Generation Technologies
Combined cycle, peakers, coal, nuclear, hydro, wind and solar, and the six numbers that decide when each one runs: heat rate, capacity factor, ramp rate, minimum run time, minimum load, start cost.
Words the industry uses are marked like this at the sentence that defines them, so you can tell a term you will hear on a desk from a phrase this book happens to be using.
The supply stack in Chapter 8 is drawn as a neat staircase of offers sorted by cost. The whole set of machines a system can call on is its fleet, and each machine in it is a unit. Underneath each step is a physical machine with a personality: a speed at which it can change output, a floor below which it cannot run, a period after shutdown before it may start again, and a bill that arrives simply for starting.
Six numbers describe almost everything a market needs to know about a generator. They come as a set, because the arguments in later chapters are usually arguments about one of them.
Table 4-1: The six numbers, and what each one settles
| Number | What it measures | What it decides |
|---|---|---|
| Heat rate | Fuel energy burned per unit of electricity produced | Marginal cost, and so the plant’s place in the dispatch order |
| Capacity factor | Share of the year spent at full output | What kind of business it is, since a plant running 13% of the year and one running 64% are not the same trade |
| Ramp rate | How fast output can change, in megawatts per minute | Whether it can follow the evening ramp of Chapter 7 |
| Minimum stable load | The floor below which it cannot run steadily | How much room it occupies in the stack that it cannot vacate |
| Minimum run time and minimum down time | How long it must stay on once started, and off once stopped | Whether it can be started for a single tight hour at all |
| Start cost | Fuel and wear consumed getting from cold to useful output | The side payments of Chapter 9 that no single price can express |
What heat rate decides, and where it decides nothing
Heat rate is the fuel energy required to produce a unit of electrical energy, conventionally expressed in British thermal units per kilowatt-hour. Since one kilowatt-hour equals 3,412 Btu (pronounced “B T U”) of energy, dividing 3,412 by the heat rate gives thermal efficiency directly.
Table 4-2: Typical American heat rates and what they imply
| Technology | Heat rate, Btu/kWh | Thermal efficiency |
|---|---|---|
| Modern combined cycle | around 6,960 | roughly 49% |
| Simple cycle turbine | around 10,000 | roughly 34% |
| Coal steam | around 10,018 | roughly 34% |
| Nuclear steam | around 10,443 | roughly 33% |
For a nat gas plant the heat rate is the most consequential number in its commercial life, because its marginal cost is essentially the heat rate multiplied by the nat gas price. A plant at 7,000 Btu/kWh burning nat gas at four dollars per million Btu has a fuel cost of about 28 dollars per megawatt-hour ($/MWh), and a plant at 10,000 has $40/MWh. That difference decides which runs, and Chapter 23 turns the same relationship into a traded instrument.
The nuclear row is there to show where heat rate stops deciding anything. A reactor has a thermal efficiency similar to a coal plant, and the number means almost nothing economically, because uranium contributes a trivial amount to the cost of a nuclear megawatt-hour. Heat rate governs dispatch only where fuel is a real marginal cost.
US averages, EIA. Two charts rather than two axes on one chart, since the units do not compare.
The other five numbers
Capacity factor is output over a period divided by what the machine could have produced running flat out. It measures how a plant is used rather than how good it is. Combined cycle units built between 2010 and 2022 averaged around 64%, while simple cycle turbines averaged near 13%, and that gap describes two entirely different businesses using similar fuel. A machine run that rarely, kept on the system for the hours when everything else is already flat out, is what the trade calls a peaking plant or a peaker, and later chapters use those words.
Ramp rate is how fast output can change, in megawatts per minute. Minimum stable load, usually shortened to minimum load, is the floor below which a unit cannot operate steadily, which matters because a plant at its floor is occupying room in the stack it cannot vacate. Minimum run time and minimum down time are the periods a unit must stay on once started and stay off once stopped, both set by thermal stress on heavy components. Start cost is the fuel and wear consumed getting from cold to useful output, and it is the reason Chapter 9 has to discuss payments that no price can express.
Table 4-3: The fleet, by temperament
| Technology | Marginal cost | Flexibility | Role in the stack |
|---|---|---|---|
| Wind and solar | Near zero, sometimes below | Down only, and only if permitted | Bottom. Sets price only when the price collapses |
| Nuclear | Very low | Poor, and expensive to cycle | Bottom. Runs continuously |
| Coal | Moderate | Slow. Hours to start, long minimum times | Middle, and shrinking |
| Combined cycle nat gas | Moderate, tracks the nat gas price | Good | Middle. Usually the marginal unit |
| Simple cycle nat gas | High | Excellent. Minutes to full load | Top. Runs rarely, at high prices |
| Hydro with a reservoir | Near zero, but with an opportunity cost | Excellent | Wherever the operator decides the water is worth most |
The two that changed the shape
Wind and solar differ from everything above them in a way that goes beyond having no fuel cost.
Their output is set by weather rather than by instruction. They can be curtailed downward but never called upward, so an operator holding a forecast of 12 gigawatts of wind is holding a forecast rather than a resource. Their capacity factors are properties of the site: a good onshore wind project might run in the thirty to forty per cent range and a fixed solar array in the twenties, against numbers that reflect resource availability rather than any commercial decision.
They also connect through power electronics rather than through a spinning mass locked to the grid, which changes what the system has available to defend itself in the first seconds after a disturbance. Chapter 6 takes up what that does and, more usefully, what it does not do.
Hydro with a reservoir
Hydro with a reservoir behaves unlike anything else in this chapter, because its fuel arrives on its own schedule and can be held.
A hydro operator has almost no marginal cost in the accounting sense and a very real one in practice, since water used this hour cannot be used in a more valuable hour later. Its offer therefore reflects an opportunity cost rather than a fuel bill, and it moves with the operator's expectations about the rest of the season. Norway, Brazil and the Pacific Northwest all run systems where the deciding question in any given year is hydrology.
Which makes reservoir hydro the closest thing to the storage every other part of this system lacks, and the reason Chapter 5 begins by asking what happens where the geography does not provide it.