Research On Superharmonics Detection Method Based On Block Compressed Sensing Theory

With the popularization of smart grids,the use of distributed power supplies and other equipment will bring extremely high-frequency super-harmonic components to microgrids.This puts forward higher requirements for the collection of data signals,and the traditional harmonic detection method is limited by the Nyquist sampling law,and the detection effect of superharmonics is not good.In order to effectively reduce the sampling complexity,this paper focuses on the problems of block acquisition,coding matrix design and reconstruction quality optimization in the super-harmonic detection algorithm,including the following aspects:1.Establish a super-harmonic signal model and draw a conclusion that superharmonic signals have good sparsity,verifying the feasibility and advantages of compressed sensing theory in the field of super-harmonic detection;2.For the first time,the block-wise compressed sensing theory is introduced in the field of superharmonic detection to solve the problem that the size of the coding matrix and the signal reconstruction time increase exponentially with the increase of signal frequency.After the block detection of super-harmonic signals,the block effect that occurs when the number of samples is small,effective improvement methods have been studied,and the signal reconstruction quality has been improved through simulation verification;3.On the basis of the traditional coding matrix,in order to further shorten the signal reconstruction time and improve the signal data coding quality,the deterministic coding matrix based on Gabor filter and the Gabor filter based on the Gabor filter,which are more suitable for the characteristics of super harmonic signals,are studied.The construction method of the random coding matrix of the device.Comparing the superharmonic signal with MATLAB simulation,it is verified that the two coding matrices based on Gabor filter have better reconstruction effect than the traditional Bernoulli coding matrix and Toplitz coding matrix;4.Aiming at the shortcomings of the traditional block compressed sensing theory in allocating the same amount of coded data to different data blocks,the adaptive sampling method based on classical wavelet transform is studied,and then a method based on saliency and two-dimensional information entropy is designed.Finally,simulations verify that the above two image block sampling methods are more in line with the characteristics of super-harmonic signals.The reconstruction effect of the sampling allocation method based on wavelet transform is stable,and the image block adaptive sampling method based on saliency and two-dimensional information entropy weighting also has excellent reconstruction accuracy under low sampling.The paper has 29 pictures,11 tables,and 95 references.