Research On Complex Fault Diagnosis Of High Voltage Power Grid Based On Protection Multi-source Information Fusion

With the development of society and economy,the scale of the power system has gradually expanded,the grid structure has become more complex and the degree of automation of the power grid has also increased,so that the acquisition of grid data is relatively convenient.When complex faults occur in the power system,a large amount of fault information floods into the dispatching operation center in a short time,the grid staff need to grasp the core information,quickly determine the location of the faulty component,and remove the fault in time to restore the power supply.However,the explosive increase of fault data and the occurrence of wrong information and missing information will inevitably lead to misjudgments.Therefore,it is necessary to apply artificial intelligence technology to fault information processing,forming fault diagnosis and deduction algorithms for complex power systems,to provide theory support for staff to quickly response to internal faults,determine faulty components in time,accurately grasp the nature of the fault,and ensure the safe and stable operation of the system.Traditional fault diagnosis algorithms are mostly based on the operation information of protection and circuit breaker devices.This type of diagnosis method is greatly affected by the completeness and accuracy of such switch information.When the fault occurs,the internal analog information will change accordingly,and contains the fault characteristics.This paper comprehensively considers the switch information and analog information collected after the fault,forming a complex power system fault diagnosis algorithm based on information fusion.This paper first takes the time sequence information of the protection and circuit breaker device action information into account,improving the traditional fault diagnosis method based on the Bayesian theory.According to timing coordination principle of the relay protection device in the power system,the switch information is filtered to optimize the data quality applied to fault diagnosis,so that the Bayesian fault probability obtained based on the switch information is more accurate,and the diagnosis result is more accurate and reliable.Then,this paper proposes the fault criteria of the line,bus,and transformer in the system,quantifies the failure probability of the component based on the analog information.The probabilities of failure based on switch information and analog information are used as the evidence,fused at the decision-making level based on D-S evidence theory to get the composite failure probability of components based on information fusion.Finally,this paper proposes a fault deduction method for complex power system.The component-related protection and circuit breaker devices are processed hierarchically,which are divided into main protection,near backup protection,far backup protection and circuit breakers;combined with the expected action probability of each protection and circuit breaker device obtained by Bayesian forward inference,evaluate the action of the protection and circuit breaker device,and find out the mis-operation,rejection device,and time stamp error device,and make timing corrections,determine the fault time domain of the component,and complete the fault deduction.