Large Pool Fires in the Process Industry. A Simple Model for Safety Distances Estimation

Abstract

In recent years, the global production growth and the complexity of processes operation have increased safety concern, especially during the design project, considering potential accidental events. Pool fire is one of the most frequent accidental events. Some procedures for safety distances (SDs) estimation from it were reported. However, there is a lack of adequate simple models to directly estimate SDs covering a wide range of substances and pool diameters, considering a continuous interval of receptor vulnerabilities. In this paper, simple models to estimate SDs are proposed, suitable to be incorporated into complex models (e.g., safety management systems, layout optimization). A surrogate model is achieved by adopting a reference one. Methodologically, the main variables that play a key role modelling a pool fire were selected and their influence on the SD was analyzed to propose simple mathematical expressions. NLP optimization problems were solved to optimize the model parameters. Models with R-squared values of more than 0.99 were obtained when compared to the reference model. A conservative, accurate and easy-to-use model was achieved with the aim of developing safer plants or for their safe operation by improving the optimal process design, or implementing safe management systems, considering risks.