Research on power quality reconstruction based on sparse segmented compression perception
With the development of smart grids and the electric internet of things,compressed sensing theory is used to compress and reconstruct power quality data to improve the efficiency of its compressed sampling.During the research process,it was found that although compressive sensing theory can reduce data redundancy,the measurement matrix used in the reconstruction process is too large,which occupies more storage space and com-putational time.The study utilizes the periodicity and redundancy of power quality data to treat it as a column matrix,evenly dividing it into columns of the same size.Meanwhile the experiment adopts pattern similarity de-tection and screening to classify segmented power quality data into standardize power quality and distort power quality based on the relationship between the data and the standard sine function.A random Gaussian matrix is constructed for the corresponding segmented screening matrix to reduce the size and quantity of the measurement matrix,which to some extent solves the problem of complex computation and long time in reconstruction algo-rithms and improves the execution speed of compressed sensing reconstruction algorithms.
NETL
NSTL
万方数据
