|About this Abstract
||MS&T22: Materials Science & Technology
||Energy Materials for Sustainable Development
||Machine Learning Methods for Predicting Microstructural Changes in Solid Oxide Cell Electrodes
||Anna Sciazko, Rena Yamagishi, Yosuke Komatsu, Zhufeng Ouyang, Junya Ohnishi, Katsuhiko Nishimura, Naoki Shikazono
|On-Site Speaker (Planned)
Solid oxide cells (SOCs) are promising electrochemical devices for the future energy system due to their high energy conversion efficiency and the ability to handle a variety of fuels. Materials choice and microstructure of multi-phase porous electrodes determine the SOCs electrochemical performance. The recent research underlines the importance of the 3D structural design of the electrodes and necessity for introducing graded micro- and nanostructures. However, obtaining an extensive experimental data of various designs is challenging, particularly for the long-time stability experiments. In this study, a computational scheme using machine learning methods is proposed to enhance the understanding of the relationship between SOCs’ microstructure and degradation. In particular, the methods for building the artificial models of complex 3D microstructures are proposed based on the generative adversarial networks. Furthermore, the microstructure evolution prediction from the limited training data is attempted by the unsupervised image-to-image translation network with physical constraints.