Multi-open-path laser dispersion spectroscopy combined with Bayesian state estimation for localizing and quantifying methane emissions

Abstract

A global effort towards improved quantitative understanding of greenhouse gas emissions is taking pace. This includes developing source identification, quantification, and apportionment in an attempt to understand global budget and trends, but also developing monitoring systems making emission reduction commitment verifiable. In this context, we demonstrate a novel approach to continuous methane emission monitoring at the spatial scale of an industrial facility. By combining multi-directional measurements of path-integrated methane concentrations with Bayesian state estimation, we show a realistic tomographic gas plume reconstruction, its evolution in time, and the associated estimation of the source map. The method is validated using measurements from controlled methane releases over a domain of area 120×40 m2. For the first demonstration, a two dimensional geometry has been used in the gas flow model; nevertheless, sources are located within 3–12 meters, and mass emission rates are estimated within <30% for 80% of the cases.

Keywords

• Methane emission
• Gas emission inference
• Open-path spectroscopy
• Bayesian inference
• State estimation