Numerical and Parameterization Sensitivity of Pitch Trajectory Optimization for Vertical-Axis Wind Turbines
Year: 2026
Type: article
Abstract: Pitch trajectory optimization supports active control of vertical-axis wind turbines (VAWTs), but its sensitivity to numerical settings and trajectory parameterization and the reliability of its aerodynamic predictions across applications remain insufficiently quantified. This study evaluates a lifting-line free vortex wake model (LLFVWM) coupled with sequential least squares programming (SLSQP). Sensitivity analysis shows that azimuthal discretization affects simulation resolution and stability, while convergence tolerance, finite-difference step size, and initialization influence convergence and the attained local solutions. Trajectory parameterization, through control-point distribution, number, and representation, defines the admissible trajectory family and influences optimized aerodynamic performance. Using settings selected through this analysis, the workflow was applied to power maximization, rated-power regulation, weighted power–thrust trade-offs, and trajectory simplification. It achieved a 27% increase in the predicted maximum power coefficient relative to fixed pitch and generated rated-power trajectories varying smoothly across adjacent wind speeds. The power–thrust analysis identified similar trade-off trends across three turbine scales, while the independently optimized stepwise trajectory preserved the main power–TSR trends and retained 94% of the smooth trajectory’s maximum power coefficient. Computational fluid dynamics (CFD) comparisons yielded Pearson correlations of 0.8289–0.9980 and relative L2 errors of 11.95–47.12% across optimized trajectories, indicating agreement in principal aerodynamic trends but case-dependent quantitative discrepancies. These findings clarify how numerical settings and trajectory parameterization influence VAWT pitch trajectory optimization and quantify the aerodynamic prediction accuracy of representative optimized trajectories.
Source: Energies
DOI: 10.3390/en19225567
Authors: Fawei He, Fue-Sang Lien, William Melek
Institution: University of Waterloo
Language: English
Field: Energy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,137
Awards: G2053342940, G7080964852
OpenAlex ID: W7236352351
