Noise in Turbulence Measurements Using Acoustic Doppler Velocimetry

Abstract

To validate the use of acoustic Doppler velocimeters (ADVs) for the measurement of turbulent flows, experiments were conducted in an axisymmetric turbulent jet and in approximately homogenous isotropic turbulence with zero mean flow. The jet experiments show that the horizontal RMS velocities measured by the ADV were overestimated when compared to both flying hot-film anemometry measurements and the accepted values in the literature. However, the vertical component of the RMS velocity agrees well with those of other studies. This was furthermore confirmed by the ADV measurements in the isotropic turbulence with zero mean flow. Given that the overestimated RMS velocities in the horizontal directions may be caused by Doppler noise and spikes, two postprocessing filters and a Doppler noise–reduction method were applied to the jet data in an attempt to correct the data. Although the RMS velocities decreased, they remained erroneously higher than the accepted values. A noise-reduction method for axisymmetric flows that corrects the RMS velocities to within the range of acceptable values is presented. Lastly, because the Doppler noise is thought to depend on the mean velocity, a series of laboratory experiments were undertaken to relate Doppler noise to the mean velocity. The results showed no clear relationship between the Doppler noise and the mean flow. In addition, subtracting the Doppler noise (measured for a given mean velocity) from the measured RMS velocities (at the same mean velocity) did not significantly decrease the RMS velocities.