Risk Prediction And Control Mechanism Of Power Integrated Data Network Based On Gray Theory

As the network size of the power integrated data network continues to expand,the network structure becomes more complex,and the probability of occurrence of network failures and the cost of interruption are also increasing.The power integrated data network needs to have the ability to monitor and analyze the state of the network in real time.However,in contradiction with the high requirements of operation and maintenance,as a special network serving the power industry,the power data network is constrained by the layout of power plants,substations,and power grid lines in the process of construction and so it has the characteristics of improper network structure and high node vulnerability.On the other hand,compared with the public network,Due to the particularity of the bearer service,the power integrated data network has higher requirements on the security and reliability of the network.Traditional network risk management mainly focuses on post-event control,which is to quickly and accurately solve problems that have occurred in the network,and reduce actual losses,such as network risk assessment research and fault location research.This is a passive emergency method.Prior control is to reduce the probability of problems by making plans and taking measures.The sooner the risk is discovered,the sooner the measures are taken,the lower the cost of risk control.Therefore,how to change from passive to active and shift the focus of the operation and management of the network from feedback control to prior control and form a scientific and reliable mechanism is the need of the times.Since the service carried by the power integrated data network has high requirements on reliability and timeliness,a risk prediction algorithm based on entropy-gray model is proposed,which is expected to achieve effective network risk prediction and accurately and quickly identify potential risks.The algorithm includes three stages:risk indicator value prediction,weight determination of indicator value and risk discrimination.Firstly,the gray model is used to predict the value of multiple network risk indicators that affect the risk,which ensures the accuracy of the risk prediction and simplifies the time and space complexity.Secondly,the dynamic evaluation of indicators based on entropy method is proposed.The weight distribution is completely dependent on the objective problem domain,and it is better adapted to the dynamic networks and complex environments.Finally,a more reasonable risk warning is given by comprehensively considering the network risk indicator value and network dispersion degree in the risk discriminating stage.On the basis of risk prediction results,in order to achieve prior risk control and effectively reduce the probability of network failure,a risk control decision mechanism based on grey correlation is proposed.First,construct an ideal measure that meets the control objectives,and then use gray correlation analysis to measure the similarity between the candidate measures and the ideal measures,and to evaluate the merits of the candidate measures.The grey correlation matrix can reflect the influence of each evaluation index of the risk control measures,and can avoid the deviation caused by the single evaluation index.Finally,when making decisions,the gray-related projection coefficients are defined as the basis for the candidate measures.The simulation results show that the decision mechanism can achieve autonomous rapid decision-making while ensuring accuracy.The risk prediction algorithm and the risk control decision-making mechanism studied in this paper have tried the idea of becoming passive and active in modern network operation and maintenance management,which has important application significance for the intelligent management of power integrated data network.