Wavelet decomposition applied to fluid leak detection and isolation in presence of disturbances

This paper deals with wavelet decomposition applied to fluid leak detection and isolation in process engineering. The developed algorithm is sensitive to minor leaks and robust to all the disturbances (hydraulic and thermal perturbations) and transient operating modes. The proposed approach is based on the Discrete Wavelet Transform (DWT) of the sensor measurements, the Fast Fourier Transform (FFT) and the properties of the cross-correlation function. Experimental results obtained on different processes show the efficiency of the proposed approach.