Evaluating The Performance Of Refrigerant Flow Distributors

Transcription

1 Purdue Unversty Purdue e-pubs Internatonal Refrgeraton and Ar Condtonng Conference School of Mechancal Engneerng 2002 Evaluatng The Performance Of Refrgerant Flow Dstrbutors G. L Purdue Unversty J. E. Braun Purdue Unversty E. A. Groll Purdue Unversty S. H. Frankel Purdue Unversty Z. Wang Purdue Unversty Follow ths and addtonal works at: L, G.; Braun, J. E.; Groll, E. A.; Frankel, S. H.; and Wang, Z., "Evaluatng The Performance Of Refrgerant Flow Dstrbutors" (2002). Internatonal Refrgeraton and Ar Condtonng Conference. Paper Ths document has been made avalable through Purdue e-pubs, a servce of the Purdue Unversty Lbrares. Please contact epubs@purdue.edu for addtonal nformaton. Complete proceedngs may be acqured n prnt and on CD-ROM drectly from the Ray W. Herrck Laboratores at Herrck/Events/orderlt.html

2 R12-7 EVALUATING THE PERFORMANCE OF REFRIGERANT FLOW DISTRIBUTORS Gang L, James E. Braun*, Eckhard A. Groll, Steven Frankel, and Zhchao Wang Purdue Unversty, 1077 Ray W. Herrck Laboratores West Lafayette, IN 47907, USA *Author for Correspondence E-Mal: ABSTRACT CFD was appled to evaluate the performance of both estng and mproved desgns for refrgerant dstrbutors. In general, t s better to utlze a sphercal base as compared wth other shapes and to locate the orfce close to dstrbutor base. These changes tend to mprove the robustness of the dstrbutor n terms of provdng even flow and phase dstrbuton n dfferent outlet branches when the orfce and/or dstrbutor are not orented optmally. In addton, eperments were performed that tend to valdate the general trends assocated wth the CFD results. Introducton Two-phase refrgerant s generally dstrbuted to ndvdual crcuts wthn the evaporator of a vapor compresson ar condtoner or refrgerator. Often tmes, an orfce s ntegrated nto a dstrbutor housng to provde a low cost method for epanson and refrgerant dstrbuton. Ideally, the mass flow rates and qualtes of the refrgerant etng the dfferent branches of the dstrbutor should be equal n order to obtan the best performance for the evaporator and the system as a whole. However, ths s generally not the case. Very lttle has been publshed regardng the analyss of refrgerant flow dstrbutors. Generally, refrgerant flow dstrbutors are desgned usng a tral-and-error process. Nakayama (2000) dd epermental nvestgatons nvolvng a refrgerant dstrbutor and proposed a new dstrbutor that was clamed to have better performance wth respect to flow dstrbuton. The mproved desgn utlzed a capllary mng space rather than an orfce. Ths desgn resulted n much more even flow rates for the ndvdual branches wthn the refrgerant dstrbutor. In a companon paper, L et. al (2002) demonstrated that commercally avalable CFD tools can be used to analyze the phase dstrbuton and separaton phenomena n refrgerant dstrbutors. The current paper descrbes results of the applcaton of CFD to evaluate the performance of both estng and mproved desgns for refrgerant dstrbutors. In addton, eperments were performed to valdate the general trends assocated wth the CFD smulatons. Dstrbutor Geometres and Performance Crtera Table 1 and Fgures 1 5 descrbe geometres that were consdered n ths study. The fgures are drawn to scale and show the nternal volumes where refrgerant flows. Each of the dstrbutors has four branches and an orfce that s located at the centerlne of the nlet to the dstrbutor. The Type 1 and Type 3 geometres are commercally avalable, whle the other 3 ncorporate desgn modfcatons. In the Type 1 geometry, the base of the dstrbutor s conve wth respect to the flow and comes to a pont. Ths desgn was conceved so as to provde a sngle pont of contact for separaton of the flow. The Type 3 desgn uses a cone-shaped base that s concave wth respect to the flow and was conceved so as to provde a mng chamber for dstrbuton of refrgerant. Type 2 and 4 are the same as Type 3, ecept the cone s replaced wth flat and sphercal surfaces, respectvely. The desgn s a modfcaton of where the orfce has been moved closer to the dstrbutor base and the depth of the chamber has been reduced. For all of the dstrbutors, the center of the base s on the centerlne of the dstrbutor along wth the center of the orfce. Results are presented n terms of uneven flow and qualty dstrbuton performance ndces. For a gven branch, the uneven flow and qualty ndces are m, = & m& m& m (1)

3 , = (2) where m, and, are the ndces for uneven flow and qualty, m& and are ndvdual branch flow rates and qualtes, and m& and are average mass flow rates and qualtes for all four branches. For a gven dstrbutor, average performance ndces for uneven flow and qualty are defned as 4 ( m& m& ) 2 = 1 m = (3) m& 4 = 1 ( ) 2 = (4) Refrgerant Dstrbutor Modelng In a companon paper, L et. al., (2002) demonstrated that FLUENT 5.5 s applcable for flow condtons encountered n refrgerant dstrbutors. Therefore, ths CFD modelng tool was used to nvestgate the performance of the dstrbutors descrbed n Table 1 and Fgures 1 5. The nflow boundary had specfed nlet velocty wth unform vod fractons at condtons assocated wth the outlet of the orfce. Constant and unform pressures were specfed for all outflow boundares. A three-dmensonal ncompressble adabatc flow was assumed wth no phase change. The goal of the smulatons was to evaluate the robustness of each of the dstrbutors wth respect to manufacturng and nstallaton defects. For vertcally nstalled dstrbutors wth no manufacturng defects, all of the dstrbutor desgns would produce even flow and phase dstrbutons n each of the branches snce the outlet branches are located symmetrcally about the centerlne of the dstrbutor. However, t s dffcult to orent the orfce perfectly and as a result the refrgerant wll et the orfce n drecton that s not along the centerlne of the dstrbutor. To smulate ths effect, the drecton of refrgerant flow was consdered to be off the centerlne by an angle of 3.7 o. In addton, gravty effects wll affect the flow and phase dstrbutons when the dstrbutor s orented horzontally. Ths case was also consdered. As a frst step to elmnate the poorer performng devces, sngle-phase smulatons were performed. Under the assumpton of homogeneous equlbrum flow, two-phase flow can be smplfed to sngle-phase flow wth proper averagng methods used for the propertes of the fluds. From a cycle analyss of a 3-ton capacty refrgeraton applcaton, average property values for the pseudo-flud enterng the dstrbutor were obtaned for the CFD smulaton, as well as boundary condtons and operatng condtons. The densty and vscosty for ths pseudo-flud were taken to be kg/m 3 4 and kg/m-s, respectvely, correspondng to an evaporatng temperature of 0 0 C and a qualty of The outlet reference pressure was 5 atm, whch s close to the evaporatng pressure for ths specfc applcaton. Fgure 6 shows flow dstrbuton results for the sngle-phase smulatons appled to dstrbutor Types 1 4, all orented upwards n the vertcal drecton wth an orfce tlted 3.7 o towards. The and y components of the nlet velocty were m/s and 3 m/s. For ths stuaton, the sharp-end (Type 1) dstrbutor performed consderably worse than the other 3 desgns. The mamum flow rate for the sharp-end dstrbutor occurs n, whch s the drecton of orentaton of the orfce. On the other hand, the blunt-end desgns tend to recrculate refrgerant pror to dstrbuton so the mamum flow rate occurs n other branches and the flow rate s more evenly dstrbuted. In general, the sharp-end desgn s very senstve to orfce orentaton and was not consdered n two-phase smulatons. Fgures 7 and 8 show flow and qualty dstrbuton results for two-phase smulatons appled to dstrbutor Types 2 5, all orented upwards n the vertcal drecton wth an orfce tlted 3.7 o towards. The boundary condtons for two-phase flow were the same as for the sngle-phase smulatons ecept there was a

4 unform vod fracton specfed at the nlet assumng a no slp condton. For these blunt end dstrbutors, the shape of the base plays an mportant role n terms of flow recrculaton and dstrbuton. Types 2 4 have smlar overall performance, but result n dfferent flow and qualty dstrbutons for ndvdual branches. The dstrbutor s much less senstve to orfce orentaton and has much more unform flow and qualty among the dfferent branches. Fgures 9 and 10 show velocty vectors and vod fracton contours for a plane that goes through the center of the dstrbutor,, and for the Type 3 and 5 dstrbutors (orented upwards n the vertcal drecton wth an orfce tlted 3.7 o towards ). For the cone-shaped base (Type 3), more of the hgher-momentum lqud refrgerant flows through than leadng to larger total flow and lower qualty (see Fgures 7 and 8) through ths branch. However, as shown n Fgures 7, the largest flows occur n and 4 even though there s a slght ncrease n qualty of the refrgerant n these branches (Fgure 8). The comple recrculaton patterns produced by ths shape result n relatvely asymmetrc velocty and phase dstrbutons n all three dmensons. On the other hand, the desgn (sphercal base wth a closer orfce locaton) results n more symmetrc recrculaton patterns and velocty and phase dstrbutons. Fgures 11 and 12 show flow and qualty dstrbuton results for the dstrbutors mounted n a horzontal drecton wth perfect orfce orentatons. The flow and qualty dstrbutons are uneven due to the effects of gravty. The sphercally-shaped bases have more even flow and phase dstrbutons than the cone and flat bases. In partcular, the flat base has much worse phase dstrbuton than the other desgns. Table 2 summarzes the overall performance of the Type 2 5 dstrbutors n terms of mass flow and qualty dstrbutons at the outlet branches. The dstrbutor s much more robust wth respect to mperfectons n orfce and dstrbutor nstallaton as compared wth the other desgns. Epermental Testng In order to valdate the trends assocated wth the CFD modelng, eperments were performed for 3 of the geometres: sharp-end base (Type 1), cone-shaped base (Type 3), and sphercal base wth closer orfce locaton (). Type 1 and 3 were commercally avalable, whereas had to be specally bult for ths project. Dfferent orfces were nvestgated, but only the best flow dstrbuton results are presented for each dstrbutor. Fgure 13 shows the epermental test stand that was developed for testng. The nlet pressure and temperature to the orfce and the pressures at the outlets of the branches are controlled usng adjustable valves. For each test, the pressures at the outlet of each branch were controlled to the same value. The refrgerant mass flow rate through an ndvdual dstrbutor branch s estmated from an energy balance on an electrcally powered superheater located after the evaporator that follows that branch. Measurements of power nput and refrgerant nlet and outlet temperature and pressure are used along wth property data to estmate mass flow rate through that branch. The qualty of the refrgerant leavng each branch s estmated from an energy balance on the evaporator assocated wth that branch. The evaporators are water-cooled and heat transfer rates are determned from watersde flow rates and temperatures. The evaporator nlet enthalpy s determned from a refrgerant-sde energy balance usng the water-sde heat transfer rate and refrgerant outlet state measurements. The nlet qualty s then determned from refrgerant property data wth the estmated nlet enthalpy and measured pressure. Based upon an uncertanty analyss, the uncertanty n the mass flow measurements ranges from about 5 to 1, whereas the uncertanty n the qualty measurements s between about 10 and 2. The dstrbutors were tested n a vertcal, downward flowng orentaton. Tests were performed at hgh, medum, and low condenser pressures and repeated twce at each condton. In addton, each of the tests was repeated wth the dstrbutor rotated through all four possble arrangements of the outlet branches feedng the dfferent evaporator crcuts of the test stand. Ths was done n order to elmnate any bas assocated wth orfce orentaton and the test stand. Average performance ndces for uneven flow and qualty (equatons 3 and 4) were determned for each dstrbutor by averagng results obtaned for all condtons and dstrbutor orentatons (24 tests for each dstrbutor). Fgure 14 shows test results for the dfferent dstrbutors consdered. Consstent wth the CFD smulatons, the sharp-end dstrbutor had the worst performance of those consdered. The performance of the sphercal-base dstrbutor was slghtly better than the cone-base dstrbutor. However, the dfferences were smaller than those determned through smulaton. Several factors could have led to these dfferences. The actual orentaton of the orfce wthn the dstrbutor housng was unknown and has a major mpact on flow dstrbuton. The epermental results were averaged for several dfferent tests where the orfce orentaton could have changed. The uncertanty n the flow measurements can be as hgh as 1. The smulatons assumed a unform vod fracton at the outlet of the orfce. The actual vod fracton dstrbuton at the orfce outlet was unknown and a non-unform dstrbuton would mpact the flow dstrbuton at the branch outlets.

5 Conclusons Epermental results confrmed some general trends arsng from CFD smulatons of estng and mproved refrgerant dstrbutors. In general, t s better to utlze a sphercal base as compared wth other shapes and to locate the orfce close to dstrbutor base. These changes tend to mprove the robustness of the dstrbutor n terms of provdng even flow and phase dstrbuton n dfferent branches when the orfce and/or dstrbutor are not orented n an optmal fashon. References G. L, S. Frankel, J.E. Braun, and E.A. Groll, Applcaton of CFD models to two-phase flow n refrgerant dstrbutors, 2002 Internatonal Refrgeraton Conference at Purdue, (2002). M. Nakayama, Y. Sumda, S. Hrakun, and A. Mochzuk, Development of a refrgerant two-phase flow dstrbutor for a room ar condtoner, Internatonal Refrgeraton Conference at Purdue (2000) pp Sharp-end dstrbutor Blunt-end dstrbutors Table 1 Descrpton of the refrgerant dstrbutors Commercally avalable desgn Type 1 Flat base Same as Type 3 ecept base surface s flat Type 2 Cone base Commercally avalable desgn Type 3 Round base (base surface center at the center of the orfce) Same as Type 3 ecept base surface s a sphercal Same as Type 3 ecept orfce poston s moved closer to chamber end; chamber depth s reduced Table 2 CFD results for average two-phase unevenness dstrbutons for dfferent dstrbutors Horzontal nstallaton m Vertcal nstallaton wth tlted nlet velocty Type % 13.89% 27.09% 13.48% Type % 6.77% 48.23% 16.02% 2.66% 2.67% 26.87% 11.26% 0.91% 2.57% 1.02% 2.25% m

6 Branch 1 Branch 1 Branch 4 Fgure 1: Type 1 dstrbutor (half shown) wth ponted base Branch 1 Fgure 4: dstrbutor wth sphercal base Branch 1 Branch 4 Fgure 2: Type 2 dstrbutor wth flat base Branch 1 Branch 4 Fgure 5: dstrbutor (half shown) wth sphercal base and orfce moved closer to base Fgure 3: Type 3 dstrbutor wth cone base

7 m,i 100 (%) Type 1 Type 2 Type 3-30 Fgure 6: Sngle-phase smulaton results for mass flow dstrbuton n vertcally nstalled dstrbutors wth mperfect orfce orentatons m, 100 (%) Type 2 Type 3 Fgure 7: Two-phase smulaton results for mass flow dstrbuton n vertcally nstalled dstrbutors wth mperfect orfce orentatons, 100 (%) 15% 1 5% -5% Type 2 Type % Fgure 8: Two-phase smulaton results for qualty dstrbuton n vertcally nstalled dstrbutors wth mperfect orfce orentatons

8 Fgure 9: Velocty vectors supermposed on vod fracton contours for Type 3 dstrbutor Fgure 10: Velocty vectors supermposed on vod fracton contours for dstrbutor 15% 1 5% -5% -1-15% Type 3 Type 2-2 Fgure 11: Two-phase smulaton results for mass flow dstrbuton n horzontally nstalled dstrbutors wth perfect orfce orentatons m, 100 (%) 12% Type 2 1 8% 6% Type 3 4% 2% -2% -4% -6% -8% Fgure 12: Two-phase smulaton results for qualty dstrbuton n horzontally nstalled dstrbutors wth perfect orfce orentatons, 100 (%)

9 Fgure 13: Epermental test stand % % 10.2% : Sphercal Base Type 3: Cone Base Type 1: Sharp-End Base Fgure 14: Comparson of average unevenness flow dstrbutons for dfferent types of dstrbutors determned from eperments