ASHRAE IJHVAC 17-2-2011 HVAC&R RESEARCH An International Journal of Heating Ventilating Air-Conditioning and Refrigerating Research.pdf

1、 EditorialRefrigerants: Energy and environmental impactsThe results from a 1999 National Academy of Engineering survey of 60 professional engineering societiesresulted in air conditioning and refrigeration being named as the tenth greatest engineering achievement ofthe twentieth century (National Ac

2、ademy of Engineering (NAE) 2010). While the HVAC accepted October 17, 2010Dr. Guadalupe Huelsz is researcher. Fabrisio Gomez, EM, is project associate researcher. Miguel Pineirua, EM, is projectassociate researcher. Dr. Jorge Rojas is researcher. Mauricio de Alba is researcher. Vctor Guerra, ESM, is

3、 researcher.100 Btu/h). The authors presented two analyticaland two experimental methods to estimate the gas-ket heat transfer load. The two analytical methodsare based on the use of coefficients obtained for aspecific gasket model, which limits their applica-bility. Each one of the two experimental

4、 methods isbasedontworeverseheatleakage(RHL)tests.InanRHL test, the refrigerator/freezer unit is put inside atemperature-controlled chamber, an electrical heatsource is placed inside of each one of the compart-ments, and the power input to each heat source ismeasured along with the inside and outsid

5、e temper-atures. In both methods, the first RHL test measuresthe total heat loss through the unit (baseline unit).The second RHL test uses modified units, whichdiffer in both methods. The first method uses twoidentical units whose doors have been removed andare connected by the gaskets, forming one

6、cavityforeachcompartment.Thismethodisnotusefulforside-by-siderefrigerator/freezersandreliesonsomespeculation regarding the door loss. In the second133HVACthe gravity force is in the z direction.Governing equationsFluid dynamics and heat transfer phenomena aregoverned by Navier-Stokes and energy equa

7、tions.The following assumptions are considered: lami-nar flow, Boussinesq approximation (i.e., densityvariation is only relevant in buoyancy terms of mo-mentumequations),Newtonianfluid,negligiblevis-cosity dissipation, constant physical properties, andnegligibleradiationheattransfer(althoughradiatio

8、nFigure 3. A cross-section plane of computational domain; grav-ity force is in the z direction.Downloaded by Tvaluesweretaken from experimental results presented below).Thereferencedensityforthebuoyancytermwasthedensity corresponding to the initial temperature ofeach subdomain. Air velocity is zero.

9、 The tempera-tures of the solid elements is 3.75C (38.75F) (themedia between TO and TI).Numerical modelThe computational fluid dynamics software em-ployed was CFXRv 11.0. It uses the finite controlvolume technique with a total temporal implicit dif-ferentiation (ANSYS 2006). The mesh is nonstruc-tur

10、ed with tetrahedron elements. The mesh was re-fined at fluidsolid interfaces to better represent theboundaryeffects.DiffusivetermsareevaluatedwithFigure 4. Temperature map at middle cross-section plane; z =50.8 mm (2 in.).a central difference scheme. In advection terms, ahigh-resolution scheme has b

11、een used; this schemeis based in an upwind scheme plus a function thatevaluatesitschangesinthecontrolvolume.Thetimestep for fluids considered was Delta1tf= 0.1 s and forsolids Delta1ts= 1.0 s. The resulting algebraic systemequation is solved in a coupled way. As steady-stateresults are sought, the i

12、terative process is truncatedwhen the rms of the residuals from the equations isless than 104.Numerical resultsFigure4showsthetemperaturemapinthemiddlecross-section plane, i.e., height 50.8 mm (2 in.).Table 1 shows a comparison between numericaland experimental temperatures for TSO and TSI.Experimen

13、tal temperatures are taken at steady stateusingthermocoupleslocatedinsection8(seeFigure5). Numerical results are taken at correspondingpositions.Numerical results for TSO and TSI differ by0.2C(0.4F) and 2.8C(5.0F) from experimentalresults, respectively; numerical results for TSO-TSIdiffer by 2.6C(4.

14、7F). Thus, the heat transfer re-sults from the numerical model can be considered agood approximation totheexperimental conditions,Table 1. Numerical and experimental temperatures comparison.TSO TSI TSOTSINumerical 22.2C (72.0F) 3.3C (38.0F) 18.9C (34.0F)Experimental 22.4 0.3C (72.3 0.5F) 6.1 0.1C (4

15、3.0 0.2F) 16.3 0.3C (61.3 0.5F)Downloaded by Tthe thermocouple-end corresponding to TO waspastedtothecylinderwhereTOwasmeasured.Eachthermocouple-end corresponding toTFO,TSO, andTSI waspastedatthewidthcenteroftheexteriorsur-faceoftheEPS,ofthegasketexternalsurface,andofthe gasket internal surface, res

16、pectively. These endswere pasted with Omega thermocouple cement andwere covered with an insulating tape. For the frameexperiments, a rectangular cross-section EPS framewasused.ThewidthoftheEPSwasthesameasthatof the gasket external surface dF= d = 20 0.01mm (0.7874 0.0004 in.), and the length LFwas9.

17、26 0.01 mm (0.3657 0.0004 in.). To achievea good thermal contact between the gasket and theEPS frame, the frame was pasted to the door andcabinet walls using sealing extruded rubber mastic.ResultsThe effective conduction coefficient keff, the filmcoefficients hoand hi, the heat transfer load perunit

18、 length qi, and the heat transfer load Qi,foreach section of the freezer gasket are presented inFigure 6. The uncertainty associated is calculatedusing the error propagation theory, taking into ac-count the uncertainty associated to the thermocou-ple calibration, but the uncertainty associated to th

19、elocation of the thermocouples is not considered dueto the impossibility to estimate it.Detailed observations indicate that the variationof keffis mainly due to small changes in the distancebetweenthedoorandcabinet.Thesmalldifferencesin keffbetween vertical and horizontal positions in-dicate that ai

20、r convection inside the gasket is notFigure 6. (a) Effective conduction coefficient keff, (b) film coefficients ho(dots) and hi(line), (c) the heat transfer load per unitlength qi, and (d) the heat transfer rate Qifor each section of the freezer gasket.Downloaded by T (12)a= 1.2 kg/m3(0.07 lb/ft3) i

21、s the air density; V= 0.25 m3(8.8 ft3) is the cavity volume; A is theinterchange rate of air between the cavity and theroom, relative to the volume of the cavity, in inversetime units. The average specific heat of the humidair, Cp, is defined asCp = Cpa+ WCpw, (13)where Cpa= 1.006 kJ/kgaC (0.240 Btu

22、/lbF) isthe specific heat of dry air; the term WCpwis thecontribution to the sensible load of water vapor inthe air, with W the average during the experimentof the absolute humidity and Cpw =1.882 kJ/kgwC(0.999 Btu/lbF) the water specific heat.The latent heat load Qlis given byQl=m hwDelta1W, (14)wh

23、ere hw= 2,700 kJ/kgw(1,162 Btu/lb) is the av-erage water vaporization enthalpy, and Delta1W is theabsolutehumiditydifferencebetweenthecavityandthe surroundings.The interchange rate of air between the cavityand the room relative to the volume of the cavityinfiltration A is evaluated using tracer gas

24、dilutionmethod, as described below.Infiltration by a tracer gas dilution methodIn this method, a small volume of gas is intro-ducedinsidethecavity,anditsconcentrationismea-sured as a function of time. The concentration, C,defined as the ratio of gas volume to air volume,decreases with time t exponen

25、tially;C = C0eAt, (15)where C0is the initial concentration, and A is anegative constant. Equation 15 can be written aslnC = At + lnC0. (16)The value of A can be obtained by fitting theexperimental data to Equation 16.Infiltration experimentsThe infiltration is a function of the temperaturedifference

26、 between the interior and exterior of thecavity Delta1T. Thus, these temperatures must be mea-suredduringtheexperiment.Thiswasdonewiththethermocouples used to measure TO and TI at H/2.Downloaded by Tthus,itisrecommended thatthesensorsupperlimitbe close to 4,000 ppm. The CO2sensor used (10 to4,500 pp

27、m) was located at the center of the freezercavity.Theambientrelativehumiditywasmeasuredusing a data logger (5 to 95%HR, 2.5%HR). Datawererecordedusinganacquisitionsystem.TheCO2was introduced through the water input tube of theicemaker. For sealed experiments, the gasket wassealed with the same EPS f

28、rame and the sealing ex-truded rubber mastic used in previous experiments.Once the temperature at the freezer cavity hadachievedasteadystate,thecorrespondingCO2con-centrationwasrecorded(thisconcentrationiscalledthebasevalue).Then,thegastankvalvewasopenedand, using the pressure regulator, the cavity

29、wasFigure 8. Photograph of the overlapped two identical gasketsused in the second RHL test.filled with gas until the desired maximum concen-tration was obtained. In order to have the best pos-sible resolution, the maximum concentration in theexperiments was about 4,450 ppm. Concentration,temperature

30、s, and relative humidity were recordedevery minute, and the experiment finished when thegasconcentrationapproachedthebasevalue.Figure7 shows the natural logarithm of CO2concentrationas a function of time for baseline and sealed exper-iments. Using Equation 16, A = 0.1123 0.0007h1for the baseline exp

31、eriment, and A = 0.0531 0.0002 h1for the gasket sealed experiment. Theuncertainty of A increases to 0.01 due to test re-peatability as far as how uniformly the door sealsafter each openingclosing process.For both experiments, the average of the abso-lute humidity was W = 0.0075 0.0004 kgw/kga(0.0075

32、 0.0004 lbw/lba), and the absolute humid-ity difference between the cavity and the surround-ings was Delta1W = 0.015 0.0008 kgw/kga(0.015 0.0008lbw/lba).Thesequantitieswereestimatedus-ing the measured ambient absolute humidity duringthe experiments and assuming, based on previousmeasurements,thatthe

33、absolutehumidityinsidethecavity is zero.Infiltration load resultsSubstituting experimental results into Equations914,thegasketinfiltrationloadisestimatedasQing= 0.44 0.19 W (1.50 0.64 Btu/h) for a temper-ature difference between freezer and room of Delta1T= 47.2 0.3C (85.0 0.5F) and an absolutehumid

34、ity difference between the cavity and the sur-roundings of Delta1W = 0.015 kgw/kga(0.015 lbw/lba).For this condition, the infiltration through the gas-ket contributes to the frost in 0.14 g/h (0.0003 lb/h).Consideringthesame Delta1T,butassuminganextremecase where Delta1W = 0.05 kgw/kga(0.05 lbw/lba)

35、 anda room temperature of 40C (104F), the gasket in-filtration load would increase up to Qing= 0.92 0.19 W (3.14 0.65 Btu/h) and the frost to 0.45 g/h(0.0010 lb/h).Reverse heat leak testsRHL tests were performed in a controlled cham-berusingavariationofthesecondmethoddescribedby Ghassemi and Shapiro

36、 (1991) instead of insulat-ing the unit around the gasket area; for the secondtest, a unit with two equal gaskets was used. Theaforementioned test uses a unit with a modifiedDownloaded by T two identical gaskets are overlapped,and the distance between the door and the cabinetis equivalent to the dou

37、ble of the original distance(using only one gasket).Reverse heat leak resultsThe temperature difference between freezer androom was the same as in previous experiments. Thetotal thermal load of the freezer obtained by the firsttest is QfzRHL= 56.80 1.01 W (829 10 Btu/h).Thegasketloadobtainedbythetwo

38、testsisQgRHL=3.0 2.4 W (10.24 8.2 Btu/h).ConclusionsA method is developed to evaluate gasket heattransfer load due to convective, radiative, andconductive mechanisms. It combines a quasi-one-dimensional theoretical model with experimentaland numerical results. This method allows esti-mating heat tra

39、nsfer coefficients along the gasketperimeter. Tracer gas dilution and humidity mea-surements were conducted to evaluate the infiltra-tion load through the gasket. The gasket total ther-mal load is calculated as the addition of these twoloads.Foratemperaturedifferencebetweenfreezerandroom Delta1T = 4

40、7.2 0.3C (85.0 0.5F) andan absolute humidity difference between the cavityand the environment Delta1W = 0.015 kgw/kga(0.015lbw/lba), the gasket total thermal load is Qg= 3.12 0.05 W (10.65 0.17 Btu/h); 86% correspondsto the heat transfer load through the freezer gasketand 14% to the gasket infiltrat

41、ion load. The gasketload obtained in RHL tests with the same Delta1T isQgRHL= 3.0 2.4 W (10.24 8.2 Btu/h). In theRHLtests,theinfiltrationloadisonlyduetothesen-sible load. Thus, to compare the results, the latentheat load is subtracted, which gives Qg= 3.08 0.05 W (10.51 0.17 Btu/h). This value is eq

42、ual toQgRHL, within the uncertainty of the latter. The de-veloped method has the advantage of giving resultswith smaller uncertainty than the RHL method andprovides information about the heat transfer alongthe perimeter of the gasket.It is known that, on RHL tests, the heat transferbetween two gaske

43、ts is different from the one be-tween gasket and cabinet; however, the results aresimilar. The total heat transfer of the freezer gas-ket obtained from the numerical simulation is morethan two times the one calculated with the quasi-one-dimensional model and is close to the upperestimation from RHL

44、tests. This overestimation bythe numerical simulation of the heat transfer couldbe due to the assumptions made at the bottom andthe top boundaries of the gasket internal air cavitiesand that all the perimeter of the gasket has the sameheat transfer behavior. Nevertheless, it can be con-sidered that

45、the overestimation is similar for all aircavities, resulting in no change in the value for thecoefficient c1.From the total thermal load of the freezer ob-tained by an RHL test, the freezer gasket total ther-mal load obtained by the developed model, Qg, rep-resents 5.3% of QfzRHL; and the gasket hea

46、t transferload Qhtrepresents 4.7% of QfzRHL. Ghassemi andShapiro (1991) reported values between 10% and30%. This indicates that the tested gasket has animproved design.NomenclatureA = interchange rate of air between the cavityand the roomc1= proportion of the heat transfer from theoutsidec2= factor

47、related to the heat transfer from thedoor and cabinet wallsC = gas concentrationC0= initial gas concentrationCp = average specific heat of the humid airCpa= air specific heatCpw= water specific heatd = depthofthetransversalsectionofthegas-ket bulkdF= depth of the transversal section of theframedi= d

48、epth of the transversal section of the in-ternal surface of the gasketdo= depth of the transversal section of the ex-ternal surface of the gaskethi= internal surface film coefficientho= external surface film coefficienthw= average water vaporization enthalpyH = freezer cavity heightkeff= effective c

49、onduction coefficientkF= frame thermal conductivityli= length of the ith section of the gasketL = gasket widthLF= frame widthm = infiltration air mass flowq = gasket heat transfer rate per unit lengthDownloaded by T soft-computing or control techniques, such as neural networks, fuzzy logic,genetic algorithms; and the fusion or hybrid of hard and soft control techniques. Part I focused on hard-cont