A GA-Based State Estimation for Active Distribution Networks in a Digital Twin Framework

This paper presents a Genetic Algorithm (GA)based State Estimation (SE) approach within a Digital Twin (DT) framework to enhance real-time monitoring of Active Distribution Networks (ADNs). Applied to the ASM Terni grid, the method leverages real-time Power Quality Analyzer (PQA) data for near real-time SE every 6.9 minutes, totaling 21,000 operations over three months. Nine different state estimation scenarios were analyzed, varying the GA parameters to assess their impact on estimation accuracy. The scenarios were evaluated using Pearson, Spearman, and Kendall correlation coefficients, along with Dynamic Time Warping (DTW) distance, to determine the most reliable configuration. The results indicate that Scenario 2, which consists of five candidate solutions per population and 100 generations, demonstrates the highest alignment between estimated and actual power values, achieving an overall score of 83.4%. The findings confirm the robustness of the GA-based SE approach in real-time grid monitoring, efficient energy management, and flexibility exploitation in modern distribution networks.