Keynote Lecture: Aspects of Two-Phase Flow Distribution at Header-Channels Assembly

Transcription

1 Proceedings Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology Engineering Conerences International Year 005 Keynote Lecture: Aspects o Two-Phase Flow Distribution at Header-Channels Assembly S. Y. Lee J. K. Lee This paper is posted at ECI Digital Archives.

2 Lee and Lee: Proceedings o Fith International Conerence on Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Eds. R.K. Shah, M. Ishizuka, T.M. Rudy, and V.V. Wadekar, Engineering Conerences International, Hoboken, NJ, USA, September 005. CHE ASPECTS OF TWO-PHASE FLOW DISTRIBUTION AT HEADER-CHANNELS ASSEMBLY Sang Yong Lee 1 and Jun Kyoung Lee Department o Mechanical Engineering, Korea Advanced Institute o Science and Technology, Science Town, Daejeon , Korea 1 sangyonglee@kaist.ac.kr, JunKyoungLee@kaist.ac.kr ABSTRACT Flow distribution pattern o two-phase mixture rom a header to parallel channels is examined extensively. The irst part o this paper introduces the low coniguration at single T-junctions that can be considered as the unit elements o the header-channels assembly o the compact heat exchangers. Experimental observations and appropriate models or prediction o low split to the branch are reported. Then the eect o the low interaction between two neighboring channels (branches on the low split pattern is considered. Finally, to simulate practical shape o the header-channels assembly o compact heat exchangers, test sections with multiple parallel channels and a partitioned header were tested. Dependence o the low distribution pattern on various operating conditions and header-channels conigurations is presented. As a way to achieve an even distribution o the low rom the header to parallel channels, depth o the channel intrusion to the header wall was adjusted; and 1/8 o the header hydraulic diameter was ound to be the optimum value o the intrusion depth. The irst part o this paper introduces the low coniguration at single, dividing T-junctions that can be considered as unit elements o the header-channels assembly o the compact heat exchangers. In other words, the header-channels assembly is simulated as an accumulation o single T-junctions with their branches having a rectangular shape. Experimental observations and appropriate models or prediction o low split to the branch are reported. Then the eect o the low interaction between two neighboring channels on the low split to the branches will be ollowed. Based on the single T-junction results, the eects o various operating INTRODUCTION Problem o low distribution rom a partitioned header to parallel channels, as shown in Fig. 1, is becoming o interest in predicting heat transer perormance o compact heat exchangers. In many cases, due to the low mal-distribution, the low rates in the channels are not the same (Kim et al. (003. For two-phase lows, the situation becomes even worse because the phase separation eect becomes prominent due to large dierences in density and viscosity between the gas and the liquid. In this paper, various aspects o two-phase low distribution at the header-channels assembly are examined based on a series o experimental and modeling works perormed up to date. Fig. 1 Compact Heat Exchanger Published by ECI Digital Archives,

3 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] Fraction o liquid separated (W 3 / W Lee and Lee (001 D h1 = 8mm j g1 = 1m/s, j 1 = 0.19m/s (Vertical low, square channel Stacey et al. (000 D 1 = 5mm j g1 = 19.06m/s, j 1 = 0.19m/s (Horizontal low, round tube Fraction o gas separated (W g3 / W g1 Fig. 3 Eects o Flow Direction and Cross-Sectional Shape o Main Channel (Lee and Lee (001 conditions and geometrical conigurations or header-channels assemblies are presented, and possibility o achieving an even distribution rom the header to parallel channels is going to be discussed. TWO-PHASE FLOW AT SINGLE T-JUNCTIONS Experimental Observations (a Test Loop (b Mixer Fig. Experimental Setup (Lee and Lee (001 Most o the previous works on the two-phase low at dividing T-junctions were or larger than 30 mm in hydraulic diameter. (Azzopardi (1999, Reimann et al. (1988 On the other hand, the hydraulic diameter o the headers o compact heat exchangers ranges mainly rom 5 mm to 10 mm, and currently there are only a limited number o experimental studies available or this size range, represented by Hong (1978, Stacey et al. (000 and Lee and Lee (001. Figure (a illustrates the experimental setup by Lee and Lee (001, typical o this type o experiments. Water and air enter the mixer that consists o concentric tubes (Fig. (b; air passes through the inner tube while water enters through the annulus. Thus an annular low is ormed at the entrance o the test section. Figure 3 shows a typical relationship between W 3 /W 1 and W g3 /W g1, that are the ractions o liquid and gas splits to the branch, or a vertical low in a square channel (Lee and Lee (001 and or a horizontal low in a round tube (Stacey et al. (000. Under these low conditions, the low pattern in the main tube is considered annular. The straight diagonal line in the igure represents the cases o the same mass qualities in the main and in the branch, i.e., =x 3 ; the data points at the lower let corner are the cases with very small low rate to the branch while those at the upper right corner indicate high low rates to the branch. The dividing annular low generally exhibits an S-shape curve as shown in the igure depending on the momentum o the liquid ilm and the entrainment rate within the main channel. With the lower momentum o the liquid ilm, the raction o the liquid split to the branch appears larger. On the other hand, with the higher liquid entrainment, the raction o the liquid split decreases because the entrained drops tend to pass through the run along the gas stream. As seen in Fig. 3, little dierence is ound between the two cases even though the channel shapes and the orientations are dierent. This is because, in small T-junctions, the liquid ilm thickness along the periphery o the main channel is 35

4 Lee and Lee: Fraction o liquid separated (W 3 / W Lee and Lee (001 D h1 = 8mm j g1 = 1m/s, j 1 = 3m/s Azzopardi (1994 D 1 = 15mm j g1 = m/s, j 1 = m/s Fraction o gas separated (W g3 / W g1 Fig. 4 Eect o T-junction Size (Lee and Lee (001 Fraction o liquid separated (W 3 / W G 1 = 00kg/m s Lee and Lee (001 = 0.1 = 0. = 0.3 = x Fraction o gas separated (W g3 / W g1 Fig. 6 Eect o Inlet Quality ( (Lee and Lee (001 almost the same, regardless o the channel shape and orientation as ar as the gas and liquid momentum luxes (or supericial velocities remain unchanged. Asymmetry o the ilm thickness caused by the gravity orce has been checked based on the result o Hurlburt and Newell (000. They quantiied the asymmetry o the liquid ilm thickness by using the ratio o the mean value to the thickness at the bottom o the horizontal pipe. For horizontal T-junctions (Stacey et al. (000 and Hong (1978, this ratio turned out to be within , which indicates a small asymmetry. Figure 4 compares the ractions o the liquid and gas low splits to the branch or a large T-junction (D 1 = 15 mm, Azzopardi (1994 with those or the small T-junction case (D h1 = 8 mm, Lee and Lee (001. More amount o the liquid is split to the branch with the smaller T-junction; this is because the liquid ilm Reynolds Fraction o liquid separated (W 3 / W j g1 = 1 m/s, j 1 = 0.16 m/s (Lee and Lee (001 B h = 1 mm (A.R. = 0.15 B h = 4 mm (A.R. = B h = 8 mm (A.R. = 1 A.R Fraction o gas separated (W g3 / W g1 Fig. 5 Eect o Branch Aspect Ratio (Lee and Lee (001 number stays below the critical value or liquid entrainment, and less amount o the entrained liquid lows through the run along with the gas stream. Thereore, the large T-junction data are not applicable to the cases with small T-junctions. Eect o the branch aspect ratio (or branch height, B h or a ixed width (or branch size has been investigated by Lee and Lee (001, as shown in Fig. 5. As the aspect ratio decreases, the raction o the liquid split also decreases, but only slightly. Azzopardi (1984 has reported the similar trend or vertical large junctions (D 1 = 3 mm with various aspect ratios o the branch ( as well. This can be explained as ollows: With the smaller aspect ratio (or size, the smaller region o the liquid ilm in the main channel is inluenced by existence o the branch, but at the same time, the suction orce to the branch also increases; and due to these competing eects, the rate o the liquid split decreases slightly with the branch aspect ratio. Figure 6 shows the increasing trend o the raction o the liquid split to the branch with the inlet quality at the main (. This trend is the same with that o Collier (1976 ( = The gas momentum increases while the liquid momentum decreases with the larger inlet quality, and the liquid stream is more likely to be split to the branch. Flow Distribution Models Several models have been developed or prediction o dividing T-junction lows (Azzopardi and Whalley (198, Shoham et al. (1987, Hwang et al. (1988, but those are basically or large-size T-junctions (D > 30 mm, which have never been tested or smaller sizes. Lee and Lee (004b have checked the applicability o those models to two-phase annular low at small, dividing T-junctions (less than 10 mm in hydraulic diameter using the experimental data by Hong (1978, Stacey et al. Published by ECI Digital Archives,

5 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] Table 1 Prediction models or annular low distribution at T-junction Authors Geometry, mm Orientation * Fluid Remarks Azzopardi and Whalley (198 D 1 = D = 3 D 3 = 6.35, 1.7, 19 V air-water a = ag Shoham et al. (1987 D 1 = D = D 3 = 51 H air-water a a r = g Hwang et al. (1988 n n g R = ρgu g ρu a / D1 ag / D1 R = D 1 = D = D 3 = 38 H air-water ( ( g * : H and V denote horizontal and vertical lows, respectively. (a Flow Coniguration (b Cross Section o Main Tube Fig. 7. Illustration o Flow Model by Hwang et al. (1988 (000 and Lee and Lee (001. Table 1 summarizes the prediction models tested. Figure 7 illustrates the low coniguration considered based on the model by Hwang et al. (1988. In this igure, the dividing streamlines o the liquid and the gas lows are shown with their radii o curvature being R and R g, respectively. Previously, Azzopardi and Whalley (198 introduced the concept o zone o inluence and assumed the boundary lines o the liquid and gas lows are the same, i.e., a = a g. Here, when a part o the gas low is split to the branch, the zone o inluence is ormed, and accordingly, a portion o the liquid low belongs to that zone is also extracted rom the main tube. The liquid low through the branch is mainly rom the ilm portion in the main tube rather than rom the entrained drop-low in the core portion because the liquid ilm has a lower velocity (momentum than the liquid drops. Thereore, the proportion o the liquid ilm entering the branch was considered dependent simply on the gas low rate lowing into the branch. Later, Azzopardi (1984 modiied this model to take account o the eect o the diameter ratio between the branch and the main tube. Based on the similar physical concept, Shoham et al. (1987 introduced a low-pattern-dependent model to extend the applicable range to the larger dividing ratio, W 3 /W 1. The model considers the inertial, centriugal and damping orces on the liquid ilm having a boundary line, a, with no liquid entrainment to the gas stream, and the dividing gas streamline has an arc shape with the radius o curvature R g. This model has an advantage o discriminating the liquid boundary line rom the gas boundary line, and an appropriate correlation or r (= a g a was proposed. In order to estimate the dividing ratio, the liquid ilm thickness should be given somehow, and Shoham et al. (1987 adopted the model by Taitel and Duckler (1976 to get this. The model works well or large T-junctions, as tested by Azzopardi (1994. Hwang et al. (1988 proposed a physical model similar to that o Shoham et al. (1987, but with a dierent approach in obtaining the dividing streamlines

6 Lee and Lee: Fraction o liquid separated (W 3 / W G 1 = 00kg/m s Experimental Data = 0.1 = 0.3 Azzopardi and Whalley (198 = 0.1 = 0.3 Shoam et al. (1987 = 0.1 = 0.3 Hwang et al. (1988 = 0.1 = Fraction o gas separated (W g3 / W g1 Fig. 8 Comparison between Experimental Results and Prediction Models: Eect o inlet quality ( (Data by Lee and Lee (001 (Lee and Lee (004b Fig. 9 Overall Evaluation o Prediction Model by Hwang et al.(1988 (Lee and Lee (004b or the gas and the liquid lows, as illustrated in Fig. 7. For an annular low, the acclerational and interacial drag orces were considered negligible, and only the centriugal orce was considered important. Finally, they derived a relationship between the R g /R as: where R R n k g ρgu = ρ U g a = D 1 n ag D 1 { 53( a / D } ng = 5 + 0exp k 1 (k = g (gas or (liquid Again, or this model, the ilm thickness and entrainment rate should be determined to predict the raction o the low split to the branch. Lee and Lee (004b adopted the approach o Whalley (1988, where the interacial riction actor (or roughness correlation, Ambrosini et al. (1991, triangular relationship (Asali et al. (1985, and entrainment correlation (Hewitt and Govan (1990 are considered. Figure 8 shows the variation o W 3 /W 1 with W g3 /W g1 using the models o Azzopardi and Whalley (198, Shoham et al. (1987 and Hwang et al. (1988 along with the experimental data by Lee and Lee (001. Among them, the model o Hwang et al. (1988 was concluded the best in predicting the dividing low behavior o two-phase mixtures. The overall accuracy in predicting the low quality inside the branch is also high as shown in Fig. 9. This is because the streamline shapes (1 ( are better represented by the Hwang et al. s model in describing the dividing low coniguration at the junction. Eect o Channel Spacing For most o the compact heat exchangers, distance between the parallel channels (S is comparable to or even smaller than the hydraulic diameter o the header. Thus the low interaction between the junctions should be taken into account careully. This eect becomes more prominent i the distance between the channels gets closer. Figure 10 illustrates the test section consists o two horizontal parallel channels and a header used by Lee and Lee (003 to check the channel (branch spacing Fig. 10 Channel Array o Compact Heat Exchanger Published by ECI Digital Archives,

7 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] (a Upstream Channel.0 Fraction o liquid separated (W 3 / W Fig. 11 Schematic Flow Coniguration W 1 = 18 kg/s, = 0. (Lee and Lee(003 S = 9 mm S = 5 mm S = 49 mm T-junction (Lee and Lee(001 = x Fraction o gas separated (W g3 / W g1 Fig. 1 Eect o Distance between Channels (S (Lee and Lee (003 eect on the low split. In their work, the second T-junction was chosen as the reerence junction because the low disturbance propagates mostly in the downstream direction and the backward low disturbance (rom the second junction to the irst one was considered insigniicant. According to Lee and Lee (003, the liquid low rate is always smaller at the second branch than at the irst branch, while the gas low rates in those two branches remain approximately the same. As illustrated in Fig. 11, it is easy or the liquid annulus to be separated out to upon arrival at the entrance o the irst branch, and the rest amount o the liquid ilm lows through the x 3 /, Predictions subsequent branch. With the smaller branch spacing, the liquid ilm reaches the next branch without being redistributed to have a uniorm thickness, and the amount o the liquid split to the second branch decreases. Figure 1 shows the eect o the branch spacing (S on the low distribution to each branch. The raction o the liquid split to the second branch decreases with the smaller branch spacing, and ar more deviates rom the single T-junction case reported by Lee and Lee (001. In other words, with a large branch spacing, amount o the low split can be predicted using the single T-junction model without any serious error, as easily imagined. The eect o the channel spacing can be taken into account in predicting the low split to the parallel branches by modiying Eq. ( originally proposed by Hwang et al. (1988 as: n k Fig. 13 Prediction o Flow Distribution (Lee and Lee (003 [ + 0exp{ 53( a / D }] = C (3 5 k % x 3 /, Experiments - 15 % S = 9 mm S = 5 mm S = 49 mm (b Downstream Channel

8 Lee and Lee: Fig. 14 Illustration o Header-Channels Flow in a -pass Heat Exchanger with C = 1 (Upstream T-junction ( Bh C (Downstream T-junction (5 = S (a Test Loop where C considers the upstream junction eect (Lee (005b. Equations (3 - (5 well represent the measured results or the upstream (Fig. 13(a and the downstream junctions (Fig. 13(b. It should be mentioned that these equations represent the channel spacing eect better than the correlation proposed by Lee and Lee (003, 004c that is based on the model by Shoham et al. (1987. TWO-PHASE FLOW DISTRIBUTION FROM A HEADER TO MULTIPLE PARALLEL CHANNELS General Flow behavior This time we extend our discussions on the low distribution o two-phase mixture rom a partitioned header to multiple parallel channels that are much closer to the practical shape o compact heat exchangers as illustrated in Fig. 14. Lee and Lee (004c have used a test loop shown in Fig. 15 that is basically the same with the experimental setup in Fig., except or the test section. The two-phase mixture lows upwards through a square vertical header connected to iteen (15 parallel, horizontal rectangular channels. An end plate (partition was installed at the most downstream o the header. The experiments were carried out or the annular low regime at the header inlet (b Coniguration o Test Section Fig. 15 Experimental Setup or Multiple-channel Experiments (Lee and Lee (004c because this low pattern is mostly probable to occur once the mass quality becomes large, say larger than 0.1. Figure 16 shows a typical distribution shape along with a low-visualized result. In Fig. 16(a, the abscissa represents the channel number counted rom the header inlet while the ordinate the low rates o the liquid (water Published by ECI Digital Archives,

9 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] ( W, c i, ( W g, c i (kg/s Region A W,in = 3 kg/s, W g,in = 08 kg/s ( W, c i ( W g, c i Region B Region C Channel # (a Flow Distributions Fig. 17 Schematic Flow Coniguration and General Trend o the Liquid Flow Distribution (Lee and Lee (004c W,in = 3 kg/s, W g,in = 08 kg/s (b Flow Visualization Fig. 16 Eect o End Plate on Flow Coniguration inside Header (Lee and Lee (004c and the gas (air separated out through each channel, denoted as (W,c i and (W g,c i, respectively. In general, at the ore part o the header (channels #1 - #4, region A, less amount o liquid low is separated out through the channels as the two-phase mixture proceeds in the downstream direction. In the middle part o the header (channels #5 - #9, region B, the trend becomes reversed. On the other hand, near the end plate (channels #10 - #15, region C, the trend appears similar to that in region A. For the gas low distribution, the trend is exactly opposite; increases, decreases and then increases again as lows along the header. This is because, to maintain the same pressure drops across the long parallel channels, the gas low rate should be small in a channel where the liquid low rate is large. However, the variation in the distribution shape o the gas low appears minor compared to that o the liquid low distribution. To help understanding the low distribution shape, the header part was visualized using a CCD camera as in Fig. 16(b, and the low coniguration is illustrated in Fig. 17. A large low recirculation is ound near the end plate o the header, which makes the low structure much complicated. In region A, the downstream eect (the end-plate eect did not appear at all. A portion o the liquid ilm lows out through the irst channel upon arrival at the corresponding entrance, and the rest amount o the liquid lows through the subsequent channels with the smaller low rates as proceeds downstream. On the other hand, in region C, a strong local recirculation occurs in the clockwise direction and the downstream eect predominates. Region B is the transition zone between regions A and C. A portion o the downward liquid low in region C collides and mixes with the upward liquid low at point X in the right hand side, and then cross the header to the stagnation point Y in the let

10 Lee and Lee: because o the strong mixing eect by the low recirculation. x 3 /, Predictions (a 1 st channel # channel #3 channel #4 channel + 0 % - 0 % x 3 /, Experiments (b # - #4 channels Fig. 18 Prediction o Flow Distribution at Header-Channel Junctions (Lee and Lee (004c hand side. (See Fig. 17. Thereater the low is divided into upward and downward streams to be separated out to the parallel channels in regions C and B, respectively. Though the liquid distributions in regions A and C have the same trend, the variation in region C is smaller compared to that in region A. This is because, due to a strong mixing eect by the low recirculation, the low distributions through the channels tend to be even in region C. Since the downstream eect does not appear in region A (channels #1-#4, the branching low rate in this region can be reasonably predicted using Eqs. (3 (5, and the results are shown in Fig. 18. However, or other regions (i.e., B and C, no appropriate prediction model has been reported or the low distribution; probably the homogeneous two-phase low model may be applicable Geometrical Shape Eects In this section, the eect o the geometrical shape o the header-channels assembly is going to be discussed; namely, the eects o the membranes in channels, the channel spacing and the intrusion depth. Figure 19 shows the cross sections o the extruded aluminum tube with six sub-channels (Lee (005b. In the same igure, a plain rectangular channel is also shown or comparison purpose. The total low area o the channel with the membranes is 16.3 mm that is almost the same with the low area o the plain channel, 16. mm. Here, a test section with 15 parallel channels similar to that in Fig. 15(b was used. The measured results on the liquid and the gas low distributions, as shown in Fig. 0, are not much dierent rom those without the membranes. The only notable thing is that, especially or the gas distribution, the low rate dierence between the neighboring channels appears smaller and the distribution pattern tends to be rather even. This implies that the two-phase low coniguration inside the header (as shown in Fig. 16 plays the major role in determining the low distribution. In the earlier part o this paper, the eect o the channel (branch spacing or a test section with two parallel branches (Fig.10 was discussed. Lee and Lee (005a have repeated the eect o the channel spacing using a test section with six parallel channels and an end plate, simulating the header-channels junction o the compact heat exchangers. At the same time, the eect o the intrusion depth o the channel entrances to the header wall on the low distribution was tested. Figure 1 shows the coniguration o the test section, where S and H denote the channel spacing and the intrusion depth to the header wall. Figure shows a typical distribution pattern o the liquid and gas lows to the channels or dierent channel distances (S, but with zero intrusion depth, H = 0. Basically the distribution pattens are the same with that shown in Fig. 16, obtained by Lee and Lee (004c or the test section with 15 channels. The only dierent thing is the relative size o each region. With the larger channel spacing, region C becomes relatively smaller (compared to the size o the entire region while region A remains unchanged. Thus region B appears to be relatively larger. Figure 3 shows a low visualization result and an illustration o the channel spacing eect on the header low coniguration. In Fig., the previous results by Lee and Lee (004a were plotted together or comparison. Again, the trend is the same even though the sizes o the header and the channel cross-sections are dierent. It should be mentioned that the header cross-section size appears less inluential than the channel spacing. The gas low distribution is almost insensitive to the channel spacing and the header size. Published by ECI Digital Archives,

11 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] Fig. 19 Cross Sections o the Channels with and without Membranes (Lee and Lee (005b 0.5 ( W g, c i / ( W g, c total ( W, c i / ( W, c total G g,in = 45 kg/m s, G,in = 8 kg/m s Without Membranes With Membranes Channel # (a Liquid distribution G g,in = 45 kg/m s, G,in = 8 kg/m s Without Membrabes With Membranes Channel # (b Gas distribution Fig.0 Eect o Membranes in Channels (Lee and Lee (005b Figure 4 is a typical result showing the eect o the intrusion depth. The dashed horizontal lines in the igure represent the case with even distributions o the liquid and the gas lows. With the zero intrusion depth (H = 0 mm, more amount o liquid is split out to the channels at the ore part o the header. On the other hand, the trend becomes reversed with a deep intrusion depth (H = 7 mm. The same trend has been reported by Lee and Lee (004a. This is because the intruded part hinders the Fig. 1 Coniguration o Test Section or Six-Channel Experiments (Lee and Lee (005a liquid low (liquid ilm inside the header rom being separated out through the channels (This can be conirmed through low visualization as demonstrated in Fig. 5(a or H/D h = 1/. This tells that there should be an optimum value o the intrusion depth or even low distribution, staying between zero and 7 mm; and the test results give 1.75 mm as the optimum value, corresponding to 1/8 o the header hydraulic diameter. The same results were obtained or H/D h = 1/8 with dierent channel spacing as shown in Fig. 6. The low behavior with the optimum value o H/D h is visualized in Fig. 5(b. The photographs show that the liquid and the gas lows inside the headers are well mixed with each other regardless o the distance between the channels. As a consequence, the liquid low distribution to the channels tends to be even. The results o Lee and Lee (004a were also plotted to show the same value o the optimum intrusion depth, D h /8. It is interesting to note that the ratio o the optimum intrusion depth to the header size (H/D h appears to be about the same, 1/8, or practical ranges o the low rate (G in = kg/m s, mass quality (x in = , channel distance (S = 10 mm and the header size (D h,in = 14 4 mm. At the same time, the gas separation rate to the channels remains even as shown in Fig. 6. CONCLUSION In the present paper, various aspects o the low distribution o two-phase mixture in small-scale header-channels assembly have been investigated. As a undamental approach, two-phase dividing (branching low at single T-junctions has been studied with the eects o the low rate (W in and quality o the mixture (x in, channel size (D h,in and orientation, and distance

12 Lee and Lee: ( W, c i / ( W, c average ( W g, c i / ( W g, c average S D h 8 4 Lee and Lee (004a Channel # Fig. Eect o Distance between Channels (S (H = 0, G in = 70 kg/m s, x in = 0.45 (Lee and Lee (005a ( W, c i / ( W, c average ( W g, c i / ( W g, c average H(mm Legend Channel # Fig. 4 Eect o Intrusion Depth (H on Flow Distribution (D h = 14 mm, S = 1.6 mm, G in = 70 kg/m s, x in = 0.45 (Lee and Lee (005a G in = 70 kg/m s, x in = 0.45 (a Flow visualization (a H/D h = 1/ (b Schematic low coniguration Fig. 3 Eect o Distance between Channels (S on Header Flow Coniguration (Lee and Lee (005a (b H/D h = 1/8 Fig. 5 Flow Visualization o Eect o Intrusion Depth (H on Flow Coniguration with Dierent Distance between Channels (D h = 14 mm, G in = 70 kg/m s, x in = 0.45 Published by ECI Digital Archives,

13 Enhanced, Compact and Ultra-Compact Heat Exchangers: Science, Engineering and Technology, Art. 39 [005] ( W, c i / ( W, c average ( W g, c i / ( W g, c average between the channels (S taken into account. Prediction models, originally developed or large T-junctions, were assessed based on the experimental results. Among the prediction models tested, the model by Hwang et al. (1988 was concluded to be the most appropriate. Then, to simulate the practical shape o the header-channels assembly o compact heat exchangers, test sections with multiple parallel channels and a partitioned header were tested. The header low consists o three regions, where the rate o the liquid low split to the channels decreases, increases and then decreases along the downstream direction. The rate o the liquid low split to the channels near the entrance o the header was well predicted by the modiied model or single T-junctions since this region is not aected by existence o the end plate (partition. For the other regions, where the local low recirculation eect predominates, no prediction model has been proposed up to date. The distribution pattern is relatively insensitive to existence o the membranes inside the channels or the spacing between the channels (S. On the other hand, the distribution pattern is strongly inluenced by the two-phase low coniguration inside the header, which is easily controlled by adjusting the depth o the channel intrusion to the header wall (H. The optimum value o the intrusion depth or even distributions o the liquid and the gas lows to the channels was ound to be about 1/8 o the header hydraulic diameter. ACKNOWLEDGEMENT S = 8 mm (Lee and Lee (004a S = 10 mm, S = 1.6 mm Channel # Fig. 6 Eect o Distance between Channels or Intrusion Depth o H/D h = 1/8 (G in = 70 kg/m s, x in = 0.45 (Lee and Lee (005a This work was supported by the Ministry o Commerce, Industry and Energy, and in part by the Brain Korea 1 Project. NOMENCLATURE a distance o streamline A-B (in Fig.7 rom the pipe wall on the branch side [m] B h branch height [m] B w branch width [m] C empirical correction actor D diameter [m] D h hydraulic diameter [m] G mass lux [kg/m s] H intrusion depth o the channel [m] j supericial velocity [m/s] N number o channels n exponent or streamline curvature correlation R radius o curvature [m] S distance between channels (branches [m] r distance between boundary lines o gas and liquid lows, a g a [m] U velocity [m/s] W mass low rate [kg/s] W / W + W x quality, ( Greek letters δ liquid ilm thickness [m] ρ density [kg/m 3 ] Subscripts 1 main tube run 3 branch liquid g gas i index (channel numbers in header (main tube inlet REFERENCES g g Ambrosini, W., Andreussi, P. and Azzopardi, B. J., 1991, A physically based correlation or drop size in annular low, Int. J. Multiphase Flow, Vol. 17, pp Asali, J. C., Hanratty, T. J. and Andreussi, P., 1985, Interacial drag and ilm height or vertical annular low. A.I.Ch.E. J., Vol. 31, pp Azzopardi, B. J., 1984, The eect o side arm diameter on two phase low split at a T junction, Int. J Multiphase Flow, Vol. 10, pp Azzopardi, B. J., 1994, The split o vertical annular

14 Lee and Lee: low at a large diameter T junction, Int. J Multiphase Flow, Vol. 0, pp Azzopardi, B. J., 1999, Phase separation at T-junctions, Multiphase Science and Technology, Vol. 11, pp Azzopardi, B. J. and Whalley, P. B., 198, The eect o low patterns on two phase low in a T junction, Int. J Multiphase Flow, Vol. 8, pp Collier, J. G., 1976, Single-phase and two-phase behavior in primary circuit components, Proc. N.A.T.O. Advanced Study Institute on Two-phase Flow and Heat Transer, Istanbul, Turkey, Hemisphere, Washington, D.C., pp Hewitt, G. F. and Govan, A. H., 1990, Phenomenological modeling o non-equilibrium lows with phase change, Int. J. Multiphase Flow, Vol. 33, pp Hong, K. C., 1978, Two phase low splitting at a pipe tee, J. Pet. Technol., pp Hurlburt, E. T. and Newell, T. A., 000, Prediction o the circumerential ilm thickness distribution in horizontal annular gas-liquid low, J. Fluids Engineering, Trans. ASME., Vol.1, pp Hwang, S. T., Soliman, H. M. and Lahey, R. T., Jr, 1988, Phase separation in dividing two-phase lows, Int. J. Multiphase Flow, Vol.14, pp Kim, M. H., Lee, S. Y., Mehendale, S. S. and Webb, R. L., 003, Microchannel heat exchanger design or evaporator and condenser applications, Advances in Heat Transer, Vol. 37, pp Lee, J. K. and Lee, S. Y., 001, Dividing two-phase annular low within a small vertical rectangular channel with a horizontal branch, Proc. 3 rd International Conerence on Compact Heat Exchangers and Enhancement Technology or the Process Industries, Davos, Switzerland, pp Lee, J. K. and Lee, S. Y., 004b, Assessment o models or dividing two-phase low at small T-junctions, Proc. 5th International Conerence on Multiphase Flow, Yokohama, Japan, Paper No Lee, J. K. and Lee, S. Y., 004c, Two-phase low behavior inside a header connected to multiple parallel channels, Proc. 3rd International Symposium on Two-phase Flow Modelling and Experimentation, Pisa, Italy. Lee, J. K. and Lee, S. Y., 005a, Optimum channel intrusion depth or uniorm low distribution at header-channel junctions, Proc. 6th KSME-JSME Thermal & Fluids Engineering Conerence, Jeju, Korea. Lee, J. K. 005b, An Experimental Study on Two-Phase Flow Distribution at Header-Channel Junctions o a Compact Heat Exchanger, Ph.D. Dissertation (in preparation, Department o Mechanical Engineering, KAIST. Reimann, J., Brinkmann, H. J. and Domanski, R., 1988, gas-liquid low in dividing T-juncton with horizontal inlet and dierent branch orientations and diameters, Kernorschungszentrum Karlsruhe, Institute ur Reaktorbauelemente Report KK 4399, pp Shoam, O., Brill, J. P. and Taitel, Y., 1987, Two-phase low splitting in a tee junction - experimental and modelling, Chem. Eng. Sci., Vol.4, pp Stacey, T., Azzopardi, B. J. and Conte, G., 000, The split o annular two phase low at a small diameter T-junction, Int. J Multiphase Flow, Vol. 6, pp Taitel, Y. and Dukler, A. E. 1976, A model or predicting low regime transition in horizontal and near horizontal gas-liquid low, AIChE J., Vol. 4, pp Whalley, P. B., 1988, Boiling, condensation and gas-liquid low, Oxord Science Publications. Lee, J. K. and Lee, S. Y., 003, Branching o two-phase low rom a vertical header to horizontal parallel channels, Proc. 4th International Conerence on Compact Heat Exchangers and Enhancement Technology or the Process Industries, Crete Island, Greece, pp Lee, J. K. and Lee, S. Y., 004a, Distribution o two-phase annular low at header channel junctions, Experimental Thermal and Fluid Science, Vol. 8, pp Published by ECI Digital Archives,