Advancements on the use of Filtered Rayleigh Scattering (FRS) with machine learning methods for flow distortion in aero-engine intakes

In-flight measurements of aerodynamic quantities are a requirement to ensure the correct scaling of Reynolds and Mach number and for the airworthiness certification of an aircraft. The ability to obtain such measurement is subject to several challenges such as instrument installation, environment, type of measurand, and spatial and temporal resolution. Given expected, more frequent use of embedded propulsion systems in the near future, the measurement technology needs to adapt for the characterization of multi-type flow distortion in complex flow, to assess the operability of air-breathing propulsion systems. To meet this increasing demand for high-fidelity experimental data, the Filtered Rayleigh Scattering (FRS) method is identified as a promising technology, as it can provide measurements of pressure, temperature and 3D velocities simultaneously, across a full Aerodynamic Interface Plane (AIP). Τhis work demonstrates the application of a novel FRS instrument, to assess the flow distortion in an S-duct diffuser, in a ground testing facility. A comparison of FRS results with Stereo-Particle Image Velocimetry (S-PIV) measurements reveals good agreement of the out of plane velocities, within 3.3 % at the AIP. Furthermore, the introduction of machine learning methods significantly accelerates the processing of the FRS data by up to 200 times, offering a substantial prospect towards real time data analysis. This study demonstrates the further development of the FRS technique, with the ultimate goal of inlet flow distortion measurements for in-flight environments.