Text

Entergy Pre-Filed Hearing Exhibit ENT000517, Determining the Least-Cost Investment for an Existing Coal Plant to Comply with EPA Regulations Under Uncertainty
	ML12090A852
Person / Time
Site:	Indian Point
Issue date:	02/29/2012
From:	Echeverri D, Gumerman E, Hoppock D; Duke Univ
To:	; Atomic Safety and Licensing Board Panel
	SECY RAS
Shared Package
ML12090A851	List: ML12090A852; ML12090A853; ML12090A854; ML12090A855; ML12090A858; ML12090A860; ML12090A861; ML12090A862;
References
	RAS 22162, 50-247-LR, 50-286-LR, ASLBP 07-858-03-LR-BD01
	Download: ML12090A852 (43)
	v • d • e

ENT000517 Submitted: March 30, 2012 Nicholas Institute for Environmental Policy Solutions Working Paper NI WP 12-03 February 2012 DETERMINING THE LEAST-COST INVESTMENT FOR AN EXISTING COAL PLANT TO COMPLY WITH EPA REGULATIONS UNDER UNCERTAINTY David Hoppocka Dalia Patino Echeverrib Etan Gumermana a

Nicholas Institute for Environmental Policy Solutions, Duke University b

Nicholas School of the Environment, Duke University

Executive Summary Low natural gas prices and forthcoming EPA regulations for coal plant emissions, coal wastes, and thermal-generation cooling systems are forcing utilities and utility regulators to decide whether to retrofit or to retire and replace existing coal plants. Complicating this task is uncertainty about fuel prices, the requirements of and timelines for the EPA regulations, and long-term federal energy and climate policy.

This uncertainty makes identification of the least-cost investment in the near and long term difficult, because incorrect assumptions about future fuel prices, energy policies, regulatory requirements, or other unforeseen factors could make a decision that is least-cost today very expensive for ratepayers in the future.

To help utility commissions and other interested parties make least-cost investment decisions and quantify cost risk for ratepayers, researchers at Duke University and the Nicholas Institute for Environmental Policy Solutions will make the Risk Based Decision Model (RBDM) available to the public. The RBDM determines an optimal investment or series of investments across multiple scenarios and future constraints specified by the user and can be employed to estimate the impact of abrupt changes (shocks) and the cost of making bad investments that are later abandoned. The RBDM is not a substitute for the comprehensive system-wide modeling typically performed by utilities but rather is a complementary tool for the utility investment decision process.

To demonstrate the RBDM, we modeled the least-cost investment decision for Louisville Gas and Electrics (LGE) Mill Creek coal-fired power plant to meet the forthcoming EPA regulations under uncertainty using publicly available data. LGE was one of the first utilities to submit a proposal to comply with the forthcoming EPA regulations for coal plant emissions, and the decision to retrofit or retire the Mill Creek plant is an example of the difficult decisions facing utilities and utility regulators. This exercise had two goals. The first was to determine the least-cost option for LGE and Kentucky ratepayers:

retrofit the Mill Creek plant to meet the regulations or retire and replace it with new natural gas or coal generation. The second goal was to demonstrate the capabilities of the RBDM.

In June 2011, LGE submitted a $1.3 billion proposal to comply with the Cross State Air Pollution Rule (CSAPR), the Mercury Air Toxics Rule (MATS), and a local air quality requirement.1 We modeled the least-cost investment decisions for the Mill Creek plant to comply with CSAPR, MATS, the Coal Combustion Residual Rule (CCR), and Cooling Water Rule 316b under natural gas price, regulatory, and climate policy uncertainty using two sets of scenarios with and without a price for carbon emissions.

Forecasts for each scenario were created using the Nicholas Institutes version of the Energy Information Administrations National Energy Modeling System. The scenario forecasts for fuel, wholesale electricity, and emissions allowances prices as well as construction capital costs are entered as inputs into the RBDM.2 1

The local air quality requirement is to achieve compliance with the 1-hour SO2 national ambient air quality standard (NAAQS).

2 The RBDM user can enter any scenario forecast data the user believes adequately represent the users scenarios.

2

Modeling scenarios with carbon prices Modeling scenarios without carbon prices Scenario Natural gas Regulation Carbon Scenario Natural Regulation Carbon price price gas price price 1 Baseline Standard No price 1 Baseline Standard No price 2 Baseline Less stringent No price 2 Baseline Less stringent No price 3 Baseline More stringent Low 2020 3w Baseline More No carbon stringent price price 4 Baseline Standard Low 2020 5 High Standard No price 5 High Standard No price 6 High Less stringent No price 6 High Less stringent No price 7 w/o High More No carbon stringent price price 7 High More stringent Mid-2020 9 Extra high Standard No price 8 High Standard Mid-2020 10 Extra high Less stringent No price 9 Extra high Standard No price 11 w/o Extra high More No carbon stringent price price 10 Extra high Less stringent No price 11 Extra high More stringent High 2020 12 Extra high Standard High 2020 For this analysis, we assumed that each scenario is equally likely at the beginning of the modeling period (2011) and that regulatory certainty is achieved in 2020. The model runs multiple times to converge on each scenario and generates least-cost investment decisions for each convergence.3 The changes in scenario probabilities and convergence on each scenario simulate the changes in uncertainty that occur as regulations and legislation are enacted as well as the ability of the decision maker to wait for new information and revise previous decisions. The model user determines the year that uncertainty is resolved, initial scenario probabilities, and how scenario probabilities change year to year as the model converges on each scenario. We included more than 30 investment options in the RBDM4 to capture the full range of retrofit investment pathways to comply with the forthcoming EPA regulations5 and new generation investment options, including options to retrofit new plants with carbon capture and storage.

For both sets of scenarios, with and without carbon prices, the RBDM determined it is least-cost to retrofit the Mill Creek plant to comply with the forthcoming EPA regulations in all scenarios. This finding indicates that LGEs proposal to retrofit the Mill Creek plant is robust under our representation of natural gas price, regulatory, and carbon policy uncertainty.

3 For 12 scenarios, for example, 12 sets of investment and operations results are generated. For 9 scenarios, the model outputs 9 sets of investment and operations results.

4 The model user can enter up to 50 investment options.

5 For example, the Mill Creek plant could be retrofit to comply with CSAPR, MATS, CCR, and Cooling Water Rule 316b at once or in multiple stages.

3

Least-cost investment decisions for 9 scenarios without carbon prices and 12 scenarios with carbon prices in selected scenarios for initial 10 years of modeling period Least Cost Investments 9 Scenarios without Carbon Price Scenarios Year 1 2 3 4 5 6 7 8 9 10 1 Mid NG, Standard EPA 0 0 MATS Sub C Sub D 0 316b E 316b I 0 0 Mid NG, Less Stringent 2 0 0 MATS Sub C Sub D 0 316b E 316b I 0 0 EPA 3 w/o Mid NG, More Stringent Ca rbon EPA MATS Sub C Sub D 316b E Pri ce 0 0 0 0 0 0 5 High NG, Standard EPA 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 High NG, Less Stringent 6 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 EPA 7 w/o High NG, More Stringent Ca rbon EPA MATS Sub C Sub D 316b E Pri ce 0 0 0 0 0 0 Extra High NG, Standard 9 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 EPA Extra High NG, Less 10 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 Stringent EPA 11 w/o Extra High NG, More Str.

Ca rbon EPA MATS Sub C Sub D 316b E Pri ce 0 0 0 0 0 0 Year 2011 2015 2020 Least Cost Investments 12 Scenarios with Carbon Price in Selected Scenarios Scenarios Year 1 2 3 4 5 6 7 8 9 10 MATS & Sub D &

1 Mid NG, Baseline EPA 0 0 Sub C 0 0 316b E 0 0 0 316b I Mid NG, Less Stringent MATS & Sub D &

2 0 0 Sub C 0 0 316b E 0 0 0 EPA 316b I Mid NG, More Stringent MATS & Sub D &

3 0 0 0 Sub C 0 316b E 0 0 0 EPA, Low Carbon Cost 316b I Mid NG, Baseline EPA, MATS & Sub D &

4 0 0 Sub C 0 0 316b E 0 0 0 Low Carbon Cost 316b I MATS & Sub D &

5 High NG, Baseline EPA 0 0 Sub C 0 0 316b E 0 0 0 316b I High NG, Less Stringent MATS & Sub D &

6 0 0 Sub C 0 0 316b E 0 0 0 EPA 316b I High NG, More Stringent MATS & Sub D &

7 0 0 Sub C 0 0 316b E 0 0 0 EPA, Mid Carbon Cost 316b I High NG, Baseline EPA, MATS & Sub D &

8 0 0 Sub C 0 0 316b E 0 0 0 Mid Carbon Cost 316b I Extra High NG, Baseline MATS & Sub D &

9 0 0 Sub C 0 0 316b E 0 0 0 EPA 316b I Extra High NG, Less MATS & Sub D &

10 0 0 0 Sub C 0 316b E 0 0 0 Stringent EPA 316b I Extra High NG, More Str. MATS & Sub D &

11 0 0 0 Sub C 0 316b E 0 0 0 EPA, High Carbon 316b I Extra High NG, Base EPA, MATS & Sub D &

12 0 0 0 Sub C 0 316b E 0 0 0 High Carbon 316b I Year 2011 2015 2020 Retrofit to comply MATS Retrofit to comply CCR Subtitle D Retrofit to comply MATS, CCR Subtitle D, 316b Impingement Retrofit to comply 316b Entrainment Retrofit to comply CCR Subtitle C Retrofit to comply 316b Impingement For all scenarios, the RBDM makes hedging investments to comply with the more stringent CCR (Subtitle C) and Cooling Water Rule 316b (entrainment) requirements. The RBDM is structured such that running additional analyses with different costs, constraints, and scenario probabilities is relatively simple.

As more information about the CCR and Cooling Water Rule 316b become available, we can rerun the RBDM to optimize investments to comply with these rules. Additionally, users can run sensitivity analyses to determine cost and probability thresholds that lead to different investment outcomes and can hide selected scenarios to model shocks such as sudden changes in fuel prices or carbon policy.

4

Introduction Over the next few years, utilities and state utility regulators will be forced to make difficult investment decisions for existing coal power plants due to existing and forthcoming U.S. Environmental Protection Agency (EPA) regulations and to price competition from natural gas-fired generation. These decisions will be made in an environment of uncertainty about regulatory requirements and timelines, long-term federal energy and climate policy, and fuel costs driven by significant new supplies of natural gas.

Ensuring low-cost power for ratepayers will require balancing short-term costs with long-term cost risk.

Unaccounted for or unforeseen factors can make an investment decision that is low cost today significantly higher cost in the future. For example, the decision to retrofit an existing coal plant could create significant cost risk for ratepayers if Congress enacts a climate policy to price greenhouse gas emissions. Likewise, ratepayers could be hit hard if the utility replaces a coal power plant with a natural gas plant and natural gas prices rise. Balancing short-term costs and long-term cost risk is especially difficult in the electric utility sector, because generation and transmission investments tend to have operating lives of 30 or more years.

In traditionally regulated states, in exchange for exclusive franchise rights in a utilitys service territory, a utility must receive utility regulatory commission approval of all major capital investments.6 This system for approving generation investments presents a number of challenges given the uncertainty in the electric utility sector. Many utility commissions lack their own modeling tools, making them reliant on utility or consultant modeling resources. This dependency makes it more difficult for commissions and commission staff to answer questions about investment options under uncertainty because the lack of direct control over a model makes it more difficult to model a wide range of scenarios and control model inputs. Utility generation proposals often address immediate requirements and may not take into account future constraints and uncertainties, leading to investments that may not be least-cost over the projected life of the investment. Most decisions about generation investments are based on scenario analysis. Scenario analysis is a useful tool but does not optimize investments across scenarios, making determination of an optimal hedging investment difficult. Moreover, scenario analysis generally does not account for managerial flexibility, such as the ability to postpone investments, make bad investments that are later abandoned, or make investments in stages (retrofits).

The Risk-Based Decision Model (RBDM) is a modeling tool that can help utilities and utility regulators tackle the challenges of selecting optimal investments under uncertainty and estimating cost risk for ratepayers. The RBDM identifies the optimal investment or series of investments across multiple scenarios and future constraints as determined by the model user and can be employed to estimate the impact of abrupt changes (shocks) and the cost of making bad investments that are later abandoned.

The Nicholas Institute for Environmental Policy Solutions and Duke University plan to make the RBDM publically available to utility commissions and others. The RBDM is not a substitute for utilities comprehensive system-wide modeling but is a complementary tool for utility investment decision making.

To demonstrate the RBDM, researchers at Duke University and the Nicholas Institute used publically available data to model the least-cost investment decision for Louisville Gas and Electrics (LGE) Mill Creek coal-fired power plant to meet the forthcoming EPA regulations. LGE was one of the first utilities to submit a proposal to comply with the forthcoming EPA regulations, and the decision to retrofit or retire the Mill Creek plant is an example of the difficult decisions facing utilities and utility regulators over the next few years. The goal of this work is both to determine if it is least-cost to retrofit the Mill Creek plant to meet the forthcoming regulations or retire and replace it with new coal or natural gas generation given 6

T. Brennan, K. Palmer, and S. Martinez, Alternating Currents Electricity Markets and Public Policy (Washington, DC:

Resources For the Future, 2002).

5

the uncertainties facing LGE and Kentucky ratepayers, as well as to demonstrate the capabilities of the RBDM. In June 2011, Louisville Gas and Electric submitted an application (Case No. 2011-00162) to the Kentucky Public Service Commission for approval of a proposed $1.3 billion retrofit of the Mill Creek plant to comply with the EPAs new 1-hour sulfur dioxide (SO2) National Ambient Air Quality Standard (NAAQS), the (then) proposed Clean Air Transport Rule, and the proposed Mercury and Air Toxics Standards (MATS).

New and Proposed EPA Regulations Affecting Electric Utilities The EPA is researching and promulgating regulations that will affect existing fossil fuel generation in the United States. These regulations include the Cross State Air Pollution Rule (CSAPR), previously named the Clean Air Transport Rule; capping SO2 and nitrogen oxide (NOx) emissions in selected states; the MATS rule for hazardous air pollutants, the Coal Combustion Residuals Rule (CCR) and the Cooling Water Intake Structures rule for thermal power plants (Cooling Water Rule 316b). Additionally, as is required by the Clean Air Act, EPA must periodically review and revise, as necessary, the NAAQS for carbon monoxide (CO), lead, nitrogen oxide (NOx), ozone, particulate matter (PM), and SO2.7 EPA recently postponed updating the ozone NAAQS to reflect the recommendation of a scientific advisory board until 2013.8 Like most other modeling of the forthcoming EPA regulations,9 we did not explicitly include the updated NAAQS in our modeling. However, the CSAPR was promulgated to facilitate downwind states compliance with the PM2.5 and ozone NAAQS10, and compliance with CSARP and MATS will significantly reduce electricity sector NOx, SO2, and PM2.5 emissions.11 The EPA is also promulgating New Source Performance Standards (NSPS) for greenhouse gas (GHG) emissions for power plants and refineries. The NSPS will apply to new and modified sources. In addition, the EPA will provide GHG emissions guidelines for regulation of existing GHG sources under state rules. We did not attempt to model the upcoming EPA regulations on GHG emissions, because the proposed rule has not yet been released and it is uncertain what it will require, but we did include uncertainty about GHG regulation in our modeling.

Modeling Overview The RBDM modeling process begins with the user selecting scenarios that capture the uncertainties the user wants to model and obtaining a forecast for each scenario. Scenario forecasts are used as inputs into the RBDM. Before running the model, users must also determine all the investment options and constraints included in the RBDM.

In formulating our modeling framework, we have attempted to capture the major uncertainties that will affect the least-cost investment decision for LGEs Mill Creek plant. We began this process by selecting scenarios that capture the likely range of future fuel price, regulatory and policy uncertainty based on 7

Standards for NOx, ozone, PM, and SO2 are the primary standards affecting the electric utility sector.

8 John Broder, Obama Administration Abandons Stricter Air-Quality Rules, New York Times, September 3, 2011, http://www.nytimes.com/2011/09/03/science/earth/03air.html?scp=3&sq=EPA%20Ozone%20delay&st=cse.

9 Paul J. Miller, A Primer on Pending Environmental Regulations and Their Potential Impacts on Electric System Reliability (Boston, MA: Northeast States for Coordinated Air Use Management, 2011).

10 U.S. EPA (Environmental Protection Agency), The Cross-State Air Pollution Rule: Reducing the Interstate Transport of Fine Particulate Matter and Ozone, http://www.epa.gov/airtransport/pdfs/CSAPRFactsheet.pdf.

11 U.S. EPA Office of Air and Radiation, Regulatory Impact Analysis for the Federal Implementation Plans to Reduce Interstate Transport of Fine Particulate Matter and Ozone in 27 States; Correction of SIP Approvals for 22 States, http://www.epa.gov/airtransport/pdfs/FinalRIA.pdf; U.S. EPA, Regulatory Impact Analysis of the Proposed Toxics Rule: Final Report, http://www.epa.gov/ttn/ecas/regdata/RIAs/ToxicsRuleRIA.pdf.

6

present knowledge about these uncertainties. These scenarios were then represented using the Nicholas Institutes version of the Energy Information Administrations National Energy Modeling System (NI-NEMS) to generate forecasts of fuel and wholesale electricity prices and of construction capital costs for each scenario. These forecasts were used as inputs into the RBDM to determine the least-cost investment option for the Mill Creek plant to comply with the forthcoming CSAPR, MATS, CCR and Cooling Water Rule 316b under uncertainty.12 The options for investment include retrofitting the existing Mill Creek plant or retiring it and building new generation capacity. The RBDM user can utilize any scenario forecast data the user believes adequately represent the users scenarios.

The RBDM was developed by Nicholas School of the Environment Professor Dalia Patino Echeverri13 and is a multi-period decision model with an embedded multi-stage stochastic optimization program that minimizes the expected total costs of plant operation, capital investments, and emissions costs over time across multiple scenarios given the forecasts for each modeling scenario. It represents a rational decision maker who is attempting to minimize costs given his or her knowledge about the likelihood of future scenarios and system constraints. The RBDM includes constraints that force the model to retrofit or retire plants that do not comply with the forthcoming EPA regulations. For this analysis, we assumed that each scenario is perceived as equally likely in 2011 and that the decision maker learns more about the probability of each scenario each year until regulatory certainty is achieved in 2020 when the model converges on a scenario.14 The model runs multiple times to converge on each scenario and generates least-cost investment and operation decisions for each convergence.15 The changes in scenario probabilities and convergence on each scenario simulate the changes in uncertainty that occur as regulations and legislation are enacted as well as the ability of the decision maker to wait for new information and revise previous decisions. The model user determines initial scenario probabilities, the year in which uncertainty is resolved, and the way that scenario probabilities change from year to year as the model converges on each scenario.

Scenario Development The goal of our NI-NEMS scenario modeling is to capture the foreseeable range of policy, technology, and fuel supply scenarios given present knowledge about uncertainties that could reasonably be expected to affect the future of the electricity sector and, specifically, the key inputs to the RBDM for the Mill Creek plant. Major uncertainties that could affect the future of the electricity sector include x Natural gas supply x Coal supply x Potential for future carbon price/policy x Delays or major adjustments to the proposed MATS rule, the CCR rule, and Cooling Water Rule 316b x Technology change, including that induced by government policies to reduce the capital cost of specific types of generation.

12 Compliance with the CSAPR and MATS is assumed to achieve compliance with the new SO 2 NAAQS.

13 The model has also been used to design optimal incentives for investment in Carbon Capture and Storage (CCS) under federal carbon legislation by Professor Patino Echeverri, Dallas Burtraw (RFF), and Karen Palmer (RFF).

14 For this modeling exercise we chose 2020 as the year to resolve uncertainty. We set the initial scenario probabilities equal to avoid giving different sources of uncertainty greater weight or making assumptions about what scenario is a more likely representation of the future.

15 Given 12 scenarios, for example, the RBDM generates 12 sets of investment and operations results. If there are 9 scenarios the model outputs 9 sets of investment and operations results.

7

Adding more uncertainty variables to the model makes it more difficult to determine which uncertainties have the greatest impact on forecasts and can quickly lead to a large number of scenarios. Therefore, for this analysis we included natural gas supply, federal climate policy, and EPA regulatory uncertainty but ignored coal supply and technological change uncertainty. Coal supply scenarios would likely account for potential regulation of mountain-top removal coal mining, which is primarily used to mine low-sulfur coal. Historically, the Mill Creek plant has burned high-sulfur coal;16 with the proposed upgrade of the plants existing SO2 scrubbers, the price of low-sulfur coal is assumed to have little to no impact on the Mill Creek plant investment decision. Prior Nicholas Institute modeling indicates that foreseeable changes in the capital costs of emerging technology17 and in government policies to reduce the costs of specific generation (e.g., renewable generation) have a minimal impact on fossil fuel prices and capital costs for fossil-fuel generation forecasts relative to any carbon policy or natural gas supply assumptions.

Based on our narrowed list of uncertainties, we created scenarios for natural gas supply/price, EPA regulations, and carbon policy. The three natural gas scenarios are x Baseline: Based on the Energy Information Administrations reference scenario in the Annual Energy Outlook 2011 x Low supply/high price x Extra-low supply/extra-high price The EPA regulation scenarios are x Standard: CSAPR, MATS rule beginning in 2016, CCR Subtitle D beginning in 2018, Cooling Water Rule 316b beginning in 2020 x Less stringent: CSAPR, MATS rule beginning in 2018, CCR Subtitle D beginning in 2020, Cooling Water Rule 316b beginning in 2020 x More stringent: CSAPR, MATS rule beginning in 2016, CCR Subtitle C beginning in 2018, Cooling Water Rule 316b beginning in 2020 and requiring cooling towers for all cooling units with a design intake flow rate of more than 125 million gallons per day (MGD) o Fossil plants have until 2022 to install cooling towers o Nuclear plants have until 2027 to install cooling towers.

For carbon policy, we ran two sets of scenarios, one without any future carbon price and one in which half the scenarios have a range of carbon prices. To account for the difficulties associated with financing and permitting new coal generation, all scenarios include EIAs carbon risk adder for new coal plants.

Using combinations of these scenarios, we created two sets of NI-NEMS modeling scenarios; the first set includes 12 scenarios, 6 of which include a future carbon price, and the second set includes 9 scenarios, none of which include a future carbon price.

16 EIA (Energy Information Administration), Monthly Utility and Nonutility Fuel Receipts and Fuel Quality Data, http://www.eia.gov/cneaf/electricity/page/eia906_920.html.

17 Emerging technologies, such as solar generation, are technologies that are not widely adopted and that have greater potential for significant capital and operating cost reductions relative to traditional fossil fuel generation technologies. Natural gas turbines and pulverized-coal steam generation are mature technologies.

8

Table 1. NI-NEMS modeling scenarios with future carbon price scenarios Scenario Natural gas Regulation Carbon price price 1 Baseline Standard No price 2 Baseline Less stringent No price 3 Baseline More stringent Low 2020 4 Baseline Standard Low 2020 5 High Standard No price 6 High Less stringent No price 7 High More stringent Mid-2020 8 High Standard Mid-2020 9 Extra high Standard No price 10 Extra high Less stringent No price 11 Extra high More stringent High 2020 12 Extra high Standard High 2020 Table 2. NI-NEMS modeling scenarios without future carbon price scenarios Scenario Natural gas Regulation Carbon price price 1 Baseline Standard No price 2 Baseline Less stringent No price 3 w/o carbon price Baseline More stringent No price 5 High Standard No price 6 High Less stringent No price 7 w/o carbon price High More stringent No price 9 Extra high Standard No price 10 Extra high Less stringent No price 11 w/o carbon price Extra high More stringent No price NI-NEMS Modeling NI-NEMS is the Nicholas Institutes version of the Energy Information Administrations National Energy Modeling System (NEMS).18 It consists of four supply-side modules, four demand-side modules, two conversion modules, two exogenous modules, and one integrating module.19 NEMS is one of the most credible national modeling systems used to forecast the impacts of energy, economic, and environmental policies on the supply and demand of energy sources and end-use sectors. Its reference case forecasts are based on federal, state, and local laws and regulations in effect at the time of the prediction.20 The baseline projections developed by NEMS are published annually in the Annual Energy Outlook, which is regarded as a reliable reference in the field of energy and climate policy. NEMS is also used to conduct the sensitivity analyses of alternative energy policies and to validate research findings conducted by other government agencies, including the Environmental Protection Agency, Lawrence Berkeley National 18 EIA, The National Energy Modeling System: An Overview, http://www.eia.gov/oiaf/aeo/overview/.

19 Ibid.

20 EIA, Assumptions to the Annual Energy Outlook 2011, http://www.eia.gov/forecasts/aeo/assumptions/pdf/0554(2011).pdf.

9

Laboratory, Oak Ridge National Laboratory, and the Pacific Northwest National Laboratory. Each year, EIA creates multiple side cases for the Annual Energy Outlook to provide a range of forecasts for likely rules and regulations and to capture uncertainty in energy markets. We used many of these side cases, sometimes in combination, for our scenario modeling in NI-NEMS.

NI-NEMS Modeling of CSAPR and MATS We modeled CSAPR using the side case developed by EIA for the Clean Air Transport Rule (CATR).21 This side case, Transport Rule Mercury MACT 20, sets regional limits on SO2 and NOx utility sector emissions and assumes a 90 percent mercury maximum achievable control technology (MACT) standard for mercury emissions from coal plants. EIA models the MATS rule with its Retrofit Required 20 side case, requiring all coal plants to install flue gas desulfurization (FGD) and selective catalytic reductions (SCR) or retire by 2020.22 This side case likely over-estimates the cost of complying with the proposed MATS rule. The proposed rule creates emissions limits that can be met at some plants using relatively low-cost technologies, such as direct sorbent injection. It also allows each plant to average its emissions rates across all units at each plant site, indicating that some infrequently utilized units may not need to be retrofitted.23 For our Standard and More Stringent regulations scenarios we set the mercury MACT, FGD, and SCR compliance deadline at 2016. For the Less Stringent regulation scenario, the compliance deadline is 2018.

NI-NEMS Modeling of Cooling Water Rule 316b and CCR The compliance requirements of the proposed Cooling Water Rule 316b are based on the design intake flow of each plants cooling system. Plants with a flow rate greater than 2 million gallons per day (MGD) are required to meet impingement24 requirements. Plants with design intake flow rates greater than 125 MGD are required to study measures to reduce entrainment25 and meet the impingement requirements.26 Therefore, we created two scenarios for the cooling water rule:

x Baseline: All cooling units with design intake flow rates greater than 2 MGD are required to comply with the impingement requirements by 2020; zero costs are assumed for entrainment compliance.

x More Stringent: All cooling units with design intake flow rates greater than 125 MGD must install wet cooling towers if none exist by 2022 for fossil plants and 2027 for nuclear plants. All plants with design intake flow rates from 2 to 125 MGD are required to comply with the impingement requirements in the Baseline scenario. Any cooling unit installing a wet cooling tower is assumed to comply with the impingement requirements at no additional cost.

Compliance costs for each scenario were estimated on the basis of the design intake flow rate for each cooling unit27 and the cost assumptions for impingement and entrainment (cooling tower) retrofits in the EPAs economic and benefits analysis of the proposed cooling water rule.28 21 EIA released its Annual Energy Outlook 2011 on April 26, before EPAs release of the finalized CSAPR on July 6, 2011.

Although their state emissions caps differ, CATR and CSAPR achieve approximately the same net national emissions reduction.

22 EIA, Annual Energy Outlook 2011, http://www.eia.gov/forecasts/aeo/pdf/0383(2011).pdf.

23 U.S. EPA, Regulatory Impact of the Proposed Toxics Rule (Washington DC: U.S. EPA, 2011).

24 Impingement occurs when aquatic life becomes trapped on the cooling water intake screens.

25 Entrainment occurs when aquatic life is sucked into a cooling system.

26 U.S. EPA Office of Water, Proposed Regulations to Establish Requirements for Cooling Water Intake Structures at Existing Facilities, http://water.epa.gov/lawsregs/lawsguidance/cwa/316b/upload/factsheet_proposed.pdf.

27 Data from Energy Information Administration 767 database, http://www.eia.gov/cneaf/electricity/page/eia767/.

10

The proposed CCR rule has two primary regulatory pathways: coal combustion waste categorized as nonhazardous Subtitle D waste and coal combustion waste categorized as hazardous Subtitle C waste.29 Compliance costs for each coal plant were based on the EPAs regulatory impact analysis (RIA) of the proposed CCR rule and on disposal cost assumptions for Subtitle D and C in North American Electric Reliability Corporations (NERC) 2010 Special Reliability Scenario Assessment: Resource Adequacy Impacts of Potential U.S. Environmental Regulations. For Subtitle D (NERCs moderate case), NERC assumes that disposal costs for all waste currently disposed in surface impoundments (ash ponds) increase by $15/ton.30 EPAs RIA provides annual estimates of disposal in surface impoundments, onsite landfills, and offsite landfills as well as estimates of total annual ash generation for each plant. For Subtitle C, NERC assumes that disposal costs for all coal combustion byproducts (including waste currently sold or reused for beneficial uses) increase by $37.5/ton.31 A complete description of NI-NEMS modeling of the proposed CCR and Cooling Water 316b rules appears in this reports appendix.

NI-NEMS Modeling of Natural Gas Scenarios The baseline natural gas scenarios use the default natural gas supply assumptions in the EIAs Reference Case for the AEO 2011.32 The High Price/Low Supply and Extra High Price/Extra Low Supply scenarios were included to capture uncertainty about unconventional supplies of shale natural gas and potential regulations restricting shale gas drilling. The High Price/Low Supply scenario uses EIAs Low Shale EUR side case. The unproved technically recoverable shale gas resource and the ultimate gas recovery per well are 50 percent lower in the Low Shale EUR side case than in EIAs Reference Case.33 In the Extra High Price/Extra Low Supply scenario, we again adopted the Low Shale EUR side case but reduced the unproved technically recoverable shale gas resource an additional 50 percent, thereby reducing the unproved technically recoverable shale gas resource 75 percent relative to the Baseline scenario.

NI-NEMS Modeling Carbon Scenarios The scenarios with carbon prices use an economy-wide carbon tax beginning in 2020. The Low Carbon Price scenario begins at $15/ton CO2 equivalent; the Mid-Carbon Price scenario, at $25/ton CO2 equivalent; and the High Carbon Price scenario, at $45/ton CO2 equivalent.34 All carbon prices increase 5 percent annually.

NI-NEMS modeling results for key national energy benchmarks, such as Henry Hub natural gas prices, and the results used for inputs into the RBDM are available in the appendix.

Overview of Louisville Gas and Electrics Proposal for the Mill Creek Retrofit On June 1, 2011, LGE submitted its application for Certificates of Public Convenience and Necessity and for approval of its 2011 Compliance Plan for Recovery by Environmental Surcharge (Case No. 2011-00162) to comply with the EPAs new one-hour SO2 NAAQS, the (then) proposed CATR, and the 28 U.S. EPA, Economic and Benefits Analysis for Proposed Section 316(b) Existing Facilities Rule, http://water.epa.gov/lawsregs/lawsguidance/cwa/316b/upload/environbenefits.pdf.

29 U.S. EPA Office of Resource Conservation and Recovery, Regulatory Impact Analysis for EPAs Proposed RCRA Regulation of Coal Combustion Residuals (CCR) Generated by Electric Utilities Industry, http://rfflibrary.files.wordpress.com/2010/05/epa-hq-rcra-2009-0640-0003.pdf.

30 NERC (North American Electric Reliability Corporation), 2010 Special Reliability Scenario Assessment: Resource Adequacy Impacts of Potential U.S. Environmental Regulations, http://www.nerc.com/files/EPA_Scenario_Final.pdf.

31 Ibid.

32 EIA, Annual Energy Outlook 2011.

33 Ibid.

34 2009 dollars 11

proposed MATS rule for hazardous air pollutant emissions from coal-fired power plants.35 The submittal includes LGEs proposal for the Mill Creek plant. On April 21, 2011, Kentucky Utilities (KU) and LGE submitted their 2011 Joint Integrated Resource Plan (IRP)36 to meet demand, maintain reliability, and comply with environmental regulations through 2025 (LGE and KU are owned by the same parent company, PPL Corporation, and their systems are operated as one). KU submitted its application for Certificates of Public Convenience and Necessity and for approval of its 2011 Compliance Plan for Recovery by Environmental Surcharge (Case No. 2011-00161) on June 1, 2011.37 With the exception of KUs E.W. Brown plant, the IRP and environmental compliance plans do not explicitly address the CCR Rule, Cooling Water Rule 316b, GHG regulations under the Clean Air Act, or any potential future carbon price.38 LGE and KU began their planning process to comply with the forthcoming EPA regulations by conducting an engineering analysis for each generating unit to determine the least-cost retrofits to comply with the one-hour SO2 NAAQS, CATR, and MATS rule.39 They then used Strategist, commercial electricity planning modeling software, to determine which generating units should be retired or retrofitted given the estimated retrofit costs.40 Strategist determines a systems least-cost dispatch and capacity additions, including demand-side management, given existing generation assets as well as future demand, fuel prices, environmental regulations, and generation resources (wind, landfill gas, 50 MW new hydro, 800 MW supercritical coal, natural gas combustion turbine, and natural gas combined cycle (NGCC) 3x1, NGCC 2x1, and NGCC 1x141). KU and LGE ran Strategist using a baseline scenario to estimate a 30-year net present value for the total cost of their system, assuming all existing coal plants retrofit to meet the forthcoming air regulations. KU and LGE then analyzed the effect of retiring each existing coal unit to determine an alternative 30-year net present value of total cost, starting with the existing unit with the highest variable operating costs and repeating the process until they reached the unit with the lowest variable operating cost. If the 30-year net present value was lower for retirement than for a retrofit, the unit was retired and removed from the existing generation assets after its retirement date for the next Strategist run. The process was repeated until all existing coal units were analyzed. On the basis of this modeling, LGE and KU plan to retire six coal units (Tyrone 3; Green River 3 and 4; and Cane Run 4, 5, and 6) and forecast construction of 907 MW of new NGCC capacity in 2016, 907 MW NGCC in 2018, and 907 MW NGCC in 2025.42 In addition, LGE proposes to retrofit the Mill Creek plant as described below. We chose to analyze this investment decision because the Mill Creek plant has the highest total retrofit costs and highest retrofit costs per kW of capacity, indicating that retiring and replacing the plant might be a more cost-effective investment if additional regulatory costs and risks are considered.

35 LGE (Louisville Gas and Electric), Application of Louisville Gas and Electric Company for Certificates of Public Convenience and Necessity and Approval of Its 2011 Compliance Plan for Recovery by Environmental Surcharge (Case No. 2011-00162),

http://psc.ky.gov/Home/Library?type=Cases&folder=2011%20cases/2011-00162.

36 LGE and KU (Kentucky Utilities) Energy LLC, The 2011 Joint Integrated Resource Plan of Louisville Gas and Electric Company and Kentucky Utilities Company Case No. 2011-00140, http://psc.ky.gov/PSCSCF/2011%20Cases/2011-00140/20110421_LG%26E-KU_IRP_Volume%20I.pdf.

37 KU Energy LLC, The Application of Kentucky Utilities Company for Certificates of Public Convenience and Necessity and Approval of Its 2011 Compliance Plan for Recovery by Environmental Surcharge (Case No. 2011-00161),

http://psc.ky.gov/Home/Library?type=Cases&folder=2011%20cases/2011-00161.

38 The KU environmental compliance plan includes a proposal for E.W. Brown to comply with the CCR rule, because the plant was in the process of expanding its ash impoundments (ponds) when the EPA released its proposed CCR rule.

39 LGE and KU Energy LLC, 2011 Air Compliance Plan Generation Planning & Analysis, http://psc.ky.gov/Home/Library?type=Cases&folder=2011%20cases/2011-00162.

40 Ibid.

41 A 3x1 NGCC unit has three combustion turbines and one steam turbine powered by a heat recovery unit. A NGCC 2x1 unit has two combustion turbines and one steam turbine.

42 LGE and KU Energy LLC, 2011 Air Compliance Plan Generation Planning & Analysis.

12

Proposed Mill Creek retrofits:

x Unit 1: Fabric filter (baghouse), activated carbon injection, lime injection for sulfuric acid mist ($155 million) x Unit 2: Fabric filter (baghouse), activated carbon injection, lime injection for sulfuric acid mist ($151 million) x Combined Units 1 and 2: Remove existing FGDs, replace with new combined FGD serving both units

($354 million) x Unit 3: Fabric filter (baghouse), activated carbon injection, lime injection for sulfuric acid mist, remove existing FGD, selective catalytic reduction (SCR) improvements ($150 million) x Unit 4: Fabric filter (baghouse), activated carbon injection, lime injection for sulfuric acid mist, upgrade existing FGD and tie to unit 3, build new FGD for unit 4, SCR improvements ($459 million)

Risk Based Decision Model The RBDM is a multi-period decision model with an embedded multi-stage stochastic optimization program that minimizes the expected total costs of plant operation, capital investments, and emissions allowances over a time horizon across multiple scenarios. The model begins in 2011.

The Multi-Period Decision Making Model (MPDM) determines the optimal sequence of decisions by a utility over a planning horizon while uncertainty is resolved. Each year the MPDM selects the investment and operation options that minimize Expected Net Present Value of total costs over the following T=30 years (planning horizon) by solving the multi-stage Stochastic Optimization Model (SOM) described below. Once the investment and operations decisions for a given year are determined, the MPDM moves to the next year by updating (1) the current conditions, which are determined by the investments made in the past and (2) the probabilities assigned to the scenarios, which change over time to converge on individual scenarios (see Table 3 for example probabilities for converging on scenario 2). Thus, for each convergence, we assume that one of our S scenarios describes reality and that this reality will be revealed in 2020. In the last year of the MPDM (i.e., 2020), the SOM has collapsed to a deterministic optimization problem describing investment and operations decisions until the end of the planning horizon in 2049.

This formulation requires the specification of fuel prices, emissions allowances prices, and capital and O&M costs for each scenario for 39 years.

The SOM is a mixed integer linear program that includes binary variables representing the decision to invest in a retrofit or new generation and the decision to operate a plant (set to 1 if the control is installed/used and set to 0 otherwise) by the decision maker. The SOM determines the capital, O&M, fuel, and emissions costs for the current year plus the net present value of the capital, O&M, fuel and emissions cost for the next 30 years for each investment and operations option for each scenario. Each investment and operations option represents a combination of investment and operating decisions.43 The SOM then calculates these expected net present values for each investment and operating option by multiplying the expected net present value of the option by the probability of the scenario and summing for each scenario, resulting in a 30-year, expected net present value for each option. The SOM selects the lowest expected net present value investment and operations option and relays this information to the MPDM.

Thus, at each point in time, the model determines the optimal sequence of investment and operating decisions that account for the possibility of retrofitting an existing plant with different emissions controls to meet new regulations, building a new plant, mothballing, purchasing power from the wholesale market, 43 There are thousands of investment and operation options.

13

and shutting down a plant at any period of the planning horizon under different environmental, capacity, and annual generation constraints. By explicitly modeling the full flexibility of installation and operation of plants, and the irreversibility of capital investments (represented by the total capital cost, including financing costs at installation and a zero salvage value at shutdown), the model effectively accounts for the options available to a utility and utility commission to comply with new environmental regulations.

The SOM model assumes all operating plants and investment options are price takers and cannot affect market prices for fuel, electricity, capital investments, or emissions allowances.

Again, for this analysis we assume that initially the decision maker believes all scenarios are equally likely to occur. Table 3 shows assumptions for way that the probabilities of each scenario change year by year as the model converges on scenario 2. For convergence on the other scenarios, the change in probabilities is similar; the convergence scenario reaches 100 percent in 2020. In this modeling exercise, the initial probabilities of each scenario are set equal to one another and linearly decrease the probability of the nonconverging scenarios, while linearly increasing the probability of the converging scenario.

However, the model is flexible, so the user can change the initial probabilities and the way that probabilities change from year to year and reach convergence.44 44 Each year, the probabilities of all the scenarios must sum to 1.

14

Table 3. Probabilities for convergence on scenario 2 Scenario 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 1 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

2 8% 19% 29% 39% 49% 59% 69% 80% 90% 100%

3 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

4 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

5 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

6 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

7 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

8 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

9 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

10 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

11 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

12 8% 7% 6% 6% 5% 4% 3% 2% 1% 0%

Mill Creek Plant Characteristics, Modeling Assumptions, and Retrofit Investment Options The RBDM begins with the Mill Creek plant operating as reported in the joint LGE and KU IRP. The model includes multiple investment options to retrofit the plant and build new generation. The model assumes it takes 3 years to complete a retrofit or any new construction. We assume that any existing plant can continue to generate while it is being retrofitted.45 Additionally, we add a constraint that requires a plant to generate electricity all 39 years of the modeling period, removing the option to not operate a plant and to instead purchase power from the wholesale market. The model can make multiple investment decisions during the modeling period, including the decision to abandon a prior investment. For example, the model may determine it is least-cost to retrofit the Mill Creek plant in 2013 to comply with the MATS, CCR and 316b rules but then invest in a NGCC plant in 2025 after the model has converged on a scenario with low natural gas prices and a carbon price.46 Mill Creek is a four-unit coal plant with 1471.5 MW net capacity (average of summer and winter capacity). The units were built between 1972 and 1982. All of the units have SO2 scrubbers and electrostatic precipitators, and units 3 and 4 have selective catalytic reduction for NOx emissions.

Projections for Mill Creeks future heat rates, capacity factor, and availability factor, assuming the proposed retrofits, are available in the IRP (see Table 4).47 The existing Mill Creek plant in the RBDM is assumed to have a heat rate of 10,304 Btu/kWh and an annual generation of 11,327,775 MWh/year. On the basis of historical coal receipts, heat rate, and sulfur content, we assumed that Mill Creek primarily utilizes high-sulfur bituminous East Interior coal, a coal supply modeled in NI-NEMS.48 Current Mill Creek emissions rates for SO2 and NOx are 0.52 lbs/MMBtu and 0.16 lbs/MMBtu, respectively.49 CO2 emissions, at 203.1 lbs/MMBtu, were set equal to EIAs assumptions for high-sulfur bituminous East 45 If the RBDM indicates that building a new plant is the least-cost option, an existing plant may continue to generate electricity while that plant is constructed as long as it meets environmental regulations (constraints).

46 This example is not based on modeling results.

47 LGE and KU Energy LLC, The 2011 Joint Integrated Resource Plan of Louisville Gas and Electric Company and Kentucky Utilities Company Case No. 2011-00140.

48 Historical coal receipts from Energy Information Administration 923 database. Monthly Utility and Nonutility Fuel Receipts and Fuel Quality Data. Retrieved from: http://www.eia.gov/cneaf/electricity/page/eia906_920.html 49 LGE and KU Energy LLC, 2011 Air Compliance Plan Generation Planning & Analysis.

15

Interior coal.50 All variable operating costs are redacted in the publically available version of the IRP (black cells in Table 4).

Table 4. Mill Creek operations forecasts from the LGE and KU 2011 IRP Winter Summer 2010 2015 2020 2025 Capacity factor (%) 75.7 87.5 78.4 91 Mill Creek 1 Availability factor (%) 84.3 91.5 78.9 91.5 303 303 Average heat rate (Btu/kWh) 10683 10313 10331 10339 Cost of fuel ($/MBTU) 1.87 Generation MWh/yr 2,009,290 2,322,495 2,080,956 2,415,395 Capacity factor (%) 79.7 79 90.4 85.2 Mill Creek 2 Availability factor (%) 88.7 86.1 91.5 86.1 299 301 Average heat rate (Btu/kWh) 10845 10456 10516 10518 Cost of fuel ($/MBTU) 1.87 Generation MWh/yr 2,094,516 2,076,120 2,375,712 2,239,056 Capacity factor (%) 85.1 69.5 90.6 86 Mill Creek 3 Availability factor (%) 89.3 86.1 91.5 86.1 397 391 Average heat rate (Btu/kWh) 10738 10196 10218 10225 Cost of Fuel ($/MBTU) 1.87 Generation MWh/yr 2,937,175 2,398,751 3,127,005 2,968,238 Capacity factor (%) 80.1 76.1 85.8 91.5 Mill Creek 4 Availability factor (%) 83.2 91.5 86.1 91.5 492 477 Average heat rate (Btu/kWh) 10520 10276 10328 10334 Cost of fuel ($/MBTU) 1.89 Generation MWh/yr 3,399,620 3,229,851 3,641,541 3,883,461 Sum Mill Creek Generation MWh/yr 10,440,602 10,027,217 11,225,213 11,506,151 Baseline operating costs, without environmental controls, are based on the NERC report, which assumes that coal plants with a generation capacity of greater than 300 MWs have fixed and variable operating costs of $18/kW per year and $3.8/MWh, respectively.51 For compliance with the CSAPR and the MATS rule, we adopted LGEs assumptions about required retrofits and their capital costs. The variable operating cost of complying with the CSAPR and the MATS rule are based on the cost estimates developed by the firm Sargent & Lundy for EPAs modeling of the forthcoming MATS rule with the Integrated Planning Model.52 Sargent & Lundy created Excel templates to estimate the capital costs and incremental fixed and variable operating costs of wet flue gas desulfurization, baghouse particulate control, activated carbon injection, selective catalytic reduction, and other environmental controls for coal 50 EIA, Assumptions to the Annual Energy Outlook 2011.

51 NERC, 2010 Special Reliability Scenario Assessment: Resource Adequacy Impacts of Potential U.S. Environmental Regulations.

52 See the appendix for capital cost estimates from Sargent & Lundy and for a comparison of those estimates with LGEs retrofit cost assumptions.

16

plants on the basis of unit capacity, heat rate, coal type, and other plant characteristics.53 Mill Creeks fixed and variable operating costs prior to the retrofit are assumed to be the same as the existing plants with the retrofits to comply with the CSAPR and the MATS rule, because the environmental controls of the existing and the retrofitted plant are similar. We assume that annual generation with LGEs proposed retrofits equals the average annual generation of the existing plant from 2016 to 2025 (11,327,775 MWh/year). We also assumes that the heat rate remains unchanged at 10,304 Btu/kWh, because the IRP indicates that unit heat rates are nearly identical before and after the proposed retrofits.

CCR and Cooling Water Rule 316b compliance costs are based on the assumptions outlined in NI-NEMS Modeling of Cooling Water Rule 316b and CCR. The EPAs RIA of the proposed CCR rule estimates that the Mill Creek plant generates 1,236,800 tons of coal combustion wastes per year;54 64,700 tons/year are disposed in a company-owned surface impoundment (ash pond), and 191,300 tons/year are disposed in a company-owned landfill.55 According to EIAs 767 database, only one of Mill Creeks cooling units, cooling Unit 1 with a design intake flow rate of 216 MGD, uses once-through cooling and is affected by the proposed CCR rule.56 The remaining cooling units (2, 3, and 4) are recirculating units (cooling towers) that are assumed to be in compliance with the proposed rule. CCS retrofit costs are based on the National Energy Technology Laboratorys estimate of the cost of a CCS retrofit of American Electric Powers Conesville Unit 5.57 The capital and incremental variable costs were adjusted to 2010 dollars using the Chemical Engineering Plant Cost Index.58 Per KU and LGEs IRP assumptions, fixed and variable O&M escalation rates are set at 2 percent per year.59 Table 5 presents configurations of the existing plant and existing plant retrofits. Retrofit investment options encompass the different investment pathways to the retrofit configurations. For example, LGE could retrofit the Mill Creek plant once to achieve compliance with the CSAPR, MATS, and Subtitle D CCR, or it could retrofit the plant to comply with CSAPR and MATS and later retrofit it to comply with Subtitle D CCR. Table 7 lists all of the retrofit investment options included in the RBDM.

53 U.S. EPA. IPM Analysis of the Proposed Toxics Rule Documentation, http://www.epa.gov/airmarkets/progsregs/epa-ipm/toxics.html.

54 U.S. EPA, Regulatory Impact Analysis for EPAs Proposed RCRA Regulation Of Coal Combustion Residuals (CCR)

Generated by Electric Utilities Industry.

55 The remaining 980,800 tons per year are recycled or disposed offsite.

56 Data are from the EIA 767 database, http://www.eia.gov/cneaf/electricity/page/eia767/.

57 NETL (National Energy Technology Laboratory), Carbon Dioxide Capture from Existing Coal-Fired Power Plants, DOE/NETL-401/110907 (Washington, DC: Department of Energy, 2007).

58 CEPCI (Chemical Engineering Plant Cost Index), Chemical Engineering (July 2011): 64. .

59 LGE and KU Energy LLC, The 2011 Joint Integrated Resource Plan of Louisville Gas and Electric Company and Kentucky Utilities Company Case No. 2011-00140.

17

Table 5. Existing plant and existing plant retrofit configurations Mill Creek Mill Creek compliance compliance Mill Creek with CSAPR, Mill Creek with CSAPR, compliance Mill Creek MATS, Mill Creek compliance MATS, Existing and with CSAPR, compliance Subtitle C Mill Creek compliance with CSAPR, Subtitle C Existing MATS, with CCR, retrofitted compliance with CSAPR, MATS, CCR, Mill Creek Subtitle D CSAPR, Entrainment plant with CSAPR, MATS, Subtitle D Entrainment Plant CCR, MATS, 316b if > 125 MATS Subtitle D CCR, 316b if > 125 configurations Impingemen Subtitle C MGD and CCR Impingemen MGD and t 316b, CCS CCR Impingemen t 316b Impingemen retrofit t 316b if t 316b if <

<125 MDG, 125 MDG CCS retrofit Gross capacity 1,471.5 1,471.5 1,471.5 1,471.5 1,471.5 1,471.5 1,471.5 1,471.5 (MW)

Net capacity 1,471.5 1,471.5 1,471.5 1,471.5 1,030.1 1,471.5 1,467.0 1,027.0 (MW)

Net heat rate 10,304 10,304 10,304 10,304 14,721 10,304 10,336 14,766 (MMBtu/kWh)

Annual 11.328 11.328 11.328 11.328 7.929 11.328 11.293 7.905 generation (TWh/yr)

Retrofit cost

- 1,268 1,268 1,270 2,991 1,268 1,339 3,059 (M$)

Annual fixed 41 41 46 46 55 46 46 55 O&M (M$)

Annual variable 78 78 79 79 150 124 125 196 O&M (M$)*

Annual emissions CO2 11,853,037 11,853,037 11,853,037 11,853,037 1,185,304 11,853,037 11,853,487 1,185,349 (tons/yr)

Annual 30,348 11,672 11,672 11,672 1,167 11,672 11,673 1,167 emissions SO2 (tons/yr)

Annual 9,338 9,338 9,338 9,338 3,502 9,338 9,338 3,502 emissions NOx (tons/yr)

Retrofit costs include financing. *Excluding electricity and fuel costs.

New Generation Investment Options New generation investment options are taken from R.W. Becks Review of Power Plant Cost and Performance Assumptions for NEMS.60 All overnight capital costs reflect the location capital-cost variability adjustments given in the Beck study, and new generation with combustion turbines reflects the studys adjustment factors for regional ambient conditions. Any new coal plants are assumed to be built to the same capacity as that of the existing plant. Natural gas turbine options have predetermined capacities.

Capacity factors for each option, other than advanced combined cycle H class combustion turbines with CCS, are set to generate the same level of energy as that of the existing plant. The advanced combined cycle H class combustion turbines with CCS are assumed to have a maximum availability factor of 90 percent, limiting annual generation. The supercritical coal plant with CCS in the Beck study has an 60 EIA, Updated Capital Cost Estimates for Electricity Generating Plants, http://www.eia.gov/oiaf/beck_plantcosts/pdf/updatedplantcosts.pdf.

18

increased capacity boiler to supply steam to the CCS system and maintain net generation output.61 To include an option to retrofit a previously constructed supercritical plant in the RBDM, we adopted the CCS retrofit assumptions from NETLs Carbon Dioxide Capture from Existing Coal-Fired Power Plants.

Becks capital costs do not include financing costs.62 Financing costs are based on the fixed-charge rates assumptions in the LGE and KU IRP, assuming 30-year financing.63 Lump sum capital costs were calculated using LGEs 6.71 percent discount rate assumption.64 All new generation investment options are assumed to be in compliance with all forthcoming EPA regulations. Again, these options are assumed to become operational in three years. Table 6 lists the new generation investment configurations included in the model. Future versions of the RBDM will allow construction times to vary by investment option.

Table 6. Comparison of LGE retrofit cost estimates and Sargent & Lundy cost estimates Two combined cycle 2x1 F- Four Four CCS retrofit Three class CT + advanced advanced New generation Supercritical of Supercritical combined one combined combined configurations coal supercritical coal w/ CCS cycle 2x1 F-advanced cycle H class cycle H class coal class CT combined CT CT w/ CCS cycle H class CT Gross capacity (MW) 1472 1472 1472 1617 1477 1596 1596 Net capacity (MW) 1472 1,030 1472 1617 1477 1596 1,356 Net heat rate 8,800 12,572 12,000 7,050 6,883 6,430 7,568 (MMBtu/kWh)

Annual generation 11,328 7,929 11,328 11,328 11,328 11,327 10,691 (TWh/yr)

Capital cost (M$) 4,532 6,252 7,273 1,743 1,600 1,752 3,024 Annual fixed O&M (M$) 43.67 53.01 93.05 23.27 21.34 23.23 41.02 Annual variable O&M 48.12 118.88 102.46 38.85 37.88 35.23 68.96 (M$)*

Annual emissions CO2 10,122,954 1,012,295 1,380,403 4,671,840 4,560,831 4,260,877 485,447 (tons/yr)

Annual emissions SO2 4,984 997 1,359 40 39 36 40 (tons/yr)

Annual emissions NOx 2,991 2,991 4,078 299 292 273 303 (tons/yr)

Capital costs include financing. *Excluding electricity and fuel costs. CT = combustion turbine.

Table 7 lists all the investment options included in the RBDM. The investment options allow the model to select among various retrofit pathways to comply with the forthcoming EPA regulations and CCS retrofits of new generation.65 Many of the investment options require prior investment in other options.

For example, the model cannot invest in option 11 before investing in option 10. See the appendix for all investment prerequisites.

61 Ibid.

62 NETL, Carbon Dioxide Capture from Existing Coal-Fired Power Plants.

63 The rates are 9.00% for coal investments and 9.01% for NGCC investments; LGE and KU Energy LLC, 2011 Joint Integrated Resource Plan.

64 Ibid.

65 CCS retrofit costs for advanced NGCC combustion turbine plants (investment option 7) were estimated by subtracting the cost of advanced natural gas combustion turbines (investment 6) from the cost of advanced natural gas turbines with CCS (investment

8) and then increased by 20 percent to account for the extra cost of retrofits relative to new plant installations.

19

Table 7. All investment options available in RBDM for Mill Creek plant Investment options Capital costs Investment options Capital costs

($2010 M) ($2010 M) 1 Supercritical coal 4,532 17 CCS retrofit on 13 1,720 2 CCS retrofit on super 1,720 18 CCS retrofit on 14 1,720 3 Super with CCS 7,273 19 CCS retrofit on 15 1,720 4 3 NGCC 2x1 F-Class CT 1,743 20 CCS retrofit on 16 1,720 5 2 NGCC 2x1 F-Class CT + 1 1,600 21 MATS + Sub D + 316b I + 2,991 advanced NGCC H-Class CT CCS retrofit on Existing 6 4 advanced NGCC H-Class 1,752 22 Sub C retrofit on 10 -

CT 7 CCS retrofit on 4 advanced 1,526 23 MATS + Sub C retrofit on 1,268 NGCC H-Class CT Existing 8 4 Advanced NGCC H-Class 3,024 24 Sub C + 316b E retrofit on 71 CT with CCS 10 9 Existing Mill Creek - 25 316b E retrofit on 22 71 10 MATS retrofit on Existing 1,268 26 316b E retrofit on 23 71 11 Sub D retrofit on 10 - 27 MATS + Sub C + 316b E 1,339 retrofit on Existing 12 MATS + Sub D retrofit on 1,268 28 CCS retrofit on 24 1,720 Existing 13 Sub D + 316b I retrofit on 2 29 CCS retrofit on 25 1,720 10 14 316b I retrofit on 11 2 30 CCS retrofit on 26 1,720 15 316b I retrofit on 12 2 31 CCS retrofit on 27 1,720 16 MATS + Sub D + 316b I 1,270 32 MATS + Sub C + 316b E + 3,059 retrofit on Existing CCS retrofit on Existing CCS Premium The capital costs of CCS are doubled at the start of the model in 2011 to account for potential difficulties in installing and operating this new technology. This capital cost multiplier linearly decreases to 1 in 2020.

Capital Cost Escalation Rates Electricity sector generation capital costs tend to change over time due to factors such as technological learning. The capital cost escalation rates used in the RBDM are based on NI-NEMS outputs for projected capital costs for conventional coal, IGCC with CCS, conventional NGCC, advanced NGCC, and advanced NGCC with CCS. The NI-NEMS capital cost projections were converted to annual percent changes and applied to the capital costs in Table 7. Capital costs from 2035 to 2049 are assumed to be constant.66 SO2 and NOx Emission Allowance Costs In all scenarios, the prices of SO2 and NOx emission allowances are based on the allowance prices in the LGE and KU IRP. According to the IRP, these prices decrease to $0/ton by 2014.67 66 NI-NEMS projections are from 2011 to 2035.

67 LGE and KU Energy LLC, 2011 Joint Integrated Resource Plan.

20

Coal Price Differential To better approximate forecasts of coal prices for the Mill Creek plant site, we compared the historical costs of coal delivered to Mill Creek68 with the historical costs for high-sulfur bituminous East Interior coal, a NI-NEMS output (see table A3 in the appendix). On the basis of this differential, we added

$0.16/MMBtu to price forecasts for high-sulfur bituminous East Interior coal from NI-NEMS.

Natural Gas Price Differential Similar to coal, we obtained historical prices for natural gas delivered to significant natural gas-consuming plants near the Mill Creek plant69 and compared them with historical natural gas prices available from NI-NEMS. These historical price differentials vary significantly from year to year and from plant to plant (see table A4 in appendix). We used the delivered East South Central Delivered Electricity Sector natural gas price forecasts plus a $0.3104/MMBtu cost adder that LGE includes for a firm gas charge (guaranteed supply).70 Modeling Results Summary In every scenario, for the modeling runs with and without future carbon prices, the RBDM determined it is least-cost to retrofit the Mill Creek plant to comply with forthcoming EPA regulations given our assumptions about uncertainty. The initial investment decisions by the RBDM are nearly identical for all scenarios for both modeling runs (tables 8 and 9). In all scenarios, the RBDM retrofits the existing Mill Creek plant to comply with the MATS, as proposed by LGE, in 2013 and makes additional retrofits to comply with the forthcoming CCR and Cooling Water Rule 316b prior to 2019. This indicates that LGEs proposed retrofit of the Mill Creek plant to comply with the CSAPR and MATS is robust under our representation of natural gas price, regulatory, and carbon policy uncertainty.

The timing and order of CCR and Cooling Water Rule 316b retrofit investments differ by scenario and whether the modeling run includes carbon risk. Therefore, we suggest that the RBDM be rerun to determine the optimal CCR and Cooling Water Rule 316b retrofit investment as additional information about these rules becomes available.

68 Historical coal receipts are taken from the EIA 923 database.

69 Historical natural gas receipts are taken from the EIA 923 database.

70 LGE and KU Energy LLC, 2011 Joint Integrated Resource Plan.

21

Tables 8 and 9. Least-cost investment decisions for 9 scenarios without carbon prices and 12 scenarios with carbon prices in selected scenarios for initial 10 years of modeling period. Results for years 2021-2049 are not shown.

Least Cost Investments 9 Scenarios without Carbon Price Scenarios Year 1 2 3 4 5 6 7 8 9 10 1 Mid NG, Standard EPA 0 0 MATS Sub C Sub D 0 316b E 316b I 0 0 Mid NG, Less Stringent 2 0 0 MATS Sub C Sub D 0 316b E 316b I 0 0 EPA 3 w/o Mid NG, More Stringent Ca rbon EPA MATS Sub C Sub D 316b E Pri ce 0 0 0 0 0 0 5 High NG, Standard EPA 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 High NG, Less Stringent 6 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 EPA 7 w/o High NG, More Stringent Ca rbon EPA MATS Sub C Sub D 316b E Pri ce 0 0 0 0 0 0 Extra High NG, Standard 9 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 EPA Extra High NG, Less 10 0 0 MATS 0 Sub C Sub D 316b E 316b I 0 0 Stringent EPA 11 w/o Extra High NG, More Str.