The Impact of Electricity Market Conditions on the

Proceedings of the ASME 2012 International Mechanical Engineering Congress & Exposition IMECE2012 November 9-15, 2012, Houston, Texas, USA

IMECE2012-88119

THE IMPACT OF ELECTRICITY MARKET CONDITIONS ON THE VALUE OF FLEXIBLE CO2 CAPTURE

Stuart M. Cohen∗ Michael E. Webber Gary T. Rochelle Department of Mechanical Engineering Department of Mechanical Engineering Department of Chemical Engineering The University of Texas at Austin The University of Texas at Austin The University of Texas at Austin Email: [email protected] Email: [email protected] Email: [email protected]

ABSTRACT Carbon dioxide (CO2 ) capture with amine scrubbing at coal-fired power plants can remove 90% of the CO2 from flue gas, but operational energy requirements reduce net electrical output by 20–30%. Temporarily reducing the load on energy intensive components of the amine scrubbing process could temporarily increase power output and allow additional electricity sales when prices are high. Doing so could entail additional CO2 emissions, or amine solvent storage can be utilized to allow increased power output without additional CO2 emissions. Priceresponsive flexible capture is studied for $0–200/tCO2 and $2– 11/MMBTU natural gas using a nominal 500 MW coal-fired facility in the 2010 Electric Reliability Council of Texas (ERCOT) grid. CO2 capture systems use a 7 molal monoethanolamine (MEA) solvent. Venting additional CO2 while increasing electrical output provides significant benefit only at $30–60/tCO2 and when natural gas prices exceed $4/MMBTU. Solvent storage can improve profitability with CO2 capture at higher CO2 emissions penalties, but primarily at low-to-moderate natural gas prices when power plant capacity factor is less than 90%.

fossil-fuel burning. Different technology and policy solutions have a range of technical, environmental, and economic implications, but CO2 emissions reductions on the order recommended by the International Panel on Climate Change (IPCC) will require an all-of-the-above approach [1]. Carbon dioxide capture and sequestration (CCS) is one carbon mitigation option that allows continued use of fossil fuels for electricity and industrial processes with greatly reduced CO2 emissions. CCS is not sustainable over millennia, but it is a critical technology while societies transition to sustainable energy systems. A primary application for CCS is coal-based power generation, though natural gas-fired power plants can also utilize the technology. Coal is typically the focus of CCS discussions because the fuel is widely used for electricity production worldwide; in the United States, coal accounts for approximately 50% of electricity and 30% of total CO2 emissions [2]. The primary disadvantages of CCS are the capital and energy costs of CO2 capture. CO2 capture systems are typically designed to achieve about 90% CO2 removal from power plant flue gas, but energy requirements for CO2 capture and compression typically reduce net power output by 20–30% [3]. Longdistance CO2 pipeline transport and underground CO2 injection is well-understood from enhanced oil recovery (EOR) operations ongoing in West Texas since the 1970s [4, 5]. Commercial CO2 removal has also been performed for several decades in the natural gas purification and ammonia production industries, but these applications are smaller scale and more economically viable than CO2 removal from power plant flue gas [5]. Nevertheless, demo-

INTRODUCTION TO CCS Growing concerns with global climate change induced by anthropogenic carbon dioxide (CO2 ) emissions have motivated the scientific and policy-making communities to explore several short- and long-term options for reducing CO2 emissions from

∗ Address

all correspondence to this author.

1 Downloaded From: http://proceedings.asmedigitalcollection.asme.org/ on 10/13/2014 Terms of Use: http://asme.org/terms

c 2012 by ASME Copyright

Post-Combustion CO2 Capture Using Amine Absorption and Stripping

Boiler

Post-Combustion Amine Scrubbing for CO2 Capture

LP Turb.

Electric Motor Steam for CO2 Capture

Flue Gas

Flue Gas With 90% CO2 Removal Other Pollution Controls

Coal

Amine scrubbing, a chemical absorption and stripping process, is one leading technology for CO2 capture in both retrofit and new build applications. A well-designed amine scrubbing system reduces net power output by ∼20%. Though substantial, this energy penalty is only about double the minimum theoretical energy requirement for CO2 removal and compression to 150 bar, a typical pipeline pressure [3]. An actual-to-minimum energy requirement ratio of 2 is much better than other separation processes in commerical operation (distillation, cryogenic air separation, reverse osmosis desalination), which have actualto-minimum energy ratios on the order of 3–5 [3]. Consistent comparisons of CO2 capture energy requirement across technologies must be sure to include CO2 compression energy to the same outlet pressure.

Electricity Generator

CO2 Compressor

CO2

Heat Ex.

Rich Solvent

CO2 for Transport & Storage

Sttripper

Steam

IP Turb.

Absorber

Steam Turbines HP for Turb. Power

scale power plant CO2 capture on a tens of megawatts (MW) scale exist today, and CCS technology is ready for commercialscale installation (hundreds of MW) given sufficient economic and policy incentives [6].

100% Capture Steam Flow

Condensate Lean Solvent Stuart Cohen & Carey King CO2 Capture and Sequestration 8 November 6, 2008

FIGURE 1. POST-COMBUSTION AMINE SCRUBBING AT COAL-FIRED POWER PLANTS CAN ACHIEVE 90% CO2 REMOVAL BUT REDUCES NET ELECTRICAL OUTPUT BY 20–30%.

Flexible CO2 Capture Motivation One other advantage of amine scrubbing is its suitability for flexible operation independent of power generation systems, where the energy-intensive stripping and compression equipment could temporarily operate at partial or zero load to increase net power output. Typically, CO2 capture is assumed to operate continuously at maximum load, so the energy required for CO2 capture is lost permanently. However, given sufficient LP turbine and generator capacity, net power output could be increased by redirecting stripping steam back to the LP turbine for power production, and less desorbed CO2 entails lower CO2 compression work. Temporarily increasing power output using flexible capture could help meet peak electricity demand and eliminate the need to replace generating capacity lost to CO2 capture energy requirements [10, 11]. Assuming components are kept at normal operating temperatures, increased power output from flexible capture could be achieved within minutes, allowing flexible capture systems to provide reliability services in response to generation or transmission outages [12]. By operating flexible capture systems in response to electricity supplied from intermittent sources such as wind, flexible CO2 capture could also complement renewable electricity. The main advantage of flexibility examined in this article is the ability to increase power output when electricity prices are high [13–15]. If doing so offsets any incremental costs of flexibility, overall CO2 capture economics could be improved. Flexible post-combustion amine scrubbing is particularly suitable for CO2 capture retrofits, where the LP turbine and generator are already sized for electrical output at pre-capture levels. Though flexible capture could involve temporary increases in CO2 emissions, future emissions restrictions can still be satisfied by operating CO2 absorption systems frequent enough to keep total CO2 or the average CO2 emissions rate within policy

Figure 1 shows a simplified diagram of an amine scrubbing process integrated into a coal-fired power plant. Coal is burned in the normal fashion to produce steam that expands through high, intermediate, and low-pressure turbines (HP, IP, LP) that drive a generator. Flue gas passes through other pollution control equipment to remove particulates, sulfur dioxide (SO2 ), and nitrogen oxides (NOx ) before entering the CO2 absorption column. Amine solvent flows countercurrently with flue gas in the absorber, where it typically removes 90% CO2 at 40–70◦ C and exits as rich solvent (high in CO2 ). Solvent then passes through a heat exchanger before entering the stripping column, where solvent is heated to 120–150◦ C to reverse the solvent-CO2 reaction and desorb the CO2 . After condensing any steam exiting the stripper, the nearly-pure CO2 stream is compressed to a pressure suitable for pipeline transport (typically 90–150 bar) and underground CO2 injection [7]. Solvent with relatively less CO2 , called lean solvent, is then recycled back to the absorber. Stripping heat is provided is provided by saturated steam slightly above the required stripper temperature. Typically, 40– 50% of the steam from the IP/LP crossover pipe is extracted, with exact flows depending on steam cycle design [8, 9]. With a high enough IP/LP crossover pressure, the extracted steam could be expanded through a let-down turbine to drive the CO2 compressor before heating the solvent, but Fig. 1 shows the compressor driven by an electric motor. Steam extraction and CO2 compression work account for the bulk of CO2 capture energy requirements, with the remainder coming from rich & lean solvent pumps and the flue gas blower.



One Flexible Capture Configuration Bypasses pp and Vents Additional CO2 the Stripper

Solvent Storage Allows Constant 90% Capture

Rich Solvent (high CO2)

Heat Ex.

Strip pper

Bypass Stream 0–100% Flow

CO2 for Transport & Storage

Steam Control Valve 100–0% Capture Steam Flow

Flue Gas In

Condensate Lean Solvent (low CO2)

Lean Solvent Storage

Rich Solvent (high CO2)

Larger? CO2 Compressor 100–0% CO2

Larger? Heat Ex.

Rich Solvent Storage

Largerr? Stripp per

Flue Gas In

100–0% CO2 Abso orber

Flue Gas With 90–0% CO2 Removal

Flue Gas With 90% CO2 Removal

Abs sorber

CO2 Compressor

CO2 for f Transport & Storage

Steam C t l Valve Control V l 100–0% Steam Flow

Lean Solvent (low CO2)

Stuart Cohen & Carey King CO2 Capture and Sequestration 18 November 6, 2008

FIGURE 2. A VENTING-ONLY FLEXIBLE CONFIGURATION ALLOWS INCREASED POWER OUTPUT BUT REQUIRES ADDITIONAL CO2 EMISSIONS.

Stuart Cohen & Carey King CO2 Capture and Sequestration 20 November 6, 2008

FIGURE 3. SOLVENT STORAGE ALLOWS INCREASED POWER OUTPUT WITHOUT INCREASING CO2 EMISSIONS BUT REQUIRES ADDITIONAL CAPITAL EXPENSE.

requirements.

larger stripping and compression equipment. A facility with solvent storage would likely maintain the ability to vent CO2 when economically desirable or necessary for maintenance.

Flexible CO2 Capture Configurations There are two basic concepts for flexible CO2 capture using amine scrubbing. One reduces the energy requirements of solvent stripping and CO2 compression while allowing the CO2 removal rate to fall. Figure 2 displays one way to implement this concept, where steam and rich solvent flow rates to the stripper are reduced equally and simultaneously during partial- or zeroload operation [16]. At partial load, rich solvent diverted from the stripper is recycled to the absorber, so CO2 removal rates in the absorber will decrease as solvent becomes saturated with CO2 . Zero load could involve recirculating all solvent through the absorber, or the CO2 capture system could be bypassed completely. Assuming a retrofit application where the low pressure turbine and generator have been sized to operate without CO2 capture, this design has negligible capital cost, but its primary disadvantage is the environmental impact and any cost of increased CO2 emissions. Another flexible CO2 capture concept uses auxiliary solvent storage tanks to maintain high CO2 removal when stripping and compression systems operate at partial or zero load (Fig. 3) [17]. When electricity prices are high, the plant can reduce stripper and compressor load while maintaining full-load CO2 absorption by feeding the absorber from a lean solvent storage tank and depositing rich solvent into another tank. When electricity prices are low, stripping and compression systems return to a higher load to treat the current process stream and the stored rich solvent. To treat both streams, either absorber load must be reduced, or stripping and compression systems must be over-sized. Absorber load can be reduced without additional CO2 emissions if base plant load is reduced at the same time. Maintaining high CO2 removal keeps operating costs down while storing rich solvent, but any operating profit improvement must be weighed against the capital cost of solvent inventory, storage tanks, and

Scope and Contribution This article seeks to deduce the market conditions where flexible CO2 capture is valuable for increasing electrical output when electricity prices are high. Earlier work by these and other authors has explored this effect for a venting-only flexible capture system over a wide range of fuel and CO2 price1 conditions; these studies have shown that venting CO2 can be valuable at intermediate CO2 prices of $20–70/tCO2 , with the incremental benefit and CO2 price range depending on fuel, CO2 , and electricity price characteristics [13, 15, 18]. Cohen et al. 2010 and 2012 presented a model that represents flexible capture with solvent storage, but these articles reported only CO2 price sensitivity analysis [15, 19]. Patiño-Echeverri and Hoppock 2012 and Husebye et al. 2010 have studied the effect of electricity price volatility on the benefits of solvent storage, but these analyses did not explicitly study the effects of underlying fuel and CO2 prices [14, 20]. This article will systematically establish the CO2 and natural gas price regimes where flexible capture is valuable and discuss economic and environmental system performance across the studied market conditions. Coal prices are not varied here because U.S. coal prices are stable relative to natural gas prices, and the effect of any future CO2 mitigation policy (tax, cap-andtrade, or regulation) is highly uncertain. Results reported here are intended to assist future investment and operational decisions on whether or not to install flexible or inflexible CO2 capture systems given expected future market conditions. The Electric Reliability Council of Texas (ERCOT) electricity system in 2010 is used as a case study. In 2010, ERCOT

1 In

the context of this article, CO2 price refers to an emissions penalty.



contained 550 generating units with 84.4 gigawatts (GW) of total capacity that produced 319 terawatt-hours (TWh) of electricity [21]. More than half the total generating capacity (57%) uses natural gas, with 23% coal, 6% nuclear, and 12% wind. Coal and nuclear fuel is typically used for base load, so their share of generation is 40% and 13% respecively, with wind at 8% and natural gas at 38% [21]. The remainder of capacity and generation consist of hydroelectric, biomass, and other unit types. The size and variety of facilities in ERCOT make it an ideal case study. While absolute results reported here are specific to ERCOT and those conditions studied, qualitative conclusions can be drawn that apply to other electricity systems. The analysis utilizes a previously developed optimization model that maximizes operating profits for a single power plant in response to a time series of input electricity prices [15]. This model allows comparison of the facility with venting-only flexible CO2 capture (Fig. 2), flexible capture with solvent storage (Fig. 3), inflexible capture, and no capture. Within the optimization model, the facility is a “price-taker,” meaning that its operation does not affect electricity prices. However, input electricity prices are adjusted for different CO2 and natural gas prices using a first-order dispatch model of the ERCOT grid. The details of this procedure are discussed in the Methodology section.

FIGURE 4. FORECAST ERROR IS APPROXIMATED BY OPTIMIZING PLANT OPERATION IN RESPONSE TO A PSEUDOFORECAST PRICE SERIES CREATED BY REMOVING OUTLIERS AND SMOOTING VOLATILE PRICE SERIES.

quantity of solvent in rich storage is used to define that state of the solvent storage system. The objective function includes base plant startup costs, fuel and CO2 prices, and base plant variable operation and maintenance (VOM) costs. When modeling a CO2 capture configuration, additional costs are included for solvent makeup to replace degraded solvent, caustic for use in degraded solvent reclaiming, degraded solvent waste disposal, additional water use for CO2 capture, and CO2 transport and storage. In addition, a cost is assessed for ramping the CO2 capture system. This cost acts as a proxy for inefficiencies during transient CO2 capture operation; these inefficiencies are not modeled directly to maintain model linearity. The model includes constraints on the minimum and maximum load of power, absorption, and stripping/compression systems as well as limitations on the ramp rate, or system response time, of each system. Ramp rates for both power and capture systems affect the net ramp rate of a flexible capture facility because power systems could ramp up while capture systems ramp down and vice versa. Minimum load constraints enforce restrictions on gas and liquid flow rates in power and CO2 capture systems. To maintain model linearity, all performance parameters are constant average values across the relevant operating range. Though a nonlinear or piece-wise linear representation would improve solution accuracy, the simple linear formulation is assumed sufficient to draw general conclusions based on aggregate results. Different CO2 capture configurations are defined by additional constraints relating power and CO2 capture systems. An inflexible capture system is defined as one that must treat all flue gas that is produced, so CO2 capture load must equal the fractional load on power systems. A facility with venting-only flexible capture (no solvent storage) can operate with capture system load below that of the base plant, but absorption, stripping, and compression systems must have equal load. With solvent storage, absorption and stripping/compression load are decou-

METHODOLOGY Optimization Model Description The optimization model is a mixed-integer linear program (MILP) created using the General Algebraic Modeling System (GAMS) modeling language. The full model formulation is described in Cohen et al. 2012; this section contains a brief summary of model characteristics [15]. Input electricity prices are imported in 15 minute time intervals for one year, with 15 minutes being the frequency of ERCOT price settlements for payment purposes. Actual historical prices are used as the basis for input electricity prices, but price series with historical volatility are not used directly when optimizing plant operation because doing so would imply that all price variations, including extreme price spikes, can be forecast. Instead, outliers are removed from volatile prices and the resulting price series is smoothed using a Savitzky-Golay polynomial filter to create a “pseudo-forecasted” price series that differs from the original price series by an amount typical of day-ahead forecasting models in the literature. Figure 4 provides a graphical example. The pseudo-forecast price series is used to optimize power and CO2 capture system operation for the year, and the original volatile price series is used to calculate profits given optimization results. This procedure approximates the effect of forecast error. The model seeks the optimal base plant, absorption, and stripping/compression load in each interval while also deciding which systems are on or off. Stripping and compression load are assumed equal. If a solvent storage system is available, the



pled and must be constrained independently. Maximum absorption load is limited by available flue gas, while maximum stripping/compression load is limited by steam availability (minimum LP turbine flow) and stripping/compression equipment size. A solvent storage system also requires specifying a maximum capacity for stored solvent and a flow balance constraint to monitor the level in solvent tanks. The solvent storage system is defined analogously to an energy storage system such as a pumped hydroelectric station, where lean solvent is a proxy for stored energy. To be consistent with presumed day-ahead forecasting ability, the quantity of solvent stored in each of the rich and lean tanks must also return to a specified set point at the end of each day.

price conditions using historical electricity demand and power plant characteristics. 2. Calculate the change in first-order electricity prices from historical to adjusted conditions. 3. Add the calculated change in first-order prices to historical electricity prices to produce an adjusted electricity price series that preserves historical volatility. First-Order Electricity Price Calculation The MATLAB program imports a time series of net electricity load (demand minus wind production), generating unit characteristics, and a daily time series of historical and adjusted fuel and CO2 prices. Electricity demand data are retrieved from the ERCOT website in hourly intervals. To synchronize with electricity prices in 15-minute intervals, demand data are up-sampled using a linear interpolation between each hourly demand data point. The first-order dispatch procedure does not account for detailed generating unit constraints such as minimum load, ramp rates, and minimum up/down time, so only the following characteristics are imported: unit name, unit type, maximum electrical output, heat rate, CO2 emissions rate, base plant VOM costs, and whether or not the facility is designated as a combined heat and power (CHP) facility. After importing requisite data, the program iterates through each 15-minute time interval and calculates first-order electricity prices in each interval under historical and adjusted fuel and CO2 prices. It does this by calculating marginal generating costs for each generating unit, placing units in cost order, choosing the least-expensive available plants to meet current electricity demand, and setting the electricity price equal to the most expensive facility required to meet current demand. Costs at fossilfueled and biomass-based facilities are the sum of fuel costs, base plant VOM costs, and CO2 emissions costs, if applicable. Nuclear and hydroelectric facilities are assigned VOM costs, and wind facilities are not represented in the plant database because wind production is subtracted from the imported demand data. Since CHP facilities typically follow relatively constant heat and process loads, CHP facility operation is approximated by forcing them to operate at their maximum capacity at all times.

Adjusting Electricity Prices for Market Conditions In previous use of this model, CO2 emissions penalty was the only independently varied electricity market condition, and the impact of CO2 price on ERCOT electricity prices was approximated by uniformly increasing all electricity prices by the average CO2 emissions penalty of ERCOT gas-fired facilities [15]. For example, the generation-weighted average CO2 emissions rate for ERCOT gas-fired facilities in 2007 is 0.43 metric tons of CO2 per megawatt-hour (tCO2 /MWh), so a CO2 price of $50/tCO2 would raise electricity prices in each time interval by $21.5/MWh [22]. This approximation is reasonable for ERCOT under moderate CO2 prices and moderate to high natural gas prices because ERCOT is a gas-dominated market where gas-based capacity will typically be the marginal, or most expensive, generation. However, this approximation does not account for variability in the emissions costs of natural gas-based generators, and its accuracy suffers under market conditions where gas begins to replace coal as base load fuel, and coal-fired facilities are more frequently marginal generator. With natural gas prices below $3 per million British thermal unit (MMBTU) at the time this article was composed, gas is already poised to displace many coal-fired plants as base load facilities [23]. A more accurate method to approximate the effect of fuel and CO2 prices on electricity prices has been developed to enable model use under a wider range of electricity market conditions or in other electricity systems. The price-adjustment procedure accounts for changes in relative dispatch order of generating facilities when estimating the impact of changing market conditions on electricity prices. However, it also preserves historical price volatility so the optimization model and profit calculations use a realistic price series. To account for the effect of fuel and CO2 price changes on dispatch order while preserving historical electricity price volatility, the following procedure has been implemented using the MATLAB environment.

Adjusted Prices with Volatility After determining first-order electricity price series under historical and adjusted fuel and CO2 price conditions, the electricity price difference is taken at each time interval. This difference is then added to historical electricity prices in each interval. Historical electricity price data are retrieved from ERCOT then averaged across load zones. The resulting adjusted electricity price series preserve historical price volatility while approximating the effect of changes to fuel and CO2 price. Adjusted volatile electricity prices can then be used to produce pseudo-forecast price series for each fuel and CO2 price condition using the outlier removal and smoothing

1. Use a first-order electricity dispatch model to calculate electricity prices under historical and adjusted fuel and CO2



TABLE 1. THE FOLLOWING HISTORICAL AND ADJUSTED FUEL AND CO2 PRICES ARE USED IN SAMPLE PRICEADJUSTMENT CALCULATIONS.

Price set

Natural Coal price Oil price CO2 price gas price ($/MMBTU) ($/MMBTU) ($/tCO2 ) ($/MMBTU)

Hist. (2010)

2.25

5.14

12.3

0

Adj. 1

2.25

5.14

12.3

50

Adj. 2

2.25

3.00

12.3

25

Adj. 3

2.25

5.14

12.3

100

FIGURE 5. A FIRST-ORDER ELECTRICITY DISPATCH PROCEDURE CALCULATES ELECTRICITY PRICES UNDER EACH SET OF INPUT MARKET CONDITIONS.

procedure discussed above and in Cohen et al. 2012 [15]. Sample This section demonstrates the electricity price adjustment procedure for a sample set of data from December 1, 2009 – November 30, 2010. A full year of 2010 data are not available because ERCOT changed its market structure on December 1, 2010. Historical electricity demand and prices are retrieved from ERCOT [24, 25]. Generating unit data are taken from the a generating unit-specific database created using the U.S. Environmental Protection Agency (EPA) eGRID database and information provided by ERCOT [22, 26]. While the procedure can accommodate daily-varying commodity prices, constant values for fuel and CO2 prices are used in this analysis. Table 1 lists historical 2010 average fuel prices CO2 prices along with three sample adjusted price cases [2]. Figure 5 plots electricity demand along with first-order electricity prices under historical and adjusted fuel and CO2 price conditions for July 15 and 16, 2010, when demand was relatively high. For the same dates, Fig. 6 plots electricity demand with historical electricity prices and adjusted volatile electricity prices for each market condition. Below ∼40 GW of net load, gas-fired capacity is usually marginal, so the typical electricity price shift reflects the change in fuel and CO2 prices at gas-based facilities. However, coal-fired capacity is marginal at higher demand, so the electricity price shift more closely reflects the emissions cost of coal-fired generation. Fig. 6 also illustrates the utility of preserving historical price volatility, as the electricity price spikes and periods of high prices on July 15 would not be reproduced using solely a first-order approach. These high prices could be very valuable for flexible CO2 capture, so including them in the model is important.

FIGURE 6. DIFFERENCES IN FIRST-ORDER PRICE SERIES ARE ADDED TO HISTORICAL PRICES TO CREATE ADJUSTED VOLATILE PRICE SERIES.

cause it is representative of a large coal-based unit, and its heat rate and CO2 emissions rate are the generation-weighted averages across ERCOT coal-fired facilities [27]. Though more recent data are available, power system characteristics are based on the 2007 eGRID database to maintain consistency with previous work. Minimum power output, ramp rate, and startup costs are estimated from literature [28, 29]. The CO2 capture system is assumed to use 7 molal (30% weight) monoethanolamine (MEA) solvent for CO2 capture. The ramp rate of CO2 capture systems is assumed slightly greater than that of power systems, and the minimum load is the same on a percent basis. A typical value of 90% CO2 removal is assumed, and the energy requirement and design capacity are taken from an AspenPlus CO2 capture system model developed at The University of Texas at Austin [30]. There is a relationship between capture energy requirement and design capacity that produces a minimum energy requirement of 0.247 MWh/tCO2 for this system at 0.12 molCO2 /molMEA. This operating point is used for

Power & CO2 Capture System Parameters Key power and CO2 capture system specifications are listed in Tab. 2. A maximum power output of 500 MW is used be-



TABLE 2. THE FOLLOWING ARE KEY POWER AND CO2 CAPTURE SYSTEM INPUT PARAMETERS FOR THE COAL-BASED FACILITY.

inflexible and venting-only flexible capture configurations. However, because a higher design capacity with solvent storage will reduce the required quantity and cost of stored solvent inventory, a slightly higher design capacity (0.16 molCO2 /molMEA) is used that increases energy requirements to 0.249 MWh/tCO2 . For all CO2 capture configurations, 1/9 of the enegy requirement is attributed to absorption systems (flue gas fan, lean amine pump), and the remainder is attributed to stripping and compression systems (including the rich amine pump) [13].

Parameter (units)

Value

Base power plant: coal-fired

The solvent storage design assumed for this analysis is based on a separate design sensitivity study that analyzed the investment value of solvent storage across a range of solvent storage capacities, equipment oversizing fractions, and design operating points. These value estimates included solvent storage capital costs. Details of that analysis are outside the scope of this article but will be submitted for future publication and included in the Cohen 2012 dissertation [31]. Oversizing stripping and compression equipment incurs substantial capital costs without necessarily adding value, so no equipment oversizing is assumed in this analysis [13]. The chosen solvent storage capacity is large enough for a maximum of 2 hours with stripping and compression systems off and absorption at full load. This storage system size is larger than that used in previous work because the greater design capacity and lack of equipment oversizing is expected to increase the size of an economically viable solvent storage system. To enforce day-ahead planning of the solvent storage system, the rich solvent tank must return to 71% of its maximum level at the end of each day. This set point was established in previous work as a typical value without a daily set point restriction [15].

Maximum power output (MW)

500

Minimum power output (MW)

150

Heat rate (without (MMBTU/MWh)

CO2

capture)

10.8

CO2 emissions rate (without CO2 capture) (tCO2 /MWh)

1.03

Ramp rate (%/min)

4

Startup cost ($/startup)

10,000

CO2 capture system: 7m MEA Minimum load (%)

30

Ramp rate (%/min)

5

CO2 removal (fractional)

0.9

Energy requirement (equivalent work) (MWh/tCO2 )

0.247 (SS: 0.249)

Design capacity (rich minus lean loading) (molCO2 /molMEA)

0.12 (SS: 0.16)

Solvent storage capacity (max hours with full absorption and no stripping)

2

TABLE 3. SYSTEM PERFORMANCE IS STUDIED FOR THE FOLLOWING FUEL AND CO2 PRICES.

Electricity Market Parameters The reference year for all analysis is 2010, so the electricity price adjustment procedure uses historical 2010 electricity demand and commodity prices and the 2010 power plant fleet. Table 3 provides the range of fuel and CO2 prices considered in this analysis. All prices are kept constant throughout each oneyear simulation. Historically, natural gas prices have been quite volatile throughout the year, and the European Union Emissions Trading System (EU-ETS) suggests CO2 prices could also be volatile in a cap-and-trade setting [32, 33]. While the modeling framework is capable of handling daily-variable fuel and CO2 prices, constant annual averages are used to establish a baseline of market conditions where capture flexibility is valuable.

Commodity (units)

Price

Coal ($/MMBTU)

2.25

Natural gas ($/MMBTU)

2–11

CO2 ($/tCO2 )

0–200

RESULTS Power System Utilization The annual base plant capacity factor provides a general indication of how power and CO2 capture systems are utilized over the range of electricity market conditions. Figure 7 plots the plant capacity factor without capture and with inflexible capture over the full range of CO2 prices and three natural gas

Since coal price is historically stable, it is kept constant at the 2010 U.S. average of $2.25/MMBTU in all analysis [2]. The ranges of natural gas and CO2 prices are chosen to bracket the conditions that might be expected over the next 0–20 years.



prices: $2/MMBTU (minimum), $11/MMBTU (maximum), and $5.14/MMBTU (actual 2010 average) [2]. Since coal price is constant, higher natural gas prices will increase electricity prices relative to coal-based operating costs, providing more opportunity for profitable operation. However, a facility without CO2 capture never exceeds a 52% capacity factor at any natural gas price. Higher capacity factors might be expected with high gas prices and low CO2 prices, but the apparent ceiling at 52% exists due to the nature of the electricity price adjustment procedure and the relatively high assumed coal price and heat rate. Regardless of natural gas price, coal-based capacity is always displaced by the large quantity of continuously operating gas-based CHP units. Thus, first-order dispatch finds coal-fired facilities to be marginal for much of the year at any gas price, and electricity prices will not change with gas price at these times. Once all non-CHP gas-fired units are displaced, capacity factors of coalfired units do not change with natural gas price, though profits will continue to increase as electricity prices rise when gas-fired units are marginal. In practice, power plant ramping limitations and other system constraints would increase the frequency that gas-fired facilities are marginal, but the first-order procedure produces a conservative estimate of electricity price adjustments that is sufficient for analyzing aggregate trends and behaviors. Changes in capacity factor with CO2 price reflect the facility CO2 emissions rate relative to that of marginal generators. When capture is unavailable, capacity factor falls with CO2 price as emissions costs at the facility increase faster than those of marginal facilities, which are often gas-fired facilities but switch to more efficient coal-fired facilities at high CO2 and low natural gas prices. The emissions rate with inflexible CO2 capture is about 0.13 tCO2 /MWh, much less than the typical values of ∼0.5 tCO2 /MWh for natural gas and ∼1 tCO2 /MWh for traditional coal, so capacity factor increases with CO2 price. Capacity factors are lower with inflexible capture than without capture at CO2 prices below about $30/tCO2 because a plant without CO2 capture is more profitable under these conditions. Figure 8 shows an analogous plot for the venting-only and solvent storage flexible capture configurations. Flexible capture facilities forgo using capture systems at low CO2 prices when capture is uneconomical and utilize capture when CO2 prices are high, so capacity factors with flexible capture roughly follow the upper envelope of the curves with inflexible or no capture. Thus, for all natural gas prices except $11/MMBTU, capacity factor first drops with CO2 price before recovering at higher CO2 prices where CO2 capture operation is economical. Capacity factors are slightly greater with solvent storage at some intermediate CO2 prices when solvent storage expands the profitable range of electricity prices. Capacity factors are often higher for the ventingonly configuration at high CO2 prices because CO2 capture usually operates at full load at these conditions, and the larger capture energy requirement with solvent storage raises marginal gen-

Inflexible

No Capture

Capacity Factor (fractional)

1

11NG

0.8 5.14NG

0.6 0.4

2NG

0.2 0 0

40

80 120 CO2 Price ($/tCO2)

160

200

FIGURE 7. CAPACITY FACTORS REFLECT EMISSIONS COSTS RELATIVE TO EMISSIONS COSTS OF MARGINAL GENERATORS (NG PRICES ARE $/MMBTU).

Venting-Only

Capacity Factor (fractional)

1

Solvent Storage

11NG

0.8

6.5NG 8NG

0.6

5.14NG 3.5NG

0.4

2NG

0.2 0 0

40


160

200

FIGURE 8. FLEXIBLE CO2 CAPTURE ALLOWS GREATER CAPACITY FACTORS OVER THE FULL CO2 PRICE RANGE (NG PRICES ARE $/MMBTU).

erating costs and reduces plant utilization. CO2 Emissions and CO2 Capture Utilization Capacity factor trends are reflected in CO2 emissions results, which are expressed as the fraction of maximum possible annual emissions in Fig. 9 for the inflexible and no capture configurations and Fig. 10 for the solvent storage configuration. The fraction of maximum possible emissions is the ratio of annual CO2 emitted to the quantity of CO2 emitted by the 500 MW facility without capture if it operated 100% of the year. Thus, emissions fractions without capture are nearly identical to capacity factors, with any discrepancy resulting from part-load operation while ramping between offline, minimum output, and maximum out-



No Capture

Inflexible

0.6

0.6

0.5

11NG

Fraction n of Max Possible CO O2 Emissions

Fraction n of Max Possible CO O2 Emissions

11NG

0.5

5.14NG

0.4

5.14NG

0.4

0.3

0.3

2NG 2NG

02 0.2

02 0.2

0.1

0.1

0

2NG

0 0

40


160

200

0

FIGURE 9. CO2 EMISSIONS FALL WITHOUT CAPTURE DUE TO LOWER CAPACITY FACTOR, WHILE INFLEXIBLE CAPTURE ENSURES LOW EMISSIONS AT ALL MARKET CONDITIONS (NG PRICES ARE $/MMBTU).

40


160

200

FIGURE 10. FLEXIBLE CAPTURE ALLOWS SUBSTANTIAL REDUCTIONS IN CO2 EMISSIONS ABOVE $30/tCO2 . (SOLVENT STORAGE SHOWN. EMISSIONS WITH VENTING-ONLY CONFIGURATION ARE SLIGHTLY GREATER UNDER MANY MARKET CONDITIONS. NG PRICES ARE $/MMBTU.)

put. With inflexible capture, the fraction of maximum possible emissions is very low at low CO2 prices because facilities have low capacity factors, and the approach to 0.1 at high CO2 prices reflects the 90% CO2 removal specification. While CO2 emissions with flexible capture and solvent storage are the same as those without capture at CO2 prices below $20/tCO2 , they transition quickly to levels achieved by inflexible capture by $40–50/tCO2 . Higher gas prices slow this transition somewhat by increasing the frequency of high electricity prices where venting CO2 is profitable. Regardless, given CO2 prices above $30/tCO2 , flexible capture achieves significant CO2 emissions reductions relative to a facility without capture. Emissions with venting-only flexible capture are nearly the same as those with solvent storage, with there being a slight increase under many market conditions because solvent storage often allows slightly greater CO2 capture utilization. The CO2 price regime where flexible capture systems transition from primarily being offline to online is highlighted in Fig. 11, which plots the average CO2 emissions rate for the flexible capture configurations from $20–50/tCO2 at the minimum, maximum, and actual 2010 natural gas prices. Though the transition region varies little, there are subtle differences. For all natural gas prices, emissions rate is slightly lower with solvent storage. Though any operation above the minimum emissions rate signifies CO2 venting, CO2 capture utilization is greater with solvent storage in the transition regime because the facility does sometimes operate flexibly without additional CO2 emissions. In addition, the transition occurs more quickly at intermediate natural gas prices such as $5.14/MMBTU than at low and high natural gas prices. This phenomenon is explained using Figs. 12 and 13. Generally, the facility is online when electric-

ity prices exceed marginal electricity production costs, both of which vary with electricity market conditions. Figure 12 plots facility marginal costs at $0–50/tCO2 when CO2 capture systems are off (0% load) and on at 100% load. At any CO2 price where costs are lower with 100% capture, there is a breakeven electricity price where additional electricity sales with capture offset the additional CO2 emissions costs. Figure 13 plots price-duration curves at $35/tCO2 for the maximum, minimum, and actual 2010 natural gas prices. Price-duration curves are generated by placing electricity prices throughout the year in order from highest to lowest, and the horizontal axis indicates the fraction of electricity prices below a given price level. Also shown is the breakeven electricity price for venting CO2 at $35/tCO2 . The solid vertical lines mark capacity factors for the venting-only flexible capture facility at each market condition, and the dashed vertical lines mark the fraction of prices when venting is economical at each market condition. At low natural gas prices, electricity prices are low, so the facility operates very infrequently, but venting is often profitable when the facility is online. At high natural gas prices, capacity factor is much higher, but higher prices provide more opportunities to economically vent CO2 . These two behaviors produce higher average CO2 emissions rates at low and high natural gas prices relative to those at intermediate natural gas prices, when venting is profitable during a relatively small portion of online hours. Operating Economics To begin discussing the economic implications of flexible CO2 capture, Fig. 14 shows annual operating profits normalized by gross power capacity with $5.14/MMBTU natural gas and



200

Solvent Storage

Pse eudo-Forecast Electricity Price ($/MWh)

Average C CO2 Emissions Rate (ttCO2/MWh)

Venting-Only

1

0.8 11NG

0.6

2NG

0.4 0 5.14NG

0.2

150 Capacity factor Fraction at breakeven

100 Breakeven price for venting CO2 11NG

50

5.14NG 2NG

0

0 20


0.0

50

0.2

0.4 0.6 Fraction of Prices

0.8

1.0

FIGURE 13. AVERAGE EMISSIONS RATE AT TRANSITION CO2 PRICES DEPENDS ON THE FRACTION OF ONLINE TIME WHEN VENTING CO2 IS PROFITABLE (CAPACITY FACTORS ARE FOR VENTING-ONLY CONFIG. NG PRICES ARE $/MMBTU).

FIGURE 11. THE TRANSITION TO PRIMARILY FULL-LOAD CO2 CAPTURE IS FASTER WITH SOLVENT STORAGE BUT IS SLOWER AT HIGH AND LOW NATURAL GAS PRICES (NG PRICES ARE $/MMBTU).

Margina al Cost (MC) or Price ($/MWh)

$35/tCO2

140

capture system, which is not likely the case at the lowest CO2 prices. At intermediate prices, profits with venting-only flexible capture exceed those of inflexible and no capture because there are times when additional electricity sales with capture systems at reduced load offset the costs of venting CO2 . Venting CO2 is uneconomical at higher CO2 prices, but benefits persist with solvent storage at higher CO2 prices because the facility can perform price arbitrage without incurring additional emissions costs. A primary economic metric for flexible capture is the difference between operating profits with flexible capture and the maximum of those with inflexible or no CO2 capture. This incremental operating profit benefit can then be weighed against any incremental capital costs of flexibility. A contour plot of the incremental operating profit benefit, normalized by gross power output capacity, is plotted for the full natural gas price range and $0–130/tCO2 in Fig. 15 for venting-only flexible capture and Fig. 16 for flexible capture with solvent storage. Figure 15 demonstrates that the emissions penalty from venting CO2 is worth the benefit of additional electricity sales only in a CO2 price range of approximately $30–60/tCO2 , with the range and benefit being greater at higher natural gas prices. Higher electricity prices that accompany high natural gas prices increase the frequency of instances where venting CO2 is profitable, and this benefit is largely absent below natural gas prices of about $4/MMBTU. Natural gas prices have been below $3/MMBTU in the United States in 2012, so price-responsive venting-only flexible capture is unlikely to provide significant benefit if these market conditions persist [23]. The analogous plot with solvent storage (Fig. 16) demonstrates that solvent storage extends the benefit of flexibility to

120 100

Breakeven price to justify venting

80

MC at 100% Capture

60 40

MC at 0% Capture

20 0 0

10


40

50

FIGURE 12. VENTING CO2 IS MORE OFTEN PROFITABLE WHEN TURNING CAPTURE OFF CAUSES ONLY A SLIGHT INCREASE IN MARGINAL COSTS.

$0–80/tCO2 for all plant configurations. Operating profit trends mimic those of base plant capacity factor; as CO2 prices increase, profits fall without CO2 capture and rise with inflexible capture because emissions costs at marginal facilities in the electricity system are typically between those of a coal-fired power plant with and without CO2 capture. Profits with flexible capture generally follow the upper envelope of the inflexible and no capture curves. Profits are slightly lower for CO2 prices below $35/tCO2 because the capture cost model includes a fixed operating and maintenance (FOM) cost to keep capture systems available. Realistically, this cost would only be incurred given a reasonable expectation to use the CO2



higher CO2 prices and increases its magnitude in the region where venting is valuable. However, benefits at high CO2 prices are greatest at low to moderate natural gas prices. The benefit of solvent storage is reduced when both CO2 and natural gas prices are high because these market conditions provide little opportunity when increased energy cost while regenerating stored solvent is worth the increased electricity sales revenue when solvent is stored at low stripping/compression load. Operational results suggest that the preferred time to regenerate stored solvent is when electricity prices are near or slightly above marginal costs at full-load CO2 capture, when increased energy costs while regenerating stored solvent result in the facility taking a small loss. These conditions occur frequently at high CO2 prices and low to moderate natural gas prices, but high natural gas prices produce high electricity prices where lost profit while regenerating stored solvent is seldom worth the increased profits while storing solvent. Under these circumstances, solvent storage is used less often and its operating profit benefit decreases. The benefit of solvent storage is closely related to the plant capacity factor. Figure 17 plots the same data from Fig. 8 for the solvent storage configuration as a contour plot. Comparing this figure to Fig. 16, solvent storage is most valuable when capacity factors are 40–80% given CO2 prices above $30/tCO2 . The high CO2 -high natural gas price region of decreasing solvent storage benefits corresponds to the high capacity factor region, suggesting that expected capacity factor might be a good metric to decide whether or not to build a solvent storage system. Figure 18 explores this relationship by plotting capacity factor and operating profit benefits from solvent storage for $0–200/tCO2 and the minimum, maximum, and actual 2010 natural gas prices. For instance, if $4/kW is the minimum acceptable annual profit benefit set by capital costs, a capacity factor can be identified for each natural gas price above which that benefit is no longer achieved. For $5.14/MMBTU and $11/MMBTU natural gas, this capacity factor is approximately 90%, indicating that in general, a capacity factor above 90% would not achieve a $4/kW annual benefit from solvent storage. This capacity factor is achieved with $2/MMBTU natural gas at $200/tCO2 , and trends suggest that the benefit from solvent storage at $2/MMBTU will fall below $4/kW at higher CO2 prices. When market conditions are not conducive to regenerating stored rich solvent, solvent storage is likely unattractive unless capital costs are very low.

Norm malize ed Annual Operatting Prrofits ($/kW) O

120

$5.14/MMBTU NG

100 No capture

g Solvent Storage

80 Venting-only

60 40 Inflexible

20 0 0

10

20

30 40 50 60 CO2 Price ($/tCO2)

70

80

FIGURE 14. PROFITS FOR EACH CONFIGURATION DEPEND ON CAPACITY FACTOR AND THE AVAILABILITY OF FLEXIBLE CO2 CAPTURE SYSTEMS.

FIGURE 15. VENTING CO2 AT HIGH PRICES IS VALUABLE ONLY WITH $40–60/tCO2 AND HIGH NATURAL GAS PRICES.

CONCLUSIONS A profit maximization model for a single power generating facility has been utilized to determine the environmental and economic implications of electricity price-responsive flexible CO2 capture with and without solvent storage for a wide range of market conditions: $2–11/MMBTU natural gas price, and $0– 200/tCO2 CO2 price. Flexible CO2 capture systems transition from rarely using capture systems to near-100% utilization at

FIGURE 16. SOLVENT STORAGE EXTENDS BENEFITS TO HIGHER CO2 PRICES, ESPECIALLY AT LOW NATURAL GAS PRICES.



Solvent storage improves incremental profits of flexible capture and broadens the region where flexible CO2 capture is valuable to higher CO2 prices at low-to-moderate natural gas prices. The conditions where solvent storage has significant value correspond to power plant capacity factors below 90%, and the economic benefit of solvent storage is greatest when expected capacity factors are 40–80% given CO2 prices high enough to justify CO2 capture operation. Under the present case study, this situation occurs with CO2 prices above $40/tCO2 and natural gas prices below ∼$6/MMBTU. The optimal time for regenerating stored solvent is when its additional energy cost produces a small profit loss, and these conditions rarely occur when both CO2 and natural gas prices are high because electricity prices are too high for the loss to be worth utilizing solvent storage. These results illustrate the expected value of flexible CO2 capture in response to electricity price variations across a wide range of electricity market conditions. Given a forecast of future electricity market conditions and an estimation of the incremental capital costs of flexible capture, this model and its results can be used to perform investment analysis and make decisions on whether or not to install flexible or inflexible CO2 capture.

27

1.0

11NG

24

0.8 0.6

Normalized Annua al Operating Profit Improvem ment ($/kW)

21 5.14NG

18

2NG

0.4 0.2

15

0.0

12 11NG

9

5 14NG 5.14NG

Base Plant Capacity Factor (fractional)

FIGURE 17. THE VALUE OF SOLVENT STORAGE IS CORRELATED TO THE POWER PLANT CAPACITY FACTOR.

ACKNOWLEDGMENT The authors gratefully acknowledge support from the United States Environmental Protection Agency (EPA) Science to Achieve Results (STAR) Graduate Fellowship Program and the Luminant Carbon Management Program. However, the views expressed within this publication might not reflect the views of the sponsoring organizations.

-0.2

2NG

6

-0.4

3

-0.6

0

-0.8

-3

-1.0 0

50 100 150 CO2 Price ($/tCO2)

200

REFERENCES [1] IPCC, 2008. Climate Change 2007 Synthesis Report. Geneva, Switzerland. [2] USDOE, 2012. Annual Energy Outlook 2012 Early Release. U.S. Department of Energy, Energy Information Administration, January 23, 2012. [3] Rochelle, G., Chen, E., Freeman, S., Van Wagener, D., Xu, Q., and Voice, A., 2011. “Aqueous piperazine as the new standard for CO2 capture technology”. Chemical Engineering Journal, 171(3), July, pp. 725–733. [4] Ambrose, W. A., Breton, C. L., Duncan, I., Holtz, M. H., Hovorka, S. D., López, V. N., and Lakshminarasimhan, S., 2006. Source-sink matching and potential for carbon capture and storage in the gulf coast. The Gulf Coast Carbon Center, Bureau of Economic Geology, Jackson School of Geosciences, The University of Texas at Austin. [5] Metz, B., Davidson, O., de Coninck, H., Loos, M., and Meyer, L., 2005. IPCC special report on carbon dioxide capture and storage. IPCC, New York, NY.

FIGURE 18. THE MINIMUM ACCEPTABLE BENEFIT OF SOLVENT STORAGE CAN BE RELATED TO A MAXIMUM CAPACITY FACTOR WHERE THAT BENEFIT WILL BE ACHIEVED (NG PRICES ARE $/MMBTU).

$30–40/tCO2 , largely indpendent of natural gas price. This transition is slightly faster when facilities may use solvent storage to perform electricity price arbitrage without increasing CO2 emissions. Average annual CO2 emissions rates are slightly higher at transition CO2 prices for low and high natural gas prices because venting CO2 is profitable for a larger fraction of online time under these conditions. Consistent with earlier results by these and other authors, venting CO2 while selling additional electricity at high prices is valuable only in a narrow CO2 price regime, $30–60/tCO2 under the conditions studied. In addition, this work finds that natural gas prices must exceed $4/MMBTU before electricity prices are consistently high enough for venting to be valuable.



[6] Javedan, H., 2012. Carbon capture and sequestration project database. http://sequestration.mit. edu/tools/projects/index.html. [7] McCoy, S. T., and Rubin, E. S., 2008. “An engineeringeconomic model of pipeline transport of CO2 with application to carbon capture and storage”. International Journal of Greenhouse Gas Control, 2(2), pp. 219–229. doi: 10.1016/S1750-5836(07)00119-3. [8] Gibbins, J. R., and Crane, R. I., 2004. “Scope for reductions in the cost of CO2 capture using flue gas scrubbing with amine solvents”. Proceedings of the Institution of Mechanical Engineers – Part A – Power & Energy, 218(4), pp. 231–239. [9] USNETL, 2007. Cost and performance baseline for fossil energy plants. August 1, 2007. [10] Cohen, S. M., Rochelle, G. T., and Webber, M. E., 2010. “Turning CO2 capture on & off in response to electric grid demand: A baseline analysis of emissions and economics”. ASME Journal of Energy Resources Technology, 132. [11] Wiley, D. E., Ho, M. T., and Donde, L., 2010. “Technical and economic opportunities for flexible CO2 capture at australian black coal fired power plant”. In 10th International Conference on Greenhouse Gas Technologies, Elsevier. [12] Chalmers, H., Lucquiaud, M., Gibbins, J., and Leach, M., 2009. “Flexible Operation of Coal Fired Power Plants with Postcombustion Capture of Carbon Dioxide”. Journal of Environmental Engineering(June 2009). [13] Chalmers, H., Leach, M., and Gibbins, J., 2010. “Built-in flexibility at retrofitted power plants: what is it worth and can we afford to ignore it?”. In 10th International Conference on Greenhouse Gas Technologies, Elsevier. [14] Husebye, J., Anantharaman, R., and Fleten, S.-E., 2010. “Techno-economic Assessment of Flexible Solvent Regeneration & Storage for Base Load Coal-Fired Power Generation with Post Combustion CO2 Capture”. In 10th International Conference on Greenhouse Gas Technologies, Elsevier. [15] Cohen, S. M., Rochelle, G. T., and Webber, M. E., 2012. “Optimizing post-combustion CO2 capture in response to volatile electricity prices”. International Journal of Greenhouse Gas Control, 8, pp. 180–195. [16] Ziaii, S., Cohen, S., Rochelle, G. T., Edgar, T. F., and Webber, M. E., 2008. “Dynamic operation of amine scrubbing in response to electricity demand and pricing”. In 9th International Conference on Greenhouse Gas Technologies, Elsevier. [17] Chalmers, H., and Gibbins, J., 2007. “Initial evaluation of the impact of post-combustion capture of carbon dioxide on supercritical pulverised coal power plant part load performance”. Fuel, 86(14), pp. 2109–2123. [18] Cohen, S., Fyffe, J., Rochelle, G. T., and Webber, M. E., 2009. “The Effect of Fossil Fuel Prices on Flexible CO2

[19]

[20]

[21]

[22] [23] [24] [25]

[26] [27] [28]

[29]

[30]

[31]

[32]

[33]

Capture Operation”. In ASME 3rd International Conference on Energy Sustainability. Cohen, S. M., Rochelle, G. T., and Webber, M. E., 2010. “Optimal operation of flexible post-combustion CO2 capture in response to volatile electricity prices”. In 10th International Conference on Greenhouse Gas Control Technologies, Elsevier. Patiño Echeverri, D., and Hoppock, D. C., 2012. “Reducing the Energy Penalty Costs of Postcombustion CCS Systems with Amine-Storage”. Environmental Science & Technology, 46(2), pp. 1243–1252. ERCOT, 2011. Ercot Quick Facts. http://www. ercot.com/content/news/presentations/ 2011/ERCOT\%20Quick\%20Facts\%20-\ %20Oct\%202011.pdf, October 2011. USEPA, 2010. Emissions & generation resource integrated database (eGRID). Bloomberg, 2012. Energy & Oil Prices. ERCOT, 2010. Ercot hourly load data. http://www. ercot.com/gridinfo/load/load_hist/. ERCOT, 2010. Balancing energy services market clearing prices for energy annual reports archives. http://www. ercot.com/mktinfo/prices/mcpea. ERCOT, 2011. Generic unit database. USEPA, 2008. Emissions & generation resource integrated database (eGRID), December 2008. Mortensen, J. H., Moelbak, T., Andersen, P., and Pedersen, T. S., 1998. “Optimization of boiler control to improve the load-following capability of power-plant units”. Control Engineering Practice, 6(12), pp. 1531–1539. doi: DOI: 10.1016/S0967-0661(98)00114-2. Pang, C. K., and Chen, H. C., 1976. “Optimal short-term thermal unit commitment”. Power Apparatus and Systems, IEEE Transactions on, 95(4), pp. 1336–1346. Wagener, D. H. V., and Rochelle, G. T., 2011. “Stripper configurations for CO2 capture by aqueous monoethanolamine”. Chemical Engineering Research and Design, 89(9), pp. 1639–1646. Cohen, S. M., 2012. “A techno-economic plant- and gridlevel assessment of flexible co2 capture (in preparation)”. PhD thesis, The University of Texas at Austin. PE-Ltd., 2011. 2010 State of the Market Report for the ERCOT Wholesale Electricity Markets. Potomac Economics, Ltd., August 2011. EU-ECX, 2008. Historical data - ECX EUA futures contract.

