blending light and denser crude oils may be considered to ease the problem However, a thorough and careful chemical analysis of the crude is necessary, since this procedure may sometimes lead to even harsher fouling [9] Thermal shock Short-time under- or overheating of the heat transfer surfaces may cause brittle deposit layers to crack due to the different thermal expansion of tubes and deposits Figure shows the impact of sudden reduction of heat flux during pool boiling of CaSO4 solutions when the thickness of deposit layer had reached about mm The whole deposit layer spalled off the surface instantaneously [10] Intermittent changes in flow direction or velocity Regular reversal of flow direction or short-time increase of the flow velocity (flow pulsation) has been used to mitigate the formation of weakly adhering deposits Generally, better performance is achieved by continuously operating at a higher flow velocity However, this technique may be effective if applied right from the beginning of operation, before the deposit layer starts to harden Gas rumbling Deposits with moderate stickability to the heat transfer surfaces (e.g., particulate and some biological deposits) can be dislodged and washed out by periodically increasing the fluid shear forces for a short time by introducing compressed air or nitrogen into the liquid system The resulting 191 Table Categorization of chemical inhibitor agents for different fouling mechanisms [4] Fouling mechanism Crystallization, precipitation Foulant Inhibitor agent Ca2+, Ion exchange Mg2+ CaCO3 CaSO4 Soft and hard scalants Soft and hard scalants Particulate Particulate matter Chemical reaction Oxygen (polymerization) Metals (reaction catalyst) Insoluble hydrocarbon particles Biofouling Micro- and macroorganisms Corrosion fouling Passivating oxide layer Ph control Scale inhibitors (e.g., ethylenediamine tetraacetic acid [EDTA]) Adsorption agents (e.g., polyphosphates) Crystalline weakening agents (e.g., polycarboxylic acid) Surfactants or dispersants Antioxidants Metal deactivators Dispersants Oxidants (biocide, chlorine) Passivating oxidants and pH control highly turbulent gas–liquid two-phase flow can provide shear forces and pressure fluctuations that are substantially higher than for single-phase flow Gas rumbling is commonly used in cooling water applications Chemical fouling mitigation methods The most widespread mitigation strategy during on-line operation of heat exchangers is the use of chemical agents or inhibitors, which is particularly useful for heat exchangers with complex geometries where no other cleaning methods are possible Commercial antifoulants are usually polyfunctional and hence more versatile and effective, as they can be designed to combat various types of precursors that may be present in any given system For instance, for crude oil fouling, various precursors such as oxygen, metals, salts, and asphaltenes may lead to different forms of deposit formation Antifoulants are designed to prevent equipment surfaces from fouling but are usually not effective in removing already formed deposits Therefore, antifoulant addition should Figure Breakage and removal of a deposit layer formed during pool boiling of a CaSO4 solution with 1.6 g/L and a heat flux of 300 kW/m2: (a) during steady-state heating and (b) immediately after switching off the heater [10] heat transfer engineering vol 32 nos 3–4 2011 192 ă H MULLER-STEINHAGEN ET AL Figure (a) Typical spiral insert (SPIRELF system) [15] and (b) hiTRAN wire matrix insert [16] be started immediately after equipment is cleaned The usage and dosage of antifoulants depend strongly on fouling mechanisms and anticipated deposit hardness Thus, information about the prevailing fouling mechanism and the influence of operating conditions such as dominant precursors, temperature and velocity are important Table lists typical inhibitor agents for different fouling mechanisms and foulants On-line chemical fouling mitigation is effective but the chemical agents may contain substances that are potentially harmful to the environment, such as chlorine, hypochlorite, polyphosphate, coagulants, etc The use of many of these chemical inhibitors has to be reduced and eventually phased out due to the implementation of restrictive environmental legislations such as the Water Framework Directive 60/2000/EC of the European Union Furthermore, compatibility of the chemistry of the inhibitors with the metallurgy of the equipment has to be checked, to avoid corrosion or cracking Increased efforts are dedicated to the monitoring of fouling, development of less toxic substitute additives, and optimization of inhibitor dosage For instance, chlorine can be replaced by other chemicals such as methylene thiocyanate or chlorophenoles [11] To reduce the dosage of treatment chemicals, Ferreira et al [12] reported that antimicrobials can be transported on micro-sized particles in much lower concentrations to target only microorganisms on the surface On-line mechanical mitigation techniques If applicable, mechanical mitigation may have some advantages over chemical methods, which often involve materials that are difficult to handle and control Their applicability is usually determined by the type of fouled heat exchanger, deposit intensity and growth rate, operating conditions, and cleaning costs The utilization of on-line mechanical fouling mitigation may lead to significantly reduced maintenance downtime, avoidance of antifouling chemicals, and more efficient plant operation These need to be balanced against investment and operating costs, e.g., for replacement of devices due to wear or increased pressure drop due to flow resistance The previous conference proceedings on heat exchanger fouling and cleaning [13] provide an extensive source of information on mitigation and cleaning Hence these techniques are only briefly addressed in this editorial Cleaning Projectiles Projectiles of different shapes, e.g., sponge balls and wire brushes, can be propelled through the heat exchanger tubes to heat transfer engineering remove deposits already during the early stage of formation The frequency and duration of application depends on the severity of fouling and the strength of interaction between cleaning projectile and deposit Typically, projectile on-line cleaning techniques are limited to aqueous systems at temperatures below about 120◦ C, due to the stability of the projectile material There may also be some limitations due to chemical incompatibility If the application of cleaning projectiles to individual tubes occurs at random (i.e., in sponge ball systems), this may lead to overand undercleaning of tubes depending on their location in the tube bundle The “CleanEx” project [14] that has recently been funded by the European Community endeavors to address some of these drawbacks The installation of mechanical systems for the continuous propulsion of cleaning devices requires modifications of the flow system and is, therefore, best implemented in the design stage of heat exchangers Tube Inserts Tube inserts, such as twisted tapes, coils, and wire matrix inserts, can significantly increase the heat transfer coefficients by acting as turbulence promoters (see Figure [15, 16]) As deposition rates for most fouling mechanisms are inversely dependent on fluid wall shear stress and heat transfer surface temperature, reduction of the viscous and thermal sublayer thickness may also considerably reduce fouling These inserts work best for flow in the laminar or transitional flow regime In combination with further reduction of flow velocity (i.e., tube passes), design variations may be possible where significant improvements of heat transfer can be achieved with no or little increase in pressure drop The selection of a particular type of insert and insert geometry depends on the type of fouling and the availability of suitable strainers or filters that may trap particulate or fibrous matters before these enter the heat exchanger Physical Mitigation Techniques Physical fouling mitigation methods attempt to reduce/avoid fouling without changing heat exchanger layout, operation, or chemical additives, by modifying the interaction of deposit forming precursors and heat transfer surface Surface modification Of several fouling mitigation techniques, surface modification is gaining increased attention due to its environmentally friendly features Surface coatings with vol 32 nos 34 2011 ă H MULLER-STEINHAGEN ET AL On-line mitigation systems Chemical Mechanical Environmental hazards Lack of effective control and timing Health hazards Ineffective distribution Increased costs Increased pressure drop Over-dosage Limited to certain chemicals Possible corrosion impacts Abrasive impacts Require modification of heat exchanger Figure Limitations of various chemical and mechanical mitigation systems organic materials such as polytetrafluoroethylene (PTFE) and Săakaphen have been shown to reduce fouling from various fluids, for example, during seawater evaporation and heat transfer to Kraft black liquor The main reason why such materials/coatings are not more widely used is that they are poor heat conductors and form an additional resistance to heat transfer that is comparable to the TEMA fouling resistance for cooling water Another drawback is the poor stickiness of coatings to the substrate If very thin coatings are used, the resistance against erosion or other mechanical stress is greatly diminished These problems may be avoided with several novel coating methods, such as ion beam implantation, magnetron sputtering, multi-arc ion plating, filtered cathodic vacuum arc plating, or electroless Ni-P-PTFE plating, which have been investigated in recent years [17–19] These thin and stable coatings have been found to reduce scale formation during convective and boiling heat transfer and to reduce the adhesion of bacteria Sonic technologies High- and low-frequency sound has successfully been used in heat exchangers for gases to dislodge and weaken particulate deposits, which can subsequently be carried away by the process gas stream In suitable cases, this can be a very cost-effective option As for the application for liquid-side fouling, several laboratory investigations have shown promising effects It is, however, questionable whether sound or vibration generators can sensibly be installed in industrial heat exchangers, and whether their effects will extend over the typically large heat transfer surfaces 193 Magnetic, electronic, or catalytic means When it comes to commercial mitigation of scale formation, one of the most frequently and emotionally discussed topics is devices that claim to reduce scaling by magnetic, electronic, or catalytic means To date, no conclusive scientific proof or theory for the mechanisms that may be responsible for the beneficial effects of such technologies has been found A considerable number of investigations have been reported in the literature; many of them claim some sort of success with the applied technology German Industry Standards (DIN) have been formulated for performance evaluation of physical water conditioners Pilot-plant and laboratory-scale investigations have provided contradicting results For example, references [20], [21], and [22] report that the installation of magnets considerably reduced cooling water fouling, whereas [23] and [24] found no effect of the water conditioner Even the mechanisms of scale inhibition are highly disputed Systematic investigations (e.g., [22]) indicate that the effectiveness of electromagnetic fouling mitigation methods may be limited to a certain window of operation COMPARISON OF CHEMICAL AND MECHANICAL MITIGATION TECHNIQUES In general, heat transfer engineers rely on chemical and mechanical approaches, as the physical systems are still in their early development Figure summarizes the limitations of chemical and mechanical mitigation techniques All systems work best if applied to an initially clean heat exchanger Some of the mechanical systems are less dependent on the type of fouling, while chemical systems are always specific to the composition of the process fluid OFF-LINE CLEANING OF HEAT EXCHANGERS Periodical cleaning of heat exchangers will be necessary, even if the heat exchanger is well designed and the fluid treatment is effective Additionally, conditions in the heat exchanger may deviate from the design conditions due to changes in flow rates and temperatures, plant failures, ingress of air and bacteria, changes in the fluid composition, or upstream corrosion, which all may promote fouling It is, therefore, advantageous Off-line cleaning Fluid cleaning Fluid blasting Ice pigging Chemical Steam soaking Mechanical Shot blasting Figure Categorization of various off-line cleaning systems heat transfer engineering vol 32 nos 34 2011 Projectiles Drilling 194 ă H MULLER-STEINHAGEN ET AL to remove nonprotective deposits soon after the onset of their formation Heat exchangers may be cleaned by various off-line methods as categorized in Figure Intense mechanical and chemical cleaning may remove not only the deposit but also part of the protective oxide layer on the pipe surfaces Under certain circumstances, this may create a corrosion problem On the other hand, regular cleaning removes deposit and avoids flow conditions that promote corrosion due to chemical reaction or stagnant flow For very severe fouling problems, a combination of chemical and mechanical cleaning may be recommended Off-line cleaning is most prevalent in petroleum, minerals, and chemicals processing industries and mainly involves manual or semiautomatic cleaning at predetermined maintenance intervals [5] Although generally effective, these techniques not mitigate the gradual performance degradation (due to fouling) between physical cleaning intervals As a result, most heat exchangers will operate at significantly less than peak efficiency Some of the most promising off-line cleaning systems are briefly discussed next Figure Conco tube cleaner in operation [26] sponge balls Using air pressure or hydro pressure, rubber plugs or metal scrapers can be shot through the tubes Rubber plugs may fail for hard deposits In general, water pressure systems are safer than air pressure systems, due to the compressibility and subsequent rapid expansion of gases Advanced systems, such as the one shown in Figure [26], are rather fast and allow cleaning of up to 15,000 tubes within 24 hours Very dirty and plugged tubes must be cleaned with drills equipped with drill bits, brushes, or bit–brush combinations Chemical Cleaning Methods Chemical cleaning methods have a number of advantages, namely: Blasting and Jetting Techniques High-pressure water or steam blasting up to 1,500 bar can be an effective way of removing unwanted deposits Of those, water jetting is probably the most effective and technologically advanced Delivery may be through multijet sprays or through high-pressure water lances The addition of wetting agents or detergents may improve the washing process If deposits are very tenacious, abrasive particles such as sand may be added to the pressurized water to increase the cleaning efficiency Similarly, air blasting with sand or solid CO2 particles is frequently used Air, steam, and hydro blasting are labor-intensive and keep the exchanger off-line for a considerable time, even though semiautomatic cleaning devices have been developed and are commonly used [5] Blasting may not completely eliminate all deposits and some significant roughness can remain The shell side of tube bundles can only be cleaned completely if the tubes are arranged in-line The particular geometry of twisted tubes provides flow lanes for pressurized water or steam which facilitates cleaning Ice pigging has also been reported as a successful technique to remove moderately adhering deposits, since the shear forces are increased by a factor of 4–5 due to the presence of the ice slurry [25] Such a system can be applied for complex geometries and is reported to have a reduced cleaning downtime Mechanical Cleaning Techniques Several cleaning methods can be used for the inside of straight tubes For example, the on-line fouling mitigation sponge ball system can also be used as a transportable, offline cleaning system, particularly if used with corundum-coated heat transfer engineering • • • • • They are relatively quick Surfaces not experience mechanical damage Chemical solutions reach normally inaccessible areas They are less labor-intensive than mechanical cleaning Cleaning can be performed in situ Problems may arise due to the danger of handling (burns, toxicity), due to elevated application temperatures, due to the costs of cleaning agents, due to the chemical attack on the heat exchanger material (overcleaning, uneven cleaning, corrosion), and due to disposal problems Acids and alkalis must be neutralized, organic materials may be burned, and fluorides must be reacted to inactive solid residues Some of the organic acids, such as citric acid and gluconic acid, are biodegradable Research on the mechanisms of chemical cleaning of heat transfer surfaces is far less developed than research on fouling mechanisms, even though similar approaches may be used Nevertheless, some first modeling has been attempted, assuming that the cleaning process is a reversed fouling process Understanding of the interactions of cleaning and fouling is less advanced in the process industry than in the food industry, to which a series of conferences have been dedicated [27, 28] SUMMARY While significant progress has been made in the mitigation of heat exchanger fouling, the challenge to reduce its impact on heat exchanger performance is still enormous Many mitigation and cleaning techniques that have found their way into regular plant operation have been developed by an empirical trial-anderror approach These antifouling strategies have few or even no links to academic research findings, since industry and academic vol 32 nos 34 2011 ă H MULLER-STEINHAGEN ET AL research institutions have traditionally approached the problem of fouling from different aspects To optimize the effectiveness of mitigation methods, which highly depends on the dominant fouling mechanisms and influential operating conditions, and to develop new approaches for fouling mitigation, closer collaboration between the two communities is essential For the past 15 years, conferences on heat exchanger fouling have been held at bi-yearly intervals to facilitate innovative thinking and to explore new theoretical and practical approaches These conferences have successfully provided a forum for experts from industry, academia, and government research centers from around the world to present their latest research and technological developments in the areas of fouling mitigation and cleaning technologies The meetings in San Luis Obispo (1995), Lucca (1997), Banff (1999), Davos (2001), Santa Fe (2003), Kloster Irsee (2005), Portugal and Tomar (2007) were organized by Engineering Conferences International The 8th conference in this series was organized by the present authors under the auspices of EUROTHERM in Schladming, Austria, in June 2009 In total, 100 participants attended this meeting, presenting 81 papers/posters, which were the highest numbers in any meeting of this series to date The following papers in this special issue of Heat Transfer Engineering have been selected from the contributions to the 2009 Fouling Conference in Schladming after a careful refereeing and revision process The full e-proceedings of the 2009 conference as well as those from the previous conferences from 2003 until 2007 can be obtained free of charge from http://heatexchanger-fouling.com They cover various aspects of heat exchanger fouling, along with updated state-of-the-art fouling mitigation and cleaning strategies Their content is of significant value for researchers, plant operators, equipment manufacturers, chemical suppliers, and heat exchanger cleaning companies This website also contains the actual information about the next conference in this series (Heat Exchanger Fouling and Cleaning IX), which is scheduled for June 5–10, 2011 on the beautiful island of Crete, Greece [5] [6] [7] [8] [9] [10] [11] [12] [13] REFERENCES [14] [1] Heat Exchanger: A Global Strategic Business Report, Global Industry Analysts, Inc., San Jose, CA, 2008 [2] Steinhagen, R., Măuller-Steinhagen, H., and Maani, K., Problems and Costs Due to Heat Exchanger Fouling in New Zealand Industries, Heat Transfer Engineering, vol 14, no 1, pp 19–30, 1992 [3] Măuller-Steinhagen, H., Malayeri, M R., and Watkinson, A P., Heat Exchanger Fouling: Environmental Impacts, Heat Transfer Engineering, vol 30 no 1011, pp 773776, 2009 [4] Măuller-Steinhagen, H., Fouling of Heat Exchanger Surfaces, VDI Heat Atlas, Section C4, VDI Gesellschaft heat transfer engineering [15] 195 Verfahrenstechnik und Chemieingenieurwesen, SpringerVerlag Berlin Heidelberg, 2010 Măuller-Steinhagen, H., Heat Exchanger Fouling Mitigation and Cleaning Technologies, Publico Publications Essen, Germany, 2000 Esawy, M., Malayeri, M R., and Măuller-Steinhagen, H., Crystallization Fouling of Finned Tubes During Pool Boiling: Effect of Fin Density, J Heat and Mass Transfer, vol 46, pp 1167–1176, 2010 Solano, J P., Garc´ıa, A., Vicente, P G., and Viedma, A., Performance Evaluation of a Zero-Fouling Reciprocating Scraped-Surface Heat Exchanger, Heat Transfer Engineering, vol 32, no 3–4, pp 2009 Klaren, D G., and de Boer, E F., Achievements and Potential of Self-Cleaning Heat Exchangers Using Untreated Natural Seawater as a Coolant, Proceedings of the ECI Conference on Heat Exchanger Fouling and Cleaning–VII, eds H Măuller-Steinhagen, M R Malayeri, and A P Watkinson, ECI Symposium Series, vol RP5, Tomar, Portugal, pp 262–274, 2007 Wiehe, I A, The Oil Compatibility Model and Crude Oil Compatibility, Energy & Fuels, vol 14, pp 56–59, 2000 Evangelidou, M., Crystallization Fouling of Structured Tubes During Pool Boiling Heat Transfer, Diploma thesis, University of Stuttgart, Stuttgart, Germany, 2010 Waite, T D., and Fagan, J R., Summary of Biofouling Control Alternatives, in Condenser Biofouling Control, ed J Garey, Ann Arbor Science, Ann Arbor, MI, pp 441–462, 1980 Ferreira, C., Sim˜oes, M., Pereira, M C., Bastos, M M S M., Nunes, O C., Coelho, M., and Melo, L F., Control of Biofouling of Industrial Surfaces Using Microparticles Carrying a Biocide, Proceedings of EUROTHERM International Conference on Heat Exchanger Fouling and Cleaning VIII–2009, eds H Măuller-Steinhagen, M R Malayeri, and A P Watkinson, Schladming, Austria, June 14–19, pp 378–383, 2009 E-proceedings of conferences on heat exchanger fouling and cleaning, http://heatexchanger-fouling.com/ proceedings.htm EU Project: A Method for On-Line Cleaning of Heat Exchangers to Significantly Increase Energy Efficiency in the Oil, Gas, Power & Chemical Process Sectors (CleanEx), 2009–2013, http://www.pro-cleanex.org Pouponnot, F., and Kreuger, A W., Heat Exchanger Tube Inserts: An Update in New Applications With Trouble Shooting Aspects in Crude Units, Residue Service, Reboilers, U-Tubes “SPIRELF, TURBOTAL, and FIXOTAL Systems” Application Examples in Chemical Plants and Refineries, Proceedings of 6th International Conference on Heat Exchanger Fouling and CleaningChallenges and Opportunities, eds H Măuller-Steinhagen, M R Malayeri and A P Watkinson, ECI Symposium Series, vol RP2, pp 221–230, 2005 vol 32 nos 3–4 2011 196 ¨ H MULLER-STEINHAGEN ET AL [16] Ritchie, J M., Droegemueller, P., and Simmons, M J H., hiTRAN Wire Matrix Inserts in Fouling Applications, Heat Transfer Engineering, vol 30, no 10–11, pp 876884, 2009 [17] Bornhorst, A., Zhao, Q., and Măuller-Steinhagen, H., Reduction of Scale Formation by Ion Implantation and Magnetron Sputtering on Heat Transfer Surfaces, Heat Transfer Engineering, vol 20, no 2, pp 614, 1999 [18] Măuller-Steinhagen, H., Zhao, Q., Helalizadeh, A., and Ren, X G., The Effect of Surface Properties on CaSO4 Scale Formation During Convective Heat Transfer and Subcooled Flow Boiling, Canadian Journal of Chemical Enginneering, vol 78, pp 1220, 2000 [19] Făorster, M., Augustin, W., and Bohnet, M., Influence of the Adhesion Force Crystal/Heat Exchanger Surface on Fouling Mitigation, Chemical Engineering and Processing, vol 38, pp 449–461, 1999 [20] Parkinson, G., and Price, W., Getting the Most Out of Cooling Water, Chemical Engineering, vol 91, no 1, pp 22–25, 1984 [21] Donaldson, J., and Grimes, S., Lifting the Scale From Our Pipes, New Scientist, vol 18, pp 43–46, 1988 [22] Cho, Y I., Lee, S H., Kim, W., and Suh, S., Physical Water Treatment for the Mitigation of Mineral Fouling in Cooling-Tower Water Applications, Proceedings of 5th International Conference on Heat Exchanger Fouling and Cleaning—Fundamentals and Applications, eds A P Watkinson, H Măuller-Steinhagen, and M R Malayeri, ECI Symposium Series, vol RP1, pp 107–114, 2003 [23] Hasson, D., and Bramson, D., Effectiveness of Magnetic Water Treatment in Suppressing CaCO3 Scale Deposition, Industrial & Engineering Chemistry Process Design and Development, vol 24, pp 588592, 1985 [24] Săohnel, O., and Mullin, J., Some Comments on the Influence of a Magnetic Field on Crystalline Scale Formation, Chemistry and Industry, vol 6, pp 356–358, 1988 [25] Ainslie, E A., Quarini, G L., Ash, D G., Deans, T J., Herbert, M., and Rhys, T D L., Heat Exchanger Cleaning Using Ice Pigging, Proceedings of 6th International Conference on Heat Exchanger Fouling and CleaningChallenges and Opportunities, eds H MăullerSteinhagen, M R Malayeri and A P Watkinson, Schladming, Austria, June 14–19, pp 433–438, 2009 [26] Saxon, G E., Jr., and Putman, R E., The Practical Application and Innovation of Cleaning Technology for Heat Exchangers, Proceedings of 5th International Conference heat transfer engineering on Heat Exchanger Fouling and Cleaning—Fundamentals and Applications, eds A P Watkinson, H MăullerSteinhagen, and M R Malayeri, ECI Symposium Series, vol RP1, pp 294–301, 2003 [27] Fryer, P., ed., Fouling and Cleaning in Food Processing, Special Topic Issue, Food and Bioproducts Processing, vol 77, issue 2, p 71, 1999 [28] Wilson, D I., and Chew, Y M J., eds., Fouling and Cleaning in Food Processing, Proceedings of Conference at Jesus College, University of Cambridge, 2224 March 2010 ă H Muller-Steinhagen is the director of the Institute of Technical Thermodynamics of the German Aerospace Centre and the director of the Institute for Thermodynamics and Thermal Engineering of the University of Stuttgart His research work covers a wide range of topics related to heat and mass transfer, multiphase flow, fuel cells, solar technology, and process thermodynamics He is the author of more than 550 articles, and was awarded the 1992 and 1993 TMS Bauxite Processing Awards, the 1994 Light Metals Award, the Beilby Medal and Prize, the UK Heat Transfer Society Mike Akrill Trophy, the Best Paper 2000 in the Canadian Journal of Chemical Engineering, and the 2008 AIChE D Q Kern Award He is a fellow of the Royal Academy of Engineering, president of EUROTHERM, member of the Executive Boards of EUREC and ICHMT, member of the Innovation Council of the Prime Minister of Baden-Wăurttemberg, chairman of the Advisory Board of the DESERTEC Industrial Initiative, and associate editor of Heat Transfer Engineering M R Malayeri is the head of the heat exchanger fouling and mitigation research group at the Institute of Thermodynamics and Thermal Engineering (ITW), University of Stuttgart, Germany He graduated from the Department of Chemical Engineering, Amirkabir University of Technology, Iran He received his Ph.D from the University of Surrey, UK His research interests include enhanced heat transfer, heat exchanger fouling and mitigation, multiphase flows, and numerical modeling A P Watkinson is a professor of chemical engineering in the Department of Chemical and Biological Engineering at the University of British Columbia, Vancouver, Canada He is involved in research on fouling of organic fluids, asphaltene precipitation, coke formation, scaling in aqueous systems, and on the pyrolysis, gasification, and combustion of biomass fuels vol 32 nos 3–4 2011 Heat Transfer Engineering, 32(3–4):197–215, 2011 Copyright C Taylor and Francis Group, LLC ISSN: 0145-7632 print / 1521-0537 online DOI: 10.1080/01457632.2010.495579 Fouling in Crude Oil Preheat Trains: A Systematic Solution to an Old Problem S MACCHIETTO,1 G F HEWITT,1 F COLETTI,1 B D CRITTENDEN,2 D R DUGWELL,1 A GALINDO,1 G JACKSON,1 R KANDIYOTI,1 S G KAZARIAN,1 P F LUCKHAM,1 O K MATAR,1 M MILLAN-AGORIO,1 ¨ E A MULLER, W PATERSON,3 S J PUGH,4 S M RICHARDSON,1 and D I WILSON Department of Chemical Engineering and Chemical Technology, South Kensington Campus, Imperial College London, United Kingdom Department of Chemical Engineering, University of Bath, United Kingdom Department of Chemical Engineering and Biotechnology, University of Cambridge, United Kingdom IHS ESDU, London, United Kingdom A major cause of refinery energy inefficiency is fouling in preheat trains This has been a most challenging problem for decades, due to limited fundamental understanding of its causes, deposition mechanisms, deposit composition, and impacts on design/operations Current heat exchanger design methodologies mostly just allow for fouling, rather than fundamentally preventing it To address this problem in a systematic way, a large-scale interdisciplinary research project, CROF (crude oil fouling), brought together leading experts from the University of Bath, University of Cambridge, and Imperial College London and, through IHS ESDU, industry The research, coordinated in eight subprojects blending theory, experiments, and modeling work, tackles fouling issues across all scales, from molecular to the process unit to the overall heat exchanger network, in an integrated way To make the outcomes of the project relevant and transferable to industry, the research team is working closely with experts from many world leading oil companies The systematic approach of the CROF project is presented Individual subprojects are outlined, together with how they work together Initial results are presented, indicating that a quantum progress can be achieved from such a fundamental, integrated approach Some preliminary indications with respect to impact on industrial practice are discussed INTRODUCTION About 6% of the energy content of each crude barrel processed in an oil refinery is used in the refinery itself With The authors gratefully acknowledge EPSRC (grants EP/D503051/1, EP/D506131/1, EP/D50306X/1) for financial support of the project and the industrial partners of the CROF consortium for valuable inputs, data, materials, and discussions Special thanks go to the CROF researchers, M Abubakar, E Al-Muhareb, M Behrouzi, C Berrueco, J Chew, S Dong, T Gu, C Hale, A Haslam, E Ishiyama, J Jover, T J Morgan, J K Pental, K Rostani, K C Sahu, H Shumba, D Sileri, F H Tay, S Venditti, and A Young Address correspondence to Prof Sandro Macchietto, Department of Chemical Engineering, Roderic Hill Building, South Kensington Campus, Imperial College London, London SW7 2AZ, United Kingdom E-mail: s.macchietto@imperial.ac.uk a global production of about 82–85 million barrels per day (bbl/day), this is roughly equivalent to the entire production of Exxon or Shell to operate the world’s 720 refineries Crude oil distillation, where the incoming crude is first heated up and split into its main fractions, accounts for a large fraction of this energy Thus, strenuous attempts are made to recover as much as possible of the energy from the product streams of the crude distillation column (and other refinery units) by means of a network of heat exchangers, often called the “preheat train” (PHT) A typical crude preheat train is illustrated in Figure Unfortunately, crude oil contains a variety of substances, which tend to deposit as fouling layers in the heat exchangers when heated The material deposited ranges from gel-like to solid-like and may change its properties with time The fouling deposit growth results over time in decreased energy recovery 197 T GU ET AL 343 liquid to flow and thereby reducing the extent of the curvature zone The difference between the two apparatuses and the two nozzles are compared In Figure 5c the characteristic mass flow rate–nozzle clearance profile is shown, where the discharge mass flow rate was divided by m∞ to yield a dimensionless profile For the same clearance, more liquid flowed through gauge 2, due to a higher pressure driving force, a smaller nozzle rim, and a larger hydraulic diameter The change in the nozzle geometry will also affect the mass flow rate, and this needs to be investigated separately The figure shows that the working range of the gauge for both nozzles lies in the range 0.06 < h/dt < 0.30 Discharge Coefficient The effect of the geometry on the nozzle performance can be quantified via Cd The effect of nozzle geometry on the asymptotic nozzle discharge coefficient, Cd,∞ , defined as the average value of Cd at large clearances (e.g., h/dt ≥ 0.6), is shown in Figure for different values of Reannulus The 45◦ and 30◦ nozzles behaved quite differently; Cd,∞ decreased slightly with increasing Reannulus for the 45◦ nozzle (apparatus 1), suggesting that the nozzle behaves less ideally at higher annular flow rates, whereas for the 30◦ nozzle the opposite trend was observed The latter has a smaller angle, which is likely to reduce the amount of recirculation within the nozzle at low gauging flow rates The apparatus nozzle rim was also narrower, which is expected to reduce (a) the frictional losses underneath the rim and (b) the disturbance to the flow in the annulus when the nozzle is far from the gauging surface Both of these factors are hypothesized to contribute to lowering the hydraulic losses across the nozzle, giving larger values of Cd,∞ For the lowest Reannulus values investigated, i.e., 90 and 250 for the 30◦ and 45◦ nozzle, respectively, both nozzles gave a common Cd,∞ value of around 0.68 Heated Surfaces Figure Effect of Reannulus on experimental FDG mass profile Regions marked: (i) curvature zone; (ii) incremental zone; (iii) asymptotic zone (a) Apparatus 1, α = 45◦ , H = 405 mm (b) Apparatus 2, α = 30◦ , H = 350 mm (c) Data from (a) solid symbols and (b) open symbols plotted as dimensionless mass flow rate–clearance profiles The diameter of the inner rod of apparatus is almost twice that to apparatus (21 mm vs 12 mm), while the nozzle throat sizes are identical (dt = mm) This means that with respect to the liquid near the nozzle throat, the geometry of apparatus is more similar to that of a parallel plate when the nozzle is very close to the surface (h/dt < 0.04) For apparatus 2, for a given clearance the surface curvature is larger, allowing more heat transfer engineering The effect of a heated surface on the discharge mass profile, studied using apparatus 2, proved to be negligible (Figure 7) Wall temperatures were maintained at 20◦ C, 50◦ C, 80◦ C, and 110◦ C for Reannulus values of 1,700, 3,000, and 10,000, representing flows in the laminar, transitional, and turbulent regime The corresponding heat fluxes lay in the range 7–25 kW/m2 A slight increase in mass flow rate (6% or less) was noticed for the higher wall temperatures when the nozzle was located near the heated surface, due to the lower viscosity of the liquid in this region Little difference was observed beyond 220 µm from the heated surface (h/dt = 0.22) These experiments were occasionally affected by bubble formation promoted by the hot surface, which affects the performance of the gauge For virtually all cases investigated, the mass flow rate through the gauge was only a fraction (less than or equal to 15%) of the vol 32 nos 3–4 2011 344 T GU ET AL 1.0 Table 1) 0.9 h/dt Reannulus Cd,exp Cd,sim Difference (%) Retube 0.10 0.14 0.20 0.10 0.14 0.20 0.10 0.14 0.20 0.10 0.14 0.20 560 0.266 0.364 0.481 0.261 0.353 0.476 0.264 0.351 0.472 0.258 0.342 0.465 0.240 0.347 0.473 0.240 0.347 0.472 0.240 0.362 0.472 0.240 0.350 0.472 10% 5% 2% 8% 2% 1% 9% −3% 0% 7% −1% −2% 542 740 973 532 718 964 538 714 956 527 696 941 0.8 Cd,∞ 0.7 0.6 0.5 0.4 Summary of CFD simulation of FDG in annular flow (apparatus 300 190 90 0.3 5000 10000 15000 Re annulus Figure Effect of Reannulus on asymptotic nozzle discharge coefficient Cd,∞ for gauge and 2, respectively Symbols: solid triangles, gauge (Figure 1); open circles, gauge total flow through the annulus The exception is that for Reannulus = 90 (apparatus 1), where the maximum flow rate through the gauge approached 30% of the flow through the annulus inlet This is unlikely to be desirable in a monitoring experiment CFD Simulation experimental and computed values of Cd agreed to 10% or better for all cases The simulated Cd values lie for most cases within, or close to, the experimental Cd error bars For the range of Reannulus investigated, the agreement appears better for the lower Reannulus flows and for larger clearances (h/dt > 0.10) The associated flow velocity distributions in the tube (ydirection) are presented in Figure The highest velocity occurs within the throat of the nozzle with expansion further along the tube Asymmetrical flow in the tube is noticed when the gauge is close to the surface, at h/dt = 0.10, where the momentum from the annular flow is more influential, in agreement with results for square ducts reported by Gu et al [8] The flow in the nozzle Table summarizes the results for a series of CFD simulations of FDG in (laminar) annular flow, for apparatus The Reannulus = 10000 m [g/s] Reannulus = 4000 Reannulus = 1700 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 h/d t Figure Apparatus Effect of wall temperature on discharge mass flow rate Bulk flow temperature 20◦ C Experimental conditions as in Figure 5b Symbols: squares, 20◦ C; circles, 50◦ C; triangles, 80◦ C; crosses, 110◦ C heat transfer engineering Figure Tube (y-wise) velocity component at the y–z plane of symmetry for Reannulus of 560 and 90 Lighter shading indicates higher velocity The simulation shows that the maximum y-wise velocity occurs at the nozzle throat vol 32 nos 3–4 2011 T GU ET AL 345 Figure Shear stresses imposed by the gauging flow on the inner surface of the annulus (i.e., at y = 0), directly underneath the gauge, as illustrated in the inset (a) Stress along the inner surface of the annulus in the z’ direction, τyz’ (b) Shear stress, τya , imposed by the gauging flow on the inner surface of the annulus at different positions along the arc, a and tube becomes more symmetrical when the gauge is further away from the surface, e.g., h/dt ≥ 0.14 Values of shear stresses acting on the surface being gauged could be extracted from the CFD simulations The surface is heat transfer engineering curved and hence the stresses acting on the surface along the annulus in the z-direction, and across the annulus in the x-direction, are different Shear stress values are presented in Figure 9a along the inner surface of the annulus in the z-direction, and in 9b along vol 32 nos 3–4 2011 346 T GU ET AL the arc length of the inner surface in the x-direction, for the simulation case h/dt = 0.10, 0.14, and 0.20, for Reannnulus 550 and 90, respectively The z-wise shear stress (τyz’ ) is approximately zero at the centerline of the tube (z’ = 0), reaches a maximum beneath the nozzle lip (z’ = 0.5 to 1.0 mm) and approaches zero asymptotically for z’ > mm The magnitude of the zwise shear stress at a given value of z’ decreases as the gauge is further away from the surface The magnitude of these shear stresses is comparable with velocities used in cleaning-in-place operations [9] As the flow rate through the annulus increases, so does the shear stress underneath the gauging nozzle, but the dominant shear stress is that caused by the proximity of the gauge to the surface When the gauge is close to the surface, i.e., at h/dt = 0.10, the shear stress is slightly higher upstream from the gauge, which is an effect of the flow in the annulus The gap between the nozzle and the surface increases along the arc (Figure 9b) The highest values of shear stress (τza ) occur within the throat and underneath the nozzle rim The magnitude and shape of τza are similar to τyz’ when the gauge is close to the surface, for h/dt of 0.10 and 0.14 However, when the nozzle is further away, at a clearance of h/dt = 0.20, the shear stress τza exerted on a curved surface by the nozzle is less pronounced, which is why the shear stress decays from the centreline of the tube, rather than displaying peaks underneath the nozzle rim Whey Protein Fouling Solutions of whey protein concentrate (WPC) at wt% aqueous were passed through apparatus and allowed to run without interruption for 12 h The bulk temperature of the solution was maintained at 54◦ C, the wall temperature at 95◦ C, and a flow rate through the annulus of 0.22 m/s, corresponding to Reannulus = 7,500 After 10 h an increase in fouling resistance (Rf ) was Figure 10 Change in fouling resistance with time Inset shows the image of the ring left by the nozzle heat transfer engineering noticed, shown in Figure 10, which indicated the formation of a fouling layer The gauge measured a thickness of 200 µm The layer was soft and gel-like, and was removed from the surface by the gauging flow (see inset in Figure 10) Lumps of the deposit were observed in the discharge flow from the tube during the gauging experiment when the nozzle was close to the deposit surface The layer dried quickly once the system had been drained The nozzle left a ring of diameter mm, in good agreement with the CFD predictions, suggesting that the dominant force of removal was caused by adhesive failure To work out the shear stress and hence the forces of removal, a CFD model needs to be implemented incorporating the effect of heat transfer from the surface of the wall CONCLUSIONS Fluid dynamic gauging has been successfully applied to laminar, transitional, and turbulent annular flows for different annular geometries, and for a heated surface, at various wall temperatures The practical working range of the gauge proved to be independent of the surface being heated The gauge measured the thickness of a soft whey protein fouling layer in situ, and because of the proximity of the gauge to the surface, the soft deposit was removed by the forces of the gauging flow near the surface CFD simulation demonstrated that the highest shear stresses are located under the rim of the nozzle, which was confirmed by the fouling experiment To be able to predict the exact forces of removal of the deposit, further fouling experiments need to be performed NOMENCLATURE a Cd Cd,∞ CFD d dt Dh FDG FEM g h H k L’ l l’ m n N p arc length along annulus cross section, m discharge coefficient, dimensionless asymptotic discharge coefficient, dimensionless computational fluid dynamics inner diameter of dynamic gauging tube, m nozzle throat diameter, m hydraulic diameter of the annulus, m fluid dynamic gauging finite element method acceleration due to gravity, m/s2 clearance between nozzle tip and gauging surface, m hydrostatic head providing pressure driving force for gauging flow, m wall thickness of gauging tube, m CFD model duct length, m length of siphon tube, m CFD model tube length, m tube discharge mass flow rate, kg/s normal vector number pressure, Pa vol 32 nos 3–4 2011 T GU ET AL ps R Rf Re r s TU-BS u v v v¯ w w ¯ x, y, z static pressure, Pa gauging tube radius, m fouling resistance, m2K/W Reynolds number, dimensionless radial coordinate of the gauging nozzle, m width of nozzle rim, m Technisch Universităat Braunschweig x-wise velocity, m/s velocity vector y-wise velocity, m/s mean y-wise velocity, m/s z-wise velocity, m/s mean z-wise velocity, m/s coordinates Greek Symbols α λ µ ρ τyz’ τza nozzle inner angle, degrees length of nozzle exit, m dynamic viscosity, Pa-s density, kg/m3 shear stress on the y-plane in the z-direction, Pa shear stress on the z-plane in the a-direction, Pa 347 [5] Bennett, C A., Kistler, R S., Nangia, K., Al-Ghawas, W., Al-Hajji, N., and Al-Jemaz, A., Observation of an Isokinetic Temperature and Compensation Effect for High Temperature Crude Oil Fouling, ECI Conference on Heat Exchanger Fouling and Cleaning, Tomar, Portugal, vol RP5, 2007 [6] Baker, D K., Vliet, G C., and Lawler, D F., Experimental Apparatus to Investigate Calcium Carbonate Scale Growth Rates, Proc International Conference on Mitigation of Heat Exchanger Fouling and its Economic and Environmental Implications, Banff, Alberta, Canada, 1999 [7] Tritton, D J., Physical Fluid Dynamics, 2nd ed., Oxford University Press, Oxford, UK, pp 58–59, 1988 [8] Gu, T., Chew, Y M J., Paterson, W R., and Wilson, D I., Experimental and CFD Studies of Fluid Dynamic Gauging in Duct Flows, Chemical Engineering Science, vol 64, issue 2, pp 219–227, 2009 [9] Timperley, D A., Cleaning in Place, Journal of Dairy Technology, vol 42, issue 2, pp 32–33, 1989 T Gu is a PhD student at the Department of Chemical Engineering and Biotechnology, University of Cambridge, UK Her research interests focus on developing the technique of fluid dynamic gauging to monitor fouling layer thicknesses Subscripts e eff exp max sim ∞ elements effective experimental maximum value simulation asymptotic REFERENCES [1] Tuladhar, T R., Macleod, N., Paterson, W R., and Wilson, D I., Development of a Novel Non-Contact Proximity Gauge for Thickness Measurement of Soft Deposits and Its Application in Fouling Studies, Canadian Journal of Chemical Engineering, vol 78, pp 925–947, 2000 [2] Tuladhar, T R., Paterson, W R., and Wilson, D I., Dynamic Gauging in Duct Flows, Canadian Journal of Chemical Engineering, vol 81, pp 279–284, 2003 [3] Hooper, R J., Liu, W., Fryer, P J., Paterson, W R., Wilson, D I., and Zhang, Z., Comparative Studies of Fluid Dynamic Gauging and a Micromanipulation Probe for Strength Measurements, Food & Bioproducts Processing, vol 84, pp 353–358, 2006 [4] Gu, T., Chew, Y M J., Paterson, W R., and Wilson, D I., Experimental and CFD Studies of Fluid Dynamic Gauging in Annular Flows, AIChE Journal, vol 55, no 8, pp 1937–1947, 2009 heat transfer engineering F Albert is a doctoral candidate at the Institute for Chemical and Thermal Process Engineering, Technische Universităat Braunschweig, Germany He worked on heat transfer characterization of highly viscous fluid flow in circular ducts for his diploma thesis and received his Dipl.-Ing Degree in 2005 from the Technische Universităat Braunschweig His research is focused on fouling in tube heat exchangers and on heat transfer performance of viscous fluid flow W Augustin is an Academic Director at the Institute for Chemical and Thermal Process Engineering, Technische Universităat Braunschweig, Germany He received his Dr.-Ing in 1992 from Technische Universităat Braunschweig His main research interests are heat and mass transfer, fouling, viscous fluid flow and surface interactions in the micro scale level Y M J Chew is a Royal Academy of Engineering/EPSRC Research Fellow based at the Department of Chemical Engineering and Biotechnology at Cambridge, UK He holds a PhD from the University of Cambridge His core research activity is the study of the formation and removal of surface layers, combining experimental investigations with numerical modeling He is due to take up a lectureship at the University of Bath in 2010 vol 32 nos 3–4 2011 348 T GU ET AL W R Paterson has retired from his Senior Lectureship in Chemical Engineering at Cambridge, and is a former Chartered Engineer His research interests include process simulation & synthesis, and fouling & cleaning He holds a PhD from the University of Edinburgh S Scholl is a Professor of Chemical and Thermal Process Engineering at the Technische Universităat Braunschweig, Germany He received his Dr.-Ing degree in 1991 from the Technical University of Munich After eleven years with BASF SE, Ludwigshafen, Germany he joined TU Braunschweig in 2002 His main research areas are heat and mass transfer with an emphasis on fouling and viscous systems, energy efficiency, separation processes in biotechnological applications, as well as micro process engineering K Wang completed his M.Eng in Chemical Engineering at Cambridge in June 2009 His final year research project was on dynamic gauging in annular flows He is now working as a trader in Citigroup D I Wilson is a Reader in Chemical Engineering at Cambridge He has worked in the area of fouling and cleaning of heat transfer systems since 1988 His research interests include rheology, paste processing and food engineering He is a chartered engineer and holds a PhD from the University of British Columbia I Sheikh completed his M.Eng in Chemical Engineering at Cambridge in June 2009 His final year research project was on dynamic gauging in annular flows He is now working with ExxonMobil Engineering (EMEE) heat transfer engineering vol 32 nos 3–4 2011 Heat Transfer Engineering, 32(3–4):349–357, 2011 Copyright C Taylor and Francis Group, LLC ISSN: 0145-7632 print / 1521-0537 online DOI: 10.1080/01457632.2010.495668 Detection of Fouling in a Cross-Flow Heat Exchanger Using Wavelets ´ HELGA INGIMUNDARDOTTIR and SYLVAIN LALOT2,3 University of Iceland, Reykjavik, Iceland Universit´e Lille Nord de France, Lille, France Universit´e de Valenciennes, et du Hainaut-Cambresis, LME, Valenciennes, France Detection of fouling in a heat exchanger experiencing perfect steady-state conditions is not very difficult But the challenge is to detect fouling when all inputs (inlet temperature of the fluids and the mass flow rates) are simultaneously varying In this paper it has been considered that the mass flow rates can vary in a ratio of 2, and that the inlet temperatures can vary by about ±20% This first approach is dedicated to show the feasibility of using the wavelet transform It has been considered that getting simulated data is the best way In fact, it is then possible to introduce an arbitrary fouling factor Thus, in the first part of the paper the model of the heat exchanger is presented It is developed using Simulink The validation is carried out on an electrical heater, for which it is possible to find an analytical solution for transient states It is also shown that steady states are accurately computed over a large range of the number of transfer units and heat capacity rate ratios Then a brief overview of the wavelet transform is given Then basic examples show that the wavelet transform can help to find the trend of time series It is then applied to the analysis of the “wavelet-transformed” effectiveness of the heat exchanger This analysis is carried out on a sliding observation window (to be able to detect fouling on-line) It is shown that fouling is detected at a very early stage INTRODUCTION Studies concerning fouling can be divided into three complementary domains: the principles of fouling (chemistry and flow conditions), e.g., [1] or [2], the mitigation of fouling (design phase, water treatment, surface treatment), e.g., [3] or [4], and fouling monitoring (model based techniques, sensors), e.g., [5]–[11] The present study belongs to the last category: online fouling monitoring It is not a model based technique such as the one presented in [12], [13], or [14] It is only based on the evolution analysis of the effectiveness of the heat exchanger Special thanks are due to Mr A Chamroo, Mr G Merc`ere, and Mr T Poinot, from the Laboratoire d’Automatique et d’Informatique Industrielle at the Universit´e de Poitiers, for their valuable advice for the Simulink implementation of differential equations This work could not have been carried out without the financial help of the European Union (Erasmus grant for one author), the French/Icelandic “Jules Verne” program (18990VL), and the DESURENEIR project supported by the CNRS All these are gratefully acknowledged Address correspondence to Professor Sylvain Lalot, Laboratoire de Mecanique et D’energetique (LME), Universit´e de Valenciennes, et du HainautCambresis, le Mont Houy, 59313 Valenciennes, Cedex 9, France E-mail: sylvain.lalot@univ-valenciennes.fr But it is well known that in transient states, the effectiveness of a heat exchanger is not defined That is why it has been necessary first to find a tool able to determine a sliding average “steady state” of the heat exchanger After a first test of the simple “slope method” presented in [15], it has been chosen to try a more complex tool, a wavelet transformation, as it is used in many fields [16–21] A wavelet transform is basically a signal processing tool [22, 23] and is introduced in this paper after the presentation of the heat exchanger model THE HEAT EXCHANGER To be able to evaluate the efficiency of the method proposed here, we chose to work using simulated data In this case, it is possible to introduce an arbitrary time variation of the fouling factor; and consequently to know what this factor is when fouling is detected The next subsection is dedicated to the description of the heat exchanger 349 350 Figure Discretization of the heat exchanger DESCRIPTION OF THE HEAT EXCHANGER Figure Simulink block for the hot fluid (one cell) We chose to model a cross-flow heat exchanger, having both fluids unmixed The fluids are separated by a plate; the other side of the fluid channels is perfectly insulated To model the exchanger using Simulink, it is necessary to divide it into “cells” as shown in Figure In each cell three energy balance equations can be written, one for the hot fluid: Mh ch d dt Th,i,J + Th,i+1,J + αI,J Ah,I,J Tp,I,J − ˙ h,J ch (Th,i,J − Th,i+1,J ) =m Th,i,J + Th,i+1,J (1) and one for the cold fluid: Mc cc d dt Tc,I,j + Tc,I,j+1 + βI,J Ac,I,J Tp,I,J − ˙ c,I cc Tc,I,j − Tc,I,j+1 =m Tc,I,j + Tc,I,j+1 (2) Figure Simulink block for the cold fluid (one cell) and one for the separating plate: Mp cp d Th,i,J + Th,i+1,J Tp,I,J = −αI,J Ah,I,J Tp,I,J − dt − βI,J Ac,I,J Tp,I,J − Tc,I,j + Tc,I,j+1 (3) ˙ h,J ch and nutI,J = Introducing, NUTI,J = αI,J Ah,I,J /m ˙ c,I cc , which are similar to number of transfer units, βI,J Ac,I,J /m ˙ h,J , which is the residence time in one cell, and τh,I,J = Mh /m with γI,J = αI,J Ah,I,J /Mp cp and δI,J = βI,J Ac,I,J /Mp cp , which are inverses of response times, it is possible to build the respective blocks in Simulink Figures 2–4 show the blocks included in one cell for the hot side, the cold side, and the wall, respectively To so, it has to be noted that it is better to use integrators than to use derivatives Using these blocks, one cell is defined as described in Figure heat transfer engineering Figure Simulink block for the separating plate (one cell) vol 32 nos 3–4 2011 ´ H INGIMUNDARDOTTIR AND S LALOT 351 Figure Comparison of the results obtained using the Simulink model and analytical results (transient state) ˙ h ch , τ = Mh /m ˙ h , and γ = hA/Mp cp where NUT = hA/m When applying a step function for the heat flux, the solution is given by Eq (5) NUT/τ × − exp (− (NUT/τ + γ) t) NUT/τ + γ NUT exp (−NUT) − τ ⎧⎛ ⎞⎫ NUT/τ ⎪ ⎪ ⎪ ⎪ exp (− (NUT/τ + γ) t) × ⎪ ⎜ ⎟⎪ ⎪ ⎪ NUT/τ+γ ⎪ ⎟⎪ ⎪⎜ ⎪ ⎪ ⎪ ⎪ ⎪ ⎝ ⎠ ⎪ ⎪ γ ⎬ ⎨ + NUT/τ + γ × ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ∗ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ exp (−γt) ⎪ ⎪ √ ⎩ ⎭ ×I0 NUTγt Th = Figure One cell of the cross-flow heat exchanger Determination of the Cell Number It is then necessary to determine how many cells are to be linked to compute an accurate solution To so, it has been decided to model an electrical heater for which an analytical solution can be found The geometry is quite simple; a thin plate is heated by the Joule effect, so that a constant heat flux is generated all along the heater; its length is m It is considered that the temperature of the plate is homogeneous in the direction perpendicular to the flow In this case, the exact solution in the Laplace space is given by Eq (4) for the transfer function Th = NUT/τ s (s + NUT/τ + γ) × − exp (−τs) exp −NUT s s+γ (4) heat transfer engineering (5) (t − τ) It is then possible to compare this analytical solution to the solution obtained with a Simulink model Figure shows that choosing 20 cells leads to a very accurate solution Note that to get this solution, the Simulink block of the plate has been slightly modified to take account of the constant heat flux It can be noted that the study of a heat exchanger having a constant plate temperature leads to the same conclusion: that 20 cells are sufficient in the flow direction It is also important to note that the steady states should be reached for whatever the temperature level and the mass flow rates are This is done comparing the effectiveness computed in a steady state for a large range of number of transfer units (NTU) values and a large range of heat capacity rate ratios vol 32 nos 3–4 2011 352 Figure Filter values (lowpass and highpass filters) Figure Comparison of the results obtained using the Simulink model and analytical results (steady states) Figure shows the comparison of analytical results [24] to results obtained using 20 × 20 cell groups (or 1,200 blocks) in Simulink THE WAVELET TRANSFORM A Short Introduction When analyzing nonstationary signals it is necessary to take careful consideration of the time and frequency domains and what compromises should be made between the two The standard Fourier transform is only localized in frequency; the shorttime Fourier transform is limited by its fixed window length On the contrary, wavelets are localized both in time and in frequency, but it is possible to control the localization Therefore, it is necessary to get a better tool In fact, wavelets separate data into frequency components and analyze each component with a resolution matched to its scale The more the wavelet is similar to the signal components, the larger is the corresponding wavelet coefficient Wavelet transform can be beneficial for feature extraction, e.g., fingerprints recognition (wavelet-based fingerprint image retrieval [25]), or for diagnosis, e.g., [26–29] Wavelets are functions that satisfy certain requirements; e.g., they should integrate to zero, waving above and below the x axis; be well localized; and there are other requirements that are technical to insure quick and easy calculations for the direct and inverse wavelet transform Wavelets are structured with a basis in discrete or continuous time, and they allow different time versus frequency resolution trade-offs Wavelet transforms are built on orthonormal bases of the form φnk (x) = {2n/2 φ(2n x − k) : (n, k) ∈ Z2 } Thus, each element of the basis is a translated and dilated version of a single wavelet heat transfer engineering Usually, φnk (x) are called daughter wavelets of the mother wavelet φ When the mother wavelet is composed of two parts (the lowpass part η and the highpass part µ), any arbitrary signal can be expressed as follows: N N νn (k) η(2n t − k) + f (t) = k=1 λn (k) µ(2n t − k) k=1 where n fixes the level of approximation available, N fixes the level of approximation used when reconstructing the signal, νn are the lowpass coefficients, and λn are the highpass coefficients The latter are computed using the inner product between the original signal and the lowpass part η of the wavelet and the highpass part µ of the wavelet, respectively For the present study, the Daubechies wavelet basis has been chosen [30] It is compactly supported and can be designed with as much smoothness as desired It relies on the iteration of the discrete filter bank that converges to a continuous time wavelet basis Daubechies wavelets are designed so that they have the minimum length of support for a given number of vanishing points Note that when the support is short, the wavelet does not interact very much with a singularity In this case, two parameters are necessary The first parameter, dim, is the length of the support; the second parameter, scale, fixes the approximation level Figure shows the lowpass and highpass parts for 16-sample-long support For a longer introduction to wavelets, see http://users rowan.edu/∼polikar/WAVELETS/WTtutorial.html Illustrative Example It is interesting to show the influence of the two main parameters used in the wavelet transform Figure shows, for two values of the parameter dim and for four values of the parameter scale, the result of the approximation process on an arbitrary signal It can be seen that the couple (dim, scale) = (2, 6) would lead to easy detection for this set of data These values are chosen as starting points for fouling detection vol 32 nos 3–4 2011 ´ H INGIMUNDARDOTTIR AND S LALOT 353 Figure 10 Partial view of the inputs (set number 80) Figure Illustration of the wavelet transform process (arbitrary units) RESULTS In a first step, 200 sets of data have been generated The first half are for a clean exchanger; the second half are for a heat exchanger where fouling occurs In all cases, the following ranges are used: In a second step, an analysis is carried out Figure 13 shows that it is not possible to try to detect fouling using the raw evolution of an “instantaneous effectiveness.” Figure 14 shows the procedure used to detect fouling The first thousand samples are skipped, just to show that the analysis can be applied on an ongoing process The next samples (in the sliding observation window) are used to compute the approximation of the effectiveness On these samples, the wavelet transform is applied to the instantaneous effectiveness To avoid the “side effect” seen in Figure 9, we choose to consider only 85% of the computed values The average value obtained over this interval is considered to be the reference value; then the ratio of the approximated effectiveness to this average value is plotted The upper bound of the observation window is moved by an offset, and the same procedure is applied (wavelet transform + • • 0.6 to 1.2 kg/s for the mass flow rates 16 to 24◦ C for the inlet temperature of the cold fluid • 56 to 64◦ C for the inlet temperature of the hot fluid The values are randomly varying (from time to time 10,000 s), as shown in Figure 10 (partial view) During this period the fouling factor increases (applied on the hot side) Figure 11 shows this evolution The corresponding outlet temperatures are shown in Figure 12 Note that for a given “clean” set, the corresponding “fouling” set takes account of exactly the same inputs heat transfer engineering Figure 11 Evolution of the fouling factor (all sets) vol 32 nos 3–4 2011 354 Figure 14 Effectiveness ratio computation procedure is detected for the 100 “fouling” heat exchangers In the latter case, fouling is detected at sample 6,500 for 93 tests, and at sample 7,300 for tests This corresponds to fouling factors of 0.55 × 10−4 m2 K/W and 0.83 × 10−4 m2 K/W Note that to get these results, it has been necessary to increase the second parameter for the wavelet transform to 10 It can also be noted that when studying the sliding average value of the instantaneous effectiveness (over a 2,000-sample observation window), although the detection can occur sooner, some tests (7%) lead to higher values of the fouling factor at detection, as shown in Figure 16 In that case the reference value is the average value for the first sliding window (ending Figure 12 Partial view of the hot and cold fluids outlet temperature (set number 80) average value + ratio) In this study, the first computed value is calculated for sample number 6,500, to be able to detect fouling when the fouling factor is just higher than 0.55 × 10−4 m2 K/W As fouling is quite slow, the offset is set to 200 samples Figure 15 shows the evolutions of the effectiveness ratio for the 100 “clean” heat exchangers and for the 100 “fouling” heat exchangers A very simple test is then carried out As soon as the effectiveness ratio is lower than a threshold, it is said that fouling occurs It can be seen in Figure 15 that choosing 0.88 as the threshold leads to no false alarm, and that fouling Figure 13 Partial view of the evolution of an “instantaneous effectiveness” (set number 80) heat transfer engineering Figure 15 Evolution of the effectiveness ratio (100 curves) vol 32 nos 3–4 2011 ´ H INGIMUNDARDOTTIR AND S LALOT c d dt dim h I0 M ˙ m Ntu NUT nut Figure 16 Distribution of the fouling factor at detection when using a sliding average value at sample 3,000), and the threshold has to be decreased to 0.85 to get no false alarm DISCUSSION Although it could be thought that the wavelet transform is a very complicated tool, on a up-to-date personal computer the approximation of a 10,000-sample time series just needs 5.2 ms using (dim, scale) = (2, 10) as done here It can be concluded that this tool can be implemented on-line The authors think that in real applications, the variation ranges would be much smaller, leading to an even more accurate detection CONCLUSIONS In a first part, it has been shown that Simulink can be used to accurately model a cross-flow heat exchanger Then it has been shown that using wavelets can lead to an early detection of fouling in a heat exchanger Nevertheless, as this is not based on a sliding window, the method could be computationally burdensome if fouling occurs over a very long period In that case, the user could get enough data to try to adapt the procedure on sliding windows where the inputs not vary too much Tests will be carried out on the test rig under construction at the Universit´e de Valenciennes et du Hainaut Cambr´esis Then the efficiency of various methods (model based identification, neural networks, time-series analysis, fuzzy models, etc.) will be compared NOMENCLATURE A area of the convection surface, m2 (or m2/m when without subscripts) heat transfer engineering s scale T t 355 specific heat, J/kg-K derivative with respect to time first parameter for the wavelet transform convection coefficient, W/m2-K modified Bessel function of first kind and of order mass of fluid in one cell, kg mass flow rate, kg/s number of transfer units, dimensionless modified number of transfer units; based on the hotside values modified number of transfer units; based on the coldside values Laplace variable second parameter for the wavelet transform temperature, ◦ C time, s Greek Symbols α β δ φ γ η λ µ ν τ convection coefficient on the hot side, W/m2-K convection coefficient on the cold side, W/m2-K inverse of response time (cold side), s−1 wavelet inverse of response time (hot side), s−1 lowpass filter part of the wavelet highpass coefficient of the wavelet highpass filter part of the wavelet lowpass coefficient of the wavelet residence time, s Subscripts c h I i J j k n p cold side hot side in the middle of the cell on the left side of the cell number I in the middle of the cell on the bottom side of the cell number J translation parameter of the daughter wavelet dilatation parameter of the daughter wavelet separating plate Superscript f Laplace transform of function f REFERENCES [1] Rosmaninho, R., Rizzo, G., Măuller-Steinhagen, H., and Melo, L F., Deposition From a Milk Mineral Solution on Novel Heat Transfer Surfaces Under Turbulent Flow Conditions, Journal of Food Engineering, vol 85, no 1, pp 29–41, 2008 vol 32 nos 3–4 2011 356 [2] Gu´erin, R., Ronse, G., Bouvier, L., Debreyne, P., and Delaplace, G., Structure and Rate of Growth of Whey Protein Deposit From In Situ Electrical Conductivity During Fouling in a Plate Heat Exchanger, Chemical Engineering Science, vol 62, no 7, pp 1948– 1957, 2007 [3] Kukulka, D J., and Devgun, M., Fluid Temperature and Velocity Effect on Fouling, Applied Thermal Engineering, vol 27, no 16, pp 2732–2744, 2007 [4] Rosmaninho, R., Rocha, F., Rizzo, G., Măuller-Steinhagen, H., and Melo, L F., Calcium Phosphate Fouling on TiNCoated Stainless Steel Surfaces: Role of Ions and Particles, Chemical Engineering Science, vol 62, no 14, pp 3821–3831, 2007 [5] Prieto, M., Vallina, J., and Suarez, I., Application of a Design Code for Estimating Fouling On-Line in a Power Plant Condenser Refrigerated by Seawater, Proc ASMEZSITS International Thermal Science Seminar, 2000, June 11–14, Bled, Slovenia, on CDROM [6] Riverol, C., and Napolitano, V., Estimation of the Overall Heat Transfer Coefficient in a Tubular Heat Exchanger Under Fouling Using Neural Networks Application in a Flash Pasteurizer, International Communications in Heat and Mass Transfer, vol 29, no 4, pp 453–457, 2002 [7] Lalot, S., and Lecoeuche, S., Online Fouling Detection in Electrical Circulation Heaters Using Neural Networks, International Journal of Heat and Mass Transfer, vol 46, no 13, pp 2445–2457, 2003 [8] Jonsson, G R., Lalot, S., Palsson, O P., and Desmet, B., Use of Extended Kalman Filtering in Detecting Fouling in Heat Exchangers, International Journal of Heat and Mass Transfer, vol 50, no 13–14, pp 2643– 2655, 2007 [9] Lalot S., Palsson O P., Jonsson G R., and Desmet B., Comparison of Neural Networks and Kalman Filters Performances for Fouling Detection in a Heat Exchanger, International Journal of Heat Exchangers vol VIII, pp 151–168, 2007 [10] Delmotte, F., Delrot, S., Lalot, S., and Dambrine, M., Fouling Detection in Heat Exchangers With Fuzzy Models, Proc 19th International Symposium on Transport Phenomena, 17–21 August, 2008, Reykjavik, Iceland, on CDROM [11] Merc`ere, G., P´alsson, H., and Poinot, T., Linear ParameterVarying Identification of a Cross Flow Heat Exchanger for Fouling Detection, Proc International Conference on Heat Exchanger Fouling and Cleaning, Schladming, Austria, June 14–19, 2009 [12] Lalot, S., On-Line Detection of Fouling in a Water Circulating Temperature Controller (WCTC) Used in Injection Moulding: Part 1: Principles, Applied Thermal Engineering, vol 26, no 11–12, pp 1087–1094, 2006 [13] Lalot, S., On-Line Detection of Fouling in a Water Circulating Temperature Controller (WCTC) Used in Injection heat transfer engineering [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] Moulding: Part 2: Application, Applied Thermal Engineering, vol 26, no 11–12, pp 1095–1105, 2006 Lalot, S., and Merc`ere, G., Detection of Fouling in a Heat Exchanger Using a Recursive Subspace Identification Algorithm, Proc 19th International Symposium on Transport Phenomena, 17–21 August, 2008, Reykjavik, Iceland, on CD-ROM Gudmusson, O., Palsson, O P., Palsson, H., and Lalot, S., Method to Detect Fouling in Heat Exchangers, Proc 11th International Symposium on District Heating and Cooling, Reykjavik, 31 August–2 September 2008 Prabhakar, S., Mohanty, A R., and Sekhar, A S., Application of Discrete Wavelet Transform for Detection of Ball Bearing Race Faults, Tribology International, vol 35, pp 793–800, 2002 Lou, X., and Loparo, K., Bearing Fault Diagnosis Based on Wavelet Transform and Fuzzy Inference, Mechanical Systems and Signal Processing, vol 18, pp 1077–1095, 2004 Purushotham, V., Narayanan, S., and Prasad, S A N., Multi-Fault Diagnosis of Rolling Bearing Elements Using Wavelet Analysis and Hidden Markov Model Based Fault Recognition, NDT&E International, vol 38, pp 654–664, 2005 Li, T.-S., Applying Wavelets Transform, Rough Set Theory and Support Vector Machine for Copper Clad Laminate Defects Classification, Expert Systems With Applications, vol 36, pp 5822–5829, 2009 Wu, J.-D., Hsu, C.-C., and Wu, G.-Z., Fault Gear Identification Using Discrete Wavelet Transform and Adaptive Neuro-Fuzzy Inference, Expert Systems With Applications, vol 36, pp 6244–6255, 2009 Wu, J.-D., and Liu, C.-H., An Expert System for Fault Diagnosis in Internal Combustion Engines Using Wavelet Packet Transform and Neural Network, Expert Systems With Applications, vol 36, pp 4278–4286, 2009 Heneghan, C., Khanna, S M., Flock, A., Ulfendahl, M., Brundin, L., and Teich, M C., Investigating the Nonlinear Dynamics of Cellular Motion in the Inner Ear Using the Short-Time Fourier and Continuous Wavelet Transforms, IEEE Transactions on Signal Processing, vol 42, no 12, pp 3335–3352, 1994 Farooq, O., and Datta, S., Phoneme Recognition Using Wavelet Based Features, Information Sciences, vol 150, pp 5–15, 2003 C¸engel, Y A., Introduction to Thermodynamics and Heat Transfer, Irwin/McGraw-Hill, New York, 1997 Montoya Zegarra, J A., Leite, N J., and da Silva Torres, R., Wavelet-Based Fingerprint Image Retrieval, Journal of Computational and Applied Mathematics, vol 227, no 2, pp 294–307, 2009 Sung, C K., Tai, H M., and Chen, C W., Locating Defects of a Gear System by the Technique of Wavelet Transform, vol 32 nos 3–4 2011 ´ H INGIMUNDARDOTTIR AND S LALOT [27] [28] [29] [30] Mechanism and Machine Theory, vol 35, pp 1169–1182, 2000 Belotti, V., Crenna, F., Michelini, R C., and Rossi, G B., Wheel-Flat Diagnostic Tool Via Wavelet Transform, Mechanical Systems and Signal Processing, vol 20, pp 1953–1966, 2006 Saravanan, N., and Ramachandran, K I., Fault Diagnosis of Spur Bevel Gear Box Using Discrete Wavelet Features and Decision Tree Classification, Expert Systems With Applications, vol 36, pp 9564–9573, 2009 Wu, J.-D., and Chan, J.-J., Faulted Gear Identification of a Rotating Machinery Based on Wavelet Transform and Artificial Neural Network, Expert Systems with Applications, vol 36, pp 8862–8875, 2009 Daubechies, I., Orthonormal Bases of Compactly Supported Wavelets, Communications on Pure and Applied Mathematics, vol 41, no 7, pp 909–996, 1988 heat transfer engineering 357 Helga Ingimundard´ottir completed a B.Sc in mathematics from the University of Iceland in 2008 She is currently finishing an M.Sc in computational engineering and is also a full-time Ph.D student under the supervision of Prof T´omas Philip R´unarsson at the aforementioned university Her research interest areas include heuristics, optimization, and computational intelligence Sylvain Lalot is a professor at Universit´e Lille Nord de France (UVHC) He received his engineer diploma from the Institut Industriel du Nord in 1982 (mechanical engineering), his master’s degree in thermal engineering from the Universit´e de Valenciennes in 1983, and his Ph.D in 1994 from the Universit´e de Valenciennes He worked for the French subsidiary of GEA from 1982 to 1984 Then he was a lecturer at the Ecole d’Ing´enieurs de Gabes (Tunisia) for years He joined VULCANIC between 1986 and 1992 to develop new products Then he joined academia at the Ecole d’Ing´enieurs du Pas-de-Calais as the head of the mechanical engineering department, where he worked for 10 years He is now a professor at the UVHC His present research deals with the transient states of heat exchangers, with a focus on the detection of fouling vol 32 nos 3–4 2011