Differences between semi-quantitative and quantitative approaches for probabilistic assessment of technological risks

TL;DR

This paper compares fault tree-based probabilistic risk assessment with semi-quantitative methods, highlighting differences in ease of use and results magnitude through a case study on a low pressure separator.

Contribution

It provides a detailed comparison of classical probabilistic and semi-quantitative risk assessment approaches using a practical industrial case study.

Findings

Semi-quantitative approach is easier and faster to implement.

Results can differ by up to a factor of 10 between methods.

Semi-quantitative method may underestimate or overestimate risk.

Abstract

The aim of this paper is to compare the probabilistic approach based on fault trees and event trees, to an approach called "semi-quantitative" as presented by the guide Omega 10 (INERIS, 2008), for the industrial risk management. A brief presentation of the ''classical'' probabilistic risk assessment as well as the ''semi-quantitative'' approach is proposed. Then, a case study is detailed to illustrate and compare their use. This case study focuses on a low pressure separator (water/gas/condensate) for which scenarios have been identified around the undesired event "gas leak due to overpressure in the separator.'' The system includes a control loop and the following safety barriers: an alarm with operator actions, a safety instrumented system, two safety relief valves, and a blast-proof partition. This case study therefore represents a relatively conventional system in the industry. Regarding to the implementation of the two approaches, the assessment according to the Omega 10 guide is easier and faster. In terms of results, differences can be observed, up to a factor 10.