The Testing Theory And Method Of Power System Harmonic And Voltage Disturbance

In recent years, with the extensive applications of nonlinear power electronic devices, the abominable effects of harmonic wave and voltage fluctuation on the power grid and crimes become increasingly serious, which result in the serious deterioration of power quality and seriously affect the entire power grid security and the economy running. Therefore, real-time, accurate detection of harmonic content and voltage fluctuation in power system, and the master of the exact power grid harmonic in the actual situation have practical significance to prevent the harm harmonics, power system security and protect important practical significance. Active power filter control power system harmonics is an effective means of pollution, and detection is the key to achieve harmonic and reactive power compensation. In recent years, with the rapid development of power electronic technology, there has been substantial progress in harmonic suppression and reactive compensation for the theoretical study and practical application home and abroad. Especially, there has been some breakthrough achievement in the emergence of active power filter. With the application of these results, it will play a significant role in reducing and eliminating hazards harmonic grid, voltage fluctuation and maintaining the green environment.This paper firstly introduces the working principle of the power system harmonic detection and control, the main cleaning methods and shortcomings; and then elaborates the research progress on the power system harmonic detection and control; after that, we focuses on the research of neural network system stability theory, the harmonic detection method based on the neuron theory, the harmonic detection methods based on the synchronous reference coordinates, the new active filter injection modeling and predictive control technology, circuit breakers mutase monitoring systems and industrial linkage program. The main work of this paper include the following areas.The design and research of power system harmonic detection and control system in this paper are based on neural network theory. As the neural network is a complex nonlinear system, therefore, Chapter2in thesis details the nature of neural network system dynamics and characteristics, and conducte a study for a class of time-varying delay with the binary neural network system stability. Compared with the existing literature, the activation functions do not require the assumption that the smoothness, monotonicity and boundedness. With the use of Brouwer and Banach fixed point theorem, and in accordance with Halanay type inequality, we looked at the existence and global exponential stability of equilibrium point, and get a group of sufficient conditions for exponential stability of the neural network system.Chapter3analyzes the thesis of power system Harmonic diagnose based on neural network. First of all, based on the Fourier basis function, we give the model of neurons for the harmonic analysis of power system, and then discusses neurons algorithm convergence in detail, at last, gives a detailed proof and the neural network training steps. Experimental results show that the proposed harmonic analysis method has high accuracy and computational speed, the lower calculation. Therefore, it has a greater value in the measurement of power system harmonics, and it is a good real-time, high accuracy for a wide range of harmonic detection.On the basis of the basic principles of positive sequence fundamental wave extractors, Chapter4proposes a new improving synchronization reference coordinates for reactive power, harmonics and negative sequence current detection method. Through the synchronous rotation tracking to the fundamental positive sequence voltage vector, It could reduce the calculation of phase-locked loop and trigonometric functions, and get the reference current signal of APF quickly and exactly. Simulation and experimental results demonstrate that the use of the detection method of active power filter has good compensation for the performance.Chapter5proposes a new hybrid active power filter IHAPF (Injection Type Hybrid Active Power Filter) as an example, and gives detailed analysis for their mathematical modeling. The model can be some reference for other types of modeling active filter. Based on it, this paper presents the theory of improved unbiased model predictive control method based on gray system GM (1,1). This method can reduce the deviation of predictive control to achieve predictive control without delay. Simulation and experimental results show that the use of gray can be a better predictive control to overcome the the impact of active injection power filter to harmonic compensation. It also shows the correctness of the mathematical model and the effectiveness of the current forecast control methods.In order to improve the voltage quality and solve the dynamic voltage quality problems, we need to detect voltage fluctuation waveform of the fundamental and the harmonic voltage of information. Based on mathematical morphology filtering and the "selective harmonic elimination" algorithm, Chapter6introduces hybrid filter algorith, and then puts forward fast detection algorithm which can separate the fundamental wave and specific sub-harmonic. After that, it proposes the hybrid filtering algorithm composed of morphological filtering and specific harmonic elimination. Simulation and experimental results show the practicability of detection algorithm.Finally, this paper summarize the main results of innovative research, and look forward to the next step.