关键词: 城市电网信息物理系统, 安全态势预警, 风险跨空间传播, 分布式电源, 元胞自动机, 核极限学习机

Abstract: In the context of the deep coupling of information and communication technology and power technology, timely early warning of the security situation of urban power grids is the key to ensure their safe and stable operation. In order to solve the problem of cyber-physical cross-space risk propagation faced by the cyber-physical system (CPS) of urban power grid in the digital transformation, this paper proposes a security situation early warning method based on Kernel Extreme Learning Machine (KELM). Firstly, combined with the cellular automata theory, the coupling model of the physical layer and the information layer of the power grid CPS was constructed to reveal the cross-space propagation mechanism of faults, and the key risk characteristics were extracted based on multi-source data such as SCADA and PMU. Secondly, the integrated KELM early warning model was designed, the deep integration of multi-dimensional data was realized through kernel function mapping, and the early warning index system covering voltage overrun, line overload and load loss was constructed simultaneously, and the entropy weight method was used to dynamically allocate the index weights and divide the security situation level. Finally, simulation experiments based on IEEE 33-node distribution network show that the proposed method has a great improvement in the accuracy of voltage fluctuation and line overload prediction compared with the traditional ELM method in the distributed generation access scenario, which verifies the efficiency and robustness of the model.

Key words: cyber-physical systems of urban power grids, security situation early warning, risk spread across spaces, distributed power generation, cellular automata, Kernel Extreme Learning Machine