Fault Type Identification Method of DC Loop Network Based on CEEMD-SVM

Aiming at the DC loop network system, this paper proposes a DC loop network fault type identification method based on CEEMD-SVM. The DC line current signal is collected as the original signal and decomposed by CEEMD. Using the combination of correlation coefficient and energy distribution, two most sensitive IMF components are selected from the IMFs components obtained by CEEMD decomposition. The statistical characteristics and frequency domain characteristics of the selected IMF components are calculated to form a feature matrix. Then the feature matrix is input into the first two-class SVM model for fault detection, and the detected fault data is input into the second multi-class SVM model for fault type identification. MATLAB simulation verifies the effectiveness of the proposed method.