Multi-objective optimal design of standalone hybrid renewable energy system using evolutionary algorithms

The most popular options for remote applications where a grid connection is not available are wind turbine (WT) generators and solar photovoltaic (PV) power systems. Selecting a desirable design of the WT/PV/Battery system among a wide range of configurations, particularly at a favorable level of reliability, lowering the total cost, and reducing surplus energy, remains a challenging task. Therefore, this study seeks to first explore the acquired approaches to solve the nonlinearity of the three-diode PV model’s equation utilizing actual measured laboratory data collected at a variety of environmental conditions. Then, the optimum design of the hybrid standalone WT/PV/Battery system is chosen while considering three conflicting techno-economic criteria for power-isolated dwellings in Malaysia and South Africa. The research presented in this thesis is divided into two phases, namely, modeling of the PV module and optimal sizing of the entire system to obtain reliability, cost-effectiveness, and reduce surplus energy for the WT/PV/Battery system. The robust adaptive arithmetic optimization algorithm based on the adaptive damping Berndt-Hall-Hall-Hausman (RAOAAdBHHH) approach is used to efficiently determine the parameters of the three-diode PV model using two types of experimental data of the PV module technologies. In optimal sizing of the hybrid standalone WT/PV/Battery system, a unique improved two-archive approach was presented to develop diversity and convergence independently to find four optimum sets of Pareto front (PF) solutions. Additionally, a new integration based on the best-worst method (BWM) and preference ranking organization for enrichment evaluations (PROMETHEE II) method, along with a group decision-making (GDM) mechanism, was addressed to sort and rank the optimal sets of PF solutions. The objective functions were loss of load probability (LLP), life cycle cost (LCC), and dump power (Pdump) to analyze and evaluate the performance of the selected optimum design using hourly meteorological data for one year in Malaysia and South Africa. Finally, the proposed improved two-archive algorithm is verified using a hybrid numerical method. The experimental results for the parameter extraction optimization problem demonstrate that the proposed RAOAAdBHHH approach successfully minimizes error to zero with rapid convergence, as determined by different statistical criteria and compared to experimental data of the two PV technologies. On the other hand, the theoretical results based on actual hourly meteorological data in Malaysia and South Africa indicated that the proposed improved two-archive-BWM-PROMETHEE II-GDM method is not only able to construct a uniform set of Pareto optimal solutions with fast convergence and high diversity but can also rank and select the most desired design for the WT/PV/Battery system. The optimum configurations of the two case studies are composed of a single WT, 105 PV modules, 69 battery storages at zero LLP, 60185 ($) of LCC, and 13548018 (KWh) of Pdump for the Malaysia case study. Meanwhile, it consists of 16 WTs, 80 PV modules, and 69 battery storages at zero LLP, 66073 ($) of LCC, and 31371 (KWh) of Pdump for the South Africa case study. Furthermore, the proposed improved two-archive-BWM-PROMETHEE II-GDM method is verified by utilizing a hybrid numerical method, which shows a high level of reliability, lowering overall cost, and reducing wasted energy. Therefore, the employment of WT in the Malaysia scenario is not recommended, while the WTs may considerably incorporate in supplying energy for the required load demand.

Text FK 2024 26 - Declaration Form.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (720kB) Text FK 2024 26 - Full Text.pdf Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (7MB) Text FK 2024 26.pdf Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (7MB) Official URL or Download Paper: http://ethesis.upm.edu.my/id/eprint/18991

Actions (login required)

View Item