Abstract
There has been growing interest in single-stage grid-connected photovoltaic (PV) systems due to their reduced losses and overall size, as they eliminate the intermediate DC-DC conversion stage. The primary objective of this research is to develop a more efficient and industry-oriented control strategy for these systems. This study proposes an artificial neural network (ANN)-based controller to address the high complexity and computational demands of traditional model predictive control (MPC) methods. The ANN-controller simplifies the process by utilizing basic linear equations, significantly reducing the computational burden. Additionally, it integrates an improved maximum power point tracking (MPPT) algorithm to ensure optimal power extraction from the PV panels while maintaining excellent transient performance. The methodology involves validating the superior control performance of the proposed ANN-based strategy in a simulation environment. This includes comparisons with a benchmark grid-tied PV system managed by three different controllers, demonstrating the robustness of the ANN-controller under realistic irradiation-temperature patterns. Results show that the ANN-controller achieves faster response and improved performance compared to traditional MPC methods. Finally, the effectiveness of the ANN-based control logic is experimentally validated using a Control Hardware-In-the-Loop (C-HIL) setup, proving its practicality and reliability in real-world applications.