Abstract: Network attacks could enter into the cyber-physical power system through the network security holes of the information layers and pose a threat to the power system operation. So it is necessary to evaluate the vulnerability of the cyber-physical power system. This paper proposed a comprehensive vulnerability assessment method of the cyber-physical power system based on the integration of Bayesian attack graphs and relay protection mechanism. This method could assign values to the risk probability of the distribution substation nodes by using the Bayesian network, vulnerability scoring system, and junction tree algorithm in the information layers. The AC cascading failure model was used to analyze the busbar fault in the physical layer and triggered the relay protection mechanism. It calculated the proportion of load removed by the relay protection mechanism to eliminate cascading faults. Combining the risk probability of information layers with the failure of the physical layer, it proposed a new load reduction probability of the vulnerability evaluation index to find the vulnerable node and evaluate the vulnerability in the cyber-physical system. The effectiveness of this method is verified by the analysis based on IEEE 14.

    Key words: cyber-physical system; vulnerability; attack graph; network attack; cascading failure; power flow calculation