Gaming the Odds on Sunshine in Nevada
NEXUS scientists develop weather prediction tools to maximize the use of solar energy
By Jane Palmer
August 1, 2016
Satellite Image of Cloud Shadow — Image from “Physically-Based Satellite Methods” by Steven Miller et al., Chapter 3 in “Solar energy forecasting and resource assessment”, book edited by Jan Kleissl)
Peering into the Crystal Ball of Clouds
On a typical summer day in Las Vegas, the atmosphere is so dry that if a scientist could condense the entirety of the moisture in the air into a glass, the water would only be 1 centimeter deep, Wilcox says. But every two weeks or so, the North American monsoon effect carries moisture from the Gulf of California into southern Nevada. “When one of these episodic events occurs you get half a day of clouds,” Wilcox says.
Numerical weather prediction models can determine when one of these weather events will arrive up to five days in advance, but these same models can’t predict when a particular cloud will move in between a solar panel array and the sun. Typically during these times, the amount of sunlight reaching a panel can vary dramatically over very short time scales.
“You can see an increase and then a subsequent decrease of four or five times magnitude in terms of Watts per meter squared in just a matter of 15 minutes,” Wilcox says. “So it is these kinds of short time scale surges that create large fluctuations in voltage and power that we’d like to anticipate.”
To do this, Wilcox and his graduate student at the University of Nevada Reno, Marco Giordano, have built a prototype sky-imaging camera that takes images of the sky in the vicinity of solar PV arrays. The weatherproof camera takes the pictures and then analyzes them to distinguish cloudy pixels from clear sky pixels. Using this information, a computer algorithm can then track the movement of a cloud and predict when it will shade the PV array.
The idea to use sky imaging is not a new one, Wilcox says. But current commercial cameras cost about $15,000. “Our goal was to see if we could do this with the components not exceeding about $100,” Wilcox says. “This is what we have done with cameras jury rigged in the field and what we have now done with a more operational tool.”
The low cost of the tool opens up a wealth of possible applications for its use, Wilcox says. The scientists could deploy the instruments at distributed solar PV sites in the city of Las Vegas and develop a shared database of sky images. This database could be used to refine the algorithms that predict the cloud movements.
Left image shows the sun is visible. The image on the right shows the sky two minutes later when the sun is partially obscured. This is exactly the type of scenario that the researchers are trying to anticipate.
–Images from camera mounted on the top of the UNR physics building by Pat Arnott
Toward a More Reliable Source of Solar Energy
The team’s ultimate goal is to investigate how the forecasting tools can be integrated into the management of the electrical grid that distributes power from solar panels to consumers. “That’s where the collaborative aspect of the NEXUS project comes into play because I’m not an engineer, but there are a lot of electrical engineers on this project,” Wilcox says. “So by having access to those aspects of this project, I can start to ask the broader question of: If you have this information, how could it be used to provide a more reliable source of solar electricity?”
The scientists’ first steps in this direction involve integrating the forecasting technology into the newly developed University of Nevada Las Vegas (UNLV) microgrid. NEXUS scientist Dr. Yahia Baghzouz
at UNLV is currently testing the technology for its ability to smooth out variations in solar power output to the electricity grid.
“It is exactly these variations that we hope to anticipate by developing forecasting tools,” Wilcox says. “So the microgrid is the natural place to see how we can combine forecasting technology with other smart grid technology with the goal of increasing the reliability of solar power on the electric grid.”