Advanced Methods for Efficient Aeroelastic Analysis of Wind Turbines and Wind Farms

Wind power plays an important role in satisfying the electricity needs of the US; however, significant maintenance costs have underscored the need to predict the unsteady loading related to turbine layout and off-design wind conditions. Current design tools omit the fluid-structure interactions that drive costly fatigue loads, and thus, the research community has started utilizing HPC solvers to investigate these phenomena. Unfortunately, such tools are too complicated and expensive for industrial use - particularly in areas related to grid generation and setup. While wind-turbine blades are geometrically simple, their proximity to other components as well as the need to refine grid spacing in high gradient flow regions, complicate the grid generation process. These difficulties are especially severe during design where mesh generation must be repeated following each change. It is exactly this hurdle that the proposed effort seeks to address by enhancing emerging CFD methods into a robust aeroelastic analysis system for wind-turbines, farms and rotating machinery. In an ongoing initial Phase II, CDI recently demonstrated a high Reynolds number adaptive cut-cell Cartesian grid CFD solver generalized to accommodate imperfect surface geometries and obtain predictions autonomously. CDIĺs approach hardens CFD by separating the volume mesh resolution from that of the surface, thus enabling non-experts to rapidly investigate design changes without the associated mesh generation impediments. In the initial Phase II, this flow solver was extended to support combinations of rotating and non-rotating components, and a first- of its kind deforming Cartesian grid concept developed for aeroelastic analysis. Moreover, preliminary overset integration with an efficient AMReX-based prototype far-field flow solver is underway, and will be completed during the remainder of the effort. The proposed Phase IIB is structured to address needs identified by customers during Phase II, and focuses on automated/simplified problem setup, data management in the distributed HPC environment, and implementation of more reliable error handling. The effort would emphasize software interfaces compatible with industry standard engineering software to promote software acceptance/adoption by building upon the industryĺs investment in legacy software. A successful effort would produce a robust and efficient computational tool for wind turbine design and analysis that directly addresses the limitations of software for predicting unsteady turbine blade loading and situational interactions. Based upon a modest market entry, combined sales and associated service work could generate >$163M in cumulative revenue within ten years, with major cost savings for manufacturers and end users from improved predictions and lower maintenance costs. Moreover, as is evidenced by the included letters of commitment, additional commercialization is anticipated through application to other vorticity dominated flows such as rotorcraft.