Quantitative Risk Assessment: Common Pitfalls in Top Event Frequency Calculation

Abstract

The main goal of process safety is to analyse and reduce risks related to industrial processes in order to ensure that the final risk on people and environment is as low as possible.
To establish whether the risk related to the process is tolerable or not it is therefore necessary to calculate the risk associated with the events under consideration and to compare the result with the selected tolerability criteria: this is the common approach used in Quantitative Risk Assessment (QRA) analysis.
The risk associated with an event, whatever is its nature (on people, on the environment or financial) is a function of the likelihood of the event (generally expressed in events/year) and the consequences (expressed in terms of damage) of the event itself.
The aim of this article is to focus on the first parameter (frequency evaluation): the main goal is not to write a math paper on frequency calculation (there are a lot of articles and specialised sources in the scientific literature that deal with the theory of probability calculation and, therefore, it would not make sense to focus on math) but to report the main conceptual errors in frequency calculation found in existing risk analysis reports.
Frequency calculation depends especially on the quality of the performed hazard identification study (e.g HAZOP) and on the proper identification of the common cause failures present within a complex system which, if not properly identified, can lead to an incorrect assessment of the likelihood of a dangerous event.
The final scope is to show how it is possible to fall into pitfalls during frequencies calculation if the hazards identification is not properly performed and if dependencies between safeguards are not properly assessed: usually these errors lead to obtain frequency values that have no physical meaning.