Reliability Modeling for Virtualized PLC Architectures with Redundant Supervision and Stealth Attack Mitigation

The virtualization of industrial systems has enabled the consolidation of multiple PLCs into centralized environments, improving scalability and efficiency. However, this also increases exposure to simultaneous failures and stealth cyberattacks. This paper presents a hybrid reliability model combining physical failures - represented by Weibull and exponential distributions - with cyberattack dynamics and imperfect detection by a central supervisor. System logic is structured using Minimal Cut Sets (MCCs) to identify critical failure combinations. Monte Carlo simulations are used to estimate system reliability as a function of physical and cybernetic factors. Sensitivity, nonlinearity, and curvature analyses reveal how reliability varies with detection effectiveness and attack frequency. The findings highlight the existence of critical regions where reliability sharply declines, underscoring the need for robust detection strategies in virtualized industrial architectures.