Three-dimensional multi-physics modelling and structural optimization of SOFC large-scale stack and stack tower

Multi-physics modelling of the Solid Oxide Fuel Cell (SOFC) stack requires significant computational resources. Design optimization of large-scale stacks and stack towers has always been a challenge in recent years. This study establishes a three-dimensional multi-physics model based on a two-step coupling using the BP neural network. The comparison between the novel model and the traditional fully coupled model in both accuracy and computing resource requirements are explored. The novel method has high effectiveness for modelling the large-scale stacks. Based on this, planar SOFC 50-cell stacks and 150-cell stack towers are simulated. The results show that, the flow uniformity of fuel distribution of the stack towers can decrease more than 30% comparing with the 50-cell stack, which leads to significant deterioration of the voltage and temperature distribution. The pa-rameters of manifold and buffer area and channel height of the stack tower is optimized to achieve better uniformity of flow and voltage distribution and lower temperature gradient simultaneously.(c) 2022 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Automated summary

This study establishes a three-dimensional multi-physics model for the Solid Oxide Fuel Cell (SOFC) stack using a two-step coupling method based on the BP neural network. Compared with the traditional fully coupled model, the novel model shows significant advantages in accuracy and computational resource requirements. The simulation of planar SOFC 50-cell stacks and 150-cell stack towers demonstrates that the flow uniformity of the stack towers decreases by more than 30% compared to the 50-cell stack, resulting in significant deterioration of voltage and temperature distribution. The optimization of the manifold and buffer area parameters and the channel height of the stack tower achieves better uniformity of flow and voltage distribution and lower temperature gradient simultaneously.