A Digital Twin Approach for Assembly Design of Proton Exchange Membrane Fuel Cell Stacks

The digital twin is a vital technology for the intelligent manufacturing of proton exchange membrane fuel cell (PEMFC) stacks. Aiming to achieve an efficient optimal design for the stack assembly process, a digital twin modeling approach combining finite element analysis (FEA) and artificial neural network (ANN) is proposed in this paper. The establishing method of finite element model (FEM) for the stack under assembly conditions, the dataset construction method for FEM simulation results, and the modeling method of ANN for mechanical behavior prediction are systematically expounded. After that, the digital twin modeling solution is verified by a case study of the contact pressure prediction inside the stack. The results show that the ANN model based on the simulation dataset has a prediction accuracy close to 98% of the FEM, and the computational efficiency is improved more than tens of thousands of times. This model can also quickly obtain the entire stack uniformity at different compression ratios, which is a prerequisite for guiding the optimal design of the stack assembly process in production scenarios.