Chasing after windmills and photovoltaics could well be the stuff of fiction.
Wind and solar cells (photovoltaics or PVs) are two renewable energy technologies that many hope will eventually provide the United States with massive amounts of clean, sustainable electric power for the indefinite future. Indeed, it is often suggested or implied that the United States can look to a future where most, if not all electric power can be provided by wind and photovoltaics [1, 2]. But careful analysis shows that will not be possible unless consumers are willing to pay five to 10 times what they pay for electricity today.
Both wind and sunlight are unpredictable and intermittent. The wind blows at different speeds at different times in different places. Sometimes the air hardly moves. Sunlight is available only in the daytime; it is weak in the morning and late afternoon and is dramatically reduced by cloudiness, which is location-dependent and unpredictable. That's nature, as we all know from our everyday experience.
On the other hand, the public requires electricity on demand, which requires dispatchable electric power generators. Lights must go on when the switch is flipped, the computer must start and operate steadily on demand, etc. Very few people would be willing to operate their homes or businesses only when the sun shines or the wind blows. So the question becomes one of how to accommodate the inherent mismatch between what we demand in electric power availability and what nature provides in light and wind, which are inherently not dispatchable [3, 4, 5].
Both wind and photovoltaics are currently too expensive in today's marketplace for widespread commercial application, so large sums are being invested in research and development and government subsidies. In the near-term, investors are building government subsidized and/or mandated wind farms that are capable of providing modest but useful amounts of electric power, which is absorbed on our large, complex, interconnected electric power grids. That arrangement has been demonstrated to be operable, but the economics are not what they appear to be, because the impact on the rest of the grid is not often calculated nor made public.
Investigating the Promise of Large-Scale, Baseload Solar Electricity
Consider a future time when wind and photovoltaics might conceivably provide a major fraction of the nation's electric power needs. Two major paths emerge that could provide large-scale electric power-on-demand. (By large-scale we mean a major fraction of U.S. electric power consumption-say many tens of percent). The first path involves the use of energy storage to bridge the gap between our requirement for power-on-demand and the intermittence of wind and photovoltaics [6, 7]. The second involves the use of wind or photovoltaics in combination with a conventional gas or coal power plant that together could supply electricity-on-demand when wind or sunlight diminishes or simply is not available.
In most of today's applications of photovoltaics, which are comparatively small-scale, energy storage (often batteries) is used to bridge the no- or low-sunlight gaps . Applications include the supply of electricity to remote locations for water pumping, vaccine refrigeration, communication, warning lights, beacons and weather monitoring.
These systems are costly for two reasons. First, electric energy storage costs are very high in spite of decades of research. Second, charging electrical storage requires additional photovoltaic cells beyond those that just meet load requirements under ideal conditions. This is because the output of photovoltaics is quoted on the basis of what they can provide under ideal circumstances, averaged over a long period of time.
But PV power is intermittent and unpredictable, so to bridge the gap to reliable power-on-demand, it is necessary to add many more solar cells to charge some form of energy storage. When one takes into account the need for electric power at night and during periods of cloudiness, the total costs mount dramatically-typically five to10 times or more than the costs quoted for a single unit  (see sidebar).
The analogy to wind is similar. Specifics vary according to location. Moreover, on an annual basis, the situation for photovoltaics is roughly twice as bad in New England, with its long periods of cloudiness, as it is in Nevada, with its generally clear skies . For wind, the situation is better in the upper Midwest than most other U.S. locations. (One wonders if politicians from these poorer wind or sun areas understand that their rush to renewables could result in much higher electricity costs for their constituents compared with people in superior sun/wind regions).
The Hybrid Approach Still Depends on the Unpredictable
The second path of interest is one in which a power plant stands by in spinning reserve to provide electric power when sunlight or wind is diminished [3, 4, 5, 7, 8]. In this situation, that power plant is actually the principal source of power-on-demand, so wind or solar cells act simply to save some fuel, when the wind blows or the sun shines. This mode of operation is then a "fuel-saver," because of the secondary nature of the renewable energy.
The problem is that across the United States, it is cloudy roughly half the time on average, and while the sun shines maybe eight to12 hours a day (winter-summer), it shines strongest from roughly 10 a.m. to 4 p.m. On that basis over an average 24-hour period, the sun might provide useful energy 20-25 percent of the time , so photovoltaics might "fuel-save" only about 20-25 percent of the fuel from a primary, power-on-demand power plant that otherwise is assumed to operate at 100 percent capacity in this analysis.
Proper accounting requires that system economics include not only the cost of the photovoltaics but also the capital and operating costs of the conventionally fueled power plant and the fuel for 75-80 percent of its operation. In this hybrid mode, the required cost of the associated photovoltaics system must be equal or lower cost than the avoided power plant costs (saved fuel and variable O&M), which will be quite low.
The same general situation holds for wind. On the positive side, the fractional contribution may be more like 30 percent in good wind locations . On the negative side, wind power can vary by factors of 10 or more on an hourly basis , which makes for a significant cycling challenge for an associated conventional power plant that must be capable of dramatic changes very rapidly.
Peak and Intermediate Power: Not Economical
Consider society's need for peak and intermediate electric power, which typically occurs during the day from early morning to late afternoon. Wind availability does not often correlate with that demand, so wind is not a good match for peak power use. On the other hand, PVs produce electric power during that period, which has led many to believe that PVs may be most useful for peaking. The problem is again the unpredictable intermittence of sunlight, which means that a conventional power plant must be available to provide power-on-demand when the PVs cannot. The most attractive power plant to have "fuel saved" by photovoltaics is a natural gas combustion turbine (CT), because of its low capital cost and non-trivial fuel costs.
Comparing capital costs between a CT and PV system is not simple. This is because a CT can produce power-on-demand under a broad range of conditions while the PV system is a daylight-only power system. Furthermore, the convention is to quote PV system costs in dollars per peak kilowatt, which is not easily comparable to CT power costs.
Current CT capital costs are roughly $400/kilowatt electric . Recent PV system costs are roughly $4000/kilowatt-peak . (The convention in photovoltaics is to designate module electricity production on a peak basis, because that is the simplest, application-independent descriptor). One simple way to transform PV peak power costs to the same basis as CT costs would be to multiply by a factor of four-a factor of two corresponding to daylight being a maximum of 12 hours out of 24 hours per day and another factor of two because year-around average cloudiness in the United States is roughly 50 percent. This yields a PV system equivalent capital cost of $16,000/kilowatt. However, in a peak and intermediate load application, we are interested in power for half a day or less, which would lower the equivalent PV cost to maybe $8000/kilowatt.
But PV systems have no fuel costs and low operations and maintenance (O&M) costs, which is not the case for CTs. Accordingly, it is necessary to do more detailed calculations that include utilization (percent of a day that the system is utilized), fuel costs, O&M costs, etc., to determine overall economics. Using GRI data  and a $4.00/MMBtu gas price, CT power costs are in the six to seven cents/kWh range for 30-40 percent utilization (seven to10 hours/day).
For peaking-intermediate duty of 30-40 percent of a 24-hour day, today's PV system costs would be of the order of 30-40 cents/kWh  or roughly five times more expensive than CT power costs. So to compete with CT power requires closing a large cost gap, even if photovoltaic systems costs were dramatically reduced and natural gas prices were to soar. And even if such dramatic cost reductions were possible, photovoltaics would only be fuel saving approximately 20 percent, which is far from the major contribution that many proponents hope for.
How does all of this add up? If the United States were decide to use wind and/or photovoltaics for a major fraction of U.S. electric power, the storage approach would be required. The resulting cost for electric power would then be five to 10 times or more expensive than quoted. If we were to use fossil-fueled power plants as the backbone of a "solar supplement" power-on-demand system, the contributions from wind or photovoltaics would be maybe 20-30 percent, which is far from the 100 percent sustainable renewables future that many were hoping for. Furthermore, the solar supplement approach will happen only with major cost reductions in photovoltaic and wind systems, along with investments in conventional power plants and controls that would allow hybrids to operate compatibly, considering the intermittent nature of wind and PV generation.
Returning to the issue of the low levels of wind-generated electricity being absorbed on today's power grid, this analysis demonstrates that the current situation is really one in which power plants on the grid are being throttled back to "make room" for the trickle of wind energy. Those throttle-backs cost us all and must be included in calculating the current cost of wind energy. If the calculations were done properly, the costs of current wind energy would be different than most people currently believe.
This whole matter is an important national issue, because a number of policy-makers are setting policy based on the expectation that wind and photovoltaics will provide a major portion of future U.S. energy supply because they are clean and will be inexpensive. What's needed is an unbiased, expert evaluation of the situation. In the view of this observer, the most believable evaluation would come from an independent committee of commercially experienced engineers empanelled by the National Research Council of the National Academies.
- Brower, M.C., et al. Powering the Midwest-Renewable Electricity for the Economy and the Environment. Union of Concerned Scientists. 1993.
- DOE. Wind Energy Information Guide. DOE/NREL. April 1996.
- EIA. Renewable Energy Issues & Trends 2000. DOE/EIA-0628(2000). February 2001.
- National Research Council. Renewable Pathways. National Academy Press. 2000.
- Johansson, T.B. et al. Renewable Energy - Sources for Fuels and Electricity. Island Press. 1993.
- Hirsch, R. Renewable Energy in Perspective. The Energy Daily. December 19, 1995.
- EPRI/DOE. Renewable Energy Technology Characterizations. EPRI TR-109496. December 1997.
- Zweibel, K. Harnessing Solar Power-The Photovoltaics Challenge. Plenum Press. 1990.
- Holtberg, P. D. et al. GRI Baseline Projection Data Book. Gas Research Institute. GRI-01/002.1. March 2001.
- Holtberg, P.D. Private Communication. June 2002.
Articles found on this page are available to Internet subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.