Sci Rep. 2026 Jun 18. doi: 10.1038/s41598-026-57578-z. Online ahead of print.
ABSTRACT
Intermittent renewable energy sources and the increasing use of Electric Vehicles (EVs) cause significant challenges for Smart Grid (SG) stability and energy management. To address these challenges, this paper proposes a hybrid EV charging coordination system. It uses a Single-Switch High Step-Up Zeta converter (S2HSZ) along with Leaf in Wind Optimization (LiWO) and Spiking Deep Residual Network (SDRN) in a unified framework. The proposed method is designed to boost grid stability, optimize converter efficiency, minimize charging errors, and improve energy use in systems with PV and wind integration. In this setup, LiWO fine-tunes the PI controller gains, while SDRN predicts ideal operating parameters for smooth performance in dynamic grid conditions. The proposed framework is implemented in MATLAB, and the LiWO-SDRN model is compared with Adaptive Interaction Artificial Neural Networks (AI-ANN), Artificial Neural Network-Particle Swarm Optimization (ANN-PSO), Multi- Agent Deep Neural Network (MADNN), Gannet Optimization Algorithm-Dilated Residual Convolutional Neural Networks (GOA-DRCNN) models. The LiWO-SDRN model delivers coordinated EV charging and steady converter operation. The experiment shows the LiWO-SDRN framework attained improved results: 99.7% converter efficiency, 98.7% energy use efficiency, and a system error of 1.5%. The LiWO-SDRN framework outperformed other methods in terms of stability, efficiency, and charging coordination. These findings demonstrate that the model significantly boosts smart grid operations, especially with EV charging and renewable energy.
PMID:42315885 | DOI:10.1038/s41598-026-57578-z
