A human reliability analysis methodology for oil refineries and petrochemical plants operation: Phoenix-PRO qualitative framework

摘要

The oil industry has grown in terms of quantity of facilities and process complexity. However, human and material losses still occur due to major accidents, and many of which involve human failures. These failures can be identified, modeled and quantified through Human Reliability Analysis (HRA). The most advanced HRA methods have been developed and applied in nuclear power plants, while the petroleum industry has mainly focused on process safety in terms of technical aspects of the operation and equipment. The existing HRA methodologies may not reflect the idiosyncrasies of refining and petrochemical plants regarding the interaction of the operators with the plant, their failure modes, and the factors that influence them. This paper builds on Phoenix HRA Methodology to develop a methodology specific for Petroleum Refining Operations (Phoenix-PRO). It uses as basis the Hybrid Causal Logic model, with Event Sequence Diagrams, Fault Trees and Bayesian Belief Networks. Phoenix-PRO development relied on interviews with HRA specialists, visitations to a refinery and its control room, and analysis of past oil refineries accidents. The use of this methodology for HRA of oil refineries and petrochemical plants operations can enhance this industry safety and allow for solid risk-based decisions.