Effect of Frequency Analysis Algorithms on Velocity Data Using Ultrasonic Doppler Method

Abstract

An ultrasonic velocity profile (UVP) method can measure instantaneous one-dimensional velocity profile along the measuring line. It is non-intrusive technique, and it can be easily applied for existing facilities. The UVP method utilizes the ultrasonic Doppler frequency reflected on tracer particles for obtaining the velocity. For analyzing the Doppler frequency, there are some algorithms such as Fast Fourier Transform (FFT), auto-correlation, and so on. However, the influence of the algorithms on measuring velocity profile has not been investigated. In this study, difference of the algorithms for velocity measurements was investigated using a developed system. It consists of an ultrasonic pulser-receiver which transmits and receives ultrasonic waves, a high-speed digitizer which records wave data, and a PC which analyzes and calculates velocity. FFT and auto-correlation method were used for the calculation. In case of the FFT, there are 2 methods to calculate the Doppler frequency. One is a method to calculate the frequency of maximum value in the power spectral density (PSD), and the other is a method to calculate the average value of the frequency in the PSD. Velocity distributions were compared to a laser Doppler velocimeter (LDV), and the appropriate algorithms were chosen depending on the flow conditions. From the results, it was clarified that the auto-correlation method could reduce the number of ultrasonic repetition for the measurement.