Radial Basis Function Kriging Model for Evaluating Cyber–Physical Grid Reliability Under Cyber Network Failures

ABSTRACT The impact of cyber network failures on the reliability of cyber–physical grids is profound and overlooking them leads to an overly optimistic assessment of reliability. Given the interdependencies between cyber and physical elements, evaluating the reliability of such systems is inherently complex. This paper addresses this challenge by employing the improved radial basis function Kriging method (RBF_KM) to assess the reliability of cyber–physical grids. Additionally, the study incorporates a communication network based on IEC‐61850 to accurately model switches, enhancing the fidelity of the assessment. The assessment metrics, loss of load expectation (LOLE) and expected energy not served (EENS), are calculated across various scenarios and configurations to evaluate the efficacy of the proposed method and compare it against the Monte Carlo simulation (MCS) approach. Additionally, the proposed model takes into account the impact of using the dynamic thermal rating (DTR) system, which allows for the safe enhancement of system reliability. Results indicate that IRBF_KM outperforms MCS and Polynomial_basis Kriging method in capturing the interdependency between elements and nonlinearity of the problem achieving 98.41% and 98.67% accuracy for LOLE and EENS, respectively, compared to MCS's 83.14% and 82.79%. Furthermore, the study reveals that employing redundant configurations for cyber networks cannot fully offset the adverse effects of cyber failures, with EENS increasing by 100.03% despite a 12.4% decrease observed with the use of redundant switches. Moreover, the reliability indices show significant improvement when considering the DTR system instead of the static thermal rating (STR) system.