- Email: [email protected]
Experimental observation of surface morphology effect on crystallization fouling in plate heat exchangers
International Communications in Heat and Mass Transfer 38 (2011) 25–30
Contents lists available at ScienceDirect
International Communications in Heat and Mass Transfer j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i c h m t
Experimental observation of surface morphology effect on crystallization fouling in plate heat exchangers☆ Chengwang Lei a,⁎, Zhongxiao Peng b, Thomas Day b, Xinping Yan c, Xiuqin Bai c, Chengqing Yuan c a b c
School of Civil Engineering, The University of Sydney, Sydney, NSW 2006, Australia School of Engineering and Physical Sciences, James Cook University, Australia Reliability Engineering Institute, Wuhan University of Technology, China
a r t i c l e
i n f o
Available online 18 November 2010 Keywords: Crystallization fouling Surface texture Antifouling Heat exchanger
a b s t r a c t In this study, stainless steel test plates with different surface roughness and textures, which are used as the heat transfer surface of a plate heat exchanger, are tested individually in calcium carbonate fouling experiments. The present experimental results clearly indicate a strong correlation between the surface roughness and the amount of crystallization fouling deposit. Through detailed image analysis, four stages of the formation of crystallization fouling are identified, and the impact of the surface morphology on the extent of crystallization fouling is described qualitatively. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction Fouling is the accumulation of an unwanted layer of scale on a surface submerging in a fluid medium. Six different forms of fouling may occur in industrial applications, which include crystallization, particulate, chemical, corrosion, biological and solidification fouling [1–3]. Fouling is a highly detrimental process which has a significant negative impact on the performance of industrial systems. In industrialized countries, fouling costs have been estimated to be as large as 0.25% of the country's gross national product [3]. Therefore, significant effort has been devoted to fouling prevention and mitigation as well as the development of antifouling techniques. Among the six different forms of fouling, crystallization fouling, which is common in industrial heat exchangers, is responsible for approximately 25% of the scale related issues. Inversely soluble salts such as calcium sulphate and calcium carbonate have been recognized as the most prolific forms of crystalline scale deposit in heat exchangers [1]. Crystallization fouling involves three basic fouling phases [2,3]. The first of the three phases of crystallization is called ‘attainment of supersaturation’. Following supersaturation, nanometre scale clusters of dissolved salts then begin to form as nuclei, which become stable once they mature to a critical size. This phase is referred to as the ‘formation of nuclei’. The final phase of crystallization is called ‘growth of crystals’, which involves subsequent development of a variety of larger, variably sized crystal shapes. Bansal et al. [3] revealed five circumstances under which supersaturation, and hence crystallization fouling can occur. In addition, ☆ Communicated by W.J. Minkowycz. ⁎ Corresponding author. E-mail address: [email protected] (C. Lei). 0735-1933/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.icheatmasstransfer.2010.10.006
crystallization fouling is affected by factors such as flow velocity, heat and mass transfer, chemical composition of the fluid medium, and material properties of the surface, etc. [4]. The occurrence of crystallization fouling may dramatically reduce the performance of industrial heat exchangers. On one hand, the poor thermal conductivity of the scale deposit slows down the heat transfer rate through the fouled surface. On the other hand, the presence of the unwanted layer of scale deposit in flow channels produces an additional resistance to the flow and increases the pressure drop across the heat exchangers and thus the pumping power of the industrial processes. Accordingly, industrial heat exchangers are commonly overdesigned in order to compensate the loss of efficiency due to crystallization fouling [2]. This method of dealing with fouling is inefficient and uneconomical, and expensive cleaning and removal procedures are desirable at routine intervals. Plate heat exchangers are among the most cost-effective types of heat transfer devices. They are comprised of flat, corrugated or finned plate-like heat transfer surfaces, which separate hot and cold process fluids, and they can be designed with single or multi-pass configurations in parallel or counter flow arrangements. The plates are generally designed as large as possible to maximize the heat transfer area. Plate heat exchangers are used for a wide range of applications, including food processing, space heating, refrigeration, central cooling systems, automotive applications, power plants, chemical plants, petrochemical plants, petroleum refineries and natural gas processing [5]. However, once fouled, plate heat exchangers become highly unreliable compared to alternative designs. This highlights the need for further study into the fouling of plate heat exchangers. In the field of biofouling, extensive studies have explored the effects of various microtopographic surface parameters in marine environments, and green and biomemetic antifouling techniques
26
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
have been developed. A recent study of Scardino et al. [6] has revealed distinct correlations between certain surface parameters and the antibiofouling properties of the surface. Their experiments show that variations in parameters such as fractal dimension, surface roughness, waviness, skewness and texture aspect ratio, can have a strong influence on the rate and degree of biofouling. Despite that extensive studies have also been undertaken in the field of crystallization fouling, research in this field has not explored the relationship between fouling and surface micro textures to the degree of detail as in biofouling research, and so far mainly the surface roughness effect on crystallization fouling has been reported [7,8]. This highlights the need for further study in this area. The present investigation addresses this issue. In this study, different surface textures and roughness are produced on stainless steel surfaces of a plate heat exchanger and tested against crystallization fouling. Qualitative observation of the formation of fouling on the plate heat exchanger is carried out based on image analysis. Details of the experiments and experimental results are described in the following sections.
exchanger surfaces. Calcium carbonate [CaCO3] and calcium sulphate [CaSO4] fouling have been identified as the most prevalent in industrial applications, and thus are the most commonly studied forms of fouling agents. Of these two forms of fouling, calcium carbonate fouling is more prevalent than calcium sulphate fouling, and is chosen as the fouling agent here. The calcium carbonate fouling additive may be introduced into the hot side solution in two different ways. One way involves directly dissolving calcium carbonate into the hot solution, and the other involves chemically mixing sodium bicarbonate [NaHCO3] and calcium chloride [CaCl2] in the heating tank to produce calcium carbonate [9]. The latter method can produce a significantly higher concentration of a calcium carbonate solution, which will produce faster crystallization fouling, and thus is adopted here. The highest percentage of supersaturation of calcium carbonate reported in the literature is 74% [9]. This value is also adopted for the present experiments.
2. Experimental details
The focus of this investigation is on the test plate of the plate heat exchanger assembly (refer to Fig. 1), and the purpose is to test the effects of surface textures and roughness under crystallization fouling. Since stainless steel is commonly used in plate heat exchangers owing to its strength and strong resistance to corrosion, AISI 316 stainless steel is selected as the test material in this study. This material has also been successfully tested by others [1,10]. Two types of surface textures and a range of surface roughness are produced on the stainless steel plates. The first type of surface texture is a mirror smooth surface, which has no particular texture pattern. The other type of surface texture has simple linear and parallel grooves on the surface. In order to prepare the stainless steel samples with the required surface textures and roughness, roughness testing is carried out prior to fouling experiments, which includes roughening the surface with sandpapers and measuring its roughness. This procedure is carried out using P1200, P600, P400, P320, P240 and P120 sandpapers and a polished surface respectively. A calibrated Mahr Perthometer M1, which is suitable for roughness measurements of most engineering surfaces, is used to measure the surface roughness of the produced samples. A trace length of 17.5 mm is selected for each measurement and a total of 3 parallel traces in the direction perpendicular to the surface textures are performed on the stainless steel test plates. The average roughness value of the three measurements is adopted as the final roughness. The results of the roughness testing are shown in Table 1.
2.1. Apparatus The fouling experiments are conducted on a plate heat exchanger apparatus, which includes a heating tank, a water pump, a plate heat exchanger assembly, a number of control valves, and flow rate and temperature control and measurement systems. Fig. 1 illustrates the configuration and working principle of the plate heat exchanger assembly. The dimensions of the heat exchanger assembly are 200 mm × 120 mm × 80 mm, and the dimensions of the heat transfer plate are 100 mm × 85 mm × 2 mm. Water in the heating tank is heated directly by electric heating coils, the operation of which is controlled by a thermostat. Hot water at a set temperature is pumped into the plate heat exchanger and circulated through the hot side of the exchanger. The main water supply from a water tap is circulated through the cold side of the heat exchanger and discharged into the drain afterwards. Two rotameters of different flow rate ranges are used to measure the flow rates of the hot- and cold-water flows respectively, and four K-type thermocouples are used to monitor the water temperatures at the inlets and outlets of the hot and cold sides of the plate heat exchanger respectively. The uncertainty of the temperature measurements by the thermocouples is ±0.75 °C over the operational temperature range after calibration. 2.2. Fouling agents As mentioned above, inversely soluble salts are responsible for crystallization fouling, which occurs mainly on the hot side of heat
Cold inlet T1
Hot outlet T3
2.3. Heat transfer surfaces
2.4. Fouling experiments Four experimental runs have been performed, including one with the polished surface and three with linearly textured surfaces of different roughness (refer to Table 1). All experiments are carried out with the hot-water temperature set to 70 °C, and the hot and coldwater flow rates are set to 200 L/h and 600 L/h respectively. Each Table 1 Roughness values of the stainless steel test surfaces.
Test plate
T2 Cold outlet
T4 Hot inlet
Fig. 1. Illustration of the plate heat exchanger assembly.
Sandpaper grade
Avg. grit size (μm)
Ra1 (μm)
Ra2 (μm)
Ra3 (μm)
Raavg. (μm)
Test code
Polished P1200 P600 P400 P320 P240 P120
0.079 0.122 0.188 0.314 0.513 1.211
0.084 0.1 0.193 0.311 0.485 1.278
0.07 0.115 0.18 0.335 0.542 1.295
0.073 0.078 0.112 0.187 0.32 0.513 1.261
1
15.3 25.8 35 46.2 58.5 125
2 3 4
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
experiment is run continuously for 168 hours (1 week). Before and after each experiment, the stainless steel test plate is weighed using a SALTER WEIGHT TRONIX HF-6000G electronic scale with an accuracy of ±0.05 g. Photographs of the test plates are also taken after each experiment. A cleaning procedure is applied before and after each experiment to remove the calcium chloride from the hot side of the experimental system.
3. Results The following sections describe the experimental results based on visual inspection and weighing of the test plate for each fouling experiment.
3.1. Polished control surface Fig. 2 shows two photographic images of the polished plate surface after the one-week fouling experiment. It is clear in Fig. 2(a) that the entire surface is covered with a thin layer of calcium carbonate fouling. The fouling deposit is weighed as 0.2 g in total. The majority of the test plate is covered with the off white calcium carbonate, which contrasts with the two darker side strips of the polished surface. The fouling appears in a grey colour, but is actually semitransparent, and its colour is influenced by the colour of the plate. The darker strips running along the sides of the plate are not fouled since they are encased in the recess of the heat exchanger assembly. There appears to be some variation in the grayness of the fouling which is likely to have been influenced by the localized flow variations within the heat exchanger. The relative thickness of the fouling layer to the thickness of the test plate (2 mm) can be seen in Fig. 2(b).
Fig. 2. Photography of the polished plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed view of the fouling.
27
3.2. P600 sandpaper The second fouling experiment is conducted on the stainless steel plate prepared using the P600 sandpaper. The plate seen in Fig. 3(a) is removed from the heat exchanger assembly after one week testing. As seen in the image, its surface is completely covered with fouling. The quantity of the fouling is substantially greater than that generated by the polished plate. The clean recess strips running along the edge of the plate highlight the difference between the unscaled and fouled surfaces. The fouling is rougher and less organised in the middle of the plate and flatter and more consistent toward the edges. At the inlet (bottom right), outlet (bottom left) and the top right corner, the fouling is visibly thinner and smoother. These flattened regions are likely to have been produced through a localized increase in the flow rate and turbulence. The surface acquires 0.7 g of fouling deposit in total, which is in a very light grey colour. Rust spots are evident on two sides of the plate. These are generated by the ingress of sikaflex onto the edges of the hot side of the heat transfer surface. Fig. 3(b) shows a zoomed view of the edge of the fouling taken at an angle to the surface. In this image, the thickness of the calcium carbonate layer can be compared with the thickness of the 2 mm plate. The growth of the crystal clusters appears to be spaced and wispy. The average thickness of the fouling is estimated using the weight of the fouling as 31 μm. Fig. 3(c) shows a zoomed top view of the fouling. The spaced nature of the fouling clusters is evident. Upon close inspection of the crystal formation, tiny, spindle shaped calcite
Fig. 3. Photography of the P600 plate after one week fouling experiment. (a) An overview of the fouled surface. (b) A zoomed side view showing the edge of the fouled plate. (c) A zoomed top view of the plate.
28
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
crystals are observed. The flatter section of fouling in the bottom right hand corner can also be clearly seen in this detailed image.
3.3. P400 sandpaper The P400 test plate has experienced a thin layer of initial fouling and then the growth and spreading of calcium carbonate crystals, which have developed from the initial layer. The fouling forming on this test plate can be seen in Fig. 4(a). The accumulated fouling has a total mass of 0.7 g. The initial fouling layer is in grey colour, covering a large portion of the plate, while the calcium carbonate crystals are the off white segments that cover the initial fouling layer. The figure shows that different sections of the plate have incurred differing fouling severities. In the centre of the plate, the fouling clusters are spaced, such that some regions encounter less than 50% cluster coverage. Along the sides of the plate, particularly at the top left and bottom right corners, the fouling clusters are less spaced. In these
regions the clusters almost completely cover the entire initial fouling layer. The progression of the fouling crystal clusters has not progressed to the very top right hand corner of the test plate after one week testing. The hot-water inlet of the plate heat exchanger is positioned near the bottom right corner of the plate in Fig. 4(a), indicated by the upward pointing arrow, and the hot-water outlet is positioned near the bottom left of the plate. This orientation suggests that the momentum of the inlet flow may be responsible for the region containing no crystal clusters at the top right corner of the plate. During the test, the hot solution passes through the inlet, travels upwards across the plate and reflects off the far side of the heat exchanger. This is observed to produce increased turbulence in the top-right region, which may be responsible for the decreased crystal growth in this region. The fouling clusters are also examined in more detail by zooming into the surface. Fig. 4(b) is taken at an angle to the fouled surface to highlight the height of the fouling clusters. The clusters shown in the figure are from the centre of the plate, where they are more spaced and easier to examine. This photo also illustrates the smooth nature of the initial fouling layer. There does not appear to be any specific pattern related to the location of the crystal formation. Fig. 4(c) and (d) show the crystal clusters formed in the centre of the plate and in the top left corner respectively. The photos show that the crystals have long spindle-like structure, which includes numerous spindles extending from centralized anchor points. These crystals are consistent with the aragonite polymorph of calcium carbonate. Comparing these two images also highlights the difference in crystal cluster density in different regions over the plate surface. 3.4. P320 sandpaper The P320 test plate has experienced severe fouling after one week testing. Fig. 5(a) shows that the fouling crystal clusters completely cover the entire heat exchange surface. The density of the fouling clusters is so severe that the initial fouling layer cannot be identified at any location on the plate. This test has produced in total 1.1 g of fouling over the 1-week period. The crystal structures present on the surface vary in different locations. Those formed near the top right hand corner of the plate appear to be relatively flat, while those at approximately one third of the way up from the bottom of the image contain more irregularity. Fig. 5(b) shows a close view of the fouling on the plate. The crystal structure is flat and clumped. However, upon a closer inspection it is revealed that the clumps have been formed by tiny spindle shaped crystals. This shape corresponds to the aragonite polymorph of calcium carbonate. There appears to be some direction associated with the growth of the crystals, which is likely to have been influenced by the flow direction. Fig. 5(c) is a side-view image of the fouled plate, which shows the thickness of the calcium carbonate scale layer relative to the 2-mm thick plate. The average thickness of the scale layer is estimated as 48 μm, which is significantly thicker than the scales formed on the other test plates. It is seen in Fig. 5(c) that the crystal structure near the inlet, which is in the right hand corner of the plate in this figure, is flatter and contains smaller crystals compared with the crystal structure in regions further away from the inlet. This difference in the crystal structure can be attributed to the influence of the flow condition on the crystal growth. 4. Discussion
Fig. 4. Photography of the P400 plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed and angled view of the fouled surface. (c) Spaced fouling clusters in the center of the plate. (d) Dense fouling clusters near the top-left corner.
Fig. 6 plots the mass of the accumulated fouling against the average roughness of the test surface. With only four data points, it is difficult to obtain a definitive correlation between the fouling deposit
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
29
surface roughness, consistent with the findings in [7]. An exception is observed for the P600 plate, which has developed more fouling than would be expected (see Fig. 6). Minute surface pitting may be responsible for the misaligned results (refer to Section 4.4 below).
4.2. Surface coverage The degree of visible fouling coverage achieved for each test is useful for defining the different stages of fouling growth. The following four stages of fouling growth have been identified: 1. Complete surface coverage with a thin non-crystalline layer of fouling. 2. Growth of minute calcium carbonate nuclei at spaced locations.
Fig. 5. Photography of the P320 plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed-in view of fouling clusters. (c) A side view of the plate showing the relative thickness of the fouling layer to the plate.
and the surface roughness. However, some general description of the result may be given as follows. 4.1. Roughness effect The roughest plate, which is sanded using P320 grade sandpapers, has developed the most fouling (1.1 g), while the smoothest surface, or the control surface, has recorded the lowest amount of fouling (0.2 g). These are confirmed by visual inspections. The result of the P400 plate (0.7 g) aligns very well with the results of the control surface and the P320 plate (indicated by the dashed line in Fig. 6). These results show that the fouling rate generally increases with the
Fig. 6. Fouling weight versus surface roughness.
Fig. 7. Microscopic images (100×) showing surface pitting on the P320 plate before the fouling experiment. (a), (b) and (c) are taken at different spots on the surface.
30
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
3. Rapid crystal growth originating from the nuclei and producing clumped crystal structures. 4. Complete surface coverage by interconnected crystal structures. After one week testing the polished plate has achieved only the first two stages of fouling, making it the slowest to progress. The P400 plate has reached the third stage of fouling, but complete surface coverage by interconnected crystal structures has not been achieved. The P600 and P320 plates have both experienced all four stages of fouling and a growth of the crystal fouling layer. This indicates that the P320 and P600 plates produce the most rapid fouling, followed by the P400 plate and the polished plate respectively. The present experiments have demonstrated that in general a rougher surface progresses through the four fouling stages at a faster rate. However, the comparison between the P400 and P600 plates contradicts this general trend. Again, minute surface pitting may be responsible for this variation (refer to Section 4.4). 4.3. Crystal size Aragonite crystal growth is observed on all surfaces except for the polished surface. The size of the aragonite crystals differs between the plates. The spindle shaped crystals are significantly smaller for the P600 and P320 plates, which have achieved complete surface coverage, compared to the crystals formed on the P400 plate. Again, minute surface pitting may be responsible for this effect (see the discussion below). 4.4. Surface pitting In the above discussion, minute surface pitting has been considered as being a possible cause of inconsistent fouling growth on three occasions. Surface pitting was visible to the naked eyes under direct sunlight, but was difficult to capture using the ordinary digital camera. Despite that effort has been made to remove all surface pitting, this cannot be fully achieved in the present experiment. Fig. 7 presents three microscopic images of the P320 test plate surface, which clearly demonstrate the presence and severity of surface pitting. A close inspection of the four plate surfaces has shown that surface pitting is more prevalent on the P320 and P600 test surfaces than the P400 plate, and the polished test surface has the least surface pitting after multiple polishing processes. During the experiments, the tiny holes in the surface (surface pitting) provide many additional nucleation points for the fouling to grow from, and thus more crystal clusters would form and cover the surface in a shorter time. Accordingly, the overall weight of the accumulated fouling, the aragonite crystal size and the distribution rate of the fouling are all affected by the surface pitting.
5. Summary The objective of this study is to determine the effects of surface roughness and textures on calcium carbonate crystallization fouling. Three stainless steel surfaces of the same linear and parallel textures but of different surface roughness, along with a mirror finish control surface for comparison purpose, are tested for a week under crystallization fouling conditions. The test plate was visually inspected and weighed before and after the experiment. Based on visual comparison of fouled surfaces with different surface textures, the present experiments have clearly demonstrated that the growth rate, the distribution and the crystal size of calcium carbonate fouling are strongly dependent on the surface texture and finish. It is found that the rate of crystallization fouling generally increases with the surface roughness and the quantity of surface pitting. Under the present experimental conditions, the mirror smooth heat transfer plate represents the best antifouling surface.
Acknowledgements The authors would like to acknowledge Mr. Peter Friedrich for designing and constructing the testing apparatus. Thanks also goes to the Mechanical Engineering Workshop in the School of Engineering and Physical Sciences, James Cook University for its technical support.
References [1] N. Andritsos, A.J. Karabelas, Calcium carbonate scaling in a plate heat exchanger in the presence of particles, International Journal of Heat and Mass Transfer 46 (2003) 4613–4627. [2] M.G. Mwaba, M.R. Golriz, J. Gu, A semi-empirical correlation for crystallization fouling on heat exchange surfaces, Applied Thermal Engineering 26 (2005) 440–447. [3] B. Bansal, X.D. Chen, H. Müller-Steinhagen, Analysis of ‘classical’ deposition rate law for crystallisation fouling, Chemical Engineering and Processing 47 (2008) 1201–1210. [4] A. Helalizadeh, H. Müller-Steinhagen, M. Jamialahmadi, Mixed salt crystallisation fouling, Chemical Engineering and Processing 39 (2000) 29–43. [5] J.A.W. Gut, J.M. Pinto, Optimal configuration design for plate heat exchangers, International Journal of Heat and Mass Transfer 47 (2004) 4833–4848. [6] A.J. Scardino, D. Hudleston, Z. Peng, N.A. Paul, R. de Nys, Biomimetic characterization of key surface parameters for the development of fouling resistant materials, Biofouling 25 (2009) 83–93. [7] S. Keysar, R. Semiat, D. Hasson, J. Yahalom, Effect of surface roughness on the morphology of calcite crystallizing on mild steel, Journal of Colloid and Interface Science 162 (1994) 311–319. [8] A. Herz, M.R. Malayeri, H. Müller-Steinhagen, Fouling of roughened stainless steel surfaces during convective heat transfer to aqueous solutions, Energy Conversion and Management 49 (2008) 3381–3386. [9] Z. Quan, Y. Chen, C. Ma, Experimental study of fouling on heat transfer surface during forced convective heat transfer, Chinese Journal of Chemical Engineering 16 (2008) 535–540. [10] D.J. Kukulka, M. Devgun, Fouling surface finish evaluation, Applied Thermal Engineering 27 (2007) 1165–1172.
Contents lists available at ScienceDirect
International Communications in Heat and Mass Transfer j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i c h m t
Experimental observation of surface morphology effect on crystallization fouling in plate heat exchangers☆ Chengwang Lei a,⁎, Zhongxiao Peng b, Thomas Day b, Xinping Yan c, Xiuqin Bai c, Chengqing Yuan c a b c
School of Civil Engineering, The University of Sydney, Sydney, NSW 2006, Australia School of Engineering and Physical Sciences, James Cook University, Australia Reliability Engineering Institute, Wuhan University of Technology, China
a r t i c l e
i n f o
Available online 18 November 2010 Keywords: Crystallization fouling Surface texture Antifouling Heat exchanger
a b s t r a c t In this study, stainless steel test plates with different surface roughness and textures, which are used as the heat transfer surface of a plate heat exchanger, are tested individually in calcium carbonate fouling experiments. The present experimental results clearly indicate a strong correlation between the surface roughness and the amount of crystallization fouling deposit. Through detailed image analysis, four stages of the formation of crystallization fouling are identified, and the impact of the surface morphology on the extent of crystallization fouling is described qualitatively. © 2010 Elsevier Ltd. All rights reserved.
1. Introduction Fouling is the accumulation of an unwanted layer of scale on a surface submerging in a fluid medium. Six different forms of fouling may occur in industrial applications, which include crystallization, particulate, chemical, corrosion, biological and solidification fouling [1–3]. Fouling is a highly detrimental process which has a significant negative impact on the performance of industrial systems. In industrialized countries, fouling costs have been estimated to be as large as 0.25% of the country's gross national product [3]. Therefore, significant effort has been devoted to fouling prevention and mitigation as well as the development of antifouling techniques. Among the six different forms of fouling, crystallization fouling, which is common in industrial heat exchangers, is responsible for approximately 25% of the scale related issues. Inversely soluble salts such as calcium sulphate and calcium carbonate have been recognized as the most prolific forms of crystalline scale deposit in heat exchangers [1]. Crystallization fouling involves three basic fouling phases [2,3]. The first of the three phases of crystallization is called ‘attainment of supersaturation’. Following supersaturation, nanometre scale clusters of dissolved salts then begin to form as nuclei, which become stable once they mature to a critical size. This phase is referred to as the ‘formation of nuclei’. The final phase of crystallization is called ‘growth of crystals’, which involves subsequent development of a variety of larger, variably sized crystal shapes. Bansal et al. [3] revealed five circumstances under which supersaturation, and hence crystallization fouling can occur. In addition, ☆ Communicated by W.J. Minkowycz. ⁎ Corresponding author. E-mail address: [email protected] (C. Lei). 0735-1933/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.icheatmasstransfer.2010.10.006
crystallization fouling is affected by factors such as flow velocity, heat and mass transfer, chemical composition of the fluid medium, and material properties of the surface, etc. [4]. The occurrence of crystallization fouling may dramatically reduce the performance of industrial heat exchangers. On one hand, the poor thermal conductivity of the scale deposit slows down the heat transfer rate through the fouled surface. On the other hand, the presence of the unwanted layer of scale deposit in flow channels produces an additional resistance to the flow and increases the pressure drop across the heat exchangers and thus the pumping power of the industrial processes. Accordingly, industrial heat exchangers are commonly overdesigned in order to compensate the loss of efficiency due to crystallization fouling [2]. This method of dealing with fouling is inefficient and uneconomical, and expensive cleaning and removal procedures are desirable at routine intervals. Plate heat exchangers are among the most cost-effective types of heat transfer devices. They are comprised of flat, corrugated or finned plate-like heat transfer surfaces, which separate hot and cold process fluids, and they can be designed with single or multi-pass configurations in parallel or counter flow arrangements. The plates are generally designed as large as possible to maximize the heat transfer area. Plate heat exchangers are used for a wide range of applications, including food processing, space heating, refrigeration, central cooling systems, automotive applications, power plants, chemical plants, petrochemical plants, petroleum refineries and natural gas processing [5]. However, once fouled, plate heat exchangers become highly unreliable compared to alternative designs. This highlights the need for further study into the fouling of plate heat exchangers. In the field of biofouling, extensive studies have explored the effects of various microtopographic surface parameters in marine environments, and green and biomemetic antifouling techniques
26
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
have been developed. A recent study of Scardino et al. [6] has revealed distinct correlations between certain surface parameters and the antibiofouling properties of the surface. Their experiments show that variations in parameters such as fractal dimension, surface roughness, waviness, skewness and texture aspect ratio, can have a strong influence on the rate and degree of biofouling. Despite that extensive studies have also been undertaken in the field of crystallization fouling, research in this field has not explored the relationship between fouling and surface micro textures to the degree of detail as in biofouling research, and so far mainly the surface roughness effect on crystallization fouling has been reported [7,8]. This highlights the need for further study in this area. The present investigation addresses this issue. In this study, different surface textures and roughness are produced on stainless steel surfaces of a plate heat exchanger and tested against crystallization fouling. Qualitative observation of the formation of fouling on the plate heat exchanger is carried out based on image analysis. Details of the experiments and experimental results are described in the following sections.
exchanger surfaces. Calcium carbonate [CaCO3] and calcium sulphate [CaSO4] fouling have been identified as the most prevalent in industrial applications, and thus are the most commonly studied forms of fouling agents. Of these two forms of fouling, calcium carbonate fouling is more prevalent than calcium sulphate fouling, and is chosen as the fouling agent here. The calcium carbonate fouling additive may be introduced into the hot side solution in two different ways. One way involves directly dissolving calcium carbonate into the hot solution, and the other involves chemically mixing sodium bicarbonate [NaHCO3] and calcium chloride [CaCl2] in the heating tank to produce calcium carbonate [9]. The latter method can produce a significantly higher concentration of a calcium carbonate solution, which will produce faster crystallization fouling, and thus is adopted here. The highest percentage of supersaturation of calcium carbonate reported in the literature is 74% [9]. This value is also adopted for the present experiments.
2. Experimental details
The focus of this investigation is on the test plate of the plate heat exchanger assembly (refer to Fig. 1), and the purpose is to test the effects of surface textures and roughness under crystallization fouling. Since stainless steel is commonly used in plate heat exchangers owing to its strength and strong resistance to corrosion, AISI 316 stainless steel is selected as the test material in this study. This material has also been successfully tested by others [1,10]. Two types of surface textures and a range of surface roughness are produced on the stainless steel plates. The first type of surface texture is a mirror smooth surface, which has no particular texture pattern. The other type of surface texture has simple linear and parallel grooves on the surface. In order to prepare the stainless steel samples with the required surface textures and roughness, roughness testing is carried out prior to fouling experiments, which includes roughening the surface with sandpapers and measuring its roughness. This procedure is carried out using P1200, P600, P400, P320, P240 and P120 sandpapers and a polished surface respectively. A calibrated Mahr Perthometer M1, which is suitable for roughness measurements of most engineering surfaces, is used to measure the surface roughness of the produced samples. A trace length of 17.5 mm is selected for each measurement and a total of 3 parallel traces in the direction perpendicular to the surface textures are performed on the stainless steel test plates. The average roughness value of the three measurements is adopted as the final roughness. The results of the roughness testing are shown in Table 1.
2.1. Apparatus The fouling experiments are conducted on a plate heat exchanger apparatus, which includes a heating tank, a water pump, a plate heat exchanger assembly, a number of control valves, and flow rate and temperature control and measurement systems. Fig. 1 illustrates the configuration and working principle of the plate heat exchanger assembly. The dimensions of the heat exchanger assembly are 200 mm × 120 mm × 80 mm, and the dimensions of the heat transfer plate are 100 mm × 85 mm × 2 mm. Water in the heating tank is heated directly by electric heating coils, the operation of which is controlled by a thermostat. Hot water at a set temperature is pumped into the plate heat exchanger and circulated through the hot side of the exchanger. The main water supply from a water tap is circulated through the cold side of the heat exchanger and discharged into the drain afterwards. Two rotameters of different flow rate ranges are used to measure the flow rates of the hot- and cold-water flows respectively, and four K-type thermocouples are used to monitor the water temperatures at the inlets and outlets of the hot and cold sides of the plate heat exchanger respectively. The uncertainty of the temperature measurements by the thermocouples is ±0.75 °C over the operational temperature range after calibration. 2.2. Fouling agents As mentioned above, inversely soluble salts are responsible for crystallization fouling, which occurs mainly on the hot side of heat
Cold inlet T1
Hot outlet T3
2.3. Heat transfer surfaces
2.4. Fouling experiments Four experimental runs have been performed, including one with the polished surface and three with linearly textured surfaces of different roughness (refer to Table 1). All experiments are carried out with the hot-water temperature set to 70 °C, and the hot and coldwater flow rates are set to 200 L/h and 600 L/h respectively. Each Table 1 Roughness values of the stainless steel test surfaces.
Test plate
T2 Cold outlet
T4 Hot inlet
Fig. 1. Illustration of the plate heat exchanger assembly.
Sandpaper grade
Avg. grit size (μm)
Ra1 (μm)
Ra2 (μm)
Ra3 (μm)
Raavg. (μm)
Test code
Polished P1200 P600 P400 P320 P240 P120
0.079 0.122 0.188 0.314 0.513 1.211
0.084 0.1 0.193 0.311 0.485 1.278
0.07 0.115 0.18 0.335 0.542 1.295
0.073 0.078 0.112 0.187 0.32 0.513 1.261
1
15.3 25.8 35 46.2 58.5 125
2 3 4
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
experiment is run continuously for 168 hours (1 week). Before and after each experiment, the stainless steel test plate is weighed using a SALTER WEIGHT TRONIX HF-6000G electronic scale with an accuracy of ±0.05 g. Photographs of the test plates are also taken after each experiment. A cleaning procedure is applied before and after each experiment to remove the calcium chloride from the hot side of the experimental system.
3. Results The following sections describe the experimental results based on visual inspection and weighing of the test plate for each fouling experiment.
3.1. Polished control surface Fig. 2 shows two photographic images of the polished plate surface after the one-week fouling experiment. It is clear in Fig. 2(a) that the entire surface is covered with a thin layer of calcium carbonate fouling. The fouling deposit is weighed as 0.2 g in total. The majority of the test plate is covered with the off white calcium carbonate, which contrasts with the two darker side strips of the polished surface. The fouling appears in a grey colour, but is actually semitransparent, and its colour is influenced by the colour of the plate. The darker strips running along the sides of the plate are not fouled since they are encased in the recess of the heat exchanger assembly. There appears to be some variation in the grayness of the fouling which is likely to have been influenced by the localized flow variations within the heat exchanger. The relative thickness of the fouling layer to the thickness of the test plate (2 mm) can be seen in Fig. 2(b).
Fig. 2. Photography of the polished plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed view of the fouling.
27
3.2. P600 sandpaper The second fouling experiment is conducted on the stainless steel plate prepared using the P600 sandpaper. The plate seen in Fig. 3(a) is removed from the heat exchanger assembly after one week testing. As seen in the image, its surface is completely covered with fouling. The quantity of the fouling is substantially greater than that generated by the polished plate. The clean recess strips running along the edge of the plate highlight the difference between the unscaled and fouled surfaces. The fouling is rougher and less organised in the middle of the plate and flatter and more consistent toward the edges. At the inlet (bottom right), outlet (bottom left) and the top right corner, the fouling is visibly thinner and smoother. These flattened regions are likely to have been produced through a localized increase in the flow rate and turbulence. The surface acquires 0.7 g of fouling deposit in total, which is in a very light grey colour. Rust spots are evident on two sides of the plate. These are generated by the ingress of sikaflex onto the edges of the hot side of the heat transfer surface. Fig. 3(b) shows a zoomed view of the edge of the fouling taken at an angle to the surface. In this image, the thickness of the calcium carbonate layer can be compared with the thickness of the 2 mm plate. The growth of the crystal clusters appears to be spaced and wispy. The average thickness of the fouling is estimated using the weight of the fouling as 31 μm. Fig. 3(c) shows a zoomed top view of the fouling. The spaced nature of the fouling clusters is evident. Upon close inspection of the crystal formation, tiny, spindle shaped calcite
Fig. 3. Photography of the P600 plate after one week fouling experiment. (a) An overview of the fouled surface. (b) A zoomed side view showing the edge of the fouled plate. (c) A zoomed top view of the plate.
28
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
crystals are observed. The flatter section of fouling in the bottom right hand corner can also be clearly seen in this detailed image.
3.3. P400 sandpaper The P400 test plate has experienced a thin layer of initial fouling and then the growth and spreading of calcium carbonate crystals, which have developed from the initial layer. The fouling forming on this test plate can be seen in Fig. 4(a). The accumulated fouling has a total mass of 0.7 g. The initial fouling layer is in grey colour, covering a large portion of the plate, while the calcium carbonate crystals are the off white segments that cover the initial fouling layer. The figure shows that different sections of the plate have incurred differing fouling severities. In the centre of the plate, the fouling clusters are spaced, such that some regions encounter less than 50% cluster coverage. Along the sides of the plate, particularly at the top left and bottom right corners, the fouling clusters are less spaced. In these
regions the clusters almost completely cover the entire initial fouling layer. The progression of the fouling crystal clusters has not progressed to the very top right hand corner of the test plate after one week testing. The hot-water inlet of the plate heat exchanger is positioned near the bottom right corner of the plate in Fig. 4(a), indicated by the upward pointing arrow, and the hot-water outlet is positioned near the bottom left of the plate. This orientation suggests that the momentum of the inlet flow may be responsible for the region containing no crystal clusters at the top right corner of the plate. During the test, the hot solution passes through the inlet, travels upwards across the plate and reflects off the far side of the heat exchanger. This is observed to produce increased turbulence in the top-right region, which may be responsible for the decreased crystal growth in this region. The fouling clusters are also examined in more detail by zooming into the surface. Fig. 4(b) is taken at an angle to the fouled surface to highlight the height of the fouling clusters. The clusters shown in the figure are from the centre of the plate, where they are more spaced and easier to examine. This photo also illustrates the smooth nature of the initial fouling layer. There does not appear to be any specific pattern related to the location of the crystal formation. Fig. 4(c) and (d) show the crystal clusters formed in the centre of the plate and in the top left corner respectively. The photos show that the crystals have long spindle-like structure, which includes numerous spindles extending from centralized anchor points. These crystals are consistent with the aragonite polymorph of calcium carbonate. Comparing these two images also highlights the difference in crystal cluster density in different regions over the plate surface. 3.4. P320 sandpaper The P320 test plate has experienced severe fouling after one week testing. Fig. 5(a) shows that the fouling crystal clusters completely cover the entire heat exchange surface. The density of the fouling clusters is so severe that the initial fouling layer cannot be identified at any location on the plate. This test has produced in total 1.1 g of fouling over the 1-week period. The crystal structures present on the surface vary in different locations. Those formed near the top right hand corner of the plate appear to be relatively flat, while those at approximately one third of the way up from the bottom of the image contain more irregularity. Fig. 5(b) shows a close view of the fouling on the plate. The crystal structure is flat and clumped. However, upon a closer inspection it is revealed that the clumps have been formed by tiny spindle shaped crystals. This shape corresponds to the aragonite polymorph of calcium carbonate. There appears to be some direction associated with the growth of the crystals, which is likely to have been influenced by the flow direction. Fig. 5(c) is a side-view image of the fouled plate, which shows the thickness of the calcium carbonate scale layer relative to the 2-mm thick plate. The average thickness of the scale layer is estimated as 48 μm, which is significantly thicker than the scales formed on the other test plates. It is seen in Fig. 5(c) that the crystal structure near the inlet, which is in the right hand corner of the plate in this figure, is flatter and contains smaller crystals compared with the crystal structure in regions further away from the inlet. This difference in the crystal structure can be attributed to the influence of the flow condition on the crystal growth. 4. Discussion
Fig. 4. Photography of the P400 plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed and angled view of the fouled surface. (c) Spaced fouling clusters in the center of the plate. (d) Dense fouling clusters near the top-left corner.
Fig. 6 plots the mass of the accumulated fouling against the average roughness of the test surface. With only four data points, it is difficult to obtain a definitive correlation between the fouling deposit
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
29
surface roughness, consistent with the findings in [7]. An exception is observed for the P600 plate, which has developed more fouling than would be expected (see Fig. 6). Minute surface pitting may be responsible for the misaligned results (refer to Section 4.4 below).
4.2. Surface coverage The degree of visible fouling coverage achieved for each test is useful for defining the different stages of fouling growth. The following four stages of fouling growth have been identified: 1. Complete surface coverage with a thin non-crystalline layer of fouling. 2. Growth of minute calcium carbonate nuclei at spaced locations.
Fig. 5. Photography of the P320 plate after one week fouling experiment. (a) An overview of the fouled plate. (b) A zoomed-in view of fouling clusters. (c) A side view of the plate showing the relative thickness of the fouling layer to the plate.
and the surface roughness. However, some general description of the result may be given as follows. 4.1. Roughness effect The roughest plate, which is sanded using P320 grade sandpapers, has developed the most fouling (1.1 g), while the smoothest surface, or the control surface, has recorded the lowest amount of fouling (0.2 g). These are confirmed by visual inspections. The result of the P400 plate (0.7 g) aligns very well with the results of the control surface and the P320 plate (indicated by the dashed line in Fig. 6). These results show that the fouling rate generally increases with the
Fig. 6. Fouling weight versus surface roughness.
Fig. 7. Microscopic images (100×) showing surface pitting on the P320 plate before the fouling experiment. (a), (b) and (c) are taken at different spots on the surface.
30
C. Lei et al. / International Communications in Heat and Mass Transfer 38 (2011) 25–30
3. Rapid crystal growth originating from the nuclei and producing clumped crystal structures. 4. Complete surface coverage by interconnected crystal structures. After one week testing the polished plate has achieved only the first two stages of fouling, making it the slowest to progress. The P400 plate has reached the third stage of fouling, but complete surface coverage by interconnected crystal structures has not been achieved. The P600 and P320 plates have both experienced all four stages of fouling and a growth of the crystal fouling layer. This indicates that the P320 and P600 plates produce the most rapid fouling, followed by the P400 plate and the polished plate respectively. The present experiments have demonstrated that in general a rougher surface progresses through the four fouling stages at a faster rate. However, the comparison between the P400 and P600 plates contradicts this general trend. Again, minute surface pitting may be responsible for this variation (refer to Section 4.4). 4.3. Crystal size Aragonite crystal growth is observed on all surfaces except for the polished surface. The size of the aragonite crystals differs between the plates. The spindle shaped crystals are significantly smaller for the P600 and P320 plates, which have achieved complete surface coverage, compared to the crystals formed on the P400 plate. Again, minute surface pitting may be responsible for this effect (see the discussion below). 4.4. Surface pitting In the above discussion, minute surface pitting has been considered as being a possible cause of inconsistent fouling growth on three occasions. Surface pitting was visible to the naked eyes under direct sunlight, but was difficult to capture using the ordinary digital camera. Despite that effort has been made to remove all surface pitting, this cannot be fully achieved in the present experiment. Fig. 7 presents three microscopic images of the P320 test plate surface, which clearly demonstrate the presence and severity of surface pitting. A close inspection of the four plate surfaces has shown that surface pitting is more prevalent on the P320 and P600 test surfaces than the P400 plate, and the polished test surface has the least surface pitting after multiple polishing processes. During the experiments, the tiny holes in the surface (surface pitting) provide many additional nucleation points for the fouling to grow from, and thus more crystal clusters would form and cover the surface in a shorter time. Accordingly, the overall weight of the accumulated fouling, the aragonite crystal size and the distribution rate of the fouling are all affected by the surface pitting.
5. Summary The objective of this study is to determine the effects of surface roughness and textures on calcium carbonate crystallization fouling. Three stainless steel surfaces of the same linear and parallel textures but of different surface roughness, along with a mirror finish control surface for comparison purpose, are tested for a week under crystallization fouling conditions. The test plate was visually inspected and weighed before and after the experiment. Based on visual comparison of fouled surfaces with different surface textures, the present experiments have clearly demonstrated that the growth rate, the distribution and the crystal size of calcium carbonate fouling are strongly dependent on the surface texture and finish. It is found that the rate of crystallization fouling generally increases with the surface roughness and the quantity of surface pitting. Under the present experimental conditions, the mirror smooth heat transfer plate represents the best antifouling surface.
Acknowledgements The authors would like to acknowledge Mr. Peter Friedrich for designing and constructing the testing apparatus. Thanks also goes to the Mechanical Engineering Workshop in the School of Engineering and Physical Sciences, James Cook University for its technical support.
References [1] N. Andritsos, A.J. Karabelas, Calcium carbonate scaling in a plate heat exchanger in the presence of particles, International Journal of Heat and Mass Transfer 46 (2003) 4613–4627. [2] M.G. Mwaba, M.R. Golriz, J. Gu, A semi-empirical correlation for crystallization fouling on heat exchange surfaces, Applied Thermal Engineering 26 (2005) 440–447. [3] B. Bansal, X.D. Chen, H. Müller-Steinhagen, Analysis of ‘classical’ deposition rate law for crystallisation fouling, Chemical Engineering and Processing 47 (2008) 1201–1210. [4] A. Helalizadeh, H. Müller-Steinhagen, M. Jamialahmadi, Mixed salt crystallisation fouling, Chemical Engineering and Processing 39 (2000) 29–43. [5] J.A.W. Gut, J.M. Pinto, Optimal configuration design for plate heat exchangers, International Journal of Heat and Mass Transfer 47 (2004) 4833–4848. [6] A.J. Scardino, D. Hudleston, Z. Peng, N.A. Paul, R. de Nys, Biomimetic characterization of key surface parameters for the development of fouling resistant materials, Biofouling 25 (2009) 83–93. [7] S. Keysar, R. Semiat, D. Hasson, J. Yahalom, Effect of surface roughness on the morphology of calcite crystallizing on mild steel, Journal of Colloid and Interface Science 162 (1994) 311–319. [8] A. Herz, M.R. Malayeri, H. Müller-Steinhagen, Fouling of roughened stainless steel surfaces during convective heat transfer to aqueous solutions, Energy Conversion and Management 49 (2008) 3381–3386. [9] Z. Quan, Y. Chen, C. Ma, Experimental study of fouling on heat transfer surface during forced convective heat transfer, Chinese Journal of Chemical Engineering 16 (2008) 535–540. [10] D.J. Kukulka, M. Devgun, Fouling surface finish evaluation, Applied Thermal Engineering 27 (2007) 1165–1172.