Performance optimization of anodic porous transport layer in proton exchange membrane electrolyzer using multilayer perceptron model

Abstract

Resulting from the capability of resisting fluctuating energy inputs, proton exchange membrane water electrolysis (PEMWE) technology holds significant potential for green hydrogen production. The performance of PEMWE is influenced by various structural parameters, in which the properties of the porous transport layer (PTL) are particularly critical. Optimizing the structural characteristics of the PTL is important for enhancing the overall performance of PEMWE. In this study, a three-dimensional steady-state PEMWE model is firstly developed. Based on the model, polarization curves of the PEMWE under different PTL parameters are computed, and the impacts of three characteristic parameters, i.e. porosity, thickness, and conductivity, on the PEMWE performance are thoroughly investigated. Then, the corresponding performance optimization strategies are proposed by incorporating a multilayer perceptron (MLP) machine learning model. It shows that porosity plays a dominant role in the PTL performance among the three parameters, followed by thickness, with conductivity having a relatively minor impact. The increasing of porosity and reducing of thickness can effectively enhance the electrolyzer performance. According to the MLP model screening, the optimal PTL structure is determined to be the porosity of 0.52, thickness of 0.2 mm, and conductivity of 4×10⁶ S/m. At 2 A/cm², the operating voltage of the PEMWE is 1.85 V.