Research on matching pursuit algorithm based-on Laplace wavelet for partial discharge pulse detection

A new method has been proposed for detecting partial discharge pulse under noisy and interferential condition, which is based on the matching pursuit algorithm. With a dictionary of Laplace functions, whose structure is similar to that of PD pulses, the proposed scheme can extract PD pulses from the signals within noise and interference successfully. In contrast to the other methods, the new way is easy to realize and can obtain more information of PD, such as the time of occurrence, frequency component, damping rate and so on, which are useful for location, classification and identification of PD pulses. The extraction results of simulation and field experiment data revealed that the effect of extracting PD pulses is very satisfactory with the proposed algorithm.