WIXLIB

A EFRIGERATORT. Koizumi, M. Takahashi, T. Uchida,Y. Kanazawa,and M. SuzukiHiratsuka Research Laborat@ry,SumitomoHeavy Industries, Ltd., JapanABSTRACTA small and light weightheat exchangerused for small heliumrefrigeratorhas been developedby SumitomoHeavyIndustries,Ltd. This heatexchangeris a laminatedmetal heat exchangerwhichconsistsof perforatedaluminummetal platesand glassfiberreinforcedplasticseparators.Thesize is from I00 mm to 28 mm in diameterand about 300 mm in length.Theweightis from 2.5 kg to 0.6 kg.Also it can be used betweenroom temperature and liquidheliumtemperature.The thermalefficiencyobtainedhasbeen more than 96%.The heat exchangerhas been practicallyused foron-boardheliumrefrigeratorin vitatedtrains.INTRODUCTIONCounterflowheat exchangerof high thermalefficiencyis requiredforrefrigerationand liquefactionsystems.In recentyears,many type heatexchangers,for example,Hampson,Collinsand Plate-fintype have beendeveloped.But these heat exchangersdo not fit in with small refrigeratorsbecausethe heat transferarea per unit volumeof these heat exchangerisnot large enough.The small refrigerator,for requiresveryefficientheat exchanger.Such heat exchangermustcharacteristics[1],[2].i. Large heat transferunit weight.surface179areaperrefrigeratorin supersmall and high thermalpossessthe followingunitvolumeandper

2. Very small longitudinalthe exchanger.3. Uniform distributionexchanger.heat conduction through the walls inof flow throughout any cross section of4. High resistance to shock and vibration.Wehave developed the small and light weight heat exchanger whichfulfilledabove requirements and applied it to on-board helium refrigeratorin JapaneseNational Railways' superconducting magnetic levitated trains.This paper describes the development of the laminated metal heatexchanger and its application for small refrigeratoruse.CONSTRUCTIONThe constructionof laminatedmetal heat exchangeris illustratedinFig. i.This heat exchangerconsistsof a large numberof parallelperforated aluminummetal plates,plasticseparatorsand epoxy adhesives.The perforatedmetal platesand separatorsare alternatelystackedand bondedwith adhesivesto build up the multilayer-body,lleaders made of aluminumare bondedto both faces of the multilayer-body.Then, headersandmultilayer-bodyare insertedinto a thin walledstainlesssteel vesselwithpipe fittings.One end of the fittingis weldedto the stainlesssteelvesseland the other end is bondedto the header.Therefore,there isno gas leakageto thea alledinIn this exchanger,gas flow passagesare dividedinto eight sectionsand gas flows longitudinallyin counterflowpattern.Heat transferslaterally from hot gas to cold gas throughthe perforatedplates.The perforatedheat transferplate is made of aluminumwhich has high thermalconductivityand is 0.3 mm in thickness,with 0.5 mm diameterholes.Athickness-to-diameterratiois 0.6inthisperforatedplate.It is representedthat the desirableratio is in the range 0.5 to 1.0 becausethermaland hydrodynamicboundarylayersare brokenup beforethey have achanceto becomefully developed,which resultsin high heat transfersurfacecoefficients[i].The separatoris made of glassfiberreinforcedplasticwhich has lowthermalconductivityand its thicknessis 0.4 mm.Therefore,the longitudinal heat conductionthroughthe walls in the heat exchangeris very samll.firstFigurestage2 is a photographof this heat exchanger.heat exchangerfor on-boardrefrigerator.180This is appliedtoThe size is 80 mm

in diameterand 250 mmtransferarea per unitin lengthvolumeinand the weightis about 2.2 kg.The heatthis heat exchangeris about i000 ma/m 3.PERFORMANCETESTAPPARATUSThe heat transferand frictionloss performancetest apparatusisshown in Fig. 3.The test has been performedby using heliumgas betweenroom temperatureand the liquid nitrogentemperature.The pressureof thehigh pressureline and low pressureline are 1.57 MPa and 0.I MPa respectively.High pressureheliumgas flows throughtest heat exchangerand isthrottledto low pressureby throttlevalve and flows back to heliumcompressorthroughthe heat exchanger.The temperatureof gas flow ismeasuredby copper-constantanthermocouple.The pressuredrop is measuredby differentialpressuregage.EXPERIMENTALDATAThe heat exchangershown in Fig. 2 has been tested.The hole diameterof perforatedplate is 0.5 mm and open area ratio of the high pressurepassageis 15% and low pressureis 30%.The gap betweenperforatedplatesisabout0.6 mm.The effectivenessand pressuredrop versusthe mass flow rateof heliumare plottedin Fig. 4.It shows that when heliumgas flow rate is1 g/sec,96.5% effectivenessis obtainedand pressuredrop is 0.008 MPa.Anothertype laminatedmetal heat exchangerwhich we have fabricatedisshownFig. 5.The size of this heat exchangeris 28 mm in diameterand300 mm in length.Headersand multilayer-bodyare insertedinto the stainless steel vesseland their weightis about 0.6 kg.Obviously,it is muchsmalland light in weightcomparedwith the heat exchangershown in Fig. e is very simple.But for heat exchangerit is importanttothe effectiveadhesivestrength,gas tight and good durability,this exchangeris constructedwith adhesivebonding.181des-

So, we have selected many kinds of adhesives and tested their adhesivestrength and gas leakage at low temperature to choose the best adhesive amongthem. The test results of epoxy adhesive which we applied to fabricatethe heat exchanger is explained in the following.ADHESIVESTRENGTHWemeasuredthe strength of two different test specimens, one isaluminum-aluminumadhesive bonded joint, the other is aluminum-GFRP-aluminumadhesive bonded joint.The shearing strength versus temperature is shownin Fig. 6. Theresults indicate that the shearing strength increases as the temperaturedecreases. The shearing strength of AI-GFRP-A1at liquid nitrogen temperature is about 15 MPaand the specimens failed in the adhesive. Also shearingstrength denoted by triangles showsthe test results after 7 thermal cyclesbetween room temperature and liquid nitrogen temperature.Fig. 7 presents the tensile strength versus temperature. The tensilestrength of AI-GFRP-AI specimens at liquid nitrogen temperature is about50 MPa. Somespecimens denoted by squares were cycled 235 times betweenroom temperature and liquid nitrogen temperature after one year has passedsince they were bonded. Their tensile strength were nearly equal to theother specimens. Also the failed surface displayed a cohesive failure inadhesive.EFFECTOF THERMALCOOLINGThe heat exchanger is also requested to have high reliabilityfor gasleakage. The thermal cycle test of the heat exchanger has been performedand examinedits reliability.The thermal cycle test apparatus is Fig. 8.The test was conducted by cooling the exchanger between room temperatureand liquid nitrogen temperature. The test vessel in which test heatexchangers are put is immersedin liquid nitrogen and the heat exchangersare cooled to liquid nitrogen temperature. Then, the vessel movesto upand heat exchangers are heated up to room temperature by the heater.The temperature of the heat exchanger is measuredby copper-constantanthermocouple. Also one thermal cycle between room temperature and liquidnitrogen temperature takes about 2 hours. During the test, high pressureand low pressure passages are pressurized with 1.57 MPaand 0.i MPaheliumrespectively.Helium gas leakage flow is measuredwith a soap bubble flowmeter attached to the low pressure side.182

Table 1 summarizesresults of the thermal cycle test.Before coolingthe heat exchanger, helium gas leakage flow high pressure passages to lowpressure passages was measuredto be less than 1X iO4Pa m3/sec.After O 4 Pa m3/sec.CONCLUSIONFrom these test results,can meet the requirementsandrefrigerator"it is confirmedthat thisfulfillthe characteristicsheatforexchanger" on-boardA photographof on-boardheliumrefrigeratorwhich we have developedis shownin Fig. 9.This refrigeratoris ClaudeCycle and consistsof tworeciprocatingexpantionenginesand five laminatedmetal heat exchangers.Therefrigeratorcapacityis5 W at4.4K.Thesizeis300mmindiameterand 800 mm in lengthand the weightis about 40 kg.This refrigerator has been installedin the test vehicleand is presentlybeing testedat MiyazakiTest Track in JapaneseNationalRailways.REFERENCESI. R.B.Fleming," A CompactPerforated-PlateCry. Eng., vol. 14, 197-204(1968)2. G.Vonk,Philips" A CompactHeatTechnicalReview,Exchangervol. 29,HeatExchangerof High Thermalno. 5, 158-162"; Adv.Efficiency(1968)";DISCUSSIONQuestionby W.L.Swift,tensilestrengthand by F.J.Kadi,potentialleakageofCreareP D; Haveof epoxy adhesiveI haveyou testedshear strength,jointsbelowliquid nitrogennot.Air Productsheliuminto& ChemicalInc.;vaccumspace?Answerby author;Thereis no leakagebecauseheadersand multilayer-bodyof heata thin walledstainlesssteel vessel.183of heliumexchangerWouldyouinto vaccumare insertedcommentspace,into

r(A[)StainlessHot gas passageLHeader(At)ColdgasiHotEpoxygas outletadhesive,8OFig.1TheFig. 2constructionlaminatedofmetallaminatedmetalheat exchanger.184heatexchanger.prate

MassflowmeterPdPHe Compressor71.57MPQ; Pressure; DifferentialPressure(E) ( ) ; Thermocouple(Boi Iof'{ " 1 ,,flow met erl J. .r. J rI rogen)Vessel- lSchematic diagram of heat transferfriction loss performance test.and1000.015HoleCLdiam; 950.01OpenareardtioHigh press,passageLI/1L/Ir 15%CLOI,"ElJ L.DI/1I/1Pw139003.005o.'250.%MassFig. 4Performancegageo:7l lorate(g/sec)flowof laminated185metal1. 5i.5heat exchanger.gage

Fig.Table5Laminatedi.Resultmetalofthermalofcyc I .Pressure(MPa)Number! htowleakage(Pom than1 x 10 -41.570.1lessthan1 x 10 .41.5'70.1lessthan1 x10.42.300.1lessthan1 x 10 - "42350235186

ooIDioX00-. 0, oooIi:Icoi,.DoooICCoi u. i-ZuO.Eo n'IIn--I"o. Io60,00o6'--.1"LP(od )ql6uaJ}s c/ID. -(c0I(D "Ul6 xl6el!SUaJ.,-I0o' o'.X 0ii8I O&oohIWI-4444 I -., i-, I , .,o6,( cllal)q } Bu'aJ6}sBu!JOeLIS187I r,I(:II:la- ,DI:II. . .!.,[I0r,rqm

I---1I11.5?MPoL4x Buf/fertctnkq- Buffi:er tank:1(, )(cMeasurem): ;: L '' 'I-\]l 'TestDewor(LiquidFig. mFig. 9\of thermalRefrigerator188cycle test.

Helium gas leakage flow is measuredwith a soap bubble flow meter attached to the low pressure side. 182. Table 1 summarizesresults of the thermal cycle test. Before cooling the heat exchanger, helium gas leakage flow high pressure passages to low pressure passageswas measuredto be less than 1X iO4Pa m3/sec. After 235