The Application Of Wavelet Analysis In De-noising And Compressing Power Quality Disturbance

There are various kinds of disturbances in power system, which impact powerquality and safe operation of the system. In order to analyze and reduce thedisturbances phenomenon, a large number of power quality monitors are applied.Generally speaking, the sampling frequency of the monitor may be very high, whichwill bring a large data storage and communication cost. Simultaneously, samplingdata are usually overlapped by noise which will impact on the analysis results. Forraising the SNR of the signal and reducing data storage, the de-noising and datacompression methods of power quality disturbance signal are studied in this paper.A model of an 110kv three-phase network is established by Matlab/simulink SPSfirstly, and the disturbance of harmonic in power system is simulated by a type-ss1electric locomotive in this model. Secondly, kinds of short faults are analyzed, thestable and transient voltage waves are obtained.Then, a de-noising method is presented, which is based on wavelet packettransform and has better resolution in time and frequency domain. The de-noisingresults of different threshold and different threshold function are analyzed to verifythe feasibility and effectiveness of the method.Furthermore, three data compression methods are introduced and compared inthe paper, which are based on Fourier transform, Wavelet transform and Waveletpacket transform. A novel synthesis algorithm for complex electric signal, based onFourier transform and Wavelet packet transform, is proposed to compress the powerquality disturbance data in this paper. It first classifies power quality disturbancesinto stable one and transient one by wavelet's multi-resolution analysis, and thencompresses the stable and transient disturbance signals by Fourier transform andorthogonal wavelet packet transform, respectively. The simulation compression resultof three methods, which are the novel synthesis algorithm, the methods based onWavelet transform and Wavelet packet transform, are performed to verify theeffectiveness of the novel synthesis algorithm.