Anda di halaman 1dari 3

1000 Islands Fluid Dynamics Meeting, 2012

WIND EFFECTS ON PHOTOVOLTAIC (PV) PANELS A CFD APPROACH


C.M. Jubayer and H. Hangan
The Wind Engineering, Energy and Environment (WindEEE) Research Institute
Western University, London, ON.

Introduction dimension of 2.48 m (B) x 7.29 m (W) x 1.65 m (H)


(Fig. 1), obtained from the specification provided by
Solar farms consist of large number of flat PV
First Solar Inc. This gives an aspect ratio (W/B) of
panels arranged in arrays and are usually developed
2.94. The inclination of the panel with the horizontal
in an open terrain to get unobstructed sunshine, due
is 25o. Two different wind directions, 0o and 180o are
to which PV panels in the farm experience higher
considered, 0o being the wind normal to the leading
wind loads and more potential damage. On the other
edge of the panel and upper surface of the panel
hand, the existing codes for estimating wind loads on
facing the oncoming flow. The computational domain
buildings/structures (NBC 2005, ASCE 7-05) do not
and mesh are created according to the COST
provide sufficient information to predict wind loads on
guidelines (Franke et al., 2007) and are generated
the ground mounted arrays of solar PV panels.
using Pointwise (Pointwise, Inc.). Dimension of the
Most of the previous works performed on wind
domain is 21H (length) x 24H (width) x 6H (height),
loading on solar collectors/PV panels mainly focused
where H is the height of the panel. The created mesh
on the panels mounted on the horizontal or inclined
has 1.7 million cells. At the domain inlet, profiles of
roof of a building (Wood et al., 2001; Kopp and Surry,
velocity and turbulence properties are obtained from
2002; Meroney and Neft, 2010; Chung et al., 2011).
ESDU (ESDU 82026, 83045) for design wind speed
Ground mounted solar panels were studied by a very
of 26 m/s at 10 m height with aerodynamics
few (Bitsuamlak et al., 2010; Shademan and Hangan,
roughness length (y0) of 0.03 m which represents an
2010). Previous studies showed a significant
open terrain. The bottom of the domain is modelled
variation of a) geometry of panels (dimension, aspect
as rough wall with roughness height and roughness
ratio of the panel), b) mounting locations (horizontal
constant provided as to be consistent with the
roof, inclined roof, ground) and c) installation
boundary layer profile.
methods (inclination angle, vertical spacing from the
surface on which the panel is installed) resulting in
different wind loads estimation.
The objective of this research is to evaluate wind
loads on the ground mounted PV panels using a
computational fluid dynamics (CFD) approach with
application to PV solar farms. This study also aims to
analyze the wind flow behaviour around the PV
panels. The focus of the present study is a stand-
alone system which is a unit system of a solar farm. (a)

Methodology
In this study 3D unsteady Reynolds-Averaged
Navier-Stokes (RANS) simulation of wind flow over a W
ground mounted stand-alone PV system using Shear
Stress Transport k- (SST k-) turbulence closures (b)
is performed. The stand-alone PV system has the Figure 1: Computational model for the PV panel
1000 Islands Fluid Dynamics Meeting, 2012

Open source CFD toolbox OpenFOAM v2.0.0 (SGI


CFD (Upper surface)
Corp.) is used for the simulation. In this study CFD (Lower surface)
pisoFoam, which is a transient solver for Experiment (Upper surface)
incompressible flows that uses PISO (Pressure Experiment (Lower surface)
Implicit with Splitting of Operators) algorithm, is 2
employed. Simulation has been run for 20 s of flow

Mean Cp
-5 1
time with time step size of 10 s. Data for the last 15
s of flow time are used to report the results. 0
Results and Discussion -1
Streamlines of mean velocity around the panel for -2
both 0o and 180o wind directions are shown in Figure 0 0.2 0.4 b/B 0.6 0.8 1
2. From Figure 2(a), it can be seen that the flow
remains attached on the upper surface of the panel. (b)
At the lower surface of the panel two separate Figure 3: Mean Cp profiles along the midline of the
vortices are observed (Fig. 2a) due to the support panel surface for wind directions of (a) 0o and (b)
structure. On the other hand, for the 180o wind 180o. Here, b is the distance from the leading edge
direction, flow separates at the leading edge and along the breadth of the panel
forms a large recirculation region rotating clockwise
on the upper surface of the panel and close to the conducted in the RWDI wind tunnel where a 1:20
trailing edge, a small recirculation region rotating scale model of a stand-alone system with aspect ratio
counter clockwise is formed due to the roll-up of the of 6.85 was placed inside the atmospheric boundary
shear layer (Fig. 2b). For both 0o and 180o wind layer (ABL)flow for open terrain. A higher mean Cp
directions, shedding of vortices are not observed. values (magnitude) are observed for experiment
This is more likely due to the RANS modelling compared to CFD. These differences in mean Cp
approach (Breuer et al., 2003). Mean pressure values may be due to the differences in aspect ratios,
coefficient (Cp) along the mid-line between two scaling of the experimental model, height of the panel
support structures on the upper and lower surfaces of from the ground or the incoming flow conditions.
the panel for both 0o and 180o wind directions are Conclusion
compared with the experimental results obtained
Unsteady RANS simulations of wind loading on a
from Dr. Girma Bitsuamlak by personal
communications (Fig. 3). This experiment was ground mounted PV panel system immersed in the
atmospheric boundary layer are performed for two
different wind directions (0o and 180o). Wind flow field
and mean Cp values along the midline on the
surfaces of the panel are analyzed.
Wind Wind
Reference
ASCE/SEI 7-05, 2006, Minimum design loads for
buildings and other structures, American Society
of Civil Engineers.
(a) (b) Bitsuamlak, G. T., Dagnew, A. K. and Erwin, J., 2010,
Figure 2: Streamlines of mean wind velocity for (a) 0o Evaluation of wind loads on solar panel modules
o
and (b) 180 wind directions using CFD, The Fifth International Symposium on
Computational Wind Engineering, Chapel Hill,
CFD (Upper surface) North Carolina, USA.
CFD (Lower surface)
Experiment (Upper surface) Breuer, M., Jovicic, N. and Mazaev, K., 2003,
Experiment (Lower surface) Comparison of DES, RANS and LES for
2 separated flow around a flat plate at high
incidence, International Journal for Numerical
Mean Cp

1 Methods in Fluids 41, 357-388.


0 Chung, K. M., Chang, K. C. and Chou, C. C., 2011,
Wind loads on residential and large-scale solar
-1 collector models, Journal of Wind Engineering
-2 and Industrial Aerodynamics 99, 59-64.
0 0.2 0.4 0.6 0.8 1 ESDU, 1982, Strong winds in the atmospheric
b/B boundary layer. Part 1: Mean hourly wind speeds,
(a) Engineering Science Data Unit Number 82026.
1000 Islands Fluid Dynamics Meeting, 2012

ESDU, 1983, Strong winds in the atmospheric


boundary layer. Part 2: Discreet gust speeds,
Engineering Science Data Unit Number 83045.
Franke, J., Hellsten, A., Schlunzen, H. and
Carissimo, B., 2007, Best practice guideline for
the CFD simulation of flows in the urban
environment, COST Action 732; Quality
assurance and improvement of microscale
meteorological models.
Kopp, G. A., Surry, D. and Chen, K., 2002, Wind
loads on a solar array, Wind and Structures 5,
393-406.
Meroney, R. N. and Neft, D. E., 2010, Wind effects
on roof-mounted solar photovoltaic arrays: CFD
and wind-tunnel evaluation, The Fifth
International Symposium on Computational Wind
Engineering, Chapel Hill, North Carolina, USA.
NBC 2005, National Building Code of Canada,
volume 1, Institute of Research in Construction,
National Research Council of Canada.
Shademan, M. and Hangan, H., 2010, Wind loading
on solar panels at different azimuthal and
inclination angles, The Fifth International
Symposium on Computational Wind Engineering,
Chapel Hill, North Carolina, USA.
Wood, G. S., Denoon, R. O. and Kwok, K. C. S.,
2001, Wind loads on industrial solar panel arrays
and supporting roof structure, Wind and
Structures 4 , 481-494.