Research On Single-phase-fault Line Detection For Small Current Grounding Power System

Small current grounding power system which can improve the reliability of power supply is widely used in distribution power systems in China. The single-phase-faults are the most probable faults .When single-phase-faults occurs, long-time earthing operation, however, will destroy the systematic safe operation as the result of two-point, more seriously, multipoint earthing as well as the systematic over voltage caused by arc light, thus, how to detect the fault line quickly and accurately is an important subject in power system operation. Because the fault current is very small and many others reasons, it has not been solved completely to detect fault line in small current grounding power system.This thesis mainly deals with this problem. It aims to find the deficiency of the existing methods and to probe for new method of fault line selection. Firstly, mechanisms and features of single-phase grounding fault that occurs in small current grounding system are discussed as a base, and the feature distinctions of fault component between fault line and health line are listed from two aspects of transient and steady, which supplies the base of theory for latter study.On the base of these theory, this thesis has analysed existing methods of fault line selection, discussed their shortcomings, and summed up the reasons why hard to detect fault line. Then, this thesis has pointed out that each method of fault line selection has respective shortcomings, and we should colligate many kinds of methods to a multi-method to detect fault line, which can offset respective shortcomings of each method.This thesis has established a multi-method to detect fault line, which bases on fuzzy theory. It colligated six methods which includes the zero-sequence current comparative method, the signal injection method, the addition of integrating multiply zero-sequence current to others, the attenuation of direct current method, wavelet-packet transform method and the new method of total harmonic distortion of zero-sequence current, and defined method of these membership functions. And, it uses fuzzy decision-making to detect fault line with structure of network and all kinds of fault characteristics. Then, a simulation model of small current grounding power system is established, and the simulations, which include different kinds of single-phase to earth faults, show that the multi-method which is proposed in this thesis is correct and effective.Finally, the traditional methods of fault line selection only utilized fault informations of pending faults but not the advantaged informations which is included the existent samples. According to the existent samples fault line selection, this thesis has proposed that use gray correlation analysis to fault line selection for pending faults, and a new method based on gray correlation analysis of fault line selection has been established, which inculdes the signal injection method and wavelet-packet transform method. Then, through a large number of numerical simulation experiments to verify that the method established in this thesis is feasible.The method of fault line selection based on gray correlation analysis, which is different from the traditional methods of fault line selection, has inaugurated a new way to detect fault line in small current grounding power system. Combines with multi-method to detect fault line, which bases on fuzzy theory, it can make a role of function on verification, and will be greatly enhanced the veracity of fault line selection. So there will be have a certain practical significance to detect fault line in small current grounding power system.