CFD-based curved tip shape design for wind turbine blades

Abstract

Abstract. This work presents a high-fidelity shape optimization framework based on computational fluid dynamics (CFD). The presented work is the first comprehensive curved tip shape study of a wind turbine rotor to date using a direct CFD-based approach. Preceding the study is a thorough literature survey particularly focused on wind turbine blade tips in order to place the present work in its context. Then follows a comprehensive analysis to quantify mesh dependency and to present needed mesh modifications ensuring a deep convergence of the flow field at each design iteration. The presented modifications allow the framework to produce up to six-digit-accurate finite difference gradients which are verified using the machine-accurate Complex-Step method. The accurate gradients result in a tightly converged design optimization problem in which the studied problem is to maximize power using 12 design variables while satisfying constraints on geometry, as well as on the bending moment at 90 % blade length. The optimized shape has about 1 % r/R blade extension, 2 % r/R flapwise displacement, and slightly below 2 % r/R edgewise displacement resulting in a 1.12 % increase in power. Importantly, the inboard part of the tip is de-loaded using twist and chord design variables as the blade is extended, ensuring that the baseline steady-state loads are not exceeded. For both analysis and optimization an industrial-scale mesh resolution of above 14×106 cells is used, which underlines the maturity of the framework.

Multi-fidelity, steady-state aeroelastic modelling of a 22-megawatt wind turbine

“…The tightly integrated multi-fidelity modelling framework provides very interesting perspectives related to design optimization and exploration of advanced blade shapes. For example, in Zahle et al [26,3], advanced tips were designed using both surrogate-based and direct optimization techniques. Extension of these works to consider the aeroelastic response of the blade can now be achieved and also greatly accelerated using well-established multifidelity optimization techniques, where the medium-fidelity BEVC-NW model can be used.…”

“…It has interfaces to the cross-sectional finite element code BECAS [24], HAWC2 for the structural modelling, HAWCStab2 [25] for the aeroelastic stability analysis, the multi-fidelity engineering rotor aerodynamics code BEVC, as well as to the CFD code EllipSys3D. Aerostructural optimization workflows currently primarily use low-to medium-fidelity aerodynamic modelling, but the CFD code EllipSys3D has previously been used to optimize blade tips using both surrogate-based and direct gradient-based methods [26,3]. The framework uses the WindIO wind turbine ontology developed in the IEA Wind Task 37 [27] to represent the turbine inputs.…”

“…This is because the influence of the blade curved geometry on the wake geometry and consequently on the inductions and the loads are not modelled. This provides challenges when using the BEM method for design optimization and load calculation of blades with large deflections and also novel tip designs with curved shapes [3,4]. There have been many recent developments on the corrections and extensions to further extend the BEM method to model blade sweep effects [5,6,7,8] and blade dihedral [2,9] on the loads.…”

Wind Turbine Technology Trends

“…Applying computational fluid dynamics, Madsen et al [8] tried to optimize the shape of the blade tip. The performance prediction of aerodynamic design of a blade using Deep Learning method was presented in the work of Du et al [9].…”

Fluid–Structure Interaction Simulations of Wind Turbine Blades with Pointed Tips

“…Over the last few years, the industry has made a great effort and achieved a considerable reduction in the levelized cost of energy by increasing the size or optimizing the blade geometries of wind turbines. For instance, O'Brien et al [1] reviewed horizontal axis wind turbine research by focusing on both numerical modeling and experimental practices; Madsen et al [2] presented a curved tip shape design to a 10 MW reference wind turbine using a high-fidelity optimization approach based on computational fluid dynamics (CFD); and Posta et al [3] presented an aeroelastic model to investigate the FSI for a reference 5 MW turbine, where both one-way and two-way coupled simulations were performed. With relatively fewer constraints in offshore applications, the trend of increasing the size of large wind turbines is expected to continue within this decade.…”