Fault Tracking Of Circuit Breaker And Relay Protection System Based On Intelligent Algorithm

At present,the power grid of our country is developing towards the direction of intelligence continuously.As an important part of smart substation,many new secondary equipment such as merging unit and intelligent terminal have been added,and information and communication have become digitalized and networked.When the power system fails,the fault equipment should be cut off from the system by opening the corresponding circuit breaker through the protection action to reduce the scope of power failure.Compared with the traditional substation,the action signal passing through the more electrical equipment:the voltage and current signals measured by the traditional electromagnetic transformer or electronic transformer are transmitted to the merging unit,which combines the data and sends the protection device through the switch.After the protection makes a judgment,the tripping signal is sent to the intelligent terminal,which controls the corresponding circuit breaker action to cut the fault.If a fault occurs in one of the above links,which causes the protection or circuit breaker to miss trip or maloperation,the fault range is likely to be extended.In order to quickly find out the fault cause after the fault occurs in the relay protection device,circuit breaker or other devices and remove the fault in time,this paper adopted the reverse thinking and used the fault tracking method to solve the problem.Fault tracking refers to the process of finding the incorrect action of circuit breaker or relay protection by fault diagnosis after the dispatching end calls the fault diagnosis algorithm for the collected alarm information,using data mining technology to classify and extract the alarm data in substation and then find out the cause of wrong operation.That is,after the incorrect operation of the device is known,the state of the device is tracked backward to find the internal cause of its failure,which is conducive to making full use of all kinds of information sources in the power system.The main contents of this paper are as follows:1.In this paper,a fault tracking model based on reasoning chain and Bayesian network is proposed.Firstly,the reasoning chain model and Bayesian network model are introduced respectively,and the characteristics and calculation methods of them are described.The combination of reasoning chain and Bayesian network is beneficial to construct the correlation between fault cause and fault symptom effectively and intuitively.Then the concept of Bayesian suspected degree is introduced,which lays the algorithm foundation for the establishment of fault tracking model.2.The structure and main components of relay protection device,circuit breaker,merging unit,intelligent terminal and switch are introduced and analyzed respectively.And the possible failures of each part of the equipment are studied.After that,the abnormal warning information related to relay protection device and other equipment and the measurement data information such as voltage and current that can be obtained at the substation end after the fault are sorted out,and these information are classified as fault symptoms.This paper summarizes the common internal fault causes of each equipment and the fault symptoms that can be obtained when the equipment has some kind of fault.The correlation between different fault causes and fault symptoms is listed in the table.3.Based on the measured characteristic parameters of the protection device and the received alarm signal,the concept of event set is introduced.According to the fault cause set and the known fault symptom set,the reasoning chain model of relay protection device,circuit breaker and other equipment is constructed to clearly and intuitively show the causal relationship between the fault cause and fault symptoms.Then a corresponding Bayesian network model is constructed to obtain the Bayesian suspected degree of possible fault causes by Bayesian reverse reasoning,and the most likely fault causes can be determined by the Bayesian suspected degree.In addition,the method takes into account the case of data loss and puts forward a solution.In addition,it also takes into account the complicated situation of multiple failures.of different modules or multiple failures of a certain part of equipment.In the end,the effectiveness of the method is proved by the example analysis,and its application value is proved.