4th Thailand International Chemical Engineering and Chemical Technology Asia Conference (TNChE Asia 2025), Thailand
This paper explores the application of artificial intelligence (AI), natural language processing (NLP) and machine learning (ML) techniques to enhance Hazard and Operability (HAZOP) studies. The methodology employs advanced natural language processing techniques including semantic similarity analysis and generative models to generate potential consequences based on identified causes. Furthermore, it utilizes NLP, feature engineering techniques, and vector embeddings for binary classification to predict consequence categories. Finally, NLP, feature engineering with vector extraction, and multi-label classification are used to predict severity based on consequence and likelihood based on cause. This approach aims to streamline HAZOP processes, speed up HAZOP analysis sessions, and improve risk assessment consistency in process plants with similar construction.