How can artificial networks enhance second-order hybrid extended kalman filtering for energy management?

Abstract

Accurate state estimation is vital for effective energy management. This study introduces an integrated approach combining Second-Order Hybrid Extended Kalman Filtering (SO-HEKF) with artificial neural networks to enhance estimation accuracy in dynamic energy environments. Simulation results indicate up to 16.7 % improvement in estimation accuracy and a 12.4 % reduction in operational cost compared to standard SO-HEKF. The adaptive learning mechanism enables real-time adjustments under varying grid conditions. Case studies across renewable integration, load forecasting, and demand response scenarios confirm the method’s effectiveness in improving resource allocation, grid stability, and robustness. This integrated approach supports more reliable and intelligent energy systems.