Research On Power Grid Fault Diagnosis Based On Multi-source Information Fusion

With the rapid development of smart grid technology and information technology in China,the power industry,which is dominated by power grids,is moving towards smart and intensive processes to meet the high requirements of contemporary social and economic development for stable operation of power grids.However,power grid accidents occur from time to time.In order to avoid further expansion of the accident and endanger the operation safety of the power grid,dispatchers need to quickly and accurately diagnose the faulty components of the power grid in order to take corresponding adjustment measures in time.However,there are many types and large quantities of information to be reported when the power grid fails.The traditional power grid fault diagnosis mainly relies on the action of the protection device before and after the fault and the tripping of the circuit breaker.The information source is relatively single,and it is difficult to apply to protection,circuit breaker refusal to operate,misoperation,and misreporting and missing information in the power grid.In the case of complex faults and multiple faults,it is more difficult to ensure the accuracy of the diagnosis results.This paper analyzes the information source of power grid fault diagnosis,and uses the characteristics of electrical and switch information changes before and after component faults to study a power grid fault diagnosis method based on multi-source information fusion.Firstly,analyze and apply research on multiple information sources that can participate in power grid fault diagnosis.Using the displacement information and charging status of the circuit breaker,a method of power grid fault area division based on breadth-first search combined with electrical quantity information is proposed.Then study the extraction of the switching quantity information and electrical quantity information fault characteristic values of the suspicious components in the fault area,and use the Bayesian network with credibility to analyze and calculate the fault degree of the components based on the switching quantity,and use the multi-discrimination wavelet decomposition and reconstruction.The method obtains the energy distortion and failure degree of the electrical quantity signal of the component,and selects it as a separate evidence body.The improved D-S evidence fusion method is used to fuse the fault degree obtained from the switching quantity information and the electrical quantity information to construct an improved fuzzy C-means model,and cluster the fusion results to make the final diagnosis conclusion.Finally,an example simulation is carried out through PSCAD software to verify the feasibility and effectiveness of the power grid fault diagnosis method,which has application value for fault diagnosis in power grid dispatching.