Single Phase Adaptive Reclosure of Transmission Lines Based on EMD Approximate Entropy and LS-SVM with BCC

Empirical mode decomposition (EMD) approximate entropy and least squares support vector machines (LS-SVM) with bacterial colony chemotaxis (BCC) were used to detect the fault signals in single phase earth, and then distinguish the transient faults and permanent faults on transmission lines. After faults occur, different kinds of faults have different complex levels of voltage transient signals which can be measured on busbar. By analyzing the characteristics of the voltage transient signals, this paper used EMD to decompose the voltage transient signals, then calculated the approximate entropy of intrinsic mode functions as the eigenvectors, BCC algorithm is used to optimize LS-SVM, faults are classified using EMD approximate entropy and LS-SVM with BCC. A lot of simulation results of EMTP and Matlab calculation show that the method can distinguish the transient faults and permanent faults accurately and verify its correctness and practicability. Compared with LSSVM with grid search, the superiority of applying LS-SVM with BCC to single phase adaptive reclosure is proved.