Simulation of PEM fuel cells by OpenFOAM

Transcription

1 Project Report 2014 MVK160 Heat and Mass Transport May 15, 2014, Lund, Sweden Simulation of PEM fuel cells by OpenFOAM Jiatang Wang Dept. of Energy Sciences, Faculty of Engineering, Lund University, Box 118, Lund, Sweden ABSTRACT Proton exchange membrane (PEM) fuel cells are known as environmental friendly energy conservation device, and have the potential to be suitable alternative power sources. The cost and durability of a PEM fuel cell are strongly affected by the involved transport phenomena and reactions and there are two of the major challenges to be overcome before commercialization. Modeling and simulation are crucial for the cell design and optimization and various add-on fuel cell modules are available in commonly-used commercial CFD codes: FLUENT, STAR-CD and COMSOL Multiphysics. However, the length scale of a PEM fuel cell s main components ranges from the micro over meso to the macro level. The various transport processes at different scales sometimes cannot be simulated simultaneously by these codes. On the other hand, physical properties of materials used in MEA (membrane electrolyte assembly, consisting of catalyst layers, gas diffusion layers and membrane) play an important role on the cell performance. Therefore coupling the multiscale structural and transport characteristics of the materials will be the most effective way to understand the electrochemical reactions and transient transport phenomena in PEM fuel cell. OpenFOAM (Open Field Operation and Manipulation) is an open source finite volume code and has an object-oriented design written in C++, which allows implementing own models and numerical algorithms. Furthermore, it is possible to insert other models, e.g., particlebased models, to the OpenFOAM CFD Toolbox. Thus OpenFOAM has the potential to meet the requirements faced in PEM fuel cell simulations as mentioned above. In this paper, various literatures simulated by OpenFOAM are outlined and reviewed. The potential methods and challenges coupling OpenFOAM with other modeling techniques are also discussed and highlighted. INTRODUCTION Proton exchange membrane (PEM) fuel cell fueled by hydrogen is considered to be the promising device for the direct conversion from chemical energy into electrical energy. The first commercial used of the typical of PEM fuel cell was in 1960 s Gemini space program by NASA [1]. Owing to its advantages of efficient energy conversion, high power density, environmental friendliness, tremendous research programs worldwide promote PEM fuel cell as power sources that could replace internal combustion engines for sub-watt and M-Watt applications in transportation, heating, manufacturing, and communication, e. g. portable and stationary applications [2, 3]. Cost and durability are supposed as the main challenges to the development and commercialization of PEM fuel cells [4]. During the past decades, the performance of PEM fuel cells have been improved significantly resulting from the research in the different aspects, i.e. implementation of innovative materials [5], optimized designing [6] and modeling [7]. The research of materials is to decrease the usage of expensive Pt, increase the density of activity sites in porous structure and improve conductivity of electrons or protons, and synthesize the porous materials with excellent transport ability for different species. Regarding to the second field, aim to design the optimized geometry for improved fuel cell performance. As the deeply understanding of transport phenomena within the fuel cell of modeling is to provide the directions on fuel cell manufacturing and operation. Even though many effects have been taken out for comprehensive improvement for the performance, the barriers for widely used of PEM fuel cells are still remained. The energy resources crisis all over the world also promotes the development of PEM fuel cells. Especially the energy crisis happened in 1973 [1]. Thus the new energy, e.g. PEM fuel cell, attracted more and more attentions in the modern society of resources limited. With the rapid development of computer technology, simulation of a desired model is becoming widely used in this kind of sustainable energy applications, In this paper, the problem remained in transport phenomena on investigating, and simulation of PEM fuel cells will be discussed. Additionally, the modeling by OpenFOAM used in PEM fuel cell will be briefly reviewed. Finally, the promising tool of OpenFOAM can be used in the multi-scale modeling on PEM fuel cells will also be evaluated. PROBLEM STATEMENT The operation principles of PEM fuel cell and the technology limitation currently cannot meet the requirement of commercialization. Large amount of experiment data have been established revealing the relationships between the different operation parameters, which are the initial and basic knowledge for understanding the mechanism of PEM fuel cells. However, the interplayed transport process inside the porous media is not

2 deeply understood. Additionally, the limitation of experiment of measurement on interplayed transport processes is apparent. Working of PEM Fuel Cell Typically PEM fuel cell operates below 100 ºC (usually at 80 ºC). Hydrogen gas mixed with water is supplied to the anode and air/water stream cathode. Anodic hydrogen is oxidized liberating electrons and producing protons. Then the electrons flow through anode gas diffusion layer (GDL), bipolar, external circuit and further to cathode catalyst layer (CL). Meanwhile, produced protons transport through solid electrolyte membrane to cathodic electrode. Oxygen is reduced within cathodic CLs, and then combines with the electrons and protons to produce water and heat. The two-dimensional (2-D) of the reactions and transport processes are shown in Fig. 1. Fig. 1. A 2-D schematic of PEM fuel cell. As regarding to materials used in PEM fuel cells, the carbon supported Pt and carbon fiber or woven cloths comprise the porous CLs and GDLs, respectively. The bipolar plate is constructed form graphite or metal, and polymer electrolyte membrane form perfluorocarbon-sulfonic acid ionomer (also called commercial Nafion). The basic parameters of these terms are listed in Table 1. It can be seen that the parameters of different terms in PEM fuel cells range from nanometers centimeters, typically multi-scale scope. Table 1 Parameters of terms in PEM fuel cells [3, 8, and 9]. Terms Thickness Particle size Pore size CLs µm Carbon particle 5-10 nm 0-10 nm Agglomerates nm nm GDLs µm 5-10 µm 10 nm-10 µm Membrane 100 µm Channel 10 cm 1 mm Transport Phenomena of PEM Fuel Cell The understanding of the transport process of different phenomena is crucial important on optimizing the performance of PEM fuel cells. Typically the transport process of water, heat, gases, charges (protons and electrons) happened simultaneously inside the porous media of fuel cells. These are (a) transport of the reactants to the reaction sites; (b) transport of protons between the reaction sites and membrane; (c) conduction of electrons between the current collectors and the reaction sites; (d) transport, primarily through the solid matrix, of heat produced by the exothermic reaction; and (e) transport of water vapor and liquid water with condensation/evaporation [4]. Transport Phenomena. The water management and heat transport are extremely important for the operation of PEM fuel cells. Water content inside the fuel cell can be determined by the balance between water production and three water transport process, which are electro-osmotic drag, back diffusion and diffusion of water from/to the fuel/oxidant gas streams. The performance of membrane, in terms of proton conductivity, strongly depends on the water content and temperature [1]. On the other hand, the membrane, with good proton conductivities in fully hydrated, highly depends on the water content and temperature. On the contrary, severe dehydration will occur when water completely evaporate from the electrolyte membrane [10]. In order to insure the membrane hydrogen, as shown in Figure 1, the water is inputted into the fuel cell with the gases. The temperature of humidifying water also affects the performance of membrane. In fact, membrane dehydration occurs if that temperature is too low, but flooding will take place on the contrary temperature [10]. It is believed that the catalyst layers, complex coupled processes unfold in, which is the toughest unknown component result from all species and all processes that occur in it: electrochemical reaction, diffusion of hydrogen (anode) and oxygen (cathode), migration and diffusion of protons, migration of electrons, water transport by diffusion, permeation, electroosmotic drag, as well as vaporization/condensation of water [3]. The species transport, e.g. gases, protons and electrons, mostly happened simultaneously with water movement. It means that the water balance inside fuel cell plays an more important in optimizing performance. Material sciences applications. The transport balance in GDLs is also important. In order to avoid flooding at high current densities, GDLs are treated with PTFE to impart hydrophobicity, which can force water droplets to agglomerate at the surface of the GDL [4]. The situation is totally different in CL, which hydrophilic Nafion in it will absorb the water in the CLs, which cannot be avoid but only can decrease the content of Nafion in prepared MEA. The effects on catalyst synthesizing is also appreciated, the porous carbon support with excellent conductivities are commonly selected. Apparently, the porous carbon has the better ability for transport in fuel cells, and the material with good conductivity is for the conduction of electrons, which can decrease the concentration polarization and ohmic polarization. Currently researchers are also interested in synthesizing ordered

3 mesoporous, macroporous carbon used as the catalyst support [11]. It is believed that the gas, species and water transport through this kind of ordered 3D porous media is more convenient. On the other hand, the Pt particles can be doped on the porous wall with more homogeneous, which is another advantage for the reaction species transport to active sites. Kim [12] hold the point that large porous with significant volume fraction plays a dominated role for the transport of gas phases in the CLs, particularly the pores between 20 and 200 nm. Moreover, a carbon layer with microspores is sandwiched into GDL and CL in recently years [13 ]. The innovation method can avoid flooding at high current densities. The inserted layer can be regarded as the bridge connecting between the CL and GDL. Pasaogullari [14] and co-wokers confirmed the innovation by simulation, they thought that placing a microporous layer between GDL and PEM enhances liquid-water removal and reduces the liquid saturation in the catalyst layer. Simulation applications. In operational PEM fuel cells all components have to cooperate well to optimize the complex interplayed transport and reactions. Due to the nano/micro scales and complex porous structure materials used in membrane electrode assembly (MEA), the interactions between fluid and solid particles within the multi-scale and porous structures become strong and should be taken into account. On the other hand, the oxygen reduction reaction (ORR) in the cathode of PEM fuel cell is especially regarded as the dominant limiting factor because of its lower reaction kinetics on electrochemical reaction compared to the one of hydrogen oxidized reaction (HOR) [15]. Therefore, reduction of polarization and keeping balance on transport species inside of fuel cell are important in the operation of PEM fuel cells. Nowadays, it is widely recognized that modeling becomes an effective way on deeply understanding the transfer phenomena and principles of PEM fuel cell, which can give us directions on fuel cell manufacturing and operation. Various continuum approaches, such as computational fluid dynamics (CFD) have been developed to evaluate the water transport processes and the effects on the reactions and other transport phenomena [16]. The Knudsen number, K n used in CFD as shown in Equation (1), K n (1) L where λ is mean free path, L representative physical length scale. The pore size inside the porous materials is widely assumed as the physical length scale in PEM fuel cells [9]. The Knudsen number, K n, based on the ratio of average distance between gas collisions or liquid water lattices and the pore size (~10nm), is about 8.2 and 0.03 for the gas mixture and liquidwater, respectively [17], for the CLs, while for the GDLs, the Knudsen number becomes in the range of 0.01 to 10. The parameters of the pore size and mean free path of gas molecule and Knudsen number on CLs, GDLs, membrane and channel are shown in Table 2. It means that the continuum approaches is not able to directly resolve the influence of the structural morphology on the water transport/state-change dynamics, particularly in CLs. Table 2 Parameters of CLs and GDLs [3, 8, and 9]. Terms Pore size K n Mean free path CLs 100 nm Carbon particle 5-10 nm 10 Agglomerates nm 1 GDLs 100 nm 10 nm-10 µm Membrane Channel 0.5 mm 1 mm 0.5 Fig. 2. SEM image of carbon paper [18]. Fig. 3. Transmission electron photograph of a membrane-catalyst layer assembly [7]. Additionally the material properties used in PEM fuel cell, such as specific heat, conductivities, diffusion coefficients, and physical dimensions, play an important role on the performance of fuel cell. Many mathematical equations applied in CFD model can solve the basic material properties mentioned above. The commonly assumptions were used is homogeneous transport process in simulation of CFD. Actually the partials distribution in functional materials of GDLs and CLs are quite non-uniform, e.g. shown by Fig. 2 and Fig. 3, the distributions of platinum CL and carbon particle carriers for fuel cell [7]. The authors also hold the point that the distribution is also not constant in time, as catalyst particles usually migrate with time due to their unintended involvement in transport processes during fuel cell operation. With microscopic structure of CLs, carbon (or other) particles with surface deposits of platinum (a preferred catalyst) may be bonded to the membrane with an ion-conducting polymer, as shown in Fig. 4. It is believed that

4 the triple-phase boundary (TPB) between catalyst/ionomer/gas pores is the place where the electrochemical reactions occur [19]. Hence the geometry, triple point locus, is not simply represented in two-dimensional or homogeneous idealization. It means that the electronic conduction and gas transport around the active sites must coincide to enable the electrochemistry of the problem. Fig. 4. A schematic drawing of the TPB [19]. Pharoah [20] and co-workers reviewed the current densities and effective transport coefficients of porous PEMFC electrodes. It concluded that a majority of the existing PEMFC models volume-average the effect of porous media using a function that assumes the electrode to be isotropic. But this ignores the factor of inherently anisotropic microstructure of the porous carbon-fiber electrode results in a distinctly different effective mass diffusivity, electronic conductivity, thermal conductivity, and hydraulic permeability in the through-plane and the in-plane directions. Above all, the homogeneous assumption sometimes cannot meet the investigating of the transport phenomena inside the porous particle and cluster, e.g. carbon cluster. The simulation on interplayed transport process concerning the multi-scale, porous functional and anisotropic structural materials used in the PEM fuel cells is required. LITERATURE SURVEY During the past decades, tremendous research has been carried out on investigating the manufacturing, design and simulation. It is widely realized that modeling and simulation are crucial for the cell design and optimization, also can decrease the research cost and time. Moreover, simulation can solve the problem cannot be solved in the experiment, e.g. the water, heat, charge, and gases inside the fuel cell related to mass and momentum transport process. Simulation Methods Modeling and simulation are crucial for the cell design and optimization and various add-on fuel cell modules are available in commonly-used commercial CFD codes: FLUENT, STAR-CD, and COMSOL Multiphysics. With the help of effective and commercial simulation method, a large numbers of literature have been published. The understanding of PEM fuel cell is more comprehensive and can give some directions to the operation of fuel cell applications. A general classification of those modeling efforts can be identified as: from one-dimensional to multi-dimensional, from isothermal to non-isothermal, from single-phase to multiphases, from single components to the cell units, and from steady state to transient [21]. The dimensions of these modeling can be divided into in several scales, i.e. large-scale [22], macro-scale [23], mecro-scales and nano-scale [24, 25] based on the length scale of computational domain. Effects in largescale and macro-scales are commonly simulated by computational fluid dynamics (CFD) method, but nano-scale by Molecular Dynamics (MD) method [24] and Lattice Boltzmann Method (LBM) [25]. Every method has its own advantages and disadvantages in investigating the transport processes of PEM fuel cell. The recently years developed simulation method is by OpenFOAM, another free tool in simulation of CFD. In the next part, we will give a short summary of OpenFOAM (Open Field Operation and Manipulation) used in PEM fuel cells. Simulation of PEM Fuel Cells by OpenFOAM OpenFOAM is a widespread open source for fluid and continuum mechanics. It has a large user related to many areas of engineering and science, in both commercial and academic organizations. The simulation of PEM fuel cell on OpenFOAM attracts some of the researchers in recent years after the modeling of solid oxide fuel cells (SOFC) by same software. The development of SOFC model on OpenFOAM was first developed in the literature by Novaresio [26] and co-workers in But the code library can only be used in the version of OpenFOAM-1.6-ext, which is initial version of OpenFOAM. Several multi-component mass-transport models were implemented in this code, e.g. Fickian, Stefan-Maxwell and Dusty Gas Model, which can be simulate both within porous media and in porosity-free domains. Almost in same year a special code for SOFC simulation called openfuelcell was developed and financial supported by Multi-Scale Integrated Fuel Cell (MuSIC) program, which can be used in the newly version of OpenFOAM The openfuelcell model was designed for the comprehensive simulation of single channel or several parallel channels regarding to five region of SOFC. Fortunately, both of the code developed for SOFC can be free downloaded from website. Moreover, several PEM fuel cell simulation papers using OpenFOAM can also be found meanwhile. But the PEM fuel cell model were developed by individuals, it cannot be competitive with the solver used in SOFC. Therefore the number of published papers related to SOFC simulated by OpenFOAM is serious more than the literatures related to PEM fuel cells. The results of simulation by OpenFOAM in the application of SOFC will not be discussed here. As regarding to the usage of OpenFOAM on PEM fuel cells. No more than ten literatures were published in journals or conference. The simulation of PEM fuel cell using OpenFOAM can be seen in literature [6, 16, 27-30]. In the work of Mustata et al. [16], two

5 kinds of gas income, e.g. U shape and Z shape are solved and focusing on distribution of air flow in a proton exchange membrane fuel cell stack. In all cases the flow pattern is adequate, although the Z configuration is preferable. But regarding to the work studied by Lozano et al. [28], an experimental and numerical study has been presented to analyze the gas flow across a GDL from different bipolar plates. The experimental results were for comparing with the predictions obtained from a computational numerical simulation. Additionally, a review made by Siegel [29], hold the point that Large-scale problems can be solved using parallel computing and MPI-protocol. OpenFOAM is well suited for fuel cell modeling beside commercial CFD software. But coding knowledge is required. It means that OpenFOAM has the advantages for simulation in large scale. Imbrioscia [6] and partner used OpenFOAM for optimizing bipolar plate geometry, such as width, depth and shape of the distributing channels (collectors) as over the main channels, in order to get homogenize the flow distribution. The numerical simulation results of optimized sample were shown in Fig. 5. (a) cross area velocity field, (b) pressure field. Valino [27] and co-wokers studied the numerically and nonhomogeneous distribution of the reacting flows at the catalyst layers of a working bipolar plate of a PEMFC in the case of pure hydrogen and oxygen. Moreover, they used OpenFOAM modeled a real single cell geometry, using a 3D finite volume discretization. Results are analyzed and validated against experimental data. The model was implemented in a developed module attached to OpenFOAM general package. Results are physically meaningful and comparisons also show a very good quantitative agreement with available in-house experimental data [30]. It can be seen that most of the work about simulation of PEM fuel cell using OpenFOAM was focused on the fluid field channel modeling. The properties of functional materials vary in special uniformity and show high of anisotropy in GDLs and CLs of PEM fuel cells. In the CLs of a PEM fuel cell, the void space is filled with the reactant gaseous species, hence, the protonic and electronic conductivity of the void space is zero. Especially the ionomer phases also act as electrical insulators [31]. Moreover, the basic fibers which consist of GDLs cause anisotropic permeability and thermal conductivity. The interplayed transport phenomena inside the fuel cell serious need to be concerned on the functional microstructure of complex fluid. Operating at low temperature, which is the advantage of PEM fuel cell turns into a complex problem compared with other kind of fuel cells. The TPBs electrochemical reaction and multiphase flow (water vapors into liquid phase) occur inside the porous media and on the surface of media. The CFD method used in PEM fuel cell is regarded as an effective way that can simulate the fluid and species transport distributions, velocity and pressure by mathematical simulation method with proper boundary parameters. However, recent analytical techniques used to capture the liquid water distribution inside the diffusion media have revealed that many of the theories and assumptions used for liquid water transport mechanism are inappropriate under normal operation conditions [32, 33]. On the other hand, the fluids have a homogenous appearance on macroscopic scales, but very disordered and heterogeneous over smaller scales due to thermal fluctuations and complex interactions inside the PEM fuel cell. Fig. 6. The reconstruction of CLs by CG method [39]. Fig. 5. Numerical simulation results of optimized sample (a) cross area velocity field, (b) pressure field [6]. PROJECT DESCRIPTION The development of numerical modeling owns to the improvement in computer technology. Form the microscopic point of view, molecular dynamics (MD) and discrete particle method, e.g. dissipative particle dynamics (DPD) and smoothed particle hydrodynamics (SPH) are realistic approach that suffers from the impossibly of exploring space and time scale that are outside the microscopic realm [34, 35]. But MD is too expensive in large numbers of molecular. The so-called pseudo particle method, such as coarse-grained (CG) MD and DPD, in these models, a particle represents a small cluster of atoms and molecules, so it can be used in larger particle size [36]. The DPD is acceptable with length and time scales ranging

6 nm and 1 ns-10 ms, respectively [37, 38]. As for SPH, is a fully Lagragian scheme where a set of macroscopic governing equations are discretized by interpolating the flow properties over a discrete set of points, i.e., pseudo particles, distributed randomly within the domain of solution, without the need to define a spatial mesh [35]. A specific example using CG-MD method is presented by Wescott and coworkers [40], in which they divided the membrane system into three components: backbone, side chain and water to analysis the water content. But interaction parameters for this model were generated using classical molecular dynamics. CG-MD method was also applied in the study of CLs in PEM fuel cell. A remarkable construction was built by Xiao et al. [2] as shown in Fig. 6. The simulation results were also compared with labor testing data. In principle, the process based CG method can provide various microscopic parameters from the reconstructed microstructure, which are relating to the importance properties and effective parameters used in the upper scale models, for comprehensive understanding of the multifunctional porous materials and the various transport processes and the electrochemical reactions in the PEM fuel cell [39]. Additionally, the LBM has become an effective tool to simulate the water transport in GDL. LBM can give more realistic pore-scale dynamic simulations of water transport in GDL with accurate pore geometry taken into consideration. Niu et al. [25] presented a LBM model in which a water-gas flow in the GDL is simulated and the saturation-dependent transport properties under different conditions are investigated. Compared to the MD simulation, the LBM approach assumes that the fluid is made of large amounts of fluid particles instead of individual molecules. A fluid particle is a large group of molecules which, although much larger than a molecule, is still considerably smaller than the smallest length scale of the simulation [21]. Thus, a desired model would have a multiscale character to couple both microscopic and macroscopic flow descriptions [35]. Despite recent progress in developing modeling multiscale [41], enormous challenges remain in bridging between atomistic simulation of realistic structures and continuum models that describe the operation of functional materials for PEM fuel cell applications. Prospective Multiscale Modeling by OpenFOAM OpenFOAM is a general purpose open-source CFD C++ code. OpenFOAM is written in and uses an object oriented approach which makes it easy to extend. The package includes modules for a wide range of applications. The numeric implemented in OpenFOAM uses the finite volume (FV) method on unstructured meshes, and provides many capabilities, including free-surface and multi-phase flow modeling, lagrangian spray model and automatic mesh motion. OpenFOAM allows implementing own models and numerical algorithms. Furthermore, it is possible to insert other models, e.g., particle-based models, to the OpenFOAM CFD Toolbox. Thus OpenFOAM has the potential to meet the requirements faced in PEM fuel cell simulations as mentioned above. There is almost no limitation according to the feature of OpenFOAM so far for the simulation of CFD. The above-mentioned microscopic approaches are limited, in terms of investigated phenomena and the number of particles even simulated by the largest supercomputers, and considered too difficult to resolve the micro- to macroscopic interactions that must captured in the analysis of any real device. The recognition of these issues has led to the development of several methods for the coupling of micro- and macroscopic descriptions in a single modeling and simulation framework. The major goal of these methods is to allow the use of a continuum-based technique, such as CFD, in the parts of the domain where such a description is valid, while using the microscopic method for the subdomain where the continuum description breaks down, such as in the porous CLs and GDLs in PEMFCs. Some attempts have been made previously to couple and bridge the approaches, but no technique exists to couple the transport dynamics in the microscopic regions with those in the continuum parts for true two-way properties coupling [45]. In 1997, Groot [43] and partner reviewed DPD as a mesoscopic simulation method and first made the link between parameters and parameters in Flory-Huggins-type models. This is possible because the equation of state of the DPD fluid is essentially quadratic in density. This link opens the way to do large scale simulations, effectively describing millions of atoms. Kojic [44] and co-workers reviewed the possibility, equations of a multiscale procedure to couple a mesoscale discrete particle model and a macroscale continuum model on incompressible fluid flow. Also they supplied a simple example to confirm possibility, which the finite element (FE) equations rely on the Navier-Stokes and continuity equations. The DPD method for the discrete particle mesoscale model was employed. As shown in Fig. 7, the entire fluid domain is divided into a local domain and a global domain. Fluid flow in the local domain is modeled with both DPD and FE method, while fluid flow in the global domain is modeled by the FE method only. Fig. 7. Two domains within a flow field and boundary conditions at the common boundary between fine (meso) and coarse (FE) models [44]. Although OpenFOAM uses the FV method on unstructured meshes, the progress made by Groot and Kojic

7 and their partners mentioned above supplied a possible direction for coupling the pseudo particle to the CFD in OpenFOAM, which is possible for inserting particle-based model to open source toolbox. On the other hand, modeling the PEM fuel cell with modified and optimized mathematic function will come true. Most importantly, the different components inside the PEM fuel cell with vary Knudsen number can be simulated simultaneously on one model. CONCLUSIONS Functional materials used in MEA of PEM fuel cell play an important role on the performance of fuel cell output. Modeling in multiscale is an effective way in deeply understanding of the interplayed transport phenomena of PEM fuel cells. OpenFOAM as an open source C++ code has the potential to insert other models, e.g., particle-based models it. Thus OpenFOAM has the potential to meet the requirements faced in PEM fuel cell simulations as mentioned above. Especially couple the dissipative particle method and continuum method. REFERENCES [1] Bhatt, S., Gupta, B., Sethi, V. K., & Pandey, M. (2012). Polymer Exchange Membrane (PEM) Fuel Cell: A Review. International Journal of Current Engineering and Technology, 2(1), [2] Xiao, Y., Dou, M., Yuan, J., Hou, M., Song, W., & Sundén, B. (2012). Fabrication Process Simulation of a PEM Fuel Cell Catalyst Layer and Its Microscopic Structure Characteristics. Journal of The Electrochemical Society, 159(3), B308-B314. [3] Zhang, J. (Ed.). (2008). PEM Fuel Cell Electrocatalysts and Catalyst Layers: Fundamentals and Applications. Springer. [4] Djilali, N. (2007). Computational Modelling of Polymer Electrolyte Membrane (PEM) Fuel Cells: Challenges and Opportunities. Energy, 32(4), [5] Fang, B., Kim, J. H., Kim, M., & Yu, J. S. (2009). Ordered Hierarchical Nanostructured Carbon as a Highly Efficient Cathode Catalyst Support in Proton Exchange Membrane Fuel Cell. Chemistry of Materials, 21(5), [6] Imbrioscia, G. M., & Fasoli, H. J. (2014). Simulation and Study of Proposed Modifications over Straightparallel Flow Field Design. International Journal of Hydrogen Energy, 39(16), [7] Reifsnider, K., Huang, X., Ju, G., & Solasi, R. (2006). Multi-scale Modeling Approaches for Functional Nano-composite Materials. Journal of materials science, 41(20), [8] Litster, S., & McLean, G. (2004). PEM Fuel Cell Electrodes. Journal of Power Sources, 130(1), [9] Becker, J., Wieser, C., Fell, S., & Steiner, K. (2011). A Multi-scale Approach to Material Modeling of Fuel Cell Diffusion Media. International Journal of Heat and Mass Transfer, 54(7), [10] Costamagna, P. (2001). Transport Phenomena in Polymeric Membrane Fuel Cells. Chemical Engineering Science, 56(2), [11] Fang, B., Kim, J. H., Kim, M., & Yu, J. S. (2009). Ordered Hierarchical Nanostructured Carbon as a Highly Efficient Cathode Catalyst Support in Proton Exchange Membrane Fuel Cell. Chemistry of Materials, 21(5), [12] Kim, S. H., & Pitsch, H. (2009). Reconstruction and Effective Transport Properties of the Catalyst Layer in PEM Fuel Cells. Journal of the Electrochemical Society, 156(6), B673-B681. [13] Wang, X. L., Zhang, H. M., Zhang, J. L., Xu, H. F., Tian, Z. Q., Chen, J.,... & Yi, B. L. (2006). Micro- Porous Layer with Composite Carbon Black for PEM Fuel Cells. Electrochimica Acta, 51(23), [14] Pasaogullari, U., & Wang, C. Y. (2004). Two-phase Transport and the Tole of Micro-porous Layer in Polymer Electrolyte Fuel Cells. Electrochimica Acta, 49(25), [15] Zhang, L., Zhang, J., Wilkinson, D. P., & Wang, H. (2006). Progress in Preparation of Non-noble Electrocatalysts for PEM Fuel Cell Reactions. Journal of Power Sources, 156(2), [16] Mustata, R., Valiño, L., Barreras, F., Gil, M. I., & Lozano, A. (2009). Study of the Distribution of Air Flow in a Proton Exchange Membrane Fuel Cell Stack. Journal of Power Sources, 192(1), [17] Djilali, N., & Lu, D. (2002). Influence of Heat Transfer on Gas and Water Transport in Fuel Cells. International Journal of Thermal Sciences, 41(1), [18] Wang, Y., Chen, K. S., Mishler, J., Cho, S. C., & Adroher, X. C. (2011). A Review of Polymer Electrolyte Membrane Fuel Cells: Technology, Applications, and Needs on Fundamental Research. Applied Energy, 88(4), [19] Khan, M., Xiao, Y., Sundén, B., & Yuan, J. Analysis of Multiphase Transport Phenomena in PEMFCS by Incorporating Microscopic Model for Catalyst Layer Structures, Proc ASME Int. Mech. Eng. Cong. and Exp., New York, IMECE , pp

8 [20] Pharoah, J. G., Karan, K., & Sun, W. (2006). On Effective Transport Coefficients in PEM Fuel Cell Electrodes: Anisotropy of the Porous Transport Layers. Journal of Power Sources, 161(1), [21] Wu, H. (2009). Mathematical Modeling of Transient Transport Phenomena in PEM Fuel Cells (Doctoral Dissertation, University of Waterloo). [22] Chang, P. A., St-Pierre, J., Stumper, J., & Wetton, B. (2006). Flow Distribution in Proton Exchange Membrane Fuel Cell Stacks. Journal of Power Sources, 162(1), [23] Jörissen, L. (2011). Hydrogen and Fuel Cells. Fundamentals, Technologies and Applications (Edited by Detlev Stolten), WILEY-VCH Publications, British. [24] Jang, S. S., Molinero, V., Cagin, T., & Goddard, W. A. (2004). Nanophase-segregation and Transport in Nafion 117 from Molecular Dynamics Simulations: Effect of Monomeric Sequence. The Journal of Physical Chemistry B, 108(10), [25] Niu, X. D., Munekata, T., Hyodo, S. A., & Suga, K. (2007). An Investigation of Water-gas Transport Processes in the Gas-diffusion-layer of a PEM Fuel Cell by a Multiphase Multiple-relaxation-time Lattice Boltzmann Model. Journal of Power Sources, 172(2), [26] Novaresio, V., García-Camprubí, M., Izquierdo, S., Asinari, P., & Fueyo, N. (2012). An Open-source Library for the Numerical Modeling of Mass-transfer in Solid Oxide Fuel Cells. Computer Physics Communications, 183(1), [27] Valiño, L., Mustata, R., Gil, M. I., & Martín, J. (2010). Effect of the Relative Position of Oxygen-hydrogen Plate Channels and Inlets on a PEMFC. International Journal of Hydrogen Energy, 35(20), [28] Lozano, A., Valiño, L., Barreras, F., & Mustata, R. (2008). Fluid Dynamics Performance of Different Bipolar Plates: Part II. Flow Through the Diffusion Layer. Journal of Power Sources, 179(2), [29] Siegel, C. (2008). Review of Computational Heat and Mass Transfer Modeling in Polymer-electrolyte- Membrane (PEM) Fuel Cells. Energy, 33(9), [30] Valiño, L., Mustata, R., & Dueñas, L. (2013). Consistent Modeling of a Single PEM Fuel Cell Using Onsager's Principle. International Journal of Hydrogen Energy. [31] Khan, M. (2011). Multiphysics Modelling of PEM Fuel Cells-with Reacting Transport Phenomena at Micro and Macroscales (Doctoral dissertation, Lund University). [32] Hartnig, C., Manke, I., Kuhn, R., Kardjilov, N., Banhart, J., & Lehnert, W. (2008). Cross-sectional Insight in the Water Evolution and Transport in Polymer Electrolyte Fuel Cells. Applied Physics Letters, 92(13), [33] Turhan, A., Kim, S., Hatzell, M., & Mench, M. M. (2010). Impact of Channel Wall Hydrophobicity on Through-plane Water Distribution and Flooding Behavior in a Polymer Electrolyte Fuel Cell. Electrochimica Acta, 55(8), [34] Español, P., Serrano, M., & Zuñiga, I. (1997). Coarse- Graining of a Fluid and Its Relation with Dissipative Particle Dynamics and Smoothed Particle Dynamic. International Journal of Modern Physics C, 8(04), [35] Filipovic, N., Ivanovic, M., & Kojic, M. (2009). A Comparative Numerical Study Between Dissipative Particle Dynamics and Smoothed Particle Hydrodynamics When Applied to Simple Unsteady Flows in Microfluidics. Microfluidics and nanofluidics, 7(2), [36] Kojic, M., Filipovic, N., & Tsuda, A. (2008). A Mesoscopic Bridging Scale Method for Fluids and Coupling Dissipative Particle Dynamics with Continuum Finite Element Method. Computer methods in Applied Mechanics and Engineering, 197(6), [37] Flekkøy, E. G., Wagner, G., & Feder, J. (2000). Hybrid Model for Combined Particle and Continuum Dynamics. EPL (Europhysics Letters), 52(3), 271. [38] Espanol, P., & Warren, P. (1995). Statistical Mechanics of Dissipative Particle Dynamics. EPL (Europhysics Letters), 30(4), 191. [39] Xiao, Y., Yuan, J., & Sundén, B. (2012). Process Based Large Scale Molecular Dynamic Simulation of a Fuel Cell Catalyst Layer. Journal of The Electrochemical Society, 159(3), B251-B258. [40] Wescott, J. T., Qi, Y., Subramanian, L., & Capehart, T. W. (2006). Mesoscale Simulation of Morphology in Hydrated Perfluorosulfonic Acid Membranes. The Journal of chemical physics, 124(13), [41] Morrow, B. H., & Striolo, A. (2007). Morphology and Diffusion Mechanism of Platinum Nanoparticles on Carbon Nanotube Bundles. The Journal of Physical Chemistry C, 111(48), [42] Jorn, R., & Voth, G. A. (2012). Mesoscale Simulation of Proton Transport in Proton Exchange Membranes.

9 The Journal of Physical Chemistry C, 116(19), [43] Groot, R. D., & Warren, P. B. (1997). Dissipative Particle Dynamics: Bridging the Gap between Atomistic and Mesoscopic Simulation. Journal of Chemical Physics, 107(11), [44] Kojic, M., Filipovic, N., & Tsuda, A. (2008). A Mesoscopic Bridging Scale Method for Fluids and Coupling Dissipative Particle Dynamics with Continuum Finite Element Method. Computer methods in applied mechanics and engineering, 197(6),