by Willem Post


The New England Electric Grid, NEEG, managed by ISO New England, ISO-NE, has a generating capacity of about 34,020 MW, power supplied is about 130,000 GWh/yr. It includes over 350 central power plants and 8,000 miles of high-voltage transmission lines to provide power to about 6.5 million customers. The NEEG power is 62% from CO2-producing fossil fuels, 26% from CO2-free nuclear, 6% from CO2-free hydro, 4% from CO2-producing wood waste, 2% from CO2-producing solid waste and less than 1% Other i.e., CO2-free wind, solar, etc. Almost all of this power is STEADY power and the T&D systems of the NEEG are designed accordingly.

Historically, electric grids have experienced varying electric demands during a day and varied the output of their generating plants to serve that demand and, at the same time, regulate frequency.

Cold, quick-starting, quick-ramping peaking plants, such as a mix of gas-fired OCGTs and CCGTs, are turned on and off each day to serve normal peak demands which occur once or twice per day. From a cold start, CCGTs take about an hour before there is enough steam pressure to operate the steam cycle. During this hour they run as OCGTs at about 25 to 30% efficiency, instead of the 55 to 59% efficiency as CCGTs.

Base-loaded coal and nuclear plants, which take about 6-12 hours from a cold start to rated output, are less suitable for variable output operation. Usually they operate continuously near rated output for about a year for coal plants, for about 1.5 years for nuclear plants, after which they shut down for several weeks for maintenance and refueling.

Base-loaded coal plants, designed for most economical, least polluting, steady operation near rated output, are often used to follow daily demand profiles and are sometimes used for the frequent cycling operations required by wind power; the coal plants need to be designed for ramp rates of 5-10 MW/min for a 500 MW plant. Cycling operations of coal plants require more fuel per kWh and emit more pollutants, including SOX, NOX, CO2, and particulate, per kWh, as shown by coal plants used for cycling in Colorado, Texas, etc. The main reason for using coal plants for cycling is because their grids lack sufficient capacity of pumped-storage hydro plants and gas-fired OCGT and CCGT plants that are utility-owned,

Base-loaded nuclear plants, designed for most economical, steady operation near rated output, are rarely used for cycling operations.

The increased generation of wind power would present additional challenges to the management of the power on the NEEG. Because wind power is variable and intermittent, additional spinning and backup plants, such as a mix of OCGTs and CCGTs, must be kept in continuous operation to supply and withdraw power as required. The plants must respond to:

– changes in electric demand of millions of users during a day.

– changes in power supply, such as from unscheduled plant outages.

– changes due to weather events, such as lightning, icing and winds knocking out power lines.

– changes in wind power.

If these changes, especially those due to wind power, are of high MW/min, the CCGTs may have to temporarily operate as OCGTs at about 25 to 30% efficiency, instead of 55 to 59%, because the heat recovery steam generators, HRSGs, of the CCGTs have slower ramp rates than OCGTs. This mode of operation, MO, increases fuel consumption and NOX and CO2 emissions per kWh.

Independent Power Producers and Power Purchase Agreements

Independent power producers, IPPs, usually sell their power to public utilities under long-term power purchase agreements, PPAs. The IPPs have greater profits, if they operate their plants steadily, near rated output. They do not make their plants available for cycling, as it would reduce their output and profits. Accordingly, only the plants owned/controlled by public utilities are available for cycling. Public utilities operate on a cost-plus basis; their justifiable costs due to cycling would be made up with rate increases, or a fee on wind facility owners, or both.

The IPPs of existing wind facilities on the NEEG have PPAs with utilities. These IPPs have their wind facilities, about 250 MW, primarily in Maine, in the best locations and have been getting a free ride, as their wind power impact on NEEG regulation and spinning plant operations is nearly undetectable.

As NEEG wind power penetration increases, the impact would be detectable and stricter requirements regarding frequency regulation and variability of wind facility outputs would be required, instead of placing the onus on ISO-NE, NEEG plants and ultimately rate payers. Compliance with such stricter requirements is being deployed in higher wind penetration nations, such as Spain, Ireland, Germany, etc.

Collecting Power Plant Operating Data

ISO-NE would be well advised to set up centralized logging of wind power production of all wind facilities at 15-minute intervals and for fuel consumption of the plants on the NEEG; it would be much better to have minute-by-minute data logging which could be implemented at about the same capital cost. Many of these plants already report hour-to-hour emissions of NOX, SOX and particulate to the US-EPA. All plants on the NEEG should also be required to report CO2 emissions at 15-minute intervals. Such baseline information is essential for any accurate analysis and comparison of power generation alternatives to reduce such CO2 emissions; flying blind regarding global warming is inexcusable. The Electric Reliability Council of Texas, ERCOT, collects such data at 15-minute intervals.


The purpose and approach of this study is to:

– determine the impacts of several large, medium and small wind power decreases on CO2 emissions during a day.

– determine the capacity of the cycling facility required to immediately ramp up or ramp down to accommodate any wind power changes.

– assume a total of 10,000 MW of onshore wind power has been implemented in the NEEG service area.

– compare wind power with CCGT power.

– propose increased energy efficiency as a more cost-effective CO2 reduction strategy than wind power.

– summarize the BENTEK study “How Less Became More” regarding using coal plants to accommodate wind power in Colorado and Texas.


Because wind power is proportional to the cube of wind speed, a doubling of wind speed, which are frequently occurring events during a day, would increase wind power 8-fold. Wind power often varies by a factor of ten or more during a few hours which is quite shock to the stability of electric grids, if an adequate capacity cycling facility is not available.

The 10,000 MW was chosen because the onshore wind power capacity of New England sites with Wind Class 3 or better winds is estimated at 10,989 MW. See page 4 of

The New England average capacity factor, CF, of 0.31 was chosen because early installed wind facilities would likely be in the best wind areas with higher CFs, such as the wind facilities in the west of Maine which have an average CF = 0.32, whereas later installed wind facilities would be in poorer wind areas with CFs of 0.30 or less.

Some wind power proponents make optimistic statements regarding wind power CFs calculated from wind speed measurements for a period of one to three years, but actual operating experience proves otherwise. The lower CFs are partially due to wind power curtailments to avoid excessive cycling stresses on the power plants connected to the grids and stresses on transmission systems and increased O&M downtime than anticipated.

For example: In Denmark, all wind turbines are monitored and controlled by Vestas from a single location to avoid excessive wind power increases that would overwhelm Denmark’s’ small grid and that are not wanted by Norway, Sweden and Germany for various reasons; Vestas feathers the blades of the rotors or stops them entirely.

The above New England average CF of 0.31 may prove to be very optimistic, because large geographical areas rarely have capacity factors greater than 0.30. For comparison: Ireland (0.323 for the 2002-2009 period, the best in Europe), UK (0.282), Texas (0.258 for 2009), Denmark (0.242 for the 2005-2009 period), the Netherlands (0.186), Germany (0.167). It would not be credible to aver onshore wind speeds in New England are comparable to onshore wind speeds in Ireland, one of the windiest areas of Europe.


The capital cost for the assumed wind power facilities would be 10,000 MW x $2,500,000/MW = $25 billion (2010 dollars)

In addition, several billion dollars for upgrading and expanding transmission and distribution, T&D, systems would be required.

A wind turbine installation cost of $2,500,000/MW is assumed for this study. It is the same as the average of the capital costs of the recently installed operating and planned wind facilities in Maine.

As the ridge lines in New England have the best winds, it is assumed almost all wind turbines would be located on them. If the wind turbines are a 50/50 mix of 2.5 MW and 3 MW units, then about 3,670 units/(7 units per mile of ridge line) = 524 miles of ridge line would be required. The wind turbines would be located in areas with the best winds, such as on the ridge lines of the north-south spine of Vermont, northern New Hampshire and western Maine; the latter two areas have greater average wind speeds than Vermont.

Such a concentration of wind power facilities would yield less of a reduction of wind power variability normally associated with widespread geographic distribution of wind facilities, i.e., the variability of wind power would be reduced if some areas temporarily seeing higher wind speeds are combined with areas simultaneously seeing lower speed winds.

The 2.5 MW and 3 MW units are about 390 to 415 ft tall to the tip of the blade, respectively, which would appear quite large if the ridge line is at 2,000 ft elevation and a person’s house is at 1,000 ft elevation and within a mile of a row of wind turbines; at night there would be an unsteady beat of whoosh sounds. Wind turbines are often made to look small on distant ridge lines using Adobe’s Photoshop software.


Quick-ramping, spinning OCGT plants are necessary to continuously maintain the 60 Hz frequency of the grid within a narrow band. The OCGTs in frequency-response mode, under nominal conditions, would run at reduced output to maintain a buffer of spare capacity and would continually alter their outputs on a second-to-second basis to maintain frequency with a so-called droop speed control.

When the demand exceeds the supply (including back-up spinning reserve), the voltage and frequency drop, increasing loss-of-load-probability (LOLP). Even small changes in frequency or voltage (either positive or negative) can significantly increase the LOLP. Loss of load implies blackouts and/or brownouts.

A demand increase, or a supply decrease, or both at the same time, would cause the generators on the grid to slow down. Their synchronizers would sense this and more steam, or fuel, is supplied to the generators which would increase their RPMs. Some generators are slow to react, others are faster, such as spinning OCGTs, which would immediately ramp up until the others catch up. Variable wind power on the grid acts as a supply increaser and supply decreaser 24/7/365. Accordingly, greater regulating capacity is required with increased wind power penetration.

With the current wind power on the NEEG, the hourly capacity used for frequency regulation varies from a low of 30 MW (overnight on weekends) to a high of 200 MW (spring morning load pickup), a 7:1 ratio. Over all hours of 2008, the weighted average hourly regulation, WAHR, was about 80 MW. The addition of wind power would increase the real-time variability and short-term uncertainty of the power supply. See Page 171 of New England Wind Integration Study, NEWIS.

Based on a statistical analysis of ISO-NE grid operating data and on various other sources of wind data, the WAHR capacity would increase from the above about 80 MW with current wind power, to a high of 315 MW at 20% wind penetration, 230 MW at 14%, 160 MW at 9% and 100 MW at 2.5%. See page 182 of NEWIS. Note that the WAHR values are AVERAGES. With very large wind power decreases, as often happens during New England’s unstable weather conditions, the required quick-ramping capacity would be much larger, as shown below.

The annual owning and O&M costs of operating a 4-fold increase in WAHR capacity would be significantly. Would owners of wind power facilities pay those additional costs as part of accommodating wind power?


Because the normal daily demand profile from hour-to-hour and day-to-day is know with some certainty, any normal rising demand is also known with some certainty and the appropriate spinning and peaking capacity could be deployed to provide power for that demand.

But the magnitude and duration of steep, large wind power decreases are not known with adequate and timely certainty, especially in an area with highly variable weather, such as New England. Hundreds of weather stations all over New England, and beyond New England, would be required to monitor atmospheric conditions to predict, using computer programs, hourly forecasts of wind power output for the next 48 hours, updated every 15 minutes, to be useful to grid operators, such as ISO-NE, for making day-ahead, unit-commitment decisions concerning which units to turn on and when to do so; such wind power prediction systems are in operation, in Denmark, Spain, Germany, etc., with mixed success.

The annual owning and O&M costs of such a facility would be significantly. Would owners of wind facilities pay those costs as part of accommodating wind power?

The appropriate deployment of cycling plants to supply power during any steep wind power decreases would likely involve some guess work that must err on the safe side to avoid brownouts, etc. Adequate capacity of quick-ramping, spinning plants must be deployed, if the monitored atmospheric conditions give any indications of sudden, steep and large wind power changes. If the cycling plants were to run out of ramping range, i.e., reach rated output, because wind power decreased too much, then wind power curtailment by feathering the rotor blades or stopping them (practiced in Denmark, Spain, Portugal, Germany, Texas, etc.) and load shedding would be required.

Modern wind facilities could be designed to present less variable power to the grid, such as by feathering their rotor blades to limit minute-by-minute ramp rates of their output, at moderate reductions of power production and at moderate additions to capital costs; Ireland enacted a new grid code in 2004 that requires wind power facilities to reduce the variability of their outputs, i.e., be more “grid-friendly.”

See Page 17 of

Because wind speeds in some geographical areas usually are higher between 9 PM and 5 AM (when demand is least), and higher during winter than during summer, such cycling would occur more frequently during the night than during the day, and during winter than during summer. In other geographical areas, the reverse happens.

Less such cycling capacity would be required if power could be drawn from and sent to other grids that are connected to the NEEG and if the hydro plants, such as of Hydro-Quebec, HQ, would be available for cycling operations. However, the other grids and the HQ service area would likely have their own wind power facilities and may not have, nor be willing to share, spare capacity for NEEG cycling operations. Hence, the generating capacity of the NEEG would need to include sufficient quick-starting, quick-ramping plant capacity to provide power when big decreases in wind power occur, especially while demand is simultaneously increasing. See fig 1 of


Two modes of operation with wind power are described:

Denmark’s Mode of Operation With Wind Power

Denmark’s wind power operations would be used as an example to learn from. Denmark’s prevailing winds are from the North Sea, across Denmark, to the Baltic Sea. The best winds are on Denmark’s northwest coast. Denmark has more than 4,000 onshore wind turbines with a capacity of about 3,150 MW, nearly unchanged since the end of 2003; the increases in capacity in 2009 and 2010 are due to offshore wind turbines. About 90% of the wind turbines are supplied by Vestas, a Danish company.

If the wind blows strongly in Denmark, and as the marginal cost of operating wind turbines is minor (i.e., ignoring major costs, such as for capital and O&M), there is a big incentive to maximize wind power even if it is not needed by Denmark; the solution to avoid congestion on Denmark’s grids is to send the excess power to Norway’s and Sweden’s pumped-storage hydro plants.

Studies of grid operating data show that Denmark exports power to Norway and Sweden, and that those exports are highest during strong wind periods. Norway and Sweden have significant hydro plant capacity, including pumped storage. The plants could be cycled at 100%/min and are highly suited to accommodate variable wind power. The power is subsequently used elsewhere in Scandinavia (which includes Denmark) and Germany.

However, sometimes Norway and Sweden refuse to take Danish wind power because reservoirs are already full, or are going to be full due to snow and rain fall, etc. In that case Vestas, using its computerized control center that controls ALL wind turbines in Denmark, would reduce the output of a percentage of them (by feathering the blades or stopping them), according to pre-planned sequences. This MO is much easier than requiring hundreds of small district heating/electrical plants or a few big central power plants to reduce THEIR outputs. In case of little snow and rain fall, hydro plant reservoir levels may be low and any wind power from Denmark, if available, would be useful to pump water from lower reservoirs to upper reservoirs.

All this back-and-forth gymnastics is inefficient and uneconomical, as various studies have shown. One indication of this inefficiency: Denmark has the highest residential electric rates in Europe, whereas its commercial rates are kept at about one third of the residential rates for international competitive reasons. France, 80% nuclear, has one of the lowest electric rates in Europe.

ISO-NE’s Mode of Operation With Wind Power

Because the NEEG has little pumped-storage hydro plant capacity connected to it, the least costly, least polluting, most suitable way of accommodating wind power is not available to the ISO-NE; it would need OTHER, quick-starting, quick-ramping power plants for accommodating wind power. The least costly, least polluting and most efficient power plants for this purpose are natural gas-fired CCGTs. The US has a secure supply of natural gas for about the next 100 years. The pollutants from natural gas are about an order of magnitude less than from coal.

If 10,000 MW of wind power facilities are implemented in the NEEG service area, the wind power production would be 10,000 MW x 8,760 hrs/yr x New England average CF 0.31 = 27,156 GWh/yr, about 27,156/130.000 = 20.9% of current consumption. The CF gradually decreases as the facility ages. It would be higher in early years, lower on later years, just as a car needs more shop time when it gets older.

Because wind speeds are not uniformly high over all of New England at the same time, the maximum output of the 10,000 MW of wind turbines is assumed at about 8,000 MW which agrees with the experience in other areas with high wind power penetration, such as Denmark, Spain, Portugal, Germany, etc.

ISO-NE would need to have a quick-starting, quick-ramping cycling facility, consisting of a mix of OCGT and CCGT plants with a capacity of 8,889 MW of which 8,889 MW x AF 0.90 = 8,000 MW would have to be available at all times. The facility would produce {(1.0 – 0.31)/0.31} x 27,156 GWh/yr = 60,444 GWh/yr of wind balancing power, about 60,444/130,000 = 46.5% of current NEEG consumption.

The estimated capital cost would be about 8,889 MW x $1,250,000/MW = $11.1 billion, plus augmentation of gas supply lines and transmission systems. Any cycling facility must be utility-owned, not IPP-owned, to ensure it is available for cycling operations.

With predicted weather conditions indicating small wind power variations, about 20% of the cycling facility would need to be in spinning mode at about 50% of rated output so it could ramp up and down all times, the rest would be in production mode near rated output.

With predicted weather conditions indicating medium wind power variations, about 40% of the cycling facility would need to be in spinning mode at about 0% of rated output at all times, the rest would be in production mode at about 50% of rated output to be able to ramp up and ramp down as needed.

With predicted weather conditions indicating large wind power variations, about 60% of the cycling facility would need to be in spinning mode at about 0% of rated output at all times, the rest would be in production mode at about 50% of rated output to be able to ramp up and down as needed.

Assume a worst case: wind power decreases from 8,000 MW to 0 MW in 6 hours, equivalent to 8,000 MW/2 x 6 hrs = 24,000 MWh. An 8,000 MW cycling facility would provide 8,000 MW /2 x 6 hrs = 24,000 MWh of ramp up power. Ramp up rate 8,000 MW/360 min = 22.2 MW/min


The gas turbines of the cycling facility, most efficient near rated output, would have to operate at a less efficient, more polluting, reduced output to be able to immediately vary their outputs to accommodate all variations of wind power, including unpredictable, sudden, large variations of wind power.

Gas turbine heat rates, Btu/kWh, and CO2 emissions, lb of CO2/kWh, increase because of increased inefficient operation below rated output of OCGTs, and CCGTs operating as OCGTs. For example: at 80, 50 and 20 percent of rated output, the heat rates are equal to the rated heat rate divided by 0.95, 0.85 and 0.55, respectively; this is for steady operation at a percentage of rated output. If ramping up, those heat rates increase even more, just as a car speeding up or going uphill has a lower mpg.

Assuming an annual-average heat rate penalty of 20% for operating the cycling facility at less than rated output, the extra CO2 emissions would be 60,444 GWh/yr x 1.298 CH4 leakage factor x 0.20 x 0.726 of CO2/kWh (CCGT basis, 55% eff) = 11.39 billion lb of CO2/yr, or (11.39 billion lb of CO2/yr)/(27,156 GWh of wind power/yr) = 0.419 lb of CO2/kWh of wind power. The daily-average extra CO2 emissions would be 11.39 billion lb of CO2/yr x 1 yr/365 days = 31.2 million lb of CO2/day.

Assuming a long-term contract cost of gas at $4/1,000,000 Btu, the extra fuel cost would be 8,889 MW x AF 0.90 x 8,760 hr/yr x 0.20 x 6,205 Btu/kWh (CCGT basis, 55% eff) x $4/1,000,000 Btu = $347,881,469/yr.

Because the annual-average CO2 emissions of the NEEG is about 1.0 lb of CO2/kWh (low compared to other grids due to CO2-free nuclear and hydro power), wind proponents claim a CO2 emissions reduction of 1.0 lb of CO2/kWh of wind power, whereas it is (1.0 – 0.419) = 0.581 lb of CO2/kWh, significantly less than claimed.

The owning and O&M costs of the 8,889 MW cycling facility would be significant. Would owners of wind facilities pay a fee as part of accommodating wind power?

What are such fees elsewhere? Nationwide wind power accommodation fees vary from about $2/MWh at low wind power penetrations to $9/MWh at high wind power penetrations. Currently, the Bonneville Power Authority charges about $5.7/MWh for cycling its hydro plants to accommodate wind power. Hydro-Quebec would likely charge a similar fee.

Because of limited US experience with greater wind power penetration, all costs of cycling to accommodate wind power may not yet be fully known, quantified and included. Therefore, studies, based on actual operating data, not statistical guesswork, need to be made to determine if these fees include all the owning and O&M costs of cycling to accommodate wind power.


Wind Power Facility With Cycling Facility:

Capital cost for 10,000 MW wind facility is about $25 billion.

Useful service life is about 25 years.

Annual production = 10,000 MW x 8,760 hr/yr x CF 0.31 = 27,156 GWh/yr*

Capital cost for 8,889 MW cycling facility is about $11.1 billion.

Capital cost for build-out T&D systems is about $3 billion.

Capital cost of wind power prediction facility is about $100 million.

Capital costs for increased frequency regulation capacity.

Extra annual fuel costs for operating CCGT cycling facility is about $347,881,469/yr.

Extra CO2 emissions from CCGT cycling facility is about 11.39 billion lb of CO2/yr.

Extra annual costs for ISO-NE operating costs to deal with wind power.

Annual owning and O&M costs of wind facilities.

Annual owning and O&M costs of CCGT cycling facility.

Annual owning and O&M costs of enlarged T&D systems.

Annual owning and O&M costs of wind prediction facility.

Annual owning and O&M costs of increased frequency regulation capacity.

Visuals and noise: significant negative which would take much costly effort and valuable time to resolve.

Net CO2 emissions reduction = 27,156 GWh/yr x (1.0 NEEG average – 0.419) lb C02/kWh = 15.78 billion lb of CO2/yr #

CCGT Power Facility Only

Capacity of CCGT facility = (27,156 GWh/yr x 1,000 MW/GW)/(8,760 hr/yr x CF 0.90) = 3,445 MW

Capital cost of CCGT facility = 3,445 MW x $1,250,000/MW = $4.31 billion.

Useful service life is about 35-40 years.

Annual production = 3,445 MW x 8,760 hr/yr x CF 0.90 = 27,156 GWh/yr*

Capital cost for build-out of T&D systems: minimal compared to wind.

Annual owning and O&M costs of CCGT facility: minimal compared to wind

Annual owning and O&M costs of built-out T&D systems: minimal compared to wind.

Visuals and noise: minimal compared to wind.

Net CO2 emissions reduction = 27,156 GWh/yr x (1.0 NEEG average – 1.298 x 0.726) lb of CO2/kWh = 1.57 billion lb of CO2/yr #

* the CF gradually decreases as the facility ages.

# this reduction reduces the NEEG average lb of CO2/kWh


Implementing CCGT power instead of wind power appears to be the best choice, by far. The enormous ADDITIONAL capital costs and annual owning and O&M costs for the wind power and cycling facilities could be much more effectively used for investments in increased energy efficiency which would reduce CO2 far more effectively per invested dollar than either CCGT or wind power.


Because the NEEG has very minor wind power penetration, there would be no data to study fuel consumption and CO2 emissions related to cycling plants to accommodate wind power, as there are in other jurisdictions. Accordingly, a recent study of Colorado and Texas, both states with significant wind facilities, would be used to illustrate some impacts of wind power on plant operations.

Power Plant Cycling In Colorado

Public Service of Colorado, PSCO, lacks sufficient gas-fired CCGT capacity for cycling to accommodate wind power. Instead, it is attempting to use coal plants for cycling for which they were not designed. The results have been significantly increased pollution and CO2 emissions.

Fuel consumption in Btu/kWh is called heat rate; for a coal plant operated near rated output it is about 10,500 Btu/kWh for power delivered to the grid. It is lowest near rated output and highest at very low outputs. If a plant is ramped up and down (cycled) at a percent of rated output, its heat rate rises. See Pages 26, 28, 35, 41 of the BENTEK study.

On Page 28, the top graph covering all PSCO coal plants shows small heat rate changes with wind power during 2006. The bottom graph shows greater heat rate changes with wind power during 2008, because during the 2006-2008 period 775 MW of wind capacity was added. For the individual PSCO plants doing most of the cycling, the heat rate changes are much higher.

On Page 26, during a coal plant ramp down of 30% from a steady operating state to accommodate state-mandated “must take” wind power, the heat rate rose at much as 38%.

On Page 35, during coal and gas plant ramp downs, the Area Control Error, ACE, shows significant instability when wind power increased from 200 to 800 MW in 3.5 hours and decreased to 200 MW during the next 1.5 hours. The design ramp rates, MW per minute, of some plants were exceeded.

On Page 41, during coal plant cycling across the PSCO system due to a wind power event, emissions, reported to the EPA for every hour, showed increased emissions of 70,141 pounds of SOX (23% of total PSCO coal emissions); 72,658 pounds of NOX (27%) and 1,297 tons of CO2 (2%) than if the wind power increase had not caused the plants to be cycled.

Those increases are due to combustion and air pollution control system instabilities which persist well beyond the wind event. PSCO does not release hourly wind generation data. Such baseline information is critical for any accurate analysis and comparison of alternatives to reduce such emissions; deliberately withholding such information is inexcusable.

Power Plant Cycling In Texas

The Texas grid in mostly independent from the rest of the US grids; the grid is operated by ERCOT. The grid has the following capacity mix: Gas 44,368 MW (58%), Coal 17,530 MW (23%), Wind 9,410 MW (12% – end 2009), Nuclear 5,091 MW (7%). Generation in 2009 was about 300 TWh. By fuel type: Coal 111.4 TWh, Gas CCGT 98.9 TWh, Gas OCGT 29.4 TWh, Nuclear 41.3 TWh, Wind 18.7 TWh. Summer peak of 63,400 MW is high due to air conditioning demand.

Wind provides 5% – 8% of the average generation overall, depending on the season. Its night contribution rises from 6% (summer) to 10% (spring). Texas capacity CF = 18.7 TWh/yr/{(9,410 + 7,118)/2) MW x 8,760 hr/yr)} = 0.258. Texas has excellent winds and should have a statewide CF of 0.30 or greater. Explanations for the low CF likely are:

– ERCOT requires significant curtailment of wind power to stabilize the grid.

– Vendors, developers and financiers of wind power, eager to cash in on subsidies before deadlines, installed some wind turbine facilities before adequate transmission capacity was installed to transmit their wind power to urban areas.

Much of the gas-fired capacity consists of CCGTs that are owned by IPPs which sell their power to utilities under PPAs. That capacity is not utility-owned and therefore not available for cycling to accommodate the more than 10,000 MW capacity of wind power. Instead, utilities are attempting to use coal plants for cycling for which they were not designed. The results have been significantly increased pollution and CO2 emissions.

Unlike PSCO, ERCOT requires reporting of fuel consumption by fuel type and power generation by technology type, including wind power, every 15 minutes. The 2007, 2008, 2009 data shows rising amplitude and frequency of cycling operations as wind penetration increased. In 2009, the same coal plants were cycled up to 300 MW/cycle about 1,307 times (up from 779 in 2007) and more than 1,000 MW/cycle about 284 times (up from 63 in 2007) from one 15-minute period to the next. The only change? Increased wind power penetration.

On Page 69: The ERCOT cycling of plants to accommodate wind power produced results similar to the PSCO system; increased cycling due to wind power resulted in significantly more SOX and NOX emissions than if wind power had been absent. CO2 emission reductions due to wind power are minimal at best.

Remedy for Colorado and Texas Cycling Problems

A way out is for PSCO and ERCOT is to retire older coal plants that have efficiencies of about 30% and emit about 2.15 lb of CO2/kWh and replace them with utility-owned, gas-fired CCGTs that have efficiencies of up to 60% and emit about 0.67 lb of CO2/kWh. The CCGTs have short installation periods and capital costs of about 1,250/kW. If wind were entirely absent, this measure would reduce the most CO2/kWh at the least $/kWh and would produce power at the least $/kWh. If some of the new units were cycled to accommodate wind power, their Btu/kWh would increase and they would produce more NOX/kWh and CO2/kWh, mostly offsetting the lb of CO2/kWh reduction due to wind power, as shown above.