At least on a cumulative basis within the sample of projects reported, average wind power prices compared favorably to wholesale electricity prices from 2003 through 2008. The increase in wind power prices in 2009, however, combined with the deep reduction in wholesale electricity prices (driven by lower natural gas prices), reversed this long-term trend in 2009.
Although low natural gas prices are, in part, attributable to the recession-induced drop in energy demand, the discovery and early development of significant shale gas deposits has resulted in reduced expectations for increases in natural gas prices going forward.
As a result, natural gas prices may not rebound to earlier levels as the economy recovers, putting the near-term comparative economic position of wind energy at some risk (especially if wind energy costs do not also decrease.
Although there is quite a bit of variability within some regions, and several regions again have limited sample size, the spread between the average wind power and wholesale electricity prices (i.e., the wind power premium) in each region is fairly consistent across the United States, suggesting that the struggle for wind energy to compete in 2009 on short-term economics alone was indeed a nationwide phenomenon.
Notwithstanding the comparisons made, it should be recognized that neither the wind nor wholesale electricity prices presented reflect the full social costs of power generation and delivery. Specifically, the wind power prices are suppressed by virtue of federal and, in some cases, state tax and financial incentives.
Furthermore, these prices do not fully reflect integration, resource adequacy, or transmission costs. At the same time, wholesale electricity prices do not fully reflect transmission costs, may not fully reflect capital and fixed operating costs, and are suppressed by virtue of any financial incentives provided to fossil-fueled generation and by not fully accounting for the environmental and social costs of that generation.
In addition, wind power prices – once established – are typically fixed and known (because wind energy is often sold through long-term, fixed-price power purchase agreements), whereas wholesale electricity prices are short-term and therefore subject to change over time.
Finally, the location of the wholesale electricity nodes and the assumption of a flat-block of power are not perfectly consistent with the location and output profile of the sample of wind power projects.
In short, comparing wind farm and wholesale electricity prices in this manner is not appropriate if one’s goal is to fully account for the costs and benefits of wind energy relative to its competition. Another way to think, however, is as loosely representing the decision facing wholesale electricity purchasers that are otherwise under no obligation to purchase additional amounts of wind energy – i.e., whether to contract long-term for wind power or to buy a flat block of (non-firm) spot power on the wholesale electricity market.
Project Performance and Capital Costs Drive Wind Power Prices
Wind power sales prices are affected by a number of factors, two of the most important of which are installed wind farm project costs and project performance.
The Installed Cost of Wind Power Projects Continued to Rise in 2009, but Reductions May Be on the Horizon
Berkeley Lab compiles data on the installed cost of wind power projects in the United States, including data on 115 wind farm projects completed in 2009 totaling 9,656 MW, or 97% of the wind power capacity installed in that year. In aggregate, the dataset includes 405 completed wind power projects in the continental United States totaling 28,522 MW, and equaling roughly 81% of all wind power capacity installed in the United States at the end of 2009.
In general, reported wind farm project costs reflect wind turbine purchase and installation, balance of plant, and any substation and/or interconnection expenses. Data sources are diverse, however, and are not all of equal credibility, so emphasis should be placed on overall trends in the data, rather than on individual project-level estimates.
The installed cost of wind power projects declined dramatically from the beginning of the industry in California in the 1980s through the early 2000s (falling by roughly $2,700/kW over this period), but have more recently increased.
Among the sample of wind farm projects built in 2009, for example, the capacity-weighted average installed cost was $2,120/kW. This average increased by $170/kW (9%) from the weighted-average cost of $1,950/kW for projects installed in 2008, and increased by $820/kW (63%) from the average cost of wind projects installed from 2001 through 2004. Project costs have clearly risen, on average, over the last five years.
Learning curves have been used extensively to understand past cost trends and to forecast future cost reductions for a variety of energy technologies, including wind energy. Learning curves start with the premise that increases in the cumulative production or installation of a given technology leads to a reduction in its costs.
The principal parameter calculated by learning curve studies is the learning rate: for every doubling of cumulative production/installation, the learning rate specifies the associated percentage reduction in costs. Based on the installed cost data and global cumulative wind power installations, learning rates can be calculated as follows: 9.4% (using data from 1982 through 2009) or 14.4% (using data only during the period of cost reduction, 1982-2004).
It is important to recognize that wind power projects were not alone in seeing upward pressure on project costs – other types of power plants experienced similar increases in capital costs. For example, the IHS CERA Power Capital Cost Index of coal, gas, and wind power plants shows a 74% capital cost increase from 2000 to the end of 2009.
Some of the cost pressures facing the industry in recent years (e.g., rising materials costs, the weak dollar, and wind turbines and component shortages) have eased since late 2008. As a result, while costs may – on average – remain high for a period of time as wind farm developers continue to work their way through the dwindling backlog of turbines purchased in early 2008 at peak prices under long-term frame agreements, there are expectations that average installed costs will decline over time.
Data compiled by Berkeley Lab show an estimated weighted-average cost for a sample of more than 4,300 MW of projects likely to be built in 2010 of $2,230/kW, or $110/kW higher than for the sample of projects completed in 2009.
Installed wind farm project costs exhibit economies of scale, at least at the low end of the project size range. Among the sample of projects installed in 2007, 2008, or 2009 – there is a significant drop in per-kW average installed project costs when moving from projects of 5 MW or less to wind farm projects in the 5 to 20 MW range. As wind farm project size increases beyond 20 MW, these data do not show continued economies of scale; the reason for this latter trend is unclear.
Regional differences in average project costs are also apparent, and may occur due to variations in development costs, transportation costs, siting and permitting requirements and timeframes, and other balance-of-plant and construction expenditures. Considering only wind farm projects in the sample that were installed in 2007, 2008, and 2009, the capacity-weighted average cost equaled $2,000/kW nationwide over this period. Texas was the lowest-cost region, while California and New England were the highest-cost regions; all other regions came in close to the nationwide average.
General Electric (GE) remained the number one manufacturer of wind turbines supplying the U.S. market in 2009, with 40% of domestic turbine installations. Following GE were Vestas (15%), Siemens (12%), Mitsubishi (8%), Suzlon (7%), Clipper (6%), Gamesa (6%), REpower (3%), Acciona (2%), and Nordex (1%). Other utility-scale (>100 kW) wind turbines installed in the United States in 2009 (and that fall into the “Other” category in Figure 9) include turbines from NedWind (6.5 MW), AAER (6 MW), DeWind (6 MW), Fuhrlander (4.5 MW), Goldwind (4.5 MW), RRB (2.4 MW), Elecon (0.6 MW), and Wind Energy Solutions (0.25 MW).
Primary authors: Ryan Wiser, Lawrence Berkeley National Laboratory, Mark Bolinger, Lawrence Berkeley National Laboratory. With contributions from Galen Barbose, Naïm Darghouth, Ben Hoen, and Andrew Mills (Berkeley Lab), Kevin Porter and Sari Fink (Exeter Associates), Suzanne Tegen (National Renewable Energy Laboratory).