A Comparative Performance Analysis of P&O and ANN Algorithm Based MPPT Energy Harvesting in PV Systems

Solar energy is a popular renewable energy source that is integrated into the electricity grid due to fossil fuel depletion and environmental concerns. Solar photovoltaic (PV) systems convert sunlight into electricity and offer economic and environmental advantages. These systems consist of solar panels, inverters, and optional battery storage. The direct current (DC) electricity generated by solar panels is converted into alternating current (AC) by inverters for use in electronic devices and appliances. Battery storage is recommended for storing excess electricity. To maximize energy generation, tracking the maximum power point (MPP) of the PV system is crucial. Various maximum power point tracking (MPPT) algorithms, including artificial neural networks (ANN), have been developed to optimize energy extraction. ANN-based MPPT methods respond quickly and accurately to rapid changes in solar irradiation. Simulations comparing ANN-based MPPT with traditional perturbation and observation (P&O) algorithms demonstrate the superior performance of ANN-based methods in tracking MPP and minimizing power oscillation. MPPT energy harvesting systems can be designed and simulated using MATLAB or Simulink, and ANN-based Levenberg-Marquardt algorithms have shown promising results. Training the model with datasets of sun irradiance, temperature, and voltages can lead to accurate results with minimal error. The robustness of the algorithm has been proved by the effectiveness of overall data processing for MPPT energy optimization through the utilization of the Levenberg-Marquardt technique, trained on a dataset consisting of 1000 instances. The ANN-based Levenberg-Marquardt technique offers swift and precise response to solar irradiation changes while minimizing power oscillation under varying conditions.