Electronic signal processing in laser doppler annemometry

Abstract

The three commonly used electronic processors for laser Doppler anemometer systems are the spectrum analyser, the frequency tracking demodulator and the period timing device. Extensive analytical studies on spectrum analyser and frequency tracking demodulator were carried out to evaluate their performance. Studies of spectrum analyser revealed that the measured power spectrum is related to the probability density of the flow velocity field. However, the quantitative assessment of relevant statistical parameters associated with flow field from this spectrum is difficult. The investigations of frequency tracking demodulator indicated that its output is a measure of the instantaneous velocity of the flow. Although it is possible to extract the mean velocity of the flow field, a correction has to be applied if higher order moments, such as mean square, autocorrelation etc., of the flow field are to be determined from this output.

Detailed experimental investigations were carried out to evaluate the performance of period timing devices with different validation schemes. The processing system used for simulating these devices consisted of a fast analog-to-digital converter and a computer. An optimum validation scheme is suggested which, when incorporated in a period timing device, is capable of measuring the Doppler frequency and hence the particle velocity correctly.

In the literature, the "biasing" error has commonly been considered an inherent property of laser Doppler anemometer. It is shown in the present work that "biasing" is due to the electronic processor and averaging procedure employed in the experimental investigations. Further, weighted averaging procedures suggested in the literature to determine the time-averaged statistical quantities from measured particle velocity data involve the measurement of the transit time of the particles through the measuring control volume. The investigations carried out here showed that the accuracy of the time-averaged quantities determined by weighted averaging procedures depends on the method employed to measure the transit time. Also, weighted averaging procedures are valid only when the particle arrival rate is correlated with particle velocity. A new averaging procedure suggested here gives correct time-averaged quantities for a three-dimensional flow field in which the particle arrival rate can be either correlated or uncorrelated with the particle velocity.

Finally, a complete electronic processing system capable of providing information on instantaneous, particle-averaged and time-averaged velocity information is suggested.