Renewable Energy Resources and Technologies
Zaiba Ishrat; Ankur Kumar Gupta; Seema Nayak
Abstract
The rapid rise in electrical energy demand and the depletion of fossil fuels have created a market for renewable energy. Among all the renewable energy resources, the most popular is solar energy, perceived as pollution-free, easily accessible, and low maintenance. In non-uniform solar irradiation or ...
Read More
The rapid rise in electrical energy demand and the depletion of fossil fuels have created a market for renewable energy. Among all the renewable energy resources, the most popular is solar energy, perceived as pollution-free, easily accessible, and low maintenance. In non-uniform solar irradiation or partial shading conditions (PSC), the photovoltaic characteristics (PVC) of a solar panel system (SPS) exhibit multiple minor peaks (MP) with one global peak power point (GPPP). To extract the utmost energy from the SPS, the authors proposed an efficient hybrid algorithm integrating the advantages of machine learning and the classical algorithm fractional open circuit voltage (FOVA) to track the GPPP. To follow the GPPP of SPS under unstable environmental surroundings, this study tests ML-based hybrid MPPT algorithms, specifically squared multiple variable linear regression algorithms (SMVLRA), using Matlab/Simulink. Simulation through Matlab is employed to validate the efficiency of the SMVLRA-MPPT approach compared to existing popular conventional and modern MPPT algorithms, namely the Perturb and Observation algorithm (P&OA), the variable step size incremental conductance (VINC) algorithm, and an intelligent algorithm, Decision Tree Regression Algorithm (DTRA). The simulation results demonstrate that SMVLRA offers higher peak power and mean peak power efficiency in less tracking time, with lower error and almost negligible steady-state fluctuation under PSC. The proposed algorithm achieves 99.99% efficiency under standard test conditions (1000w/m2, 25°C), 99.95% under PSC1 (1000w/m2, 800w/m2, 25°C), and 98.89% under PSC2 (1000w/m2, 800w/m2, 600w/m2, 25°C)
Renewable Energy Resources and Technologies
Zaiba Ishrat; Ankur Kumar Gupta; Seema Nayak
Abstract
Solar power energy continues to be a renewable and sustainable source of energy in the coming year due to its cleaner nature and abundant availability. Maximum Power Point Tracking (MPPT) is a technique used in solar power systems to extract maximum power from photovoltaic (PV) modules by tracking the ...
Read More
Solar power energy continues to be a renewable and sustainable source of energy in the coming year due to its cleaner nature and abundant availability. Maximum Power Point Tracking (MPPT) is a technique used in solar power systems to extract maximum power from photovoltaic (PV) modules by tracking the operating point of the modules. MPPT is essential for achieving optimal power output from a solar panel, particularly in variable weather conditions. Traditional MPPT techniques are subject to limitations in handling the partial shading conditions (PSC). To ensure the tracking of maximum power point while boosting the MPPT's overall efficacy and performance, Machine Learning must be integrated into MPPT. As per the reviewer work, ML techniques have the potential to play a crucial role in the development of advanced MPPT systems for solar power systems operating under partial shading conditions and to compare the performance of existing ML-MPPT in terms of accuracy, response time, and efficacy. These review papers technically analyze the result of ML-MPPT techniques and suggest the optimum ML-MPPT tactics that are Q learning, Bayesian Regularization Neural Network (BRNN), and Multivariate Linear Regression Model (MLIR) to achieve optimum outcomes in MPPT under PSC. Further, these techniques can offer efficiency greater than 95%, tracking duration less than 1sec, and error threshold of 0.05. In the future, the reviewer may propose simulation work to compare the optimal algorithms.
