Source term estimation with deficient sensors: Traceability and an equivalent source approach

Abstract

In air pollution monitoring, source term estimation (STE) is a problem to inversely estimate the source parameters, such as their locations and emission rates, with available information of meteorology, ground geometry, as well as concentration measurements and their locations. For chemical industrial parks (CIPs), a major challenge of the estimation is the ill-posedness caused by the lack of sufficient measurements to match a vast number of undetermined source parameters. This ill-posedness results in incapacity in estimating unique solution for unknown source parameters, hence they are untraceable. In this paper, a theoretical definition of traceability is introduced to give a general criterion to determine whether a STE problem has a unique solution. This is derived from an atmospheric transport and dispersion model with description of linear sensitivity of measurements to source parameters. Based on this traceability criterion, an equivalent source method to convert an untraceable problem to traceable is proposed. The performance of this method is evaluated with a numerically simulated case whose configuration is based on information collected from a CIP in Shangyu, China. The results suggest that traceability is vital to STE problems. The method is effective for solving daily STE problems with sensor deficiency encountered in CIPs.