News•February 26, 2013

Solar Forecasts Could Help Electric Utilities and Climate

As Jimmie Johnson screamed past the checkered flag to win the Daytona 500 this past Sunday — few NASCAR fans were probably thinking about solar radiation.

But Sheldon Drobot probably was. Drobot, deputy director of the Weather Systems and Assessment Program at the National Center for Atmospheric Research (NCAR), is part of a team working to create solar forecasts. Those forecasts would predict how much sunlight would reach the ground in a given location every 15 minutes for the next 36 hours. That’s important to auto racers because the stronger the sunlight, the hotter the track surface gets, which in turn affects how well tires perform.

Credit: Department of Defense

It’s even more important to electric utilities that want to tap into solar power as a non-polluting, climate-friendly alternative to the coal- or gas-fired power plants that spew heat-trapping carbon dioxide into the atmosphere to warm the globe. Solar is certainly better for the climate, but as Drobot explained, it can also be risky to count on.

“Say you’re getting x-number of kilowatts from your solar farm, when suddenly a cloud rolls in,” Drobot said in an interview. “Now you’re in a crisis situation.”

It’s a crisis, he said, because the amount of electricity the farm is providing suddenly dips. And since customers can’t have their electricity dropping without warning, there needs to be a backup in place. That usually comes from traditional power sources — coal, gas, hydroelectric, nuclear — humming along on standby, wasting energy, ready to pick up the slack. Failing that, they would have to buy electricity from another utility on the spot market, which can be very expensive.

An accurate sunlight forecast, however, could allow utilities to rely more confidently on solar by allowing them to power down their backup plants when the chance of random clouds is at a minimum. Ordinary weather forecasts give only a general idea of how clouds will behave, and that’s fine for most purposes. As Drobot said, “If solar radiation reaching the ground drops by 10 or 15 percent over the next 20 minutes, it doesn’t make a big difference to most people.”

But to a solar plant, it does matter, so, armed with a $4.1 million grant from the Department of Energy, the lab, along with a consortium of government labs, universities, commercial weather forecasting firms and power utilities, is putting together a prototype system that should be a lot more precise. Each forecasting station will use three sensitive, wide-field cameras known as “total sky imagers,” that will monitor the sky in all directions, noting the height and motions of clouds from horizon to horizon.

The system will also bring in satellite imagery and ground-based lidars — a kind of radar that uses laser beams rather than radio waves — to gauge the thickness and altitude of clouds directly overhead. All of that information will then be fed into computers armed with weather models and artificial-intelligence programs that can predict what kinds of clouds are heading for what locations on what timetable. “Yes, it’s technically difficult,” Drobot said. “But within a year, 14 months, tops — we want to have a functioning system in place.”

Initially, that system will consist of one installation each in the Northeast, Florida, the Southwest and California, each of which has different types of weather systems and cloud patterns. If it works, the technology will be offered to utilities and meteorologists.

It would probably be too expensive to have individual sun-forecasting systems on location at every solar farm and racetrack and anywhere else that could benefit from the forecasts. Drobot said part of the testing will involve figuring out how close the forecasting installations have to be to each other to give accurate forecasts.

None of this will allow utilities to rely entirely on solar power, of course. Cloudy or not, you can’t trap the Sun’s energy at night. And having better forecasts won’t prevent clouds from coming along. But knowing when they’re coming and when they’re not could save utilities enough money to be more willing to invest in solar.

It’s already been shown to work, in fact, for another carbon-free energy source. A year or two ago, NCAR put together a program to create accurate wind forecasts — saving Xcel Energy some $6 million in a single year in operating the company’s wind farms. Hard-nosed businesspeople might not switch to wind or solar purely for the good of the planet. If the costs were comparable, however, that could make all the difference.