Machine learning based analysis of metal support co-sintering process for solid oxide fuel cells

Porous metals are promising substrate supports of solid oxide fuel cell for the applications of mobile and high power density. The process conditions for a suitable mixture of pore former and stainless-steel particles, and also for the co-sintering with oxide layers of SOFC are investigated by machine learning (ML) methods. Half-cell metal supported cell (MSC) process with ScSZ electrolyte and NiO-ScSZ anode layer coated on the metal substrate of SUS430 was predicted by two ML methods of random forest (RF) and neural network (NN) were investigated. RF of multi output methods predicted the target variables, such as shrinkage, bend and pore structures with the better accuracy compared to NN, revealing that not only sintering temperature but also pore former and degreasing temperature of metal support influenced the process by feature importance analysis. The power generation performance of MSC tested from 550 degrees C to750 degrees C, and the relation to the microstructure of MSC is discussed.