Research On Security And Reliability Of Electric Power Communication Network Based On Machine Learning Theory

Electric power communication network is a communication network for power system service. The electric power communication network is divided into four major categories:voice service, data service, video service and multimedia service. Running the above four types of electric power communication business needs the power communication network has a high reliability and security. In the application of smart grid, electric power communication network plays a very important role. In recent years, the electric power system has changed greatly, at the same time, the scale of the power system is constantly expanding, and the maintenance and management level of the power system has been greatly improved, but there is no corresponding reduction and reduction of power communication network. The main reason is that:a large number of power system business are loaded in the power communication network to run, transmit, leading to excessive pressure of electric power communication network, and there is no way to improve the reliability and security of the communication network greatly by the electric power communication department. In the past, the methods of simply relying on updating equipment, remolding the communication network, upgrading and expanding to reduce the number of communication interruption and failure rate, however these methods are not enough. Because only when the communication interrupt or business operation problems occurred, we just realize that there is a fault in the communication network, but then the loss has been caused. This is the way of sense after afterwards the knowledge, and it can not effectively avoid the loss of power companies.In view of the above problems, this article makes research work in the following aspects. The purpose is really effective to increase the security and reliability of electric power communication network, and the electric power communication network faults nip in the bud:Identifying the key points of the network topology in the power communication network by using complex network. Using rapid density clustering algorithm to unsupervised learning classification, rank the importance of nodes and it can be effectively used in the evaluation of the importance of the nodes in the power communication network, which is a support for electric power communication network planning.Using optimization theory to realize the distribution of power communication network bandwidth. In practical planning, to achieve the allocation of bandwidth between each link is a very complex problem. In this paper, optimization theory is used to realize the optimization of communication links, and the reasonable allocation of bandwidth is realized on the basis of optimization.Using big data analysis technology to predict the link bandwidth of power communication network. Bandwidth prediction of electric power communication network is an important aspect of power communication network planning, In this paper, we use big data analysis technology, through the analysis of the historical data of the actual use of bandwidth, build forecasting model, realize bandwidth prediction of electric power communication network.