LSTM-Based MPPT Algorithm for Efficient Energy Harvesting of a Solar PV System Under Different Operating Conditions

Solar energy is one of the most favorable renewable energy sources and has undergone significant development in the past few years. This paper investigates a novel concept of harvesting the maximum power of a photovoltaic (PV) system using a long-short term memory (LSTM) to forecast the irradiance value and a feedforward neural network (FNN) to predict the maximum power point (MPP) voltage. This study paves a way to mitigate avoidable inefficiencies that hinder the optimal performance of a PV system, due to the intermittent nature of solar energy. MATLAB/Simulink software platform was used to validate the proposed algorithm with real irradiance data from different geographical and weather conditions. Furthermore, the maximum power point tracking (MPPT) algorithm was implemented in a laboratory setup. The simulation results portray the superiority of the proposed method in terms of tracking performance and dynamic response through a comprehensive case study conducted with five other state-of-the-art MPPT methods selected from conventional, AI based, and bio-inspired MPPT categories. In addition to that, faster response time and lesser oscillations around the MPP were observed, even during volatile weather conditions and partial shading.