Detection of Voltage Variation Signal Based on Double Wavelet

A method based on db1 and db10 wavelet to detect voltage variation (VV) signal was proposed. At first the signal was decomposed by dbl wavelet and low frequency coefficients modulus of the decomposed signal were obtained, on this basis probable time intervals of any VV events could be determined and the beginning and end points of each time interval were recorded simultaneously. Then the signal was decomposed by db10 wavelet to obtain the first layer high-frequency decomposition coefficients, and within half signal period before and after the beginning and end points of each time interval determined by dbl wavelet the modulus maxima value of first layer high-frequency decomposition coefficients were detected to determine the beginning and end points of non-zero-cross points and non wave peak points. According to the concrete judgment rule the accurate beginning and end moment of VV would be determined. Simultaneous the amplitude of the voltage variation would be determined by above-mentioned low frequency coefficients modulus. Using the proposed technique one will have the ability to measure the varied magnitude and duration of VV events. The information gained by it can be used to distinguish among similar VV problems. Test and simulation result show that the method is feasible and effective