Study On The Classification And Location Method Of Power Quality Disturbance

With the large-scale application of new energy and power electronics technology,power quality problems of power grid have met a series of new challenges.At the same time,all kinds of precision load which sensitive to power quality are more and more access to the power system,and the demand for power quality is increasing sharply.The classification and location of power quality disturbances is an important part of power quality analysis,and it can also provide theoretical support for power quality control.In this paper,disturbance feature extraction,classifier design,harmonic source location and voltage sag source location are studied.The content of the work is as follows:(1)A feature extraction scheme based on sparse decomposition for disturbance signals is proposed.Based on the characteristics of the disturbance signal,the corresponding overcomplete dictionary is constructed,and the eigenvectors suitable for the disturbance classification are extracted from the result of the sparse decomposition.MATLAB simulation proves that the eigenvector can map common disturbance signals to linearly separable high-dimensional space,thus effectively simplifying the difficulty of the design of the classifier.(2)The rating system for multi label classification is introduced.The subtraction clustering method is used to improve the training method of RBF neural network,and the multi label classification of 47 kinds of composite power quality disturbances is realized by using RBF neural network.Simulation results show that this method has low Hamming loss and high average accuracy.(3)A harmonic source localization method based on harmonic node equation is proposed.When the harmonic node equation is positive definite or overdetermined,the least square method is used to solve the nodal equation.When the harmonic node equation is underdetermined,particle swarm optimization is used to estimate the injected harmonic current of nodes.The IEEE-13 node system is built by SIMULINK,and the validity of this method under different working conditions is verified.(4)A direction discrimination method for voltage sag source based on positive sequence equivalent impedance is proposed.The accuracy of the method is high and there is no inconsistent result of three-phase judgement.The disturbance energy is used to evaluate the reliability of each monitoring device,and a weight based location method for voltage sag source localization is put forward.The PSCAD simulation proves that the method has faulttolerant ability.