Insulator leakage current data compression based on EMD and compressed sending

Abstract

In order to increase transmission and processing speed, it is essential to compress the non-stationary leakage current in the insulator monitored system. Compressed Sensing (CS) combines sampling and compression with a small amount of sample to reconstruct signal well, which not only reduces hardware requirements, but also improves compression efficiency. However, computational complexity is high in CS recovery process of non-stationary signals. In this paper, CS combines with empirical mode decomposition method (EMD). Decompose Non-stationary leakage current into a finite number of stationary intrinsic mode functions (IMF), and then CS processes stationary IMFs. The experiment results indicate that it enhances not only the speed of process and operational efficiency but also compression ratio and reconstruction accuracy in the case that CS processes leakage current signal by EMD decomposition.