Research On Information Security Evaluation And Intelligent Protection Of Wind Farm SCADA System

With the increasing maturity of wind power development technology,an increasing number of countries regard wind power as a transition from traditional energy to renewable energy.However,we connect the wind farm SCADA system to the Internet to realize the remote management of the wind farm,which will make the network security problem of the information security hidden in the SCADA system gradually appear.This problem will pose a serious threat to the information security of the wind farm SCADA system.Our paper aims to improve the autonomous defense and automated protection of the wind farm SCADA system.The work is shown below.(1)Safety analysis and vulnerability analysis of wind farm SCADA system.Firstly,the basic functions of the wind farm SCADA system are introduced,and its information security structure is explained.Secondly,the threats and risks of industrial control SCADA systems are analyzed,and common forms of industrial control system attacks are discussed.Then,the security vulnerability of the wind farm SCADA system is pointed out from the three aspects of network,communication protocol and host.Finally,we summarized the assessment and defense methods by analyzing he functions of the wind farm SCADA information security.(2)Quantitatively evaluate the security vulnerability of wind farm SCADA system based on evidence reasoning method.Our paper constructed an interpretable hypothesis chain of wind farm SCADA system security vulnerability based on evidence reasoning.And we build a fragile Bayesian network for each path in the explanatory hypothesis chain.Based on subjective Bayes to quantify the relevance and credibility of the evidence points,we use the MCMC method to obtain the posterior distribution of the overall vulnerability of the system through the Bayesian network.(3)Realization of information security cognitive detection agent for wind farm SCADA system.By using artificial intelligence diagnostic technology to analyze the characteristics of network attacks,feasible hypotheses are found through inductive reasoning,evidence is found through deductive reasoning,and the reliability of the hypothesis is evaluated through inductive reasoning.We have compared and analyzed five probabilistic assessment methods of evidence fusion.The wind farm SCADA system network intrusion detection cognitive agent is designed through seven processes of agent problem design,ontology design and development,problem decomposition and plan synthesis,evidence collection and evidence reasoning,rule learning and Rule refinement,generation of problem solutions,and similar problem solving.(4)Wind farm SCADA system network security situation assessment.By analyzing the factors that affect network security,the network security situation assessment model is established based on the confidence rule base(BRB).The evidence reasoning algorithm is used to fuse the situation elements,and an example is used to analyze the local network security situation assessment based on the confidence rule.The results prove that BRB can effectively process semi-quantitative information.It can integrate various uncertain factors to reflect the security status of the wind farm SCADA network system.