Implementation of Maximum-Likelihood Expectation-Maximization Algorithm for Tomographic Reconstruction of TDLAT Measurements

Abstract

Harsh flowfields, such as those found within a scramjet, can present great difficulties for conventional mechanical diagnostics. The high speed, high enthalpy, combusting flowfields make probing the flow very challenging. Optical diagnostics, such as tunable diode laser (TDL) techniques, have an advantage over their mechanical counterparts because of their non-intrusive nature. No hardware comes in contact with the flow itself, and therefore the integrity of the diagnostic is preserved and the flow is undisturbed. Traditional implementations of TDL, such as Tunable Diode Laser Absorption Spectroscopy (TDLAS), are indeed non-intrusive, but are limited by their path-integrated line-of-sight (LOS) nature. The harsh flowfields of a scramjet are highly three-dimensional and thus an optical diagnostic technique capable of producing spatially resolved measurements is required. Tunable Diode Laser Absorption Tomography (TDLAT) is a non-intrusive optical technique which combines TDLAS and computed tomography (CT) to produce a 2-D spatially resolved measurement of two key combustion diagnostic properties: temperature and species number density. Data have been collected at the exit plane of the University of Virginia’s Supersonic Combustion Facility (UVaSCF) in both the dual-mode and scram-mode of operation. Data from the scram-mode operation are presented here. The fuel used in the experiment dictates the important species to be studied in the exhaust flow. Hydrogen was used in these experiments and therefore water vapor, its major combustion product, is the targeted molecule. This work details the development and assessment of the tomographic inversion algorithm, Maximum-Likelihood Expectation-Maximization (ML-EM), and its integration with TDLAT data analysis.