Adaptive degenerate space method for source term estimation using a backward Lagrangian stochastic model

• Main Authors: Buchman, Omri, Fattal, Eyal
• Format: Journal Article
• Language: English
• Published: 13.04.2020
• Subjects: Physics - Atmospheric and Oceanic Physics; Physics - Computational Physics; Physics - Fluid Dynamics; Physics - Geophysics; Physics - Instrumentation and Methods for Astrophysics
• DOI: 10.48550/arxiv.2004.06526

The general problem of characterizing gas source parameters based on concentration measurements is known to be a difficult task. As many inverse problems, one of the main obstacles for accurate estimation is the non-uniqueness of solution, induced by the lack of sufficient information. As the number of detectors is lowered, which is more than a plausible scenario for many practical situations, the number of possible solutions that can characterize the source increases dramatically, leading to severe errors. In this paper, a Lagrangian stochastic based method for identifying these suspected points, which will be referred to as 'degenerate space', is formulated and analysed. Then, a new procedure for quantitative prediction of the effect of deploying a new detector in space is used to design an adaptive scheme for source term estimation. This scheme has been tested for several scenarios, differing by the location of the initial detectors, and is shown to reduce dramatically the degeneracy formed by insufficient information. The combined formulation of degenerate space with the new adaptive scheme is shown to give improved accuracy, and in particular for a relatively small number of detectors.