The Research On Fault Feature Extraction And Information Fusion Approach For Fault Diagnosis Of Grounding Grid

Grounding grid is the important facilities to ensure the electrical equipment and personal safety. Circuit branch of grounding grid, as buried under ground, subjecting to soil erosion and the discharge current, will lead to erosion and crack of grounding conductor, which caused huge economic losses. Large area excavation is usually adopted in the gird fault diagnosis; this method has the disadvantages of blindness,heavy workload and the slow speed. Therefore, the research on the fault diagnose of grounding grid is of great significance. Using grounding grid as the main research object, this paper deeply studied the establishment of the model on the fault diagnosis of grounding grid, the fault diagnosis of grounding grid based on DC testing, the fault feature extraction and recognition methods based on high frequency test and the fault diagnosis of grounding grid based on information fusion.Under the affection of DC excitation source, the grounding grid is considered as a linear network of resistor. Using the voltage of node as the fault feature, and apply RBF neural network to train and identify for fault location of grounding grid. This paper establishes the simulation model of grounding grid under the high-frequency excitation, which solve the problem of the limited number of accessible nodes caused by fault characteristic information deficiency On the basis of wavelet packet decomposition voltage signal feature extraction of accessible nodes, combined with the method of fault diagnosis based on RBF neural network to diagnose the fault of grounding grid.Based on DS evidence theory, a fault diagnosis method is also proposed to improve the fault diagnosis accuracy of grounding grid. Using fault diagnosis results under DC excitation and High-frequency excitation as evidence sources, the proposed method combined the above two fault diagnosis method by evidence theory. The simulation results show that the accuracy of the fault diagnosis method of grounding grid based on information fusion can be improved obviously.