Solar Array Analysis Suite

Abstract

The University of Michigan Solar Car Team (UMSCT) was founded in 1989 with their first car, Sunrunner, taking first place in the 1990 GM Sunrayce and third place in the 1990 World Solar Challenge. In the 31 years since, the team has built 14 solar-powered electric race cars, taken eight more national championships, one international championship, and earned a podium finish in the World Solar Challenge six additional times. In 2021, concerns over the COVID-19 virus led the South Australian government to cancel the World Solar Challenge for the first time in the history of the event. Thus, the team decided to attempt a completely solar-powered Cannonball Run from coast to coast across the United States of America. Due to the increased distance and presence of mountains, this route will put even more importance on the performance of the solar array than the World Solar Challenge would have. Furthermore, the sun will now be largely on one side of the car rather than mostly behind it, which will increase the amount of energy lost to shading and cause light distribution across the array to be uneven. Thus, I created a software suite that integrates with the team's existing workflow to enable CAD files of a potential car and prospective array layout to be analyzed in greater depth than the team's prior methods involving hand calculations and spreadsheets would allow. Each cell's exact location and normal vector can be determined, and ray tracing is used to determine how much sunlight will be incident on each cell at any given point of a race, allowing array losses due to the curvature and angle of each cell to be determined. Then, an approximation for the losses suffered by the array due to their temperature is calculated and applied, giving an estimate for the power output by the array. This power output estimate is then fed into a mathematical model which can approximate the efficiency of the Maximum Power Point Trackers (MPPTs) that boost the output voltage of the array up to the voltage of the battery. By using a 4D hyperplane fit, we can get a very good approximation of the efficiency that will be experienced under any given set of conditions, allowing us to account for changes in the MPPT efficiency over the course of an entire race and determine how much power is provided to our powertrain, enabling much more accurate predictions of array efficiency than ever before and enabling the team to make more informed decisions about the shape of their car and the layout of their solar cells.