Estimation of turbulence power spectra for bubbly flows from Laser Doppler Anemometry signals

Abstract The accuracy of the estimation of turbulence power spectra from Laser Doppler Anemometry (LDA) signals in bubbly flows is studied. Special attention is paid to the influence of the gaps in the signal created by the bubbles. The performance of reconstruction and slotting techniques for the estimation of the power spectrum is determined by application to synthetic bubbly flow LDA signals, and checked by application to real data. Estimation of spectra of bubbly flow signals with reconstruction techniques is found to give poor results. In general, the application of reconstruction techniques to LDA signals gives bias in the spectrum due to the addition of noise and low-pass filtering. The nature of the filtering and noise for bubbly flow signals differs from that for single phase flows. For bubbly flow signals, the spectrum is reliable up to a cut-off frequency which is lower than the cut-off frequency for single-phase flow signals with similar data rates. In addition, slopes close to - 5 3 may appear for signals which in reality have flat spectra. For single phase flow, it is possible to correct for these artefacts. The work shows that it is not possible to create similar correction techniques for bubbly flow signals. The application of the slotting technique for estimation of power spectra of bubbly flow signals leads to much better results. Estimation of the spectrum beyond the mean data rate is well possible. The performance of several improvements of the slotting technique is discussed.