Date of Award
Doctor of Philosophy (PhD)
Demand is increasing for a system based on renewable energy sources that can be employed to both fulfill rising electricity needs and mitigate climate change. Solar energy is the most prominent renewable energy option. However, only 30%-40% of the solar irradiance or sunlight intensity is converted into electrical energy by the solar panel system, which is low compared to other sources. This is because the solar power system's output curve for power versus voltage has just one Global Maximum Power Point (GMPP) and several local Maximum Power Points (MPPs). For a long time, substantial research in Artificial Intelligence (AI) has been undertaken to build algorithms that can track the MPP more efficiently to acquire the most output from a Photovoltaic (PV) panel system because traditional Maximum Power Point Tracking (MPPT) techniques such as Incremental Conductance (INC) and Perturb and Observe (P&Q) are unable to track the GMPP under varying weather conditions. Literature (K. Y. Yap et al., 2020) has shown that most AIbased MPPT algorithms have a faster convergence time, reduced steady-state oscillation, and higher efficiency but need a lot of processing and are expensive to implement. However, hybrid MPPT has been shown to have a good performance-to-complexity ratio. It incorporates the benefits of traditional and AI-based MPPT methodologies but choosing
the appropriate hybrid MPPT techniques is still a challenge since each has advantages and disadvantages. In this research work, we proposed a suitable hybrid AI-based MPPT technique that exhibited the right balance between performance and complexity when utilizing AI in MPPT for solar power system optimization. To achieve this, we looked at the basic concept of maximum power point tracking and compared some AI-based MPPT algorithms for GMPP estimation. After evaluating and comparing these approaches, the most practical and effective ones were chosen, modeled, and simulated in MATLAB Simulink to demonstrate the method's correctness and dependability in estimating GMPP under various solar irradiation and PV cell temperature values. The AI-based MPPT techniques evaluated include Particle Swarm Optimization (PSO) trained Adaptive Neural Fuzzy Inference System (ANFIS) and PSO trained Neural Network (NN) MPPT. We compared these methods with Genetic Algorithm (GA)-trained ANFIS method. Simulation results demonstrated that the investigated technique could track the GMPP of the PV system and has a faster convergence time and more excellent stability. Lastly, we investigated the suitability of Buck, Boost, and Buck-Boost converter topologies for hybrid AI-based MPPT in solar energy systems under varying solar irradiance and temperature conditions. The simulation results provided valuable insights into the efficiency and performance of the different converter topologies in solar energy systems employing hybrid AI-based MPPT techniques. The Boost converter was identified as the optimal topology based on the results, surpassing the Buck and Buck-Boost converters in terms of efficiency and performance.
Keywords—Maximum Power Point Tracking (MPPT), Genetic Algorithm, Adaptive Neural-Fuzzy Interference System (ANFIS), Particle Swarm Optimization (PSO)
Prairie View A&M University
© 2021 Prairie View A & M University
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Date of Digitization
John B Coleman Library
City of Publication
Taylor, A. R. (2023). Performance Analysis Of Hybrid Ai-Based Technique For Maximum Power Point Tracking In Solar Energy System Applications. Retrieved from https://digitalcommons.pvamu.edu/pvamu-dissertations/12