Research On Denoising And Location Of Transient Power Quality Signal Based On Wavelet Transform

A lot of sensitivity and non-linear loads were used in the grid. When power quality problems occurred, those loads will not work, and thus impact on industrial production and people’s daily lives, even causing huge economic losses. In order to improve the level of power quality, reduce the impact of the disturbance, we must study the characteristics of power quality problems and their causes.Signal denoising and disturbances location were two main areas in the study of transient power quality, but those problems were separated and studied separately in most researches. Only to study signal denoising, may be able to achieve better denoising effect, but may also filter out disturbed mutation point, so that to provide inaccurate positioning for disturbance location. Only to study disturbances location under ideal conditions, the interference caused by noise was ignored, so that the result is too ideal. To solve this problem, an improved threshold function was proposed, utilized stationary wavelet transform(SWT) for signal denoising and locating disturbances, which overcome the shortage of both independent, the main work and achievements as follows:At first, introduced the definition and classification of power quality problems, and gave a detailed description for four transient power quality disturbances. According to the current advances, we discussed the main methods of research power quality problems. Explained Fourier Transforms firstly, and then analyzed the development of wavelet transform, at last in-depth study of the basic principles of Wavelet Transform.Obtained the mutation point information accurately was the premise for analyzing transient power quality disturbance signal with noise, so it must be accurate to retain the mutation point features when denoised signals. To solve this problem, an improved threshold function of wavelet denoising method was proposed, which provided de-noising and smoothing ability by changing the adjustable parameters, and then selected Stationary Wavelet Transform(SWT) to signal decomposition. Described the rules of selecting decomposition layers and the wavelet basis function, modified the general threshold estimation method, calculated wavelet threshold layer by layer, and combined with the improved threshold function for wavelet coefficient denoising of each layer, at last using wavelet modulus scaling coefficients and denoised coefficient for signal reconstruction. Simulation experiments show that compared with the soft and hard threshold denoising method and other improved threshold function method, the denoising method of improved threshold function proposed in this paper can better filter out noise and preserve mutation point features, the disturbance moment of beginning and ending can be clearly observed from the processed wavelet coefficients.Stationary wavelet transform has good time- frequency localization capabilities and translation invariance, and the wavelet coefficients of denoised transient signal decomposed by SWT can reflect the signal singular. Among them, the maximum value of the mold correspond to the disturbance point directly, so the starting and ending time of transient power quality disturbance signal were accurately located by utilizing the theory of singularity. Simulation experimental results show that wavelet transform can accurately locate begin-end time of different transient disturbance, and further prove that the proposed denoising algorithm can retain the mutation point information effectively.