Solar - Curtailment Affecting Capacity Factor

Separately, we analyzed the capacity factor trend for solar PV power plants and found the median overall fleet annual degradation rate was -1.5%. Several trends were as expected: sunniness and high temperatures accelerate panel degradation and tracking systems with more moving parts have higher degradation. A third factor, curtailment, is analyzed here.   

As more solar power plants are added to the grid, the system reaches a point where there is too much solar production. When a less expensive plant has to curtail to make room for a more expensive plant something seems amiss, and historically it was. However, solar is a unique type of power plant since it very predictably will be ramping down as the sun goes down. The other types of power plants that will need to ramp up later in the day usually run at minimum output during the middle of the day. But, the minimum output level can actually be fairly high due to technical and permitting constraints. (See Duck Curve)

This oversupply is a foreseeable grid situation, and the California ISO has been talking about it since 2012. Once a grid is oversupplied, solar plants are curtailed to immediately address the issue. If even more solar is planned to be added to an oversupplied grid, better solutions will be needed such as grid energy storage, bidirectional home and car charging, transmission connections to neighboring grids, shifting load patterns, etc. However, what is not well understood is that some level of solar curtailment is optimal from a ratepayer perspective. Ideally, policy makers and regulators in each grid area would understand where they are in the cycle from no curtailment to optimal curtailment to excess curtailment to make better decisions for ratepayers and constituents.

Likewise, project developers may make better decisions with more insight. For example, once a grid operates with some solar curtailment (which initially is seasonal, spring and fall) new solar plants could be optimized to produce power during summer and winter which would be a relatively minor design change from current designs which maximize production all year.

In this discussion, I am using curtailment broadly to also include congestion which is similar but on a localized basis often due to poor site selection within the grid.

Even though curtailment is an expected outcome of PV power plant build out and some level of curtailment is actually optimal, there is a lack of analysis where each grid stands relative to an optimal level of solar curtailment. LBL takes the opposite approach and adjusts its data set to add back curtailed energy for California and Texas. Before figuring out the optimal level of curtailment, a basic understanding of solar curtailment levels is needed.  

I tabulated the total solar PV MWs added from 2020-2024 for each state and ranked the states. I then compared the rankings with the median annual capacity factor (CF) decline rate rankings. Of the 7 highest states with the highest solar MWs installed, 5 are in the top 7 for worst CF decline and another is in the top 10. This is not a statistically robust analysis, but an indicator that there is quite a bit of curtailment going on outside of TX and CA.

Florida is an interesting example. It’s median solar capacity factor trend was -3.3%, the worst of any state. It also had the 3rd highest solar MWs added from 2020-2024. I looked more closely at a fleet of about 100 solar power plants each rated at 74.5 MW PV that Florida Power & Light has built over several years (with more coming along, as well as batteries). 48 of those plants have enough operating history for this analysis and the median CF annual change was -3.99%. It seems highly likely the reason for the worse than average CF trend is curtailment. Will FP&L hit the optimum level of curtailment for its ratepayers before the current build out plan is realized?

While curtailment may be significant in the data set, it’s impact on the CF trend could be muted if it occurs uniformly, for example if it occurs to the same degree in Year 1 and Year 2, it would not show up as a declining CF trend.

Since this analysis is based on 2 full years of operating data, the analyzed plants came on line in 2022 or earlier. ERCOT added large amounts of solar power plants in 2023-2025, so the high levels of curtailment are expected to go even higher, but offset somewhat by high ERCOT load growth.

California has added large amounts of storage which helps keep their curtailment lower than it otherwise would be. But, it is likely the amount of storage is beyond the optimal ratepayer level since California has close to the highest electric rates in the US.

Policy makers need to evaluate curtailment against its alternatives such as modifying other existing power plants to enable lower minimums, adding batteries to the grid, enabling real time energy rates to shift customers usage patterns to align with periods of excess supply, adding transmission lines, etc. Adding storage (batteries) to the grid is necessary in a future scenario where the vast majority of electricity supply is intermittent (solar and wind), but in the interim a combination of all alternatives may be optimal.

More discussion and analysis of solar curtailment would help PV power plant developers and investors in those projects to forecast the inevitable capacity factor degradation that will occur when large numbers of solar power plants are added to the grid.