The Application Of Information Fusion Method Based On Fuzzy Inference And Neural Network For Fault Diagnosis In Distribution Network

When the grounding fault occurs in Pow Distribution Network(PDN), the voltagewill fall down sharply, the decrease of the voltage will push the asynchronous motorelectromagnetic torque significantly reduced, cause product damage or scrapped, theshort circuit current is also likely to burn out devices, and even cause the whole systemunstable, lead to blackouts. The accurate and rapid distribution network fault diagnosissystem, can reduce the outage time and reduce the loss caused by power outages, canmaximum guarantee continuous, reliable and safe operation of the distribution network,to improve the quality of PDN. So the study of distribution network fault diagnosis hasimportant theoretical significance and practical value, has considerable social andeconomic benefits.There are many researches about distribution network fault diagnosis for now,mainly include the fault line selection, fault phase selection, fault location and the faultdiagnosis for equipment of PDN. This article mainly aims at the analysis and researchof the fault phase selection in PDN, the grounding fault is the most common in the faultphase selection, According to the of the slow speed, misdiagnosis and false negativecase in the present situation of the fault phase selection, proposed the fault diagnosismethod based on information fusion. Single information is not enough to fully reflectthe characteristics of the object, and fusion method can give full play to the full andcomprehensive role, so in theory based on information fusion method is very good atprocessing the uncertain information, it can improve the diagnostic accuracy andreliability of the system. the fusion result is better than single diagnosis.The clue of is article is, based on information fusion, neural network and fuzzyreasoning theory, took the ground fault as diagnosis object, According to the the lowspeed, misdiagnosis and false negatives of the present diagnosis method, werespectively used based on neural network information fusion method, based on thetheory of fuzzy information fusion method and information fusion method based on thecombination of both in grounding fault diagnosis, in order to improve diagnostic speed,eliminate miscalculation and false negatives. The main research content of this article isas follows:①Deep understanding the current situation of diagnosis method through readingliterature about the fault diagnosis of PDN,according to the slow speed、 misjudgment and false negatives of current diagnosis methods, we proposed the based oninformation fusion of fault diagnosis method.②Research the information fusion theory and the fault diagnosis theory,proposed the information fusion method based on neural network and based on thefuzzy theory information fusion method applied to phase selection fault of thedistribution network. As the grounding fault an example for the simulation to verify thefeasibility and effectiveness of the method based on information fusion, and analysis thesimulation results.③Set up simulation model of the distribution network in matlab, simulate andanalysis the common faults, to ensure the feasibility and effectiveness of the model.④Proposed the neural network information fusion method combined with fuzzytheory, took the grounding fault as an example, simulate and analysis on the fault.compare the the effect when single use a method, by comparing the simulation resultsbased on the combination of these two information fusion methods on phase selection ofdistribution network fault diagnosis can improve the diagnostic speed, eliminatingmisjudgment situation, to totally improve the diagnostic performance.This article according to present problems of grounding fault diagnosis method inPDN this article proposed based on the information fusion method applied to groundingfault diagnosis. Simulated and verified the method based on neural network informationfusion and based on fuzzy reasoning of information fusion method and the method ofcombination both. The results show that the method which based on information fusionapplied in fault diagnosis of power distribution network is feasible, at the same timecombining neural network and fuzzy theory can foster strengths and circumventweaknesses, improve the speed of system diagnosis, eliminates the misjudgment andfalse negative situation, to improve the performance of the system overall.