Research On Risk Assessment For Industrial Cyber Physical Systems Based On Model-Data Integrated Method

With the deep integration of industrialization and informatization and the introduction of new technologies such as big data and cloud computing,the industrial field has undergone profound changes and gradually developed into Industrial Cyber Physical Systems(ICPS).However,the increasing openness of the system makes cyber security threats more serious.Once the cyber attack breaks through the security defense and penetrates into the ICPS control network,it will lead to a decline in the system's output or even downtime,which will cause safety accidents and endanger personnel safety.Therefore,it is of great significance to implement safety protection for ICPS to ensure the safe and stable operation of the system.Identifying and evaluating the risk of ICPS is a key part of system security protection.It can assist security operators to grasp the current security situation of the system and take appropriate response measures in a timely manner.For this purpose,a model-data integrated risk assessment method is presented in this paper to quantify the impact of cyber-attacks on the physical system of ICPSs.The method uses a Bayesian network to model the risk propagation process and infers the probabilities of sensors and actuators to be compromised.Then interdependencies among sensor readings and controller signals,through which risk propagates in process control system and thus causes impact on these controller parameters,are studied.Based on probabilities of being compromised and association analysis,the impact on a group of controller parameters that are critical for process system can be calculated and the system risk can finally be quantified.It is noted that datasets of various sources are utilized to drive the above risk assessment activity at almost every stage,from initial model setup by analyzing the system's weakness to the model's subsequent updating through machine learning,and when evaluating the impact on physical system of the attack,correlation analysis of the operation data from the target system is also conducted.To verify the effectiveness of the proposed approach,a case study is conducted on a coupling tanks control system to evaluate its timeliness,validity and adaptability.The simulation results demonstrate that the proposed method can accurately evaluate the security situation of the system and meet the requirements of timeliness for security decision.In addition,the data and model driven manner makes the method have better scalability and adaptability.