Logo image
Optimization of the capacity auxiliary winding in a self-excited three-phase synchronous reluctance generator on stochastic-based methods
Thesis   Open access

Optimization of the capacity auxiliary winding in a self-excited three-phase synchronous reluctance generator on stochastic-based methods

Tapiwanashe Kudoma
M.Eng., University of Johannesburg
2025
Handle:
https://hdl.handle.net/10210/520733

Abstract

The global shift to sustainable energy has heightened interest in rare-earth-free generator technologies, especially for decentralized wind energy in remote or off-grid settings, where efficiency losses hinder reliable power delivery. A key research gap remains in systematically optimizing the capacitive auxiliary winding of a self-excited three-phase synchronous reluctance generator (SynRG) to improve excitation stability and minimize copper losses through advanced stochastic approaches. This study addresses this gap by optimizing the capacitive auxiliary winding of a self-excited three-phase SynRG using advanced stochastic methods. The primary objective was to apply and validate an integrated optimization framework that leveraged Particle Swarm Optimization (PSO) and the Bacterial Foraging Optimization Algorithm (BFOA) to identify auxiliary winding parameters that maximize efficiency. The methodology combined analytical machine modeling, high-fidelity finite element analysis (FEA), and stochastic optimization. Design variables included the number of windings turns, conductor cross-sectional area, and capacitive values, with constraints imposed by torque and manufacturability criteria. A comparative analysis benchmarked PSO and BFOA in terms of convergence rate and robustness, and all models are validated against FEA and published data. Results demonstrated that the optimized auxiliary winding significantly reduced copper losses and improves generator efficiency (up to 78% in FEA), with both PSO and BFOA yielding robust, high-performance designs featuring stable output voltage and satisfactory torque characteristics. PSO, with an efficiency of 82%,which converged rapidly, while BFOA, with an efficiency of 84%, explored the design space more thoroughly at the cost of increased computation time. These findings establish actionable, efficiency-driven design guidelines for rare-earth-free SynRGs, highlighting substantial improvements in excitation stability, efficiency, and environmental sustainability. The research advances both theoretical understanding and practical design, supporting affordable, climate-resilient electrification and aligning with the goals of sustainable development and climate action.
pdf
T_KUDOMA_2210008535.81 MBDownloadView
Open Access

Metrics

1 Record Views

Details

Logo image