Future power systems are evolving into large-scale, highly complex and heterogeneous cyber-physical systems, integrating increasing numbers of renewable generators, power-electronic converters, energy storage systems, electric vehicles, flexible loads, data centers, and distributed energy resources. Their operation involves interactions across multiple spatial and temporal scales, from fast device-level dynamics to network-level coordination and energy management. The resulting high-dimensional state space, large number of operating conditions, uncertainties, contingencies, and controllable devices create computationally demanding problems in modelling, simulation, optimization, and decision-making. This project develops AI-enabled digital twin and intelligent operation methods for future power systems. The objective is to combine physics-based modelling, data-driven methods, and advanced computing to improve system monitoring, stability assessment, resilience analysis, and operational decision-making across large and complex networked energy systems. The research will investigate AI-enabled digital twins for modelling, monitoring, prediction, and analysis of power systems under a wide range of operating conditions and network configurations. Reinforcement learning and safe learning methods will also be developed for coordinated operation of energy storage systems, microgrids, energy communities, and other flexible energy resources while respecting network and system-level constraints. The project will address multiple complementary research problems spanning component, subsystem, and network levels. The computational studies will involve large-scale simulation and data generation, machine-learning model training, optimization, hyperparameter studies, Monte Carlo analysis, and extensive operating-condition and contingency sweeps. Multiple computational studies will be conducted in parallel and evaluated on systems with increasing numbers of components, operating scenarios, and controllable devices to assess scalability, robustness, and generalizability. The expected outcomes are scalable AI-enabled methodologies and computational tools for efficient analysis and intelligent operation of large-scale future power systems, supporting secure integration of renewable generation, power-electronic resources, and distributed flexibility.