Inferred Binary Mixture Fraction via Physics-Informed Filtered Rayleigh Scattering
M. Bundy(1), E. Warner(2), S. Powers(2), K.T. Lowe(1)
1: Kevin T. Crofton Department of Aerospace and Ocean Engineering, Virginia Tech, Blacksburg, VA, United States
2: Advanced Measurements Reasearch & Development, Pratt & Whitney, East Hartford, CT, United States
DOI:
Understanding mean passive scalar transport in non-reacting flows is a complex task, in part due to measurement difficulties outside of controlled, lab-scale environments. This work seeks to use a combination of filtered Rayleigh scattering to yield non-intrusive, high spatial resolution measurements of the mean field of a Mach 0.8 air-helium mixture and physics-informed machine learning to denoise the experimental data and provide mass fraction of helium. This yields a simultaneous result for axial and radial velocity, static temperature and density, and mass fraction of helium without measurements of a reference Rayleigh scattering signal of each gas in the mixture. Specifically, it was found that the derived mass flow from probe measurements, CFD, and the physics-informed FRS method agree within ±0.01 kg/s.
