Like every other team currently taking part in the World Solar Challenge, an arduous 3,000 km (1,864 mi) solar-powered race across the Australian outback, the University of Michigan (UM) Solar Car Team will look to keep its car chugging along by exposing it to as much sunlight as possible. But unlike the others, it has brought along a little piece of extra technology it hopes will offer an edge. Developed by IBM, the solar forecasting system tracks the clouds moving overhead so the team knows where they need to be and when to draw maximum energy from the sun.
Trying to position a solar-powered race car under clear, energizing blue skies without compromising on speed is a constant juggling act for the teams taking part in the WSC. Participants are regularly faced with decisions when hurtling down a highway with clouds moving in the same direction overhead. Do they speed up to try and get out ahead of the cloud, slow down and let it pass (sacrificing crucial seconds), or simply do nothing and ride it out?
IBM's Solar Forecasting Technology aims to take some of the guesswork out of this dilemma. It uses a basketball-shaped camera mounted on the hood of UM's weather scouting vehicle, which drives around ten or fifteen minutes ahead of its race car Aurum. Equipped with a super-wide angle lens, a solar pyranometer, GPS sensor and onboard processor, this "Sky Camera" system gathers data about the cloud movement and instant solar radiation readings.
This information is combined with weather data about wind speed and direction and a predictive model to to produce near-term weather forecasts for the team, around ten to fifteen minutes ahead. This then helps to inform them about whether they should adjust their speed in order to avoid cloud cover. IBM says that according to its simulations, the technology could save the team up to fifteen minutes per day. Last night at the end of day two of racing, 15 minutes separated the top three, with the UM team sitting in third place.
In addition to the short-term gain, the system also uses historical data and various weather and solar prediction models to offer more accurate forecasts along the race route two or three days ahead. IBM says this means the team can determine how much solar energy will be available at different locations and times, helping it choose where and when to park the car for recharging during the four-day-long cross county haul.
While the World Solar Challenge offers a very useful test bed for IBM's forecasting technology, it could come to have much wider applications. It is part of a larger program supported by the US Department of Energy's SunShot Initiative, which aims to improve the accuracy of solar forecasts so grid operators can better manage the incoming energy from their various sources, such as wind, coal, gas or solar. IBM claims the system centered around its Sky Camera is 30 percent more accurate than most current approaches, which rely on weather models that don't consider all the variables.
You can hear Leda Daehler, strategist for the UM Solar Car Team speak about the forecasting technology in the video below.