Research On Power System Harmonic Analysis Based On HHT Algorithm

With the wide application of power electronic devices, it brings serious harmonicpollution problems. So the research of appropriate harmonic analysis method, then takeeffective harmonic suppression measures based on the analysis results, has importantpractical significance.Firstly, this paper studies the basic theories of harmonic, and the advantages anddisadvantages of the existing methods of harmonic analysis. The Hilbert-Huangmethod is proposed when analyzing harmonic signal against the deficiencies of thetraditional methods. By doing theoretical analysis and simulation comparison with thefollowing three methods: Hilbert-Huang Transform, Fourier Transform and WaveletTransform, the Hilbert-Huang Transform shows its advantages in analyzingnon-stationary signals, which overcomes the deficiencies of analyze mutations inharmonic signal by traditional methods.Secondly, researches and improvements have been made to the problems whenanalyzing harmonic signals by the Hilbert-Huang algorithm.(1) Influence of samplingfrequency and sampling number on HHT analysis. In order to make sure the analysisresult is not influenced by low sampling frequency or non-whole cycle sampling, thesampling frequency is taken at least8times above the highest frequency signal, and fullperiod sampling is adopted to the greatest extent.(2) The problem of EMD terminationcriterion. Using modal amplitude and estimate function as the component of thetermination conditions, the frequency number of FFT analysis frequency signal as thedecomposition, thus the decomposition efficiency is improved and the programming inMatlab is easily reachable.(3) The problem of modal aliasing. Add different white noiseto the signal to carry out EMD decomposition, and make average to offset the addedwhite noise for many times the average offset added white noise, thus avoid the mixtureof high frequency signal components into the low frequency component. This helps tosuppress the modal aliasing phenomenon effectively.(4) The problem of HHT endpoint.The method of extreme point continuation is proposed to improve the end effect, andshows its magnificent improvement by comparison with the commonly used mirrorcontinuation method.Finally, improved HHT algorithm is used to do simulated analysis on typicalharmonic signals. A lot of analysis has been made here and based on the analysis resultof multiple sets of data, the average deviation of signal starting time is0.489%, thedeviation of termination time is0.256%, the deviation of duration time is2.15%, thedeviation of instantaneous frequency is2.486%, the deviation of instantaneousfrequency is0.373%. Compared with the error limit of5%, They are allowable error.Therefore, the improved method can detect various information such as the mutation,the time, instantaneous frequency and instantaneous amplitude of non-stationarydisturbance signal, and meet the requirement of all parameters needed to detectinstantaneous harmonic signals. The improved HHT method can be well applied toanalyze non-steady harmonic signal, and has a practical application value.