Improving acoustic leak localization accuracy using an adaptive denoising algorithm in water distribution pipes

Leaks in water distribution pipes have been localized using acoustic or vibration measurements with sensors placed on access points along the pipeline systems. Significant challenges arise for leak noise correlators when plastic pipes are increasingly used for the renovation of aging systems and the installation of new ones. In this paper, a denoising algorithm is proposed to transform the cross-correlation process into an adaptive determination of finite impulse response coefficients. The proposed algorithm is shown to outperform traditional cross-correlation methods by being less reliant on the statistical characteristics of measured leakage signals, even when the signal-to-noise ratio (SNR) is significantly low. Furthermore, theoretical analysis and numerical simulations demonstrate that the SNR of the weight coefficients is substantially enhanced when the original signals are processed through the adaptive algorithm. The effectiveness of this adaptive algorithm for accurately locating leaks is validated through experimental data from both cast iron and plastic water pipes.