Analysis of low-fidelity models for aerofoil design and optimisation in curvilinear flows

Abstract

Classical aerodynamics often struggle to predict the performance of devices operating in curvilinear flows, such as Vertical Axis Turbines (VATs) and cycloidal rotors. These applications exhibit aerodynamic phenomena such as virtual camber, complex unsteady effects, and wake interactions that traditional theories cannot fully model. This study systematically compares low-to-high-fidelity aerodynamic models against experimental data from a cyclorotor operating in hover (Re ≈ 30, 000), highlighting each methods’ limitations. The cyclorotor uses three different aerofoils to assess the methods’ suitability for ranking designs in a design optimisation process. Results show that classical unsteady models, such as Theodorsen’s theory, are ineffective for achieving the correct instantaneous force/torque profile. A quasi-steady approach using a single stream-tube model for cyclorotor downwash was the most effective analytical method for predicting force and torque for standard symmetric aerofoils. However, this approach struggles to capture the impact of varying aerofoils. Unsteady panel methods can rank aerofoil thrust performance but not torque. Only the more computationally intensive Unsteady Reynolds-Averaged Navier-Stokes (URANS) simulations reliably ranked designs for thrust, torque, and figure of merit. While 2D URANS achieved similar time-averaged metrics to experiments, its instantaneous force/torque profiles differed significantly from the more accurate, but costly, 3D URANS results.

More information

Identifiers

ORCID for Benjamin Patrick Ibrahim Irwin: orcid.org/0000-0001-7939-6682

ORCID for David Toal: orcid.org/0000-0002-2203-0302

ORCID for Swathi Krishna: orcid.org/0000-0003-4316-7428

Catalogue record

Export record