Year: 2025
Type: conference-paper
Abstract: The ducted wind turbine designed and built at Clarkson University is unique because of its ability to harness wind energy more effectively through aerodynamic optimization guided by Reynolds-Averaged Navier-Stokes (RANS) and actuator disk models. Similar to other diffuser-augmented wind turbines, the Clarkson design is based on increased wind velocity and higher power output from relatively low wind speeds. Clarkson’s ducted wind turbines have a rotor diameter D of 3.0m and have performed considerably under ambient conditions below the initial wind tunnel data. These ducted wind turbines employed the Eppler 423 airfoil at an angle of attack (AoA) of 25 degrees for the diffuser (Eppler25). The rotor blades were created based on cambered plate GOE417a airfoils that work very well for small wind turbines. These airfoils can reduce the blade production cost substantially because of their low tooling and manufacturing costs. Their hub diameter is 0.155D. In this study, a new diffuser is designed based on the Selig airfoil at a 20-degree AoA (Selig20). The present study employs a flow-driven-rotor model of a commercial CFD package, Simerics-MP+, based on unstructured-grid finite volume solutions of URANS equations for the flow field that are two-way fully coupled with a dynamic solution of the rigid-body rotation of the turbine rotor. This study investigates the aerodynamic performance of a ducted wind turbine at a Reynolds number of 2.05 million based on the rotor diameter and freestream velocity. The solution starts with the fluid solver by solving three-dimensional incompressible Navier -Stoke equations with the standard k-ε turbulence models to compute a fluid torque fluid over the rotor. The second-order implicit Euler scheme was used for the time marching of the momentum equations. Simerics-MP+ uses pressure-velocity coupling techniques similar to PISO and SIMPLE for solving incompressible flows, ensuring that pressure and velocity fields are consistently updated to satisfy mass conservation and momentum balance. Subsequently, the fluid torque is used as an input to a rigid-body-rotation dynamics solver for computing the rotor’s angular speed rotor, using an explicit time-marching scheme. The rotational rate of the rotor is then fed into the fluid solver to determine the rotational rate of the spinning domain and moving wall boundary conditions. Finally, the results from the flow-driven rotor simulations are compared to those from the prescribed-rotation model described in Safford et al (2024) to evaluate differences in flow characteristics, rotor dynamics, and turbine efficiency for Eppler25 and then Selig20 ducted wind turbines. The first outcome of this study is that the flow-driven rotor model of Simerics-MP+ can model a ducted wind turbine self-starting and predict a fully developed flow for a saturated tip speed ratio for both rotor-at-throat and rotor-at-rear cases. The flow-driven-rotor model successfully predicts the optimal thrust coefficient while the prescribed-rotation model fails to do so. While the optimal Cp predicted by the flow-driven rotor model is fairly close to the prediction from the prescribed-motion model, the flow-driven-rotor model is generally more accurate in predicting underperformance under ambient wind conditions away from the optimal tip speed ratio. The flow-driven rotor model does offer a new path to simulate ducted wind turbines under ambient wind conditions. The second major finding is that the rotor-at-rear configuration gives higher Cpt than the rotor-at-throat configuration for the Eppler25 ducted wind turbines. Both configurations have gap flows strongly attached to the duct's inner wall. The tip vortices are pushed to the duct inner wall and propagate with a longer distance along the gap flow streamlines for the rotor-at-throat case. The third outcome of this study is the Selig20 ducted wind turbine gives a slightly higher Cpt than the Clarkson Ducted Wind Turbine. The Selig20 ducted wind turbine has a wider range of optimal tip speed ratios enabling it to harvest more wind energy under ambient wind conditions. We anticipate the Selig20 ducted wind turbine produces less flow separation than Epper25 and will check into the surface pressure distributions for the final manuscript.
DOI: 10.2514/6.2025-97819
Authors: Seyi Oluwadare, Dr. Chunlei Liang
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4413191244