Lattice Boltzmann modeling and artificial intelligence

Electrode is the heart of proton-exchange membrane fuel cells, where reactive transport, two-phase flow, and electrochemical processes simultaneously occur during the operation. In order to optimize the porous electrode structure for better performance and lower cost, the lattice Boltzmann method (LBM) is widely used as a mesoscopic approach for a deeper insight into the correlation between transport mechanisms and microstructure. However, the high computational cost for simulating the mass transport mechanism of porous electrodes at mesoscopic scales with LBM has greatly limited the application of LBM for multi-scale studies in porous electrodes. In contrast, artificial intelligence (AI) methods can give results efficiently yet with heavy demand on datasets. The combination of LBM and AI is promising to mitigate the time- and space-scale limitations of LBM in solving real physics problems and to further optimize the electrode design in a fast and accurate manner. This chapter reviews advanced researches where LBM is applied to study the reactive transport and two-phase flow in the porous electrode, as well as the representative application of AI methods in fuel cells, including parameter optimization, model predictive control, prediction and health management, and fault diagnosis. Finally, it is discussed the way and possibility of combining the two methods based on the current research results.