PULSED FLOW AND SURFACE COATINGS TO MITIGATE FOULING

Transcription

1 Proceedings o International Conerence on Heat Exchanger Fouling and Cleaning (Peer-reviewed) June 09-14, 2013, Budapest, Hungary Editors: M.R. Malayeri, H. Müller-Steinhagen and A.P. Watkinson Published online PULSED FLOW AND SURFACE COATINGS TO MITIGATE FOULING H. Föste, F. Stehmann, W. Augustin and S. Scholl Technische Universität Braunschweig, Institute or Chemical and Thermal Process Engineering, Langer Kamp 7, Braunschweig, Germany, h.oeste@tu-braunschweig.de ABSTRACT Crystallization ouling on heat transer suraces is a major problem in many ields o the process industry. Adverse eects, such as decreasing thermal eiciency, increasing pressure drop and production loss in consequence o cost intensive and persistent cleaning procedures, are to be mentioned. Crystallization ouling in heat exchangers is a superimposed process o deposition and removal. The objective o this work is to mitigate ouling by enhancing removal applying two complementary approaches: (1) Increased luid orces using pulsed low (2) Decreased adhesion orces between wall and crystals using dierent surace coatings with varied ree surace energies. Experimental examinations were carried out with a calcium sulate solution in a rectangular low channel (Re ). Two exchangeable stainless steel plates were mounted on a heating element, so that dierent surace coatings could be examined. The coatings applied in this work were SICAN and SICON. A pulsed low was imposed in intervals during the heat transer process. Dierent modes o pulsation were tested. The examinations showed that applying pulsed low to an adhering process leads to a decreased inal ouling resistance. The decrease o the inal ouling resistance results rom increasing luid orces. For certain parameter combinations a balance between deposition and removal occurs and leads to an asymptotic ouling resistance. The experiments using surace coatings revealed to lengthen the induction period. INTRODUCTION Crystallization ouling on heat transer suraces is a major problem in the process industry. There are several disadvantages linked to the appearance o ouling. The growth o crystals leads to an increase o the heat transer resistance, which can be quantiied by the thermal ouling resistance R. R 1 U 1 U = (1) 0 U describes the overall thermal resistance o the ouled plate and U 0 the overall resistance o the clean plate. The ouling resistance may also be quantiied using a mass based approach (Mayer et al., 2010), m describes the ouling mass, λ thermal conductivity o the ouling layer and ρ the density o the ouling layer: R m λ ρ (2) Additionally an increase o the pressure drop occurs due to roughness eects (Albert et al., 2011) and the reduction in low area o the duct. Both mechanisms reduce the eiciency o the whole process. Oversized heat exchangers and peripheral devices or periodic cleaning procedures are common approaches to handle this issue. This results in a reduced economic and ecological eiciency o a process. The ouling process can be divided into two sections, the induction period and crystal growth period (Geddert, 2009). During the induction period no signiicant decrease o the heat transer can be detected. During the crystal growth period an increase o the thermal resistance occurs. Both periods contain deposition and removal processes. Depending on process parameters dierent ouling progressions can appear: (a) Linear progression: The ratio o deposition and removal is constant or the removal can be neglected (b) Decreased progression: The rate o growth decreased over time, no inal value will be achieved (c) Asymptotic progression: Deposition and removal are in balance, a inal value will be achieved (d) Sawtooth progression: Removal rate is not constant, pieces o the deposit are being removed on the microscale. FOULING MITIGATION Fig. 1 Deposition and removal on a heat transer surace [Geddert, 2011] 397

2 Föste et al. / Pulsed low and surace coatings to mitigate ouling Numerous actors aect ouling and cleaning [Geddert, 2011], see Fig. 1. As the crystallization ouling process is a superposition o deposition and removal, two general mitigation approaches are conceivable: (1) Reduction o deposition (2) Enhancement o removal In the ollowing examination, the enhancement o removal shall be observed. The ouling mass rate m is deined as the dierence between the deposition mass rate m d and the removal mass rate m r. Förster (2001) speciied the removal rate using the removal probability Γ which is a unction o the ratio o shear orce F τ and adhesion orce F ad (Chen et al., 1995). m F 1 Γ F & = m& d m& r = m& d τ (3) ad Reerring to this equation the removal rate can be increased by increasing the acting shear orce on the crystals and decreasing the adhesion orce between crystal and surace. Besides the possibility o increasing the main low velocity to enhance the removal rate, Augustin (2003) introduced the approach o a pulsed low to reduce the ouling tendency o calcium sulate. The superposition o a steady base low and an oscillating low movement results in a pulsed low. The experiments with pulsation showed a continuous removal o crystals, which led to a quasi-stationary state o the ouling resistance. In urther investigations a non-continuous pulsation was applied where the period o pulsation was interrupted by a period o steady low conditions. It can be seen that or short enough intervals the results are similar to the continuous pulsation (see Fig. 2). The application o pulsed low was also used or the enhancement o cleaning processes. Gillham et al. (2000) revealed the cleaning time o whey protein ouling to be signiicantly decreased by pulsed low in comparison to a steady cleaning procedure. Blel et al. (2009a, 2009b) examined bacterial removal under pulsed low conditions and noticed an increase o the constant removal rate in comparison with steady low conditions. modiications (DLC, SICAN, SICON ). The main characteristic o these suraces is the reduction o adhesion orces between crystals and surace. Combined with shear orces driven by the luid low the removal rate can be enhanced (Geddert, 2009). Geddert (2009) also examined the ree surace energies o the used suraces (see Fig. 3), but did not ound a clear correlation between ree surace energy and induction period. Zhao et al. (2005) examined the ouling behavior o a Ni-Cu-P-PTFE composite coating and the inluence o ree surace energy on the adhesion orce. They identiied a ree surace energy range in which the adhesion orce is minimal and concluded the potential o such a coating to reduce mineral and bioouling [Zhao et al. 2005]. Zettler et al. (2005) examined several coatings based on dierent coating technologies. They did not ind a correlation between surace energy and ouling behavior, but it tends to reduced ouling at reduced surace energies. Gao et al. (2006) described an extension o the induction period and enhanced anti-ouling properties by the use o a Ni-based implanted tube at boiling conditions. Al-Janabi et al. (2011) examined solvent based, water based and electroless Ni-P-BN coatings whereby the solvent based and Ni-P-BN coatings showed a signiicantly decreased inal ouling resistance. Fig. 3 Surace energy o the dierent coatings on untreated stainless steel [Geddert, 2009] Based on these observations the two ouling mitigation strategies were examined. In the presented work both approaches are investigated separately to attain a deeper understanding o the main parameters inluencing the ouling process. In uture work both approaches will be combined applying the most eective parameters o the present work. Fig. 2 Fouling curves or dierent delay times, w stat = 0.25 m/s, W = 1 [Augustin, 2003] The possibilities o surace modiication are described by several authors. Förster (1999) and Geddert (2009) showed increased induction periods by the use o energetic surace EXPERIMENTAL SETUP The experiments were carried out with a plate heat exchanger, see Fig. 4. The dierent suraces were electrically heated and the temperature was measured by two thermocouples to calculate the thermal ouling resistance. To eliminate the inluence o smaller luctuations o the electrical power on the ouling resistance, the electrical power at each time step was used or the calculation o R. For parallel testing, two test suraces could be mounted on 398

3 Heat Exchanger Fouling and Cleaning 2013 the heating element, one on each side. This led to the possibility o examining dierent surace modiications and their inluence on the ouling process. The outer body o the low channel contained two transparent sections made o PMMA to observe the ouling process visually. Coming rom a storage tank, 150 L o aqueous calcium sulate solution was pumped in circle. Ater the centriugal pump, the solution passed through a heat exchanger, which enabled a constant inlet temperature. Finally, it passed through the investigated test section. A detailed description o the test rig can be ound in (Förster, 1999). period o steady low conditions. Dierent parameters were varied (see Fig. 6): (i) The waviness, which describes the ratio o the maximum oscillating velocity w os,max and the steady velocity w stat, was adjusted between 0.5 and 1.5, (ii) the pulsation interval t p, which is deined as the sum o one pulsation time plus one steady interval, was varied between 10 s and 1000 s and (iii) the number o strokes were varied between 2 and 8. Fig. 4 Test section [Förster, 1999] For pulsating experiments, the plant was equipped with a pulsator. The pulsator consisted o a piston pump with the suction side shut. A pulsed low consists o a steady low (generated by the centriugal pump) which is superimposed by an oscillating low movement (generated by the pulsator). The resulting ideal velocity unction is shown in Fig. 5. I the maximum oscillating low velocity is greater than the steady low, a temporary low reversal occurs. The characteristic parameter is the dimensionless waviness W: W w os, max = (4) w Fig. 6 Schematic diagram o the main pulsation parameters The experiments were carried out with a standard stainless steel plate. The concentration o calcium sulate was 24 mmol/l and the luid velocity was set to 0.12 m/s. The experiments with surace coatings were carried out at steady low conditions (w stat = 0.1 m/s) and a calcium sulate concentration o 27 mmol/l. The surace coatings applied in this work were SICAN and SICON. SICAN is a diamond like carbon (DLC) surace where silicon is built into the a- C:H matrix, SICON includes the combination o silicon and oxygen inside the DLC layer. The thickness o the coatings is 3 µm and did not change the original roughness o the stainless steel plate. The coatings were developed at the Fraunhoer Institute or Surace Engineering and Thin Films in Braunschweig, Germany. Tab. 1: Overview o the test parameters Parameter Range Steady low velocity w stat [m/s] Max. oscillating low velocity w os,max [m/s] Waviness W [-] Pulsation interval t p [s] Number o strokes t p 2-8 [-] Pulsation requency 1.7 [Hz] Reynolds number [-] Concentration CaSO [mmol/l] Fig. 5 Stationary and oscillating component o pulsating low [Augustin, 2003] In order to leave the energy input in range o steady low conditions, the experiments shown here were carried out with temporary pulsation instead o permanent pulsation. This means that a period o pulsation is interrupted by a RESULTS AND DISCUSSION The ollowing results show a parameter study with a variation o the waviness, the number o strokes and the pulsation intervals. Inluence o the pulsation interval Experiments were carried out with t p = 10 s, 100 s, and 1000 s with a number o strokes o t p = 2. As a reerence, the ouling curve o a constant low velocity at w = 0.12 m/s is 399

4 Föste et al. / Pulsed low and surace coatings to mitigate ouling presented. The comparison o the pulsation intervals shows no signiicant dierences during the irst 12 h. Ater that, an increase o the ouling resistance in each case occurs. All ouling curves at pulsed low conditions show a more or less pronounced sawtooth progression where crystal growth is ollowed by a sudden removal. Generally it can be seen that with a decreasing pulsation interval the inal ouling resistance is also decreased (Fig. 7). I the interval is long ( tp = 1000 s) the pulsation has no inluence on the ouling behavior and the inal ouling resistance is approximately equal to the steady experiment. The ouling curve at tp = 10 s reached a constant value and is a 65 % reduction o the steady case. Considering the induction period, it can be seen that pulsed low does not lead to an increase o the induction period, even a more requent pulsation reduces the induction time. To urther reduce the energy input, the pulsation interval was increased by a actor o ten. The ouling curves in Fig. 9 show experiments at tp = 1000 s. It can be seen that the eect o pulsation is much smaller. The dierence in the inal ouling resistance is insigniicant between the steady case and tp = 2 and tp = 4, respectively. Only at tp = 8 a decline occurs. In conclusion, the pulsation interval or present parameters appears to be too long to achieve a positive eect on the ouling behavior. This is in good agreement with results described by Augustin et al. (2003). Fig. 9 Fouling curves at constant pulsation interval tp=1000 s and a constant waviness o W = 1.5, variation o number o strokes tp Fig. 7 Fouling curve at constant number o strokes tp = 2 and a constant waviness o W = 1.5, variation o pulsation interval tp Inluence o the number o strokes Experiments were carried out with tp= 2, 4, 8 at constant tp = 100 s respectively 1000 s. The ouling curves presented in Fig. 8 are examinations at tp = 100 s. In general the tendency is similar to Fig. 7. Increased pulsation leads to a decrease o the inal ouling resistance, which is signiicant in each case, and to shorter induction periods. Furthermore the sawtooth developing can be seen. Fig. 8 Fouling curve at constant pulsation interval tp=100 s and a constant waviness o W = 1.5, variation o number o strokes tp Inluence o waviness The inal pulsation parameter that was examined was the waviness W. The comparison includes the waviness W = 1.5, which leads to a temporary low reversal based on equation (4) and the waviness W = 0.5 where no back low is occurring. The resulting ouling curves, see Fig. 10, reveal that both wavinesses lead to a decrease o the inal ouling resistance in comparison to the steady case. Furthermore, it is apparent that the inal ouling resistance between both pulsation parameters does not dier much, but the sawtooth developing is more pronounced or W = 1.5. Fig. 10 Fouling curve at constant number o strokes tp = 2 and a constant pulsation interval tp = 100 s, variation o waviness W 400

5 Heat Exchanger Fouling and Cleaning 2013 Fouling layer properties As a result o the unsteady low movement it could be determined that bigger parts o the orming crystal deposit were removed, which can be seen by the sawtooth progression o the ouling curves. Fig. 11 shows a picture o the ouling layer ater the test run at steady low conditions. The layer is uniorm and adheres completely onto the heat transer surace. In contrast, Fig. 12 shows the inal ouling layer grown under pulsed low conditions. Fig. 11 Final ouling layer at steady low o w = 0.12 m/s, test duration 40.6 h Fig. 12 Final ouling layer at waviness W = 1.5, t p = 100 s, t p = 2, test duration 42.3 h under pulsed low conditions, which is in good agreement with equation (2). The required deposited mass to reach a speciic ouling resistance is less than under pulsed low condition, due to the denser layer structure. Inluence o surace coatings The ollowing section illustrates the inluence o dierent surace coatings on the ouling behavior under steady low conditions. Fouling experiments were carried out at accelerated conditions to obtain results in relatively short time exposures due to laboratory restrictions. The concentration o calcium sulate was increased to 27 mmol/l and the luid velocity was decreased to 0.1 m/s. Fig. 14 illustrates the ouling curves o SICAN and SICON. Both runs are still in the induction period ater 50 h. The experiment with SICAN was continued until 350 h. Ater that time the measured ouling resistance was still close to zero and no visible crystals were present. The extended induction period can be explained by the low adhesive characteristic o the coatings combined with acting luid orce (Geddert, 2009). The slight increase o the ouling resistance can be explained by the deposition o crystal layers starting rom a small gap at the beginning and end o the test plate. These layers covered a small part o the plate and also inluenced the luid low. In current investigations, the experimental conditions are being modiied in order to achieve ouling in a reasonable time rame. Based on this, the combination o both ouling mitigation approaches can be examined. During the pulsation period, parts o the orming crystal deposit were removed completely or peeled o partly. In the gap between the wall and the partly peeled o layer a new ouling layer grew, which results in a layer composition. It can be hypothesized that an optimized low coniguration can lead to a complete removal o these layers. Fig. 14 Fouling curve o SICAN and SICON at steady low conditions Fig. 13 Correlation o the inal ouling resistance and inal mass o the ouling layer For a urther examination o these phenomena, the mass o the ouling layer was measured at the end o each experiment and compared to the inal ouling resistance. Fig. 13 illustrates the correlation between both values. A linear it seems to be a reasonable correlation or the results CONCLUSIONS In this study the inluence o pulsed low and surace modiications on the ouling process was examined. I the pulsation time in comparison to the steady low velocity period is long enough, the application o a temporary pulsation leads to a signiicant reduction o the inal ouling resistance in comparison to the steady case. The induction period could not be increased by the application o pulsed low, even a more requent pulsation reduces the induction period. This phenomenon is not ully understood and is currently being investigated in more detailed

6 Föste et al. / Pulsed low and surace coatings to mitigate ouling The experiments using surace coatings remain in the induction period during the whole experimental time. Currently, investigations are underway to examine the combination o both approaches. Especially i the process parameters are in a critical range so that ouling occurs on surace coatings, the combination o both seems to be a promising combination to a urther reduction o ouling in heat exchangers. NOMENCLATURE F ad adhesion orce, N F τ shear orce, N R thermal ouling resistance, m 2 K W -1 Re Reynolds number, dimensionless U 0 overall thermal resistance o the clean plate, W m -2 K -1 U overall thermal resistance o the ouled plate, W m -2 K -1 W waviness, dimensionless m d deposited mass, kg m -2 m d deposition mass rate, kg m -2 s -1 m ouling mass, kg m -2 m ouling mass rate, kg m -2 s -1 m r removed mass, kg m -2 m r removal mass rate, kg m -2 s -1 t p number o strokes, dimensionless w os,max max. oscillating velocity, m s -1 w mean/ steady velocity, m s -1 Γ removal probability, dimensionless t p pulsation interval, s λ thermal conductivity o the ouling layer, W m -1 K -2 ρ density o the ouling layer, kg m -3 REFERENCES Al-Janabi, A., Malayeri, R. M., Müller-Steinhagen, H., 2011, Minimization o CaSO 4 deposition through surace modiication, Heat Transer Eng., 32:3-4, Albert, F., Augustin, W., Scholl, S., 2011, Roughness and constriction eects on heat transer in transer in crystallization ouling, Chem. Eng. Sci. 66, Augustin, W., 2003, Crystallization on Heat Transer Suraces under Unsteady Flow Conditions, Turbulence, Heat Mass Transer 4, Begell House Inc. Blel, W., Le Gentil-Lelièvre, C., Bénézech, T., Legrand, J., Legentilhomme, P., 2009a, Application o turbulent pulsating lows to the bacterial removal during a cleaning in place procedure. Part 1: Experimental analysis o wall shear stress in cylindrical pipe, J. Food Eng., 90, Chen, P., Chen, X. D., Free, K. W., 1995, A computational luid mechanics analysis o the removal orces acting on crystal deposits at the suraces o plate heat exchangers, Proceeding o the sixth Asian congress o luid mechanics, Singapore Förster, M., 2001, Verminderung des Kristallisationsoulings durch gezielte Beeinlussung der Grenzläche zwischen Kristallen und Wärmeübertragungsläche, Dissertation, TU Braunschweig, Cuvillier, Göttingen, Germany Förster, M., Augustin, W., Bohnet, M., 1999, Inluence o adhesion orce crystal/heat exchanger surace on ouling mitigation, Chem. Eng. Process., 38, Gao, M., Sun, F., Shi, Y., Lei, S., Niu, Z., 2006, Experimental research and theoretical analysis o antiouling characteristics on the surace o Ni-based implanted tube, Int. Commun. Heat Mass Transer, 33, Geddert, T., Augustin, W., Scholl, S., 2011, Inluence o Surace Deects and Aging o Coated Suraces on Fouling Behavior, Heat Transer Eng., 32:3-4, Geddert, T., Bialuch, I., Augustin, W., Scholl, S., 2009, Extending the Induction Period o Crystallization Fouling Through Surace Coating, Heat Transer Eng., 30:10-11, Gillham, C. R., Fryer, P. J., Hasting, A. P. M., Wilson, D. I., 2000, Enhanced cleaning o whey protein soils using pulsed lows, J. Food Eng., 46, Mayer, M., Bucko, J., Benzinger, W., Dittmeyer, R., Augustin, W., Scholl, S., 2010, Investigation and visualization o crystallization ouling in a micro-heat exchanger, in: Proceedings o the 2nd European Conerence on Microluidics, Toulouse, France Zettler, U. H., Weiß, M., Zhao, Q., Müller-Steinhagen, H., 2005, Inluence o surace properties and characteristics on the ouling in plate heat exchanger, Heat Transer Eng., 26(2): 3-17 Zhao, Q., Liu, Y., Wang, C., Wang, S., Müller- Steinhagen, H., 2005, Eect o surace ree energy on the adhesion o bioouling and crystalline ouling, Chem. Eng. Sci, 60, Blel, W., Legentilhomme, P., Bénézech, T., Legrand, J. Le Gentil-Lelièvre, C., 2009b, Application o turbulent pulsating lows to the bacterial removal during a cleaning in place procedure. Part 2: Eects on cleaning eiciency, J. Food Eng., 90,