My initial effort to write this blog ended up way too long and way too detailed. If you want that level of detail, read my post Peak Power Generators and Grid Reliability instead of this post.
Electricity demand fluctuates dramatically throughout the day. When wind turbines slow or clouds pass over solar panels, grid operators deploy specialized power plants to prevent blackouts. Let’s dive into peak power and in particular how it handles intermittent generators.
Peak electricity demand occurs when consumption spikes beyond baseline levels, typically during mornings as businesses start operations, evenings when households activate appliances, and extreme weather events requiring heating/cooling.1 In 2021, peaker plants accounted for 19% of U.S. power capacity but only 3% of annual generation—highlighting their specialized role.2
Grid operators must ensure there’s enough capacity to handle these spikes without causing blackouts. This requires not only forecasting demand accurately but also having flexible generation resources ready to step in at a moment's notice.
The Options
When renewable sources like wind and solar falter—due to calm winds or cloudy skies—the grid turns to conventional power plants to fill the gap. The choice of generator depends on several factors, including cost, availability, environmental impact, and how quickly the plant can respond. Here’s a breakdown of common peak power sources:
1. Natural Gas Turbines
Natural gas is the go-to resource for many grids because it’s relatively clean compared to coal, abundant, and highly flexible. Two primary types of natural gas plants are used:
Simple-Cycle Gas Turbines (SCGT): SCGTs operate much like jet engines, burning natural gas to spin a turbine directly connected to a generator. They can start producing electricity in minutes, making them ideal for sudden spikes in demand. However, their efficiency is lower than other options, typically around 30–40%.
Combined-Cycle Gas Turbines (CCGT): CCGTs add a secondary steam cycle that captures waste heat from the initial combustion process, boosting overall efficiency to 50–60%. While more efficient, they take longer to ramp up, usually 30 minutes to an hour, so they’re better suited for sustained increases in demand rather than instantaneous spikes.
Grid operators prefer SCGTs for short-term needs and CCGTs for extended periods of elevated demand. For example, if wind turbines slow down unexpectedly, SCGTs might be deployed first while CCGTs come online later to maintain stability.
2. Coal Plants
Coal-fired power plants are less flexible than natural gas facilities, taking hours to reach full output. Their inefficiency and higher carbon emissions make them less desirable for peak load management. Nevertheless, some regions still rely on coal during prolonged high-demand events, especially where gas infrastructure is limited.
3. Hydroelectric Power
Hydropower offers unparalleled flexibility. Reservoir-based hydro plants can adjust output almost instantly by releasing water through turbines. Pumped-storage hydropower (PSH), a form of energy storage, pumps water uphill during low-demand periods and releases it downhill to generate electricity during peaks. PSH is widely regarded as one of the most effective solutions for managing daily fluctuations.
4. Batteries
Battery storage systems have gained traction in recent years, particularly for addressing the "duck curve" associated with solar power. During midday, excess solar energy charges batteries, which then discharge in the evening when solar production drops off. Lithium-ion batteries dominate this space due to their fast response times and declining costs.
Efficiency and Emissions Tradeoffs
Ramping power plants up and down comes with tradeoffs in terms of efficiency and emissions. There’s trade-offs for each technology:
Gas Turbines: SCGTs lose efficiency when operated below full capacity, leading to higher fuel consumption per unit of electricity generated. Similarly, frequent cycling increases wear and tear, raising maintenance costs. CCGTs mitigate these issues somewhat but are less agile than SCGTs.
Coal Plants: Cycling coal plants is expensive and inefficient, as they require significant energy just to stay warm and ready. Additionally, partial-load operation increases CO₂ emissions per megawatt-hour produced.
Hydropower: Hydropower suffers minimal efficiency losses during ramping, though reservoir levels and seasonal water availability constrain its flexibility.
Batteries: While batteries don’t emit CO₂ during operation, their lifecycle emissions depend on manufacturing processes and grid mix during charging. They excel in rapid response scenarios but face limitations in duration and scale.
Decision-Making in Real Time
Grid operators generally follow a merit order during shortages:
Demand response: Industrial users reduce consumption3
Battery discharge: 0-4 hour gaps
SCGT activation: 15-minute to 4-hour needs
CCGT/Hydro: Multi-hour deficits
Emergency imports: Neighboring grids4
Solar/Wind: At times the BA will take solar or wind offline, then later back online5
When Texas faced a 2023 wind drought, SCGTs provided 42% of peak capacity versus 12% from batteries.6 Conversely, California now uses batteries to shave 5GW off evening peaks—the equivalent of 10 SCGT plants.7
Ultimately, decisions hinge on local constraints, such as available infrastructure, regulatory policies, and market dynamics. Advanced forecasting tools and real-time data analytics increasingly guide these choices, ensuring optimal performance under varying conditions.
But the bottom line is that until we get widespread nuclear or much cheaper batteries, SCGTs are with us as the go to for intermittent sources. Even for something as predictable as the duck curve, without batteries, the interval is too short for CCGTs. So again, SCGTs are the go to provider.
General Resources
Historically not many industrial customers are willing to do this
A BA operator shared with me that when he does take wind offline the wind operator always replies “but we’re wind.”
Very well done - a couple of points:
"Hydropower offers unparalleled flexibility." Actually, not as much as you might think - hydropower has limits to what it can discharge and when, based on requirements for instream resources like fish. You can ramp, but within strict limits.
A second issue - we already have spikes in demand due to consumers. What wind and solar do is add supply spikes. And one problem makes the other worse, so it isn't arithmetic, it is a logarithmic increase in the difficulty in management.
What about some combination solutions--say nukes with pumped hydro? I know the TVA has had success with that.
I'm surprised the Army Corps of Engineers and Georgia Power/SoCo haven't done more with that in Georgia, although both have done some. There are two Georgia Power lakes and three USACE lakes that aren't far from Vogtle.
Nukes are probably the worst for adjusting to demand except for coal.