基于正交实验和熵值法耦合人工神经网络的PEMFC多目标优化
Multi-Objective Optimization of PEMFC Based on Orthogonal Experiment, Entropy Method, and Artificial Neural Network
摘要: 质子交换膜燃料电池(PEMFC)性能指标多样，通过多目标优化可以提升PEMFC的综合性能。本文综合运用正交实验(OEM)、方差分析(ANOVA)以及熵值法(EWM)与人工神经网络(ANN)，提出了一种综合评价PEMFC性能的多目标优化方法。首先，选择几个常用的PEMFC性能指标，针对PEMFC的几何参数和操作参数初步选择对PEMFC性能有影响的参数设计正交实验，随后使用COMSOL软件模拟计算获取对应工况下各个性能指标结果。通过方差分析筛选出对目标有显著影响的变量，可以达到减小优化空间、提高优化效率的目的。其次，根据显著变量再次进行正交实验，获取优化空间中具有全局代表性的数据集，随后提出了代表PEMFC综合性能(CP)的表达式，通过熵值法分析了CP中多目标的权重，进一步计算出CP的数学表达式并作为适应度函数。最后将ANN作为计算CP的代理模型，对显著变量进行优化，继而得到一种最优的参数组合使得对应的CP值达到最大，从而最大限度地提升PEMFC的综合性能。利用该方法对梯形挡板流道模型进行了多目标优化分析，在0.4 V电压下，优化后流道模型相比较基础流道模型而言，功率密度和净功率分别提升了53.71%和53.70%，然而阴极压降却仅仅只上升了7.93 Pa，氧气不均匀度降至0.19。上述方法可有效降低优化计算时间，对今后PEMFC的结构设计与操作运行具有一定的指导意义。
Abstract: Proton exchange membrane fuel cell (PEMFC) has various performance indicators, and the comprehensive performance of PEMFC can be improved by multi-objective optimization. In this paper, a multi-objective optimization method for comprehensively evaluating the performance of PEMFC is proposed by comprehensively applying orthogonal experiment (OEM), analysis of variance (ANOVA), and entropy value method (EWM) with artificial neural network (ANN). Firstly, several commonly used PEMFC performance indexes are selected, and orthogonal experiments are designed for the parameters that have an impact on the PEMFC performance by initially selecting the geometrical and operational parameters of the PEMFC, and then simulations are performed using COMSOL software to obtain the results of the performance indexes under the corresponding working conditions. Through the analysis of variance to screen out the variables that have a significant impact on the target, it can achieve the purpose of reducing the optimization space and improving the optimization efficiency. Secondly, orthogonal experiments were conducted again according to the significant variables to obtain a globally representative data set in the optimization space. Subsequently, an expression representing the comprehensive performance (CP) of PEMFC was proposed, and the weights of multiple objectives in the CP were analyzed by entropy method, and the mathematical expression of the CP was further computed and used as the fitness function. Finally, the ANN is used as a proxy model for calculating the CP to optimize the significant variables, and then an optimal combination of parameters is obtained to maximize the corresponding CP value, thus maximizing the comprehensive performance of the PEMFC. A multi-objective optimization analysis of the trapezoidal baffle runner model is carried out using this method, and the power density and net power of the optimized runner model are increased by 53.71% and 53.70%, respectively, compared with the base runner model at 0.4 V, while the cathode pressure drop is increased by only 7.93 Pa, and the oxygen inhomogeneity is reduced to 0.19. The above method can effectively reduce the optimization computation time, which is very useful for the design of future PEMFC structures and the development of the PEMFC.
文章引用：王则善, 徐洪涛, 李建军. 基于正交实验和熵值法耦合人工神经网络的PEMFC多目标优化[J]. 物理化学进展, 2025, 14(2): 370-386. https://doi.org/10.12677/japc.2025.142035

参考文献

[1] Perng, S. and Wu, H. (2011) Non-Isothermal Transport Phenomenon and Cell Performance of a Cathodic PEM Fuel Cell with a Baffle Plate in a Tapered Channel. Applied Energy, 88, 52-67. [Google Scholar] [CrossRef] 
[2] Perng, S. and Wu, H. (2015) A Three-Dimensional Numerical Investigation of Trapezoid Baffles Effect on Non-Isothermal Reactant Transport and Cell Net Power in a PEMFC. Applied Energy, 143, 81-95. [Google Scholar] [CrossRef] 
[3] 孙峰, 苏丹丹, 殷宇捷, 等. PEMFC翅脉流道传质及输出性能分析[J]. 太阳能学报, 2022, 43(6): 414-419.
[4] Cai, G., Liang, Y., Liu, Z. and Liu, W. (2020) Design and Optimization of Bio-Inspired Wave-Like Channel for a PEM Fuel Cell Applying Genetic Algorithm. Energy, 192, Article 116670. [Google Scholar] [CrossRef] 
[5] Yin, Y., Wu, S., Qin, Y., Otoo, O.N. and Zhang, J. (2020) Quantitative Analysis of Trapezoid Baffle Block Sloping Angles on Oxygen Transport and Performance of Proton Exchange Membrane Fuel Cell. Applied Energy, 271, Article 115257. [Google Scholar] [CrossRef] 
[6] Wang, X., Qin, Y., Wu, S., Shangguan, X., Zhang, J. and Yin, Y. (2020) Numerical and Experimental Investigation of Baffle Plate Arrangement on Proton Exchange Membrane Fuel Cell Performance. Journal of Power Sources, 457, Article 228034. [Google Scholar] [CrossRef] 
[7] Zhang, S., Mao, Y., Liu, F., Xu, H., Qu, Z. and Liao, X. (2023) Multi-Objective Optimization and Evaluation of PEMFC Performance Based on Orthogonal Experiment and Entropy Weight Method. Energy Conversion and Management, 291, Article 117310. [Google Scholar] [CrossRef] 
[8] Yan, W., Chen, C., Mei, S., Soong, C. and Chen, F. (2006) Effects of Operating Conditions on Cell Performance of PEM Fuel Cells with Conventional or Interdigitated Flow Field. Journal of Power Sources, 162, 1157-1164. [Google Scholar] [CrossRef] 
[9] Kim, Y.B. (2011) Study on the Effect of Humidity and Stoichiometry on the Water Saturation of PEM Fuel Cells. International Journal of Energy Research, 36, 509-522. [Google Scholar] [CrossRef] 
[10] Liu, Z., Zeng, X., Ge, Y., Shen, J. and Liu, W. (2017) Multi-Objective Optimization of Operating Conditions and Channel Structure for a Proton Exchange Membrane Fuel Cell. International Journal of Heat and Mass Transfer, 111, 289-298. [Google Scholar] [CrossRef] 
[11] Kwan, T.H., Wu, X. and Yao, Q. (2018) Multi-Objective Genetic Optimization of the Thermoelectric System for Thermal Management of Proton Exchange Membrane Fuel Cells. Applied Energy, 217, 314-327. [Google Scholar] [CrossRef] 
[12] Zhang, G., Fan, L., Sun, J. and Jiao, K. (2017) A 3D Model of PEMFC Considering Detailed Multiphase Flow and Anisotropic Transport Properties. International Journal of Heat and Mass Transfer, 115, 714-724. [Google Scholar] [CrossRef] 
[13] Zhou, Y., Chen, B., Chen, W., Deng, Q., Shen, J. and Tu, Z. (2022) A Novel Opposite Sinusoidal Wave Flow Channel for Performance Enhancement of Proton Exchange Membrane Fuel Cell. Energy, 261, Article 125383. [Google Scholar] [CrossRef] 
[14] Sezgin, B., Caglayan, D.G., Devrim, Y., Steenberg, T. and Eroglu, I. (2016) Modeling and Sensitivity Analysis of High Temperature PEM Fuel Cells by Using Comsol Multiphysics. International Journal of Hydrogen Energy, 41, 10001-10009. [Google Scholar] [CrossRef] 
[15] Lin, R., Diao, X., Ma, T., Tang, S., Chen, L. and Liu, D. (2019) Optimized Microporous Layer for Improving Polymer Exchange Membrane Fuel Cell Performance Using Orthogonal Test Design. Applied Energy, 254, Article 113714. [Google Scholar] [CrossRef] 
[16] Öztürk, A., Fıçıcılar, B., Eroğlu, İ. and Bayrakçeken Yurtcan, A. (2017) Facilitation of Water Management in Low Pt Loaded PEM Fuel Cell by Creating Hydrophobic Microporous Layer with PTFE, FEP and PDMS Polymers: Effect of Polymer and Carbon Amounts. International Journal of Hydrogen Energy, 42, 21226-21249. [Google Scholar] [CrossRef] 
[17] Antonacci, P., Chevalier, S., Lee, J., Ge, N., Hinebaugh, J., Yip, R., et al. (2016) Balancing Mass Transport Resistance and Membrane Resistance When Tailoring Microporous Layer Thickness for Polymer Electrolyte Membrane Fuel Cells Operating at High Current Densities. Electrochimica Acta, 188, 888-897. [Google Scholar] [CrossRef] 
[18] Zhou, M., Wei, P. and Deng, L. (2022) Research on the Factorial Effect of Science and Technology Innovation (STI) Policy Mix Using Multifactor Analysis of Variance (ANOVA). Journal of Innovation & Knowledge, 7, Article 100249. [Google Scholar] [CrossRef] 
[19] Li, Q., Hu, H., Ma, L., Wang, Z., Arıcı, M., Li, D., et al. (2022) Evaluation of Energy-Saving Retrofits for Sunspace of Rural Residential Buildings Based on Orthogonal Experiment and Entropy Weight Method. Energy for Sustainable Development, 70, 569-580. [Google Scholar] [CrossRef]