Quantification of Liquid Water Saturation in a Transparent Single-Serpentine Cathode Flow Channel of PEM Fuel Cell by Using Image Processing

Transcription

1 Quantification of Liqui Wate Satuation in a Tanspaent Single-Sepentine Cathoe Flo Channel of PEM Fuel Cell by Using Image Pocessing S. Niunsin* an Y. Khunaton Depatment of Mechanical Engineeing Chiang Mai Univesity Chiang Mai Thailan *Coesponing Autho: niunsin@hotmail.com Tel. (66) Fax. (66) Abstact:The objective of this eseach is to quantify the ate content on the cathoe sie in vaious PEM fuel cell opeations. It can be eveale by iect visualization in an opeational tanspaent single-sepentine PEM fuel cell. Images of liqui ate accumulate insie the cathoe flo channel ee ecoe by a igital camea to stuy ate flooing in PEM fuel cell. The ate coveage aea in the cathoe flo channel as estimate by an image pocessing technique. The effects of oxygen flo ate cell tempeatue an time evelopment on the ate flooing ee stuie. The esults inicate that excessive lo o high cell tempeatue cause ate to floo into the PEM fuel cell. The inceasing of oxygen flo ate can emove moe liqui ate out of the cathoe flo channel. Hoeve too high oxygen flo ate cause the insufficient ate content to maintain the membane in the hyate state an the amatic ecease of fuel cell s pefomance. The ate flooing i not appea in a single-sepentine cathoe flo channel hen a tanspaent PEM fuel cell has opeate ithin 4 minutes. Keyos: PEM fuel cell; ate management; iect visualization; tanspaent single-sepentine image pocessing.. Intouction Fuel cells ae pesently egae as pomising enegy convesion systems fo electical vehicles an poe stations. PEM fuel cells in paticula have many avantages such as using soli polyme electolytes lo opeating tempeatues col stat-up high enegy efficiency an poe ensity [-3]. The polyme membane in a PEM fuel cell shoul be in a hyate state to facilitate poton tanspot acoss the membane. If thee is not enough ate the membane becomes y an its esistance inceases shaply. Hoeve if too much ate is pesent flooing may occu hich blocks the tanspot of eactants to the eaction sites. Thus ate management in PEM fuel cells is vey impotant an has been mentione in many stuies [3]. In a PEM fuel cell the ate istibution in the membane is etemine by to main mechanisms: electo-osmotic ag an iffusion. In pactice thee is much moe ate esulting in flooing in the cathoe than in the anoe especially at high cuent ensity an lo tempeatue conitions. If the geneate ate is not emove fom the electoe an flo channels at a sufficient ate flooing appeas an the tanspot of eactants is hinee [4]. Many stuies on ate tansfe an ate management have been publishe. Thee ae seveal moels that can peict PEM s pefomance base on iffeences in the level of ate flooing [5-7]. Even though thee have been ne evelopments in selfhumiifying polyme electolyte membanes hich can hyate a PEM fuel cell by geneating ate fom the electochemical eaction [8-9]. Hoeve these stuies i not pesent the images an pattens of the liqui ate in PEM fuel cell. At pesent thee ae many techniques to etect cathoe flooing. One can use global tools such as fully satuate ai at the exit an inceasing the pessue op. Flooing is also associate ith a fuel cell s pefomance. Local infomation about flooing can be given by cuent an tempeatue istibution measuements []. Physical inicatos of flooing such as cuent tempeatue pessue op an elative humiity ae also tools to etect ate flooing in a fuel cell. Vaious imaging techniques can be use to investigate to-phase flo ynamics insie a fuel cell. These techniques ae iect visualization [-5] neuton aiogaphy [6] X-ay mico tomogaphy [7] an magnetic esonance imaging [8]. Although iect visualization equies a special cell esign it is vey attactive expeimental technique as visual infomation offes the avantage of stuying the tophase phenomena at iffeent levels of opeating conition on ate istibution an ate flooing in PEM fuel cells [2]. Hoeve the visualization pimaily povies qualitative ata an thee have been only a fe epots about quantifying the ate content though this technique []. The pupose of this pape is to quantify the accumulate ate in a tanspaent single-sepentine cathoe flo channel. The effects of cell tempeatue cathoe gas flo ate an time evelopment on the ate flooing ee also examine. The ate images ee ecoe by igital camea. Since the igital images pesente to-imensional infomation the ate content in this pape must be measue in tems of ate coveage aea hich can be quantifie by using image pocessing technique. 2. Image pocessing A igital image is efine as a to-imensional function f(xy) hee x an y ae spatial cooinates an the amplitue of f at any pai of cooinates (xy) is the intensity o gay level (-255) of the image at that point. A igital image is compose of a finite numbe of elements each of hich has a paticula location an gay level value. These elements ae efee to as image elements o pixels. Pixel is the tem most iely use to enote the elements of a igital image. The igital image can be consiee as a lage aay of pixels. Thus the igital image can be manipulate ith matix opeations in image pocessing outines to evaluate any impotant infomation. The amount of ate in the cathoe flo channel of a PEM fuel cell can be quantifie by using an image pocessing outine. The ate appeaance in the cathoe flo channel as epesente by the ate pixels in the spatial omain of the igital image. The ate image W as sepaate fom backgoun image I by subtacting the image of eaction state I escibe in Eq. () fom y state image I in Eq. (2). The calculation of ate content fom the images may be itten as the folloing equations. I x y I x y x y () I W I W x I I (2) (3) 2 By Jounal of Sustainable Enegy an Envionment 29

2 hee x an x ae noise functions of the y an eaction states espectively. These may be occue fom eflection an istubance of light souce fequency (5 Hz). Hoeve the tem of noise function in Eq. (3) may be negligible because the subtaction of both noise functions is vey small. So that Eq. (3) can be eitten as Eq. (4) ithout the tem fo noise function. W I I (4) The numbe of ate pixels in Eq. (4) can be etemine by using a histogam of the image x W. The histogam is a gaph inicating the numbe of times each gay level occus in the image. The pobability ensity of ate pixels P can be obtaine by nomalize histogam ith iviing all ate pixels nk by the total numbe of pixels n in the image as shoe in Eq. (5). P nk n (5) hee k is gay level of ate. Theefoe the ate coveage aea A on the cathoe sie can be quantifie fom the image by multiplication of A P ith active aea size of A c. A P (6) c Hoeve the capable ate coveage aea is only a pat of gas flo fiel channel aea hich ose not inclues the ibs aea. Theefoe fom a single-sepentine flo fiel esign of cathoe sie the maximum aea is covee by ate about 55 pecent of the hole active aea. The bass plate as machine as a gas flo fiel plate in the cathoe sie an conventional gaphite as use as a gas flo fiel plate in the anoe sie. Tanspaent acylic ith thickness of 5 mm as use as the en plate at the cathoe sie. Membane electoe assembly (MEA) as mae fom Goe in 562 seies. Thee as.6 mg/cm 2 at cathoe an.4 mg/cm 2 at anoe. The membane thickness as 25 m. The size of physical active aea as 25 cm 2. When the machine bass plate as assemble beteen the tanspaent acylic an MEA the 3M 468MP ahesive film as use as sealing stip beteen the flo fiel channel an the tanspaent acylic to obstuct the gas coss ove the ibs an leak out fom the cathoe active aea. A tanspaent ino as set up in the cathoe clamping plate so that e ee able to investigate the ate buil-up an flooing insie the PEM fuel cell as shon in Fig. 2. To cicle electical heate pas ee attache on the anoe en plate an cathoe flo fiel plate. A tempeatue contolle an tempeatue senso ee use to contol the cell tempeatue. 3. Expeimental A tanspaent PEM fuel cell as esigne in a singlesepentine flo channel (.5 mm in channel ith 2 mm in epth an.5 in ib ith). Fig. shos the schematics aing of the cathoe sie of a tanspaent PEM fuel cell. Figue. The schematic aing of the cathoe flo fiel plate. Figue 2. A tanspaent single-sepentine PEM fuel cell. Fig. 3 shos the expeimental set-up. The expeimental set-up consists of: (i) a fuel an oxiant eactant supply system; (ii) a fuel cell test system; (iii) a igital camea ecoing system. Pue hyogen an oxygen ee use as fuel an oxiant eactant espectively. Figue 3. The iagam of expeimental set-up in a tanspaent PEM fuel cell. Both hyogen an oxygen gas to the fuel cell ee not humiifie so that the ate that as obseve in the cathoe flo channel of a tanspaent fuel cell has been geneate only by the electochemical eaction. Gas flo ates ee contolle by mass flo contolles (MKS Type MB-) ith a pecision of.%. A nitogen puge system as use to clean out the fuel an oxiant eactant left in the test system pipelines an fuel cells. The fuel cell cuent loaing as contolle by electonic loa (TDI RBL448 Seies). The ate flooing in the cathoe flo channel as ecoe by igital camea (Cannon GT). Duing each expeiment the eaction aea of the cathoe sie as fist photogaphe in its y state. Photogaphs ee taken evey 5 minutes afte the fuel cell eaction occue at a constant cuent loaing of 5 A (2 ma/cm 2 ). In oe to pevent the voltage of the fuel cell opeation opping too lo the cuent loaing in this expeiment must be loe than the nomal opeating cuent ensity (4-6 ma/cm 2 ) because the fuel cell as opeate in a non-humiifie state an the tanspaent configuation has moe ohmic loss fom metallic coosion of the bass plate [9 2] than oes the conventional gaphite configuation. The accuacy of using the image pocessing technique as achieve by isposition of all image components in the same fame efeence an also une the same intensity of light souces. As a esult the subtaction image beteen the eaction s state image an y state image alloe only the ate image in the cathoe flo channel. Finally befoe changing the opeating conitions nitogen gas as use to puge the gas flo channel to econition it back to its y state. 3 2 By Jounal of Sustainable Enegy an Envionment

3 4. Results an iscussion The ate flooing images on the iffeent fuel cell opeating conitions ee ecoe by igital camea. Image pocessing as use to manipulate the igital images in this expeiment. Fig. 4 (a) an 4 (b) sho an example of the ate image hich as manipulate by image pocessing techniques. 4. The effect of cell tempeatue The effect of cell tempeatue on a tanspaent PEM fuel cell has been escibe in othe papes. Hoeve most of them have analyze this phenomenon ith qualitative ata only [- 3]. Fig. 5 shos the effect of cell tempeatue in the cathoe flo channel. All images ee ecoe at 2 minutes afte the fuel cell as loae at a constant cuent of 5 A. A tanspaent fuel cell as opeating at the conitions of atmosphee pessue oxygen gas flo ate of 4 ml min - (stoichiometic atio of 2.8) an hyogen flo ate of 6 ml min - (stoichiometic atio of.63). Thee as no extenal humiifie supplie to eithe sie of the PEM fuel cell in oe to ensue that any ate as geneate fom the fuel cell eaction. Fig. 5 shos that the liqui ate in the flo channel at lo tempeatue as much moe than at high tempeatue. When the fuel cell opeate at a tempeatue loe than 6ºC thee as a geate ate coveage aea in the cathoe flo channel. The accumulate ate insie a tanspaent fuel cell i not only euce the cathoe flo channel aea an effective electochemical eaction site but it also obstucte mass tanspot. At the opeating tempeatue i.e. 25ºC the cell s pefomance oul be lo because the flo channel oul be fille ith the liqui ate. By inceasing the cell s tempeatue the ate coveage aea in the flo channel as ecease accoing to the ate quantifie by image pocessing in Fig. 6. Inteestingly Fig. 6 shos the ate coveage aea at 25ºC to be less than that at 4ºC. The eason is that hen the nitogen gas as use to puge the accumulate liqui ate in the cathoe flo channel to econition to the y state befoe changing the tempeatue fom 25ºC to 4ºC thee as a small amount of liqui ate emaining in the gas iffusion laye (GDL). So this ate as ae to the ate pouct of the next fuel cell opeating conition to aise the amount of ate coveage aea at 4ºC highe than at 25ºC. But this phenomenon as less ominating at the highe opeating tempeatue because the emaining liqui ate ha eceive moe themal enegy fom the heate of the tempeatue contolle. So moe ate vapo as geneate hich oul be moe easily leave fom the poous electoe an ive faste out of the cathoe flo channel befoe the electonic loa as tune on to geneate the (a) ate image befoe pocess (b) ate image afte pocess Figue 4. Example of ate image manipulate by image pocessing technique. (a) 25ºC (b) 4ºC (c) 6ºC () 7ºC Figue 5. Effect of cell tempeatue on ate appeaance in cathoe flo channel. 2 By Jounal of Sustainable Enegy an Envionment 3

4 electochemical eaction an ate pouct in the next cell tempeatue conition. As a esult of the opeating tempeatue ising the vapo conensation ate as much sloe than that at lo tempeatue. Thee as only a little ate in the cathoe flo channel at the highest cell tempeatue of 7ºC as shon in Fig. 6. In Fig. 7 hen the cell tempeatue as inceasing fom 25ºC to 6ºC the fuel cell s pefomance oul be highe but the ate content began to ecease a little at tempeatue of 6ºC because thee as moe ate vapo geneate in the cathoe flo channel. The ate content ecease apily at the highest tempeatue of 7ºC. As the loest content of ate at 7ºC the membane oul become y an its esistance oul incease shaply hich gave the loest pefomance. Theefoe the appopiate cell tempeatue oul be in the ange of 4 6ºC hich povie the best fuel cell pefomance. 4.2 The effect of oxygen flo ates The cathoe gas flo ate can contibute to ate emoval. In oe to euce the concentation losses fom ate flooing the stoichiometic atio must be at least 2 [2] but if the cathoe eactant gas is in the fully hyate state the stoichiometic atio must be moe than 43 to avoi ate flooing in the cathoe gas flo channel [5]. In this stuy case the oxygen gas as not humiifie so that the loest stoichiometic atio as efine at highe than 2. Fig. 8 shos the images of conensation of liqui ate in the cathoe flo channel at iffeent oxygen flo ates. These photos ee taken at a cuent of 5 A afte the fuel cell ha opeate fo 2 minutes at the ambient pessue an cell tempeatue of 25ºC. The oxygen flo ates ee 4 ml min - 7 ml min - 5 ml min - an 28 ml min - an accoingly the stoichiometic atios ee an 5.24 espectively. Wate coveage aea (cm 2 ) Tempeatue ( o C ) Voltage (V) Figue 6. The ate coveage aea in cathoe flo channel of iffeence cell tempeatues. This moeate ange of cell tempeatue as also ecommene by Liu et al. [3]. Wate management poblems such as ate flooing o membane ehyation can lea to a op in cell s pefomance. Theefoe ate management is impotant to fuel cell pefomance hich is also elate to themal management. 25 oc 4 oc 6 oc 7 oc Figue 7. The cell voltage in iffeence cell tempeatues at a constant cuent loaing of 5 A. When the stoichiometic atio of oxygen as incease fom 2.8 to 3.8 the ate content in the cathoe flo channel ecease to almost fifty pecent an tene to moe ecease in the oxygen stoichiometic atio highe than 3.8 as shon in (a) Stoi 2.8 (b) Stoi 3.8 (c) Stoi 8.6 () Stoi 5.24 Figue 8. Effect of oxygen flo ate on ate appeaance in cathoe flo channel By Jounal of Sustainable Enegy an Envionment

5 Fig. 9. Theefoe the ate flooing in a flo channel as moe seious at the oxygen stoichiometic atio of 2.8 because oxygen flo ate as too lo to emove the ate occuing fom the electochemical sites of the PEM fuel cell. This esult coespons to the stuy of Weng et al. [4] ho epote that ate flooing as obvious at a stoichiometic atio of 2 on the fuel cell opeation ith y oxygen flo ate conition. In the case that the oxygen stoichoimetic atio as highe than 3.8 thee is not enough ate content to keep the membane in hyate state accoing to the expeimental esult on the oxygen stoichiometic atio of 5.24 hich povie the loest cell pefomance as shon in Fig.. In this case since the oxygen gas has not been humiifie befoe enteing the fuel cell the geneate ate s effect on fuel cell opeation as insignificant. So an appopiate amount of oxygen flo ate must be supplie to emove the excess liqui ate out of the cathoe flo channel but still some ate emaine fo fuel cell opeation. Wate coveage aea (cm 2 ) Oxygen stoichiometic atio Figue 9. The ate coveage aea in cathoe flo channels ith iffeent oxygen flo ates. It is impotant to note that the ate coveage aea at the oxygen stoichiometic atio of 3.8 is less than that at 8.6 as shon in Fig. 9. Hoeve this esult contaicts the stuy of ate-flooing behavio in cathoe flo channel by Weng et al. [4] ho epote that the liqui ate as easily emove at a high cathoe gas flo ate an the eaction aea as not hinee by flooing. To explain this expeimental eviation the image pocessing technique has a limite ability to etection some tanspaency oplets [2] so that the quantifie ate coveage aea at oxygen stoichiometic atio of 3.8 as less than the eal existent ate accumulate in the cathoe flo channel. This limitation mae euce the ate content at oxygen stoichoimetic atio of 3.8 less than at that 8.6. Hoeve this poblem of limite etection of tanspaency oplets can be coecte by ege etection techniques [22] hich ill be pesente in a futue epot. Otheise the effect of the emaining ate as montione befoe in section 4. as less ominant on the ate content in this eviation case because the fuel cell as opeate in the constant tempeatue ith a highe oxygen flo ate. Voltage (V) Stoi 2.8 Stoi 3.8 Stoi 8.6 Stoi 5.24 Figue. The cell voltage in iffeence oxygen flo ates at a constant cuent loaing of 5 A. 4.3 Effect of fuel cell opeation time The ate geneation occue continuously as the fuel cell as in opeation. It as monitoe an ecoe by igital camea egaless of tempeatue an oxygen flo ate. A tanspaent PEM fuel cell as loae at a constant cuent of 5 A constant cell tempeatue of 5ºC non-humiifie eactant gases an atmospheic pessue conition. The oxygen gas flo ate as 4 ml min - (stoichiometic atio of 2.8) an the hyogen flo ate as 6 ml min - (stoichiometic atio of.63). Fig. shos the pocesse images of liqui ate in the tanspaent cathoe flo channel at iffeent times. All images ee ecoe at minutes 2 minutes 3 minutes an 4 minutes at constant cuent loaing. (a) min (b) 2 min (c) 3 min () 4 min Figue. Effect of opeation time on cathoe ate buil-up in the flo fiel channels. 2 By Jounal of Sustainable Enegy an Envionment 33

6 It as obseve that the small ate oplets conense on the inne suface of tanspaent acylic afte cuent loaing fo minutes. Beteen 2 minutes an 4 minutes of the fuel cell s opeation the ate oplets incease in size an accumulate in the flo channel. The quantification of ate coveage aea iffeent at times is shon in Fig. 2. The ate coveage aea in the cathoe flo channel ecease gaually hile the ate oplets incease in size afte 2 minutes of opeation. Due to a single-sepentine flo fiel patten being use in this eseach as the Reynols numbe as highe than ith multiple sepentine channels at the same oxygen flo ate because the flo ate pe coss section of a single channel as highe [2] an also because the fuel cell as assemble in a single cell configuation. It as easy to ive the lage size of ate oplets in the cathoe flo channel une an oxygen stoichiometic atio of 2.8. Thus ate flooing in the tanspaent cathoe flo channel as not obseve an the fuel cell pefomance cuve i not sho any signal of voltage op uing fuel cell opeation time ithin 4 minutes as shon in Fig. 3. At this steay state (constant voltage) the fuel cell as opeate at the same balance conitions. In spite of the ate content eclining continuously as shon in Fig. 2 it still maintaine the ionic conuctivity of the membane to keep the fuel cell s pefomance constant. The effect of the fuel cell s opeating time as also iscusse by Liu et al. [3] ho epote that the fuel cell opeate at a constant voltage ithin the fist peio of opeating time (about 3 4 minutes) befoe the voltage as oppe to a ne balance opeating conition at a futhe peio of opeating time. Hoeve the steay state in this stuy as maintaine fo a longe peio than Liu et al. [3] ha epote because the fuel cell as opeating at a loe cuent ensity. Wate coveage aea (cm 2 ) Figue 2. The ate coveage aea on iffeent fuel cell opeating time in cathoe flo channel. 5. Conclusion The pesent stuy has exploe the possibility of using image pocessing techniques to quantify the ate content in tems of ate coveage aea on the cathoe flo channel of a PEM fuel cell. Quantification of ate coveage aea in the cathoe flo channel is citical to establishing a basic unestaning of the to-phase flo an flooing occuence in PEM fuel cells. The effect of cell tempeatue oxygen flo ates an time evelopment on the ate flooing ee analyze by iect visualization in a tanspaent single-sepentine cathoe flo channel. The igital images fom cathoe flo channel ee ecoe to quantify the ate coveage aea at iffeent opeating conitions by using image pocessing outines. The pocesse images povie infomation about the ate flooing at iffeence fuel cell opeating conitions as follos: () The image of the ate at the loest cell tempeatue of 25ºC shos thee as a lot of ate coveage aea in the cathoe flo channel. This coul lea to mass tansfe limitation because the conensation of the liqui ate stays in the channel an occupies the path of the gas to the eaction sites. Hoeve at the highest cell tempeatue of 7ºC thee as insufficient ate content to humiify the membane in the hyate state. Thus excessive lo o high tempeatue can lea to eteioation of the fuel cell s pefomance. The appopiate cell tempeatue as in the moeate ange of 4 6ºC hich povie the highe pefomance. (2) The cathoe gas flo ate can contibute to ate emoval. By inceasing the stoichiometic cathoe flo ate fom atio of 2.8 to 3.8 the ate content accumulate in the cathoe flo channel can be euce by almost fifty pecent an also lea to highe cell pefomance. In this stuy the eactant as not humiifie. When the stoichiometic atio of oxygen eache 8.6 the cell s pefomance ecease ith the loest pefomance at the oxygen stoichiometic atio of 5.24 because the membane ha became too y an its conuctivity to poton ions ha been euce. (3) The time evelopment of ate flooing i not affect on the cell s pefomance an the ate flooing i not appea in the single-sepentine cathoe flo channel. In spite of the ate coveage aea in the cathoe flo channel slightly eceasing ithin 4 minutes this ate content as able to maintain the ionic conuctivity of the membane to keep the fuel cell opeating at the same balance conitions. Despite thei initial pomise image pocessing techniques equie much moe evelopment an efinement fo application to a PEM fuel cell. A futhe evelopment of ege etection techniques an also impovements in the esolution of image an photogaphy technique oul incease the accuacy of this stuy. Acknolegements Voltage (V) The autho oul like to acknolege the folloing suppotes of this stuy; Depatment of Mechanical Engineeing Faculty of Engineeing Chiang Mai Univesity; The Gauate School Chiang Mai Univesity an Enegy Policy an Planning Office Ministy of Enegy Thailan. Refeences Figue 3. The cell pefomance uing opeation of a tanspaent single-sepentine PEM fuel cell. [] Chalk SG Milliken JA Mille JF The US Depatment of Enegy investing in clean tanspot Jounal of Poe Souces 7 (998) [2] Mehta V Coope JS Revie an analysis of PEM fuel cell esign an manufactuing Jounal of Poe Souces 4 (23) By Jounal of Sustainable Enegy an Envionment

7 [3] Pasaogullai U Wang CY Liqui Wate Tanspot in Gas Diffusion Laye of Polyme Electolyte Fuel Cells Jounal of Electochem. Soc. 5 (24) A339. [4] Miachon PA Intenal hyation H 2 /O 2 cm 2 polyme electolyte membane fuel cell Jounal of Poe Souces 56 (995) [5] Tung VN Ralph EW Wate an Heat Management Moel fo Poton-Exchange-Membane Fuel Cells J. Electochem. Society 4 (993) [6] Spinge TE Wilson M Gottesfel S Moeling an expeimental iagonostic in polyme electolyte fuel cells J. Electochem. Society 4 (993) [7] Spinge TE Zaozinski TA Gottesfel S Polyme electolyte fuel cell moel J. Electochem. Society 38 (99) [8] Yang TH Yoon YG Kim CS Kak SH Yoon KH A novel pepaation metho fo a self-humiifying polyme electolyte membane Jounal of Poe Souces 6 (22) [9] Kaka SH Yang TH Kimb CS Yoona KH The effect of platinum loaing in the self-humiify polyme electolyte membane on ate uptake Jounal of Poe Souces 8 (23) [] Hakenjos A Muente H Wittstat U Hebling C A PEM fuel cell fo combine measuement of cuent an tempeatue istibution an flo fiel flooing Jounal of Poe Souces 3 (24) [] Li H Tang Y Wang Z Shi Z Wu S Song D Zhang J Fatih K Zhang J Wang H Liu Z Abouatallah R Mazza A A evie of ate flooing issue in the poton exchange membane fuel cell Jounal of Poe Souces 78 (28) 3-7. [2] Spenjak D Pasa AK Avani SG Expeimental investigation of liqui ate fomation an tanspot in a tanspaent sing-sepentine PEM fuel cell Jounal of Poe Souces 7 (27) 334. [3] Liu X Guo H Ma C Wate flooing an to-phase flo in cathoe channels of poton exchange membane fuel cells Jounal of Poe Souces 56 (26) [4] Weng FB Su A Hsu C-Y Lee C-Y Stuy of ateflooing behaviou in cathoe channel of a tanspaent poton-exchange membane fuel cell Jounal of Poe Souces 57 (26) [5] Ous T Acoumanis C Visualisation of ate oplets uing the opeation of PEM fuel cells Jounal of Poe Souces 73 (27) [6] Pekula N Helle K Chuang PA Tuhan A Mench MM Benize JS Ünlü K Stuy of ate istibution an tanspot in a polyme electolyte fuel cell using neuton imaging Nuclea Instuments an Methos in Physics Reseach Section A: Acceleatos Spectometes Detectos an Associate Equipment 542/-3 (25) 34-4 [7] Sinha PK Halleck P Wang CY Quantification of liqui ate satuation in a PEM fuel cell iffusion meium using X-ay micotomogaphy Electochemical an Soli- State Lettes 9/7 (26) A344-A348. [8] Tsushima S Teanishi K Hiai S Magnetic esonance imaging of the ate istibution ithin a polyme electolyte membane in fuel cells Electochemical an Soli-State Lettes 7/9 (24) A269-A272. [9] Yoon W Huang X Fazzio P Reifsnie KL Akkaoui MA Evaluation of coate metallic bipola plates fo polyme electolyte membane fuel cells Jounal of Poe Souces 79 (28) [2] Tafik H Hung Y Mahajan D Metal bipola plates fo PEM fuel cell-a evie Jounal of Poe Souces 63 (27) [2] Laminie J Dicks A Fuel Cell Systems Explaine (2) Wiley West Sussex Englan pp [22] Mcane A Digital Image Pocessing ith MATLAB (24) Couse Technology a ivision of Thomson Leaning Inc Unite States of Ameica pp By Jounal of Sustainable Enegy an Envionment 35