Digitally-assisted structure design of a large-size proton exchange membrane fuel cell

Abstract

The flow field plays a significant role in the performance of proton exchange membrane (PEM) fuel cells. However, its complex structure leads to unacceptable development costs and time commonly using the trial-and-error method based on many experiments. Herein, we propose a digitally-assisted method to accelerate the development process and reduce costs. Comprehensive experiments and tests are conducted using the commercial-size PEM fuel cell with an active area of 332 cm², including the investigation of polarization curves, five sensitivity parameters under seven different current densities, and spatial distributions. A high-resolution printed circuit board with 408 segments of 0.8 cm² is employed to explore the current density distribution. The commercial-size PEM fuel cell is further digitalized with a self-developed fuel cell numerical model, which is strictly verified in terms of all experimental data. The digital multi-physics information inside PEM fuel cells is obtained and evaluated via this efficient numerical model in order to search for the structure defects quickly and accurately. Afterwards, targeted structure optimization is effectively carried out to achieve a better performance, with the maximum deviation of oxygen concentration in the channels decreasing from 26.33% to 3.78%. This digital method is very valuable for the forward design of flow field structures to considerably reduce the development cost and time.