High-Fidelity Aerostructural Design Optimization of Wind Turbine Rotors

Abstract

The mitigation of climate change effects through the decarbonization of the global production system is one of the most pressing challenges of our time. Wind and solar power will account for tenths of petawatt-hour yearly by the second half of this century. Thus, efficient energy harvesting systems need to be rapidly developed to match the demand for low-cost-per-kilowatt electricity from renewable sources.

Wind energy technology has matured prominently in the last few decades. However, the growing size and complexity of modern wind turbines is pushing the boundaries of conventional design tools and methodologies. Large and highly flexible rotors are characterized by complex multiphysics interactions and unconventional design features that require novel investigation strategies. High-fidelity design tools promise to address these arising challenges. Their application however has been limited by their computational and implementation costs so far.

The outcome of my doctoral work is the first set of large-scale high-fidelity wind turbine rotor aerostructural design optimization studies. I use a framework that couples computational fluid dynamics (CFD) and computational structural mechanics (CSM) analyses with an efficient gradient-based optimization algorithm. The coupled-adjoint sensitivity solver implementation makes the problem numerically tractable and highly scalable. The optimization studies I perform account for multiple representative steady-state below-rated operating conditions and use hundreds of structural and geometric design variables simultaneously.

First, I adapt an existing aircraft design framework to handle a wind turbine rotor and its system sensitivities. Both the aerodynamic solver and geometry manipulation tool are enhanced, and a newly developed aerostructural rotor model is verified.

Following that, I demonstrate the framework optimization capabilities on this benchmark rotor model. I dissect optimal design features and explore tradeoffs between steady-state aerodynamic efficiency and structural weight, quantifying the impact of design freedom and objective formulation on the final design.

A more refined aerostructural model is then used to perform composite fiber angle tailoring. The optimizer leverages material anisotropic properties to increase local stiffness and reduce the sandwich panel thickness. The resulting designs are more than 10% lighter than the benchmark for the same planform shape.

Next, I optimize the rotor for multiple wind and rotation velocities. Single-point designs are outperformed over a larger operational envelope and flow separation at off-design conditions is mitigated. Enabling modifications to airfoil geometry leads to radical rotor redesign. Using the full set of geometrical and structural variables leads to an aeroelastically tailored rotor with more than 10% higher average torque and a 40% lighter structure than the baseline model.

Finally, I collaborated to the development of a mixed-fidelity framework combining our high-fidelity code with conventional design tools. This prototype software enables high-fidelity optimization including life-cycle sizing constraints. We use state-of-the-art software to predict extreme loads and fatigue damage from time-accurate multiphysics simulations. Two different approaches are proposed to integrate these constraints within our original framework.

Modelling the full range of operating conditions and improving the tradeoffs between accuracy and cost remain open challenges. Nevertheless, this work demonstrates how coupled CFD-CSM approaches can capture unconventional design tradeoffs and provide an unprecedented level of insight on the optimal blade design features. High-fidelity optimization can complement conventional design approaches using more realistic blade-resolved rotor models and leveraging a larger design space. The results I present pave the way for more comprehensive design studies and holistic multifidelity design optimization workflows.