WIND FLOW CHARACTERISTICS AROUND ROOFTOP SOLAR ARRAY - A NUMERICAL STUDY

The Eighth Asia-Pacific Conference on Wind Engineering, December 10 14, 2013, Chennai, India WIND FLOW CHARACTERISTICS AROUND ROOFTOP SOLAR ARRAY - A NUMERICAL STUDY D.Ghosh 1, A. K. Mittal 2, S. Behera 3, A. Gupta 4 1 M. Tech Student, CSIR-Central Building Research Institute, Roorkee, India, debdutta001@gmail.com 2 Principal Scientist, CSIR-Central Building Research Institute, Roorkee, India, akmittal@cbri.res.in 3 Scientist, CSIR-Central Building Research Institute, Roorkee, India, siddharth@cbri.res.in 4 ECP Consultant, Roorkee, India, amitgupta@ecproorkee.com ABSTRACT Photovoltaic (PV) Solar panel systems are very popular alternative energy sources all over the world. These systems are usually mounted on building roof tops or in the open ground and are often vulnerable to wind forces. It may be analyzed like a mono-slope roof but location wise it may not be the correct procedure. Limited wind tunnel studies are carried out on this aspect worldwide. In this paper the wind action on solar array is investigated using Computational Fluid Dynamics (CFD). Recently, a solar power generation facility has been developed at CSIR-CBRI with roof mounted solar arrays. The wind effects on one solar array situated centrally on the roof is considered in the study. Results of the CFD simulation are presented in the paper and may be used as an approximate estimate of wind load on solar array. ANSYS FLUENT v14.5 is used for the simulation of the problem. The results reveal that the solar array is subject to significant suction under wind load which may lift the array. Keywords: Solar array, Wind force calculation, CFD, ANSYS FLUENT Introduction The design of solar arrays is often governed by wind loads. It needs to be designed for extreme wind conditions at a particular region. Large number of wind tunnel tests for ground and roof mounted solar arrays are studied by researchers considering parameters viz. wind direction ( ), building height(h), tilt( ), etc. highlighted by : Chevalien et al.(1979), Radu et al. (1986), Radu and Axinte (1989), Wood et al. (2001), Kopp et al. (2002), Chung et al.(2008), Shademan et al.(2009), Ruscheweyh and Windhovel (2011), Stathopoulos et al. (2012, 2013) etc. Uematsu et al. (2008), performed several wind tunnel testing on free standing canopy roofs to determine the characteristics of wind loads on solar array like structures. Although wind tunnel testing is recommended for the determination of wind load on rooftop panels, but with the help of CFD simulation (by creating same wind condition inside the numerical domain around the object to be tested) the characteristics of wind can be assessed. CFD technique has been adopted for the simulation of wind conditions around CBRI building and the effect of wind pressure on roof top solar array is estimated. The solar arrays are considered as mono-slope roof. In the present study, a roof top solar array of CBRI situated at different locations is considered. Besides, another solar array on top of a square shaped building is also considered for the comparison of pressure coefficients. The general topography and aerial view of CBRI building is shown in Fig. 1. The typical locations of the solar array on CBRI rooftop is shown in Fig. 2. Proc. of the 8th Asia-Pacific Conference on Wind Engineering Nagesh R. Iyer, Prem Krishna, S. Selvi Rajan and P. Harikrishna (eds) Copyright c 2013 APCWE-VIII. All rights reserved. Published by Research Publishing, Singapore. ISBN: 978-981-07-8011-1 doi:10.3850/978-981-07-8012-8 328 674

Modelling of Wind flow around CBRI building The dimensions of the computational domain are fixed according to the guidelines mentioned in Blocken et al. (2007) and Revuz et al. (2012). A geometric scale of 1:40 is adopted for both building and solar array with a blockage ratio of 0.2%. The actual dimensions of computational domain are L D x W D x H D = 820 x 560 x 720 m 3 in full scale (20.5x 14 x 18 m 3 in simulation scale). The dimension of the solar array and the building are depicted in Fig. 3. The dimension of CBRI building is 77.66 x 17.36 x 9.0 m 3. A single solar array of 5.02 x 1.96 m 2 is situated both at edge as well as at the centre of the roof of CBRI building. Besides, another building with square plan of size 40 x 40 x 9 m 3 with solar array situated at the centre of its roof. The solar array inclination is kept constant at 25. The unstructured hexahedral computational mesh is made with various sizes. Near wall treatment is used in the domain as well as around the model. The distance (y p ) between the centre point of the wall-adjacent cells and the ground surface i.e. first layer thickness is decided according to the Reynolds number and y+ (30< y+<100) value and is fixed at 0.005 m (simulation scale) which is equivalent to 0.2m (Fig.4). Total number of hexahedral cells is nearly 15 x 10 5, are employed for the numerical domain. A commercial CFD code, Fluent is used to solve the 3D Reynolds- Averaged Navier Stokes equations and the continuity equation using the control volume method. Realizable k- model is used for turbulence modelling and SIMPLE algorithm is used for the pressure velocity coupling. First-order upwind discretization schemes are used for solving both convection and viscous terms of the governing equations. Calculations were repeated until small residuals were obtained (~1000 iterations). CSIR-CBRI Building Fig. 1 CSIR-CBRI building G-Earth image Fig. 2 Rooftop solar array of CSIR-CBRI Fig. 3 Geometry of CSIR-CBRI building 675

Fig. 4 Meshing of Computational domain Numerical simulation of Atmospheric Boundary Layer Boundary layer velocity profile has been created with the help of user defined function (UDF) in ANSYS FLUENT. Turbulence energy and turbulence dissipation rate is varied accordingly through UDF. In boundary layer simulation, a very accurate description near the ground surface is required. Roughness height decided to simulate sand grain roughness is 2.5 x 10-6 (scaled) or 0.0001m in full scale, which is equivalent to equivalent sand grain roughness of a smooth floor. This value is smaller than the y p (0.005) in simulation scale. Velocity distribution throughout the numerical domain can be seen in Fig. 5. The velocity along the height at domain inlet is measured for boundary layer depiction (Fig. 6), where U 0 =5.9. Fig. 5 Velocity distribution throughout the domain 676

Fig. 6 Atmospheric Boundary layer in inlet Results Pressure Coefficient (C P ): The coefficient of pressure is calculated as following. Cp = P 1 2 P ρu 0 2 0 ρ is free stream density, u 0 is free stream velocity, P 0 free stream static pressure, P static pressure at the point of interest. The pressure coefficient along the width of the solar array is compared for different locations on rooftop of the buildings considered in Fig.7. Fig.7 Pressure coefficient on the rooftop of solar array situated at different location. Pressure contours The pressure contours on single solar array placed at different locations is shown in Fig. 8. 677

5.02m Fig. 8 (a) Solar array at the edge (b) Solar array at the centre (c) Solar array at the centre of of CBRI Building of CBRI Building a square building Pressure in the front face of the building is more than other faces of building. Solar array is subjected to lesser pressure than the building face towards the wind direction, but array may suffer lift due to negative pressure created around the array. Pressure contours on building front face and solar array and pressure field near building periphery is depicted in corresponding Fig. 9. Wind Flow Velocity streamlines Fig. 9 Pressure contour on building and solar array Velocity stream around the building periphery and solar array is depicted in Fig. 10 (Side elevation). Fig. 11 depicts the wind flow streamline around building (Plan view) at a particular plane. Vortex is clearly visible in the leeward side of CBRI building. Solar array at the edge of CBRI building subjected to less negative pressure (Fig. 8 (a)) than centrally located solar array at CBRI building (Fig. 8 (b)) because the windward face of the building saparates the flow and it starts reattaching near the leeward edge (Fig. 10 (a) and 10 (b)). 678

Fig.10 (a) Velocity streamline solar array at the edge of CBRI building. (b) Velocity streamline solar array at the centre of CBRI building. (c) Velocity streamline solar array at the centre of square building. Velocity contour Fig.11 Velocity streamline around building in horizontal plane Velocity contours are showing low pressure (negative velocity) region in leeward side of the building (Fig.12). Solar array is subjected to high velocity due to amplification of wind flow at roof. The solar array, situated at the centre of the CBRI building is subjected to comparatively high velocity (negative) than the other two cases which agrees with pressure coefficient values (C p ). Fig. 12 (a) Solar array at the edge of CBRI building. Fig. (b) Solar array at the centre of CBRI building Fig. (c) Solar array at the centre of square building 679

Conclusions 1) Pressure coefficient (C p ) on the solar array varies from -6 to -10 for centrally located solar array on CBRI building as compared to solar array located at edge (-0.6 to -5.6) and centre of square building (-3 to -7). 2) Average C p on the centrally located solar array is two times solar array located at leeward edge of the building (Fig. 7). It happens due to wind flow separation by the windward edges of the building and creation of significant negative pressure at the position of centrally located solar array. 3) Most of the local C p values as per IS 875 Part-3, for a 25 mono-slope roof with solidity ratio of 1, coincide with the C p values obtained for a array situated at the edge of the CBRI building in our study (Fig. 7). 4) Windward face of the building is subjected to maximum positive pressure as expected. 5) Low pressure zone has been created on roof and corners of the building. 6) The solar array situated at the top of a building is subjected to considerable amount of negative pressure as expected, for which arrays have to be designed. Future Scope: 1) This way numerical modelling can be an alternative approach to the costly and time consuming wind tunnel experiment to assess the wind characteristics around building. 2) Multiple solar arrays can be placed and CFD simulation can be possible for actual load prediction. 3) Most critical location of the solar array on the roof top is to be studied. References: Blocken, B., Stathopoulos, T. and Carmeliet, J. (2007), CFD simulation of the atmospheric boundary layer: wall function problems, Atmos. Environ., 41(2), 238-252. Chevalien, L., Norton, J. (1979), Wind loads on solar collector panels and support structure Aerospace Engineering Department, Texlas A&M University Chung, K., Chang, K., Liu, Y. (2008). Reduction of wind uplift of a solar collector model, Journal of Wind Engineering and Industrial Aerodynamics, 96, 1294-1306 Indian Standards (IS). (1987). Code of practice for design loads (other than earthquake) for buildings and structures.(second Revision) IS 875 (Part 3) Wind Loads, New Delhi, India. Kopp, G., Surry, D., Chen, K. (2002), Wind loads on solar array Wind and Structures 5, 393-406. Radu, A., & Axinte, E. (1989). Wind forces on structures supporting solar collectors, Journal of Wind Engineering and Industrial Aerodynamics, 32(1-2), 93-100 Revuz, J, Hargreaves D.M., Owen J.S. (2012). On the domain size for the steady-state CFD modelling of a tall building. Wind and structures, 15 (4), 313-329. Revuz, J. (2011), Numerical simulation of the wind flow around a tall building and its dynamic response to wind excitation, PhD Thesis, The University of Nottingham, Nottingham, UK. 680

Ruscheweyh, H. & Windhovel, R. (2011) Wind loads at solar and photovoltaic modules for large plants 13th International Conference on Wind Engineering (ICWE13), (pp. 1-8). Amsterdam, The Netherlands Shademan, M. & Hangan, H. (2009). Wind Loading on Solar Panels at Different inclination Angles The 11th American conference on Wind Engineering. San Juan, Puerto Rico. Stathopoulos,T., Xypnitou, E., Zisis, I. Wind Loads on Rooftop Solar Panel Systems: A Contribution to NBCC 2015 Uematsu, Y., Stathopoulos,T., Iizumi, E., (2008) Wind loads on free-standing canopy roofs: Part 2 overall wind forces Journal of Wind Engineering and Industrial Aerodynamics, 96, 1029-1042. Wood, G. S., Denoon, Roy O; Kenny, C.S. Kwok (2001). Wind loads on industrial solar panel arrays and supporting roof structure Wind and Structures, 4(6), 481-494. 681