Studies of statistical bias in turbulent flow parameters measured with a laser velocimeter

Bragg diffracted individual realization laser velocimeter measurements were taken in a jet flow in which efforts were made to attain the conditions that theoretically result in bias. Arithmetic mean velocities and turbulence intensities were determined from the measurements and compared with pitot and hot-wire measurements, but no clear evidence of statistical bias was found. In addition several bias correcting algorithms were employed to evaluate the laser velocimeter measurements and it was found that routinely correcting for bias can produce serious errors. To analyze effects that might reduce bias under experimental conditions a numerical experiment was devised to show the relation of seeding rate and flow structure to bias. Further analysis showed the bias reduction that can occur when the velocities of some particles crossing the probe volume are not recorded and when the seed distribution was not perfectly uniform. In addition, the bias reduction due to three dimensional flow was considered.