High-fidelity forward model for determining combustion efficiency of industrial flares

The petroleum refining process utilizes hydrocarbon combustion by industrial flares to dispose of waste gas streams. These flares emit substantial quantities of greenhouse gases into the atmosphere. As a result, there is an ongoing need for monitoring and control technology to minimize the products of incomplete combustion, including ozone-forming highly-reactive volatile organic compounds (HR-VOCs) and human carcinogens. Remote sensing of emission products using a combination of a hyperspectral infrared imager and a high-performance retrieval algorithm can provide a robust, low-cost solution. Here, we outline the construction of a high-fidelity forward model to be used by our retrieval algorithm. This model takes user-supplied layer temperatures and molecular concentrations and provides the retrieval algorithm with bandpass-specific spectral radiances (for least-squares fitting against measured data) and derivatives with respect to the dependent variables (to adjust the next iterative step).