Study On Transient Power Quality Disturbance Detection And Recognition Methods

In modern power system,various power quality problems have caused with the increasing use of various power electronic devices and nonlinear,impact,and fluctuating loads.With the development of society and the improvement of living standards,new sensitive electrical loads,which are more sensitive to the transient power quality problem in the power grid,are rapidly growing.The economic losses of sensitive load users are increasing year by year because of the transient power quality problems.Therefore,the requirements for power quality are getting higher and higher.Long-term monitoring and intelligent analysis for transient power quality disturbances(TPQDs)can be benefit to find out and correct the transient power quality problems,so as to minimize economic losses.This paper focuses on four aspects of TPQDs monitoring and analysis: the power quality signal denoising,TPQDs detection and localization,feature extraction of TPQDs,and TPQDs classification.For the power quality signal denoising,the BM3D(Block-matching and 3D collaborative filtering)algorithm and iterative adaptive kernel regression method with strong abilities to protect the image detail features are first introduced into the power quality signal denoising,and a adaptive denoising method based on the improved BM3 D algorithm and a denoising method based on the improved iterative adaptive kernel regression for PQDs are proposed in this paper.The two proposed methods not only have the advantages of avoiding estimating the noise variance and setting the filtering threshold artificially but also can well overcome the difficulty of power quality signal denoising(which can effectively suppress the noise and well protect the PQD features at sudden change points).In addition,the denoising method based on the improved iterative adaptive kernel regression,which has small computational complexity,is easy to implement.The simulation experiments are performed and the comparisons with the widely used wavelet threshold denoising method are analyzed.The experiment results show that the proposed two methods are effective.For the TPQDs detection and localization,this paper presents a novel transient disturbance detection method based on differential singular value decomposition(SVD)and a new disturbance detection method based on the Hilbert transform and the slip-SVD-based noise-suppression algorithm in order to compensate for the shortcomings of traditional disturbance detection methods and improve the TPQDs detection accuracy in low-SNR environment.To demonstrate the effectiveness of the two proposed methods,extensive tests are conducted on the diverse simulation and actual disturbances,and some disturbance detection methods are compared.Test results show that the proposed two methods have a small amount of computation,fewer parameters,good noise immunity,and good detection results for single andcombined disturbances even if they occur to the zero crossing of the voltage waveform.In addition,the disturbance detection method based on the Hilbert transform and the slip-SVD-based noise-suppression algorithm has low false detection rate,practicability,and compatibility,and can provide some important features for classification.Such advantages let us believe that the proposed methods can be a good choice for real-time power quality monitoring system,and be easily integrated into a digital fault recorder,and then the rate of false and miss triggering of the digital fault recorder in noisy environments will be reduced.For the feature extraction of TPQDs,multiple technical means are used to extract multiple frequency and time domain features from the decomposition waveforms of the method based on the Hilbert transform and the slip-SVD-based noise-suppression algorithm.These extracted features can represent the characteristics of many types of single and combined disturbances,and have good classification ability and noise immunity.In addition,they are benefit to design the disturbance classification system based on simple rules.For the TPQDs classification,this paper presents a new multi-label classification method for transient disturbances.The proposed multi-label classification method includes the modules of amplitude and additive disturbance determination,and the classification result is determined according to the output label value.Extensive tests are conducted on 10 kinds of simulated single disturbances(including normal voltage)and 13 kinds of simulated combined disturbances(including double and triple PQDs)with different SNR values and actual transient disturbances.The test results show that the proposed classification method can effectively classify single and multiple combined disturbances,even in low-SNR environments.At the same time,the proposed method has very small computational complexity,which has superiority in the application environments with higher real-time requirements and lower hardware performance.The research results of this paper improve and enrich the theoretical system of transient power quality monitoring and analysis,and provide an important theoretical basis and effective way to develop the real-time TPQDs monitoring and intelligent analysis system.