Research On Fault Diagnosis And Protection Alarm Method Based On Switch Quantity Information In Power Grid

With the continuous development of society,the scale of interconnection between power grids continues to expand,the topology of the power grid is becoming more and more complex,when the power grid fails,a large number of switching information is uploaded to the dispatching center,dispatchers are difficult to identify the faulty components in a timely and accurate manner,so it is necessary to introduce artificial intelligence technology to assist diagnosis.After the failure,due to the relay protection configuration and relay protection equipment defects,the protection switch information will be distorted,resulting in misdiagnosis,but the current research is mainly for the identification of abnormal protection before the fault,and the lack of abnormal protection identification after the failure.Therefore,this paper is studied from three aspects:fault diagnosis using local information,fault diagnosis using global information,and abnormal protection device identification after failure.The main contents and innovative achievements of this paper are as follows:Aiming at the problem of low fault diagnosis accuracy of traditional Bayesian network,this paper uses the collected switching information and the timing information of device action to generate Bayesian network for each suspected fault component,use its corresponding local information,perform timing reasoning,and correct the prior probability of each node in the Bayesian network,and propose a fault diagnosis method of Bayesian network grid based on prior probability correction.At the same time,considering that when using Bayesian forward reasoning for device action analysis,there will be misjudgments,and expert rules are used for correction.The simple power system model and the actual power grid in a certain area of southern China were verified,and the corrected method improved the accuracy of fault diagnosis.In view of the misoperation rejection or false alarm and false alarm of the key protection or circuit breaker device,only using the local information Bayesian network method and the existing fault diagnosis and analysis model for fault diagnosis,which may occur misdiagnosis,using the collected global information,on the basis of the traditional fault diagnosis analysis model,the difference between the expected state and the actual state of the protection startup is increased,and a power grid fault diagnosis and analysis model considering the protection of the startup information is proposed.Finally,the method of inference chain is used to carry out fault inference.The improved IEEE39 node standard model and the actual power grid in a certain area of Guangdong are verified,which shows that the improved analytical model has greatly improved the sensitivity and accuracy,and solves the problem of misdiagnosis in some cases.Aiming at how to identify the abnormal protection device problem after the fault,by analyzing the applicability of the improved Bayesian network method and the improved analytical model method under different conditions,the final fault element is obtained,and the timing constraint conditions are used to filter the protection device startup information related to the fault,and combined with the knowledge of graph theory,a relay protection alarm method based on wide-area protection startup information is proposed.At the same time,by querying the timing information and startup status of the abnormal protection device,different protection abnormal alarms are issued.By analyzing the fault cases in different cases,it is verified that the method can quickly and accurately find out the abnormal protection device after the fault occurs.