University of Colorado Boulder Background
Laser Doppler velocimetry (LDV) has long been a standard technique in experimental fluid mechanics labs. Researchers at the University of Colorado extend the capabilities of this technique by reshaping the intensity profile of the optical probe beam and by developing a machine learning-based signal processing scheme to analyze the expected signals which can be more complicated than those from LDV.
The light scattered by a particle passing through a probe beam carries with it a history of the particle’s trajectory through the beam. When the beam is patterned, the scattered light signal is matched with the properties which gave rise to the motion via a machine learning model.
E. F. Strong, A. Q. Anderson, M. P. Brenner, B. M. Heffernan, N. Hoghooghi, J. T. Gopinath, and G. B. Rieker, “Angular velocimetry for fluid flows: an optical sensor using structured light and machine learning,” Opt. Express 29, 9960-9980 (2021)
Stage of Development
Technology Readiness Level (TRL): 4
Signal processing technique makes no compromise between spatial and temporal resolution
Uses readily available seeding particles and requires only a low seeding density
May function with existing LDV hardware
Flow facilities (wind tunnels, water channels)