Identification of line source of environmental pollutants using Bayesian inference with Super-Gaussian function

Abstract

This research proposed a new method for the line source identification by combining the Bayesian inference with the Super-Gaussian function. A simple boundary layer was simulated with CFD (Computational Fluid Dynamics) to obtain the measurements and adjoint concentration flow field. The difference between the measurements and the adjoint concentrations was formulated according to Bayesian theorem to estimate the source parameters stochastically. The Super-Gaussian function was employed to approximate the line source with several tractable coefficients. The results showed that this method is capable of identifying all the parameters without any prior geometric information about the line source as long as enough information is provided by sensors.