Partial Discharge Signal Denoising Of Synchrosqueed Wavelet Transform Based On Iterative Layered Threshold

Partial discharge(PD)is usually in transformer or high voltage cable fault,its mechanism is due to the insulator internal or insulator surface local electric field caused by a highly centralized,and the discharge signals of duration of not more than 1 ps,tens of thousands of HZ frequency range to several hundred MHZ,generally for the random signal.Transformer partial discharge are easily cause of insulation degradation,which may cause the transformer accidents,even threatening the safety of the power system operation,so by on-line monitoring in real time reflects the state of transformer insulation has important practical significance.In the actual situation of the monitoring system,there is a strong noise interference(such as white noise).Therefore,it is necessary to eliminate noise of the collected integrated signal to obtain the accurate signal.In many of the existing algorithms,there are theoretical defects or limitations in application.In this paper,two improvement methods are proposed by analyzing the shortcoming of single threshold in the process of SST.First,firstly,according to the distribution model of SST decomposition coefficient of signal and noise,the mean square error calculation formula of the amplitude threshold weight coefficient of SST chaotic denoising is derived.Furthermore,according to the minimum criterion of mean square error,the optimal value of the weight coefficient of the range threshold is calculated.Finally,the hierarchical threshold of SST chaotic denoising is determined according to the optimal threshold weight coefficient and noise standard deviation.The second is to construct SST denoising threshold: a SST partial discharge signal denoising method which ADAPTS to the stratified threshold.Firstly,a threshold function with second order continuous derivation is constructed.Then,based on SURE unbiased estimation and minimum mean square error,constructs the new threshold function is used,the SST threshold iteration,and the SST partial discharge signals denoising threshold of each layer of the optimal estimate.Finally,based on the simulation test and measurement test,comparing various methods,verified the synchronous extrusion based on iterative layered threshold de-noising method and the feasibility and effectiveness of that based on iterative layered SST partial discharge signal de-noising threshold discharge denoising algorithm is better than others.