A Cyber-Physical Digital Twin Framework for State Estimation of Dendrite-Risk Prediction and Resilient Control in Solid-State Batteries.

Solid-state batteries (SSBs) offer high energy density and improved safety but remain vulnerable to hidden electro-chemo-thermo-mechanical degradation, lithium dendrite formation, and cyber-physical attacks that cannot be reliably detected using conventional battery-management systems. This work presents a cyber-physical digital twin framework for real-time state estimation, dendrite-risk prediction, and resilient control of SSBs. A physics-regularized reduced-order model integrated with Moving Horizon Estimation reconstructs unmeasurable internal concentration, potential, temperature, stress, and interfacial degradation states from limited terminal measurements. A physics-informed Dendrite-Risk Forecast Index (DRFI) is developed by combining stress evolution, current-density variance, interfacial impedance growth, and thermal-gradient severity to provide early degradation warning. Physics-consistent anomaly detection identifies measurement manipulation and cyber-attacks, while a DRFI-aware Model Predictive Controller adaptively regulates battery operation to mitigate degradation. Simulation studies under normal operation, accelerated degradation, and cyber-attack scenarios demonstrate state-estimation errors of 5%-7% during normal cycling and less than 10% under accelerated ageing, 0.8-1.2 cycles of early degradation prediction, 40%-55% reduction in stress concentration, and attack detection within 80-110 ms. These results demonstrate that the proposed physics-guided cyber-physical digital twin provides an interpretable and computationally efficient framework for predictive battery management, degradation forecasting, and resilient operation of next-generation solid-state batteries.