對質子交換膜燃料電池而言，反應堆中反應氣體的分佈對於質子交換膜燃料電池的效率至關重要，反應堆中反應氣體的不均勻分佈可能會導致不良的電流密度，較低的性能以及材料降解。為了理解和準確預測質子交換膜燃料電池系統中的流場並提升質子交換膜燃料電池壓力均勻度，本研究內容可以分為兩個部分。第一部分為分析燃料電池堆中的壓降；第二部分則是研究不同幾何尺寸的圓管和中間區域通道對電池電堆中的流場變化。
第一部份研究目的為開發一種簡單的關係式來分析燃料電池堆中的壓降，將電堆中每個電池中的流道視為多孔介質，並使用冪律模型來近似多孔介質動量源項。對於電池數較少的電堆，即1、5和10顆電池，根據實驗數據建立冪律中的參數。然後，建立關係式以模擬流量並預測具有更高電池串聯數（即20和40顆電池）的電池堆中的壓降。模擬顯示，電堆中每個電池的壓降幾乎不變，並且平均壓降隨著電池數的增加而減小。使用無因次壓降和壓降比來評估具有不同電池數的電池堆中的流動均勻性。結果顯示電池數越低，壓降越均勻。所開發的模型有助於有效地設計具有大電池數量的燃料電池堆的流道。
第二部分研究主要目的在研究燃料電池底板中不同幾何尺寸的圓管和中間區域通道對電池電堆中的流場變化。調整圓管與中間區域的長度比，中間區域的寬度和圓管的直徑以設計幾何尺寸。為了找到具有最高壓力均勻性的最佳幾何尺寸，總共使用了田口法，神經網絡（NN）和多元適應性回歸線（MARS）這三種方法。結果顯示，根據三種方法，在提高壓力均勻性的三個因子中，管徑影響最大。當圓管與中間區域的長度比為9，中間區域的寬度為14 mm，圓管的直徑為9 mm時，圓管和中間區域的最佳幾何尺寸設計的壓力均勻性有所提升。此外，使用神經網絡和多元適應性回歸線可以預測優化組合，相對誤差分別為1.62％和3.89％。因此，在進行的三種方法中，神經網絡是相對穩定的方法，具有較低的相對誤差來預測壓力均勻性。

For proton exchange membrane fuel cells, the distribution of reactant streams in the reactor is critical to the efficiency. Anon-uniform distribution of reactants within the flow channel may reduce the efficiency of PEMFC significantly, which also may cause non-uniform current density, low performance, and material degradation. The content of this study can be divided into two parts. The first part is to analyze the pressure drop in the fuel cell stack; the second part is to investigate the flow field changes in the fuel cell stack with different geometric dimensions of tube and intermediate zone (IZ).
For first part, it aims to develop a simple correlation to analyze the pressure drop in fuel cell stacks. The flow channel in each cell of a stack is treated as a porous medium, and a power-law model is used to approximate the porous medium momentum source term. For the stacks with fewer cell numbers, namely, 1, 5, and 10 cells, the parameters in the power law are established based on the experimental data. Then, a correlation is developed to simulate the flow and predict the pressure drop in the stack with higher cell numbers (i.e., 20 and 40 cells). The simulations show that the pressure drop in each cell of a stack is almost invariable, and the average pressure drop decreases with increasing the number of cells. The flow uniformity in the stacks with different cell numbers is evaluated using the dimensionless pressure drop and the pressure drop ratios. It suggests that the lower the cell number, the more uniform the pressure drop. The developed model is conducive to efficiently designing the flow channel for a fuel cell stack with large cell numbers.
For second part, this part aims to investigate the changes of flow field in the fuel cell stack with different geometric dimensions of tube and intermediate zone (IZ). The length ratio of tube to IZ, the width of IZ, and the diameter of tube are the three factors considered to adjust and optimize the geometric dimensions with highest pressure uniformity. Three methods are used in the analyses including Taguchi method, neural network (NN), and multiple adaptive regression lines (MARS). The results indicate the tube diameter is the most impactive one among the three factors to improve the pressure uniformity. The pressure uniformity of the optimal geometric design has improved when the length ratio of tube to IZ is 9 with the width of IZ of 14 mm and the diameter of tube of 9 mm. Additionally, the optimized combination is predictable using NN and MARS with relative error 1.62 % and 3.89 %, respectively. Thus, NN is a relatively stable method with lower relative error to predict pressure uniformity among the three methods.

[1] Baykara SZ. Hydrogen: A brief overview on its sources, production and environmental impact. International Journal of Hydrogen Energy. 2018;43:10605-14.
[2] Chen W-H, Chen C-Y. Water gas shift reaction for hydrogen production and carbon dioxide capture: A review. Applied Energy. 2019:114078.
[3] Nikolaidis P, Poullikkas A. A comparative overview of hydrogen production processes. Renewable and Sustainable Energy Reviews. 2017;67:597-611.
[4] Wilberforce T, Khatib FN, Ijaodola OS, Ogungbemi E, El-Hassan Z, Durrant A, et al. Numerical modelling and CFD simulation of a polymer electrolyte membrane (PEM) fuel cell flow channel using an open pore cellular foam material. Sci Total Environ. 2019;678:728-40.
[5] Wensheng He JSY TVN. Two‐phase flow model of the cathode of PEM fuel cells using interdigitated flow fields. AIChE Journal. 2004;46;2053-64.
[6] Joseph S, McClure J, Chianelli R, Pich P, Sebastian P. Conducting polymer-coated stainless steel bipolar plates for proton exchange membrane fuel cells (PEMFC). International Journal of Hydrogen Energy. 2005;30:1339-44.
[7] Saha LK, Tabe Y, Oshima N. Effect of GDL deformation on the pressure drop of polymer electrolyte fuel cell separator channel. Journal of Power Sources. 2012;202:100-7.
[8] Qin Y, Du Q, Fan M, Chang Y, Yin Y. Study on the operating pressure effect on the performance of a proton exchange membrane fuel cell power system. Energy Conversion and Management. 2017;142:357-65.
[9] Fan L, Zhang G, Jiao K. Characteristics of PEMFC operating at high current density with low external humidification. Energy Conversion and Management. 2017;150:763-74.
[10] Shao H, Qiu D, Peng L, Yi P, Lai X. In-situ measurement of temperature and humidity distribution in gas channels for commercial-size proton exchange membrane fuel cells. Journal of Power Sources. 2019;412:717-24.
[11] Park J, Li X. Effect of flow and temperature distribution on the performance of a PEM fuel cell stack. Journal of Power Sources. 2006;162:444-59.
[12] Huang F, Qiu D, Lan S, Yi P, Peng L. Performance evaluation of commercial-size proton exchange membrane fuel cell stacks considering air flow distribution in the manifold. Energy Conversion and Management. 2020;203:112256.
[13] Shah AA, Luo KH, Ralph TR, Walsh FC. Recent trends and developments in polymer electrolyte membrane fuel cell modelling. Electrochimica Acta. 2011;56:3731-57.
[14] Dong J, Xu X, Xu B. CFD analysis of a novel modular manifold with multi-stage channels for uniform air distribution in a fuel cell stack. Applied Thermal Engineering. 2017;124:286-93.
[15] Le AD, Zhou B. A numerical investigation on multi-phase transport phenomena in a proton exchange membrane fuel cell stack. Journal of Power Sources. 2010;195:5278-91.
[16] Kodým R, Drakselová M, Pánek P, Němeček M, Šnita D, Bouzek K. Novel approach to mathematical modeling of the complex electrochemical systems with multiple phase interfaces. Electrochimica Acta. 2015;179:538-55.
[17] Kandlikar SG, Lu Z, Domigan WE, White AD, Benedict MW. Measurement of flow maldistribution in parallel channels and its application to ex-situ and in-situ experiments in PEMFC water management studies. International Journal of Heat and Mass Transfer. 2009;52:1741-52.
[18] Manso AP, Marzo FF, Barranco J, Garikano X, Garmendia Mujika M. Influence of geometric parameters of the flow fields on the performance of a PEM fuel cell. A review. International Journal of Hydrogen Energy. 2012;37:15256-87.
[19] Mustata R, Valiño L, Barreras F, Gil MI, Lozano A. Study of the distribution of air flow in a proton exchange membrane fuel cell stack. Journal of Power Sources. 2009;192:185-9.
[20] Drakselová M, Kodým R, Šnita D, Beckmann F, Bouzek K. Three-dimensional macrohomogeneous mathematical model of an industrial-scale high-temperature PEM fuel cell stack. Electrochimica Acta. 2018;273:432-46.
[21] Lim BH, Majlan EH, Daud WRW, Rosli MI, Husaini T. Numerical analysis of modified parallel flow field designs for fuel cells. International Journal of Hydrogen Energy. 2017;42:9210-8.
[22] Chen C-H, Jung S-P, Yen S-C. Flow distribution in the manifold of PEM fuel cell stack. Journal of Power Sources. 2007;173:249-63.
[23] Shimpalee S, Ohashi M, Van Zee JW, Ziegler C, Stoeckmann C, Sadeler C, et al. Experimental and numerical studies of portable PEMFC stack. Electrochimica Acta. 2009;54:2899-911.
[24] Lim BH, Majlan EH, Daud WRW, Rosli MI, Husaini T. Numerical analysis of flow distribution behavior in a proton exchange membrane fuel cell. Heliyon. 2018;4:e00845.
[25] Lim BH, Majlan EH, Daud WRW, Rosli MI, Husaini T. Three-dimensional study of stack on the performance of the proton exchange membrane fuel cell. Energy. 2019;169:338-43.
[26] Shen J, Tu Z, Chan SH. Enhancement of mass transfer in a proton exchange membrane fuel cell with blockage in the flow channel. Applied Thermal Engineering. 2019;149:1408-18.
[27] Li M, Duan K, Djilali N, Sui P-C. Flow sharing and turbulence phenomena in proton exchange membrane fuel cell stack headers. International Journal of Hydrogen Energy. 2019;44:30306-18.
[28] Maharudrayya S, Jayanti S, Deshpande AP. Pressure losses in laminar flow through serpentine channels in fuel cell stacks. Journal of Power Sources. 2004;138:1-13.
[29] Malhotra S, Ghosh S. Effects of channel diameter on flow pattern and pressure drop for air–water flow in serpentine gas channels of PEM fuel cell-An Ex situ experiment. Experimental Thermal and Fluid Science. 2019;100:233-50.
[30] Wang J. Pressure drop and flow distribution in parallel-channel configurations of fuel cells: U-type arrangement. International Journal of Hydrogen Energy. 2008;33:6339-50.
[31] Jackson JM, Hupert ML, Soper SA. Discrete geometry optimization for reducing flow non-uniformity, asymmetry, and parasitic minor loss pressure drops in Z-type configurations of fuel cells. Journal of Power Sources. 2014;269:274-83.
[32] Chen W-H, Lin Y-Y, Liu H-C, Baroutian S. Optimization of food waste hydrothermal liquefaction by a two-step process in association with a double analysis. Energy. 2020;199:117438.
[33] Chen WH, Chen CJ, Hung CI. Taguchi approach for co-gasification optimization of torrefied biomass and coal. Bioresour Technol. 2013;144:615-22.
[34] Chen W-H, Chen C-Y, Huang C-Y, Hwang C-J. Power output analysis and optimization of two straight-bladed vertical-axis wind turbines. Applied Energy. 2017;185:223-32.
[35] Chen W-H, Huang S-R, Lin Y-L. Performance analysis and optimum operation of a thermoelectric generator by Taguchi method. Applied Energy. 2015;158:44-54.
[36] Wu H-W, Ku H-W. The optimal parameters estimation for rectangular cylinders installed transversely in the flow channel of PEMFC from a three-dimensional PEMFC model and the Taguchi method. Applied Energy. 2011;88:4879-90.
[37] Karthikeyan P. Parametric analysis of proton exchange membrane fuel cell (PEMFC) performed by the taguchi method2016.
[38] A.U. Levin KSN. Control of nonlinear dynamical systems using neural networks – part II: observability, identification, and control. IEEE Transactions on Neural Networks. 1996:30-42.
[39] Wang G-J, Tsai J-C, Tseng P-C, Chen T-C. Neural-Taguchi method for robust design analysis. 1998;19:223-30.
[40] Lin H-L, Chou C. Optimisation of the GTA welding process using the Taguchi method and a neural network. Science and Technology of Welding & Joining. 2006;11:120-6.
[41] Wu S-J, Shiah S-W, Yu W-L. Parametric analysis of proton exchange membrane fuel cell performance by using the Taguchi method and a neural network. Renewable Energy. 2009;34:135-44.
[42] Chang K-Y, Lin H-J, Chen P-C. The optimal performance estimation for an unknown PEMFC based on the Taguchi method and a generic numerical PEMFC model. International Journal of Hydrogen Energy. 2009;34:1990-8.
[43] Chang K-Y. The optimal design for PEMFC modeling based on Taguchi method and genetic algorithm neural networks. International Journal of Hydrogen Energy. 2011;36:13683-94.
[44] Yu W-L, Wu S-J, Shiah S-W. Experimental analysis of dynamic characteristics on the PEM fuel cell stack by using Taguchi approach with neural networks. International Journal of Hydrogen Energy. 2010;35:11138-47.
[45] Zheng G, Yang P, Zhou H, Zeng C, Yang X, He X, et al. Evaluation of the earthquake induced uplift displacement of tunnels using multivariate adaptive regression splines. Computers and Geotechnics. 2019;113:103099.
[46] Zheng G, Zhang W, Zhou H, Yang P. Multivariate adaptive regression splines model for prediction of the liquefaction-induced settlement of shallow foundations. Soil Dynamics and Earthquake Engineering. 2020;132:106097.
[47] Lokuge W, Wilson A, Gunasekara C, Law DW, Setunge S. Design of fly ash geopolymer concrete mix proportions using Multivariate Adaptive Regression Spline model. Construction and Building Materials. 2018;166:472-81.
[48] Chou S-M, Lee T-S, Shao YE, Chen IF. Mining the breast cancer pattern using artificial neural networks and multivariate adaptive regression splines. Expert Systems with Applications. 2004;27:133-42.
[49] Lee T, Chen I. A two-stage hybrid credit scoring model using artificial neural networks and multivariate adaptive regression splines. Expert Systems with Applications. 2005;28:743-52.
[50] Lu C-J, Lee T-S, Lian C-M. Sales forecasting for computer wholesalers: A comparison of multivariate adaptive regression splines and artificial neural networks. Decision Support Systems. 2012;54:584-96.
[51] Adoko A-C, Jiao Y-Y, Wu L, Wang H, Wang Z-H. Predicting tunnel convergence using Multivariate Adaptive Regression Spline and Artificial Neural Network. Tunnelling and Underground Space Technology. 2013;38:368-76.
[52] Cheng M-Y, Cao M-T. Evolutionary multivariate adaptive regression splines for estimating shear strength in reinforced-concrete deep beams. Engineering Applications of Artificial Intelligence. 2014;28:86-96.
[53] Zhang W, Goh ATC. Multivariate adaptive regression splines and neural network models for prediction of pile drivability. Geoscience Frontiers. 2016;7:45-52.
[54] Wu H-W, Gu H-W. Analysis of operating parameters considering flow orientation for the performance of a proton exchange membrane fuel cell using the Taguchi method. Journal of Power Sources. 2010;195:3621-30.
[55] Chen W-H, Lin Y-X, Chiou Y-B, Lin Y-L, Wang X-D. A computational fluid dynamics (CFD) approach of thermoelectric generator (TEG) for power generation. Applied Thermal Engineering. 2020;173:115203.
[56] Zhong W, Xu K, Li X, Liao Y, Tao G, Kagawa T. Determination of pressure drop for air flow through sintered metal porous media using a modified Ergun equation. Advanced Powder Technology. 2016;27:1134-40.
[57] Chen W-H, Tsai C-W, Lin Y-L, Chein R-Y, Yu C-T. Reaction phenomena of high-temperature water gas shift reaction in a membrane reactor. Fuel. 2017;199:358-71.
[58] Özçelep Y, Sevgen S, Samli R. A study on the hydrogen consumption calculation of proton exchange membrane fuel cells for linearly increasing loads: Artificial Neural Networks vs Multiple Linear Regression. Renewable Energy. 2020;156:570-8.
[59] Li Y, Li J, Huang J, Zhou H. Fitting analysis and research of measured data of SAW micro-pressure sensor based on BP neural network. Measurement. 2020;155:107533.
[60] Guo Y. Credit Risk Assessment of P2P Lending Platform towards Big Data based on BP Neural Network. Journal of Visual Communication and Image Representation. 2019:102730.
[61] Wang S, Wu TH, Shao T, Peng ZX. Integrated model of BP neural network and CNN algorithm for automatic wear debris classification. Wear. 2019;426-427:1761-70.
[62] Friedman JH. Multivariate Adaptive Regression Splines. Annals of Statistics. 1991;19(1):1-67.
[63] Qi X, Wang H, Pan X, Chu J, Chiam K. Prediction of interfaces of geological formations using the multivariate adaptive regression spline method. Underground Space. 2020.
[64] Chen W-H, Cheng C-L, Show P-L, Ong HC. Torrefaction performance prediction approached by torrefaction severity factor. Fuel. 2019;251:126-35.
[65] Tonomura O, Tanaka S, Noda M, Kano M, Hasebe S, Hashimoto I. CFD-based optimal design of manifold in plate-fin microdevices. Chemical Engineering Journal. 2004;101:397-402.
[66] Koh J-H, Seo H-K, Lee CG, Yoo Y-S, Lim HC. Pressure and flow distribution in internal gas manifolds of a fuel-cell stack. Journal of Power Sources. 2003;115:54-65.
[67] Bayat J, Nikseresht AH. Thermal performance and pressure drop analysis of nanofluids in turbulent forced convective flows. International Journal of Thermal Sciences. 2012;60:236-43.
[68] Singh A, Prasad BVSSS. Influence of novel equilaterally staggered jet impingement over a concave surface at fixed pumping power. Applied Thermal Engineering. 2019;148:609-19.
[69] Yin L, Li Q, Chen W, Wang T, Liu H. Experimental analysis of optimal performance for a 5 kW PEMFC system. International Journal of Hydrogen Energy. 2019;44:5499-506.
[70] Matulić N, Radica G, Barbir F, Nižetić S. Commercial vehicle auxiliary loads powered by PEM fuel cell. International Journal of Hydrogen Energy. 2019;44:10082-90.
[71] Wang J, Wang H. Discrete approach for flow field designs of parallel channel configurations in fuel cells. International Journal of Hydrogen Energy. 2012;37:10881-97.
[72] Wang K, Wang L, Yan D, Zhang YS, Pu J, Jian L. Investigation on Flow Distribution in an External Manifold SOFC Stack by Computational Fluid Dynamics Technique. Fuel Cells. 2015;15:61-70.
[73] Chu W-x, Bennett K, Cheng J, Chen Y-t, Wang Q-w. Numerical study on a novel hyperbolic inlet header in straight-channel printed circuit heat exchanger. Applied Thermal Engineering. 2019;146:805-14.
[74] Zhao C, Yang J, Zhang T, Yan D, Pu J, Chi B, et al. Numerical modeling of manifold design and flow uniformity analysis of an external manifold solid oxide fuel cell stack. International Journal of Hydrogen Energy. 2020;45:14440-51.
[75] Kumaraguruparan G, Kumaran RM, Sornakumar T, Sundararajan T. A numerical and experimental investigation of flow maldistribution in a micro-channel heat sink. International Communications in Heat and Mass Transfer. 2011;38:1349-53.
[76] Riedmiller M, Braun H. A direct adaptive method for faster backpropagation learning: the RPROP algorithm. IEEE International Conference on Neural Networks1993. p. 586-91 vol.1.
[77] Elliot. DL. A better Activation Function for Artificial Neural Networks. 1993.