The solar industry is rapidly expanding, providing enough clean electricity to power millions of homes across the country. Still, solar variability — fluctuations in solar power output based on the availability of sunlight — presents a challenge for the continued expansion of the industry. Some aspects of solar variability (think sunrise and sunset) are simple to predict but other aspects — most notably cloud cover — are much less so. Unanticipated changes in solar power plant output as a result of solar variability can strain the electric grid — a key concern for grid operators.

That’s why Sandia National Laboratories’ Matt Lave and Josh Stein — along with University of California, San Diego professor Jan Kleissl — developed the Wavelet Variability Model: a simulation tool to more accurately predict solar variability throughout the day. The Wavelet model is particularly useful for helping grid operators estimate how much energy must be stored to make up for cloud-caused fluctuations in solar power output. Their research is supported by funding from the Energy Department’s SunShot Initiative, which announced today $15 million to help communities boost solar deployment across the U.S.

So, how does it work? The Wavelet model uses data from pyranometers — devices that measure available sunlight — and scales it up to accurately represent the entire power plant. For its simulations, the Wavelet model incorporates several different data points — including the arrangement and number of solar panels in the plant and the daily local cloud speed. In testing — both at a 2-megawatt (MW) photovoltaic plant in Japan and the 48 MW Copper Mountain solar facility in Nevada — the model’s simulations have proven to match up well with actual conditions. Now that the testing and validation phases are complete, the researchers are preparing to make the Wavelet model publicly available.

To learn more about solar variability and the Wavelet Variability Model, check out Clearing up cloudy understanding on solar power plant output from Sandia National Laboratories.