Optimization of an urban monitoring network for retrieving an unknown point source emission
Résumé
In this study, we present a new methodology, basing on a new entropic approach for designing monitoring networks in urban environments. The objective of the optimal design is the source reconstruction of atmospheric releases. The methodology is presented by coupling the optimization algorithm Simulated Annealing (SA) with the Renormalization Inversion (RI) technique and the Computational Fluid Dynamics (CFD) modelling approach. A detailed validation of the unknown source reconstruction is performed using the concentration measurements from the sensors in the obtained optimal networks in 20 trials of the MUST tracer field experiment. This study exhibits the importance of a new optimality criterion inspired from the Renormalization Inversion (RI) approach for optimising the deployment of a limited number of sensors in urban scale in order to estimate the position and the intensity of unknown emissions.