Reconstruction of power fault signal based on big data fusion

Abstract

In order to solve the storage difficulty of power failure signals under big data fusion, which leads to the waste of hardware resources, this paper proposed a Dice-gradient projection for sparse recovery (Dice-GPSR) algorithm based on the compressed sensing theory to reconstruct the fusion signals. Firstly, the algorithm optimizes the support set based on the Dice coefficient measure matching criterion, changes the search direction and iterative search path of the gradient projection sparse reconstruction algorithm, and improves the convergence speed. Secondly, the sparsity of the signal is estimated and the flow of Dice-GPSR algorithm is given. Finally, Dice-GPSR algorithm is used to reconstruct the original signal and compare the reconstruction performance with other reconstruction algorithms. The simulation results show that Dice-GPSR algorithm has higher reconstruction accuracy and reconstruction signal to noise ratio than other reconstruction algorithms.