The Energy Collective Group

This group brings together the best thinkers on energy and climate. Join us for smart, insightful posts and conversations about where the energy industry is and where it is going.

9,742 Subscribers

Article Post

Wind Energy CO2 Emissions Reductions are Overstated

Governments have passed laws that provide various subsidies to promote build-outs of wind and solar power systems to reduce CO2 emissions from fossil-fueled energy generators; CO2 is one of many contributors to global warming.


If CO2 is so important, why are real-time, 1/4-hr grid operations data not reported by grid operators to determine just how effective wind and solar energy is for reducing CO2 emissions and how effective one balancing generator is versus another? If Ireland and Texas can do it, so can Germany and every other nation with wind energy on their grids.


Instead, elaborate systems of emission factors are applied to fuel consumption data or energy production data for a week, or a month, or a year to calculate CO2 emission reductions, i.e., nothing is measured, monitored and reported on a real-time, 1/4-hr basis. 


Government statistics end up showing their CO2 emissions are declining month-to-month or year-to-year, i.e., our subsidizing policies are working, let’s charge ahead and tell everyone to do the same.


Various power systems engineers, with decades of experience designing and operating power plants and grids, some retired and finally free to speak their minds, have doubts whether the CO2 emissions reductions claimed by government officials and wind energy promoters are actually true.


The purpose of this article is to examine the issue of CO2 emissions reduction by wind energy in some detail. 




All generators on a grid, not dependent on the weather or the sun, are available to produce energy up to rated output throughout the year, except during scheduled and unscheduled outages. The scheduled outages are for annual maintenance, usually a 3-4 week period of the year, usually scheduled during low demand periods of the year. The unscheduled outages occur at random, such as due to a failure of an item of equipment. Such outages are kept to a minimum by having multiple identical items of equipment that perform the same function, such as having three 50% pumps of which only two are required at plant rated output.


Wind and solar energy generators, dependent on the weather and the sun, are available to produce energy at only a fraction of their rated outputs; in case of wind energy, only when the wind blows above about 7.5 mph to turn the rotors; in case of solar energy, during about 8-10 daytime hours.


The variability and intermittency of wind and solar energy present grid operators with extra challenges to maintain grid frequency and voltages within prescribed ranges. When too much wind energy, usually at night when demand is low, and too much solar energy, always during the day, are fed into the grid, other quick-ramping generators, such as OCGTs and CCGTs, need to operate at part-load (which is inefficient) to be able to ramp down when wind energy surges and ramp up when it ebbs (which is also inefficient and more quickly wears out equipment).


If grids do not have sufficient capacity of such quick-ramping generators, wind energy needs to be curtailed by feathering the blades and/or stopping the rotors, or exported to other grids, if THEY have the capacity to deal with it. Solar energy presents some special challenges, because it cannot be turned off, which is an issue in areas with many PV solar systems, such as in Southern Germany, Southern California, Spain, etc.


Much of those challenges would not exist, if economically-feasible, utility-scale, energy storage systems had been invented AND were deployed in sufficient energy storage capacity, GWh. The US capital cost for energy storage systems would be many tens of billions of dollars and deployment in meaningful energy storage capacity would not exist for at least 10-15 years AFTER it has been invented.




Dispatch Value: Wind energy is significantly different from conventional gas, coal, nuclear and hydro energy. Conventional generators, such as coal, gas and nuclear, are available for service 24/7/365, except during scheduled and unscheduled outages, i.e., reliable, staffed, fueled and controllable. Some are base-loaded or load-following, others can be put in service, i.e., dispatched, on short notice, whereas the “fuel” of wind turbines is a product of weather-dependent, variable wind speeds, i.e., its supply is unpredictable, unreliable and uncontrollable, and therefore, it has zero-dispatch value to a grid operator. Wind energy DISPLACES conventional energy on the grid, but in an inefficient manner. Wind energy does not REPLACE conventional energy.


Real-time wind speed prediction has become more accurate in recent years. It gives a grid operator a few hours notice regarding ESTIMATES of wind speed changes, which will give him time to more efficiently order the starting or stopping of OCGTs and CCGTs to maintain adequate spinning and ramping capacity at all times. This extra “juggling” of generators just to accommodate wind energy is less efficient, i.e., consumes extra fuel/kWh and emits extra CO2/kWh, than without wind energy.


Capacity Value: A grid operator needs to have available an adequate mix of generating capacity to serve peak demands for long-term planning purposes. The mix varies from grid to grid. Wind turbine systems have a capacity value in this mix.


Example: For summer peak capacity planning, ERCOT, the operator of the Texas grid, counts 8.7 percent of the Texas wind turbine rated capacity as dependable capacity at peak demand, in accordance with ERCOT’s stakeholder-adopted methodology. According to ERCOT, the capacity value is a statistical concept created for generator planning purposes. It is based on multiyear averages of wind energy generation at key peak demand periods.


ERCOT’s capacity planning value of 8.7% does not mean the ENERGY of 8.7% of wind turbine rated capacity would be available at any specified “time-ahead” period. Because of the randomness of wind speeds, no one can accurately predict available wind energy at any future time. Hence, it’s not available “on-demand”, i.e., not dispatchable.


Variability: Because wind energy increases by the cube of the wind speed, any change in wind speed creates significant surges and ebbs of wind energy. If such energy were fed into the grid, it would create chaos.


Thus, wind energy cannot stand on its own, has no value on its own, is completely useless, unless the grid has an adequate capacity of quick-ramping gas turbines and/or hydro plants that are required to inefficiently operate at part-load to be able to ramp up when wind energy ebbs and ramp down when it surges, which happens at least 100 times per day, to maintain grid frequency and voltage within required limits. If a grid does not have adequate capacity of such ramping plants, it either must acquire it, or connect to grids that do have it and do not need it for their own variable wind and solar energy.

During periods of high wind energy generation, many grids, such as of Germany, the Bonneville Power Authority, Texas, Colorado, Germany, Spain, etc., do not have a sufficient capacity of such quick-ramping generators. As annual wind energy percents on the grids increase, the grid operators are unable to balance the wind energy and need to transfer it to neighboring grids for balancing, if possible, and/or implement curtailments, which upsets wind turbine owners, because subsidy payments may be at risk; in the US, the production tax credit, PTC, is 2.2 cent per kWh produced.


Example: German wind power output peaked at about 12,000 MW on July 24, 2011, four days later the peak was 315 MW. Germany’s wind turbines are located mostly in Northern Germany which lacks adequate transmission facilities to Southern Germany, where the unpredictable, excess wind energy is likely not needed, because it usually occurs at night when demands are minimal. 


Intermittency: Wind energy usually is minimal during summer (almost non-existent in New England), moderate during spring and fall, and maximal during winter. Almost all the time, it is maximal at night. 


In the US Great Plains, with excellent wind conditions most of the year, about 10-15 percent of the hours of a year near-zero wind energy is produced, because wind speeds are insufficient (less than 7.5 mph) to turn the rotors, or too great for safety. During these hours, wind turbines draw self-use (parasitic) energy FROM the grid, and also during hours with slowly turning rotors when self-use energy exceeds the generated energy. Rotors are often kept turning with grid energy to prevent the rotor shaft from “taking a set”, or to not disappoint visiting lay public, including legislators, etc.


New England, with good wind conditions only on 2,000-ft or higher ridge lines, about 30 percent of the hours of the year near-zero wind energy is produced, because wind speeds are insufficient, or too great for safety, as would be the case during windy weather fronts or tropical storms, such as Sandy, passing over the ridge lines. About 60% of the annual wind energy is produced during about 30% of the hours of the year, mostly at night, and mostly during winter.


There are times, not predictable, when wind energy is minimal, sometimes for days, even in New England, and then conventional capacity, MW, plus non-wind RE capacity, such as biomass, need to be sufficient to serve demand; remember, economically viable, utility-size energy storage, other than hydro, has not yet been invented.


Offshore, New England wind energy production is technically feasible, but its subsidized energy cost would be at least 20 c/kWh, 2 times ridge line energy cost, excluding wind energy integration costs of at least 1.0 c/kWh. This compares with New England annual average grid prices of about 5 c/kWh, unchanged for the past 3 years.


The capital costs of a utility-scale offshore project is about $4,200,000/MW, of which about 35% is for the capital cost of new transmission systems to connect the offshore wind turbines to each other and to shore. The capital cost of reinforcing the onshore grid and the capital cost of the balancing plants, etc., are not included. See URLs.

This compares with $2,500,000/MW for ridge line projects, such as on Lowell Mountain, Vermont.




Dispatch Value: Solar energy is significantly different from conventional gas, coal, nuclear and hydro energy. Conventional generators, such as coal, gas and nuclear, are available for service 24/7/365, except during scheduled and unscheduled outages, i.e., reliable, staffed, fueled and controllable. Some are base-loaded or load-following, others can be put in service, i.e., dispatched, on short notice, whereas the “fuel” of PV solar panels is a product of weather-dependent, variable solar insolation, i.e., its supply is unpredictable, unreliable and uncontrollable, and therefore, it has zero-dispatch value to a grid operator. Solar energy DISPLACES conventional energy on the grid, but in an inefficient manner. Solar energy does not REPLACE conventional energy.


Real-time solar insolation prediction has become more accurate in recent years. It gives a grid operator a few hours notice regarding ESTIMATES of solar insolation, which will give him time to more efficiently order the starting or stopping of OCGTs and CCGTs to maintain adequate spinning and ramping capacity at all times. This extra “juggling” of generators just to accommodate solar energy is less efficient, i.e., consumes extra fuel/kWh and emits extra CO2/kWh, than without solar energy.


Capacity: A grid operator needs to have available an adequate mix of generating capacity to serve peak demands for long-term planning purposes. The mix varies from grid to grid. Solar systems could have a capacity value in this mix, but insufficient systems are in place to determine it, except possibly in Southern Germany, Southern California, Spain, etc.


Variability: As scattered clouds move over a large number of PV systems, as in Southern Germany, Southern California, Spain, etc. they cause rapid, local decreases in output which adversely affects grid stability.


Example: One thousand roof- and field-mounted PV systems blanketed by a moving a square-mile cloud could cause a wavelike output decrease of several MW that moves with the cloud. With multiple clouds, the grid voltage and frequency would become unstable over a large geographical area.


Unlike wind energy, solar energy CANNOT be turned off or curtailed, as in Southern Germany with about 1 million PV systems, when on sunny summer days solar output surges from 0 MW in the morning to about 12,000 – 14,000 MW at noon and the energy in excess of demand has to be partially exported to France and the Czech Republic at fire sale prices, 5.5 euro cent/kWh or less, after having been subsidized at an average of about 50 euro cent/kWh. Any leftover/unwanted/unusable energy is grounded, i.e., wasted.


Example: Germany’s peak solar power is as little as 2% of rated capacity, or 340 MW (2% of end 2010 capacity), on cloudy days and when snowand ice covers the panels. 


Intermittency: In sunny areas, such as the US Southwest, about 65% of the hours of a year near zero solar energy is produced, more than 70% of the hours in New England and Germany. Solar energy is minimal in the morning, maximal at noon about 3-5 hours before the daily peak demand, minimal in the afternoon, minimal during foggy, overcast, snowy days, and zero at night.




The above indicates there are many hours during a year when little or no wind and solar energy is generated. Therefore, almost all conventional generator units would still need to be kept in good operating condition, AND staffed 24/7/365, AND fueled to serve the daily demand when wind and solar energy is insufficient. 


Without economically-viable, utility-scale energy storage systems, wind turbines and solar systems cannot replace any conventional units. All the units that would be needed WITHOUT the existence of wind turbines and solar systems, would also be needed WITH the existence of those systems. 


Some of the conventional units would have less energy production with wind and solar energy on the grid, thereby adversely affecting their profitability, which becomes worse as annual wind energy percent increases on the grid, requiring increased start/stop, part-load and part-load-ramping operations, all of which requires extra fuel consumption/kWh and emits extra CO2 emissions/kWh. At greater annual wind energy percent on the grid, say 4%, the extras significantly offset what wind energy was meant to reduce.




ECONOMICALLY-VIABLE, utility-scale energy storage (other than hydro) has not yet been invented and it would not be deployed in meaningful capacities (GWh) for a few decades AFTER it has been invented, i.e., by 2030 or later, as a recent German study concludes. In the meantime OTHER measures for balancing wind energy, listed below, are required.


What is generally not understood by lay people, i.e., at least 99% of the people, is that increasingly greater annual wind energy percent on the grid requires:


– inefficient operation of the conventional generators on the grid

– less efficient operation of the grid to maintain voltage and frequency within desired ranges.


At greater annual wind energy percent, say 12%, both mostly offset the reduction of fuel consumption and CO2 emissions due to wind energy, as explained below. 


Because wind energy is variable and intermittent, it requires hydro plants and/or quick-ramping, gas turbine plants and/or quick-ramping coal plants to ramp up when wind energy ebbs and ramp down when it surges which occurs at least 100 times per day, 24/7/365. 


A greater annual wind energy percent on the grid requires a greater capacity of generators to be in:


– starting/stopping mode (which is less efficient; more fuel and CO2 emissions/kWh)

– spinning mode (which produces no grid energy, but consumes fuel and emits CO2)

– decreased part-load mode (which is less efficient; more fuel and CO2 emissions/kWh)

– increased part-load-ramping mode (which is less efficient; more fuel and CO2 emissions/kWh)

Note: synchronous spinning is at 3,600 rpm, uses about 6-8 % of rated fuel consumption, sends no energy to the grid.  


The net result is increased fuel consumption/kWh and CO2 emissions/kWh of the fossil units that significantly offsets the fuel and CO2 emissions wind energy was meant to reduce, as proven by studies of the grid operations data of Ireland, the Netherlands, Texas and Colorado. The studies of Ireland are based on real-time, 1/4-hour grid operations data, of Texas and Colorado on 1-hour grid operations data, and of the Netherlands on historic fuel consumption and energy production data.


Example: Wind energy does displace coal energy, i.e., coal plants will not be operating at as high an output, as they would without wind energy (which is less fuel-efficient and adversely affects their economics), but wind energy increases a grid’s need for gas energy, because the grid’s OCGTs will need to be used not only for peaking and wind energy balancing during the day, but also for wind energy balancing at night, when wind speeds are greatest and demands are least, i.e., with increasing wind energy on the grid, an increasing capacity, MW, of generator units are needed to perform more hours of inefficient operation in spinning, part-load and part-load-ramping mode. 


Grids Using Primarily Gas Turbine Plants for Balancing 

The wind energy balancing in the Irish, the Netherlands and New England grids is performed by gas turbines, as described below. With wind energy on the grid, any CO2 emissions reduction would be almost entirely due to the changed operation of the gas turbines.


The Irish Grid: Ireland has more than a decade of experience with “juggling” generators to accommodate wind energy. EirGrid, the operator of the Irish grid, uses real-time wind speed prediction and publishes real-time, 1/4-hour grid operations data. Analysis of the November 2010 to August 2011 EirGrid grid operations data shows at an average wind energy penetration of 12.6%, the average efficiency of reducing CO2 emissions is about 70.6%, i.e., a ratio 1 : 0.7, for that 10-month period. 

See Table 2 in


However, this ratio is much less, because Eirgrid accounts only for efficiency reductions due to part-load operation of generators, but does not account for the extra fuel/kWh and CO2 emissions/kWh due to:


– Increased spinning plant operations; fuel and self-use energy is consumed, no output to the grid

– Increased start/stop operations; fuel and self-use energy is consumed, minimal output to the grid

– Increased part-load-ramping operations with wind energy than without.

– less than optimum economic scheduling of generating units with wind energy than without.

– increased line losses to gather the distributed wind energy from wind turbines.

– self-use energy drawn from the grid by the wind turbines during low/no-wind periods.


If EirGrid had accounted for all of the above items, wind energy’s CO2 reduction effectiveness would have been significantly less than the calculated 70.6%.


The Netherlands’ Grid: The Netherlands’ grid has a similar component of gas turbines as the Irish grid. A calculation of the effectiveness of wind energy was performed using the Netherlands’ historic fuel consumption and energy production data which INCLUDE the effects of spinning, start/stop, part-load-ramping, etc., unlike the EirGrid data.


For the Netherlands’ annual wind energy of only 3.36% in 2010 (3.972 TWh wind energy of 118.2 TWh total production), the effectiveness was calculated at 70%. If the Netherlands would have an annual wind energy of 12.6%, the same as Eirgrid’s, this ratio would have been significantly less, which proves Eirgrid omits much of the extra CO2 emissions due to wind energy being accommodated onto the grid.


Note: The Netherlands CF = 3.972 TWh/{(2,222 MW, end 2009 + 2,230 MW, end 2010)/2 x 8,760 hr/yr} = 0.204; mediocre, not profitable, but better than Germany’s 2007-2011 average CF of 0.187; Germany is as unsuitable for onshore wind energy as it is for PV solar energy. 


The New England Grid: ISO-NE personnel stated they do not yet “notice” wind energy on the grid, because it is only 1.0% of the annual supply. Inefficiencies due to spinning, start/stop, part-load ramping, etc., are minimal and the effectiveness would likely be about 0.95.


Summary of CO2 Emission Reduction Effectiveness: Government officials and wind energy promoters, such as the EWEA, BWEA, etc., usually claim one MWh of “clean” wind energy offsets one MWh of “dirty” fossil fuel energy, which is true regarding energy, but not regarding CO2 emissions, because of the inefficient operation of the other generators on the grid due to wind energy.


Below is summary of wind energy CO2 emission reduction effectiveness versus annual wind energy percent, for various grids:


1.0 at 0% wind energy on any grid.

0.97 (my assumption) at 1.0%, New England grid.

0.70 (calculated by Dr. LePair) at 3.36%, the Netherlands grid; based on at least 10 years of actual fuel and production data.

0.78 (calculated by Dr. Wheatley) at 3.5%, Australian grid. 

0.706 (calculated by Dr. Udo) at 12.6%, Ireland grid; based on deficient EirGrid data.

0.526 (calculated by Wheatley) at 17%, Ireland grid; based on SEMO data of individual generators, including increased start/stop CO2 emissions, and increased capacity and hours of spinning plant CO2 emissions, i.e., better than Eirgrid data.


Example: Ireland has an island grid with a minor connection with the UK grid. Eirgrid, the operator of the grid, publishes ¼-hour data regarding CO2 emissions, wind energy production, fuel consumption and energy generation. Several analyses of the Irish grid operations data by Dr. Udo and Wheatley offer clear evidence of the offset percentages increasing with increasing wind energy percentages.


The Wheatley study of the Irish grid shows: Wind energy CO2 reduction effectiveness = (CO2 intensity, metric ton/MWh, with wind)/(CO2 intensity with no wind). Ireland = (0.279, @ 17% wind)/(0.53, @ no wind) = 0.526, based on SEMO data.

If 17% wind energy, wind energy promoters typically claim a 17% reduction in CO2, i.e., 83% is left over.

If 17% wind energy, actual performance data of the Irish grid shows, 0.526 x 17% is reduced = 8.94%, i.e., 91.06% is left over.

NOTE: Dr. Udo used 1/4-hour Eirgrid data of the Irish system, and Wheatley used 1/4-hour SEMO data of individual generators, including: 

1. part-load operation output and hours of each generator

2. start/stop operation output and hours of each generator

3. spinning operation hours of each generator


With the characteristics of each generator, fuel consumption and CO2 emissions can be calculated for each 1/4 hour, as Wheatley has done in his article. Udo used a similar approach with the Eirgrid data that does not include the level of detail of the SEMO data. Eirgrid does not account for Items 2 and 3.


The Wheatley method is vastly superior to the government-sanctioned methods used in the US and Europe.


The method used by LePair for the Dutch system is based on about 10 years of government-published fuel consumption statistics.


Here is a rough CO2 emission estimate:


If no wind energy, emissions would be 100%.

If on the Irish system there is 17% wind energy, emissions would be about 83%, using government-sanctioned methods (near-100% effectiveness), for a net reduction of about 17%, but emissions would be about 100 – 0.526 x 17 = 91% using 1/4-hour SEMO data, for a net reduction of about 9%, i.e., not 17% less, but only 9% less.


On a graph of effectiveness versus wind %, a straight line drawn from 1.0 at zero wind %, to 0.526 at 17 wind % will show about about 0.97 at 1%, 0.90 at 3%, etc. As the Irish grid and New England grid have some hydro energy, their effectiveness will be greater, than if hydro energy were absent, as is the case with many other grids. 


NOTE: In general, at up to about 3% annual wind energy percent, the existing conventional generators and the “inertia” of most grids usually are sufficient to accommodate most of the wind energy variations. 


Above 3%, this is usually not the case and additional measures are required, such as:


– increased rotor blade “feathering”

– increased start/stops of wind turbines

– increased use of synchronous condenser systems for power factor control

– increased quick-ramping OCGT/CCGT capacity, GW, to balance wind energy

– Increased spinning plant capacity, GW, to quickly supplement wind energy

– increased energy storage, GWh; not yet invented

– increased interconnection with neighboring grids

– increased demand management enabled by the “smart grid”

– increased weather forecasting 


All those measures require energy and emit CO2 that offsets what wind energy was meant to reduce.


NOTE: In addition to the above CO2 emission reduction effectiveness, there are CO2 emissions associated with the materials, transport, construction and operation of the IWTs which equate to 1 – 2 years of CO2 emission reduction, which need to be offset before any CO2 emission reduction takes place. 


According to a study of UK and Denmark IWT replacements and decommissionings, the energy production useful service life of IWTs is 15 – 20 years, i.e., the effective CO2 emission reduction useful service life is 1 – 2 years less.

Grids Using Primarily Hydro Plants for Balancing


Balancing wind energy with hydro plants incurs the least cost/kWh and CO2 emissions/kWh. The outputs of hydro plants are controlled by varying the water flow to the turbines. The turbines need to operate in part-load-ramping mode for balancing wind energy which is less efficient, i.e., more waterflow/kWh, and incurs more wear and tear than if they were operated to follow daily demands without wind energy on the grid.  


Example: Danish wind energy in excess of Danish demand is fed into the Nordpool grid and absorbed and balanced by the hydro plants of Norway and Sweden thereby maintaining their reservoirs at higher levels than they would have been. 


Norway and Sweden buy the Danish energy from the Nordpool grid mostly at very low nighttime rates. They use the saved water in their reservoirs to generate energy to serve the Nordpool grid daytime demands when rates are higher. Denmark exporting subsidized energy, is, in fact, a flow of foreign aid from Danish to other Nordic consumers of up to one billion DKK during a “big export year”.


A good deal for Norway and Sweden, a bad deal for Denmark. The extra costs are rolled into Danish household electric rates (25.62 euro cent/kWh in 2012, highest in Europe), while industrial rates are kept low for international competitive reasons, as are Germany’s household (24.06 euro cent/kWh in 2012, second highest) and industrial rates.


Note: To make matters worse, Denmark is proudly aiming to have 50% of its energy production from on- and offshore wind turbines, almost all of which will need to be exported, as the output of Danish central plants and CHP plants, which have limited ramping capability, cannot be sufficiently reduced to keep all of the energy in Denmark.


Denmark will become a regional exporter of variable energy with balancing by Norway and Sweden. Whether this on-going, very expensive, offshore capacity expansion will be profitable remains to be seen.


All this is justified, because the nearly-bankrupt Vestas, a national “champion” whose stock has declined about 95%, is creating wind turbine jobs in Denmark.


NOTE: Minute-to-minute graphs of energy flows over international connections show, Denmark EXPORTS energy when winds are stronger and imports when winds are weak. The more wind energy Denmark produces (mostly at night when its own demand is minimal), the more it exports.


When ANNUAL wind energy was 20% of Denmark’s production, about 50% of that was consumed in Denmark, the rest was exported.


Now that ANNUAL wind energy is about 35% of Denmark’s production, about 30% of that is consumed in Denmark, the rest is exported.


Almost all generators in Denmark are “must run”, which mostly precludes operating in part-load-ramping mode just to balance wind energy. 


Grids Using Primarily Coal Plants for Balancing 


Older coal plants were designed to be base-loaded, not designed to have the high ramping rates, MW/minute, required for wind energy balancing. Newer coal plants, if designed for higher ramping rates, are more suitable for wind energy balancing. Whereas the operating range of gas turbines is about 40 – 100 % of rated output, of coal plants it is about 50 -100 % of rated output.


The operation of coal plants in part-load-ramping mode for balancing wind energy, especially during high-wind-speed periods, may destabilize combustion control systems causing extra fuel consumption, CO2 emissions and NOx emissions/kWh, and destabilize air quality control systems causing extra particulate, NOx and SOx emissions/kWh.


The extra fuel consumption and CO2 emissions causes the wind energy CO2 reduction effectiveness to significantly decrease during high-wind-speed periods, as shown by the Texas and Colorado grids when older coal plants of various vintages were used for wind energy balancing during high-wind-speed periods, because of insufficient available capacity of quick-ramping gas turbines.


During a wind energy surge, a coal plant in part-load-balancing mode would need to reduce steamflow to the turbine to reduce its output. This is done by blowing off steeam through a valve and by decreasing the pulverized coal flow to the burners and/or shutting off burners. 


During a wind energy ebb, a coal plant in part-load-balancing mode would need to increase steamflow to the turbine to reduce its output. This is done by opening wider the valve to the turbine and increasing the pulverized coal flow to the burners and/or turning on burners. 


All is done slowly to minimize combustion system and air quality control system instabilities and minimize thermal stresses which might result in boiler tube leaks and unscheduled outages. Usually, the limiting factor is the thermal stresses; unscheduled outages are expensive.


Grids dominated by coal plants of various vintages and an ANNUAL wind energy percentage of about 4% or greater, have significant operational challenges regarding frequency and voltage regulation and balancing of wind energy, especially during high-wind-speed periods when INSTANTANEOUS wind energy on the grid may be about 20% or greater during periods of low demand, such as at night when wind speeds usually are greatest.


In a system dominated by coal, wind energy primarily displaces gas turbine energy and coal energy which have CO2 emissions of about 400 g/kWh and 2.15 lb/kWh x 1 kg/2.205 lb = 975 g/kWh, respectively. 


Note: Modern subcritical boilers, supercritical boilers, ultra-supercritical boilers are more efficient and have CO2 emissions of 838 g/kWh, 800 g/kWh, 770 g/kWh, respectively.


Any CO2 emissions reduction in such a coal-dominated grid would depend on the weather-dependent annual wind energy %, the fuel types and consumption, and the changes of: 


– start/stop operations, and the type of units 

– spinning plant operations, and the type of units 

– part-load operations, and the type of units

– part-load-ramping operations, and the type of units

– scheduling of units to integrate wind energy; likely less economical than without wind energy.

Example: A study of the coal-dominated State of Illinois grid by Argonne National Laboratory, “Grid realities cancel out some of wind power’s carbon savings”, dated May 29, 2012, shows the 1 : 1 ratio is not valid. The study shares the usual flaws of other such NREL studies by being based on estimates, probabilities, algorithms, assumptions, grid operations modeling, weather and wind speed forecasts, etc., but, to its credit, it included CO2 emissions estimates of the increased start/stop operations and increased spinning plant operations with wind energy than without wind energy.




The Irish grid will be a major focus of this article because EirGrid, the grid operator, makes available the most complete real-time, 1/4-hour grid operations data for study.


Ireland’s Energy Generation: Ireland’s total electricity production was about 26,000 GWh in 2010. Gas-fired OCGTs and CCGTs provided about 65.5%, coal 13.2%, peat 8.2%, wind 9.8%, hydro 2.5% of which 1.7%, or 442 GWh, was impounded/run-of-river hydro. Ireland imports 100% of its coal, about 90% of its gas and produces 100% of its peat.


Wind Energy: In Ireland, good wind energy months are April, May, June, November and February.  On the west coast of Ireland, wind energy is greatest during summer daytimes, because of increased wind speeds as the lands warms up. The west coast wind energy coincides with greater daytime demands which is fortuitous. However, much of the energy needs to be transmitted to the east coast (line and transformer losses), as few people live on the west coast. 


This video, based on EirGrid data, shows wind output, MW, and total system output, MW, versus time, from 2001-2011. As Irish wind ouput increased from year-to -year, it became an increasingly larger fraction of the total system output, especially during very windy nighttime periods when demand is minimal.


Coal/Peat: The below website shows coal/peat plants are base-loaded, i.e., not used for balancing wind energy, i.e., their CO2 emission intensities are essentially constant.


Hydro: Ireland has many small hydro plants and a few larger plants, such as the Ardnacrusha power plant, built 1929, capacity 85 MW, output 332 GWh/yr, Cathaleens Falls 45 MW, Poulaphuca 30 MW and Inniscarra 19 MW. The below website shows hydro plant outputs follow daily demand, i.e., not used for balancing wind energy.


The almost 40-year old, 292 MW Turlough Hill pumped-storage facility pumps to add to its upper reservoir during low nighttime demand and produces energy during peak daytime demand. Its net effect is to “flatten” the daily demand profile. It is not used for balancing wind energy. Currently, it operates at about 50% of capacity, because of ongoing modifications.  


Combined-Heat-Power: Ireland has about 195 units totaling about 282 MW of operating combined-heat-power, CHP, plants of which a few larger units totaling 248 MW are dedicated to industrial processes, such as food, manufacturing and pharmaceutical. The output of these units is independent of the weather.


CHP energy generation was 6.3% of Ireland’s total energy generation in 2008 (latest data). 

Only 11 CHP units (mostly associated with industrial processes) exported 1,013 GWh to the grid in 2008, or 1,013/260 = 3.9% of total production. Eirgrid includes the exported energy and associated CO2 emissions of these units in its 1/4-hour data sets.


CHP heat generation was 4% of Ireland’s total heat generation in 2008 (latest data).


The above indicates CHP operations have no material impact on the 1/4-hour CO2/kWh posted by EirGrid.


OCGTs/CCGTs: A part of the OCGT/CCGT capacity serves base-load, follows daily demand, provides peaking power and performs voltage and frequency regulation. It also performs wind energy balancing, i.e., ramps down with smaller wind energy surges and ramps up with small wind energy ebbs.


Because larger wind energy surges and ebbs are unpredictable, additional OCGT/CCGT capacity needs to be in spinning and part-load-ramping mode for balancing wind energy; the greater the wind energy, the greater the additional  spinning and balancing capacity. 


Because of much degraded heat rates, Btu/kWh, and their combustion process becoming unstable, gas turbines are rarely operated below 40% of their rated output which limits their ramping range from 40 to 100 percent of rated output.


Fossil Units Less Efficient With Wind Energy: In the gas-energy-dominated Irish system, wind energy displaces mostly CCGT energy which, at zero wind energy on the grid, has CO2 emissions of 117 lb of CO2/(million Btu x 1 kWh/7,000 Btu) = 0.819 lb/kWh x 1/2.205 = 371 g/kWh, at an average turbine efficiency of (3,413 Btu/kWh)/(heat rate of 7,000 Btu/kWh) = 48.85%.; Ireland has mostly newer model CCGTs. 


The addition of wind energy to the Irish grid causes less efficient operation of the fossil units; i.e., extra Btu/kWh and extra CO2 emissions/kWh which significantly offset what wind energy was meant to reduce.


How EirGrid Calculates CO2 Emissions/kWh: The following is a direct quote from the EirGrid website:


“EirGrid, with the support of the Sustainable Energy Authority of Ireland, has developed together the following methodology for calculating CO2 Emissions.


The rate of carbon emissions is calculated in real time by using the generators MW output, the individual heat rate curves for each power station and the calorific values for each type of fuel used.


The heat rate curves are used to determine the efficiency at which a generator burns fuel at any given time.


The fuel calorific values are then used to calculate the rate of carbon emissions for the fuel being burned by the generator“


Grid operators know the heat rate curves of the plants on their grids which were obtained by testing. They need to know this for economic dispatch.


Eirgrid takes the percent of rated output each plant is operated at and multiplies it by the heat rate for that output percentage (from the above mentioned heat rate curve) to calculate the fuel consumption/kWh and CO2 emissions/kWh every 1/4 hour. It posts the grid CO2 intensity (CO2 emissions of all plants/total kWh produced by all plants) as gram CO2/kWh on its website every 1/4 hour.




The Irish grid was selected to determine the CO2 emission reductions due to wind energy on the grid. 


Eirgrid, the grid operator, is one of the few operators that publishes the following real-time, 1/4-hr grid operations data which are, for study purposes, superior to the 1-hr data published by Texas and Colorado.


– grid CO2 emission intensity, gram/kWh

– wind energy produced, GWh

– total energy produced, GWh 


Dr. Fred Udo, a graduate of the Technical University of Delft, the MIT of the Netherlands, spent a good part of his career at CERN, Switzerland, performing analyses of engineering and scientific data. He is retired, has no financial interest in RE. He performed several studies of the real-time, 1/4-hr data published by EirGrid.


Analysis of the November 2010 to August 2011 EirGrid grid operations data shows at an average wind energy penetration of 12.6%, the average efficiency of reducing CO2 emissions is about 70%, i.e., a ratio 1 : 0.7, for that 10-month period.


Wind energy: 12.6%

System, with wind energy: CO2 =  451.3 g/kWh

System, without wind energy: CO2 = 495 g/kWh

Fossil plants only: CO2 = 518.1 g/kWh

Reduction: (495 – 451.3)/495 = 8.9%

Efficiency: 8.9/12.6 = 70.6%


See Table 2 in


How EirGrid Understates CO2 Emissions/kWh: The 1/4-hour reported CO2 emissions/kWh are understated, as Eirgrid does not account for the extra fuel/kWh and CO2 emissions/kWh due to:


– Increased spinning plant operations; fuel and self-use energy is consumed, no output to the grid 

– Increased start/stop operations; fuel and self-use energy is consumed, minimal output to the grid 

– Increased part-load-ramping operations with wind energy than without.

– less than optimum economic scheduling of generating units with wind energy than without.

– increased line losses to gather the distributed wind energy from wind turbines.

– self-use energy drawn from the grid by the wind turbines during low/no-wind periods.


If EirGrid had accounted for all of the above items, wind energy’s CO2 reduction effectiveness would have been significantly less than the calculated 70.6% using the reported EirGrid data.


Note: There are significant CO2 emissions associated with RE build-outs, especially wind energy systems that need to be supported by additional transmission systems and OCGT/CCGT balancing gas turbines.

Example: Element Power claims an estimated total project cost of 8 billion euro (2,667,000 euro/MW, or $3,467,000/MW) and claims a completion date by the end of 2018. This estimate appears low compared to similar projects in the US. See URL.


Wind turbine installed capital cost = 3,000 MW x 1,800,000 euro/MW = 5.4 billion euro

Roads, transmission and other installation costs = 2.6 billion euro


Element Power does not mention the capital costs of reinforcing the Wales onshore grid to take the additional energy and the capital cost of the UK OCGT/CCGT balancing plants, etc. Are they assumed to be provided “by others”, i.e., UK rate payers?

Note: In my discussions with Mr. O’Sullivan, energy systems analyst of Eirgrid, he confirmed:


– Eirgrid’s reported CO2 emissions accounts for part-load operations, but not the ramping operations and the other above-listed factors.


– CO2 emissions reduction is secondary, as there are other reasons for building out wind energy, such as the Brussels’ mandated renewable energy percentages that provide Ireland with subsidies for wind turbine facilities.


– Ireland wants to reduce its fuel imports and increase its wind energy exports to Britain.


The analysis of the EirGrid data also found:

– the greater the wind energy percent on the grid, the lower the ratio, i.e., adding still more wind energy becomes less and less effective for CO2 emissions reduction.


– at very high wind energy percent on the grid, the ratio will ultimately go to zero and then become negative, i.e., adding still more wind energy to the grid will not further decrease CO2 emissions.

See Figure 1 in

The decreasing CO2 emission reduction effectiveness was verified by preparing a scatter diagram of the EirGrid data. The fit lines of the scatter diagrams of CO2 emission intensity, g/kWh, versus wind energy, %, show increasing CO2 emissions/kWh of the fossil units as wind energy percent increases. Where the fit line intersects the Y-axis, i.e., no wind energy, is their lowest CO2 emissions/kWh.

See Figure 1 of 


Opinions of Experienced Power Systems Engineers: The above findings appears entirely reasonable to power system engineers who know the more their power generators are operated in part-load and part-load-ramping mode, the less efficient they become and the less efficient the whole grid becomes.


Here is the testimonial of a UK power systems engineer with decades of experience in the utility industry. He is retired, i.e., finally free to speak up, and claims CO2 emission reduction due to wind energy is minimal.


Here are two articles by William Palmer, a retired power systems engineer of the Ontario Power System.


Just as a car, if operated at 20 mph, then accelerated to 50 mph and back down again a few hundred times during a 24-hour trip would use more gas and pollute more than operated at a steady speed, so would the balancing CCGTs and OCGTs.


However, gas turbines operating in part-load-ramping mode have even greater degradations of heat rates, Btu/kWh, than gasoline and diesel engines. The extra fuel consumed and extra CO2 emitted by the gas turbines are so much that they significantly offset what wind energy was meant to reduce. 



A national wind energy build-out would: 


– result in a littering of the US landscape with several hundred thousand, 450-ft to 550-ft tall, noise-making, health-damaging, industrial wind turbines everywhere there is wind. See note.


– require connecting them all with highly-visible, transmission systems.


– require integrating their variable, intermittent, zero-dispatch-value energy into the grid using less-efficiently operated hydro plants in some places and inefficiently-operated, CO2-emitting, gas turbine plants almost everywhere else.


– require rolling a part of the above mentioned LCOEs/kWh into rate schedules; households and small businesses would bear the brunt, because bigger businesses would be largely exempted for international competitive reasons, as is the case in Denmark and Germany.


– the CURRENT US energy production is about 4,000 TWh/yr, would be 5,724 TWh/yr in 2052, at a 0.9%/yr growthrate. Wind energy produced by 425,000 MW of onshore and offshore IWts would be 1,116 TWh/yr in 2052, or 19.5% of production. 

Note: Transmission and distribution losses, i.e., gather wind energy from remote offshore/onshore windy regions and deliver it to consumers in population centers, will reduce the energy delivered to consumers.


– the capital cost (2012$) would be trillions of dollars (wind turbines, gas turbines, transmission). 


Wind turbines: 425,000 MW  of IWTs would cost about $1.473 trillion

Gas turbines for balancing: 300,000 MW x $1,500,000/MW = $0.45 trillion 

Transmission and Distribution: $0.3 trillion

Total = $2.223 trillion.  


As the build-out would take at least 40 years at a rate of 21,250 MW/yr and the IWT useful service life is about 15 – 20 years (based on Netherlands, Denmark and UK experience), wind turbines built in the first 20 years would need to be replaced during the next 20 years, etc.


– most of existing energy plants would still need to be staffed 24/7/365, fueled and kept in proper operating condition to provide energy when wind and solar energy are insufficient. With less production, the economics of these plants would be adversely affected; a politically untenable situation requiring electric rates (mostly of households) be raised to compensate owners. 


– Assuming conventional coal plants would be phased out during those 40 years, the other 80.5% of energy would primarily be from natural gas, nuclear, hydro, coal gas, and PV solar. Energy from coal gas, wind, and PV solar likely would not be competitive with natural gas, unless subsidized or given other preferred treatment. 

Note: For the $2.22 trillion capital cost the US could build 500,000 MW of nuclear plants that would last about 60 years to produce 500,000 MW x 8,760 hr/yr x CF 0.90 = 3,942 TWh/yr, 69% of 2052 production, of relatively low-cost, CO2-free, steady, 24/7/365 energy and that would use most of the existing grid system, i.e., 3.5 times the energy production for about the same capital cost AND none of the cost and ever-present risk of grid stability deterioration due to variable, intermittent wind energy being added to the grid.


Note: Before variable, intermittent wind and solar energy, Germany had the most reliable and stable power system in the world that was ideally suited for its various heavy industries. Now, with just 7% annual wind energy and 3% annual solar energy, the German power system needs to be rescued hundreds of times a year to prevent power outages. What would happen with 20% annual wind energy and 6% annual solar energy? 


Note: According to Tennet, one of the four system operators in Germany, the number of network interventions to stabilize the power system rose to several hundred in 2011. Just a few years ago, grid operators only had to intervene around 15 times a year to ensure a reliable and stable power supply. The grids of Texas, Colorado, etc., have similar experiences.




Wind Energy Integration Fees


For a proper evaluation of wind energy cost, the total would have to include not only the LCOE of the wind turbines, but also all or part of the LCOEs of:


– Increased regulating plant operation for grid stability; extra fuel and CO2

– Increased spinning plant operation; extra fuel and CO2

– Increased start/stop operations; extra fuel and CO2

– Increased part-load operation; less efficient, extra fuel and CO2 

– Increased part-load-ramping operation; less efficient, extra fuel and CO2

– Increased wear and tear of equipment of generating units  

– staffing, fueling and operation of most of the existing generating units 

– less than optimum economical scheduling of plants due to wind energy on the grid

– less economical operation of existing plants due to wind energy on the grid

– expanded transmission and distribution systems

– increased grid management systems, staffing and operation 

– increased weather and wind speed forecasting systems, staffing and operation


Rarely are any of these costs identified, quantified and charged to wind turbine owners as wind energy integration fees, i.e., they are getting a free ride.  


The above costs are not yet separately identified and quantified by grid operators, generator owners and utilities, because heretofore they have been relatively minor. But as wind energy percent increases, they will be come increasingly greater expenses, as experienced by other grids with greater than about 3% annual wind energy.


Grid operators typically add their extra costs to the invoices sent to utilities and generator owners that supply the grid.


Utilities typically add their extra costs to their other costs to justify rate increases. How generator owners will be compensated for the adverse impact of wind energy on the economics of their generators remains an open question.


Legislators, who wear the “RE” label to get votes, and utilities, dependent on rate increases from legislatures, are loathe to investigate, as it would be considered adverse to RE. They usually work together to make these costs “disappear”, i.e., “socializing” them, by rolling any RE costs mostly into household rate schedules.


Denmark, an RE role model, has done it for decades and Danish households “enjoy” the highest electric rates in Europe (about 31.5 euro cent/kWh), Germany’s households “enjoy” the second highest rates (about 27 euro cent/kWh), France enjoys the lowest (about 12 euro cent/kWh).


The lowest-cost wind energy balancing is with hydro plants. Higher cost wind energy balancing is with gas-fired, quick-ramping gas turbines. The below costs do not cover increased wear and tear, and grid extensions and reinforcements, etc.


– Denmark, “borrowing” Norway’s hydro plants, claims the cost at about 1 – 4 euro/MWh

– Hydro-Quebec, using its hydro plants, charges wind turbine owners $5/MWh. 

– The Bonneville Power Authority, BPA, using its hydro and gas turbine plants, charges $5.7/MWh.

– The Netherlands, using its gas turbine plants, charges 10 euro/MWh.

– Public Service of Colorado claims balancing costs at about $10/MWh


Below is a recent study of wind integration costs for various nations that includes the costs of balancing, increased wear and tear, and grid extensions and reinforcements. In the US, at 10% annual wind energy on the grid, the “grid level cost” is estimated at $16.30/MWh. See page 8 of below URL


Parasitic Energy Demand of Wind Turbines


Wind turbines need energy for their own operation 24/7/365. The parasitic energy demand can be 10%-15% of rated output on cold winter days, whether operating or not. 


Example: The average Danish Vestas-V82 wind turbine produces about 1,650 kW x 0.228 (2007 CF) = 376 kW. The AVERAGE power draw from the grid to keep itself running is about 50 kW and at times up to 80 kW. Thus, a V82 operating in Denmark has an annual average brochure output of about 376 kW +50 kW = 426 kW, but an actual output of about 376 kW, about 13%  less than advertised in Vendor brochures. No wonder Vendors keep quiet about parasitic energy; at 426 kW, the CF would have been 426/1,650 = 0.258.


Example: The Enercon-82, capacity 2 MW, hub height 138 m (460 ft), rotor diameter 82 m (273 ft), for a total height of about 600 ft. The unit requires a substantial foundation. The installed cost is about $2,600/kW. 


The units have a fan with an electric heater and duct system in each blade to circulate warm air through the uninsulated, hollow blade to keep it warm in winter to prevent icing that impairs blade aerodynamics, as on an airplane wing. Total power draw of the blade heating system is about 60 kW, plus about 50 kW for other parasitic loads. 


Comparison of Wind Energy with Advanced CCGT Energy 


The US Energy Information Administration projects levelized production costs (national averages, excluding subsidies) of NEW plants coming on line in 2016 as follows (2009$):


Offshore wind $0.2432/kWh 


PV solar $0.2107/kWh (significantly greater in marginal solar energy areas, such as New England) 


Onshore wind $0.097/kWh (significantly greater in marginal wind energy areas with greater capital and O&M costs, such as on ridge lines in New England; less in the Great Plains states) 


Conventional new coal (base-loaded) $0.0948/kWh 


Advanced 60%+ efficient CCGT (base-loaded) $0.0631/kWh.


Without subsidies, the US average LCOE of onshore wind energy is about 0.097/0.0631= 0.52 times greater than adv. CCGT. 


Without subsidies, the US average LCOE of offshore wind energy is about 0.2432/0.0631 = 3.85 times greater than adv. CCGT, because of much greater (owning + O&M) costs. 


The below table summarizes capital costs, O&M, capacity factors and unsubsidized LCOEs for recently-built/proposed IWT systems in different regions. The Great Plains, GP, has the least cost O&M; it is set at 1; New England, ridgeline O&M is about 2x GP; New England, offshore O&M is 3-4 times GP.


                                                    Cap Cost            O&M       CF       LCOE       LCOE

                                                        $/kW                                      $/kW      $/kW

                                                                                                      WO/sub  W/sub


New England grid price                                                                                   0.055                 

Great Plains                                     1,800                1         0.40     0.090     0.700

New England, ridgeline               2,500 – 2,800          2         0.32     0.150     0.100

New England, offshore                     4,200                3-4      0.40     0.243     0.170


Note: The above wind energy LCOEs do not include all of the LCOEs listed under “Wind Energy integration Fees” in this article.


Note: The New England energy costs are based on CFs of 0.32 or better, and IWT useful service lives of 25 years. Actual CFs on New England ridge lines are 0.25 or less, based on FERC data, and actual lives of IWTs are 15 – 20 years, say 17.5 years, based on actual lives of IWTs in the UK and Denmark. See this article for details.


Without subsidies, the LCOE, NE ridgelines, would be at least 0.15/0.0631= 2.38 times greater than adv. CCGT.


With subsidies, the LCOE, NE ridgelines, would be at least 0.10/0.0631= 1.58 times greater than adv. CCGT.


If the production tax credit of $0.022/kWh expired, the LCOE would be about {(0.10 + 0.022)/0.0631} = 1.93 times greater than adv. CCGT.



– Maine wind turbine facilities have an average installed cost of about $2,500/kW and a vendor-predicted average capacity factor of 0.32; actual ridge line capacity factor is 0.25 or less. 

– The Granite Reliable Power Windpark, Coos County, NH, has 33 Vestas units @ 3 MW each, capital cost $2,778/kW. 


A Lack of Real-time, 1/4-hr Data Hampers Proper Inquiry


The above study of the Irish grid could only be performed, because EirGrid publishes real-time, 1/4- hour grid operations data. Almost all grid operators HAVE those data, as they need them to properly operate their grids and for economic dispatch, but do not publish them because:


– they are not required to, or they do not want to.

– wind turbine owners claim their data are proprietary.

– wind turbine vendors and owners have lobbied legislatures to maintain the “do-not-tell” status quo. 


Note: RE departments of governments and other organizations are filled with people who would not have their wind energy jobs and federal and state wind energy subsidies would be at risk, if grid operators were required by law to publish real-time, 1/4-hr grid operations data,  


Because the real-time, 1/4-hr grid operations data is generally not made public, it became possible for government leaders and wind energy promoters to make unrealistic CO2 emissions reduction claims, such as the 1 : 1 ratio, using studies based on estimates, probabilities, algorithms, assumptions, grid operations modeling, weather and wind speed forecasts, etc., and thereby maintain a spell of deception and delusion regarding the claimed CO2 emission reduction benefits of wind energy.


Whereas such studies are costly, complex, look impressive, and create the appearance of serious inquiry to the lay public (many of those studies are performed and/or financed by wind energy promoters, such as the AWEA, US DOE, NRELs, et al), simulation studies usually introduce subjective elements (such as minor tweaking of the values of the study parameters or omitting certain aspects) that skew the results and conclusions more favorable to wind energy. 


The performers of such studies usually claim they have to do them the simulation way, because real-time, 1/4-hr data is not available. However, they are well aware, as are most energy systems analysts managing electric grids, real-time, 1/4-hour grid operations data, as published by EirGrid, are superior to any simulations for performing wind energy studies, an inconvenient truth for wind energy promoters. 


As a result of the similar methodologies, these simulation studies tend to produce similar outcomes in favor of wind energy, which reinforces the orthodoxy of wind energy promoters, such as the AWEA, US DOE, NRELs, et al. Any study at variance with their orthodoxy, such as of the Irish grid, are readily denounced/debunked as biased, a special case, cherry-picking, using statistical trickery, etc.


The above shows:


– the lay public has been led to believe by government leaders and wind energy promoters that wind energy is “fighting climate change and global warming”. Building out the wind energy “solution” would be much less attractive regarding CO2 emissions reduction and more costly/kWh than meets the eye.


– too many renewable energy certificates, RECs, are being granted to wind energy producers than is warranted based on their actual CO2 emissions reduction.


Additional References Showing a Lack of CO2 Emissions Reductions: 


Bentek Energy LLC, How Less Became More: Wind, Power and Unintended Consequences in the Colorado Energy Market,


Institute for Energy Research, June 2010:


Argonne National Laboratory, System-Wide Emissions Implications of Increased Wind Power Penetration, March 5, 2012;


Argonne National Laboratory, Grid realities cancel out some of wind power’s carbon savings, May 29, 2012;


Forbes, Wind Power May Not Reduce Carbon Emissions As Expected: Argonne, May 30, 2012;


Aol Energy, A Brave New World: Renewable Energy Without Subsidies, June 6, 2012;


Bloomberg, Renewable-Power Boom Leaves Nations Without Backup, Report Shows, June 8, 2012;


Climate Wire, Renewable Energy: Wind power may not reduce carbon emissions as expected, June 1, 2012;



Content Discussion

Michael Goggin's picture
Michael Goggin on July 3, 2012

Unfortunately for Mr. Post, repeating a myth does not make it true. Regular readers of EnergyCollective have probably seen a nearly identical posting from Mr. Post almost a dozen times already. In fact, since everything in this posting has already been debunked, I’ll just copy and paste the previous debunkings below. However, I would be curious how Mr. Post would attempt to explain away this inconvenient truth, which conclusively tests and refutes his entire hypothesis:

The Department of Energy collects detailed data on the amount of fossil fuels consumed at power plants, as well as the amount of electricity produced by those power plants. By comparing how the efficiency of power plants has changed in states that have added significant amounts of wind energy against how it has changed in states that have not, one can test the hypothesis that wind energy is having a negative impact on the efficiency of fossil-fired power plants. The data clearly shows that there is no such relationship, and in fact states that use more wind energy have seen greater improvements in the efficiency of their fossil-fired power plants than states that use less wind energy. Specifically, coal plants in the 20 states that obtain the most electricity from wind saw their efficiency decline by only 1.00% between 2005 and 2010, versus 2.65% in the 30 other states. Increases in the efficiency at natural gas power plants were virtually identical in the top 20 wind states and the other states, at 1.89% and 2.03% improvements respectively. The conclusion that adding wind energy actually increases fossil plant efficiency makes intuitive sense, because adding wind energy to the grid displaces the output of the most expensive, and therefore least efficient, fossil-fired power plants first.


Previous debunking of playing statistical tricks with Irish grid operator data:

Regarding Fred Udo’s report on wind and emissions in Ireland, it only took a few minutes to unravel the statistical trick Mr. Udo was using to get his results, which might explain why his analysis wasn’t published in a peer-reviewed journal and rather appears on an obscure Dutch anti-wind website.

This appears to be a classic case of a lurking (or confounding) variable being used to misleadingly present correlation as causality; a comparable example is arguing that cigarette lighters cause lung disease since people who buy them tend to develop lung disease. In this case, the lurking variable that is the actual causal factor appears to be cold weather and its impact on heating demand, data that Mr. Udo should have had access to but that (for reasons we can only speculate) he chose not to use in his correlational analysis.

What tipped me off was part 3, Figure 3 of his text, where Mr. Udo called out an event in Ireland around June 9-12, 2011, when the carbon intensity of Ireland’s electricity production surged. I was curious as to what might have caused that event so, on a hunch, I pulled weather records for Ireland. Sure enough, there was an abnormally cold spell when temperatures fell into the 30’s and 40’s F, 10 to 20 degrees below normal for that time of year. Aha! Cold temperatures cause a spike electric heating demand, causing the grid operator to turn on more expensive, less-efficient fossil plants to operate to meet the abnormally high electric demand.

Another factor is that cold weather could force some of Ireland’s fossil-fired combined heat and power (CHP) plants to fire up and run at a high level of heat production (and subsequently more emissions per megawatt-hour, MWh, of electricity, since CHP plants relative to the rest of the fleet are not optimized for electricity production, and CHP plants being run to produce maximum heat are not being operated in a way that is optimized for electricity production; moreover, it appears that the emissions associated with heat production are rolled into the data that Mr. Udo is using, so a CHP plant producing only or mostly heat and little or no electricity under cold conditions like these would score at infinite emissions/MWh).

As one would expect, cold spells and home heating demand often correlate with high wind speeds, which is how Mr. Udo was able to draw his false conclusion that wind was the causal factor. Sure enough, a closer examination of the spikes in emissions/MWh in his data show that all are associated with cold spells, and only some are associated with an increase in wind output. It doesn’t take a statistician to tell you which is the causal factor in that relationship. Had Mr. Udo himself been more interested in finding the actual causal relationship at play here, he might have noted that the correlations between wind output and emissions intensity varied widely from month to month (as one would expect for weather-driven seasonal changes in electric demand), usually a strong indication that another variable may be the actual causal factor.

I should also point out that, contrary to Mr. Udo’s claims, the method Irish utility system operator EirGrid uses to calculate emissions savings from wind is accurate. The plant-specific heat rate curve that they are using would account for all of the impacts wind energy would have on the efficiency of the fossil fleets under all operating conditions.

Summaries of government and grid operator data showing emissions have decreased as or more than expected in Colorado, Texas, and other regions as wind energy has been added to the grid, directly refuting Mr. Post’s claims to the contrary:

Michael Goggin

American Wind Energy Association

Jorge Montero's picture
Jorge Montero on July 3, 2012

The analysis is good for Ireland.  But it really depends on the composition of your generation base, you cannot generalize.  Here in Costa Rica, we provide backup and other auxiliary services to the network with hydropower.


Michael Goggin's picture
Michael Goggin on July 3, 2012


I see you were unable to offer any response at all to the DOE data that completely disproves your entire argument. If adding wind energy really does decrease the efficiency of fossil power plants, why has the efficiency of fossil power plants improved more in states that have added wind energy versus those that haven’t? If you are unable to answer that simple question, I’d suggest that you stop repeatedly spamming this website with the same claims that have been thoroughly debunked and which you yourself cannot defend, if for nothing else than for the sake of your own credibility.

Also, you should be rather embarrassed about claiming that your work or Udo’s work has been peer-reviewed. Udo’s article did not appear in European Physics Review as you claim, which is a legitimate peer-reviewed journal, but rather in the Europhysics News, which, as the title suggests, is a non-peer-reviewed news magazine that posts news articles on physics topics submitted by volunteers. As the Europhysics News website explains, it publishes “review articles, features on advances topics, news reports and items of general interest.” Regarding your claim that your own work is peer-reviewed, I hope that you realize that having an associate simply read something you are going to post on a blog does not constitute peer review.

Finally, in my initial post I made it clear that I was linking to AWEA’s summaries of government and independent grid operator data and analysis. I’ll assume that you chose to make an ad hominem attack on the messenger (me) because you have no substance on which to refute the dozens of pieces of evidence from independent third parties that I linked to in those summaries. Please follow the dozens of links to those government and independent grid operator data and analyses, read the analyses and look at the data on your own, and please explain to me what is wrong with all of them, as I have done for every piece of fossil industry-funded junk analysis you’ve attempted to pawn off on the readers of this website. I hope you have enough faith in the readers of this website to understand that your incessant re-posting of debunked junk will not win them over. I will be posting a similar comment on each article you post, until you can provide a substantive rebuttal to the facts I have cited.


Paul O's picture
Paul O on July 3, 2012


As a member of American Wind Energy Association, do you have access to real-time, 1/4-hour grid operations data. If so would you be willing to publish them?

This would go a great long way toward resolving the question of CO2 emmissions reductions of wind farms.

Kimi Arima's picture
Kimi Arima on July 5, 2012

Hi all, 

interesting discussion. I must, however, point out that there is an established alternative to using OCGTs for wind balancing, namely, combustion engines. A modular power plant based on a multi-unit approach yields superior part-load efficiency, as individual units can be shut down and started up regardless of the other units, and all bar one of the units running can be running on full load. 

Moreover, from start up to full output takes less than 5 minutes for a combustion engine, further reducing the efficiency penalty of cyclic operation. Finally, combustion engines exhibit little to no additional wear and tear from starts/stops and cyclic operation. For instance, there’s a 200 MW plant in Colorado, US, called Plains End, that gets hundreds of starts and stops annually, yet shows little additional cost in maintenance. 

If you’d like to know more on the subject, we published a book on this about a year ago, a free PDF version of which is available on our website 

Kimi Arima's picture
Kimi Arima on July 6, 2012

A study by Black & Veatch a while back showed that for intermediate and peaking operating profiles (< 5000 hours per year), gas-fired combustion engine plants beat any OCGT plant in lifetime feasibility. This is just not a well-known fact, as the “standard” approach for energy consultancies is to test feasibility on the assumption of 8000 hours per year base load operation (and why not, since all GTs behave rather similarly in cyclic operation, and combustion engines are usually not considered as an option).

Regarding pure base load plants, you are correct, modern CCGTs are the most efficient solution. However, looking at recent developments in some Western countries, it is uncertain how much of pure base load capacity will be required. For instance, Spanish utilities built vast amounts of CCGT capacity during the last decade, but due to the increase of wind and solar, the average capacity factor for these brand new, very efficient CCGTs now sits around 30%. In private discussions some utility people have confided that they’re incurring heavy losses and only running the CCGTs because of take-or-pay gas contracts. 

Regarding capacity, 1/4 of Wärtsilä’s installed base is in power plants. And, regarding the common conception of “diesel engines”: nowadays a majority of the engines Wärtsilä sells for power plants are gas-fired.

Kimi Arima's picture
Kimi Arima on July 6, 2012

I agree with you on O&M. Nevertheless, Wärtsilä has some very nice references in North America, and the tide might be turning, at least in states with a lot of installed wind capacity. 

I’d venture a guess and say that moving downstream and into the generation business has been considered at the highest levels at Wärtsilä, but that would perhaps necessitate an entirely different set of competencies, one that might not be so easily acquired. 

Enjoyed the chat, wish you a nice weekend. Best regards, Kimi Arima

Eirik Johnson's picture
Eirik Johnson on July 6, 2012

Mr Post’s observation that traditional steam generators waste fuel when called upon to shift on and off to accomodate demand shifts due to the intermittent nature of wind power is an indictment of fuel-based generators, not an attack on wind power.  The existing power grid was not designed as a back-up for a superior technology, and so there’s no surprise if it’s not ideal for that purpose.  This is simply one more reason to demand a new grid, one designed for modern purposes, rather than to support investing more and more money in the overcentralized nineteenth-century technology of the old grid.  I think I’ve seen several posts explicating better than I can why the investment level appropriate for maintaining the old grid as a back-up system is much lower than the level appropriate for a new one–whether or not it is rooted partly in renewable-energy technology.  I can’t recall the name, but  I wish I could thank the contributor who pointed out that Mr. Post’s numbers have been “debunked,” which seems reasonable.  A single massive wind generator field might have the sort of irregularity that Mr. Post describes, but multiple small-scale generators spread out over a large area wouldn’t.  If it isn’t windy on the North Side of Chicago at the moment, it probably is on the West Side or the South Side.   It’s true that a whole week or month might be pretty calm citywide, but turning steam generators on or off once a week or a month is likely part of safety procedures in any case. 

Paul Lindsey's picture
Paul Lindsey on July 6, 2012

I wish the AWEA and wind farm operators published near-real time and historical data in 15-minute blocks. As far as I know, only the BPA publishes their data, and everyone can watch over 4400 MW of installed wind power CAPACITY ebb and flow, Wow, since 0200, July 4, 2012, to present, BPA’s wind has been running at less than 10% capacity and was near 0% for 12 hours on July 4th! AWEA must hate that the BPA publishes this data.

The only other way to find out what a wind farm is producing is to wait for the owner to post the previous quarter’s data to the FERC website. Unfortunately, most of the data is in monthly increments and only occasionally in hourly chunks, completely masking the inherent variability of wind power. Sellers of power aren’t required to submit their data until the end of the month foloowing the end of the quarter (i.e. July 31st for the Apr-June qtr). The data also has potential “holes”, because if the power generator is producing nothing, then the hour isn’t reported. This typically happens with solar PV arrays. So don’t just run an AVERAGE function on a column without making sure that there are enough hours accounted for in the data. Of course, data reported in monthly chunks has the zero-power periods completely hidden.

Eirik Johnson's picture
Eirik Johnson on July 6, 2012

  It would appear that my unfortunately casual response to Willem Post’s post has aggravated what seems to be well-founded annoyance at being ignored, and I hope that he will accept this follow-up by way of apology.  How-ever, the fact remains that the inability of traditional power grids to cope with appropriate technology is a shortcoming of the power grid, not a problem with appropriate technology.

   It is also true, as Mr. Post points out, that his concern is not a mere “myth” and does make it misleading to claim that wind power can “replace” old-fashioned technology in simple, direct proportion.  But as SmartPower suggests in its response, ICs (internal combustion) power might.  As best I can understand, Mr. Post has given good reason to think that Wind and Solar do not directly reduce the need for fuel-based power capacity, but concedes that the capacity which it seems to replace could do so if be paired with IC.  Thus, the financial cost of using Wind and Solar to patch up the grid must include the cost of IC units which will have very low capacity ratings, but that would not change the ability of appripriate technology to provide power with significantly lower environmental cost than the currrent grid does.  Appropriate technology may not be able to solve the financial or reliability problems of the grid at all.

   And those problems seem to be far more grave than I have ever considered before: according to the example offered by Mr. Post, the best power grid in the world can’t cope with variability in power contributions that account for only 10% of its total size.  If we have a power grid which is worse than Germany’s, and that one can’t handle 10%, then we desperately need something better.  And frankly, the condition of US grid at the moment, with millions of East-coasters going without power, suggests that the implications of Mr. Post’s report are right.  If Germany’s is the best a grid can do, and it reacts to a sunny day by dumping power in a panic on the neighbors, then the time has come to try something better than a power grid–with or without IC or wind power.

   I don’t see how a little extra help from wind and solar can make it any harder to nurse the old grid along until we get something better—unless making money rather than providing power is the ultimate purpose of the grid, in which case that PR guy that Mr. Post’s been quarrelling with is probably more of an expert than I am, given that my degrees are merely in history and law. 

Alain Verbeke's picture
Alain Verbeke on July 7, 2012
Dear michaelgoggin,

I wholeheartedly agree with your points.

Don’t even bother to reply to Mr Post. He doesn’t have a case. Spain achieves 25% of it’s electricity supplied by wind power only, Spain being a 40 million country with very few transmission grid connections with neighboring countries (the Iberic island). This 25% penetration is measured over a year. On very windy days, their wind turbine electricity feed is so huge, that they have to shut down wind turbines. Spain is building huge pumped up hydro storage plants to backup wind turbine parks, to balance their grid.

I myself am supplied 100% day and night with electricity sourced from renewable fuels, and produced by EcoPower in belgium, them using wind turbines, biodiesel piston generators, run-of-river hydropower, biomass cogen plants and biogas piston / turbine generators to send power to their 35 000 customers base that is growing at a 15% clip each year.

So whatever ‘study’ Mr Post references, the european customer is voting in the real world with his euro’s to get electricity sourced from proper energy resources supplied by a market that allows that.

Met vriendelijke groeten,

Alain from Belgium

Editor’s Note: This comment has been modified from it’s original form.

Eirik Johnson's picture
Eirik Johnson on July 7, 2012

I won’t waste space here justifying an apology. 

Simon Friedrich's picture
Simon Friedrich on July 8, 2012

Dear Alain, 

Please note biodiesel and other biomass fuels generate greenhouse gases. Also, the use of biodiesel in Europe has contributed to deforestation in Southeast Asia and additional greenhouse gas emissions.

The use of the word “Pest” (Post) is not a polite way to address somebody, particularly in Flemish (Dutch). 


Simon Friedrich

Paul O's picture
Paul O on July 13, 2012

Before joining TEC, I had personally refused to participate in sites where GW and RE were the subjects of debate.

Frequently on these sites extremely harsh and strident tone of comments are used, and pejorative labels such as "denier" was being employed to keep people who wanted to dispute the widely accepted beliefs about GW and RE.

Since joining TEC however, I am unable to recall another comment that has given me as much indignation as this one referring to Mr. Post as "Pest". I am confident that the people in  charge of overseeing this site will act accordingly.


David Thorpe's picture
David Thorpe on July 24, 2012


Why do you keep repeating things that other commentators have debunked? Your incapacity to listen, learn and develop your argument does you little credit.


David Thorpe's picture
David Thorpe on July 24, 2012

I know all that, you've said it before. You're repeating yourself. I do read what you write. I like especially the paragraph on energy efficiency. But please answer my question.

Paul O's picture
Paul O on July 24, 2012


I too would like to hear/read a response from Willem...

Would you kindly care to be more specific, POINT BY POINT, on precisely which debunked statement/comments that Willem is/has been repeating.

IOTW, please tell us which comment he made has been debunked, and if possible link us to the debunk iteslf. I think this kind of connversation is sorely needed and very very helpful in these Energy Policy debates.

Thanks Dave.

David Thorpe's picture
David Thorpe on July 25, 2012

Here is an example of a point which you do not address: the definition raised of what counts as peer reviewed. You don't seem to understand what this means!

It's not that what you say isn't interesting, is that if we want to find out the truth, which your contribution might help us towards, we need to develop our understanding.

Just because your post is the most popular doesn't mean you are automatically completely in the right!

Your point seems to be that emission savings are inaccurately attributed to the addition of wind turbines to the National Grid. But by my understanding emissions are calculated from actual emissions from actual fossil fuel powered power stations. Therefore, there is no false accounting.

The technology that we are introducing is evolving all the time, the 1st industrial revolution did not happen overnight. Bicycles were faster than the first cars, but that was not a reason to discontinue their development. As the smart grid is introduced, as more kinds of energy storage are developed, and as other forms of more efficient renewable energy enter the system, over the next 30 to 40 years, we will move to a situation where most of our energy comes from renewable sources. Decarbonisation of the grid is happening simultaneously on many fronts, at least here in Europe.

David Thorpe's picture
David Thorpe on July 25, 2012

What you don't take into account the cost of not doing all these things. Not just now, but for the sake of the future.