(탄소배출 없는 기저부하 전력)
What Is “Baseload?”
by Amory B. Lovins
“Baseload” has at least five prevalent meanings. Three are used by electricity professionals, who seldom distinguish clearly enough between them, while the other two are widespread urban myths:
1. For someone who analyzes utility loads: the steady (8,766-hours-per-average-year) portion of demand, below the shoulder of the load-duration curve
2. For the system planner: the least-levelized-cost resource that can be planned and built (regardless of unit size)
3. For the system operator: the lowest-marginal-operating-cost resource, dispatched whenever available
4. For the journalist, politician, and layperson: a gigawatt-scale thermal power plant (typically coal or nuclear), but may also include a big hydroelectric dam
5. For the nuclear advocate (who wants to exclude variable renewables by definition and confine choices to coal vs. nuclear to distract attention from the market's rejection of both): a hypothetical, though nonexistent, “24/7” power plant that runs at all times
These definitions are not compatible. For example:
• A least-cost [supply-side] resource can be most-cost to run, e.g. gas combined-cycle
• A most-cost [supply-side] resource can be least-cost to run, e.g. photovoltaics
Also, big thermal plants can have low operating cost but also limited operating flexibility that can impose significant cost on other resources (e.g., wind that must be spilled if nuclear has a must-dispatch privilege; that’s an income transfer from wind to nuclear operators that should be compensated). Nor do big thermal plants necessarily run steadily: many combined-cycle gas plants run with 20- to 40-odd percent capacity factor despite their high technical availability, because at high gas prices, their operating cost exceeds that of old coal and nuclear units.
Baseload plants in sense 5 do not exist. All kinds of generators are intermittent, varying only in the size, duration, frequency, cause, and predictability of their outages. Even normally reliable big thermal plants do have both planned and forced outages. For example, NERC data for all U.S. generators during 2003–07 show that coal capacity was shut down an average of 12.3% of the time (4.2% without warning); nuclear, 10.6% (2.5%); gas-fired, 11.8% (2.8%). Nuclear units’ cumulative worldwide forced outage rate was 6.4% through 2008.
For over a century, utilities have routinely used controls and grids to combine intermittent resources, including sufficient reserve margin, so as to ensure supply/demand balance despite fluctuations in both supply and demand. RMI’s hourly simulations show that a typical system dominated by big thermal plants can accommodate large amounts of variable renewables simply by using them, when producing, to ramp down coal and plant plants within normal ramp-rate constraints and without system planning or stability issues. This follows orthodox principles of economic dispatch (per sense 3), and merely requires existing assets to be operated differently than they were before nearly-zero-marginal-operating-cost resources became widely installed and attractively priced in sense 2.
End-use efficiency is a baseload resource in senses 2 and 3. It can even be considered a baseload resource in sense 1, because it automatically dispatches itself whenever the service demand that it reduces is being served. (End-use efficiency and demand response are also self-“insuring” resources in that if the service demand expands, e.g. because you add a factory shift or extend a store’s opening hours, the savings automatically expand in proportion to the load.)
However, a load that looks steady may actually be the sum of multiple separate loads that are not steady: even if some occur while others don’t, they can add up to a steady apparent load. Many of what appear to be baseload demands in sense 1 do not in fact run continuously; only their aggregate does. That is, sense 1 has different meanings at the level of utility dispatch than it does at the level of individual customers or end-use devices.
Why worry about all these disparate definitions? Because much nonsense is being published about how variable renewables (wind and photovoltaics) cannot contribute significantly to reliable electricity supply because they aren't “baseload” in sense 4 or 5. In fact, that doesn’t matter, because:
A. Baseload (steady) demand does not require a steady generator (that’s good, because there’s no such animal).
B. Steady output is thus a statistical attribute of the aggregate of generators on the grid, not a physical requirement for any single generating unit.
That is, baseloads in sense 1 do exist. They are met by a mix of resources chosen according to sense 2 and run according to sense 3. The resulting supply-side portfolio need not include sense 4—it may, but whether it will depends empirically on competition under senses 2 and 3—and cannot include sense 5.
Seattle City Light has no generating resource with a capacity factor above ~25%, equivalent to a pretty poor wind turbine. Yet it provides cheap and reliable electricity, serving all loads—baseload (sense 1), intermediate-load-factor, and peaking. If the people who don’t understand points A and B were correct, this would be impossible.
* * *