Assessment of ignition probabilities - Mathematical models

Many industrial facilities store, synthesize or transfer flammable products. Depending on their flammability characteristics, these products are likely to form flammable mixtures in air. In the event of ignition, dangerous phenomena such as fire, flames or explosions can occur.

As part of a quantified risk assessment, it is therefore necessary to estimate the probability of a cloud igniting on contact with ignition sources. An error at this stage can have significant consequences for the conclusions of a risk analysis. It's also important to remember that quantified risk analyses are increasingly used for plant sizing purposes. In other words, facilities (equipment, structures, buildings) are designed to withstand stress intensities corresponding to a certain frequency.

In 2017, as part of quantified risk analyses, there are three ways of assessing ignition probabilities:

1. database evaluation ;

2. evaluation using semi-quantitative approaches ;

3. the application of more or less complex mathematical models.

Approaches 1) and 2) are presented in [SE 4 020] . This article reviews the mathematical models available for estimating the probability of delayed inflammation. The focus is on so-called temporal mathematical models, which are now recognized as the most advanced approach for estimating inflammation probabilities.