Appendix 12B: Scale in Power Systems

For about a century after the first power station was commissioned in 1882, skilled power engineers used hotter steam, higher pressures, better alloys, more sophisticated design, and higher voltages to double the unit size of the largest generating machines every 6.5 years through five orders of magnitude. Had this continued, around 2064 the largest unit would have produced eight billion kilowatts, or about what everyone used for everything (electricity and all other forms of energy) in the mid-1970s.

For obvious reasons, it didn't happen.1 Around 1960 big generating units stopped getting more efficient, and they stopped getting bigger around 1970 at around 1,200 megawatts. This proved to be roughly ten times the economic optimum, because the bigger units proved less reliable. Construction-cost economies of scale disappeared in the 1960s and reversed in the 1970s as gigantic scale brought longer lead times, greater complexity and conflict, less flexibility, and more financial risk. By optimizing for the wrong thing—projected capital cost per kilowatt of generating capacity, rather than observed whole-system cost of electricity delivered to retail customers—the industry nearly destroyed itself. Only a few utilities actually went broke, typically from violating Miss Piggy's Fourth Law—"Never try to eat more than you can lift"—but scores more had near-death experiences, and their balance sheets will take another decade or two to recover.

Worse, by interrupting for a decade the previously steady decline in real electricity prices since 1890, the industry created a political backlash now being played out—despite the resumption of falling real prices in 1983—in moves to restructure utilities. This is chiefly as a ruse to shuffle or ignore obligations incurred in the 1960s and 1970s to pay for now uncompetitive, mainly nuclear, stations. To be sure, the original proposals for restructuring, the "big dogs eat first" approach, is hardly being discussed any more. Of the modest number of states that are significantly restructuring their utilities, most of the leaders are preserving such public goods as farsighted R&D and cost-effective efficiency investments. They are continuing to reward utilities for cutting customers' bills rather than for selling more energy, and are thus fostering healthy wholesale competition without sacrificing the much larger potential benefit of helping customers to use electricity productively. Early returns from most restructuring experiments also suggest little interest among most customers in having a whole swarm of vendors make them "an offer they can't understand." With California participation so far around 1% despite a $100-million "public education" campaign, RMI's early characterization of the supposedly tsunami-like trend of "retail wheeling" as "Big noise on stairs, nobody come down" seems prophetic.2

Largely unnoticed, however, there was an epochal shift in the choice of power plants in the more competitive wholesale market. Combined-cycle gas-fired power plants with half the capital cost, two-year lead time, doubled efficiency, and railcar portability promptly grabbed half the utility market for new capacity. The era of ordering new central steam plants ended, at least in the U.S. Even more surprisingly, utilities' planned capacity shifted to units that were ten to twenty times smaller, reapproaching a unit-size distribution last seen in the 1940s. Nonutility generation, having fallen steadily from 60% of the nation's electricity in 1900 to 3% in 1980, rebounded twice as fast, returning to 12% by 1996. It also strongly emphasized hundred- or ten-megawatt rather than thousand-megawatt unit sizes. Even smaller units—plug-in manufactured products down to the kilowatts range—were rapidly approaching over the horizon.

The electric utility business got into its mess in large part because it was an engineering-and-accounting-dominated business. Now that new, often financially trained actors and increasingly competitive conditions are bringing new disciplinary perspectives to the utility industry it has become painfully obvious that million-kilowatt, billion-dollar, decade-to-build stations incurred costly diseconomies of scale. In a country where 75% of the households don't need more than 1.5 kilowatts of average power and 74% of commercial customers don't need more than 10, a million kilowatts is simply too much in one place, often hundreds of miles from its loads. Unnoticed, because it wasn't clearly shown in the utilities' statistics, the cost of building the grid—which had dominated utilities' construction budgets for most of the 20th century—and the cost of running it came to exceed the cost of generating the power for the grid to deliver. The U.S. grid's average delivery cost is about 2.4 cents per kilowatt-hour to the average customer in 1996. In short, people who bought and ran power stations and grids are starting to realize that a five- or six-order-of-magnitude mismatch between production and usage makes no more sense than in, say, the cola-can business. What matters, and what they had overlooked to their peril, was that even cheap power stations, if they were too big and too remote, cost too much when the rest of the whole power system is properly included.

In 1998, a new introduction to these scale issues systematically organized more than 500 studies published over the previous two decades, and included some sanitized versions of proprietary findings, to explore what factors determine the right size, how to calculate them, and what the calculations showed. Its findings were explosive. Properly counting approximately 75 documented and measurable diseconomies of scale, not just the few well-known economies of scale, will typically make decentralized ways to make, store, or save electricity around ten times more valuable than conventionally scale-blind comparisons had long shown. This would mean, among other things, that photovoltaics for making solar electricity are cost-effective now in most applications.

The scale effects that yielded this finding came in four main categories. The most important came from financial economics—the risk-assessing discipline that underpins the world's financial markets. For example:
  • Because small modules and short lead times reduce financial risk, if you're willing to pay $1,000 per kilowatt for a 50-megawatt generator that takes two years to come online, you should be equally willing to pay $2,700 per kilowatt for a 10-kilowatt generator that can be installed by next Tuesday. They'll both have identical risk/return financial performance.
  • Accounting costs can't fairly compare, say, a windfarm and a gas-fired power station, because the windfarm has no fuel, while the gas plant uses fuel whose volatile market price creates financial risks. To adjust for those risks, you should increase the calculated fuel cost by about half.
  • Several kinds of operational financial risk are reduced by many renewables' tendency to produce the most power when demand is highest, such as on hot summer afternoons, so their supply is more valuable.
Just these scale effects increase the economic value of small, fast, modular resources by the best part of a factor of ten.

The second-biggest economic benefits that decentralized resources offer typically come from electrical engineering, and relate to the costs, losses, operations, and repair of the power grid. These benefits collectively increase the value of renewables by about a factor of about 2–3. They include:
  • the smaller costs and losses of distributing electricity over shorter distances or no distance (from your roof, backyard, driveway, or basement);
  • the greater reliability of that shorter haul length;
  • the potential to defer or avoid costly expansions of the grid wherever it's already fully loaded;
  • similar avoidance of grid extensions to remote sites;
  • the longer life of utility lines, transformers, switchgear, etc. that gets relieved of some grid load so they can run cooler;
  • the higher power quality of many distributed resources, which can't pick up electrical noise from other customers, lightning, or drunks skidding into power poles;
  • the utility's greater flexibility in managing grid failures and rerouting power around them pending repairs;
  • inverter-driven distributed generators' ability, when needed, to make electrical current out-of-phase with voltage in a way that helps support the voltage and stability of the utility grid; and
  • such generators' ability to ramp up instantly to meet unexpected demand, thus reducing the need to maintain costly "spinning reserve" for this role in evening out fluctuations.
Further benefits come from many diverse terms, ranging from solar cells' shading your roof and thus reducing air-conditioning loads to their ability to bring power to remote sites at a cost below the discount that real-appraisers apply to land with no utilities; from their use of local waste fuels to their convenient provision of waste heat onsite. These many small terms are typically worth up to about a twofold increase in value. And finally, of course, there's the ability of efficiency and renewable sources to avoid environmental and social costs, which are hard to quantify but may be socially important and politically decisive.

1 Hirsch, Richard F 1989: "Technology and Transformation in the American Electric Industry," Cambridge University Press, New York.
2 Lovins, A.B. 1996: "Negawatts: Twelve Transitions, Eight Improvements, and One Distraction," En. Pol. 24(4), April, RMI Publication #U96-11,

To make comments or report problems with this site, please contact

© All rights reserved. Published by Rocky Mountain Institute.
2317 Snowmass Creek Road  |  Snowmass, CO 81654-9199  |  Ph: 970.927.3851


Powered by Intrcomm Technology's SMC