Wednesday, 31 January 2024: 11:30 AM
347/348 (The Baltimore Convention Center)
Nearly three weeks prior to the total solar eclipse on August 21, 2017, Maxar predicted that national clear-sky solar power production during the peak of the eclipse would be reduced to as low as ~12 gigawatts (GW), approximately 46% less than the typical potential generation of 22 GW. In the six years between 2017 and 2023, total installed solar capacity in the U.S. has more than doubled, and grid operators have become increasingly reliant on timely and accurate solar power forecasts to properly balance variable and dispatchable energy sources. Six months prior to the October 14, 2023, annular solar eclipse, Maxar estimated that clear-sky solar power production in the Electric Reliability Council of Texas (ERCOT) regional transmission organization (RTO) would be reduced by 87% during the peak of the solar eclipse, with a maximum ramp rate of 180 megawatts per minute. This presentation extends that clear-sky analysis of solar power production in Texas to include predictions of all-sky solar power production with and without the effects of the eclipse for seven U.S. RTOs. We will quantify the reduction in utility-scale solar power due to the eclipse in forecasts leading out 15-days, and present actual solar power reductions attributed to the eclipse and the forecast accuracy. Beyond the forecast and its accuracy, this presentation will also highlight the impact of this event on grid operations for ERCOT, an RTO with high solar penetration, and show the inherent value of physics-based solar forecasting systems for modeling infrequently observed events, which may be poorly predicted in purely machine-learning based models.

