Design-Space Exploration for Inverse-Design of Wind Turbine Blades Using Data-Driven Methods
The state-of-the-practice aerodynamic design methods for wind turbine blades is typically based on Blade-Element Momentum (BEM) theory using a pre-designed frozen family of airfoils. The airfoils are themselves typically designed using panel methods. The design of the next-generation of large flexible rotors will need to capture the non-linear aerodynamics and three-dimensional flow to reduce the levelized cost of wind energy. Data-driven methods for aerodynamic design using data generated by computational fluid dynamics offer an attractive alternative to BEM-based methods that captures the non-linear aerodynamics of the component airfoils as well as the root and tip sections. In this work, we develop and demonstrate a framework to "smartly" explore the relevant design space in combination with an appropriate automated CFD pipeline to evaluate the aerodynamics of each design. The design-space exploration framework uses appropriate perturbations to the airfoil shape and induction profile from a baseline shape in combination with the inverse-design using BEM. The resulting designs are evaluated using an automated CFD pipeline using the in-house CFD solver framework "Mercury". We perform verification and validation to establish the capability of the Mercury framework to predict the aerodynamic performance of wind turbines. The CFD simulation of the perturbed blade shapes are optimized to restart from the converged baseline simulation to reduce the computational time. Finally, we demonstrate the design-space exploration technique for the design of the outboard section and the full rotor using perturbations to the shape and operating conditions of the NREL 5-MW turbine.
- Research Organization:
- National Renewable Energy Laboratory (NREL), Golden, CO (United States)
- Sponsoring Organization:
- USDOE Advanced Research Projects Agency - Energy (ARPA-E)
- DOE Contract Number:
- AC36-08GO28308
- OSTI ID:
- 1845894
- Report Number(s):
- NREL/CP-5000-82142; MainId:82915; UUID:255a7e5a-685b-4b1c-bdfe-b22c9f5e35d4; MainAdminID:63881
- Resource Relation:
- Conference: Presented at the AIAA SCITECH 2022 Forum, 3-7 January 2022
- Country of Publication:
- United States
- Language:
- English
Similar Records
Invertible Neural Networks for Aerodynamic Design of Wind Turbine Blades
Airfoil Computational Fluid Dynamics - 2k shapes, 25 AoA's, 3 Re numbers