Studies of the Statistical Bias and Signal Characteristics of Laser Velocimeters

摘要

Analyses are presented of (1) the statistical bias individual realization laser velocimeters (LV's) and (2) the theoretical signal amplitudes obtained from dual-scatter (fringe) LV's. The analyses are compared to experimental data obtained from a subsonic, confined, seeded jet. The statistical bias study reviews the various formulations relating LV sampling characteristics to the velocity field and shows that, theoretically, bias is reduced by random sampling, introduction of fixed-time delays between samples, and nonuniformities in seeding. These concepts are tested by a bias simulation program which simulates LV response to a turbulent flow by applying various sampling laws to hot-wire anemometer data. The finite-difference algorithms suggested by various investigators to eliminate bias from LV data fail if the sampling rate is less than the local integral time scale of the turbulent flow. Velocity data for the near field of a uniformly seeded subsonic jet exhausting into a plenum were obtained by pitot probes, hot-wire anemometers, and a Bragg-diffracted laser velocimeter. No consistent bias of the data was observed, indicating that routinely correcting LV data for statistical bias may lead to significant errors.