Research On Inter-harmonic Detection And Identification Method Of Power Quality Compound Disturbance In Power Grid

With the widespread use of non-linear load and time-varying load, the load of grid is becoming more and more complex, meanwhile the power quality problem is getting worse, which brings more and more damage to power grid and power users. The steady-state disturbances like harmonic and inter-harmonic and the transient disturbances like voltage swell, voltage sag and interruption caused by the problems have brought more damage to the grid, thus searching for more effective power quality analysis methods, especially the inter-harmonic detection and compound disturbance location and classification method, to achieve comprehensive power quality analysis and evaluation are of important theoretical and practical significance.Aiming at the challenge of inter-harmonic detection in power grid, the spectrum analysis theory was introduced, and the MDL criterion, beam space transform, root-MUSIC with extended Prony algorithm was used in inter-harmonic detection. Firstly the MDL of the information theory was introduced to determine the total number of input signal source firstly, then the beam space transform preprocessed root-MUSIC by which carried out the spectral peak search was used to determine the frequency of inter-harmonics, whose output is sorted by the spectral peak. On the basis, the extended Prony method was used to get the amplitude, phase and attenuation factor of inter-harmonic, so as to get the complete signal information. The characteristic parameters of the inter-harmonic frequency and the attenuation factor obtained by the above algorithm can be used in the following classification of compound disturbance.Aiming at the problem of low accuracy of transient power quality disturbance time identification, the transient power quality disturbance detection algorithm based on integrated empirical mode decomposition method(EEMD) and improved TK energy operator was proposed. Firstly, the EEMD algorithm was used to decompose the initial signal contained transient disturbance, by whom get the intrinsic mode functions(IMF) from high frequency to low frequency. Then the information of extracted high frequency disturbance could be decomposed by improved TK energy operator, so as to obtain the amplitude and time value occurred in the complex transient power quality disturbance. Meanwhile, the Hilbert transform was used to analyze the high frequency IMF component, and the Hilbert spectrum of the signal was obtained. Simulation results showed that the compound disturbance time and amplitude can be detected by this algorithm effectively, with high accuracy and strong anti-noise performances. The characteristic parameters of compound disturbance amplitude and time obtained by the above algorithm can be used in the following classification of compound disturbances.Aiming at the challenge of power quality compound disturbance identification, the compound disturbance classification was attributed to the category of spatial analysis, and a new classification method on solving power quality disturbance was proposed, which based on EEMD, improved TK energy operator, Hilbert transform,extended Prony algorithm and decision tree classifier. Based on the above algorithm, the characteristics of main frequency components, disturbance amplitude,disturbance amplitude and attenuation factor was obtained, with the threshold of each branch of decision tree classifier determined, a simple decision tree for fast disturbance identification was designed. It can avoid the error caused by lack of training samples, meanwhile the recognition time was shortened to a large extent. 17 kinds of power quality disturbance signals including 10 compound disturbances were selected, and simulation results showed that the recognition rate is high, with strong anti-noise ability, and it can both applied for single and compound power quality disturbance signal classification.A power quality monitoring and analysis system based on Lab VIEW platform was built. The HIOKI 3196 harmonic analyzer was used to test the power quality of 10 k V power distribution system in Shanxi Dehui Steel Tube Plant, to compare with the designed platform, meanwhile the platform was simulated with the transient pulse and voltage sag for example, and the correctness and feasibility of the algorithm were both verified.