Operational Risk Analysis of LNG Loading

Abstract

Abstract In recent years focus has increased on conducting risk analyses that are specifically tailored to the operational phase of a selected infrastructure. This interest is the result of several operators wanting to properly understand their risks in the operational phase and in this way, through proper preparedness, being able to avoid or mitigate incidents and major accidents and to improve resilience towards these unwanted events. Consequently, we have seen a move towards the use of more advanced risk modelling tools to handle the increasing complexity of the oil and gas producing systems. In the present paper we present the results of a detailed risk analysis of an LNG export loading operation at a Jetty. The analysis takes into account the interaction between the vessel crew and the land based operators. The unwanted consequences considered vary from time limited disruptions of LNG transfer, which results in production delays, to the unintended overfilling of the cargo tanks where the worst case scenarios not only cause significant economic losses but human fatalities as well. The paper introduces a method for analysing risk that has several novel characteristics i.e. it brings together the possibility of technical failures, the errors caused by individual human performance, and the organisational influences into the operation. We have used Bayesian Networks as the platform to construct and analyse this test case. An accurate representation of the loading operation has been carefully constructed to account for the system in real life operation. The results of the assessment express the expected loss as an aggregated loss into a single quantifiable monetary measure. By using the so-called max-propagation facility of Bayesian Network the most likely configuration that generates a specific unwanted event may be identified. We use this to identify specific contributors of risk and the mitigation measures that carry the best utilization. Introduction The objective of operational risk management is to satisfy multiple stakeholders' requirements in controlling losses from operation. We are focusing on the owners and hence the owner's risk management during operation of a complex technical installation is of interest. One particular objective of an operational risk management system is that it should provide a strong ability to " forecast?? conditions and give input to management prior to the development of dangerous conditions. A second objective is to approach risk in a holistic way in assessing risk from unwanted losses due to loss of life, property and environment. This requires strong understanding of the reliability of the technical systems, the organizational performance and of how the interaction between the two affects the total systems performance. This implies that a practical requirement for risk management of unwanted losses during the operation phase is that the risk modelling shall encompass the systemic interactions covering technical and human organizations. The solution for systemic risk requires a holistic perspective in hazard identification, risk analysis and risk mitigation. This may provide the operator with a useful set of risk based indicators to guide the risk management process. To properly understand how systemic risk modelling can be applied with a holistic perspective encompassing technical and human organisational aspects we have within the project at DNV-Research & Innovation undertaken a comparative study of 4 dominating contemporary risk perspectives, namely:Normal Accident Theory (Perrow /1/)Epidemiologic Models (the barrier theory, Reason /2/, /3/)High Reliability Organisation (Weick et al. /4/) andResilience Theory (Hollnagel /5/, Hollnagel, Watts & Leveson /6/)