SOLAR TODAY®
SEPTEMBER/OCTOBER 2011
VOL. 25, NO. 7
Though conditions at the location of any given
photovoltaic system can be highly variable, empirical
evidence demonstrates that the aggregated output of
several locations mitigates short-term variability.
By R. PeRez and T. e. Hoff
Figure 1 10-second irradiance at one location in Napa, Calif., on Nov. 21, 2010
richard Perez is a research professor at the University at Albany’s Atmospheric Sciences Research
Center. He sits on the advisory board of the
G W Solar Institute at George Washington University
in Washington, D.C., and has served multiple
terms on the board of the American Solar Energy
Society (ASES). He has produced more than 200
journal articles, conference papers and technical
reports and holds two U.S. patents on methods
of load management using photovoltaics. Perez
holds a doctorate in atmospheric sciences from the
University at Albany.
10-second irradiance averaged across 25 locations in Napa, Calif., on Nov. 21, 2010
Tom Hoff is the founder of Clean Power Research
and president of its research and consulting group.
Hoff assists Clean Power Research in pursuing its
mission of powering intelligent energy decisions by
taking an analytical approach to solving problems.
Hoff has published extensively for more than 25
years. He began his career at Pacific Gas and Electric Co. He holds a Ph.D. from Stanford University’s
School of Engineering.
Short-term variability has two components:
One is predictable, and the other is not. The
predictable component is caused by solar
geometry — the sun’s apparent motion in the
sky induces changes in the resource. Predictable changes are generally not noticeable for
very short time intervals (seconds to minutes)
but become influential when the time interval
extends beyond tens of minutes, particularly
near sunrise/sunset. In this article we focus on
the non-predictable — random — part of variability caused by the motion and evolution of
cloud fields.
Short-term variability is relevant to the
operation of solar energy systems and their
impact on the power grids to which they are
connected. As suggested by the rapid change
in irradiance in figure 1, a cloud passing in front
of the sun may cause a small photovoltaic (PV)
installation to go from full production to no
production and back to full production in a
matter of seconds. Grid operators are very con-
cerned with this sort of variability.
Acknowledgements: The authors wish to thank
Sergey Kivalov for his contribution to the analysis underlying the results discussed in this article.
Portions of this study were funded under a California
Solar Initiative grant agreement titled “Advanced
Modeling and Verification for High Penetration PV.”
The California Public Utilities Commission is the
funding approver, Itron is the program manager
and the California investor-owned utilities are the
funding distributor. The authors also wish to thank
Dave Renné of the National Renewable Energy
Laboratory for reviewing this article and offering
constructive suggestions.
