Installation of gas detectors based on a quantified risk analysis

Fixed detection of flammable and toxic gases in the event of accidental leaks plays a crucial role in preventing the risks of explosion, fire and dispersion of toxic products in oil, petrochemical and chemical installations.

However, the provision and location of flammable and toxic gas detectors, based solely on operator standards and/or applicable regulations, is still somewhat empirical and dependent on feedback from operators of similar facilities and operating philosophies.

Faced with this bias, authorities and operators are increasingly calling for detector layout studies ("Gas Mapping" studies for gas detection), the aim of which is to optimize the number and location of detectors and justify the effectiveness of detector layout plans.

At present, there are several methods of designing gas detectors: the one commonly used for offshore installations considers spherical gas clouds of fixed dimensions, which must be detectable with a certain target percentage.

This method of studying the layout of gas detectors is known as the geographical method, and is used in particular by Detect3D and Effigy, two commercially available software packages.

However, the geographical method models cloud dispersion using theoretical spherical clouds of fixed dimensions, which do not take into account the various leakage conditions (leakage size, operating conditions, product released, etc.) or dispersion parameters (meteorological conditions, wind direction, etc.), which is difficult to justify physically; and the results derived from this method when applied to an onshore project require an oversized number of detectors, sometimes economically unacceptable for an onshore investment.

The method developed by TechnipFMC and presented in this article offers an alternative to this geographical approach, based on a leak scenario approach. It is inspired by the Quantified Risk Assessment (QRA) method, whose concepts it extends and adapts to the detection of flammable and toxic clouds.

To make it easier to understand, each point of methodology covered in this article will be illustrated by examples of simple applications.