Data-driven Evaluation Method for Cyber-Physical System Reliability of Integrated Energy System

Abstract

Background:: In order to achieve the social goal of low carbon and high efficiency, integrated energy system (IES) with deep integration of cyber-physical systems (CPS) are being developed rapidly. The reliability evaluation is an important basis for IES planning and operation. Considering the deep coupling relationship between various systems, energy networks, and cyber networks, and the impact of multiple uncertainties on the integrated energy system, the traditional single index is difficult to quantify the reliability of the system, and the current evaluation methods rely on mechanism analysis and accurate modeling. It is difficult to deal with the coupling relationship between integrated energy systems, energy networks and information networks. So, the quantitative evaluation of the reliability level index of integrated energy systems under the cyber physical system is established, and a data-driven evaluation method is proposed. Methods:: In this article, the energy entropy and information entropy theory are combined to establish the reliability evaluation index of integrated energy system integration of cyber-physical systems, and a data-driven evaluation method based on random matrix theory (RMT) is proposed. Results:: A reliability evaluation model is established through RMT, and the single-ring law is used to analyze the statistical properties of the matrix model to identify the abnormal state of the system. Then the reliability of cyber and physical layers are evaluated based on energy entropy and information entropy, respectively. In addition, sensitivity analyses on the impact of node interruption probability, network topology, access load rate, and other factors are carried out. Conclusion:: The examples of IES are given to verify the feasibility and effectiveness of the proposed method.