Research On Complex Fault Diagnosis And Fault Deduction Technology Of Power Grid Based On Bayesian Network

With the expansion of power grid scale and the complexity of its topological structure,it is very difficult for dispatchers to diagnose real-time faults and analyze the causes of accidents after faults,especially when complex faults occur,because of the huge amount of fault alarm data collected by the data acquisition and detection system and the influx of large amounts of data into the dispatching center.Based on this background,from the perspective of large data and artificial intelligence technology,this paper presents a Bayesian network-based complex fault diagnosis and deduction technology for power grid.Firstly,this paper improves the traditional Bayesian network fault diagnosis model generation,and proposes to form Bayesian network structure by the form of correlation matrix,synthesize the information of circuit breaker and protection action of the assurance system,and determine the fault elements by Bayesian backward reasoning.Through Bayesian forward reasoning,preliminary analysis of the action behavior of the protection device and circuit breaker is carried out.In the preliminary analysis of interruption,whether there exists any action that makes the misjudgement of the remote and near backup protection,and formulate the rules of expert system to correct the expected probability of the target node,and finally determine the misoperation and rejection of the protection device and circuit breaker.Secondly,a power system fault deduction method based on Bayesian network method is proposed.According to the logic of prior probability assignment of Bayesian network nodes,protection devices and circuit breakers are classified in the process of building Bayesian network,and combined with the action information of protection devices and circuit breakers sent by the assurance system,the main protection,near-backup protection,circuit breaker and far-back protection are formed.The sequence of fault events of standby protection-circuit breaker is combined with the results of fault diagnosis such as protection device and circuit breaker rejection and malfunction analysis to form hierarchical action information,rejection information and malfunction information.The hierarchical information is integrated to form the fault deduction process.Finally,based on the A5000 system platform to achieve the above-mentioned high-level functional applications,through the platform FIS function to achieve the application of Bayesian network-based power grid fault diagnosis technology and fault deduction technology,the results will be practical,through computer display.