Source term determination in real time, i.e., location, rate of release and its use for the simulation of dispersion of toxic substances would be extremely valuable information as support for emergency actions on industrial sites and their vicinity. For that purpose, CFD is used to model the dispersion while the Bayesian inference algorithm enables the estimation of the source term in real time. The modelling platform based on these developments can produce real time 3D concentration fields owing to precalculated wind fields patterns selected from a database on the basis of real time meteorological data acquisition. The industrial site of Lacq (France) has been chosen as a pilot area, and the key hazardous substance considered in this study is hydrogen sulphide (H2S).
Keywords: source term determination, Bayesian inference, real time simulation, industrial sites, toxic substances, H2S, hydrogen sulphide, France, emergency response, CFD, computational fluid dynamics, dispersion modelling, air pollution, air quality