Supervisorer: Mathias Stolpe, DTU Vindenergi - Lars Risager, Suzlon Blade Science Center
Ekstern eksaminator: Professor Erik Lund, Aalborg University.
Titel: Development of framework for conceptual optimization of wind turbine blades and basic wind turbine configuration
Recent advances in wind turbine technology are pushing down the cost of electricity from wind power. To compete in this market, there is a need for reliable and robust optimization to streamline and enhance the wind turbine design process. This thesis presents a framework for wind turbine optimization in cooperation with the Suzlon Blade Science Center to meet this demand. The Matlab optimization package minimizes the levelized cost of energy of a wind turbine via optimization of design variables related to the blade geometry and basic wind turbine parameters. The package consists of a blade geometry module created by the Master’s student coupled to blade element momentum, structural, and levelized cost of energy modules provided by the Suzlon Blade Science Center. The bundle of modules define an objective function for which Response Surface Methodology and pattern search optimization techniques are applied to find the blade design and turbine configuration which minimize the levelized cost of energy. Response Surface Methodology under-performs when the design space is large, however reductions in the design space lead to consistent and desirable results. Pattern search performs well with a large design space and displays reasonable computational time, though the scaling of computational cost with problem size is unpredictable. Each method is evaluated and pattern search is chosen as the preferred method for future optimization. Suggestions for maximizing the utility of Response Surface Methodology and improving the performance of pattern search are provided, with an aim towards larger optimization problems and global optimization.
Significant improvements in the levelized cost of energy of the optimized designs bolster the methodologies employed in this package and suggest a promising avenue for future wind turbine optimization.