Study On The Fault Line Detection For Small Current Grounding System In10kV Power System

Fault line detection for the non-effective single-phase earth system is importantand difficult in distribution network. At present, there are many ways to detect faultline. However, most of them are ineffective. Because a single line selection methodhas many limits, it is not suitable for all kinds of complex conditions difficultly andthe accuracy of fault line selection is very low. Therefore, the integration of a varietyof line detection principle to select fault line is the development trend in thefuture. In order to improve the accuracy of fault line detection and the quality ofpower supply, many selection methods are fused by neural network and the D-Sevidence theory in this paper. It is the deep study to detect fault line for smallcurrent grounding system.Firstly, the research status at home and abroad for fault line detection isintroduced. Next, the current main methods and the principle of fault line detectionare described and the simulation model is build by the software which namedMATLAB2010. At the same time, the traditional methods of fault line detection aretested on this model. By using this, we can lay the solid theoretical foundationfor the fuse of fault line detection in the latter part of the paper. Meanwhile, it canprovide all kinds of data for future fusion verification.Secondly, according to the characteristics of steady and transient value, we canconstruct the fault measure function. It can be used to deal with the data whichobtained from the simulation. At the same time, the processed data can be used asinput values for the fusion algorithm to detect fault lines. Then, the theory ofBP neural network, genetic algorithm based BP neural network and theD-S evidence theory are introduced in this paper. Combining fault line detection forsmall current grounding system with the intelligent algorithm is the new way tosolve this problem.Finally, in this paper, using examples to introduce the method which combinedwith genetic algorithm based on BP neural network and the theory of D-Sevidence to detect fault line, and simulation results show the advantages of thismethod and the high precision of fault line detection. Compared with othermethods, the proposed method has a great advantage.