Ethernet Fault Diagnosis Of Traveling Wave Substation In Yunnan Area Based On ITLO-SOM

Traveling wave fault location devices have been widely used because of their high location accuracy and applicability,which have formed a multi-manufacturer and multi-platform scenario in the Yunnan Power Grid.With the continuous promotion of the concept of “Strong Smart Grid”,Yunnan Power Grid builds a network system consisted of dispatching center and substations equipped with traveling wave fault location devices.It breaks the situation of the data between different manufacturers cannot cooperate with each other.The power communication network provides great support for the development of the network system of traveling wave fault location devices in the direction of intelligence and information interaction.However,there are large number of substations in Yunnan Power Grid,Ethernet faults in substations must be able to diagnose faults quickly and intelligently.This is essential for maintaining the stable interaction of the data in the network system consists by traveling wave fault location devices.In order to improve the stable operation capability of Ethernet in substations,it is necessary to continuously develop and improve the fault diagnosis technology related to Ethernet.In view of this background needs,this paper studies the method of diagnosing of Ethernet faults in substations equipped with traveling fault location devices in Yunnan.Proposes a diagnostic scheme for Ethernet based on an interactive teach-learn optimal self-organizing feature mapping neural network.Firstly,it analyzes and studies the Ethernet structure and Ethernet fault characteristics in the traveling wave station in Yunnan.The performance data changed in real time in the network equipment is regarded as the feature vector of the Ethernet operating state,that is,the diagnosis basis.After comparing the existing analysis methods of various fault diagnosis methods,it is decided to use the self-organizing feature mapping neural network as a diagnostic method to established a fault diagnosis model.Then,for the problem that self-organizing feature map neural network is susceptible to the quantity and arrangement of training data in the modeling process,which leads to a low diagnostic accuracy,this paper proposed an interactive teaching-learning optimization based on the original teaching-learning optimized the self-organizing feature map neural network fault diagnosis model is verified by the fault history data of the Ethernet interface in the original teaching-learning algorithm to optimize the self-organizing feature map neural network fault diagnosis model and perform fault diagnosis of the communication network.Finally,the diagnosis model and perform fault diagnosis of the Ethernet interface in the substations equipped with traveling wave fault location devices of Yunnan region's power grid.The results show that this method can greatly improve the accuracy of the fault diagnosis of the communication network in the equipped with traveling wave fault location devices of Yunnan Power Grid.