Technologies for Sustainable Built Environments Centre
Click
to edit
Master
title style
Dynamic
Stall
in Vertical
Axis Wind Turbines
Rosario Nobile | Dr Maria Vahdati | Dr Janet Barlow | Dr Anthony Mewburn-Crook
In Figure 4, strong instability is observed for large angle of
Mesh
Overview
Dynamic stall is an intrinsic phenomenon of Vertical Axis
Wind Turbines (VAWTs) at low tip speed ratios (TSRs)
and its nature can affect fatigue life and energy output
of a wind turbine. A two-dimensional Vertical Axis Wind
Turbine (VAWT) is explored. The analysis is conducted
by using Computational Fluid Dynamics (CFD) tools. The
attacks and low TSRs due to deep dynamic stall. In addition,
The mesh, as shown in Figure 2, is mainly composed of
the development of several peaks, especially for negative
angle of attacks and low TSRs can be associated with the
three sub-domains: one fixed sub-domain outside the
rotor, one dynamic sub-domain around the blades of the
development of upstream wakes that interact with the
rotor and one fixed sub-domain for the remaining part of
downstream blades.
the rotor.
Conclusions
numerical results are compared with experimental data.
Fixed Sub-domain
The key conclusion of this numerical study is that a CFD tool
will allow the visualisation of the flow aerodynamics
involved during the operation of a VAWT that is not
Dynamic Sub-domain
Introduction
The last few years have proved that Vertical Axis Wind
Figure 3. An example of dynamic stall for the 2-D simulation at low TSR and different positions
of the blades.
Turbines (VAWTs) are more suitable for urban areas than
Horizontal Axis Wind Turbines (HAWTs) 1, 2, 3. However, the
aerodynamic analysis of a VAWT is very complicated than
Recently, to study a full scale wind turbine in the wind
tunnel is an infeasible task due to size limitations and costs
involved. Therefore , a Computational Fluid Dynamics
(CFD) Software, ANSYS 12.0, is selected for this study and
code adopted is able to show dynamic stall that is typical
found in VAWTs at low TSRs. Also, from this numerical
Results
analysis appears that in order to achieve a good agreement
The numerical simulations obtained during the present
selection of the turbulent method is fundamental. However,
study are mainly compared with an experimental study
it is strongly suggested to develop a more sophisticated 3-D
In Figure 2, the local mesh around the blades of the rotor is
carried out in 2010 6. The SST method shows a good
model that is more realistic than 2-D.
refined for accurate and efficient resolution of the
boundary layer and wakes.
agreement with the experimental data than the k- and k-ε
conventional wind turbines and at low tip speed ratios
(TSRs<5), VAWTs are subjected to a phenomenon called
dynamic stall. This can really affect the fatigue life of a
VAWT if it is not well understood.
possible with ordinary wind tunnel tests. In general the CFD
Fixed Sub-domain
Figure 2. Mesh and sub-domains for the two-dimensional .VAWT
Boundary conditions and turbulence method
Symmetrical boundaries were used for the top and bottom
methods. Figure 4 shows how the lift and the drag
between numerical and experimental data , the right
References
coefficients, Cl and Cd, are affected by different angles of
1. S. Mertens, Wind energy in the built environment: concentrator effects of buildings.
TU Delft, 2006, pp. 3-14.
attack and TSRs. The curve shapes are in good agreement
2. S. Stankovic, N. Campbell, and A. Harries, Urban Wind Energy. Earthscan, 2009
with the experimental data, which is the red line on the right
side.
3. C. J. Ferreira, G. van Bussel, and G. van Kuik, 2D CFD simulation of dynamic stall on a
Vertical Axis Wind Turbine: verification and validation with PIV measurements,
presented at the 45th AIAA Aerospace Sciences Meeting and Exhibit, 2007, pp. 1-11.
in order to reduce time and memory costs only a 2-D case
is explored.
parts of the 2-D model with no-slip boundary conditions at
the two sides. An opening boundary was chosen for the
Method
output and a constant wind was defined for the inlet.The
three Reynolds-Averaged Navier-Stokes (RANS) turbulence
5. J. Larsen, S. Nielsen, and S. Krenk, Dynamic stall model for wind turbine airfoils, Journal
of Fluids and Structures, vol. 23, no. 7, 2007, pp. 959-982.
Geometry
methods are: the standard k-ω model, the standard k-ε
model and the SST model 4.
6. S. Wang, D. B. Ingham, L. Ma, M. Pourkashanian, and Z. Tao, Numerical investigations
on dynamic stall of low Reynolds number flow around oscillating airfoils, Computers &
Fluids, vol. 39, no. 9, 2010, pp. 1529-1541.
As shown in Figure 1, the 3-D solid model of the rotor was
generated with ProEngineer 4.0. And the 2-D model of the
Dynamic stall
The author would like to thank my academic supervisors Dr M. Vahdati and Dr J. Barlow and
my industrial supervisor Dr A. Mewburn-Crook for their supports for this work.
As shown in Figure 3, dynamic stall is mainly characterised
•
by flow separations at the suction side of the airfoil 5. This
The author is also grateful to the EPRSC and MatildasPlanet for funding this project.
can be summarised in four crucial stages:
2-D
•
Leading edge separation starts,
•
Vortex build-up at the leading edge,
•
•
Detachment of the vortex from leading edge and
build-up of trailing edge vortex,
Department of Technologies for Sustainable Built Environments, University of Reading,
Whiteknights, RG6 6AF
•
Email: r.nobile@reading.ac.uk
•
www.reading.ac.uk/tsbe
•
Figure 1. Three dimensional rotor of a straight-bladed Darrieus wind turbine obtained with
ProEngineer 4.0 and two dimensional rotor extrapolated from middle plane.
Acknowledgements
•
VAWT was generated from the middle plane and imported
into ANSYS CFX 12.0.
3-D
4. D. C. Wilcox, Turbulence Modeling for CFD. DCW industries La Canada, 2006.
Detachment of trailing edge vortex and breakdown of
leading edge vortex
Contact information
Figure 4. Lift and Drag coefficient (Cl and CD ) from numerical studies and experimental data.