Net Energy Payback and CO₂ Emissions from Three Midwestern Wind Farms: An Update

This paper updates a life-cycle net energy analysis and carbon dioxide emissions analysis of three Midwestern utility-scale wind systems. Both the Energy Payback Ratio (EPR) and CO₂ analysis results provide useful data for policy discussions regarding an efficient and low-carbon energy mix. The EPR is the amount of electrical energy produced for the lifetime of the power plant divided by the total amount of energy required to procure and transport the materials, build, operate, and decommission the power plants. The CO₂ analysis for each power plant was calculated from the life-cycle energy input data.

A previous study also analyzed coal and nuclear fission power plants. At the time of that study, two of the three wind systems had less than a full year of generation data to project the life-cycle energy production. This study updates the analysis of three wind systems with an additional four to eight years of operating data.

The EPR for the utility-scale wind systems ranges from a low of 11 for a two-turbine system in Wisconsin to 28 for a 143-turbine system in southwestern Minnesota. The EPR is 11 for coal, 25 for fission with gas centrifuge enriched uranium and 7 for gaseous diffusion enriched uranium. The normalized CO₂ emissions, in tonnes of CO₂ per GW_eh, ranges from 14 to 33 for the wind systems, 974 for coal, and 10 and 34 for nuclear fission using gas centrifuge and gaseous diffusion enriched uranium, respectively.

INTRODUCTION

Rising levels of atmospheric carbon dioxide (CO₂) have led a consensus of scientists to conclude that human activity is leading to global climate change (Oreskes, 2004). Although debate continues as to how the climate will change, many leaders seek ways to reduce anthropogenic CO₂ emissions. Reducing fossil fuel use through conservation efforts, improved energy efficiency and switching to less carbon-intensive fuels are areas of particular importance.

This paper analyzes two data points that can be used in this discussion. One is the Energy Payback Ratio (EPR), which is the ratio of energy produced for the normal life span of a power plant divided by the energy embodied in the construction, operations, fuel procurements and materials of that plant. The higher the EPR, the greater the energy gain a technology has. The second data point is the CO₂ emissions factor that is computed directly from the embodied energy calculations of the EPR analysis. Even “carbon free” energy sources, such as wind and nuclear, have a CO₂ emission factor because of the energy-consuming processes required to construct and operate the power plants, as well as procuring fuel and materials. Such information can be used in policy discussions concerned with future energy sources.

This analysis updates an earlier study comparing the net energy balance and life-cycle carbon emissions of three Midwestern wind farms of deffering size and turbine models, to those of base-line power plants that used coal and fission to generate electricity (White and Kulcinski, 2000). At the time of the original study, two of the three wind farms had been producing electricity for less than a year, which meant all energy production projections were based on the developer's projected capacity factor and not actual production data. This analysis therefore factors in the new production data, as well as new information about the wind farms’ operations. The turbines analyzed are no longer considered state-of-the-art. They do, however, represent a large number of turbines that continue to operate and produce electricity. The coal and nuclear technologies represent those power plants that produce power today, but not those that will be built in the near future.

Table 1. Comparison of Wind Data

BACKGROUND

Growth of Wind Power in the U.S.

Installed capacity of wind power in the United States more than quadrupled between 1998 and 2006, increasing from 1,698 Megawatts (MW) to 9,972 MW (EIA, 2001; NREL, 2006). Installed wind capacity in the Midwestern states of Illinois, Iowa, Minnesota, and Wisconsin alone increased from 129 MW in 1998 to 1,752 MW in 2006.

Wind power currently is the fastest growing sector for electricity production (EIA, 2006), in large part resulting from the renewal of the Federal Production Tax Credit, renewed in 2005 through December 2007 (109th Congress, 2005); the PTC is worth ∼$0.02/kwh. With, continued growth expected in the wind energy sector, understanding its energetics and carbon contribution is important for future energy policy considerations. Of the three wind projects analyzed here, one is located in northeastern Wisconsin in a “low wind speed” area, and two are located on the Buffalo Ridge of southwestern Minnesota. It should be noted that these projects are not representative of the best wind sites in the U.S. With average wind speeds ranging from 13.6 to 16.1 mph (see Table 1), this equates to class 2 and class 4 winds. Many newer wind facilities are being built in locations with class 5 and 6 winds. Nevertheless, the analysis here takes advantage of some of the most abundant data available.

Update on Wind Projects

 

Glenmore

The Glenmore wind project is a two-turbine experimental wind farm in Brown County, Wisconsin. This project originally was part of the Wind Turbine Verification Program (TVP), a national study cosponsored by the Department of Energy (DOE) and Electric Power Research Institute (EPRI, 2001) that evaluated prototype advanced wind turbines at several sites in the U.S. (McGowin and others, 1997).

The project in Glenmore was set up as a low wind speed site (average wind speeds of 13.6 miles per hour) and was the first utility wind project in Wisconsin. Since the original study (White and Kulcinski, 1999), 8.5 years of additional wind data has been collected. Most projects are selected and built because of high, steady wind speeds, so the low, steady wind speeds at this location produce a lower rating in the factors measured here.

Buffalo Ridge

The Buffalo Ridge (BR) wind farm is located in southwestern Minnesota near the town of Tyler. It consists of 73 Kenetech KVS-33 wind turbines for a total capacity of 25 MW. LG&E Power operated the wind farm at the time of the original study (White and Kulcinski, 1999), though FPL Energy is the current operator.

Since the original study, 4.5 years of additional production data is available. Early in the project, a major blade failure necessitated replacement of all blades. This problem was factored into the original analysis.

Lake Benton I

Lake Benton I (LB) was originally named Buffalo Ridge Phase II. This wind farm is located on the Buffalo Ridge, near the town of Lake Benton, Minnesota. For a brief time it was the largest wind farm in the U.S., consisting of 143 Zond Z-46 wind turbines, each of which produces 750 KW. Built by Enron Wind Corporation, it now is owned by FPL Energy (Windpower Monthly News, 2000).

Since the original study, 7.5 years of additional production data are available. Also, in 1999, Enron Wind had to replace all 143 generators (Windpower Monthly News, 1999). The energy consumed in this process (embodied and otherwise) was factored into this analysis.

APPROACH AND METHODS

Energy Payback Ratio

Calculating the EPR is straightforward. First, all the useful energy produced by an electrical power plant for its lifetime is determined. Second, the total amount of energy embodied in procuring all the fuel and construction materials, and the energy required to construct, operate, and decommission the plant is calculated. Third, the energy payback ratio (EPR) is determined by the relationship in Equation (1) as follows:

$$ EPR = \frac{{E_{n,\,L} }}{{(E_{{\rm mat},\,L} + E_{{\rm con},\,L} + E_{{\rm op},\,L} + E_{{\rm dec},\,L} )}} $$

(1)

where

• E _n,L  = the net electrical energy produced over a given plant lifetime, L.

• E _mat,L  = total energy invested in materials used over a plant lifetime L.

• E _con,L  = total energy invested in construction for a plant with lifetime L.

• E _op,L  = total energy invested in operating the plant over the lifetime L.

• E _dec,L  = total energy invested in decommissioning a plant after it has operated for a lifetime L.

In practice, step one (calculating the energy produced) is less complex than step two. Further details of this process are given in previous papers on this topic (White 1998; White and Kulcinski, 1998).

CO₂ Emissions

Once the EPR is determined, the energy input data are used to calculate the emissions of a specific pollutant (i.e., CO₂ per kg of fuel, metal, or concrete for each GW_ey of net electricity sent to consumers). In calculating the energy inputs and CO₂ emissions associated with plant materials, the mass of each material is multiplied by the corresponding energy and CO₂ emission factor for each. For other activities, the amount of fuel required to perform the task (for transportation, construction, operation and maintenance, and decommissioning) is multiplied by the emission factor for that fuel type. This analysis is limited only to CO₂, the emission factor of which is stated mathematically in Equation (2) as follows:

$$ \frac{{{\rm kg}\,{\rm CO}_2 }}{{{\rm GW}_{\rm e} {\rm y}}} = \frac{{\sum_i {\left( {\frac{{{\rm kg}\,{\rm CO}_2 }}{{{\rm kg}\,M_i }}} \right){\rm kg}\,M_i } }}{{E_{n,L} }} + \frac{{\sum_j {\left( {\frac{{{\rm kg}\,{\rm CO}_2 }}{{{\rm unit}\,F_j }}} \right){\rm unit}\,F_j } }}{{E_{n,L} }} $$

(2)

where

• E _n,L  = the net electrical energy generated over a given plant lifetime, L.

• $\frac{{{\rm kg}\,{\rm CO}_2 }}{{{\rm kg}\,M_i }}$ = kg of CO₂ emitted per kg of material i produced

• $\frac{{{\rm kg}\,{\rm CO}_2 }}{{{\rm unit}\,F_j }}$ = kg of CO₂ emitted per unit of fuel j produced

• kg M _i  = kg of material i needed to construct and/or operate the plant for life L

• unit F _j  = units of fuel type j needed in various activities of the plant for life L

Further details of this process also are given in previous papers on this topic (White, 1998; White and Kulcinski, 1998).

Data Sources

Energy Production

The monthly average electricity production data for each wind farm are listed in Table 2. Generation data came from a number of sources, including developers, project web-sites, and the Energy Information Administration (EIA). To accommodate gaps in some of the data, the annual generation average was calculated using the monthly average of all available data. This method sums all data for a given month and divides by the number of months. The sum of each month's average provides the annual average.

Table 2. Monthly Average Electricity Production, MWh

Electricity production data for the Buffalo Ridge wind farm is available for the period March 1994 through July 1998 (R. Sykes, pers. comm., 1997; R. Sykes, pers. comm., 1998), February 2001 through December 2003 (EIA, 2004), and January 2005 through May 2006 (EIA, 2004/6). There are gaps in the data from August 1998 through January 2001 and all of 2004. The capacity factor is calculated by dividing the annual average of electricity (56.0 GW_eh) by the amount of electricity that would be generated if the power plant produced at its rated capacity for a full year (25 MW_e*8760 hours/year = 219 GW_eh). The actual capacity factor of BR using the available data is 25.6%.

Electricity production data for the Lake Benton I wind farm is available for the period January 1999 through May 2006 (EIA, 2004; EIA, 2004/6). Only one-month of data (January 2002) was not available through EIA. The capacity factor for LB using the available data is 28.6%.

Electricity production data for the Glenmore wind farm is available from the project web-site (WPSC, 2005) for all months of operation, from March 1998 through July 2006. The capacity factor for Glenmore is 19.9%.

Embodied Energy and Energy Inputs

Greater detail for the energy requirements are given in the original analysis (White and Kulcinski, 1999). Data sources used to calculate the energy inputs are varied. The energy and CO₂ emission factors for materials from the wind plants are listed in Table 3 and detailed in White (1998).

Table 3. Summary of Energy and CO₂ Emission Factors for Power Plant Materials (White and Kulcinski, 1999)

The material requirements from each wind plant are listed in Table 4. These data were collected through data released by wind farm operators (WPSC, 2005) and personal communications with project representatives (R. Sykes, pers. comm., 1998; A. Zalay, pers. comm., 1998).

Table 4. Mass Requirements of Wind Turbine/Tower Assemblies (tonnes/unit)

The energy requirements for wind farm construction include transportation of the components to the construction site, and the actual onsite construction. The distances and related data for transporting the wind turbine and tower components from the manufacturing site to the wind plants are listed in the original analysis (White and Kulcinski, 1999). The manufacture site data for each component of Buffalo Ridge, Lake Benton I and Glenmore were obtained from LG&E (R. Sykes, pers. comm., 1998), Zond (A. Zalay, pers. comm., 1998), and Huron Wind Power (A. Rast, pers. comm., 1998) respectively. The distances between these sites were calculated using the Mapquest^TM map-generating program (Mapquest^TM, 1998), which determined the distances between two towns or addresses. The energy requirements per transportation mode are listed in White and Kulcinski (1999). To calculate the CO₂ emissions from component transportation, it was assumed that all energy requirements were from diesel fuel. The diesel fuel emission factor then was multiplied by the amount of fuel that would be needed to provide this energy. Heating values for diesel and other fuels also are listed in White and Kulcinski (1999).

The energy requirements to construct the respective wind plants were available only for the Glenmore Wind Project (H. Wittholz, pers. comm., 1998). Data used for the analysis of the BR and LB wind farms were scaled from data to construct the Glenmore wind farm.

The turbine/tower assemblies at the Glenmore site are both significantly larger in mass than those at BR and have a higher installed capacity (600 342.5 kW_e). The mass and material requirements are listed in Table 4. In calculating the CO₂ emissions from construction activities, it was assumed that all energy came from diesel fuel.

The construction data for the Glenmore wind project came in the form of economic costs, as did the operation and maintenance (O&M) data. Only data for Glenmore and BR wind projects O&M were made available. All costs were translated into 1995 dollars, using the Consumer Price Index (U.S. city average; all items; base period 1982–84 = 100; data are not seasonally adjusted (BLS, 1998)).

The cost data for construction and O&M for all wind farms is available in White and Kulcinski (1999). The energy requirements for both of these processes were calculated from the cost data using the Input/Output (I/O) method (Spreng, 1988). Carbon dioxide emissions were calculated by assuming that all O&M energy was in the form of diesel fuel. The energy units were converted to gallons of diesel, which in turn were multiplied by the CO₂ emission factor for the fuel (White and Kulcinski, 1999).

The O&M energy requirements for the LB wind farm were calculated from the Glenmore data scaled on the ratio of number of turbines (143/2) times O&M energy requirements for Glenmore. Carbon dioxide emissions for LB were calculated similarly to those of the Glenmore and BR wind farms.

Data on the energy requirements to decommission the three wind plants were not available at the time of this study. For this reason, an assumption was made that it would take approximately the same amount of energy to dismantle the turbine/tower-assembly as it would to erect and assemble the nacelles and towers. At the same time, it is assumed that while the nacelles with all their moving parts will last between 20 and 30 years, the towers that support the nacelles will last longer. For this analysis, it is assumed that a tower will last for the life of two wind turbines. The only energy requirements essential to assemble the second nacelle will involve removing the old and placing the new one on top of the tower, as well as necessary electrical hook-up. For this reason, it is assumed that the total energy required to dismantle one turbine will be half the energy required for on-site assembly, since the energy required for dismantlement can be amortized for two turbines. It is also assumed that the fuel used to dismantle the turbines will be diesel.

RESULTS AND DISCUSSION

Energy Analysis

The three wind farms analyzed here had energy payback ratios that varied widely. As noted in Table 5, the EPR's ranged from a low of 11 for the Glenmore wind project to 28 for Lake Benton I. Several factors seem to explain this: wind resource and capacity factor, life expectancy of the turbines, scales of economy, and material failure and replacement.

Table 5. Comparison of Energy Investments for Energy Systems, by Process (GJth/GW_eh)

The difference in the projected capacity factors of Glenmore and both Buffalo Ridge sites has to do with the wind resource at each site. The Buffalo Ridge average wind speed has been measured at 16.1 mph (Xcel Energy, 2004), which is considerably higher than the 13.6 mph originally measured at Glenmore (WPSC, 1998). However, Glenmore's average wind-speed during the first three years of operations was 15.4 mph (6.9 meters per second) (EPRI, 2001). When normalizing for capacity factor to equal that of Lake Benton I (28.6%), the EPR of BR and Glenmore increased to 27 and 16 respectively.

The manufacturers of the three wind turbine types each projected different life expectancies for their turbines. As shown in Table 1, the Kenetech KVS-33 wind turbine has a life expectancy of 25 years, while the Zond Z-46 and Tacke 600e have life expectancies of 30 and 20 years respectively. Each of these different life expectancies were used as part of the analysis for the respective wind turbines. Ultimately, it will not be known how long these turbines last until they are decommissioned. Even then, it is possible they will be refurbished and used again elsewhere. If, however, the life expectancy were normalized to 25 years, that alone would change the “actual” energy payback ratio (using existing generation data), to 24 and 14 for LB and Glenmore, respectively.

If both the capacity factor and turbine life expectancy are equalized, the EPR for BR, LB, and Glenmore would be 27, 24, and 19, respectively. These are the two major factors in the overall success of the technologies. Although the life expectancy will not affect the effectiveness of the turbines while operating, it will impact the amount of electricity they are able to generate over their lifetime. Any deviation in the actual number of years a turbine operates will affect the final results.

The results of the energy analyses as shown in Tables 6 and 7 are normalized in two ways. In Table 6, the energy requirements are normalized per megawatt of installed-capacity for the entire wind farm. Table 6's data are normalized per gigawatt-hour of electricity generated. Each method has advantages.

Table 6. Comparison of Energy Investments for Energy Systems, by Process (GJth/MWe-Installed)

Table 7. Comparison of CO₂ Emissions from Energy Systems, by Process (tonne CO₂/GW_e-Installed)

Normalization per installed-megawatt eliminates differences in the location of each wind turbine. Some sites have a better wind resource, which serves as an advantage to any wind turbine located there. This is shown by the fact that although both BR and LB have similar energy requirements per installed MW, Lake Benton has a significant advantage to BR when normalized per unit of electricity produced. The energy requirements per installed MW at Glenmore are ≈20% greater than either BR or LB resulting in part to the economy of scale, taller towers, and a more robust foundation.

Normalizing the data per gigawatt-hour produced factors in each site's wind resource, wind turbine capacity factors, as well as the conversion efficiencies of the nacelles. This method of normalization also coincides with economic analyses, which standardly measures cost per unit of electricity generated (e.g., GWh). The rest of this analysis and all the figures refer to data normalized per GW_eh, unless otherwise noted.

As seen in Figure 1, the majority of the energy requirements for all three wind projects is embodied in the materials. This is highlighted in Figure 2, which shows that the energy embodied in materials is responsible for a large share of the energy requirements of BR (49%), LB (65%), and Glenmore Wind Project (71%). Figure 1 shows the normalized energy requirements embodied in the Glenmore wind-turbine/tower-assembly materials are 50% higher than the total energy requirements of LB and nearly as much as the total for BR. The larger mass of materials is the primary reason for the greater share of energy going toward materials at Glenmore.

Figure 1.

Energy requirements per GW_eh for Glenmore wind farm are more than twice those of other two wind farms.

Figure 2.

Materials production dominates energy requirements for wind farms.

The energy requirements for operations and maintenance (O&M) comprise ≈34% of the total energy at BR. The processes involved in O&M are similar for the three wind farms. Wind farms are modular by nature and each nacelle has numerous moving parts. Maintaining a wind farm is similar to maintaining a fleet of cars in that each turbine requires regular maintenance, including lubricating oil for the generators and fuel for service vehicles. Service vehicles may require long drives for service personnel because of the typical remoteness of wind turbines. The O&M energy requirements for BR are considerably greater than those of the other two wind farms, because of the smaller output of the turbines. The processes and energy expenditures required to service any wind turbine is similar, be it a 330 kW or a 750 kW turbine. One megawatt of installed capacity at BR is comprised of three turbines, whereas at LB it is 4/3. The energy requirements for both LB and Glenmore were based on the data for BR and scaled based on the ratio of wind turbines and the ratio of number of years of expected operation. Because BR and LB have full-time dedicated crews on-site to take care of all O&M and Glenmore requires technicians to service the turbines on an annual schedule and on an “as needed” basis, the actual O&M for Glenmore is likely higher.

The energy requirements for material transportation are highest for BR. This is because in part that each turbine has a lower rated capacity and also because of the longer distance the nacelle was transported. The BR nacelles were transported 4,100 miles (R. Sykes, pers., comm., 1998), while the LB and Glenmore nacelles were transported 1,800 and nearly 5,000 miles, respectively (A. Rast, pers., comm., 1998; A. Zalay, pers., comm., 1998). Although Glenmore's Tacke nacelles came from Germany, they were transported by ship and rail, which are less energy-intensive modes of transport than truck. Also, each of the Lake Benton gearboxes had to be replaced—shipped back to California with new gearboxes shipped to Minnesota (Windpower Monthly News, 1999). This tripled the total distance attributed to this equipment in the analysis.

Plant construction data were not available for either the BR or LB wind farms. The data for these two wind farms were scaled from the data for the Glenmore wind farm, factoring in both the number of turbines and the mass of each turbine. Plant construction energy requirements are highest for the Glenmore Wind Project. This can be attributed in part to the larger mass of the Tacke components and taller, heavier towers that are needed to tap into the wind resource. It is likely that had complete construction data been available for the two Minnesota projects, the scales of economy would have factored against Glenmore.

A summary of the overall energy payback ratios (EPR) is given in Figure 3. The results of this study found the Glenmore Wind Project produces 11 times more energy than is required to make it over the lifetime of the plant. The EPR is significantly higher for BR (24) and LB (28).

Figure 3.

Lake Benton I wind farm has energy payback ratio that is more than twice that of Glenmore wind farm.

For comparative purposes, Figure 4 shows the EPR of BR compared to baseload coal, and nuclear fission. The energy gain of wind technologies, as measured through the energy payback ratio, is favorable compared with two baseload technologies, coal and nuclear fission using gaseous diffusion enrichment of the uranium. For nuclear fission, when enriching the uranium via the gas centrifuge—the United States’ first gas centrifuge enrichment facility received a license to construct and operate in New Mexico on 23 June 2006 (NRC, 2006)—the energy payback ratio is similar to wind.

Figure 4.

Energy payback ratio differs by more than factor of two between coal and wind power plants (White 1998).

A fair comparison of wind power plant technologies to baseload technologies would include energy storage for wind. Wind and other intermittent technologies will never be able to compete fully with baseload technologies without a means to store energy for the times when they are not directly producing electricity. At this time, energy storage is not needed because the amount of electricity produced by wind power is small enough that all of the electricity can be incorporated into the electrical grid. When wind comprises a sizeable share of the electricity market, it will be necessary to use some form of energy storage, and the inclusion of this component will degrade the energy payback ratio (by increasing energy requirements) as well as increase the emissions of CO₂ (Denholm, 2004).

Likewise, the energy requirements for operations and maintenance for both Glenmore and LB are based on results from BR. Operations and Maintenance make up a larger percentage of total energy requirements for the wind farms, from 12% for Glenmore to 36% for BR.

Other papers have reported EPR's for wind turbines ranging from four (Uchiyama and Yamamoto, 1991) to 80 (Danish Wind Turbine Manufacturers Association, 1997). The lower EPR was for a small 100 kW_e wind turbine while the higher one was for a 600 kW_e turbine performed by the Danish Wind Turbine Manufacturers Association, which included energy credits for the recycling of the steel at the end of the turbines’ useful life.

CO₂ Analysis

The results of the carbon dioxide analyses as shown in Tables 7 and 8 are normalized in two ways. In Table 7, the CO₂ requirements are normalized per gigawatt of installed-capacity for the entire wind farm. The data in Table 8 are normalized per gigawatt-hour of electricity generated. For the most part, the CO₂ emissions parallel the energy requirements.

Table 8. Comparison of CO₂ Emissions from Energy Systems by Process (tonne CO₂/GWh_e)

Figure 5.

Lifetime emissions per GW_eh from Glenmore wind farm are more than twice those of Buffalo Ridge.

Figure 5 shows that the CO₂ emissions from materials production are the dominant source in the wind plants’ life-cycle. Highlighted in Figures 5 and 6, materials production is responsible for the greatest share of CO₂ from Buffalo Ridge (59%), Lake Benton (82%), and Glenmore (80%). Materials production are responsible for a greater share of the total CO₂ emissions than their share of the total energy requirements because of the use of coal to produced these materials as opposed to natural gas or diesel fuel in other processes. The full mixture, for these processes, which includes coal (for ore smelting) and electricity (55% of which is from coal in the United States). The percentage of CO₂ in these fuels is higher per unit of energy than that of diesel fuel. The share of CO₂ emissions from the other processes largely parallels those in the energy analysis. As shown in Figure 6, the normalized CO₂ emissions from the Glenmore wind project materials’ production are greater than the overall totals of either BR wind project. The reason for this is partly due to the greater mass of materials for the Glenmore nacelles and towers as well as the shorter projected lifetime of the nacelles, which, results in less generated electricity.

The amount of CO₂ emitted per GWh from decommissioning, on-site assembly and materials transportation are small; each comprising less than 10% of the total emissions of the respective wind projects. The normalized emissions from operations and maintenance follow the same trend of the energy requirements, which was discussed previously.

Figure 6.

Wind farm CO₂ emissions are dominated by materials production.

Figure 7.

CO₂ emission rates of electrical power plants are dominated by coal (White, 1998).

The amount of CO₂ emitted per GWh for the Glenmore project is about twice as much as the emissions from LB. As stated previously, the projected lifetime of the nacelles, and the differences in capacity factors are the largest factors.

In Figure 7, the CO₂ emissions of the wind farms are compared to other electricity generation technologies. The amount of CO₂ from LB and BR are slightly higher than fission using gas centrifuge enrichment. Although conventional coal plants are responsible for significantly greater amounts of CO₂ (30–75 times more), CO₂ from wind is favorable compared to other low-carbon sources of electricity. Only the future nuclear fission technologies, using gas centrifuge enrichment, produce fewer emissions than the wind farms analyzed here.

The wind power plants in this analysis emit between 14 and 34 tonnes of CO₂ per GW_eh. These results compare favorably with results from other studies, which are in similar units; 7.4 from San Martin (1989), 18 from Friedrich and Marheineke (1994), and 73 from Uchiyama and Yamamoto (1991).

Although CO₂ emissions from wind are small compared to coal, they are responsible for some emissions. The amount of electricity produced per turbine, which is a factor of the number of years the turbine operates and the capacity factor (which is a factor of both wind and turbine availability), has the greatest impact on the CO₂ emission factor of wind-generated electricity.

CONCLUSIONS

The wind sites analyzed here are not considered the best wind resources in the U.S. Better sites have been developed and are planned for development. Yet, these three wind sites, especially the two in Minnesota, yield better Energy Payback Ratio's than the two main baseload technologies in the U.S. today—coal and nuclear fission. Wind farms with better wind resources such as many on the Great Plains—with measured capacity factors in excess of 35–40%—are expected to be even better. At the same time, future coal and nuclear technologies also are likely to be more efficient than those currently operating, most of which were built more than 20–30 years ago.

Wind's energy payback ratio—ranging from 11 for the Glenmore Wind Project to 28 for Lake Benton I—compares favorably to baseload electricity mainstays coal (11) and nuclear fission using gaseous diffusion enrichment (7). Nuclear energy using gas centrifuge enrichment is projected to have an EPR of 25. Low wind-speed and a poor economy of scale are the likely reason for the low EPR of Glenmore.

Carbon dioxide emissions from wind-generated electricity range from 14 tonnes per GWh of electricity produced at LB to 18 at BR to 34 at the Glenmore wind project. Emissions are low compared to the 974 tonnes of CO₂/GWh for a conventional coal plant and comparable, yet higher than nuclear fission with gas centrifuge enrichment - 10 tonnes of CO₂/GWh.

Differences in the capacity factor and projected life expectancies of the nacelles, as determined by the manufacturers, are the biggest factors in the range of results for wind. All material and processes being equal, the wind resource and availability—and the resulting capacity factor—are the primary differences. Also, to make a pure apples-to-apples comparison between wind and baseload power sources, wind power would need to be paired with some form of energy storage. In such a situation for wind power, the EPR of wind power certainly would decrease and the emissions per unit of electricity produced would increase.

The main reason wind power plants are not significantly better than baseload power plants in terms of both energy payback and CO₂ emissions is because of their low-capacity factor. Despite a capacity factor of 20–29% for the wind farms analyzed here, which is less than one-third that of coal and nuclear technologies (75%+), the EPR of wind power plants is better than coal and fission as they currently operate in the U.S.. The CO₂ emission factor for wind is also in the same range as fission. A higher capacity factor would mean more generated electricity, but would not require any additional energy input.

There is no single Energy Payback Ratio or CO₂ emission factor for wind farms. Unlike most other electric power technologies, wind power is highly variable and location dependent. No two wind farms can be expected to have the same EPR results even if the technology is the same, because of differences in the wind at the different locations. Many factors, including wind resource, turbine reliability and longevity, economy of scale, parts failure, and the differences in make and model of the turbine, will effect the final numbers in both of these categories.

Wind turbine technology continues to evolve and improve. These results for wind turbines built 8–12 years ago, do not completely represent wind farms built in the last few years, today, or the future. Today's wind turbines are taller and have more than twice the electrical capacity of the largest turbines analyzed here. A wind farm with a similar capacity today can have less than half as many turbines as those analyzed here. At the same time, today's turbines are larger, requiring more materials, taller towers and specialized equipment to erect the structures—equipment which rarely occurs locally and must be trucked in from elsewhere. It is not known how this equates to EPR and CO₂ emissions.

Although it is not known how today's wind farms compare under the parameters measured here, it can be expected that wind turbine technology will continue to improve and become more efficient. Up to a point, larger wind turbines and wind farms will take advantage of economies of scale, which will decrease the amount of energy expended in materials and construction, while producing more power per unit. When used in conjunction with other studies, these results can show potential of wind power and trends of the technologies’ energetics.