Time-resolved WMS tomography with velocimetry for high-enthalpy flows

Experimental characterization of high-enthalpy flows entails a host of instrumentation and signal processing challenges due to facility vibrations, limited optical access, elevated temperatures, high data rates, interference from condensation and particulate matter, and more. Instantaneous velocity, pressure, and temperature fields are needed to calculate performance metrics like mass capture, localize shocks, and resolve flow instabilities. Multi-beam wavelength modulation spectroscopy (WMS) can measure these quantities at a high repetition rate, is robust to harsh environments, and requires minimal optical access for 2D sensing. Existing algorithms for WMS tomography presume constant flow properties along each beam over a scan, which is often violated, e.g., during unstart in a high-speed inlet. This talk presents a neural-implicit WMS tomography, which yields velocity, pressure, and temperature fields that are continuous in (x,y,t) and processed at the photodiode acquisition rate. This approach can potentially resolve dynamics that are faster than the scan rate. Physics-based priors may be included to promote piecewise spatial and temporal smoothness and constant stagnation properties, when appropriate. The method is demonstrated through a representative phantom study.