An Industrial Application of a Large Language Model to Enhancing Asset Integrity and Process Safety Management

This research incorporates artificial intelligence (AI) into asset integrity and process safety (AIPS) management, aiming to revolutionize conventional methods. It facilitates the automation of risk assessments, enhances predictive analytics, and supports the development of proactive measures to mitigate potential incidents. It explores the application of a generative pre-trained transformer (GPT) based large language models (LLM) to analyse and classify AIPS indicators from vast datasets to generate actionable recommendations to prevent future incidents. A comparative study between two onshore liquefied natural gas (LNG) plants; one utilizing AI-driven AIPS management and the other relying on manual data analysis is presented. The results indicate that AI-driven approaches significantly enhance the accuracy and speed of incident classifications, reducing data processing times. The test model effectively predicts potential future failures by analysing past incident patterns, enabling informed decision-making to prevent and mitigate future failures. The findings highlight the importance of adopting AI-driven AIPS management as a standard practice. It also emphasises the need for stronger collaboration between academia and industry in AI solutions to drive technological advancements for sustainability.