Application of signal-adapted wavelet to the detection of PD

The paper presents a method of selecting the optimum wavelet that is adapted to the partial discharge (PD) signals to make the wavelet transform work well. The method maximizes the energy of the projection of the signal on the scales by optimizing a certain cost function. Wu's elimination algorithm is used to construct compacted orthogonal wavelets during the optimization. The physical experiment results and analysis show that the PD signals can be extracted more efficiently and deform less by using this method, especially in the case of detecting weak PD signals.