Interfacing the Probabilistic Bowtie Analysis with the Regulatory Risk Matrix

Abstract

Probabilistic bowtie risk analysis is a quantitative risk analysis method widely used by industry and recommended by regulatory bodies for studying major hazard scenarios. Its standard use consists in estimating the mean frequency of occurrence of the top event and subsequent consequences by propagating the mean frequency of occurrence of basic events through fault and event trees. Such a practice does not take full advantage of the probabilistic nature of bowtie risk analysis. In this paper, we use an industrial case study to introduce a new and simple graphical representation of probabilistic Bowtie (BT) results that provides a readily interpretable interface with the regulatory risk matrix used by decision-makers. Instead of collapsing the probability distribution of occurrence of an undesirable event into a single value, which may be misleading, it expresses the risk of occurrence as a likelihood of belonging to each level of the risk matrix. This representation is very informative in terms of risk assessment. Slightly more complex than a single value in a risk matrix, this representation more informative in terms of risk assessment and some benefits can be expected for decision-making. It could typically lead to optimization of the asset integrity management policy (periodic inspection and maintenance) or an improved definition of the testing strategy for the safety barriers.