Prime Photonics (Prime), in conjunction with Virginia Tech (VT), are proposing a novel approach to quantitative detection of airborne ice crystals and supercooled water droplets. The envisioned instrumentation system, dubbed the Temperature and Phase of Water via AI-based Detection using Raman scattering—or TAP WADR, builds upon recent breakthroughs at Virginia Tech (VT) for incorporating machine learning-based classification and regression models for non-spherical particle sizing, while recognizing the value of spontaneous Raman scattering spectroscopy for temperature and phase measurements.
In Phase I, Prime and VT will develop a library for both water structure/size and temperature and incorporate it into VT’s existing AI/machine learning architecture. A single-particle proof-of-concept experimental apparatus will be developed to characterize supercooled water droplets, and at least two forms of ice that will be used for AI model test and validation. Finally, a concept for TAP WADR system integration will be developed for a NASA facility.
In Phase II, the TAP WADR system will be scaled up to perform thermo-structural measurements across a 4+ foot water-ice cloud test section. In the first year, risk reduction testing will occur at a VT wind tunnel. In the second year of Phase II, the instrument will be transitioned to a Glenn icing facility to validate performance for mixed phase icing clouds.
The first product planned for the proposed technology will be an ice/water cloud characterization system for ground test applications. It will be marketed to icing test facilities, specifically NASA facilities such as the Glenn Icing Test Facilities, Propulsion Systems Laboratory, and Icing Research Tunnel.
Industrial facilities such as GE Aviation’s Icing Wind Tunnel, and private test laboratories such as NTS or MDS facilities. In the future, TAP WADR technology may be used to detect airborne ice in flight. Beyond aircraft applications, the technology is applicable to atmospheric research, and chemical/food process control.