Identification And Analysis Of Transient Power Quality Detection In Power System

With the continuous advancement of science and technology,the continuous upgrading of various power equipment,the requirements for power quality of equipment using electric terminals are constantly improving.At the same time,various power quality problems are constantly appearing in the power system due to the continuous application of various rectifier devices and inverter devices.In order to optimize the power quality in the power system and ensure the safety of power consumption of various types of electrical equipment,it is necessary to detect,identify and analyze the power quality disturbance signals.Therefore,this thesis proposes a method for detecting and analyzing transient power quality detection in power systems.The combination of wavelet transform and S transform is used to detect and analyze the disturbance signal.The BP neural network algorithm is used to classify and identify the disturbance signals.In this thesis,eight basic models of disturbance signals and one normal signal are constructed,including sinusoidal signals,six single disturbance signals and two composite disturbance signals.Firstly,four high-frequency sequences and one low-frequency sequence obtained by four-layer decomposition of wavelet transform are used to detect the start and end time and duration parameters of various disturbance signals.The signal can be clearly observed through the four-layer decomposition characteristic signal map of the disturbance signal.The point condition is abrupt,the corresponding start and end time,duration of the disturbance are obtained.Then,the S transform is used to detect the characteristics of various disturbance signals,the start time,the stop time and amplitude changes of the disturbance signal can be obtained according to the three-dimensional grid diagram,the contour map,the envelope diagram and the amplitude curve diagram.Harmonic components and content can also be detected for disturbance signals containing harmonics.Finally,for the classification and identification of disturbance signals,a three-layer BP neural network is constructed.The input,implicit and output layers are 5,12 and 8 dimensional structures respectively.The feature vector is constructed by the feature quantity of DB4 wavelet 4 layer decomposition,and the feature vector is used as the input of BP neural network to classify and analyze the disturbance signal.The simulation results show that the combination of wavelet transform and S transform can accurately detect the start time,the stop time,duration,amplitude change,harmonic components and content of various disturbance signals.The accuracy of classification and recognition of all kinds of disturbance signals by BP neural network algorithm is above 92.5%.It is proved that the proposed method has high detection and recognition accuracy and algorithm feasibility,and can achieve detection and recognition targets for various types of disturbance signals.This paper consists of 32 figures,7 tables and 63 references.