Research On Risk Assessment Of Industrial Control System Based On Belief Rule Base

Industrial control system is an industrial manufacturing or processing system with automatic control ability,which has been widely used in various modern industries and become an important cornerstone of national security and economic development.With the deepening of the policy of "integration of industrialization and industrialization",the safety of industrial control system has a profound impact on the development of industrial control system and related industries,and has become one of the research focuses.Among them,intrusion detection and risk assessment are effective methods for the safety management of industrial control systems,which can monitor complex industrial control systems.Due to the characteristics of industrial control system,such as wide range of parts,complex system,strong coupling and so on,there has not been a relatively perfect and mature theory and method of industrial control system intrusion detection and risk assessment.Therefore,this research has extremely important theoretical research and practical application value.Aiming at the above problems,the aim is to improve the security and availability of industrial control system and the ability of active defense against intrusion.This paper discusses the current research background and significance of industrial control system,and introduces the commonly used intrusion detection and risk assessment methods in industrial control system.In this paper,the modeling method of industrial control system based on the belief rule base is studied,and the theory research is successfully applied to the intrusion detection and risk assessment of industrial control system.This method effectively combines the quantitative monitoring information and qualitative knowledge of various uncertainties in industrial control system,solves the problem of lack of failure data in industrial control system,improves the accuracy of intrusion detection and risk assessment of industrial control system,ensures the safe and stable operation of industrial control system,and more importantly,provides measures for decision makers to avoid losses Valid basis.The main contents of this paper are as follows:(1)Aiming at the problem of intrusion detection in industrial control system,this paper analyzes the high-risk intrusion behaviors,such as denial of service attack,instruction injection attack,response injection attack,etc.Considering the quantitative monitoring data and expert qualitative knowledge of industrial control system,an intrusion detection model of industrial control system based on belief rule base is proposed.In this model,a linear combination method is first proposed to construct the rules in the belief rule base,which solves the problem of "combination explosion" when the number of premise attributes in the belief rule base is too large.Secondly,we use the evidence reasoning algorithm to combine the belief rules in the belief rule base,and optimize the initial parameters of the belief rule base.Finally,taking a gas industry control system as an example,it is verified that the accuracy of the optimized intrusion detection model based on belief rule base is higher than that of the original one.(2)Aiming at the problem that it is difficult to establish an accurate risk assessment model of industrial control system,a risk assessment model of industrial control system based on belief rule base is proposed by using semi quantitative information.In this method,the qualitative knowledge of industrial control system is combined with quantitative monitoring data by using the expert system of belief rule base.In order to improve the accuracy of risk assessment of industrial control system,the initial parameters of the established belief rule base model are optimized.Finally,taking a gas industry control system as an example,the experimental results show that the evaluation accuracy is higher than SVM and BP neural network models.