Research On Improved Detection Method Of Voltage Fluctuation And Flicker

With the continuous development of the national economy,shock loads and non-linear loads are increasing,which makes the voltage fluctuation and flicker in the power grid more and more serious.Voltage fluctuations will have a huge impact on some sensitive equipment,while flicker is caused by the voltage fluctuations that make the incandescent lamp's illuminance unstable and the visual sense deviation caused by the human eye,causing visual fatigue.In order to effectively suppress voltage fluctuations and flicker,and improve power quality,it is necessary to detect voltage fluctuations and calculate flicker values.Extracting the envelope is the key to detecting voltage fluctuation signals,four common detection methods for extracting the envelope are introduced first,namely the square detection method,the rectification detection method,the effective value detection method,and the Hilbert-Huang algorithm.Through simulation verification,it can be concluded that the first three methods will have some problems such as phase shift,waveform distortion and transient oscillation;while Hilbert-Huang will suffer from under-envelope problems.Because noise effects often occur in the operation of power systems,noise interference is often detected in the detection of voltage fluctuations and flicker signals.Therefore,two denoising algorithms are introduced,which are empirical mode decomposition and wavelet.Threshold two denoising methods.It is verified by simulation that although these two denoising methods can effectively remove a part of the noise,the accuracy of denoising is convenient and does not reach the ideal value.Based on the above shortcomings,this paper proposes a local mean deco mposition(LMD)algorithm to extract the envelope map of the voltage fluctuat ion signal.Since the LMD algorithm cannot effectively extract the envelope m ap for the signal with noise,the algorithm is improved.An LMD algorithm ba sed on variational modal decomposition(VMD)+ wavelet improved threshold denoising was proposed to extract the envelope map,and the improvement was proved by the simulation verification of noise-containing single frequency volt age fluctuation signals and multi-frequency voltage fluctuation signals.The method can effectively extract the envelope map of noise-containing voltage fluctu ation signals,which makes up for the shortcomings of the LMD algorithm.Finally,the principle of flickermeter recommended by IEC is introduced in detail,and a model is built to calculate its short-term flicker value.Then,on the premise that the LMD algorithm based on VMD + wavelet threshold denoising can effectively extract the envelope,the short-term flicker value of the detected signal is calculated using the spectrum analysis method.According to the short-term flicker value calculated by the flicker meter recommended by IEC and the flicker value calculated by the method proposed in this paper,a comparative analysis is performed,and the method proposed in this paper is compared with the flicker calculated by the flicker meter at low frequencies.The value is more accurate and effective.