Accurate measurement of the velocity field and associated flow properties, using LDV, demands the use of an optimized signal processing system with the ability to extract signals from noisy photodetector output. Although, ostensibly, both auto-correlation and spectrum analysis techniques convey the same information, the unique nature of the Doppler signal provides a special situation where auto-correlation has certain advantages. The theoretical approach to examine the analysis techniques and the potential errors that can be caused by some simplifying assumptions are discussed. To overcome the limitations of the theoretical approach, results from both computer simulation and experimental measurements are provided. Special characteristics of the Doppler signal make the sampling process and the signal detection to be crucial to getting accurate measurements. It is shown that in the DFT approaches used partial fill of data block occurs resulting in a strong impact on the lowest measurable signal quality and the associated measurement accuracy. The ability of the auto- correlation approach to overcome these problems and the robustness of the processing scheme are discussed. In addition, the auto-correlation based processors (IFA 750 & 650) provide the ability to optimize the sampling rate for each burst and the system in real time to provide higher measurement rate, better accuracy and the ability to measure lower turbulence. These capabilities are borne out by results from actual measurements.
|