Cloud Computing: NREL Studies Solar Vs. Clouds

A solar power array needs one very large, very obvious thing to operate: the sun. Since they cloak the rays of the sun, clouds are the enemies of of large scale solar systems, thieving efficiency, and causing surges, fluctuations, and headaches for both the power grid and the utility operator.

Now those pesky clouds may have met their match due to a recent study conducted by the U.S. Department of Energy’s (DOE) National Renewable Energy Laboratory (NREL). The (NREL) has produced a massive data set showing what happens, second-by-second, when clouds pass over a solar power installation.

image via Shutterstock

The study utilized 17 measurement stations near Hawaii’s Honolulu International Airport on the island of Oahu and collected data at one-second intervals over the course of an entire year. The collected data “allows us to set up a solar-monitoring network that simulates exactly how clouds would impact a large photovoltaic (PV) system,” said NREL Senior Scientist David Renne. “The time-synch data are unique. All 17 stations make a one-second measurement at exactly the same time. This allows the array to “see” clouds moving through and simulates how a PV system might behave.”

By understanding the characteristics of cloud shadows that pass across a large PV system, utility officials can devise strategies to better manage those fluctuations so the grid isn’t adversely impacted. The data set is of great interest to utilities, developers of large-scale  PV systems, forecasters, system operators, laboratories and universities.

Steve Duda lives in West Seattle, WA with three dogs and a lot of outdoor gear. A part-time fly fishing fishing guide and full-time writer, Steve’s work has appeared in Rolling Stone, Seattle Weekly, American Angler, Fly Fish Journal, The Drake, Democracy Now! and many others.

  • Bill

    There are commercial plants installed in Spain and Germanyu00a0that record the output of each Array string each second. The NREL could have used actual array data instead of messing around calculating it! It is hard to imagine how they would haveu00a0accurately estimatedu00a0the spike in production that comes after the array is allowed to cool down as a cloud passeses over. Real world data is much better.