Weighting leak frequencies in industrial risk analyses

Pipeline leaks account for a significant proportion of the dreaded events identified in risk analyses carried out as part of French regulatory studies (hazard studies), but also as part of so-called QRA (quantitative risk assessment) studies.

To quantify the occurrence of a pipe leak, one of the most commonly used methods is to extract a so-called linear frequency of occurrence (/year · m) from a database, and multiply it by the pipe metre. The main advantage of this approach is that it is relatively simple to implement. However, a number of questions arise regarding the use of "average" values reported in databases. We need to distinguish between two types of database:

• sector-specific databases, such as EGIG 6 ^th for gas transmission pipelines, or UKOPA 4 ^th for hydrocarbon transmission pipelines;

• multi-sector" databases, such as the CPR 18 E, better known as the Purple Book.

Using the leakage frequencies reported by EGIG 6 ^th as part of a risk analysis for a gas transmission pipeline is not a priori problematic, since the values reported in this database are specific to this sector of activity. In other words, the values reported in this database are good estimators of leak frequencies. On the other hand, when you want to assess the frequency of pipe leaks in an industry not covered by a specific database, the question arises as to how representative the values are. Indeed, how can we judge whether the value reported in the database is too pessimistic, or on the contrary too optimistic, with regard to the various causes of leakage and prevention measures identified during the risk analysis? Being able to deviate upwards or downwards from the values reported in the databases is therefore necessary if we wish to carry out a specific risk analysis, taking into account the particularities of the system under study (nature and intensity of the causes that could lead to the leak, performance of the prevention measures implemented by the manufacturer, etc.).

This article reviews the current state of knowledge on the subject of weighting leakage frequencies in quantified risk analysis.