Multi-fidelity Optimization of Horizontal Axis Wind Turbines
Michael Kenneth McWilliam, Frederik Zahle, Christian Pavese, José Pedro Albergaria Amaral Blasques · 2017
This paper is concerned with the numerical design optimization of wind turbines. Many examples of wind turbine design optimization in literature rely on simplified analysis in some form. This may lead to sub-optimal design, because the optimizer does not see the full fidelity of the problem. To overcome these challenges, this research will explore the multi-fidelity Approximation and Model Management Framework (AMMF) optimization algorithm. AMMF is similar to conventional gradient based optimization, except in the search phase of the optimization, the analysis is replaced with a fast low-fidelity model that has been corrected to give C1 consistency with the high-fidelity model. AMMF was first explored with a simple preliminary investigation based on analytic equations. Second, it was applied to a single-discipline design optimization problem to find the internal structure with the least weight. Finally, AMMF was used in full aero-elastic wind turbine rotor design optimization problem based on the DTU 10 MW reference wind turbine design. Mixed results were achieved for the final study and further work is needed to find the best configuration for AMMF.