Research On Harmonics Generation And Detection In The Power System

As the power electronic technology has a wide range of applications in various industries and domestic equipment and the number of the nonlinear loads is increasing, such as the LED lighting, high-speed rail electrification traction locomotive, the steel mill, electric arc furnace, electric welding machine and other equipment used in the heavy industries. All of them will create a large number of harmonics to inject into the power grid, which causes the power grid harmonic pollution and makes the power system voltage and current waveform distortion to seriously affect the quality of power supply. These harmonics on the grid also cause a bad effect at the operation of power generation and supply equipment with safety and economy. At the meantime, the harmonic may be cause a wrong operation by the electrical equipment like relay protection and automatic device. In addition, the general watt-hour meters most lack of harmonic measurement function for now, which leads to the power charge unfair. The non-harmonics source users not only absorb the harmonics from the power grid but still have to pay for the extra charges. On the contrary, the users of the harmonics source convert part of the fundamental wave power to harmonic to feedback grid but pay less. Therefore, further systematic research on harmonics in the power system is very necessary under the condition of the market economy and safety in production.This paper firstly introduces background and significance of the research, and the direction of power harmonics and research status at home and abroad, especial the power harmonic source load model and the harmonics detection methods. Then the theoretical knowledge of power system harmonics was introduced in detail, including parameter definition, harm, power metering, etc. After that, the paper analyzes the mechanism of the common loads in power system and builds the load models based on MATLAB/ Simulink platform, including rectifier, inverter, dc motor, electric locomotive and an integrated circuit model based on the port crane load. Finally, the article detailedly analyzes the harmonics detection methods including Fourier transform, instantaneous reactive power theory, the wavelet transform and neural network theory, and then put forward a kind of harmonics detection method by fast Fourier transform (FFT) and wavelet packet transform combined.Using the measured voltage harmonics data from a public substation room to verify the feasibility and accuracy of this method. In addition, paper proposes a method of neural network combined with Back Propagation and Adaline neural net work, whose accuracy is higher in amplitude and phase than the FFT with adding window double-spectrum-line interpolation. In another hand, the data of LED driver harmonic generation is measured by double-spectrum-line interpolation FFT with adding different window functions. The results found that the precision of FFT improves a lot by adding double spectral lines interpolation. The experimental results are also different by adding different windows, which can get a suitable window for the harmonics characteristic, and provide a good reference for the power system harmonics detection management.