A Wind Tunnel Study of Atmospheric Boundary Layer Effects on Roof Mounted Solar Panel Arrays

With the rising cost of conventional fossil fuels, alternative energies such as solar power are becoming more attractive. With government incentives, private companies are being encouraged to transfer their power demand from the electrical grid to on site power generation from solar sources. Incentives are also in place to enable developers to push solar panel array products to market for widespread and large scale installations. Installing solar arrays on roof top structures requires confirmation from the National Building Code that the roof top installation is safe and legal. However, due to the unique geometry and dynamic nature of solar panels, specifically those that track the movement of the sun, the National Building Code requires experimental validation of the installations' safety. This thesis examines the current nature of the solar industry, and the means in which the national building code of Canada supports it. A review of fluid fundamentals, the atmospheric boundary layer, and computational principles will prepare the reader for understanding the project's experiments. The research will examine the operating conditions that such a solar panel array on the roof of a large retail building may experience, and comment on the limitations that exist for practical full scale models while exposed to wind, known as the earth's atmospheric boundary layer. Using a load cell, measurements of forces and bending moments on scaled solar panel models were taken inside a low speed atmospheric boundary layer wind tunnel with the panel mounted on the roof of a scaled retail building. These measurements were used to validate two dimensional large eddy simulations on a computer. Differences present in the results are mainly due to the nature of the solar panel's and retail building's three dimensional effects. Vortex shedding was also observed and commented upon. It was confirmed that force and moment coefficients obtained during wind tunnel testing are insensitive to Reynolds numbers, and may thus be used for full scale model calculations. A variety of trends were noted from the wind tunnel measurements regarding the effect of numerous panel configurations on the roof top. A conclusion is finally made regarding the feasibility of the research and its application to future studies.