Investigation of Main Contributors in Solid Oxide Cell (SOC) by Multivariate Regression

Abstract

Durability and degradation-related issues affect the commercialisation of Solid Oxide Cell (SOC) technologies. Over the last decades, SOC technologies have been developed with significant progress in material development, understanding of degradation phenomena and performance-related issues. However, individual operating parameters' influence on the overall SOC degradation is still not fully understood. This thesis aims to investigate the main contributors to SOC degradation using multivariate regression analysis. Different load operations from stack experiments with homogenous properties were collected, and the degradation rate for each load operation with their corresponding operating conditions, such as current density, conversion rate and stack temperature, were determined. After consolidation of the dataset, a multivariate regression analysis was used to examine each contributor's relevance to SOC degradation. To quantify the level of uncertainty, a Bayesian multivariate regression model using PyMC3 was employed. This analysis reveals that operating current density is the main contributor to SOC degradation. The influence of conversion rate, however, cannot be neglected as the conversion rate is the second leading contributing factor to SOC degradation.