Development of an AI Prediction System for Corrosion Mechanisms of Non-Metallic Materials

Abstract

For safe and stable operation of chemical plants, we studied the implementation of artificial intelligence to determine and evaluate damage mechanisms for damage of non-metallic materials. Existing damage cases of non-metallic materials were collected, data cleansing was performed, and decision tree analysis was conducted. In addition, the presence or absence of over-learning was also examined. As a result, with a certain number of cases, it is expected to be possible to extract conditions for determining the damage mechanism and to predict possible damage for mechanisms.