ASHRAE LV-11-C018-2011 CFD Simulation of Single-Phase Flow in Plate Heat Exchangers.pdf

1、Steven OHalloran is an assistant professor in Mechanical Engineering at the University of Portland, Portland, Oregon. Amir Jokar is a consultant at ThermoFluids Tech, Vancouver, Washington. CFD Simulation of Single-Phase Flow in Plate Heat Exchangers Steven OHalloran, Ph.D. Amir Jokar, Ph.D. Member

2、ASHRAE Member ASHRAE ABSTRACT Chevron plate type heat exchangers have widely been studied through experimental analysis; however, less computational work has been reported on these types of heat exchangers due to the complexity of their interior configuration with corrugated plates. This study has a

3、pplied computational thermal and fluid dynamics methods to simulate single-phase flow in three brazed plate corrugated heat exchangers. The heat exchangers have plates with different chevron angles. The three heat exchangers simulated are: 60/60, 27/60, and 27/27. For this purpose, a commercially av

4、ailable CFD software package (Fluent) has been utilized and simulations for different temperature and velocity boundary conditions have been performed. In the numerical simulations, the k- SST turbulence model has been used. The resulting outlet temperatures have been found as well as the heat trans

5、fer rates between the fluids in the heat exchanger. The heat transfer rates obtained from the simulations are reported and compared for the three different chevron angles. The CFD model established in this study can be used for a variety of design conditions and practical applications, such as inves

6、tigation of nanofluids in complex geometries. INTRODUCTION This study presents the Computational Fluid Dynamics (CFD) analysis of single-phase flow in Plate Heat Exchangers (PHE), which is an essential part of a larger effort that will investigate the effects of nanofluids for HVAC chevron, elliptic

7、, and round embossing types. The heat exchangers consisted of 60 plates with dimensions of 1.839 m in length and 0.294 m in width, and mean channel gap of 1.6 mm. They simulated heat transfer and fluid flow, using Fluent CFD software, to obtain temperature and pressure distributions within the three

8、 heat exchangers. They also conducted experiments on the heat exchangers and compared the results with the CFD simulations, which were reported in good agreement. The plate with the elliptical shape represented better overall performance than the other two types. The CFD simulation of PHEs under thi

9、s investigation is unique since it has three channels and eliminates the above mentioned boundary conditions restrictions often used for previous work. Also, multiple chevron angle configurations have been tested, 60/60, 27/60, and 27/27. Due to complexity of the system, the CAD modeling and CFD sim

10、ulation have been conducted step by step, considering flat plates for the heat exchanger as the first step, followed by a simplified version of corrugated plates. The details of results obtained on the flow of water as a base fluid through the PHEs are presented in this manuscript. 148 ASHRAE Transa

11、ctionsSYSTEM CONFIGURATION The objective of this study was to conduct CFD simulations on a PHE that had previously been analyzed experimentally by the coauthor, as presented in Hayes and Jokar (2009), so that experimental data could be used for comparison and simulation verification. The PHE include

12、d three channels, where hot water flowed in the middle channel downward and cold water on the two side channels upward. The plate configuration, schematic flow diagram, the actual PHE, and its cutaways are shown in Figure 1. The details of interior geometries of the plate are also given in Table 1.

13、Typical experimental results from Hayes and Jokar (2009) for the L-plate configuration are given in Table 2. Three test points for the experimental results are given representing different inlet conditions. (a) (b) (c) (d) Figure 1 (a) Plate configuration, (b) flow diagram of the hot and cold fluids

14、, (c) entrance/exit ports, and (d) cutaway along and perpendicular to flow in plate heat exchangers. The corrugated plate, as shown in Figure 1(a), does not look sophisticated by itself; however, as the plates are attached together, complex and three dimensional flow channels are formed. Figure 1(d)

15、 shows cutaways from a cross section of the PHE; one along the flow with wavy shapes and the other perpendicular to the flow with honeycomb shape. Table 1. Interior geometries of the plate heat exchangers Parameter L Plate M Plate H Plate , degree 60/60 27/60 27/27 Ltotal, mm 533.4 533.4 533.4 Lcorr

16、ugated, mm 444.5 444.5 444.5 Lport, mm 476.25 476.25 476.25 W, mm 127 127 127 b, mm 2 2 2 , mm 6.27 6.19 6.03 1.2 1.2 1.2 Aprojected, m2 0.05645 0.05645 0.05645 Aeffective, m2 0.06774 0.06774 0.06774 Table 2. Experimental results for corrugated plate heat exchanger (L plate) Parameter Test Point #1

17、Test Point #2 Test Point #3 Rehot 7612 4390 2080 Recold 1657 3449 2381 g1865g4662 g3035g3042g3047 (kg/s) 0.3086 0.1915 0.0943 g1865g4662 g3030g3042g3039g3031, single channel (kg/s) 0.1887 0.4281 0.3023 Thot_in (K) 319.6 319.2 319.2 Thot_out (K) 309.2 301.7 297.6 Tcold_in (K) 289.4 290.0 289.5 Tcold_

18、out (K) 305.8 297.9 296.6 Q (W) 13422 14023 8544 2011 ASHRAE 149SYSTEM SIMULATION Due to the complexity of flow passages within the PHE and limitations on computer hardware and software, it was decided to start developing the CAD and CFD models from simpler geometries and enhance them stage by stage

19、. For this reason, flat plate heat exchangers with similar dimensions as in the PHE were generated and assembled. This simpler heat exchanger, as described in the following section, could also be used as a reference for comparison with the enhanced versions. Simulation of a Flat Plate Heat Exchanger

20、 A simplified two-channel flat plate heat exchanger was generated to create a baseline for further, more complicated simulations. The simulation was run and tested for variety of conditions and it showed reasonable results, although they are not reported here. The next step was to generate a model f

21、or a three-channel flat plate heat exchanger that could more closely resemble the real PHE, as presented in Figure 1. The middle channel contained the hot water while the side channels contained the cold water in a counter-flow arrangement. The dimensions of each channel were 1275332 mm (x,y,z), as

22、presented in Table 1. A hexahedral mesh was generated within Gambit 2.4.6 with a total of 724,200 elements. The walls separating the channels were set as stainless steel with a thickness of 0.4 mm and a thermal conductivity of 14.0 W/mK. The working fluid in the model was water with constant propert

23、ies selected at 293 K. The k- SST turbulence model was used for the simulations with transitional flows enabled. The inlets were set as constant velocity inlets. For the inlet turbulence parameters, a turbulence intensity of 5% was used along with a hydraulic diameter of 4 mm. Three different test p

24、oints were simulated, each with different inlet conditions. These test points were selected to make future comparisons to experimental data from Hayes and Jokar (2009). Inlet conditions for each test point as well as results from the simulations are shown in Table 3. The outlet temperatures were cal

25、culated during the simulation as well as the heat transfer rate (Q) between the fluids. The heat transfer rates were found to be 7460 W, 8968 W, and 6266 W for Test Point #1, #2, and #3, respectively. Table 3. Simulation results for the flat plate heat exchanger Parameter Test Point #1 Test Point #2

26、 Test Point #2 Rehot 7612 4390 2080 Recold 1657 3449 2381 g1865g4662 g3035g3042g3047 (kg/s) 0.3085 0.1915 0.0943 g1865g4662 g3030g3042g3039g3031, single channel (kg/s) 0.1886 0.4282 0.3022 Thot_in (K) 319.6 319.2 319.2 Thot_out (K) 313.0 307.3 302.8 Tcold_in (K) 289.4 290.0 289.5 Tcold_out (K) 299.2

27、 295.5 294.9 Q (W) 7460 8968 6266 Simulation of a Corrugated Plate Heat Exchanger The next step was to make the flat plates corrugated in order to move one step closer to the real plate heat exchangers. The goal was to make the corrugations similar to those in real heat exchangers and at the same ti

28、me be able to make such plate configurations in the shop. The corrugations with sine shape cross sectional area were almost impossible to make with the available facilities; however, a rectangular shape would be possible using the CNC machines. A three-channel PHE using the corrugated plates require

29、s fine meshing using tetrahedral cells. For this reason, it was decided to start with a smaller size model by shortening the length of corrugated plates but keeping other dimensions the same as the PHEs. Figure 2 shows a single plate that was modeled to create the corrugated plate heat exchanger. As

30、 mentioned above, the length of the heat exchanger was shortened in order to make the number of elements in the simulation reasonable. Three corrugation angle configurations were selected, 60/60, 27/60, and 27/27. These configurations will 150 ASHRAE Transactionsalso be referred to as L-plate, M-pla

31、te, and H-plate, respectively. The length of the corrugated section was 56 mm in each case. This length is 1/8th the length of the actual heat exchanger in the experimental setup (a representation of the full-length plate is shown on the left in Figure 2). An inlet and outlet flat plate section with

32、 a length of 50 mm was also included on the corrugated plate. This allowed an inlet and outlet to be set up in the simulation software, and the length helped to reduce reverse flow at the outlet of the heat exchanger. The width of the plate is 127 mm, and the material thickness was 0.4 mm as shown i

33、n the detail view of Figure 2. The channel depth and width were 2 and 2.5 mm, respectively, and the pitch between channels was 6.27 mm. Figure 2 Single plate for the corrugated plate heat exchanger (L-plate). Four plates were assembled in the three-dimensional model to create the solid model of the

34、corrugated plate heat exchanger. The assembly is shown in Figure 3(a), while an exploded view of the assembly is shown in Figure 3(b). When the plates are assembled, each adjacent plate is rotated 180 during assembly (the plates colored green in the assembly in Figure 3 are rotated 180 compared to t

35、he gray colored plates). The inlet and exit section (flat sections) were created with a thermal conductivity of approximately zero; as a result, heat transfer occurs only in the corrugated section of the heat exchanger with properties of stainless steel (thermal conductivity of 14 W/m.K). (a) (b) Fi

36、gure 3 (a) Assembled corrugated plate heat exchange, (b) exploded view of L plate heat exchanger assembly. 2011 ASHRAE 151After the solid model was created in Pro/Engineer, the assembly was exported in STEP format and imported into Gambit 2.4.6 to create the fluid model, mesh the volumes, and set bo

37、undaries for the CFD model. A mesh of approximately 1.5 million elements was generated for the geometry. A hexahedral mesh was used for the flat inlet and outlet sections while a tetrahedral mesh was used for the more complicated corrugated section of the heat exchanger. In addition to the mesh gene

38、ration in Gambit, boundary conditions were also set. Velocity inlets were specified for the three inlets (two cold fluid inlets and one hot fluid inlet), and pressure outlets were set for all three outlets. After mesh generation and boundary conditions were completed in Gambit, FLUENT 6.3.26 was use

39、d to conduct the simulations. Constant properties for water were used at a temperature of 293 K. The k- SST turbulence model was used with transitional flows enabled. Three different sets of inlet conditions were used and these matched the three test points listed previously in Table 3 (and matched

40、up to experimental data available for the full-length heat exchanger). Approximately 3000 iterations were required to reach steady state for all three inlet conditions. In all cases, some small reverse flow was present at the edges of the outlets; however it was determined that this did not affect t

41、he heat transfer rates calculated in the simulation since the heat transfer occurs only in the corrugated section. Grid independence was checked by varying the grid size. A finer mesh with two million elements was created for one of the models (L plate) and the resulting heat transfer and pressure d

42、rop were compared. The variation for all three parameters was within 4% between the original mesh and the finer mesh, therefore the original mesh size was used for the final simulations. Figure 4(a) shows velocity contours in one of the cold channels of the heat exchanger for Test Point 1. Figure 4(

43、b) shows pathlines in the same cold side. Each pathline has a unique color so that each path can be traced from inlet to outlet. Figure 5 shows temperature contours for the cold channel and hot channel for each of the heat exchanger configurations. (a) (b) Figure 4 (a) L-plate velocity plot in top c

44、old channel, (b) uniquely colored pathlines inside of top cold channel. 152 ASHRAE Transactions(a) (b) (c) (d) (e) (f) Figure 5 Temperature plots for the top cold (left) and hot plate (right) (a,b) L-plate, (c,d) M-plate, (e,f) H-plate. 2011 ASHRAE 153Numeric results for the corrugated heat exchange

45、r simulations are shown in Table 4. The Reynolds Number, mass flow rate, and inlet temperatures are the same as Table 3 and are therefore omitted. The outlet temperatures, pressure drops and heat transfer rates are found from the simulations. It can be seen that the H-plate configuration has the hig

46、hest heat transfer for any given test point. The H-plate also has the highest pressure drop since it has the most restrictive chevron angle. Although quantitative comparisons cannot be made between the experimental results previously collected (Hayes and Jokar 2009) and the current simulations at th

47、is time since the simulation geometry is shorter than the experiment, qualitative comparison can made comparing the L-plate data in Table 2 (experimental) and L-plate data in Table 4 (numerical). It can be seen that for the L-plate configuration Test Point #2 has a higher heat transfer rate compared

48、 to the Test Point #1 or Test Point #3 for the L-plate. Table 4. Simulation results for the flat plate heat exchanger Test Point #1 Test Point #2 Test Point #3 Parameter L-Plate M-Plate H-Plate L-Plate M-Plate H-Plate L-Plate M-Plate H-Plate Thot_out (K) 318.1 317.9 317.1 316.5 316.3 315.7 315.0 314.6 313.8 Tcold_out (K) 291.9 292.1 293.5 291.2 291.3 291.6 290.8 290.9 291.2 phot (Pa) 23,301 28,481 159,169 9411 11,629 59,514 2510 3177 15,277 pcold (Pa) 2469 10,170 15,594 11,474 50,765 73,460 5926 25,596 37,