Articles
Publication Date: 2026
Engineering Applications of Artificial Intelligence (09521976)168
As the global transition toward intelligent, resilient energy infrastructures accelerates, integrating artificial intelligence (AI) into power system protection has become a critical enabler of operational efficiency and reliability. A major challenge is the accurate identification and classification of high impedance faults (HIFs) in energy distribution grids (EDGs), where low current levels often cause conventional protection devices to malfunction. Traditional schemes require extensive training data, which is often difficult or impossible to obtain. This study presents a novel protection scheme that combines a pre-trained compact convolutional neural network (SqueezeNet) with Wigner-Ville distribution (WVD) and S-transform-based feature fusion to detect and classify HIFs using minimal data. Leveraging transfer learning, the approach reduces model retraining needs and accelerates deployment. Simulation results on modified IEEE 13-bus and 34-bus EDGs show F 1-scores exceeding 97 %, successful cross-network knowledge transfer without retraining, and rapid detection within 20 msec using only 250 training samples, highlighting its suitability for lightweight, scalable, and real-time smart-grid protection. © 2026 Elsevier Ltd.
Mehdipour birgani, A.,
Shams, M.,
Jannati, M.,
Hatami aloghareh, F. Publication Date: 2025
Engineering Applications of Artificial Intelligence (09521976)139
As the primary protection method for transmission lines, distance relays are prone to malfunction during power swings. In fact, the inability of distance relays to differentiate between power swings and short-circuit faults imposes a significant risk to power system stability that can result in blackouts. In recent years, there has been increasing interest in leveraging machine learning techniques to identify various types of faults and power swings in electrical systems. However, previous works mainly focus on fault classification, which is mostly done after a long period from the moment of fault initiation. This is the reason for requiring extensive post-fault data for diagnosis. To address this challenge, this study proposes a predictive protection strategy utilizing deep learning methodologies, specifically a sequence-to-sequence model, to monitor electrical power systems continuously. The objective is to effectively detect power swings from short-circuit faults with minimal reliance on post-fault data and accurately identify short-circuit faults during power swings. In the proposed approach, features are extracted from grid current signals using the Hilbert transform and empirical mode decomposition algorithms. These features are then fed into the sequence-to-sequence model, which issues block/unblock commands upon confirming the presence of a power swing or fault during the power swing. Results from various simulations conducted on an IEEE 39-bus grid in DIgSILENT and MATLAB environments demonstrate that the proposed scheme outperforms baseline methods in the detection of short-circuit faults, power swings, and short-circuit faults occurring during the power swings. The timely and correct operation of the proposed protection scheme contributes to the stability of transmission lines and power systems. © 2024 Elsevier Ltd
