A Digital Twin and Diffusion Model Integrated Method for Power System Operational State Generation and Risk Assessment

Abstract

The integration of renewable energy sources and the increasing frequency of extreme weather events have introduced significant uncertainties into power system operation, challenging traditional risk assessment approaches that rely on limited historical data. This study proposes a novel method that integrates digital twin technology with diffusion models for operational state generation and risk assessment in power systems. The digital twin framework constructs a high fidelity virtual replica of the physical power grid, enabling real time simulation of system dynamics under diverse operating conditions. The diffusion model, specifically a conditional denoising diffusion probabilistic model, is employed to generate realistic and diverse operational scenarios, including extreme events such as high renewable variability, load peaks, and cascading failure conditions. These generated scenarios are subsequently fed into the digital twin simulation platform to assess system risks quantitatively using probabilistic metrics. The method is validated on the IEEE 39 bus and IEEE 118 bus test systems using publicly available operational datasets. Experimental results demonstrate that the proposed approach generates high quality scenarios that capture both typical and extreme operational patterns, achieving superior performance in risk identification and decision support compared to conventional generative methods. This study contributes a comprehensive framework that bridges generative artificial intelligence and physical simulation for enhanced power system resilience.