On the mixed acoustic and vibration sensors for the cross-correlation analysis of pipe leakage signals

Abstract

Pipe leak detection can be carried out by various acoustic/vibration sensors, such as acoustic pressure, velocity and acceleration sensors. The purpose of this study is to investigate the effectiveness of the mixed acoustic/vibration sensors for the determination of pipe leakage. Firstly, a propagation and attenuation model of leakage signal in the liquid-filled pipeline is established based on frequency response functions. Further, the expressions of the cross-correlation function of different sensor types are derived with respect to the differential properties of Fourier transform. Then, the influence of signal-to-noise ratio (SNR) on the cross-correlation function is analyzed. Numerical simulation is employed to generate pressure, velocity and acceleration signals by differentiating the leakage source signals. Simulation results indicate that the use of pressure sensor acts as low-pass filtering operation, whereas the velocity and acceleration sensors act as band-pass filtering. Finally, experimental work is carried out on a medium density polyethylene (MDPE) pipe. Leakage signals are collected synchronously by hydrophones and accelerometers. Experimental results show that the cross-correlation function of hydrophones fluctuates slowly with peak value being less obvious, while the cross-correlation function of accelerometers fluctuates rapidly and has a sharper peak. In comparison, there is a compromise in the fluctuation for the sake of peak sharpness in the cross-correlation function of the mixed sensors, indicating that the corresponding cross-correlation result are less affected by background noise.