Research On Flicker Measurement Method Based On Improved Energy Operator And FastICA

With the continuous development of power generation technology and power electronics technology,more and more new energy systems are beginning to be connected to the grid.The voltage fluctuations and flicker phenomenon of the power grid are becoming more and more serious,which has a strong impact on people’s social life and stable industrial production.After years of research by scholars,a complete flicker measurement system has been established.However,the existing method cannot accurately measure the grid flicker signal with complex and high noise level environment.Therefore,study a simple,effective and anti-noise flicker detection method is of great practical significance.This thesis studies the flicker measurement method based on improved energy operator and FastICA.The main research contents include the following:Firstly,the thesis introduces the basic concepts of voltage flicker and fluctuation and the feature quantity and evaluation index of flicker.The method of voltage flicker envelope extraction and parameter detection at home and abroad are summarized,and the research status of flicker detection device are concluded.The flicker mathematical model of sinusoidal modulation is constructed.And it summarizes the basic principles,advantages and disadvantages of the existing envelope extraction algorithm.Aiming at the lack of noise resistance of Teager energy operator,an improved 5-point symmetric differential energy operator detection method is proposed,and the mathematical expression of envelope calculation is obtained.The method is simple to implement,and the anti-noise performance is improved before the improvement.Finally,a simulation test is conducted for verifying the effectiveness of the improved energy operator extraction envelope,which lays a foundation for the subsequent accurate measurement of envelope parameters.Secondly,this thesis proposes a method based on improved FastICA voltage flicker parameter measurement.Principal component analysis is used to judge the number of independent sources of signals,and a multi-channel observation matrix is constructed by single-channel delay method to ensure the accuracy of signal separation.Estimating the frequency of each component using a three-line interpolated FFT on the separated signal and using LSM to perform envelope amplitude correction on the separated signal,solving the problem that the separation amplitude and initial phase of FastICA are uncertain.The simulation experiments of envelope separation show that the algorithm has high precision for the detection of noisy single-frequency and multifrequency signals.Base on these research,the improved energy operator detection method is used to replace the square detection process,and the FastICA is used instead of the bandpass filtering and the visual sensitivity weighting process to simplify the flow of flicker measurement.The simulation experiment shows that the simplified measurement method is simple and effective.Subsequently,the thesis proposes a method based on Monte Carlo to evaluate the uncertainty of voltage flicker measurement.The hardware error of the acquisition process is analyzed,and the expression of synthetic uncertainty is derived.Using probability event simulation technology to generate random numbers in the probability distribution interval of each error input quantity,and obtain the size of each error input quantity.Control the number of error input quantities,and analyze the uncertainty caused by a single input by performing multiple independent experiments on a single input.Then input all the error quantities at the same time and analyze the synthetic uncertainty.Finally,the thesis establishes a laboratory platform for flicker measurement.The Fluke three-phase standard source generates a voltage flicker signal,and uses the LabVIEW platform for data acquisition.Based on the method proposed in this thesis,the flicker parameter measurement is performed and the measurement result is evaluated for Class A uncertainty.The experimental results show that when the algorithm performs flicker parameter measurement on grid data,the relative error is up to 0.3% for frequency measurement.The algorithm measures the frequency relative error up to 0.3%.Measuring amplitude,the relative error is less than 0.98% when the modulation depth is 0.01 pu,and the relative error is less than 0.2% when the modulation depth is greater than 0.02 pu.Its detection accuracy meets the requirements of the IEC standard.Compared with HHT,wavelet transform and Teager energy operator combined with FFT measurement,the proposed algorithm has higher precision.This also verifies the effectiveness and accuracy of the proposed improved energy operator and FastICA flicker measurement algorithm in practical applications.