Use of accident precursor event investigations in the understanding of major hazard risk potential in the Norwegian offshore industry

The recent offshore accidents at the Macondo and Montara fields in the US and Australia have demonstrated the importance of learning from major accident precursors in order to appraise the risk potential involved in critical offshore operations. This is fully realised by the Petroleum Safety Authority in Norway, which has a specific requirement for such learning in its regulations. However, an unfortunate practice has been developed by the major players in the Norwegian offshore industry, whereby potential is severely and systematically downplayed, probably to limit the negative exposure if the actual potential consequences were known. The present article analyses 45 major accident precursor investigations in order to demonstrate the effect of downplaying the potential of major accidents. It demonstrates how the risk potential classified in investigation reports has a random relationship to the more objective risk potential, as shown in the national risk indicator project conducted by the Petroleum Safety Authority. This is further demonstrated by comparing company investigations with authority investigations in four cases where parallel investigations were performed.

Risk Assessment and Management Using Accident Precursors Modeling in Offshore Process Operation

Offshore oil and gas operations are located in remote and often harsh marine environments. An offshore development can never be completely safe; however the degree of safety can be increased by selecting the optimum design, and developing proactive risk management strategies. This requires the identification and assessment of major risk contributors, which can be accomplished using quantitative risk assessment techniques. Dynamic failure assessment is a new approach in process safety management, which enables the real time failure analysis of a process. This approach uses Bayesian and joint probability theories to develop a predictive failure model for a given process. As a process proceeds and generates incidents and accident precursors, the accident occurrence probability is predicted. This paper presents a methodology based on the concept of dynamic failure assessment and its use in revising the risk profile for a process system based on accident precursor data modeling. An event tree is formed for a given abnormal event. Subsequently, using accident precursor data from the facility prior and posterior failure probabilities of events are calculated. A predictive model is developed using joint probability theory. Accident likelihood is combined with consequence analysis results to estimate posterior risk profile. Application of the proposed methodology is demonstrated on a process in an offshore process facility.