Optimisation of MPPT techniques: conventional, AI-based, and hybrid approaches in photovoltaic systems

Abstract

Maximum power point tracking (MPPT) techniques are crucial for PV systems, which are sensitive to irradiation and temperature variations. This research introduces a high-performance hybrid approach that boosts efficiency by combining the perturb and observe (P&O) technique with artificial neural networks (ANN) to enhance PV power extraction. The hybrid ANN-P&O algorithm dynamically adjusts the step size based on real-time solar irradiance conditions, making the tracking process more adaptive and efficient. To evaluate this approach, a PV system is simulated under varying weather conditions. The hybrid method's performance is compared to standard P&O and ANN techniques using MATLAB/Simulink simulations. Results highlight the enhanced response time of the hybrid ANN-P&O method (0.16 s), demonstrating a faster and more stable tracking process compared to ANN (0.18 s) and P&O (0.24 s). Additionally, the hybrid ANN-P&O achieves 98% efficiency, outperforming ANN (96%) and P&O (95%). These findings confirm the superiority of hybrid strategies in optimising energy output, ensuring higher power conversion and greater system stability for photovoltaic applications.