By Kevin Wagter
The August 21, 2017, eclipse presented a unique opportunity to study how Distributed Energy Resources (DER), such as solar PV, affect the electric distribution grid.
To observe the effects that the eclipse had on the U.S. grid, we relied on our PQube 3 power analyzer and microPMU devices. The PQube 3 and microPMU are small enough to fit in the palm of your hand, but capable of measuring thousands of electrical power quality parameters. We wanted to know how well the nation’s grid operators were able to compensate for renewable energy’s intermittent nature as the eclipse crossed the U.S. from Oregon to North Carolina.
Two grid parameters help grid operators match the load/generation balance: diversity factor and predictability. The eclipse created a unique situation where significant solar power loss happened simultaneously on a large scale, reducing the diversity factor. Predictability was also reduced due to the lack of knowledge of the exact location and output of residential and other small solar arrays. If these factors interfered with a grid operator’s ability to match loss and generation exactly, the frequency of the grid – which must remain at a steady 60 cycles – would have been impacted.
We set up data collection sites in Windsor, California, and Zebulon, North Carolina, using our PQube 3 and microPMU devices. We measured several parameters, both from the grid and from the environment affecting the grid, including “solar irradiance,” an indicator of the amount of photovoltaic (PV) power available, grid frequency, and ambient local temperature. California and North Carolina were selected for the study because they are the states with the greatest reliance on solar power and are part of the Western and Eastern Grids. California has more solar power generation than North Carolina, but only experienced a 50% to 70% loss of irradiance during the eclipse. North Carolina relies less on solar, but lost over 90% of its solar power generating capacity.
PQube 3’s real-time output gave macro scale data that allowed us to analyze the effect the eclipse had on the grid, and how well grid operators responded. The graphs in figure 1 show key data for North Carolina. The first graph tracked the loss and recovery of solar irradiance caused by the eclipse. The spiked drops are clouds passing in front of the sun, and the large “V” shape represents the eclipse. The second graph measured the changes in frequency experienced by the grid as grid operators juggled the loss of PV generation with on-demand generation. The general evenness of the graph is an indicator that they did their job well. The third graph shows a significant temperature drop that closely tracks the solar irradiance. The lowering of ambient temperature probably helped grid operators by reducing the load caused by air conditioners.
Our microPMU’s ability to collect data down to the millisecond range provided a microscope-level view of voltage magnitude and phase angles, and frequency variations (red trend line in Figure 2) as grid operators compensated for the loss of solar power. The slightly higher frequency during the first half of the eclipse might indicate that the grid operators were being cautious by maintaining slightly more generation than absolutely necessary. The frequency dropped along with solar irradiance and rose again (blue trend line in Figure 2), likely due to reduced air conditioner use.
Our conclusion was that operators on both the Western and Eastern Grids did an excellent job of maintaining frequency stability during the eclipse. They preceded the event with slight excess generation, and compensated well as solar power generation declined.
Power Standards Lab
980 Atlantic Ave.
Alameda, California 94501
US: 1.888.SEMI.F47 or 1.888.736.4347