Two novel doppler signal detection methods for laser doppler and phase doppler anemometry

In laser Doppler and phase Doppler anemometry the detection and identification of the Doppler signal in the noisy background signal is essential in order to perform fast and reliable measurements. In the present paper two novel concepts for Doppler signal identification are introduced. The first method is based on the evaluation of the Doppler burst envelope in the spectral domain using a real-time discrete Fourier transform (DFT). This method was found to be very reliable even for signal-to-noise ratio (SNR) values as low as -10 dB; however, it still requires the specification of a certain threshold in the power spectrum above which the data processing is initiated. The second method is based on the continuous estimation of the SNR value of the signal and is realized in an analogue electronic circuit using band-pass filters of different bandwidth. This method may be regarded as a spectrum modulator. The advantage of a continuous SNR estimation is that only signals or the portion of the signal which is above a pre-set SNR value is transferred for further processing. Since the accuracy of the parameter estimation (e.g. signal frequency) by any processor is dependent on the SNR level, this method allows one to specify the lower limit of the SNR to achieve a required accuracy. The developed 'SNR trigger' does not require the setting of any amplitude trigger level and can therefore be reliably applied to complex two-phase flows with strong spatial variations of the particle concentration and background noise. The robustness of both novel signal identification schemes is demonstrated by computer simulations and measurements.