Smart Integrity: Leveraging Knowledge-Base-Driven AI Chatbots for Safety and Operational Reliability in Petroleum Refining

Abstract

Safety critical decision making in the petroleum industry heavily relies on the retrieval of accurate and context specific technical knowledge. However, the mechanisms used currently for obtaining information manually typically cause delays, inconsistent guidance and increase the possibility of human factor risk in routine and high hazard situations. This paper introduces smart Integrity, a knowledge base driven AI chatbot which is designed to be a dependable, real-time refinery and safety support by utilizing the validated operational knowledge. This study examines the challenges that occur within large language models when they are used without the authoritative and domain specific sources, it additionally highlights how connecting the AI in refinery manuals and approved safety procedures significantly improve the answering reliability and consistency. By integrating the documentation, safety standards, and operational guidelines into a structured knowledge base, the system ensure that all the responses are aligned with the company practices and guarantees regulatory requirements. The proposed approach shows how a knowledge supported assistance system can shorten information retrieval time, reduce interpretation errors, reinforce compliance with already established procedures, and improve decision making in both day-to-day operations and high-risk situations within the petroleum facilities.

Ähnliche Arbeiten