International Communications in Heat and Mass Transfer

Transcription

1 International Communications in Heat and Mass Transfer 38 (2011) Contents lists available at ScienceDirect International Communications in Heat and Mass Transfer journal homepage: Benefit of filtration in physical water treatment for the mitigation of mineral fouling in heat exchangers Wontae Kim 1, Young I. Cho Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA 19104, USA article info abstract Available online 19 May 2011 Keywords: Fouling Physical water treatment Filtration Heat exchanger Fouling prevention Mineral scale The objective of the present study was to investigate the benefit of filtration in the physical water treatment used to mitigate mineral scale build-up in heat exchangers. Two types of filters were used: a 5-μm cartridge fabric filter and a sand filter with 20-μm pores, both of which were used at a side-stream loop in a laboratory cooling tower. Heat transfer fouling experiments were conducted using a double-pipe heat exchanger with cooling water at 5 and 8 cycles of concentration with a make-up water hardness of ppm. The test results demonstrated that the filtration enhanced the performance of the PWT synergistically, resulting in a near initial peak heat transfer performance in the double-pipe heat exchanger Elsevier Ltd. All rights reserved. 1. Introduction Fouling in general can be described as the formation of unwanted deposits on a heat transfer surface. Mineral fouling or the formation of scales such as calcium carbonate is common in cooling-water applications, where pure water evaporates to remove heat gained from condenser [1]. Hence, circulating cooling water quickly becomes supersaturated even if make-up water is soft. As the supersaturated cooling water is heated inside condenser, the calcium ions precipitate due to the inverse solubility characteristics of CaCO 3 salt. A number of studies [2 9] have been conducted to better understand mineral fouling process and its mitigation. Since the scale deposits have very low thermal conductivities and thus behave as an insulating layer, the heat transfer rate is diminished at the heat exchanger with scale deposits [2]. As the deposits grow on the heat transfer surface, the opening inside the heat exchanger becomes smaller, increasing the pressure drop. Subsequently, more pumping power is needed to maintain flow, which will result in higher energy consumption. Currently, industry-standard practices for preventing mineral fouling call for the addition of scale-inhibiting chemicals to circulating cooling water [1]. The use of such chemicals contributes to fresh water pollution, a major environmental concern. Several physical water treatment (PWT) methods [10 13] have been introduced in recent years as alternatives to standard chemical water treatment; some of the best results have been documented using solenoid-type PWT technologies using oscillating fields. PWT means any non-chemical methods that are utilized for the mitigation Communicated by W.J. Minkowycz. Corresponding author. address: choyi@drexel.edu (Y.I. Cho). 1 Present address, Aritec Corp., Yongin-Si, Gyeonggi-Do, Republic of Korea. of mineral fouling. The present study utilized a solenoid-coil device with a square-wave pulsing current to create time-varying magnetic fields, which in turn produced an induced pulsating electric field in the circulating water, a process known as Faraday's law [14]. Excess mineral ions such as calcium and magnesium in cooling water precipitate in the form of mineral salts, providing nucleation sites for other dissolved mineral ions [12]. As the cooling water is continuously circulated, the precipitated seed crystals grow into larger particles. Thus, it is hypothesized that if the large particles could be removed from the cooling water, fouling at the heat exchanger could be prevented or significantly mitigated. Furthermore, the amount of blowdown could be reduced, thus increasing the cycle of concentration (COC) with substantial water savings. The objective of the present study was to investigate the benefit of filtration used in conjunction with PWT by conducting fouling heat transfer experiments in a laboratory cooling-tower system at two different COC levels of 5 and 8 with a make-up water hardness of ppm. 2. Experimental facility and methods Fig. 1 shows a schematic diagram of the present experimental setup. It consisted of a laboratory cooling tower, a double-pipe heat transfer test section, an automatic blowdown system based on electric conductivity measurement, four thermocouples, a flow meter, a pump, a data acquisition system (DAS), a PWT unit using a solenoid coil and power supply, and a side-stream filtration. The detail view of the heat transfer test section is shown in Fig. 2, which was made of two concentric circular tubes: a copper tube at the inside and a quartz tube at the outside for visual inspection. The inner tube was a typical copper tube used in industrial chillers manufactured by companies like York, Carrier, and Trane. The outside diameter of /$ see front matter 2011 Elsevier Ltd. All rights reserved. doi: /j.icheatmasstransfer

2 W. Kim, Y.I. Cho / International Communications in Heat and Mass Transfer 38 (2011) Nomenclature A o Outside surface area of copper tube (m 2 ) c p Specific heat of water (J/kg K) d Diameter of copper tube (mm) ṁ Mass flow rate of water (kg/s) Q Heat transfer rate (W) R f Fouling resistance (m 2 K/W) T i Inlet temperature ( C) T o Outlet temperature ( C) ΔT LMTD Log-mean-temperature difference ( C) U Overall heat transfer coefficient (W/m 2 K) Subscripts c Cold side f Fouled state h Hot side i Initial clean state the copper tube was 1.27 cm, whereas the inside diameter of the quartz tube was 1.76 cm. Hot water moved inside the copper tube, while the cold water moved outside the copper tube, i.e., through the annulus gap made by the two tubes. The hot and cold water were flowing in the opposite directions, forming a counter-flow heat exchanger. The inlet temperatures at the hot and cold sides were maintained at 95 C and 27 C, respectively, while the outlet temperatures at the hot and cold sides were at approximately 88 C and 36 C, respectively. The volume flow rate of the cold water was m 3 /s (1.5 gpm), with the corresponding flow velocity at the heat transfer test section of 0.81 m/s (Re=4000). The durations for the fouling tests were 90 and 140 h for the cases of COCs of 5 and 8, respectively. The fouling tests were conducted using a relatively high heat flux of kw/m 2 in order to accelerate the fouling process, a practice which was common for fouling researches. Table 1 shows the test conditions of the previous fouling studies [2 8], which used high levels of heat flux as in the present study. The heat transfer rates at the hot and cold water sides were calculated as follows [15]: Q h = ṁ h c ph ΔT h ðhot water sideþ ð1þ Q c = ṁ c c pc ΔT c ðcold water sideþ ð2þ The specific heat c p was based on the average values of the inlet and outlet temperatures for both hot and cold water. The overall heat transfer coefficient U was calculated as [15] Q c U = A o ΔT LMTD The heat transfer surface area A o was calculated by A o =πd o L e, where d o is the outer diameter of the copper tube and L e is the effective heat transfer length. The log-mean-temperature-difference ΔT LMTD was determined by the following equation [15]: T h;o T c;i T h;i T c;o ΔT LMTD = 2 3 T h;o T c;i ln4 5 T h;i T c;o The fouling resistance was calculated using the universal heat transfer coefficients U f and U i corresponding to the fouled and initial states, respectively as R f = 1 U f 1 U i Fig. 2. Sketch of the heat transfer test section. ð3þ ð4þ ð5þ Fig. 1. Schematic diagram of the fouling test facility in the present study.

3 1010 W. Kim, Y.I. Cho / International Communications in Heat and Mass Transfer 38 (2011) Table 1 Data for heat flux used in the previous fouling studies. References Heat flux (kw/m 2 ) Flow velocity (m/s) Concentration (ppm) Foulant Hasson, 1968 [3] Calcium Kim & Webb, 1991 [4] Aluminum oxide/ferric oxide Somerscales, ,500 MgO 1991 [5] Nasrazadani, Calcium 1994 [6] Sheikholeslami & Calcium Watkinson, 1986 [7] Morse & Knudsen, TDS 1997 [8] Helalizadeh, 2000 [2] / Calcium sulphate/ calcium carbonate Present study Total hardness Fig. 3. Fouling resistance over time for 5 COC without filtration. Fouling experiments were conducted at two different COCs of 5 and 8. Table 2 shows the water analyses of circulating cooling water at the two COCs, including those of make-up water. Side-stream loop was used for the installation of filter. For 5 COC, a 5-μm cartridge fabric filter with a length of 25.4 cm was used at 10% of the main flow rate. At 8 COC, a sand filter was utilized at 5% of the main flow rate. In the case of the sand filter, cooling water was introduced to the top of the sand filter and fell through sand media by gravity. Hence, calcium carbonate particles produced by the PWT continued to be accumulated on the top level of the sand filter, creating a concrete-like layer, which acted as another layer of filter. For all tests, the fouling resistance, water analyses, and photographic images of the fouled and clean tubes were recorded. To check the repeatability of the present fouling test procedure, the test for the case of 5 COC without filtration was conducted twice, and the results were found to be satisfactory. The accuracy of thermocouple measurements was ±0.15 C. Detailed uncertainty analysis proposed by Kline and McClintock [16] was conducted and provided elsewhere [17]. In short, the flow rate measurement had 2.3% error; temperature measurements had 0.4% error; heat transfer rate measurement had 2.4% error; surface area measurement had 0.2% error; universal heat transfer coefficient measurement had 2.3% error; fouling resistance measurement had 10% error. 3. Results and discussion 3.2. Tests for 5 COC Case without filtration Fig. 3 shows the results of fouling resistance over time for the notreatment case and the case with PWT treatment at 5 COC. For the first 30 h, the fouling curves for both cases had negative values, reaching peak negatives at around 15 h. For the no-treatment case, the fouling resistance returned to zero at t=30 h and remained there until t=40 h, when the fouling resistance drastically began to increase until the end of the test to a maximum value of about m 2 K/W. In contrast, the fouling resistance for the PWT case remained negative until t =80 h, when it slowly began to increase, slightly going over zero at the end of the test. The fouling resistance for the PWT case was about m 2 K/W at the end of the test (i.e., t=90 h), approximately 90% reduction compared to the no-treatment case. The fouling resistance for the no-treatment case reached the industry standard allowance value of m 2 K/W [18] at approximately t=50 h, whereas the fouling resistance for the PWT case remained significantly below the industry standard allowance value even at the end of the test. Fig. 4 shows the photographs of fouled heat transfer tubes. The heat transfer surface for the no-treatment case was completely covered with scale, whereas that for the PWT case was barely covered with scale Tests for 5 COC Case with 10% filtration Fig. 5 shows the fouling resistance curves for two cases with filtration using 5-μm cartridge fabric filter: one with filter only, and the other with the simultaneous use of PWT and filter. While the filter-only case gave an impressive result, the fouling resistance for the PWT+filtration case provided a better fouling curve. The fouling resistance curve for the PWT+filtration case remained almost below zero during the entire test period, reaching a fouling resistance value of m 2 K/W at the end of the test as compared to m 2 K/W for the filter-only case. The fouling resistances for both cases remained significantly below the industry standard allowance value of m 2 K/W throughout the entire tests. Fig. 6 shows photographs of fouled heat transfer tubes for both cases. The heat transfer surface for the filter-only case had a relatively clean copper tube surface with a thin coating of scale, whereas that for the Table 2 Water analyses for the make-up and circulating water used in the present study. 5 COC 8 COC Make-up Circulating Cycle Make-up Circulating Cycle Conductivity (μmhos/cm) Hardness (ppm) Calcium (ppm) Magnesium (ppm) Chloride (ppm) Total alkalinity (ppm) ph

4 W. Kim, Y.I. Cho / International Communications in Heat and Mass Transfer 38 (2011) Fig. 6. Photographs of copper tubes at the end of fouling test, (a) 5-μm fabric filter only and (b) PWT+filter. Fig. 4. Photographs of copper tubes for 5 COC at the end of fouling test, (a) no treatment and (b) PWT, both without filtration m 2 K/W at the end of the test, resulting in approximately 27% drop compared to the no-treatment case. Photographs of fouled heat transfer surfaces for the two cases are shown in Fig. 8, where the heat transfer surface obtained for the no-treatment case was almost completely covered by CaCO 3 scale, whereas that for the PWT case was relatively free from scale Tests for 8 COC Case with filtration case with the PWT+filtration case depicted the original Coppertone color of copper tube in most parts of the tube with slightly stained spots with scale Tests for 8 COC Case without filtration Fig. 7 shows the variations of fouling resistances over time for the no-treatment case and PWT treatment case at COC of 8. For the notreatment case, the fouling resistance began to increase linearly after reaching a peak minimum value at t=15 h, passed the industry standard allowance value of m 2 K/W at t=60 h, and continued to rise, reaching a value of m 2 K/W at the end of the test. On the other hand, the fouling resistance for the PWT case remained around zero until t=70 h, and then began to rapidly increase, passing the industry standard allowance value at t=100 h. After this time, the fouling resistance increased much slowly, reaching Fig. 9 shows variations of fouling resistances over time for two cases using 20-μm sand filter: one with filter only, and the other with the simultaneous use of PWT and filter. The fouling resistance for the filter-only case almost linearly increased throughout the entire test period except a sharp drop seen briefly at the beginning of the test, passing the industry standard allowance value of m 2 K/W at the end of the test. For the PWT+filtration case, the fouling resistance steadily increased until t=70 h. However, it suddenly began to drop at t=70 h, reaching a value well below zero at t=85 h, indicating a significant removal of scales from the heat transfer tube during this period. Then, it began to increase steadily, reaching m 2 K/ at the end of the test, which represents 77% drop compared to the filter-only case. Photographs of fouled tubes in Fig. 10 show the difference in the heat transfer surfaces obtained for the above two cases. The filter-only case showed a number of relatively large CaCO 3 particles across the copper tube, whereas the PWT+filtration Fig. 5. Fouling resistance over time for 5 COC with a 5-μm fabric filter. Fig. 7. Fouling resistances over time for 8 COC without filtration.

5 1012 W. Kim, Y.I. Cho / International Communications in Heat and Mass Transfer 38 (2011) Fig. 8. Photographs of copper tubes for 8 COC at the end of fouling test, (a) no treatment and (b) PWT, both without filtration. Fig. 10. Photographs of copper tubes for 8 COC at the end of fouling test, (a) sand filter only and (b) PWT+sand filter. case showed no such particles, but almost a negligibly thin layer of scale. 4. Discussion The fouling resistances for the PWT case were significantly smaller (i.e., 90% less) than those for the no-treatment case, both without filtration. This can be explained as follows: for the PWT case, the dissolved calcium ions precipitate from the supersaturated cooling water into seed crystals in bulk water, a process that can be described as bulk precipitation. The seed crystals grow in size due to the adhesion of adjacent mineral ions, becoming large-size calcium particles which tend to adhere to the heat transfer surface in the form of particulate fouling [4,19], which can be relatively easily removed by flow shear forces. The flow velocity at the test section in the present study was at 1.1 m/s, which may be large enough to remove most of the large particles deposited on the surface through the particulate fouling. The steep increase rate of the fouling resistance observed after t=40 h for the no-treatment case (i.e., see Fig. 3) suggests that due to the supersaturated conditions of cooling water, the deposition rate of scales must be much greater than the removal rate by shear force. Compared to the no-treatment case, the PWT case must have had a Fig. 9. Fouling resistances over time for 8 COC with a sand filter (20-μm size). much greater removal rate due to the formation of particulate fouling as the deposition rates for the two cases should be about the same. Negative fouling resistances observed at the beginning of the tests could be attributed to an increased surface roughness from the deposition of CaCO 3 particles, thus enhancing convective heat transfer rate [20]. The enhanced heat transfer performance was maintained during the entire test period for the PWT cases particularly for the PWT+filtration case, a phenomenon which can be attributed to periodic adhesion and removal of large CaCO 3 particles on the heat transfer surface. In order to assess the benefit offiltration in the use of PWT for cooling water management, one can compare the present results with the industry standard fouling allowance value of m 2 K/W [18]. The results of the fouling curve obtained with no treatment for COC of 5 without filtration was significantly above this level as shown in Fig. 3. However, the PWT case for COC of 5 without filtration was substantially below this level (see Fig. 3). The results given in Fig. 5 depict that the filtration alone can bring the fouling curve well below this industry allowance level for 5-COC Case, while the combined use of PWT and filtration further reduced the fouling curve. For the case of COC of 8, the fouling curves for both no treatment and PWT cases without filtration were above the industry allowance level at the end of the tests (see Fig. 7), probably due to the extremely high hardness of cooling water, i.e., ppm. For the case of the filtration alone, the fouling curve was slightly above the industry allowance level, while the PWT+filtration gave the fouling curve well below the allowance level. Thus, the present study clearly demonstrated the benefits of the filtration and the combined use of PWT and filtration in the mitigation of mineral fouling. Although filtration is known to be useful in the cooling water management in general, the filtration is not widely used because of the need of frequent replacements of filter and associated costs. Alternatively, a periodic back-wash could be used to clean clogged filters. However, the back-wash system requires complex piping arrangement as well as clean filtered water for the operation. Furthermore, in most cooling water systems, the size of pipe diameters is relatively large due to large volumes of cooling water. Hence, it is often economically not feasible to install a back-wash system. In this regard, Yong et al. [21] recently introduced a selfcleaning filter utilizing plasma spark discharge directly in water. With such a filter, the benefit of the combined use of PWT and filtration could be sustained almost indefinitely without having to replace filter. For the no-treatment cases, the fouling resistance obtained for COC of 5 was much greater than that obtained for COC of 8 (see Figs. 3 and 7), a finding which is counter-intuitive. The authors previously

6 W. Kim, Y.I. Cho / International Communications in Heat and Mass Transfer 38 (2011) attributed this peculiar result to the extremely supersaturated state of COC of 8, where calcium ions spontaneously precipitate in bulk water [12]. As a result, particulate fouling could form on the heat transfer surface for the case of COC of 8, thus forming a relatively soft sludge coating of calcium particles compared to the case of COC of Conclusions The purpose of the present study was to investigate the benefit of filtration in the use of PWT for the mitigation of mineral fouling in a double-pipe heat exchanger. Fouling resistances were obtained over time for no-treatment, PWT, and PWT+filtration cases at two different COCs of 5 and 8. Fouling resistances for the PWT case were significantly less compared to those for the no-treatment. However, with increasing COC, the improvement in the fouling mitigation with PWT was less. When filtration was used in conjunction with the PWT, a substantially better fouling mitigation was obtained as the filtration removed the mineral particles that were produced by the PWT. Acknowledgement The work was partially funded by the California Energy Commission, the EISG Program, Grant number: # References [1] J.C. Cowan, D.J. Weintritt, Water-formed scale deposits, Gulf Publishing Company, Houston, TX, 1976, pp [2] A. Helalizadeh, H. Muller-Steinhagen, M. Jamialahmadi, Mixed salt crystallization fouling, Chem. Eng. Process. 39 (2000) [3] D. Hasson, H. Sherman, M. Biton, Prediction of calcium carbonate scaling rates, Proceedings 6th International Symposium Fresh Water from the Sea, 2, 1978, pp [4] N.H. Kim, R.L. Webb, Particulate fouling of water in tubes having a two-dimensional roughness geometry, Int. J. Heat Mass Transfer 34 (11) (1991) [5] E.F.C. Somerscales, A.F. Pontedure, A.E. Bergles, Particulate fouling of heat transfer tubes enhanced on their inner surface, HTD-Vol. 164, Fouling and Enhancement Interactions, ASME, 1991, p. 17. [6] S. Nasrazadani, T.J. Chao, Laboratory evaluations of ozone as a scale inhibitor for use in open recirculating cooling systems, ASHRAE Research Project 765-RP, [7] R. Sheikholeslami, A.P. Watkinson, Scaling of Plain and Externally Finned Heat Exchanger Tubes, J. Heat Transfer 108 (2) (1986) [8] R.W. Morse, J.G. Knudsen, Effect of alkalinity on the scaling of simulated cooling tower water, Can. J. Chem. Eng. 55 (3) (1977) [9] T.M. Paakkonen, M. Riihimaki, E. Puhakka, E. Muurinen, C.J. Simonson, R.L. Keiski, Crystallization fouling of CaCO 3 effect of bulk precipitation on mass deposition on the heat transfer surface, Proceedings of International Conference on Heat Exchanger Fouling and Cleaning 2009 (ed. By H. Muller-Steinhagen et al), Schladming, Austria, June 14 19, 2009, 2009, pp [10] L.D. Tijing, B.C. Pak, D.H. Lee, Y.I. Cho, Heat-treated titanium balls for the mitigation of mineral fouling, Exp. Heat Transfer 21 (2) (2007) [11] L.D. Tijing, H.Y. Kim, D.H. Lee, C.S. Kim, Y.I. Cho, Physical water treatment using RF electric fields for the mitigation of CaCO 3 fouling in cooling water, Int. J. Heat Mass Transfer 53 (2010) [12] Y.I. Cho, S.H. Lee, W.T. Kim, Physical water treatment for the mitigation of mineral fouling in cooling-towerwater applications, ASHRAE Trans. 109 (1) (2003) [13] Y.I. Cho, A. Fridman, W. Kim, S. Lee, Physical Water Treatment for fouling prevention in heat exchangers, Advances in Heat Transfer, 38, Academic Press, 2004, pp [14] R.A. Serway, Physics for Scientists and Engineers, 6th Ed., Saunders College Publishing, Philadelphia, PA, 2001, pp [15] F.P. Incropera, D.P. DeWitt, Fundamentals of Heat and Mass Transfer, 6th ed., John Wiley and Sons, New York, 2007, pp [16] S.J. Kline, F.A. McClintock, Describing uncertainties in single-sample experiments, Mech. Eng. (Jan. 1953) 3. [17] W.T. Kim, A study of physical water treatment methods for the mitigation of mineral fouling, Ph.D. Thesis, Drexel University, 2001, Philadelphia, PA. [18] Y.I. Cho, J. Lane, W.T. Kim, Pulsed-power treatment for physical water treatment, Int. Commun. Heat Mass Transfer 32 (2005) [19] T.R. Bott, The fouling of heat exchangers, Elsevier Science, New York, [20] F. Albert, W. Augustin, S. Scholl, Enhancement of heat transfer in crystallization fouling due to surface roughness, Proceedings of Heat Exchanger Fouling and Cleaning 2009, Schladming, Austria, June 14 19, 2009, [21] Y. Yang, H.S. Kim, A. Fridman, Y.I. Cho, Effect of a plasma-assisted self-cleaning filter on the performance of PWT coil for the mitigation of mineral fouling in a heat exchanger, Int. J. Heat Mass Transfer 53 (2010)