| Recently, with the development of high-speed railway, power electronics devices and nonlinear equipments are widely applied. Therefore, the distorted waveform of currents in traction power system is more and more deteriorated, which impacts the power quality seriously. If the harmonic which is not handled in time is poured into power system, it maybe leads harmonic resonance and harmonic enlargement. Hence, it will bring over-voltage and over-current, which will damage the safety operation of pawer system.Traditional method for detecting the harmonic of traction power system mainly includes Fourier transform and wavelet transform. Fourier transform method is only proper for stationary signal test. Also, this method has the problems including spectrum leakage and fences phenomenon. Further, the time-frequency window is fixed in Fourier transform, which is not adaptively adjusted. Hence, the resolution ratio is low. Naturally, wavelet transform is a band-pass filter, which cannot effectively extract harmonics signal of different frequency at discretion. Moreover, wavelet transform has the disadvantage of spectrum aliasing leakage phenomenon. As a result, it's important to find a proper way to analyze the harmonic signal of traction power system.The time-frequency atom decomposition method is employed for detecting and analyzing the harmonic signal of traction power system in this paper. Firstly, the characteristics of harmonic signal in electrified railway are analyzed. Then, the measured data of harmonic current in electrified railway is detected and analyzed by time-frequency atom decomposition method. Through the simulations on these measured harmonic data, the results show that time-frequency atom decomposition method is effective to detect the harmonic of traction power system. The main contributions and research contents are given as follows:(1) This paper analyzes and summarizes the present development situation of harmonic simulation and harmonic detection method in traction power system. The application fields and the basic theories of the time-frequency atom decomposition method are understanded. And then introduces the method of atom library construction, and analyzes the advantages and disadvantages of this algorithm.(2)In order to solve the high computational complexity of time-frequency atom decomposition method, the time-frequency atomic decomposition method based on intelligent optimization algorithm is given in this paper. According to the characteristics of harmonic signal, the atom library is build using sine atom. The best atoms are obtained quickly by intelligent optimization algorithm which has the strong search ability; as well the computational complexity of the time-frequency atom decomposition method is reduced. This paper adopts the particle swarm optimization algorithm and differential evolution algorithm to optimize time-frequency atom decomposition method. Through simulations, we can see that the method based on intelligent optimization algorithms has more advantages in computing speed than traditional time-frequency atom decomposition method. The method which is based on differential evolution algorithm is better in decomposing effect and algorithm stability, and using it in harmonic signal analysis.(3) The typical SS9 electric locomotive and CRH2 EMU are selected, and then this paper introduces and analyzes the main circuits and traction control system. Using Matlab/Simulink software to establish the simulation model and simulate the network side harmonic of SS9 electric locomotive and CRH2 EMU. Analyze and summarize the characteristics of harmonics.(4)This paper analyzes the harmonic characteristics of the traction power system in details. Since the energy of the fundamental is much larger than that of the other harmonic in electrified railway, modal aliasing will be appeared if we use time-frequency atom decomposition method directly. Thus, Fourier transform filter based on designated frequency is applied firstly to separate harmonic signal. Then the time-frequency atom decomposition based on differential evolution algorithm and the sine atom library is utilized to decompose the harmonic signal of the simulated and the measured in traction power system. The results show that this method is effective to detect the harmonic with any frequency in traction power system. Also, this method provides an effective way for harmonic detection of traction power system. In a word, this paper expands the application field of time-frequency atom decomposition method. |
PDF Full Text Request