Affiliations: [a] State Grid Hebei Electric Power Research Institute, Shijiazhuang, Hebei, China
| [b] State Grid Hebei Electric Power Co. Ltd., Shijiazhuang, Hebei, China
Correspondence:
[*]
Corresponding author. Ze Chen, State Grid Hebei Electric Power Research Institute, Shijiazhuang, Hebei, 050021, China. E-mail: chengyejingdy1@163.com.
Abstract: In order to improve the effect and accuracy of risk source identification, this paper studied the network security risk source identification model of power CPS system based on fuzzy artificial neural network. The network security risk source index system of power CPS system was constructed, and the dimension of index data was reduced by principal component analysis. Fuzzy theory is used to process the index data after dimension reduction, and the comprehensive membership vector of each index is obtained. The dynamic clustering algorithm is used to determine the number of hidden layer units of radial basis function neural network, and the network security risk source identification model is established. Finally, the quantitative value of risk source identification is output. The experimental results show that the model can effectively reduce the dimension of the network security risk source index data of the power CPS system. The optimal distance threshold of the hidden layer is 4.2, and the optimal number of units is 6. In the final identification results, four severe risk sources and five moderate risk sources were obtained, and the quantitative values of risk source identification of each index were 63, 70, 71, 77, 65, 89 and 96, respectively, indicating that the model can effectively identify network security risk sources of power CPS systems. With the increase of the proportion of communication nodes removed, when there are various types of security vulnerability information, the mean square error value of the model is always lower than the set threshold, indicating that the model has high recognition accuracy.
Keywords: Fuzzy theory, artificial neural network, power CPS system, network security, risk sources, identification model
