The integration of renewable energy sources (RESs) into power grids presents significant challenges to system stability, primarily due to the reduced inertia typically supplied by synchronous generators (SG). This study addresses the urgent need for accurate and real-time inertia estimation methods to ensure reliable grid operation amid evolving dynamic conditions. An advanced algorithm is proposed, which fuses artificial neural networks with the Modified Equal Area Criterion and concepts of kinetic energy. By incorporating the maximum mechanical power as a novel input feature, the methodology enhances the accuracy of inertia estimation. Additionally, a new index for identifying optimal fault locations is introduced, further refining precision. This research potentially revolutionises grid monitoring and control by delivering robust, noise-resistant and computationally efficient real-time inertia estimates. Key applications include real-time frequency management and contingency planning within modern power systems characterised by high RES penetration. Validation of the proposed approach is conducted through extensive simulations on the IEEE 39-bus New England test system, demonstrating consistently low estimation errors (less than 1%) and superior performance compared to traditional methodologies.
