&EPA
United States Air and
Environmental Protection Radiation November 2018
Agency (6204J)
Documentation for
EPA's Power Sector Modeling
Platform v6
Using the Integrated Planning
Model
WECC.PNW
SPP_WAUE
WECC_WY
Ml DA
PIM_WMAC
PJM_COMD
PJM_EMAC
WFCC_SNV
WECC_SCE
WECC_AZ
MIS_AMSO
-------
Cover: EPA's Power Sector Modeling Platform v6 is used by the U.S. Environmental Protection Agency
as a platform to conduct various scenario and sensitivity analysis on the key drivers of the power sector
behavior and to project the impact of emissions policies on the electric power sector in the 48 contiguous
states and the District of Columbia in the lower continental U.S. Representation of the electric power
sector in Canada is also included for purposes of integrated projections. The map appearing on the cover
shows the 67 model regions used to characterize the operation of the U.S. electric power system in the
lower continental U.S. and 11 model regions in Canada. EPA's Power Sector Modeling Platform v6 using
the Integrated Planning Model (IPM®) was developed by EPA's Clean Air Markets Division with technical
support from ICF, Inc. The IPM is a product of ICF, Inc. and is used in support of its public and private
sector clients. IPM® is a registered trademark of ICF Resources, L.L.C.
-------
Documentation for
EPA's Power Sector Modeling Platform v6
Using the Integrated Planning Model
U.S. Environmental Protection Agency
Clean Air Markets Division
1200 Pennsylvania Avenue, NW (6204J)
Washington, D.C. 20460
(www.epa.gov/airmarkets)
November 2018
-------
Acknowledgment
This document was prepared by U.S. EPA's Clean Air Markets Division, Office of Air and Radiation. ICF
Incorporated, an operating company of ICF, provided technical support under EPA Contract EP-W-13-
009 and EPA Contract 68HE0C18D0001.
i
-------
TABLE OF CONTENTS
Acknowledgment i
1. Introduction 1-1
1.1 Executive Summary 1-1
1.2 Review and Ongoing Improvement of the Integrated Planning Model 1-6
2. Modeling Framework 2-1
2.1 IPM Overview 2-1
2.1.1 Purpose and Capabilities 2-1
2.1.2 Applications 2-2
2.2 Model Structure and Formulation 2-2
2.2.1 Objective Function 2-3
2.2.2 Decision Variables 2-3
2.2.3 Constraints 2-4
2.3 Key Methodological Features of IPM 2-4
2.3.1 Model Plants 2-5
2.3.2 Model Run Years 2-6
2.3.3 Cost Accounting 2-6
2.3.4 Modeling Wholesale Electricity Markets 2-7
2.3.5 Load Duration Curves (LDC) 2-7
2.3.6 Dispatch Modeling 2-9
2.3.7 Fuel Modeling 2-10
2.3.8 Transmission Modeling 2-10
2.3.9 Perfect Competition and Perfect Foresight 2-11
2.3.10 Scenario Analysis and Regulatory Modeling 2-11
2.4 Hardware and Programming Features 2-11
2.5 Model Inputs and Outputs 2-12
2.5.1 Data Parameters for Model Inputs 2-12
2.5.2 Model Outputs 2-13
3. Power System Operation Assumptions 3-1
3.1 Model Regions 3-1
3.2 Electric Load Modeling 3-2
3.2.1 Demand Elasticity 3-7
3.2.2 Net Internal Demand (Peak Demand) 3-7
3.2.3 Regional Load Shapes 3-8
3.3 Transmission 3-8
3.3.1 Inter-regional Transmission Capability 3-8
3.3.2 Joint Transmission Capacity and Energy Limits 3-9
3.3.3 Transmission Link Wheeling Charge 3-10
3.3.4 Transmission Losses 3-10
3.4 International Imports 3-10
3.5 Capacity, Generation, and Dispatch 3-11
3.5.1 Availability 3-11
3.5.2 Capacity Factor 3-12
3.5.3 Turndown 3-14
-------
3.6 Reserve Margins 3-14
3.7 Power Plant Lifetimes 3-15
3.8 Heat Rates 3-16
3.9 Existing Environmental Regulations 3-17
3.9.1 SO2 Regulations 3-17
3.9.2 NOx Regulations 3-18
3.9.3 Multi-Pollutant Environmental Regulations 3-22
3.9.4 CO2 Regulations 3-24
3.9.5 Non-Air Regulations Impacting EGUs 3-25
3.9.6 State-Specific Environmental Regulations 3-27
3.9.7 New Source Review (NSR) Settlements 3-27
3.9.8 Emission Assumptions for Potential (New) Units 3-28
3.9.9 Energy Efficiency and Renewable Portfolio Standards 3-28
3.9.10 Canada CO2 and Renewable Regulations 3-28
3.10 Emissions Trading and Banking 3-29
3.10.1 Intertemporal Allowance Price Calculation 3-29
4. Generating Resources 4-1
4.1 National Electric Energy Data System (NEEDS) 4-1
4.2 Existing Units 4-1
4.2.1 Population of Existing Units 4-1
4.2.2 Capacity 4-4
4.2.3 Plant Location 4-5
4.2.4 Online Year 4-5
4.2.5 Unit Configuration 4-5
4.2.6 Model Plant Aggregation 4-6
4.2.7 Cost and Performance Characteristics of Existing Units 4-10
4.2.8 Life Extension Costs for Existing Units 4-17
4.3 Planned-Committed Units 4-18
4.3.1 Population and Model Plant Aggregation 4-18
4.3.2 Capacity 4-20
4.3.3 State and Model Region 4-20
4.3.4 Online and Retirement Year 4-20
4.3.5 Unit Configuration, Cost, and Performance 4-20
4.4 Potential Units 4-20
4.4.1 Methodology Used to Derive the Cost and Performance Characteristics of
Conventional Potential Units 4-21
4.4.2 Cost and Performance for Potential Conventional Units 4-21
4.4.3 Short-Term Capital Cost Adder 4-21
4.4.4 Regional Cost Adjustment 4-22
4.4.5 Cost and Performance for Potential Renewable Generating and Non-Conventional
Technologies 4-30
4.5 Nuclear Units 4-47
4.5.1 Existing Nuclear Units 4-47
4.5.2 Potential Nuclear Units 4-48
5. Emission Control Technologies 5-1
5.1 Sulfur Dioxide Control Technologies - Scrubbers 5-1
5.1.1 Methodology for Obtaining SO2 Controls Costs 5-2
iii
-------
5.2 Nitrogen Oxides Control Technology 5-5
5.2.1 Combustion Controls 5-5
5.2.2 Post-combustion NOx Controls 5-5
5.2.3 Methodology for Obtaining SCR Costs for Coal 5-6
5.2.4 Methodology for Obtaining SCR Costs for Oil/Gas Steam Units 5-8
5.2.5 Methodology for Obtaining SNCR Costs for Coal 5-8
5.2.6 SO2 Controls for Units with Capacities from 25 MW to 100 MW (25 MW < capacity <
100 MW) 5-8
5.3 Biomass Co-firing 5-8
5.4 Mercury Control Technologies 5-9
5.4.1 Mercury Content of Fuels 5-9
5.4.2 Mercury Emission Modification Factors 5-10
5.4.3 Mercury Control Capabilities 5-13
5.4.4 Methodology for Obtaining ACI Control Costs 5-17
5.5 Hydrogen Chloride (HCI) Control Technologies 5-17
5.5.1 Chlorine Content of Fuels 5-17
5.5.2 HCI Removal Rate Assumptions for Existing and Potential Units 5-19
5.5.3 HCI Retrofit Emission Control Options 5-19
5.6 Fabric Filter (Baghouse) Cost Development 5-21
5.7 Coal-to-Gas Conversions 5-23
5.7.1 Boiler Modifications for Coal-To-Gas Conversions 5-23
5.7.2 Natural Gas Pipeline Requirements for Coal-To-Gas Conversions 5-24
5.8 Retrofit Assignments 5-24
6. CO2 Capture, Storage, and Transport 6-1
6.1 CO2 Capture 6-1
6.1.1 CO2 Capture for Potential EGUs 6-1
6.1.2 CO2 Capture via Retrofitting Existing EGUs 6-2
6.2 CO2 Storage 6-3
6.3 CO2 Transport 6-8
7. Coal 7-1
7.1 Coal Market Representation in EPA Platform v6 7-1
7.1.1 Coal Supply Regions 7-2
7.1.2 Coal Demand Regions 7-3
7.1.3 Coal Quality Characteristics 7-4
7.1.4 Coal Emission Factors 7-5
7.1.5 Coal Grade Assignments 7-11
7.2 Coal Supply Curves 7-11
7.2.1 Nature of Supply Curves Developed for EPA Platform v6 7-11
7.2.2 Cost Components in the Supply Curves 7-12
7.2.3 Procedures Employed in Determining Mining Costs 7-13
7.2.4 Procedure Used In Determining Mine Productivity 7-14
7.2.5 Procedure to Determine Total Recoverable Reserves by Region and Type 7-14
7.2.6 New Mine Assumptions 7-15
7.2.7 Other Notable Procedures 7-15
7.2.8 Supply Curve Development 7-17
7.2.9 EPA Platform v6 Assumptions and Outlooks for Major Supply Basins 7-19
7.3 Coal Transportation 7-20
iv
-------
7.3.1 Coal Transportation Matrix Overview 7-21
7.3.2 Calculation of Coal Transportation Distances 7-21
7.3.3 Overview of Rail Rates 7-22
7.3.4 Truck Rates 7-25
7.3.5 Barge and Lake Vessel Rates 7-26
7.3.6 Transportation Rates for Imported Coal 7-26
7.3.7 Other Transportation Costs 7-26
7.3.8 Long-Term Escalation of Transportation Rates 7-27
7.3.9 Market Drivers Moving Forward 7-29
7.3.10 Other Considerations 7-31
7.4 Coal Exports, Imports, and Non-Electric Sectors Demand 7-32
8. Development of Natural Gas Supply Curves for EPA Platform v6 8-1
8.1 Introduction 8-1
8.2 Brief Synopsis of GMM 8-3
8.3 Resources Data and Reservoir Description 8-6
8.3.1 U.S. Resources and Reserves 8-6
8.3.2 Upstream Cost and Technology Factors 8-8
8.3.3 Historical Gas Production 8-9
8.3.4 Treatment of Frontier Resources and Exports 8-10
8.4 Oil Prices 8-12
8.5 Demand Assumptions 8-13
8.6 Discussion of GMM Results Underlying the Natural Gas Supply Curves 8-17
8.6.1 Supply Curves for EPA Platform v6 8-20
8.6.2 Basis 8-21
8.6.3 Delivered Price Adders 8-21
9. Other Fuels and Fuel Emission Factor Assumptions 9-1
9.1 Fuel Oil 9-1
9.2 Biomass Fuel 9-2
9.3 Nuclear Fuel 9-3
9.4 Waste Fuels 9-3
9.5 Fuel Emission Factors 9-4
10. Financial Assumptions 10-1
10.1 Introduction and Summary 10-1
10.2 Introduction to Risk 10-1
10.2.1 Deregulation - Market Structure Risks 10-2
10.3 Federal Income Tax Law Changes 10-3
10.4 Calculation of the Financial Discount Rate 10-5
10.4.1 Introduction to Discount Rate Calculations 10-5
10.4.2 Summary of Results 10-5
10.5 Discount Rate Components 10-7
10.6 Market Structure: Utility-Merchant Financing Ratio 10-7
10.7 Capital Structure: Debt-Equity Share 10-8
10.7.1 Introduction and Shares for Utilities and IPPs 10-8
10.7.2 Utility and Merchant 10-8
v
-------
10.7.3 Merchant by Technology 10-8
10.8 Cost of Debt 10-9
10.8.1 Merchant Cost of Debt 10-10
10.8.2 Utility Cost of Debt 10-10
10.9 Return on Equity (ROE) 10-10
10.9.1 Introduction and Beta 10-10
10.9.2 Risk-Free Rate and Equity Risk Premium 10-11
10.9.3 Beta 10-12
10.9.4 Equity Size Premium 10-12
10.9.5 Nominal ROEs 10-12
10.9.6 WACC/Discount Rate 10-13
10.10 Calculation of Capital Charge Rate 10-13
10.10.1 Introduction to Capital Charge Rate Calculations 10-13
10.10.2 Capital Charge Rate Components 10-15
vi
-------
LIST OF TABLES
Table 1-1 Key Updates in the EPA Platform v6 November 2018 Reference Case 1-2
Table 1-2 Plant Types in EPA Platform v6 1-4
Table 1-3 Emission Control Technologies in EPA Platform v6 1-5
Table 2-1 Run Year and Analysis Year Mapping Used in EPA Platform v6 2-6
Table 2-2 Load Duration Curves used in EPA Platform v6 2-13
Table 3-1 Mapping of NERC Regions and NEMS Regions with EPA Platform v6 Model Regions 3-3
Table 3-2 Electric Load Assumptions in EPA Platform v6 3-5
Table 3-3 Regional Electric Load Assumptions in EPA Platform v6 3-5
Table 3-4 National Non-Coincidental Net Internal Demand 3-7
Table 3-5 Annual Joint Capacity and Energy Limits to Transmission Capabilities between Model Regions
in EPA Platform v6 3-9
Table 3-6 International Electricity Imports (billions kWh) in EPA Platform v6 3-10
Table 3-7 Availability Assumptions in EPA Platform v6 3-11
Table 3-8 Seasonal Hydro Capacity Factors (%) in EPA Platform v6 3-12
Table 3-9 Planning Reserve Margins in EPA Platform v6 3-15
Table 3-10 Lower and Upper Limits Applied to Heat Rate Data in EPA Platform v6 3-17
Table 3-11 State-of-the-Art Combustion Control Configurations by Boiler Type 3-21
Table 3-12 CSAPR Update State Budgets, Variability Limits, and Assurance Levels for Ozone-Season
NOx(Tons) 3-23
Table 3-13 NY Minimum Oil Burn Rule Plant Level Oil Capacity Factor Requirements 3-27
Table 3-14 Canada Renewable Electricity Requirements (%) in EPA Platform v6 3-28
Table 3-15 Trading and Banking Rules in EPA Platform v6 - Part 1 3-29
Table 3-16 CASPR Trading and Banking Rules in EPA Platform v6 - Part 2 3-30
Table 3-17 Emission and Removal Rate Assumptions for Potential (New) Units in EPA Platform v6....3-31
Table 3-18 Recalculated NOx Emission Rates for SCR Equipped Units Sharing Common Stacks with
Non-SCR Units 3-32
Table 3-19 Renewable Portfolio Standards in EPA Platform v6 3-33
Table 3-20 Regional Net Internal Demand in EPA Platform v6 3-34
Table 3-21 Annual Transmission Capabilities of U.S. Model Regions in EPA Platform v6 - 2021 3-34
Table 3-22 Turndown Assumptions for Coal Steam Units in EPA Platform v6 3-34
Table 3-23 State Power Sector Regulations included in EPA Platform v6 3-34
Table 3-24 New Source Review (NSR) Settlements in EPA Platform v6 3-34
Table 3-25 State Settlements in EPA Platform v6 3-34
Table 3-26 Citizen Settlements in EPA Platform v6 3-34
Table 3-27 Complete Availability Assumptions in EPA Platform v6 3-34
Table 3-28 BART Regulations included in EPA Platform v6 3-34
Table 4-1 Data Sources for NEEDS v6 for EPA Platform v6 4-2
Table 4-2 Rules Used in Populating NEEDS v6 for EPA Platform v6 4-2
Table 4-3 Summary Population (through 2017) of Existing Units in NEEDS v6 4-3
Table 4-4 Hierarchy of Data Sources for Capacity in NEEDS v6 4-4
Table 4-5 Capacity-Parsing Algorithm for Steam Units in NEEDS v6 4-5
Table 4-6 Data Sources for Unit Configuration in NEEDS v6 4-6
Table 4-7 Aggregation Profile of Model Plants as Provided at Set up of EPA Platform v6 4-7
Table 4-8 VOM Assumptions in EPA Platform v6 4-11
Table 4-9 FOM Assumptions in EPA Platform v6 4-13
Table 4-10 Life Extension Cost Assumptions Used in EPA Platform v6 4-18
vii
-------
Table 4-11 Summary of Planned-Committed Units in NEEDS v6 for EPA Platform v6 4-18
Table 4-12 Planned-Committed Units by Model Region in NEEDS v6 for EPA Platform v6 4-19
Table 4-13 Performance and Unit Cost Assumptions for Potential (New) Capacity from Conventional
Technologies in EPA Platform v6 4-23
Table 4-14 Short-Term Capital Cost Adders for New Power Plants in EPA Platform v6 (2016$) 4-25
Table 4-15 Regional Cost Adjustment Factors for Conventional and Renewable Generating Technologies
in EPA Platform v6 4-26
Table 4-16 Performance and Unit Cost Assumptions for Potential (New) Renewable and Non-
Conventional Technology Capacity in EPA Platform v6 4-28
Table 4-17 Offshore Shallow Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class in
EPA Platform v6 4-30
Table 4-18 Offshore Mid-Depth Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class in
EPA Platform v6 4-32
Table 4-19 Offshore Deep Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class in EPA
Platform v6 4-33
Table 4-20 Onshore Average Capacity Factor by Wind TRG 4-35
Table 4-21 Onshore Reserve Margin Contribution by Wind TRG 4-35
Table 4-22 Offshore Shallow Average Capacity Factor by Wind TRG 4-35
Table 4-23 Offshore Shallow Reserve Margin Contribution by Wind TRG 4-35
Table 4-24 Offshore Mid Depth Average Capacity Factor by Wind TRG 4-35
Table 4-25 Offshore Mid Depth Reserve Margin Contribution by Wind TRG 4-36
Table 4-26 Offshore Deep Average Capacity Factor by Wind TRG 4-36
Table 4-27 Offshore Deep Reserve Margin Contribution by Wind TRG 4-36
Table 4-28 Capital Cost Adder (2016$/kW) for New Offshore Shallow Wind Plants in EPA Platform v64-36
Table 4-29 Capital Cost Adder (2016$/kW) for New Offshore Mid Depth Wind Plants in EPA Platform v64-
38
Table 4-30 Capital Cost Adder (2016$/kW) for New Offshore Deep Wind Plants in EPA Platform v6 ...4-39
Table 4-31 Example Calculations of Wind Generation Potential, Reserve Margin Contribution, and Capital
Cost for Onshore Wind in WECC_CO at Wind Class 3, Cost Class 1 4-40
Table 4-32 Solar Photovoltaic Reserve Margin Contribution by Resource Class 4-41
Table 4-33 Regional Assumptions on Potential Geothermal Electric Capacity 4-42
Table 4-34 Potential Geothermal Capacity and Cost Characteristics by Model Region 4-42
Table 4-35 Performance and Unit Cost Assumptions for Potential (New) Battery Storage 4-45
Table 4-36 Energy Storage Mandates in the November 2018 Reference Case 4-46
Table 4-37 Nuclear Uprates (MW) as Incorporated in EPA Platform v6 4-47
Table 4-38 Onshore Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class 4-49
Table 4-39 Wind Generation Profiles 4-49
Table 4-40 Capital Cost Adder (2016$/kW) for New Onshore Wind Plants 4-49
Table 4-41 Solar Photovoltaic Regional Potential Capacity (MW) by Resource and Cost Class 4-49
Table 4-42 Solar Thermal Regional Potential Capacity (MW) by Resource and Cost Class 4-49
Table 4-43 Hourly Solar Generation Profiles 4-49
Table 4-44 Capital Cost Adder (2016$/kW) for New Solar PV Plants 4-49
Table 4-45 Capital Cost Adder (2016$/kW) for New Solar Thermal Plants 4-49
Table 4-46 Solar Photovoltaic Average Capacity Factor by Resource class 4-49
Table 4-47 Solar Thermal Capacity Factor by Resource Class and Season 4-49
Table 4-48 Potential Electric Capacity from New Landfill Gas Units (MW) 4-49
Table 4-49 Characteristics of Existing Nuclear Units 4-49
Table 4-50 Generating Units from EIA Form 860 Not Included 4-49
Table 4-51 Generating Units Not Included Due to Recent Announcements 4-49
-------
Table 5-1 Summary of Emission Control Technology Retrofit Options in EPA Platform v6 5-1
Table 5-2 Summary of Retrofit SO2 Emission Control Performance Assumptions in EPA Platform v6 ....5-2
Table 5-3 Illustrative Scrubber Costs (2016$) for Representative Sizes and Heat Rates under the
Assumptions in EPA Platform v6 5-4
Table 5-4 Summary of Retrofit NOx Emission Control Performance Assumptions in EPA Platform v6 ....5-5
Table 5-5 Illustrative Post-combustion NOx Control Costs (2016$) for Coal Plants for Representative
Sizes and Heat Rates under the Assumptions in EPA Platform v6 5-7
Table 5-6 Post-Combustion NOx Controls for Oil/Gas Steam Units in EPA Platform v6 5-8
Table 5-7 Coal Units with Biomass Co-firing Option in EPA Platform v6 5-9
Table 5-8 Assumptions on Mercury Concentration in Non-Coal Fuel in EPA Platform v6 5-10
Table 5-9 Mercury Emission Modification Factors Used in EPA Platform v6 5-11
Table 5-10 Definition of Acronyms for Existing Controls 5-12
Table 5-11 Keyto BurnerType Designations in Table 5-9 5-13
Table 5-12 Assignment Scheme for Mercury Emissions Control Using Activated Carbon Injection (ACI) in
EPA Platform v6 5-15
Table 5-13 Illustrative Activated Carbon Injection (ACI) Costs (2016$) for Representative Sizes and Heat
Rates under the Assumptions in EPA Platform v6 5-18
Table 5-14 HCI Removal Rate Assumptions for Potential (New) and Existing Units in EPA Platform v6 ...5-
19
Table 5-15 Summary of Retrofit HCI (and SO2) Emission Control Performance Assumptions in EPA
Platform v6 5-20
Table 5-16 Illustrative Dry Sorbent Injection (DSI) Costs (2016$) for Representative Sizes and Heat Rates
under Assumptions in EPA Platform v6 5-22
Table 5-17 Illustrative Particulate Controls Costs (2016$) for Representative Sizes and Heat Rates under
the Assumptions in EPA Platform v6 5-22
Table 5-18 Cost and Performance Assumptions for Coal-to-Gas Retrofits in EPA Platform v6 5-23
Table 5-19 First Stage Retrofit Assignment Scheme in EPA Platform v6 5-25
Table 5-20 Second and Third Stage Retrofit Assignment Scheme in EPA Platform v6 5-26
Table 5-21 Cost of Building Pipelines to Coal Plants in EPA Platform v6 5-28
Table 6-1 Cost and Performance Assumptions for Potential USC and NGCC with and without Carbon
Capture 6-2
Table 6-2 Performance and Unit Cost Assumptions for Carbon Capture Retrofits on Coal Plants 6-3
Table 6-3 Lower-48 CO2 Sequestration Capacity by Region 6-7
Table 6-4 CO2 Storage Cost Curves in EPA Platform v6 6-9
Table 6-5 C02 Transportation Matrix in EPA Platform v6 6-9
Table 7-1 Coal Supply Regions in EPA Platform v6 7-2
Table 7-2 Coal Rank Heat Content Ranges 7-4
Table 7-3 Coal Grade SO2 Content Ranges 7-4
Table 7-4 Coal Quality Characteristics by Supply Region and Coal Grade 7-5
Table 7-5 Coal Clustering by Coal Grade - SO2 Emission Factors (Ibs/MMBtu) 7-8
Table 7-6 Coal Clustering by Coal Grade - Mercury Emission Factors (Ibs/TBtu) 7-8
Table 7-7 Coal Clustering by Coal Grade - Ash Emission Factors (Ibs/MMBtu) 7-9
Table 7-8 Coal Clustering by Coal Grade - HCI Emission Factors (Ibs/MMBtu) 7-9
Table 7-9 Coal Clustering by Coal Grade - CO2 Emission Factors (Ibs/MMBtu) 7-10
Table 7-10 Example of Coal Assignments Made in EPA Platform v6 7-11
Table 7-11 Basin-Level Groupings Used in Preparing v6 Coal Supply Curves 7-11
Table 7-12 Rail Competition Definitions 7-23
Table 7-13 Assumed Eastern Rail Rates for 2020 (2016 mills/ton-mile) 7-24
Table 7-14 Assumed Midwestern Rail Rates for 2020 (2016 mills/ton-mile) 7-24
ix
-------
Table 7-15 Assumed Non-PRB Western Rail Rates for 2020 (2016 mills/ton-mile) 7-25
Table 7-16 Assumed PRB Western Rail Rates for 2020 (2016 mills/ton-mile) 7-25
Table 7-17 Assumed Truck Rates for 2020 (2016 Real Dollars) 7-25
Table 7-18 Assumed Barge Rates for 2020 7-26
Table 7-19 Assumed Other Transportation Rates for 2020 (2016 Real Dollars) 7-27
Table 7-20 EIA AEO Diesel Fuel Forecast, 2020-2040 7-29
Table 7-21 Summary of Expected Escalation for Coal Transportation Rates, 2020-2050 7-31
Table 7-22 Coal Exports (Million Short Tons) 7-32
Table 7-23 Residential, Commercial, and Industrial Demand (Million Short Tons) 7-33
Table 7-24 Coal Import Limits (Million Short Tons) 7-33
Table 7-25 Coal Transportation Matrix in EPA Platform v6 7-35
Table 7-26 Coal Supply Curves in EPA Platform v6 7-35
Table 7-27 Coal Demand Regions in EPA Platform v6 7-35
Table 8-1 Supply/Demand Balance and Henry Hub Price for a GMM Run Underlying the Natural Gas
Supply Curves in EPA Platform v6 8-18
Table 8-2 Delivered Price Adders 8-22
Table 8-3 EIA Style Gas Report for EPA Platform v6 8-22
Table 8-4 Natural Gas Basis for EPA Platform v6 8-22
Table 8-5 Natural Gas Supply Curves for EPA Platform v6 8-22
Table 9-1 Fuel Oil Prices by NEMS Region in EPA Platform v6 9-1
Table 9-2 Waste Fuels in NEEDS v6 and EPA Platform v6 9-3
Table 9-3 Fuel Emission Factor Assumptions in EPA Platform v6 9-4
Table 9-4 Biomass Supply Curves in EPA Platform v6 9-4
Table 10-1 Summary Tax Changes 10-4
Table 10-2 Financial Assumptions for Utility and Merchant Cases 10-6
Table 10-3 Weighted Average Cost of Capital 10-7
Table 10-4 Share of Annual Thermal Capacity Additions by Market 10-8
Table 10-5 Capital Structure Assumptions 10-9
Table 10-6 Nominal Debt Rates 10-10
Table 10-7 Utilities Used to Calculate Cost of Debt 10-10
Table 10-8 Estimated Annual Levered Beta forS15ELUT Utility Index Based on Daily Returns 10-12
Table 10-9 Real Capital Charge Rate - Blended (%) 10-13
Table 10-10 Real Capital Charge Rate - IPP (%) 10-14
Table 10-11 Real Capital Charge Rate - Utility (%) 10-14
Table 10-12 Book Life, Debt Life and Depreciation Schedules for EPA Platform v6 10-15
x
-------
LIST OF FIGURES
Figure 1-1 Modeling and Data Structures in EPA Platform v6 1-6
Figure 2-1 Hypothetical Chronological Hourly Load Curve and Seasonal Load Duration Curve in EPA
Platform v6 for Summer 2-8
Figure 2-2 Stylized Depiction of a Six Segment Load Duration Curve Used in EPA Platform v6 2-9
Figure 2-3 Stylized Dispatch Order in Illustrative Load Segments 2-10
Figure 3-1 EPA Platform v6 Model Regions 3-2
Figure 3-2 Scheduled Retirements of Existing Nuclear Capacity Under 80-Year Life Assumption 3-16
Figure 3-3 Modeling Process for Obtaining Projected NOx Emission Rates 3-19
Figure 3-4 How One of the Four NOx Modes Is Ultimately Selected for a Unit 3-21
Figure 4-1 Derivation of Plant Fixed O&M Data 4-13
Figure 7-1 Map of the Coal Supply Regions in EPA Platform v6 7-3
Figure 7-2 Coal Mine Productivity (2000-2015) 7-16
Figure 7-3 Average Annual Cost Growth Assumptions by Region (2021-2050) 7-16
Figure 7-4 Maximum Annual Coal Production Capacity per Year (Million Short Tons) 7-17
Figure 7-5 Illustration of Preliminary Step in Developing a Cumulative Coal Supply Curve 7-17
Figure 7-6 Illustration of Final Step in Developing a Cumulative Coal Supply Curve 7-18
Figure 7-7 Example Coal Supply Curve in Stepped Format 7-18
Figure 7-8 Calculation of Multi-Mode Transportation Costs (Example) 7-21
Figure 7-9 Rail Cost Indices Performance (2Q2011-2Q2016) 7-28
Figure 7-10 Long-Run Marginal Cost Breakdown by Transportation Mode 7-29
Figure 8-1 GMM Gas Quantity and Price Response 8-1
Figure 8-2 IPM/GMM Interaction 8-2
Figure 8-3 Geographic Coverage of GMM 8-3
Figure 8-4 Example Pipeline Discount Curve 8-4
Figure 8-5 GMM Natural Gas Storage Regions 8-6
Figure 8-6 GMM U.S. and Canada Projected Gas Production by Source 8-9
Figure 8-7 Production Comparison for San Juan and Raton Basins 8-10
Figure 8-8 Existing and Proposed Marine LNG Terminals as of May 2017 8-11
Figure 8-9 LNG Export Volumes versus Capacity 8-11
Figure 8-10 U.S. Pipeline Exports to Mexico 8-12
Figure 8-11 Refiners'Acquisition Cost of Crude (RACC) 8-13
Figure 8-12 GMM Residential/Commercial Gas Demand Regions 8-14
Figure 8-13 GMM Industrial Gas Demand Regions 8-15
Figure 8-14 GMM Power Generation Gas Demand Regions 8-16
Figure 8-15 GMM U.S. and Canada Gas Demand Projection 8-17
Figure 8-16 Demand Region Definition 8-19
Figure 8-17 Supply Region Definition 8-19
Figure 8-18 Supply Curves for 2021, 2023, 2025, 2030, 2035, 2040, 2045, and 2050 8-20
xi
-------
LIST OF ATTACHMENTS
Attachment 3-1 NOx Rate Development in EPA Platform v6 3-34
Attachment 4-1 Nuclear Power Plant Life Extension Cost Development Methodology 4-49
Attachment 5-1 Wet FGD Cost Methodology 5-28
Attachment 5-2 SDA FGD Cost Methodology 5-28
Attachment 5-3 SCR Cost Methodology 5-28
Attachment 5-4 SNCR Cost Methodology 5-28
Attachment 5-5 DSI Cost Methodology 5-28
Attachment 5-6 Hg Cost Methodology 5-28
Attachment 5-7 PM Cost Methodology 5-28
Attachment 6-1 CO2 Reduction Cost Development Methodology 6-9
Attachment 7-1 Mining Cost Estimation Methodology and Assumptions 7-34
xii
-------
1.
Introduction
1.1 Executive Summary
This document describes the nature, structure, and capabilities of the Integrated Planning Model (IPM)
and the assumptions underlying the EPA's Power Sector Modeling Platform version 6 (EPA Platform v6)
that was developed by the U.S. Environmental Protection Agency (EPA) with technical support from ICF,
Inc. IPM is a multi-regional, dynamic, deterministic linear programming model of the U.S. electric power
sector. It provides forecasts of least cost capacity expansion, electricity dispatch, and emission control
strategies while meeting energy demand, environmental, transmission, dispatch, and reliability
constraints. IPM can be used to evaluate the cost and emissions impacts of proposed policies to limit
emissions of sulfur dioxide (SO2), nitrogen oxides (NOx), carbon dioxide (CO2), mercury (Hg), and HCI
from the electric power sector.
This introduction chapter summarizes the key modeling capabilities and major data elements that are
described in greater detail in the subsequent chapters.
EPA Platform v6 incorporates important structural improvements and data updates with respect to the
previous version (v5.13). A new version number (moving from v5 to v6) indicates a substantial change to
the architecture (such as this version's significantly more detailed representation of the load segments
and seasons). EPA Platform v6's November 2018 Reference Case uses Energy Information Agency's
(EIA) Annual Energy Outlook (AEO) 2018 demand projections.
EPA Platform v6 documentation includes assumptions and data values that were used to produce the
November 2018 Reference Case; for subsequent runs that examine various future scenarios, we include
separate documentation that makes clear where any assumptions or data values differ from the
November 2018 Reference Case conditions shown in this core documentation for Platform v6. EPA
Platform v6 November 2018 Reference Case serves as the starting point against which key drivers of the
power system dynamics (such as level of fuel prices, high or low costs for generation technologies and
high or low demand growth) are compared and analyzed. EPA Platform v6 is coupled with a Results
Viewer to facilitate easy comparison of different scenario projections and linking them with historical data.
When policy analysis is conducted using EPA Platform v6, relevant assumptions and documentation will
be provided elsewhere accordingly.
EPA Platform v6 November 2018 Reference Case is a projection of electricity sector activity that takes
into account only those Federal and state air emission laws and regulations whose provisions were either
in effect or enacted as documented in Section 3.9. Section 3.9 contains a detailed discussion of the
environmental regulations included in EPA Platform v6, which are summarized below.
• EPA Platform v6 includes the Cross-State Air Pollution Rule (CSAPR) Update Rule, a federal
regulatory measure affecting 22 states to address transport under the 1997 and 2006 National
Ambient Air Quality Standards (NAAQS) for ozone and fine particles.
• EPA Platform v6 reflects the Standards of Performance for Greenhouse Gas Emissions from New,
Modified, and Reconstructed Stationary Sources: Electric Utility Generating Units.1
• EPA Platform v6 includes the Mercury and Air Toxics Rule (MATS),2 which was finalized in
2011. MATS establishes National Emissions Standards for Hazardous Air Pollutants (NESHAP) for
the "electric utility steam generating unit" source category.
1 80 FR 64510
2 82 FR 16736
1-1
-------
• EPA Platform v6 reflects current and existing state regulations. A summary of these state regulations
can be found in Table 3-23.
• EPA Platform v6 reflects the final actions EPA has taken to implement the Regional Haze
Regulations and Guidelines for Best Available Retrofit Technology (BART) Determinations Final
Rule3. This regulation requires states to submit revised State Implementation Plans (SIPs) that
include (1) goals for improving visibility in Class I areas on the 20% worst days and allowing no
degradation on the 20% best days and (2) assessments and plans for achieving Best Available
Retrofit Technology (BART) emission targets for sources placed in operation between 1962 and
1977. Since 2010, EPA has approved SIPs or, in a very few cases, put in place regional haze
Federal Implementation Plans for several states. The BART limits approved in these plans (as of
summer 2017) that will be in place for EGUs are represented in the EPA Platform v6.
• EPA Platform v6 also includes three non-air federal rules affecting EGUs: National Pollutant
Discharge Elimination System-Final Regulations to Establish Requirements for Cooling Water Intake
Structures at Existing Facilities and Amend Requirements at Phase I Facilities, Hazardous and Solid
Waste Management System; Disposal of Coal Combustion Residuals From Electric Utilities; and the
Effluent Limitation Guidelines and Standards for the Steam Electric Power Generating Point Source
Category.4
Table 1-1 lists key updates included in EPA Platform v6 incremental to the previous major platform (v5)
with the corresponding data sources. Highlighted items are the updates incremental to the previous
release of EPA's v6 Platform (May 2018). The updates are listed in the order in which they appear in the
documentation.
Table 1-1 Key Updates in the EPA Platform v6 November 2018 Reference Case
Description
For More
Information
Modeling Framework
Modeling time horizon out to 2050 with eight model run years (2021, 2023, 2025,
2030, 2035, 2040, 2045, 2050)
Table 2-1
Incorporation of three seasons
Section 2.3.5
Increasing the number of load segments to 72 per year
Section 2.3.5
All costs and prices are in 2016 dollars
Power System Operation
Updates based on recent data from EIA, NERC, and FERC
Chapter 3
Updated inventory of state emission regulations
Section 3.9
CSAPR, MATS, and BART are reflected
Section 3.9.3
Updated RPS standards for CT and NJ
Table 3-19
Updated inventories of NSR, state, and citizen settlements (as of May 2018)
Table 3-24, Table
3-25, and Table 3-26
Updated transmission Total Transfer Capability's (TTC) and regional reserve
margins (2015-2016 ISO/RTO NERC Reports)
Table 3-5 and Table
3-21
AEO 2018 NEMS region level electricity demand is disaggregated to IPM model
region level. IPM model region level peak load projection is based on the future load
factors from NERC 2017 ES&D and AEO 2018
Section 3.2
Implemented the NY minimum oil burn rule through facility level minimum generation
constraints
Table 3-13
3 70 FR 39104
4 79 FR 48300, 80 FR 21302, 80 FR 67838
1-2
-------
Description
For More
Information
Updated ELG costs
Section 3.9.5
Generating Resources
Updates to NEEDS planned units, retirements, and emission control configurations
(July 2018 EIA Form 860m, 2017 EIA Form 860 ER, AEO 2018, AMPD 2017 and
Table 4-1
recent lists of deactivations from PJM, MISO, and ERCOT)
Updates to unit level NOx rates (EPA ETS 2017, 2016 CARB, and 2014 NEI)
Section 3.9.2
Providing life extension costs to allow existing nuclear units to continue operation
over the extended 80 year life (Sargent & Lundy 2017)
Section 4.5.1
Updated cost and performance characteristics for potential (new) conventional,
renewable, and nuclear generating units (AEO 2017 and NREL ATB 2017)
Table 4-13 and Table
4-16
Wind and solar technologies have revised cost and resource base estimates,
capacity credit calculation methodology, hourly generation profiles, and time of day
based load segments to improve curtailment modeling (NREL 2017)
Section 4.4.5
Implemented energy storage options based on AEO 2018 cost and performance
assumptions. Included the energy storage mandates for CA, NY, NJ, OR and MA.
Table 4-35 and Table
4-36
Emission Control Technologies
Complete update of cost and performance assumptions for SO2, NOx, Hg, HCI and
CO2 emission controls based on engineering studies by Sargent & Lundy
Chapter 5
Inclusion of cost and performance assumptions for coal-to-gas conversion and
capability to model heat rate improvement technologies
Section 5.7
Carbon Capture, Transport, and Storage
Updated CO2 storage cost curves based on a $75 crude oil price, average EOR
efficiency of 10 Mcf of CO2 per incremental barrel of crude oil and adjustment to
geologic storage curve for industrial uses of storage capacity
Table 6-4
Updated CO2 transportation cost adders
Table 6-5
Coal
Complete update of coal supply curves and transportation matrix (Wood Mackenzie
2016 and Hellerworx2016)
Table 7-25 and Table
7-26
Natural Gas
Natural gas assumptions modeled through annual gas supply curves and IPM region
level seasonal basis differentials (ICF 2017)
Section 8.6
Other Fuels
Incorporation of biomass supply curves at a state and IPM region level (DOE 2016)
Section 9.2
Update of price assumptions for fuel oil, nuclear fuel, and waste fuel (AEO 2017)
Chapter 9
Financial assumptions
Update of discount and capital charge rate assumptions based on a hybrid capital
cost model of utility and merchant finance structures for new units
Chapter 10
Use of separate capital charge rates for retrofits based on utility and merchant
finance structures
Section 10.4.2
Cost adder for new non-peaking fossil units associated with future CO2 emissions
Section 10.7.3
Incorporated the implications of the Tax Reform Bill in the discount rate and capital
charge rate calculations
Section 10.3
1-3
-------
Table 1-2 lists the types of plants included in the EPA Platform v6.
Table 1-2 Plant Types in EPA Platform v6
Conventional Technologies
Coal Steam
Oil/Gas Steam
Combustion Turbine
Combined-Cycle Combustion Turbine
Integrated Gasification Combined-Cycle (IGCC) Coal
Ultra-Supercritical Coal with and without Carbon Capture
Fluidized Bed Combustion
Nuclear
Renewables and Non-Conventional Technologies
Hydropower
Pumped Storage
Energy Storage
Biomass
Onshore Wind
Offshore Wind
Fuel Cells
Solar Photovoltaics
Solar Thermal
Geothermal
Landfill Gas
Other1
Note:
1 Included are fossil and non-fossil waste plants.
1-4
-------
Table 1-3 lists the emission control technologies available for meeting emission limits in EPA Platform v6.
Table 1-3 Emission Control Technologies in EPA Platform v6
Sulfur Dioxide (SO2)
Limestone Forced Oxidation (LSFO)
Lime Spray Dryer (LSD)
Nitrogen Oxides (NOx)
Combustion controls
Selective catalytic reduction (SCR)
Selective non-catalytic reduction (SNCR)
Mercury (Hg)
Combinations of SO2, NOx, and particulate control technologies
Activated Carbon Injection
Hydrogen Chloride (HCI)
Dry Sorbent Injection (with milled Trona)
Carbon Dioxide (CO2)
Heat rate improvement
Coal-to-gas
Carbon Capture and Sequestration
Notes:
Fuel switching between coal types is also a compliance option
for reducing emissions in EPA Platform v6.
Figure 1-1 provides a schematic of the components of the modeling and data structure used for EPA
Platform v6. The document contains separate chapters devoted to all the key components shown in
Figure 1-1. Chapter 2 provides an overview of IPM's modeling framework (also referred to as the "IPM
Engine"), highlighting the mathematical structure, notable features of the model, programming elements,
and model inputs and outputs. The remaining chapters are devoted to different aspects of EPA Platform
v6. Chapter 3 covers the power system operating characteristics captured in EPA Platform v6. Chapter 4
explores the characterization of electric generation resources. Emission control technologies (chapter 5)
and carbon capture, transport and storage (chapter 6) are then presented. The next three chapters
discuss the representation and assumptions for fuels in the EPA Platform v6. Coal is covered in chapter
7, natural gas in chapter 8, and other fuels (i.e., fuel oil, biomass, nuclear fuel, and waste fuels) in chapter
9 (along with fuel emission factors). Finally, Chapter 10 summarizes the financial assumptions.
1-5
-------
Figure 1-1 Modeling and Data Structures in EPA Platform v6
Outputs for Air Quality
Modeling
Criteria Air Pollutants
Post-Processor
Parsing Outputs
Individual Boiler Level Data
Retail Price Projection
IPM Engine**
(Chapter 2)
Coal
(Chapter 7)
Coal Supply Curves
Coal Transportation Matrix
C02 Capture, Transport,
and Storage
(Chapter 6)
Capture Technologies
Transportation
Storage Cost Curves
Model Outputs
Emissions
Costs
Capacity Expansion
Retrofit Decisions
Fuel Consumption and Prices
Electricity Generation and
Prices
Generation Resources
(Chapter 4)
Existing EGUs*
Planned EGUs*
Potential New EGUs
Short-term Capital Cost Adder
Regional Cost Adjustments
Financial Assumptions
(Chapter 10)
Discount Rate
Capital Charge Rate
Book Life
Adder for Climate Uncertainty
Other Fuel Assumptions
(Chapter 9)
Fuel Oil
Nuclear Fuel
Biomass
Waste Fuels
Emission Factors
Emission Control Technologies
(Chapter 5)
Sulfur Dioxide and HCI
Nitrogen Oxides
Mercury
Carbon Capture and Storage
Particulate Controls
Coal-to-Gas Conversion
Power System Operation
(Chapter 3)
Regional Configurations
Capacity and Dispatch Assumptions
T ransmission Assumptions
Turndown Constraints
Reliability Constraints
Electricity Demand Growth
Environmental Regulations
Natural Gas
(Chapter 8)
Synopsis of the Gas Market Model (GMM)
Resources Data and Reservoir Description
Treatment of Frontier Resources and Exports
Fuel Prices
Demand Assumptions
Supply Curves
Regional Basis
Notes:
* Information on existing and planned electric generating units (EGUs) is contained in the National Electrical
Energy Data System (NEEDS) data base maintained for EPA by ICF, Inc. Planned EGUs are those which
were under construction or had obtained financing at the time EPA's Platform v6 was finalized.
**IPM Engine is the model structure described in Chapter 2
Outputs for Air Quality
Modeling
Criteria Air Pollutants
Non-criteria Air Pollutants
Toxics Air Pollutants
Point Source Locators
1.2 Review and Ongoing Improvement of the Integrated Planning Model
A customized, fully documented version of the data assumptions underlying IPM has been developed and
used by EPA to help inform power plant air regulatory and legislative efforts for over 20 years, following
the enactment of the Clean Air Act Amendments of 1990. The model has been tailored to meet the
unique environmental considerations important to EPA, while also fuliy capturing the detailed and
complex economic and electric dispatch dynamics of power plants across the country. It has been EPA's
goal to thoroughly explain and document the agency's use of the model in a transparent and publicly
accessible manner, while also providing for concurrent channels for improving the model's assumptions
and representation by soliciting constructive feedback so that the model may be continually improved.
This includes making ail inputs and assumptions to the model, as well as all output files from the model,
publicly available on EPA's website (and, when applied to inform a rulemaking, in the relevant publicly
accessible regulatory docket).
EPA's use of IPM depends upon a variety of environmental, policy, and regulatory considerations.
Generally, EPA's version of the model input assumptions has undergone significant updates and
architectural improvements every 2-4 years in order to best reflect the evolving dynamics of the power
sector, and smaller ongoing updates (1-2 times a year) to reflect changes in fleet composition
1-6
-------
(retirements, new capacity builds, and installed retrofits). Currently, EPA's implementation of IPM is in its
sixth major version, not including Coal and Electric Utility Model (CEUM), the model used by EPA before
its use of IPM.
Federal Regulatory efforts:
EPA has used IPM for many regulatory efforts affecting the power sector, including:
• The NOx SIP Call, the Clean Air Interstate Rule (2004-2006), the Clean Air Visibility Rule, the
Clean Air Mercury Rule (2005), the Cross-State Air Pollution Rule and Updates (2010-2016), the
Mercury and Air Toxics Rule (2012), the Clean Power Plan (2015), and various Ozone, PM
NAAQS, and regional haze regulatory efforts.
National Legislative efforts:
EPA has used IPM to support legislative efforts that affect the power sector, including:
• The Clear Skies Act (2002-2005), the Clean Air Planning Act (2002-2005), the Clean Power Act
(2002-2005), the Climate Stewardship and Innovation Act (2007), the Low Carbon Economy Act
(2007-2008), the Lieberman-Warner Climate Security Act (2007-2008), and the American Clean
Energy and Security Act (2008-2009).
Notable Versions and Updates/Improvements/Enhancements:
EPA Base Case using IPM - 1996
• Designed for projections covering the US with 4 run years
• Disaggregated the US into 17 IPM model regions
• Modeled coal and gas markets through coal and gas supply curves
EPA Base Case using IPM - 1998
• Updated unit inventory of power plants
• Increased the number of IPM model regions covering the US from 17 to 21
• Disaggregated New York into 4 IPM model regions
• Increased the number of run years from 4 to 6
EPA Base Case 2000 using IPM Version 2.1 (2000-2003)
• Updated unit inventory of power plants
• Increased the number of IPM model regions covering the US from 21 to 26
• Increased the modeling time horizon to 2030
• Increased the overall number of emission control technology options modeled
• Incorporated Activated Carbon Injection (ACI) retrofit options for mercury control modeling
• Expanded coal supply representation
EPA Base Case 2004 using IPM Version 2.1.9 (2004)
• Updated unit inventory of power plants
• Improved the characterization of SO2 and NOx emissions
• Revised coal choice assumptions for individual coal units
• Updated natural gas supply curves, incorporating recommendations from the natural gas peer
review
1-7
-------
EPA Base Case 2006 using IPM Version 3 (2005-2009)
• Updated unit inventory of power plants
• Improved environmental pollution control retrofit assumptions
• Increased the number of IPM model regions covering the US from 26 to 32 to enhance regional
representation
• Increased the number of load segments from 5 to 6 to enhance electric load representation
• Updated natural gas supply curves based on ICF's North American Natural Gas Systems
Analysis (NANGAS) model
• Updated coal supply curves
• Enhanced electric transmission capabilities and imports/exports
• Enhanced power plant representation detail
EPA Base Case using IPM Version 4.10 (2010-2013)
• Updated unit inventory of power plants
• Integrated Canada into the modeling framework
• Incorporated HCI emissions and Dry Sorbent Injection retrofit options
• Improved resolution of Carbon Capture and Storage, including regional storage representation
and transportation network
• Updated coal supply modeling with significantly more resolution of coal mine data
• Incorporated natural gas resource model for North America to reflect emerging shale resource
• Enhanced power plant representation detail to support toxic air pollutant emissions and controls
EPA Base Case using IPM Version 5 (2014-2017)
• Updated unit inventory of power plants
• Doubled the number of IPM model regions from 36 to 64
• Revised environmental pollution control retrofit assumptions for conventional pollutants and toxic
emissions
• Incorporated additional technology options for new power plants
• Overhauled coal supply assumptions, with even further resolution to reflect mine-by-mine
geography and coal characteristics
• Improved coal transportation network by modeling each individual coal plant as its own coal
demand region
• Updated gas modeling assumptions to reflect natural gas shale supply/trends and pipeline
capacity expansion
Background on EPA Base Case using IPM Review:
Peer Reviews:
EPA conducts periodic formal peer review of the EPA Base Case application of IPM. These reviews have
included separate expert panels on the model itself, and EPA's key modeling input assumptions. For
example, separate panels of independent experts have been convened to review IPM's coal supply and
transportation assumptions, natural gas assumptions, and model formulation.
EPA Base Case v5.13 Data Assumption Review
In 2015, an independent peer review panel provided expert feedback on whether the analytical
framework, assumptions, and applications of data in IPM were sufficient for the EPA's needs in estimating
the economic and emissions impacts associated with the power sector. The panel identified a number of
strengths associated with the model and underlying data and assumptions. For example, the report
1-8
-------
stated that EPA's platform exceeds other model capabilities in providing a relevant feedback mechanism
between the electric power model and key fuel inputs that drive simulation results5.
Other strengths the panel identified include:
• The detail with which pollution control technology options and costs are represented
• The level of detail at which federal Clean Air Act (CAA) regulations are represented
• The ability of the model to allow for the detailed representation of a variety of potential changes in
energy and environmental policies, including important features of market-based programs
• The accuracy of the emissions control costs and their relationship to retirement decisions
• The expansion of model regions from 32 to 64, which allows the model to better represent current
power market operations and existing transmission bottlenecks even within regional transmission
organization (RTO) regions
• Continuous updates of the representation of domestic coal and natural gas market conditions
The peer review panel has also provided several areas for investigation and additional recommendations
for the EPA's consideration, including:
• Improved documentation of the input assumptions
• Changes to certain cost functions and financial assumptions
• Consideration of certain improvements to the Base Case architecture (additional seasonal
representation, representation of electric demand, transmission considerations, and renewable
energy representation among others)
The updated EPA Platform v6 using IPM addresses many of these recommendations (seasons,
renewable energy representation, regional representation, etc.), and this peer review has also lead to
additional work at EPA to further understand and better represent some of the emerging issues in the
power sector. EPA intends to add more capabilities and continue to refine the modeling platform to
reflect these comments, and adopt those changes at an appropriate time after further research and
testing of the model.
Coal Market Assumptions Review
In 2003, a group of experts in the field of cost, quality, reserves, and availability of coal were selected as
peer reviewers to assess whether the choice, use, and interpretation of data and methodology employed
in the derivation of the IPM coal supply curves was appropriate and analytically sound. The peer
reviewers were charged with:
• Evaluating the appropriateness of the overall methodology used to develop the new coal supply
curves,
• Assessing the adequacy of the individual components employed in building the coal supply
curves in terms of both the approach and data used,
• Assessing the technical soundness of the resulting coal supply curves for each coal type and
supply region in terms of the cost/quantity relationship and the characteristics associated with the
coal (e.g., sulfur, heat, and mercury content), and
• Assessing the appropriateness of the use of this set of supply curves for use in production cost
models in general (of which IPM is a particular example).
The review process produced useful and specific recommendation for improvements and updates to the
coal supply information that is represented in IPM, which were subsequently incorporated into the model.
5 https://www.epa.aov/airmarkets/response-and-peer-review-report-epa-base-case-version-513-usina-ipm
1-9
-------
Gas Market Assumptions Review
In 2003, a peer review of the natural gas supply assumptions implemented in EPA Base Case using IPM
v.2.1.6 (2003) was performed. The peer reviewers were charged with evaluating the following:
• The appropriateness of the representation of all the key natural gas market fundamentals in
NANGAS,
• The reasonableness of the natural gas supply curves, non-electricity demand assumptions and
transportation adders, and
• The reasonableness of the iteration process between NANGAS and IPM.
The review commended the comprehensiveness of the approach used to generate the gas supply curves
implemented in the EPA Base Case. The review further identified assumptions that could be revised in
generating a new set of natural gas supply curves, as well as nonelectric-sector gas demand curves, for
the next update of the EPA Base Case.
IPM Formulation Review
Conducted in 2008, this peer review focused on IPM's core mathematical formulation. The objective of
the review was to obtain expert feedback on the adequacy of the formulation in representing the
economic and operational behavior of the power sector over a modeling time horizon of 20-50 years.
The panel identified several strengths of IPM, including:
• The model's ability to compute optimal capacity that combined short-term dispatch decisions with
long-term investment decisions.
• The model's integration of relevant markets, including the electric power, fuel, and environmental
markets, into a single modeling framework.
• And the model's ability to represent a very detailed level of data with regard to the emissions
modeling capability.
The peer review panel also provided several areas for investigation and recommendations for the EPA's
consideration. These peer reviews led to concrete changes, enhancements, and updates to the IPM
framework to better represent the power sector and related markets (i.e., fossil fuels).
Regulatory Review:
The formal rulemaking process provides opportunity for expert review and comment by key stakeholders.
Formal comments as part of a rulemaking are reviewed and evaluated, and changes/updates are made to
IPM where appropriate. Stakeholders to EPA regulatory efforts are a diverse group, including regulated
entities and impacted industries, fuel supply companies, states, environmental organizations, developers
of other models of the U.S. electricity sector, and others. The feedback provides a highly detailed review
of input assumptions, model representation, and model results.
Other Uses and Reviews:
• IPM has been used by many regional organizations for regulatory support, including the Regional
Greenhouse Gas Initiative (RGGI), the Western Regional Air Partnership (WRAP), and the Ozone
Transport Assessment Group (OTAG). IPM has also been used by other Federal agencies (e.g.,
FERC, USDA), environmental groups, and many electric utilities,
• The Science Advisory Board reviewed EPA's application of IPM as part of the CAAA Section 812
prospective study 1997-1999.
• The President's Council of Economic Advisors (2002-2003) performed head-to-head comparison
of IPM and ElA's NEMS system for use in multi-pollutant control analysis.
1-10
-------
• IPM has been used in a number of comparative model exercises sponsored by Stanford
University's Energy Modeling Forum and other organizations.
EPA Platform v6 using IPM represents another major iteration of EPA's application of IPM, with notable
structural and platform improvements/enhancements, as well as universal updates to reflect the most
current set of data and assumptions.
1-11
-------
2. Modeling Framework
ICF developed the Integrated Planning Model (IPM) to support analysis of the electric power sector. The
EPA, in addition to other state air regulatory agencies, utilities, and public and private sector entities, has
used IPM extensively for various air regulatory analyses, market studies, strategy planning, and economic
impact assessments.
IPM is a long-term capacity expansion and production-costing model of the electric power sector. Its
mathematical formulation is based on a Linear Programming (LP) structure. The structure provides for
several advantages, one of which is the guarantee of a globally optimal solution. Fast and efficient
commercial solvers exist to solve LP models. The solved dual variables (also known as shadow prices)
of each constraint modeled in IPM inform EPA rule-making or policy analysis process in regards to the
marginal cost pricing of energy, capacity, fuels, and emission allowances. Also, reasonable solution
times for an LP model allow EPA to gain insights by modeling a large number of scenarios in a relatively
short period of time.
The first section of this chapter provides a brief overview of the model's purpose, capabilities, and
applications. The following sections are devoted to describing the IPM's model structure and formulation
(2.2), key methodological characteristics (2.3), and programming features (2.4), including its handling of
model inputs and outputs. Readers may find some overlap between sections. For example, transmission
decision variables and constraints are covered in the discussion of model structure and formulation in
section 2.2, and transmission modeling is covered as a key methodological feature in section 2.3.8. The
different perspectives of each section are designed to provide readers with information that is
complementary rather than repetitive.
2.1 IPM Overview
IPM is a well-established model of the electric power sector designed to help government and industry
analyze a wide range of issues related to this sector. The model represents economic activities in key
components of energy markets - fuel markets, emission markets, and electricity markets. Since the
model captures the linkages in electricity markets, it is well suited for developing integrated analyses of
the impacts of alternative regulatory policies on the power sector. In the past, applications of IPM have
included capacity planning, environmental policy analysis and compliance planning, wholesale price
forecasting, and power plant asset valuation.
2.1.1 Purpose and Capabilities
IPM is a dynamic linear programming model that generates optimal decisions under the assumption of
perfect foresight. It determines the least-cost method of meeting energy and peak demand requirements
over a specified period. In its solution, the model considers a number of key operating or regulatory
constraints that are placed on the power, emissions, and fuel markets. The constraints include, but are
not limited to, emission limits, transmission capabilities, renewable generation requirements, and fuel
market constraints. The model is designed to accommodate complex treatment of emission regulations
involving trading, banking, and special provisions affecting emission allowances (e.g., bonus allowances
and progressive flow control), as well as traditional command-and-control emission policies.
IPM represents power markets through model regions that are geographical entities with distinct
operational characteristics. The model regions are largely consistent with the North American Electric
Reliability Council (NERC) assessment regions, and with the organizational structures of the Regional
Transmission Organizations (RTOs), and the Independent System Operators (ISOs) that handle dispatch
on most of the U.S. grid. IPM represents the least-cost arrangement of electricity supply (capacity and
generation) within each model region to meet assumed future load (electricity demand) while constrained
by a transmission network of bulk transfer limitations on interregional power flows. All utility-owned
existing electric generating units, including renewable resources, as well as independent power producers
and cogeneration facilities selling electricity to the grid, are modeled.
2-1
-------
IPM provides a detailed representation of new and existing resource options. These include fossil,
nuclear, renewable, and non-conventional options. Fossil options include coal steam, oil/gas steam,
combined cycles, and gas-fired simple cycle combustion turbines. Renewable options include wind,
landfill gas, geothermal, solar thermal, solar photovoltaic, and biomass. Non-conventional options include
fuel cell, pump storage, and battery storage.
IPM can incorporate a detailed representation of fuel markets and can endogenously forecast fuel prices
for coal, natural gas, and biomass by balancing fuel demand and supply for electric generation. The
model also includes detailed fuel quality parameters to estimate emissions from electric generation.
IPM provides estimates of air emission changes, regional wholesale energy and capacity prices,
incremental electric power system costs, changes in fuel use, and capacity and dispatch projections.
2.1.2 Applications
IPM's structure, formulation, and set-up make it adaptable and flexible. The necessary level of data,
modeling capabilities exercised, and computational requirements can be tailored to the particular
strategies and policy options being analyzed. This adaptability has made IPM suitable for a variety of
applications. These include:
Air Regulatory Assessment: Since IPM contains extensive air regulatory modeling features, state and
federal air regulatory agencies have used the model extensively in support of air regulatory assessment.
Integrated Resource Planning: IPM can be used to perform least-cost planning studies that
simultaneously optimize demand-side options (load management and efficiency), renewable options and
traditional supply-side options.
Strategic Planning: IPM can be used to assess the costs and risks associated with alternative utility and
consumer resource planning strategies as characterized by the portfolio of options included in the input
data base.
Options Assessment: IPM allows industry and regulatory planners to "screen" alternative resource
options and option combinations based upon their relative costs and contributions to meeting customer
demands.
Cost and Price Estimation: IPM produces realistic estimates of energy prices, capacity prices, fuel prices,
and allowance prices. Industry and regulatory agencies have used these cost reports for due diligence,
planning, litigation, and economic impact assessment.
2.2 Model Structure and Formulation
IPM employs a linear programming structure that is particularly well-suited for analysis of the electric
sector to help decision makers plan system capacity and model the dispatch of electricity from individual
units or plants. The model consists of three key structural components:
• A linear "objective function,"
• A series of "decision variables," and
• A set of linear "constraints".
• The sections below describe the objective function, key decision variables, and constraints included
in IPM for EPA Platform v6.
2-2
-------
2.2.1 Objective Function
IPM's objective function is to minimize the total, discounted net present value of the costs of meeting
demand, power operation constraints, and environmental regulations over the entire planning horizon.
The objective function represents the summation of all the costs incurred by the electricity sector on a net
present value basis. These costs, which the linear programming formulation attempts to minimize,
include the cost of new plant and pollution control construction, fixed and variable operating and
maintenance costs, and fuel costs. Many of these cost components are captured in the objective function
by multiplying the decision variables by a cost coefficient. Cost escalation factors are used in the
objective function to reflect changes in cost overtime. The applicable discount rates are applied to derive
the net present value for the entire planning horizon from the costs obtained for all years in the planning
horizon.
2.2.2 Decision Variables
Decision variables represent the values for which the IPM model is solving, given the cost-minimizing
objective function described in section 2.2.1 and the set of electric system constraints detailed in section
2.2.3. The model determines values for these decision variables that represent the optimal least-cost
solution for meeting the assumed constraints. Key decision variables represented in IPM are described in
detail below.
Generation Dispatch Decision Variables: IPM includes decision variables representing the generation
from each model power plant.6 For each model plant, a separate generation decision variable is defined
for each possible combination of fuel, season, model run year, and segment of the seasonal load duration
curve applicable to the model plant. (See section 2.3.5 below for a discussion of load duration curves.)
In the objective function, each plant's generation decision variable is multiplied by the relevant heat rate
and fuel price (differentiated by the appropriate step of the fuel supply curve) to obtain a fuel cost. It is
also multiplied by the applicable variable operation and maintenance (VOM) cost rate to obtain the VOM
cost for the plant.
Capacity Decision Variables: IPM includes decision variables representing the capacity of each existing
model plant and capacity additions associated with potential (new) units in each model run year. In the
objective function, the decision variables representing existing capacity and capacity additions are
multiplied by the relevant fixed operation and maintenance (FOM) cost rates to obtain the total FOM cost
for a plant. The capacity addition decision variables are also multiplied by the investment cost and capital
charge rates to obtain the capital cost associated with the capacity addition.
Transmission Decision Variables: IPM includes decision variables representing the electricity
transmission along each transmission link between model regions in each run year. In the objective
function, these variables are multiplied by variable transmission cost rates to obtain the total cost of
transmission across each link.
Emission Allowance Decision Variables: For emission policies where allowance trading applies, IPM
includes decision variables representing the total number of emission allowances for a given model run
year that are bought and sold in that or subsequent run years. In the objective function, these year-
differentiated allowance decision variables are multiplied by the market price for allowances prevailing in
each run year. This formulation allows IPM to capture the inter-temporal trading and banking of
allowances.
Fuel Decision Variables: For each type of fuel and each model run year, IPM defines decision variables
representing the quantity of fuel delivered from each fuel supply region to model plants in each demand
6 Model plants are aggregate representations of real-life electric generating units. They are used by IPM to model the
electric power sector. For a discussion of model plants in EPA Platform v6, see section 4.2.6.
2-3
-------
region. Coal decision variables are further differentiated according to coal rank (bituminous, sub-
bituminous, and lignite), sulfur grade, chlorine content and mercury content (see Table 7-4). These fuel
quality decision variables do not appear in the IPM objective function, but in constraints which define the
types of fuel that each model plant is eligible to use and the supply regions that are eligible to provide fuel
to each specific model plant.
2.2.3 Constraints
Model constraints are implemented in IPM to accurately reflect the characteristics of and the conditions
faced by the electric sector. Among the key constraints included in EPA Platform v6 are:
Reserve Margin Constraints: Regional reserve margin constraints capture system reliability requirements
by defining a minimum margin of reserve capacity (in megawatts) per year beyond the total capacity
needed to meet future peak demand that must remain in service to that region. These reserve capacity
constraints are derived from reserve margin targets that are assumed for each region based on
information from NERC, RTOs or ISOs. If existing plus planned capacity is not sufficient to satisfy the
annual regional reserve margin requirement, the model will "build" the required level of new capacity.
Please see Section 3.6 for more information on reserve margin assumptions.
Demand Constraints: The model categorizes regional annual electricity demand into seasonal load
curves which are used to form winter (December 1 - February 28), winter shoulder (March 1 - April 30,
October 1 - November 30), and summer (May 1 - September 30) load duration curves (LDC). The
seasonal load segments when taken together represent all the hourly electricity load levels that must be
satisfied in a region in the particular season for a particular model run year. As such, the LDC defines the
minimum amount of generation required to meet the region's electrical demand during the specific
season. These requirements are incorporated in the model's demand constraints.
Capacity Factor Constraints: These constraints specify how much electricity each plant can generate (a
maximum generation level), given its capacity and seasonal availability.
Turn Down Constraints: The model uses these constraints to take into account the cycling capabilities of
the units, i.e., whether or not they can be shut down at night or on weekends, or whether they must
operate at all times, or at least at some minimum capacity level. These constraints ensure that the model
reflects the distinct operating characteristics of peaking, cycling, and base load units.
Emissions Constraints: IPM can endogenously consider an array of emissions constraints for SO2, NOx,
HCI, mercury, and CO2. Emission constraints can be implemented on a plant-by-plant, regional, or
system-wide basis. The constraints can be defined in terms of a total tonnage cap (e.g., tons of SO2) or a
maximum emission rate (e.g., Ib/MMBtu of NOx). The scope, timing, and definition of the emission
constraints depend on the required analysis.
Transmission Constraints: IPM can simultaneously model any number of regions linked by transmission
lines. The constraints define either a maximum capacity on each link, or a maximum level of transmission
on two or more links (i.e., joint limits) to different regions.
Fuel Supply Constraints: These constraints define the types of fuel that each model plant is eligible to
use and the supply regions that are eligible to provide fuel to each specific model plant. A separate
constraint is defined for each model plant.
2.3 Key Methodological Features of IPM
IPM is a flexible modeling tool for obtaining short- and long-term projections of production activity in the
electric generation sector. The projections obtained using IPM are not statements of what will happen,
but they are estimates of what might happen given the assumptions and methodologies used. Chapters
3 to 10 contain detailed discussions of the cost and performance assumptions specific to EPA Platform
2-4
-------
v6. This section provides an overview of the essential methodological and structural features of IPM that
extend beyond the assumptions that are specific to EPA Platform v6.
2.3.1 Model Plants
Model plants are a central structural component that IPM uses in three ways: (1) to represent
aggregations of existing generating units, (2) to represent retrofit and retirement options that are available
to existing units, and (3) to represent potential (new) units that the model can build.
Existing Units: Theoretically, there is no predefined limit on the number of units that can be included in
IPM. However, to keep model size and solution time within acceptable limits, EPA utilizes model plants to
represent aggregations of actual individual generating units. The aggregation algorithm groups units with
similar characteristics for representation by model plants with a combined capacity and weighted-average
characteristics that are representative of all the units comprising the model plant. Model plants are
defined to maximize the accuracy of the model's cost and emissions estimates by capturing variations in
key features of those units that are critical in the November 2018 Reference Case and anticipated policy
case runs. For EPA Platform v6, IPM employed an aggregation algorithm, which allowed 21,931 actual
existing electric generating units to be represented by 5,747 model plants. Section 4.2.6 describes the
aggregation procedure used in the EPA Platform v6.
Retrofit and Retirement Options: IPM also utilizes model plants to represent the retrofit and retirement
options that are available to existing units. EPA Platform v6 provides existing model plants with a wide
range of options for retrofitting with emission control equipment as well as with an option to retire. (See
Chapter 5 for a detailed discussion of the options that are included in the EPA Platform v6.) EPA
Platform v6 model plants that represent potential (new) units are not given the option to take on a retrofit
or retire.
The options available to each model plant are pre-defined at the model's set-up. The retrofit and
retirement options are themselves represented in IPM by model plants, which, if actuated in the course of
a model run, take on all or a portion of the capacity initially assigned to a model plant, which represents
existing generating units7. In setting up IPM, parent-child-grandchild relationships are pre-defined
between each existing model plant (parent) and the specific retrofit and retirement model plants (children
and grandchildren) that may replace the parent model plant during the course of a model run. The "child"
and "grandchild" model-plants are inactive in IPM unless the model finds it economical to engage one of
the options provided, e.g., retrofit with particular emission controls or retire.
Theoretically, there are no limits on the number of child, grandchild, and even great-grandchild model
plants (i.e., retrofit and retirement options) that can be associated with each existing model plant.
However, model size and computational considerations dictate that the number of successive retrofits be
limited. In EPA Platform v6, a maximum of three stages of retrofit options are provided (child, grandchild
and great-grandchild). For example, an existing model plant may retrofit with an activated carbon
injection (ACI) for mercury control in one model run year (stage 1), with a selective catalytic reduction
(SCR) control for NOx in the same or subsequent run year (stage 2), and with a CCS for CO2 control in
the same or subsequent run year (stage 3). However, if it exercises this succession of retrofit options, no
further retrofit or retirement options are possible beyond the third stage.
Potential (New) Units: IPM also uses model plants to represent new generation capacity that may be built
during a model run. All the model plants representing new capacity are pre-defined at set-up,
differentiated by type of technology, regional location, and years available. When it is economically
advantageous to do so (or otherwise required by reserve margin constraints to maintain electric
7 IPM has a linear programming structure whose decision variables can assume any value within the specified
bounds subject to the constraints. Therefore, IPM can generate solutions where model plants retrofit or retire a
portion of the model plants capacity. IPM's standard model plant outputs explicitly present these partial investment
decisions.
2-5
-------
reliability), IPM "builds" one or more of these predefined model plants by raising its generation capacity
from zero during the course of a model run. In determining whether it is economically advantageous to
"build" new plants, IPM takes into account cost differentials between technologies, expected technology
cost improvements (by differentiating costs based on a plant's vintage, i.e., build year), and regional
variations in capital costs that are expected to occur over time.
Parsing: Since EPA Platform v6 results are presented at the model plant level, EPA has developed a
post-processor "parsing" tool designed to translate results at the model plant level into generating unit-
specific results. The parsing tool produces unit-specific emissions, fuel use, emission control retrofit and
capacity projections based on model plant results. Another post-processing activity involves deriving
inputs for air quality modeling from IPM outputs. This entails using emission factors to derive the levels of
pollutants needed in EPA's air quality models from emissions and other parameters generated by IPM. It
also involves using decision rules to assign point source locators to these emissions. (See Figure 1-1 for
a graphical representation of the relationship of the post-processing tools to the overall IPM structure.)
2.3.2 Model Run Years
Another important structural feature of IPM is the use of model 'run years' to represent the full planning
horizon being modeled. Although IPM can represent an individual year in an analysis time horizon,
mapping each year in the planning horizon into a representative model run year enables IPM to perform
multiple year analyses while keeping the model size manageable. IPM takes into account the costs in all
years in the planning horizon while reporting results only for model run years. (See section 2.3.3 below
for further details.)
The analysis time horizon for EPA Platform v6 extends from 2021 through 2054. The eight years
designated as "model run years" and the mapping of calendar years to run years is shown in Table 2-1.
Table 2-1 Run Year and Analysis Year Mapping Used in EPA Platform v6
Run Year
Years Represented
2021
2021
2023
2022 - 2023
2025
2024 - 2027
2030
2028 - 2032
2035
2033 - 2037
2040
2038 - 2042
2045
2043 - 2047
2050
2048 - 2054
Often models like IPM include a final model run year that is not included in the analysis of results. This
technique reduces the likelihood that modeling results in the last represented year will be skewed due to
the modeling artifact of having to specify an end point in the planning horizon, whereas, in reality,
economic decision-making will continue to take information into account from years beyond the model's
time horizon. This should be considered when assessing model projections from the last output year.
2.3.3 Cost Accounting
As noted earlier in the chapter, IPM is a dynamic linear programming model that solves for the least cost
investment and electricity dispatch strategy for meeting electricity demand subject to resource availability
and other operating and environmental constraints. The cost components that IPM takes into account in
deriving an optimal solution include the costs of investing in new capacity options, the cost of installing
and operating pollution control technology, fuel costs, and the operation and maintenance costs
associated with unit operations. Several cost accounting assumptions are built into IPM's objective
function that ensures a technically sound and unbiased treatment of the cost of all investment options
offered in the model. These features include:
2-6
-------
• All costs in IPM's single multi-year objective function are discounted to a base year. Since the
model solves for all run years simultaneously, discounting to a common base year ensures that
IPM properly captures complex inter-temporal cost relationships.
• Capital costs in IPM's objective function are represented as the net present value of levelized
stream of annual capital outlays, not as a one-time total investment cost. The payment period
used in calculating the levelized annual outlays never extends beyond the model's planning
horizon: it is either the book life of the investment or the years remaining in the planning horizon,
whichever is shorter. This approach avoids presenting artificially lower capital costs for
investment decisions taken closer to the model's time horizon boundary simply because some of
that cost would typically be serviced in years beyond the model's view. This treatment of capital
costs ensures both realism and consistency in accounting for the full cost of each of the
investment options in the model.
• The cost components informing IPM's objective function represent the composite cost over all
years in the planning horizon rather than just the cost in the individual model run years. This
permits the model to capture more accurately the escalation of the cost components over time.
2.3.4 Modeling Wholesale Electricity Markets
Another methodological feature worth noting about IPM is that it is designed to simulate electricity
production activity in a manner that would minimize production costs, as is the intended outcome in
wholesale electricity markets. For this purpose, the model captures transmission costs and losses
between IPM model regions, but it is not designed to capture retail distribution costs. However, the model
implicitly includes distribution losses since net energy for load,8 rather than delivered sales,9 is used to
represent electricity demand in the model. Additionally, the production costs calculated by IPM are the
wholesale production costs. In reporting costs, the model does not include embedded costs, such as
carrying charges of existing units, which may ultimately be part of the retail cost incurred by end-use
consumers.
2.3.5 Load Duration Curves (LDC)
IPM uses Load Duration Curves (LDCs) to provide realism to the dispatching of electric generating units.
Unlike a chronological electric load curve, which is simply an hourly record of electricity demand, the
LDCs are created by rearranging the hourly chronological electric load data from the highest to lowest
(MW) value. In order to aggregate such load detail into a format enabling this scale of power sector
modeling, EPA applications of IPM use a 24-step piecewise linear representation of the LDC.
IPM can include any number of user-defined seasons. A season can be a single month or several
months. EPA Platform v6 contains three seasons: summer (May through September), winter (December
through February), and a winter shoulder season (October, November and March, April). The summer
season corresponds to the ozone season for modeling seasonal NOx policies. The residual seven
months are split into a three-month winter and four-month winter shoulder seasons to better capture
winter peak and seasonality in wind and solar hourly generation profiles. Separate summer, winter, and
winter shoulder season LDCs are created for each of IPM's model regions. Figure 2-1 below presents
side-by-side graphs of a hypothetical chronological hourly load curve and a corresponding load duration
curve for a summer season.
Use of seasonal LDCs rather than annual LDCs allows IPM to capture seasonal differences in the level
and patterns of customer demand for electricity. For example, in most regions air conditioner cycling only
8 Net energy for load is the electrical energy requirements of an electrical system, defined as system net generation,
plus energy received from others, less energy delivered to others through interchange. It includes distribution losses.
9 Delivered sales is the electrical energy delivered under a sales agreement. It does not include distribution losses.
2-7
-------
impacts customer demand patterns during the summer season. The use of seasonal LDCs also allows
IPM to capture seasonal variations in the generation resources available to respond to the customer
demand depicted in an LDC. For example, power exchanges between utility systems may be seasonal in
nature. Some air regulations affecting power plants are also seasonal in nature. This can impact the
type of generating resources that are dispatched during a particular season. Further, because of
maintenance scheduling for individual generating units, the capacity and utilization for these supply
resources also vary between seasons.
Figure 2-1 Hypothetical Chronological Hourly Load Curve and Seasonal Load Duration Curve in
EPA Platform v6 for Summer
Chronological Hourly Load Curve
Seasonal Load Duration Curve
MW
Hours in Season
3672
Hours in Season
3672
In EPA Platform v6, regional forecasts of peak and total electricity demand from AEO 2018 and hourly
load curves from FERC Form 714 and ISO/RTOs10 are used to derive future seasonal load duration
curves for each IPM run year in each IPM region. The results of this process are individualized seasonal
LDCs that capture the unique hourly electricity demand profile of each region. The LDCs change over
time to reflect projected changes in load factors because of future variations in electricity consumption
patterns.11
Within IPM, LDCs are represented by a discrete number of load segments, or generation blocks, as
illustrated in Figure 2-2. EPA Platform v6 uses 24 load segments in its seasonal LDCs. Figure 2-2
illustrates and the following text describes the 24-segment LDCs used in EPA Platform v6. Length of time
and system demand are the two parameters, which define each segment of the load duration curve. The
load segment represents the amount of time (along the x-axis) and the capacity that the electric dispatch
mix must be producing (represented along the y-axis) to meet system load. In EPA Platform v6, the
hours in the LDC are initially clustered into six groups. Group 1 incorporates 1% of all hours in the
season with the highest load. Groups 2 to 6 have 4%, 10%, 30%, 30%, and 25% of the hours with
progressive lower levels of demand. Each of these 6 groups of hours are further separated into four time
of day categories to result in a possible maximum of 24 load segments. This approach better accounts
for the impact of solar generation during periods of high demand. The four time-of-day categories are
8PM - 6AM, 6AM - 9AM, 9AM - 5PM and 5PM - 8PM. Plants are dispatched to meet this load based on
economic considerations and operating constraints. The most cost effective plants are assigned to meet
10 The 2016 load curves are used for IPM model regions in ERCOT. The 2011 load curves are used for all remaining
model regions. For further details, see Section 3.2.3.
11 For further details in regards to the source of the load factors used in EPA Platform v6, see Section 3.2.2.
2-8
-------
load in all 24 segments of the load duration curve. This is discussed in greater detail in section 2.3.6
below.
Table 2-2 contains data of the seasonal 2021 load duration curves in each of the 67 model regions in the
lower continental U.S. for EPA Platform v6.
Figure 2-2 Stylized Depiction of a Six Segment Load Duration Curve Used in EPA Platform v6
Stylized Six Segment Load Duration Curve
Load
(MW)
Segment
Segment 5
Segment 4
Segment 6
Duration (Hours)
Segment 1 Segment 2
2.3.6 Dispatch Modeling
In IPM, the dispatching of electricity is based on the variable cost of generation. In the absence of any
operating constraints, units with the lowest variable cost generate first. The marginal generating unit, i.e.,
the power plant that generates the last unit of electricity, sets the energy price. Physical operating
constraints also influence the dispatch order. For example, IPM uses turndown constraints to prevent
base load units from cycling, i.e., switching on and off. Turndown constraints often override the dispatch
order that would result based purely on the variable cost of generation. Variable costs in combination
with turndown constraints enable IPM to dispatch generation resources in a technically realistic fashion.
Figure 2-3 below depicts a highly stylized dispatch order based on the variable cost of generation of the
resource options included in the EPA Platform v6. In Figure 2-3, two hypothetical load segments are
subdivided according to the type of generation resource that responds to the load requirements
represented in that segment. Notice that the generation resources with the lowest operating cost (i.e.,
hydro and nuclear) respond first to the demand represented in the LDC and are accordingly at the bottom
of "dispatch stack." They are dispatched for the maximum possible number of hours represented in the
LDC because of their low operating costs. Generation resources with the highest operating cost (e.g.,
peaking turbines) are at the top of the "dispatch stack," since they are dispatched last and for the
minimum possible number of hours. In the load segment with non-dispatchable generating capacity such
as solar, the conventional power plants are dispatched to the residual load level where residual load is
defined as the difference between the total load and the load met by non-dispatchable resources.
2-9
-------
Figure 2-3 Stylized Dispatch Order in Illustrative Load Segments
MW
Solar and
Wind
¦—Total Load —
Residual Load
Hydro
Hydro
Hours
Gas Combined
Cycle
Combustion
Turbine
Coal Steam
Oil/Gas
Steam
Nuclear
Gas Combined
Cycle
Combustion
Turbine
Coal Steam
Oil/Gas
Steam
Nuclear
Without Intermittent With Intermittent
Capacity Capacity
Note: Figure 2-3 does not include all the plant types that are modeled in EPA Platform v6. Intermittent
renewable technologies such as wind and solar are considered non-dispatchable and are assigned a
specific hourly generation profile.
2.3.7 Fuel Modeling
Another key methodological feature of IPM is its capability to model the full range of fuels used for electric
power generation. The cost, supply, and (if applicable) quality of each fuel included in the model are
defined during model set-up. Fuel price and supply are represented in EPA Platform v6 in one of two
alternative ways: (1) through a set of supply curves (coal, natural gas, and biomass) or (2) through an
exogenous price stream (fuel oil and nuclear fuel). With the first approach, the model endogenously
determines the price for that fuel by balancing the supply and demand. IPM uses fuel quality information
(e.g., the sulfur, chlorine or mercury content of different types of coal from different supply regions) to
determine the emissions resulting from combustion of that fuel.
EPA Platform v6 includes coal, natural gas, fuel oil, nuclear fuel, biomass, and fossil and non-fossil waste
as fuels for electric generation. The specific base case assumptions for these fuels are examined in
chapters 7 to 9.
2.3.8 Transmission Modeling
IPM includes a detailed representation of existing transmission capabilities between model regions. The
maximum transmission capabilities between regions are specified in IPM's transmission constraints. Due
to uncertainty surrounding the building of new transmission lines in the U.S., EPA Platform v6 does not
exercise IPM's capability to model the building of new transmission lines. However, that capacity of the
model is described here in case it is applied in future analyses. Additions to transmission lines are
represented by decision variables defined for each eligible link and model run year. In IPM's objective
function, the decision variables representing transmission additions are multiplied by new transmission
line investment cost and capital charge rates to obtain the capital cost associated with the transmission
addition. The specific transmission assumptions in EPA Platform v6 are described in section 3.3.
2-10
-------
2.3.9 Perfect Competition and Perfect Foresight
Two key methodological features of IPM are its assumptions of perfect competition and perfect foresight.
The former means that IPM models production activity in wholesale electric markets on the premise that
these markets subscribe to all assumptions of perfect competition. The model does not explicitly capture
any market imperfections such as market power, transaction costs, informational asymmetry or
uncertainty. However, if desired, appropriately designed sensitivity analyses or redefined model
parameters can be used to gauge the impact of market imperfections on the wholesale electric markets.
IPM's assumption of perfect foresight implies that agents know precisely the nature and timing of
conditions in future years that affect the ultimate costs of decisions along the way. For example, under
IPM there is complete foreknowledge of future electricity demand, fuel supplies, and other variables
(including regulatory requirements) that in reality are subject to uncertainty and limited foresight.
Modelers frequently assume perfect foresight in order to establish a decision-making framework that can
estimate cost-minimizing courses of action given the best-guess expectations of these future variables
that can be constructed at the time the projections are made.
2.3.10 Scenario Analysis and Regulatory Modeling
One of the most notable features of IPM is its detailed and flexible modeling features enabling for
scenario analysis involving different outlooks of key drivers of the power sector and environmental
regulations. Treatment of environmental regulations is endogenous in IPM. That is, by providing a
comprehensive representation of compliance options, IPM enables environmental decisions to be made
within the model based on least cost considerations, rather than exogenously imposing environmental
choices on model results. For example, unlike other models that enter allowance prices as an exogenous
input during model set-up, IPM obtains allowance prices as an output of the endogenous optimization
process of finding the least cost compliance options in response to air regulations. (In linear
programming terminology, they are the "shadow prices" of the respective emission constraints — a
standard output produced in solving a linear programming problem.) IPM can capture a wide variety of
regulatory program designs including emissions trading policies, command-and-control policies, and
renewable portfolio standards. IPM's representation of emissions trading policies can include allowance
banking, trading, borrowing, bonus allowance mechanisms, and progressive flow controls. Air regulations
can be tailored to specific geographical regions and can be restricted to specific seasons. Many of these
regulatory modeling capabilities are utilized in EPA Platform v6.
2.4 Hardware and Programming Features
IPM produces model files in standard MPS linear programming format. The model runs on most
PC-platforms. Hardware requirements are highly dependent on the size of a particular model run. For
example, with almost 19.7 million decision variables and 1.32 million constraints, EPA Platform v6 is run
on a 64 bit Enterprise Server - Windows 2008 R2 platform with two Intel Xeon X5675 3.07 GHz
processors and 72 GB of RAM. Due to the size of the EPA Platform v6, a commercial grade solver is
required. (Benchmarking tests performed by EPA's National Environmental Scientific Computing Center
using research grade solvers yielded unacceptable results.) For current EPA applications of IPM, the
FICO Xpress Optimization Suite 8.3 (64 bit with multi-threads barrier and MIP capabilities) linear
programming solvers are used.
Two data processors - a front-end and the post-processing tool - support the model. The front-end
creates the necessary inputs to be used in IPM, while the post-processing tool maps IPM model-plant
level outputs to individual electric generating units (a process called "parsing," see section 2.3.1) and
creates input files in flat-file format as required by EPA's air quality models.
In preparation for a model run, IPM requires an extensive set of input parameters. The input parameters
are discussed in Section 2.5.1 below. Results from a model run are presented in a series of detailed
reports. The reports are described in Section 2.5.2 below.
2-11
-------
2.5 Model Inputs and Outputs
2.5.1 Data Parameters for Model Inputs
IPM requires input parameters that characterize the U.S. electric system, economic outlook, fuel supply
and air regulatory framework. Chapters 3-10 contain detailed discussions of the values assigned to these
parameters in EPA Platform v6. This section simply lists the key input parameters required by IPM:
Electric System
Existing Generating Resources
• Plant Capacities
• Heat Rates
• Fuels Used
• Emissions Limits or Emission Rates for NOx, SO2, HCI, CO2, and Mercury
• Existing Pollution Control Equipment and Retrofit Options
• Availability
• Fixed and Variable O&M Costs
• Minimum Generation Requirements (Turn Down Constraint)
• Output Profile for Non-Dispatchable Resources
New Generating Resources
• Cost and Operating Characteristics
• Resource Limits and Generation Profiles
• Limitations on Availability
Other System Requirements
• Regional Specification
• Inter-regional Transmission Capabilities
• Reserve Margin Requirements for Reliability
• System Specific Generation Requirements
Economic Outlook
Electricity Demand
• Firm Regional Electricity Demand
• Load Curves
Financial Outlook
• Capital Charge Rates
• Discount Rate
Fuel Supply
Fuel Supply Curves for Coal, Gas, and Biomass
• Fuel Price
• Fuel Quality
• Transportation Costs for Coal, Natural Gas, and Biomass
2-12
-------
Regulatory Outlook
Air Regulations for NOx, SO2, HCI, CO2, and Mercury
• Other Air Regulations
• Non-air Regulations (affecting electric generating unit operations)
2.5.2 Model Outputs
IPM produces a variety of output reports. These range from extremely detailed reports, which describe
the results for each model plant and run year, to summary reports, which present results for regional and
national aggregates. Individual topic areas can be included or excluded at the user's discretion.
Standard IPM reports cover the following topics:
• Generation
• Capacity mix
• Capacity additions and retirements
• Capacity and energy prices
• Power production costs (capital, VOM, FOM, and fuel costs)
• Fuel consumption
• Fuel supply and demand
• Fuel prices for coal, natural gas, and biomass
• Emissions (NOx, SO2, HCI, CO2, and Mercury)
• Emission allowance prices
List of tables that are uploaded directly to the web:
Table 2-2 Load Duration Curves used in EPA Platform v6
2-13
-------
3. Power System Operation Assumptions
This section describes the assumptions pertaining to the North American electric power system as
represented in EPA Platform v6.
3.1 Model Regions
EPA Platform v6 models the U.S. power sector in the contiguous 48 states and the District of Columbia
and the Canadian power sector in the 10 provinces (with Newfoundland and Labrador represented as two
regions on the electricity network even though politically they constitute a single province12) as an
integrated network13.
There are 67 IPM model regions covering the U.S. 48 states and District of Columbia. The IPM model
regions are approximately consistent with the configuration of the NERC assessment regions in the
NERC Long-Term Reliability Assessments. These IPM model regions reflect the administrative structure
of regional transmission organizations (RTOs) and independent system operators (ISOs). Further
disaggregation of the NERC assessment regions and RTOs allows a more accurate characterization of
the operation of the U.S. power markets by providing the ability to represent transmission bottlenecks
across RTOs and ISOs, as well as key transmission limits within them.
The IPM regions also provide approximate disaggregation of the regions of the National Energy Modeling
System (NEMS) to provide for a more accurate correspondence with the demand projections of the
Annual Energy Outlook (AEO). Notable disaggregations are further described below:
NERC assessment regions MISO, PJM, and SPP cover the areas of the corresponding RTOs and are
designed to better represent transmission limits and dispatch in each area. In IPM, the MISO area is
disaggregated into 14 IPM regions, PJM assessment area is disaggregated into 9 IPM regions, and SPP
is disaggregated into 5 IPM regions, where the IPM regions are selected to represent planning areas
within each RTO and/or areas with internal transmission limits.
New York is now disaggregated into 8 IPM regions, to better represent flows around New York City and
Long Island, and to better represent flows across New York State from Canada and other U.S. regions.
The NERC assessment region SERC is divided into Kentucky, TVA, AECI, the Southeast, and the
Carolinas. New England is disaggregated into CT, ME, and rest of New England regions. ERCOT is also
disaggregated into three regions.
IPM retains the NERC assessment areas within the overall WECC regions, and further disaggregates
these areas using sub-regions from the WECC Power Supply Assessment.
The 11 Canadian model regions are defined along provincial political boundaries.
Figure 3-1 contains a map showing all the EPA Platform v6 model regions.
Table 3-1 defines the abbreviated region names appearing on the map and gives a crosswalk between
the IPM model regions, the NERC assessment regions, and regions used in the Energy Information
Administration's (ElA's) National Energy Model System (NEMS) that is the basis for ElA's Annual Energy
Outlook (AEO) reports.
12 This results in a total of 11 Canadian model regions being represented in EPA Platform v6.
13 Because United States and the Canadian power markets are being modeled in an integrated manner, IPM can
model the transfer of power in between these two countries endogenously. This transfer of power is limited by the
available transmission capacity in between the two countries. Hence, it is possible for the model to build capacity in
one country to meet demand in the other country when economic and is operationally feasible.
3-1
-------
3.2 Electric Load Modeling
Net energy for load and net internal demand are inputs to IPM that together are used to represent the
grid-demand for electricity. Net energy for load is the projected annual electric grid-demand, prior to
accounting for intra-regional transmission and distribution losses. Net internal demand (peak demand) is
the maximum hourly demand within a given year after removing interruptible demand. Table 3-2 shows
the electricity demand assumptions (expressed as net energy for load) used in EPA Platform v6. It is
based on the net energy for load in AEO 2018.14
Figure 3-1 EPA Platform v6 Model Regions
' NENG_ME
WECC.
WECCJD
MIS_WUMS
NENG_CT
Mi DA
PJMWMAC
PJM_COMD
WECCJJT
WECC_CO
wecc_snv
S_VACA
WKC_AZ
WECC_NM
MIS_AMSO
For purposes of documentation, Table 3-2 and Table 3-3 present the net energy for load on a national
and regional basis respectively. EPA Platform v6 models regional breakdowns of net energy for load in
each of the 67 IPM U.S. regions in the following steps:
• The net energy for load in each of the 22 NEMS electricity regions is taken from the NEMS reference
case.
• NERC balancing areas are assigned to both IPM regions and NEMS regions to determine the share
of the NEMS net energy for load in each NEMS regions that falls into each IPM region. These shares
are calculated in the following steps.
14 The electricity demand in EPA Platform v6 for the U.S. lower 48 states and the District of Columbia is obtained for
each IPM model region by disaggregating the Total Net Energy for Load projected for the corresponding NEMS
Electric Market Module region as reported in the Electricity and Renewable Fuel Tables 73-120 at
http://www.eia.gov/forecasts/aeo/tables ref.cfm.
3-2
-------
• Map the NERC Balancing Authorities/ Planning Areas in the US to the 67 IPM regions.
• Map the Balancing Authorities/ Planning Areas in the US to the 22 NEMS regions.
• Using the 2007 data from FERC Form 714 for non WECC regions and 2011 data for WECC
regions on net energy for load in each of the balancing areas, calculate the proportional share of
each of the net energy for load in 22 NEMS regions that falls in each of the 67 IPM Regions.
• Using these shares of each NEMS region net energy for load that falls in each IPM region, calculate
the total net energy for load for each IPM region from the NEMS regional load in AEO 2018.
Table 3-1 Mapping of NERC Regions and NEMS Regions with EPA Platform v6 Model Regions
AEO 2017 NEMS
NERC Assessment Region
Region
Model Region
Model Region Description
ERCT (1)
ERC_REST
ERCOT_Rest
ERCT (1)
ERC_GWAY
ERCOT_Tenaska Gateway Generating Station
ERCOT
ERCT (1)
ERC_FRNT
ERCOT_Tenaska Frontier Generating Station
ERCT (1)
ERC_WEST
ERCOT_West
ERCT (1)
ERC_PHDL
ERCOT_Panhandle
FRCC
FRCC (2)
FRCC
FRCC
MAPP
MROW (4)
MIS_MAPP
MISO_MT, SD, ND
SRGW (13)
MISJL
MISO_lllinois
RFCW (11), SRCE (15)
MISJNKY
MISO_lndiana (including parts of Kentucky)
MROW (4)
MISJA
MISOJowa
MROW (4)
MIS_MIDA
Ml SOJowa-MidAmerican
RFCM (10)
MIS_LMI
MISO_Lower Michigan
SRGW (13)
MIS_MO
MISO_Missouri
MISO
MROE (3), RFCW (11)
MIS_WUMS
MISO_Wisconsin- Upper Michigan (WUMS)
MROW (4)
MIS_MNWI
MISO_Minnesota and Western Wisconsin
SRDA (12)
MIS_WOTA
MISO_WOTAB (including Western)
SRDA (12)
MIS_AMSO
MISO_Amite South (including DSG)
SRDA (12)
MIS_AR
MISO_Arkansas
SRDA (12)
MIS_D_MS
MISO_Mississippi
SPSO (18)
MIS_LA
MISO_Louisiana
NEWE (5)
NENG_CT
ISONE_Connecticut
ISO-NE
NEWE (5)
NENGREST
ISONE_MA, VT, NH, Rl (Rest of ISO New
England)
NEWE (5)
NENG_ME
ISONE_Maine
NY UP (8)
NY_Z_C&E
NY_Zone C&E
NY UP (8)
NY_Z_F
NY_Zone F (Capital)
NY UP (8)
NY_Z_G-I
NY_Zone G-l (Downstate NY)
NYISO
NYCW (6)
NY_Z_J
NY_Zone J (NYC)
NYLI (7)
NY_Z_K
NY_Zone K (LI)
NY UP (8)
NY_Z_A
NY_Zone A (West)
NY UP (8)
NY_Z_B
NY_Zone B (Genesee)
NY UP (8)
NY_Z_D
NY_Zone D (North)
RFCE (9)
PJM_WMAC
PJM_Western MAAC
RFCE (9)
PJM_EMAC
PJM_EMAAC
PJM
RFCE (9)
PJM_SMAC
PJM_SWMAAC
RFCW (11)
PJM_West
PJM West
RFCW (11)
PJM_AP
PJM_AP
3-3
-------
NERC Assessment Region
AEO 2017 NEMS
Region
Model Region
Model Region Description
RFCW (11)
PJM_COMD
PJM_ComEd
RFCW (11)
PJM_ATSI
PJM_ATSI
SRVC (16)
PJM_Dom
PJM_Dominion
RFCE (9)
PJM_PENE
PJM_PENELEC
SERC-E
SRVC (16)
S_VACA
SERC_VACAR
SRCE (15)
S_C_KY
SERC_Central_Kentucky
SERC-N
SRDA (12)
S_D_AECI
SERC_Delta_AECI
SRCE (15)
S_C_TVA
SERC_Central_TVA
SERC-SE
SRSE (14)
S_SOU
SERC_Southeastern
MROW (4)
SPP_NEBR
SPP Nebraska
SPNO (17), SRGW (13)
SPP_N
SPP North- (Kansas, Missouri)
SPP
SPSO (18)
SPP_KIAM
SPP_Kiamichi Energy Facility
SPSO (18), SRDA (12)
SPP_WEST
SPP West (Oklahoma, Arkansas, Louisiana)
SPSO (18)
SPP_SPS
SPP SPS (Texas Panhandle)
MROW (4)
SPP_WAUE
SPP_WAUE
CAMX (20)
WEC_CALN
WECC_Northern California (not including BANC)
California/Mexico (CA/MX)
CAMX (20)
CAMX (20)
WEC_LADW
WEC_SDGE
WECC_LADWP
WECC_San Diego Gas and Electric
CAMX (20)
WECC_SCE
WECC_Southern California Edison
NWPP (21)
WECC_MT
WECC_Montana
CAMX (20)
WEC_BANC
WECC_BANC
Northwest Power Pool
(NWPP)
NWPP (21)
NWPP (21)
AZNM (19)
WECCJD
WECC_NNV
WECC_SNV
WECCJdaho
WECC_Northern Nevada
WECC_Southern Nevada
NWPP (21)
WECC_UT
WECC_Utah
NWPP (21)
WECC_PNW
WECC_Pacific Northwest
Rocky Mountain Reserve
RMPA (22)
WECC_CO
WECC_Colorado
Group (RMRG)
NWPP (21), RMPA (22)
WECC_WY
WECC_Wyoming
Southwest Reserve Sharing
Group (SRSG)
AZNM (19)
AZNM (19)
AZNM (19)
WECC_AZ
WECC_NM
WECCJID
WECC_Arizona
WECC_New Mexico
WECCJmperial Irrigation District (IID)
CN_AB
Canada_Alberta
CN_BC
Canada_British Columbia
CN_MB
Canada_Manitoba
CN_NB
Canada_New Brunswick
CN_NF
Canada_New Foundland
Canada
CN_NL
CN_PE
CN_NS
CN_ON
CN_PQ
CN_SK
Canada_Labrador
Canada_Prince Edward island
Canada_Nova Scotia
Canada_Ontario
Canada_Quebec
Canada_Saskatchewan
3-4
-------
Table 3-2 Electric Load Assumptions in EPA Platform v6
Year
Net Energy for Load (Billions of kWh)
2021
4,076
2023
4,121
2025
4,167
2030
4,282
2035
4,393
2040
4,542
2045
4,692
2050
4,872
Notes:
The data represents an aggregation of the model-region-specific
net energy loads used in the EPA Platform v6.
Table 3-3 Regional Electric Load Assumptions in EPA Platform v6
IPM Region
Net Energy for Load (Billions of kWh)
2021
2023
2025
2030
2035
2040
2045
2050
ERC_FRNT
0
0
0
0
0
0
0
0
ERC_GWAY
0
0
0
0
0
0
0
0
ERC_PHDL
0
0
0
0
0
0
0
0
ERC_REST
352
360
366
383
400
419
437
456
ERC_WEST
28
29
29
30
32
33
35
36
FRCC
240
243
247
256
267
279
292
308
MIS_AMSO
33
34
35
36
38
40
41
43
MIS_AR
39
40
41
43
45
47
48
50
MIS_D_MS
23
24
24
25
26
27
28
29
MIS_IA
22
22
22
23
24
24
25
26
MIS IL
46
47
47
48
50
51
53
54
MISJNKY
93
94
95
97
100
103
105
109
MIS_LA
48
49
50
52
54
57
59
61
MIS_LMI
102
103
104
106
108
111
114
117
MIS_MAPP
8
8
9
9
9
9
10
10
MIS_MIDA
30
30
31
32
32
34
35
36
MIS_MNWI
90
91
92
95
98
101
104
108
MIS_MO
39
40
40
41
42
43
45
46
MIS_WOTA
35
36
36
38
39
41
43
44
MIS_WUMS
65
66
67
68
70
72
74
76
NENG_CT
30
29
29
29
29
29
29
29
NENG_ME
10
10
10
10
10
10
10
10
NENGREST
77
76
76
76
75
75
76
77
NY_Z_A
16
16
16
16
16
16
16
16
NY Z B
10
10
10
10
10
10
10
10
3-5
-------
IPM Region
Net Energy for Load (Billions of kWh)
2021
2023
2025
2030
2035
2040
2045
2050
NY_Z_C&E
25
25
24
24
24
24
25
25
NY_Z_D
7
7
7
6
6
7
7
7
NY_Z_F
12
12
12
12
12
12
12
12
NY_Z_G-I
19
19
18
18
18
18
18
19
NY_Z_J
47
47
47
46
45
45
46
47
NY_Z_K
20
20
20
20
19
20
20
20
PJM_AP
45
46
46
48
49
50
51
53
PJM_ATSI
67
68
68
70
72
74
76
78
PJM_COMD
98
98
99
102
104
107
110
113
PJM_Dom
97
99
101
105
109
114
118
124
PJM_EMAC
138
139
139
140
142
145
148
153
PJM_PENE
17
17
17
17
17
18
18
19
PJM_SMAC
63
63
64
64
65
66
68
70
PJM_West
203
205
208
213
218
224
230
237
PJM_WMAC
55
55
55
56
57
58
59
61
S_C_KY
31
32
33
34
35
36
37
39
S_C_TVA
173
176
180
186
192
199
205
213
S_D_AECI
18
18
18
18
19
19
20
21
S_SOU
238
242
247
257
265
276
287
299
S_VACA
224
228
232
242
251
262
273
285
SPP_KIAM
0
0
0
0
0
0
0
0
SPP_N
71
72
73
75
77
80
82
86
SPP_NEBR
34
34
35
36
37
38
39
40
SPP_SPS
29
30
30
31
33
34
36
37
SPP_WAUE
23
23
24
24
25
26
27
27
SPP_WEST
129
131
134
140
146
153
159
166
WEC_BANC
14
14
14
14
14
14
14
15
WEC_CALN
111
110
109
108
107
109
111
116
WEC_LADW
27
27
27
26
26
27
27
28
WEC_SDGE
21
21
21
21
21
21
21
22
WECC_AZ
91
92
93
96
100
105
109
115
WECC_CO
66
67
69
71
74
77
81
85
WECCJD
22
23
23
23
23
24
25
26
WECCJID
4
4
4
4
4
5
5
5
WECC_MT
13
13
13
13
13
14
14
15
WECC_NM
24
24
24
25
26
27
29
30
WECC_NNV
13
13
13
13
13
13
14
14
WECC PNW
173
174
174
176
179
185
191
199
3-6
-------
IPM Region
Net Energy for Load (Billions of kWh)
2021
2023
2025
2030
2035
2040
2045
2050
WECC_SCE
108
108
107
106
105
106
109
113
WECC_SNV
27
27
28
29
30
31
32
34
WECCJJT
28
28
28
28
29
30
31
32
WECC_WY
17
18
18
18
18
19
20
21
3.2.1 Demand Elasticity
EPA Platform v6 has the capability to consider endogenously the relationship of the price of power to
electricity demand. However, the capability is exercised only for sensitivity analyses where different price
elasticities of demand are specified for purposes of comparative analysis. The default assumption is that
the electricity demand shown in Table 3-2, which was derived from EIA modeling that already considered
price elasticity of demand, must be met as IPM solves for least-cost electricity supply. This approach
maintains a consistent expectation of future load between the EPA Platform and the corresponding EIA
Annual Energy Outlook reference case (e.g., between EPA Platform v6 and the AEO 2018 reference
case).
3.2.2 Net Internal Demand (Peak Demand)
EPA Platform v6 has separate regional winter, winter shoulder, and summer peak demand values, as
derived from each region's seasonal load duration curve (found in Table 2-2). Peak projections for the
2021-2027 period were estimated based on NERC ES&D 2017 load factors15, and the estimated energy
demand projections shown in Table 3-3. For post 2027 years when NERC ES&D 2017 load factors were
not available, the NERC ES&D 2017 load factors for 2027 were projected forward using growth factors
embedded in the AEO 2018 load factor projections.
Table 3-4 illustrates the national sum of each region's seasonal peak demand and Table 3-20 presents
each region's seasonal peak demand. Because each region's seasonal peak demand need not occur at
the same time, the national peak demand is defined as non-coincidental (i.e., national peak demand is a
summation of each region's peak demand at whatever point in time that region's peak occurs across the
given time period).
Table 3-4 National Non-Coincidental Net Internal Demand
Year
Peak Demand (GW)
Winter
Winter Shoulder
Summer
2021
653
586
769
2023
660
592
776
2025
669
599
786
2030
690
618
812
2035
714
638
843
2040
745
664
880
2045
779
692
923
2050
818
724
972
Notes:
This data is an aggregation of the model-region-specific peak demand loads.
15 Load factors can be calculated at the NERC assessment region level based on the NERC ES&D 2017 projections
of net energy for load and net internal demand. All IPM regions that map to a particular NERC assessment region
are assigned the same load factors. In instances where sub regional level load factor details could be estimated in
selected ISO/RTO zones, those load factors were assigned to the associated IPM region.
3-7
-------
3.2.3 Regional Load Shapes
As of 2013, EPA has adopted year 2011 as the meteorological year in its air quality modeling. In order
for EPA Platform v6 to be consistent, the year 2011 was selected as the "normal weather year"16 for all
IPM regions except for ERCOT, where 2016 data was used. The proximity of the 2011 cumulative annual
heating degree days (HDDs) and cooling degree days (CDDs) to the long-term average cumulative
annual HDDs and CDDs over the period 1981 to 2010 was estimated and found to be reasonably close.
The 2011 and 2016 chronological hourly load data were assembled by aggregating individual utility load
curves taken from Federal Energy Regulatory Commission Form 714 data and individual ISOs and RTOs.
3.3 Transmission
The United States and Canada can be broken down into several power markets that are interconnected
by a transmission grid. As discussed earlier, EPA Platform v6 characterizes the US lower 48 states, the
District of Columbia, and Canada into 78 different model regions by means of 64 power market regions
and 3 power switching regions17 in the US and 11 power market regions in Canada. EPA Platform v6
includes explicit assumptions regarding the transmission grid connecting these modeled power markets.
This section details the assumptions about the transfer capabilities, wheeling costs and inter-regional
transmission used in EPA Platform v6.
3.3.1 Inter-regional Transmission Capability
Table 3-2118 shows the firm and non-firm Total Transfer Capabilities (TTCs) between model regions.
TTC is a metric that represents the capability of the power system to import or export power reliably from
one region to another. The purpose of TTC analysis is to identify the sub-markets created by key
commercially significant constraints. Firm TTCs, also called Capacity TTCs, specify the maximum power
that can be transferred reliably, even after the contingency loss of a single transmission system element
such as a transmission line or a transformer (a condition referred to as N-1, or "N minus one"). Firm
TTCs provide a high level of reliability and are used for capacity transfers. Non-firm TTCs, also called
Energy TTCs, represent the maximum power that can be transferred reliably when all facilities are under
normal operation (a condition referred to as N-0, or "N minus zero"). They specify the sum of the
maximum firm transfer capability between sub-regions and incremental curtailable non-firm transfer
capability. Non-firm TTCs are used for energy transfers since they provide a lower level of reliability than
Firm TTCs, and transactions using Non-firm TTCs can be curtailed under emergency or contingency
conditions.
The amount of energy and capacity transferred on a given transmission link is modeled on a seasonal
basis for all run years in the EPA Platform v6. All of the modeled transmission links have the same Total
Transfer Capabilities for all seasons, which means that the maximum firm and non-firm TTCs for each link
is the same for winter, winter shoulder, and summer. The maximum values for firm and non-firm TTCs
were obtained from public sources such as market reports and regional transmission plans, wherever
available. Where public sources were not available, the maximum values for firm and non-firm TTCs are
based on ICF's expert view. ICF analyzes the operation of the grid under normal and contingency
16 The term "normal weather year" refers to a representative year whose weather is closest to the long-term (e.g., 30
year) average weather. The selection of a "normal weather year" can be made, for example, by comparing the
cumulative annual heating degree days (HDDs) and cooling degree days (CDDs) in a candidate year to the long-term
average. For any individual day, heating degree days indicate how far the average temperature fell below 65
degrees F; cooling degree days indicate how far the temperature averaged above 65 degrees F. Cumulative annual
heating and cooling degree days are the sum of all the HDDs and CDDs, respectively, in a given year.
17 Power switching regions are regions with no market load that represent individual generating facilities specifically
configured so they can sell directly into either ERCOT or SPP: these plants are implemented in IPM as regions with
transmission links only to ERCOT and to SPP.
18 In the column headers in Table 3-21, the term "Energy TTC (MW)" is equivalent to non-firm TTCs and the term
"Capacity TTC (MW)" is equivalent to firm TTCs.
3-8
-------
conditions, using industry-standard methods, and calculates the transfer capabilities between regions.
ICF uses standard power flow data developed by the market operators, transmission providers, or utilities,
as appropriate.
Furthermore, each transmission link between model regions shown in Table 3-21 represents a one-
directional flow of power on that link. This implies that the maximum amount of flow of power possible
from region A to region B may be more or less than the maximum amount of flow of power possible from
region B to region A, due to the physical nature of electron flow across the grid.
3.3.2 Joint Transmission Capacity and Energy Limits
Table 3-5 shows the annual joint limits to the transmission capabilities between model regions, which are
identical for the firm (capacity) and non-firm (energy) transfers. The joint limits were obtained from public
sources where available, or based on ICF's expert view. A joint limit represents the maximum
simultaneous firm or non-firm power transfer capability of a group of interfaces. It restricts the amount of
firm or non-firm transfers between one model region (or group of model regions) and a different group of
model regions. For example, the New England market is connected to the New York market by four
transmission links.
Table 3-21, the transfer capabilities from New England to New York for the individual links are:
• NENG_CT to NY_Z_G-I: 600 MW
• NENGREST to NY_Z_F: 800 MW
• NENG_CT to NY_Z_K: 760 MW
Without any simultaneous transfer limits, the total transfer capability from New England to New York
would be 2,160 MW. However, current system conditions and reliability requirements limit the total
simultaneous transfers from New England to New York to 1,730 MW. ICF uses joint limits to ensure that
this and similar reliability limits are not violated. Therefore, each individual link can be utilized to its limit
as long as the total flow on all links does not exceed the joint limit.
Table 3-5 Annual Joint Capacity and Energy Limits to Transmission Capabilities between Model
Regions in EPA Platform v6
Region Connection
Transmission Path
Capacity
TTC
(MW)
Energy
TTC
(MW)
NY_Zone G-l (Downstate NY) & NY_Zone J (NYC) to NY_Zone K (LI)
NY_Z_G-I to NY_Z_K
NY Z J to NY Z K
1,528
NY_Zone K(LI) to NY_Zones G-l (Downstate NY) & NY_Zone J (NYC)
NY_Z_K to NY_Z_G-I
NY Z K to NY Z J
282
ISONEto NYISO
NENG_CT to NY_Z_G-I
NENGREST to NY_Z_F
NENG CT to NY Z K
1,730
NYISO to ISO NE
NY_Z_G-I to NENG_CT
NY_Z_F to NENGREST
NY_Z_K to NENG_CT
1,730
PJM West & PJM_PENELEC & PJM_AP to PJM_ATSI
PJM_West to PJM_ATSI
PJM_PENE to PJM_ATSI
PJM AP to PJM ATSI
7,881
12,000
PJM_ATSI to PJM West & PJM_PENELEC & PJM_AP
PJM ATSI to PJM West
7,881
12,000
3-9
-------
Region Connection
Transmission Path
Capacity
TTC
(MW)
Energy
TTC
(MW)
PJM_ATSI to PJM_PENE
PJM ATS I to PJM AP
PJM_West & PJM_Dominion to SERC VACAR
PJM_West to S_VACA
PJM DomtoS VACA
2,208
3,424
SERC VACAR to PJM_West & PJM_Dominion
S_VACA to PJM_West
S VACA to PJM Dom
2,208
3,424
MIS_MAPP & SPP_WAUE to MIS_MNWI
MIS_MAPP to MIS_MNWI
SPP WAUEto MIS MNWI
3,000
5,000
MIS_MNWI to MIS_MAPP & SPP_WAUE
MIS_MNWI to MIS_MAPP
MIS MNWI to SPP WAUE
3,000
5,000
SERC_Central_TVA & SERC_Central_Kentucky to PJM West
S_C_TVA to PJM_West
S C KYto PJM West
3,000
4,500
PJM West to SERC_Central_TVA & SERC_Central_Kentucky
PJM_West to S_C_TVA
PJM West to S C KY
3,000
4,500
MISJNKYto PJM_COMD & PJM_West
MISJNKY to PJM_COMD
MIS INKY to PJM West
4,586
6,509
PJM_COMD & PJM_West to MlS_ INKY
PJM_COMD to MIS_INKY
PJM West to MIS INKY
5,998
8,242
3.3.3 Transmission Link Wheeling Charge
Transmission wheeling charge is the cost of transferring electric power from one region to another using
the transmission link. The EPA Platform v6 has no charges within individual IPM regions and no charges
between IPM regions that fall within the same RTO. Charges between other regions vary to reflect the
cost of wheeling. The wheeling charges in 2016 mills/kWh are shown in Table 3-21 in the column labeled
"Transmission Tariff'.
3.3.4 Transmission Losses
The EPA Platform v6 assumes a 2.8 percent inter-regional transmission loss of energy transferred in the
WECC interconnect and 2.4 percent inter-regional transmission loss of energy transferred in ERCOT and
Eastern interconnects. This is based on average loss factors calculated from standard power flow data
developed by the transmission providers.
3.4 International Imports
The US electric power system is connected with the transmission grids in Canada and Mexico and the
three countries actively trade in electricity. The Canadian power market is endogenously modeled in EPA
Platform v6 but Mexico is not. International electric trading between the US and Mexico is represented by
an assumption of net imports based on information from AEO 2017. Table 3-6 summarizes the
assumptions on net imports into the US from Mexico.
Table 3-6 International Electricity Imports (billions kWh) in EPA Platform v6
2021
2023
2025
2030
2035
2040
2045
2050
Net Imports from Mexico 6.34
6.34
6.34
6.34
6.34
6.34
6.34
6.34
Note 1: Source: AEO 2017
Note 2: Imports & exports transactions from Canada are endogenously modeled in IPM.
3-10
-------
3.5 Capacity, Generation, and Dispatch
While the capacity of existing units is an exogenous input into IPM, the dispatch of those units is an
endogenous decision that the model makes. The capacity of existing generating units included in EPA
Platform v6 can be found in the National Electrical Energy Data System (NEEDS v6), a database which
provides IPM with information on all currently operating and planned-committed electric generating units.
NEEDS v6 is discussed in full in Chapter 4.
A unit's generation over a time period is defined by its dispatch pattern over that duration of time. IPM
determines the optimal economic dispatch profile given the operating and physical constraints imposed
on the unit. In EPA Platform v6, unit specific operational and physical constraints are represented
through availability and turndown constraints. However, for some unit types, capacity factors are used to
capture the resource or other physical constraints on generation. The two cases are discussed in more
detail in the following sections.
3.5.1 Availability
Power plant availability is the percentage of time that a generating unit is available to provide electricity to
the grid. Availability takes into account both scheduled maintenance and forced outages; it is formally
defined as the ratio of a unit's available hours adjusted for derating of capacity (due to partial outages) to
the total number of hours in a year when the unit was in an active state. For most types of units in IPM,
availability parameters are used to specify an upper bound on generation to meet demand. Table 3-7
summarizes the availability assumptions used in EPA Platform v6. They are based on data from NERC
Generating Availability Data System (GADS) 2011-2015 and AEO 2017. NERC GADS summarizes the
availability data by plant type and size class. Unit level availability assignments in EPA Platform v6 are
made based on the unit's plant type and size as presented in NEEDS v6. Table 3-27 shows the
availability assumptions for all generating units in EPA Platform v6.
Table 3-7 Availability Assumptions in EPA Platform v6
Unit Type
Annual Availability (%)
Biomass
Coal Steam
Combined Cycle
Combustion Turbine
Energy Storage
Fossil Waste
Fuel Cell
Geothermal
Hydro
IGCC
Landfill Gas
Municipal Solid Waste
Non-Fossil Waste
Nuclear
Oil/Gas Steam
Offshore Wind
Onshore Wind
Pumped Storage
Solar PV
Solar Thermal
69-89
83
76-85
85
84-91
90
90
87
87
79-84
79-85
90
90
90
75-97
95
95
82
90
90
Notes:
Values shown are a range of all of the values modeled
within the EPA Platform v6.
3-11
-------
In the EPA Platform v6, separate (seasonal winter, winter shoulder, and summer) availabilities are
defined. For the fossil and nuclear unit types shown in Table 3-27, seasonal availabilities differ only in
that no planned maintenance is assumed to be conducted during the onpeak- summer (June, July, and
August) months for summer peaking regions and onpeak - winter (December, January, and February)
months for winter peaking regions. Characterizing the availability of hydro, solar, and wind technologies
is more complicated due to the seasonal and locational variations of the resources. The procedures used
to represent seasonal variations in hydro are presented in section 3.5.2 and of wind and solar in section
4.4.5.
3.5.2 Capacity Factor
Generation from certain types of units is constrained by resource limitations. These technologies include
hydro, wind, and solar. For such technologies, IPM uses capacity factors or generation profiles, not
availabilities, to define the upper bound on the generation obtainable from the unit. The capacity factor is
the percentage of the maximum possible power generated by the unit. For example, a photovoltaic solar
unit would have a capacity factor of 27% if the usable sunlight were only available that percent of the
time. For such units, explicit capacity factors or generation profiles mimic the resource availability. The
seasonal capacity factor assumptions for hydro facilities contained in Table 3-8 were derived from EIA
Form-923 data for the 2007-2016 period. A discussion of capacity factors and generation profiles for
wind and solar technologies is contained in section 4.4.5 and Table 4-20, Table 4-22, Table 4-24, Table
4-26, Table 4-46 and Table 4-47.
Table 3-8 Seasonal Hydro Capacity Factors (%) in EPA Platform v6
Model
Region
Winter Capacity
Factor
Winter Shoulder
Capacity Factor
Summer Capacity
Factor
Annual Capacity
Factor
ERC_REST
10%
11%
17%
13%
FRCC
51%
42%
35%
42%
MIS_AR
44%
40%
46%
43%
MISJA
42%
48%
57%
50%
MIS IL
56%
61%
60%
59%
MISJNKY
70%
76%
84%
78%
MIS_LA
62%
56%
64%
61%
MIS_LMI
61%
76%
48%
60%
MIS_MAPP
76%
76%
84%
79%
MIS_MIDA
26%
29%
32%
29%
MIS_MNWI
47%
57%
62%
57%
MIS_MO
36%
43%
55%
47%
MIS_WOTA
20%
20%
20%
20%
MIS_WUMS
51%
62%
54%
56%
NENG_CT
41%
42%
37%
40%
NENG_ME
65%
58%
57%
59%
NENGREST
39%
43%
33%
38%
NY_Z_A
70%
66%
63%
66%
NY_Z_B
35%
31%
24%
29%
NY_Z_C&E
53%
52%
51%
52%
NY_Z_D
71%
75%
79%
76%
NY_Z_F
55%
54%
49%
52%
NY_Z_G-I
34%
34%
33%
33%
PJM_AP
64%
56%
50%
55%
PJM_ATSI
17%
20%
25%
21%
PJM_COMD
38%
42%
50%
44%
3-12
-------
Model
Region
Winter Capacity
Factor
Winter Shoulder
Capacity Factor
Summer Capacity
Factor
Annual Capacity
Factor
PJM_Dom
24%
19%
15%
18%
PJM_EMAC
44%
40%
24%
35%
PJM_PENE
58%
57%
36%
48%
PJM_West
34%
31%
29%
31%
PJM_WMAC
41%
40%
23%
33%
S_C_KY
31%
25%
22%
25%
S_C_TVA
52%
36%
30%
37%
S_D_AECI
13%
18%
21%
18%
S_SOU
30%
22%
16%
21%
S_VACA
27%
20%
17%
20%
SPP_N
13%
16%
20%
17%
SPP_NEBR
30%
34%
43%
37%
SPP_WAUE
32%
34%
43%
37%
SPP_WEST
26%
26%
32%
29%
WEC_BANC
16%
19%
31%
23%
WEC_CALN
21%
26%
40%
31%
WEC_LADW
12%
13%
21%
16%
WEC_SDGE
25%
30%
49%
37%
WECC_AZ
27%
28%
32%
29%
WECC_CO
30%
24%
34%
30%
WECCJD
31%
32%
46%
38%
WECC_IID
30%
37%
61%
45%
WECC_MT
37%
37%
50%
43%
WECC_NM
23%
24%
32%
27%
WECC_NNV
38%
49%
55%
49%
WECC_PNW
44%
41%
45%
43%
WECC_SCE
19%
25%
46%
32%
WECC_SNV
19%
24%
26%
24%
WECC_UT
28%
29%
39%
33%
WECC WY
15%
22%
53%
34%
Note: Annual capacity factor is provided for information purposes only. It is not used in modeling.
Capacity factors are also used to define the upper bound on generation obtainable from nuclear units.
This rests on the assumption that nuclear units will dispatch to their availability, and, consequently,
capacity factors and availabilities are equivalent. The capacity factors (and, consequently, the
availabilities) of existing nuclear units in EPA Platform v6 vary from region to region and over time.
Further discussion of the nuclear capacity factor assumptions in EPA Platform v6 is contained in Section
4.5.
In EPA Platform v6, capacity factors for oil/gas steam units are treated separately and assigned minimum
capacity factors under certain conditions. These minimum capacity factor constraints reflect stakeholder
comments that if left unconstrained, IPM does not project as much operation from oil/gas steam units as
stakeholders expect will continue to occur based on observed market outcomes to date. These
comments note that these units often operate due to local transmission constraints, unit-specific grid
reliability requirements, or other drivers that are not captured in EPA's modeling. EPA examined its
modeling treatment of these units and introduced minimum capacity factor constraints to reflect better the
real-world behavior of these units where drivers of that behavior are not fully represented in the model
itself. This approach is designed to balance the continued operation of these units in the near term while
also allowing economic forces to influence decision-making over the modeling time horizon; as a result,
3-13
-------
the minimum capacity factor limitations are imposed for limited time horizons (and are terminated even
earlier if the capacity in question reaches 60 years of age). Historical operational data indicate that
oil/gas steam units with high capacity factors have maintained a high level of generation over many years;
in order to reflect persistent operation of these units, minimum capacity factors for higher capacity factor
units are phased out more slowly than those constraints for lower capacity factor units. The steps in
assigning these capacity constraints are as follows:
1) For each oil/gas steam unit, calculate an annual capacity factor over a ten-year baseline (2007-2016).
2) Identify the minimum capacity factor over this baseline period for each unit.
3) Terminate the constraints in the earlier of (a) the run-year in which the unit reaches 60 years of age,
or (b) based on the assigned minimum capacity factor and the model year indicated in the following
schedule:
• For model year 2021,
• For model year 2023,
• For model year 2025,
• For model year 2030,
• For model year 2035,
• For model year 2040,
3.5.3 Turndown
Turndown assumptions in EPA Platform v6 are used to prevent coal and oil/gas steam units from
operating strictly as peaking units, which would be inconsistent with their operating capabilities.
Specifically, the turndown constraints in EPA Platform v6 require coal steam and oil/gas steam units to
dispatch no less than a fixed percentage of the unit capacity in the 23 base and mid-load segments of the
load duration curve in order to dispatch 100% of the unit in the peak load segments of the LDC. Oil/gas
steam units are required to dispatch no less than 25% of the unit capacity in the 23 base and mid-load
segments of the LDC in order to dispatch 100% of the unit capacity in the peak load segment of the LDC.
The unit level turndown percentages for coal units were estimated based on a review of recent hourly Air
Markets Program Data (AMPD) data and are shown in Table 3-22.
3.6 Reserve Margins
A reserve margin is a measure of the system's generating capability above the amount required to meet
the net internal demand (peak load) requirement. It is defined as the difference between total dependable
capacity and annual system peak load divided by annual system peak load. The reserve margin capacity
contribution for renewable units is described in Section 4.4.5; the reserve margin capacity contribution for
other units is the capacity in the NEEDS for existing units or the capacity build by IPM for new units. In
practice, each NERC region has a reserve margin requirement, or comparable reliability standard, which
is designed to encourage electric suppliers in the region to build beyond their peak requirements to
ensure the reliability of the electric generation system within the region.
In IPM, reserve margins are used to depict the reliability standards that are in effect in each NERC region.
Individual reserve margins for each NERC region are derived either directly or indirectly from NERC's
electric reliability reports. They are based on reliability standards such as loss of load expectation
(LOLE), which is defined as the expected number of days in a specified period in which the daily peak
load will exceed the available capacity. These margins are imposed throughout the entire time horizon.
EPA Platform v6 reserve margin assumptions are shown in Table 3-9.
remove minimum constraint from units with capacity factor < 5%
remove minimum constraint from units with capacity factor < 10%
remove minimum constraint from units with capacity factor < 15%
remove minimum constraint from units with capacity factor < 25%
remove minimum constraint from units with capacity factor < 35%
remove minimum constraint from units with capacity factor < 45%
3-14
-------
Table 3-9 Planning Reserve Margins in EPA Platform v6
Model Region
Reserve Margin
CN_AB
11.0%
CN_BC
12.1%
CN_MB
12.0%
CN_NB
20.0%
CN_NF
20.0%
CN_NL
20.0%
CN_NS
20.0%
CN_ON
17.00%
CN_PE
20.0%
CN_PQ
12.70%
CN_SK
11.00%
ERC_FRNT
13.8%
ERC_GWAY
13.8%
ERC_PHDL
13.8%
ERC_REST
13.8%
ERC_WEST
13.8%
FRCC
18.6%
MIS_AR
15.2%
MIS_D_MS
15.2%
MIS_IA
15.2%
MIS IL
15.2%
MISJNKY
15.2%
MIS_LA
15.2%
MIS_LMI
15.2%
MIS_MAPP
15.2%
MIS_MIDA
15.2%
MIS_MNWI
15.2%
MIS_MO
15.2%
MIS_AMSO
15.2%
MIS_WOTA
15.2%
MIS_WUMS
15.2%
NENG_CT
15.9%
NENG_ME
15.9%
NENGREST
15.9%
NY_Z_A
15.0%
NY_Z_B
15.0%
NY_Z_C&E
15.0%
NY_Z_D
15.0%
NY Z F
15.0%
Model Region
Reserve Margin
NY_Z_G-I
15.0%
NY_Z_J
15.0%
NY_Z_K
15.0%
PJM_AP
16.5%
PJM_ATSI
16.5%
PJM_COMD
16.5%
PJM_Dom
16.5%
PJM_EMAC
16.5%
PJM_PENE
16.5%
PJM_SMAC
16.5%
PJM_West
16.5%
PJM_WMAC
16.5%
S_C_KY
15.0%
S_C_TVA
15.0%
S_D_AECI
15.0%
S_SOU
15.0%
S_VACA
15.0%
SPP_KIAM
12.0%
SPP_N
12.0%
SPP_NEBR
12.0%
SPP_SPS
12.0%
SPP_WAUE
12.0%
SPP_WEST
12.0%
WEC_BANC
16.3%
WEC_CALN
16.2%
WEC_LADW
16.2%
WEC_SDGE
16.2%
WECC_AZ
15.8%
WECC_CO
14.1%
WECC_ID
16.3%
WECC_IID
15.8%
WECC_MT
16.3%
WECC_NM
15.8%
WECC_NNV
16.3%
WECC_PNW
16.3%
WECC_SCE
16.2%
WECC_SNV
16.3%
WECC_UT
16.3%
WECC WY
14.1%
3.7 Power Plant Lifetimes
EPA Platform v6 does not include any pre-specified assumptions about power plant lifetimes (i.e., the
duration of service allowed) except for nuclear units. All conventional fossil units (coal, oil/gas steam,
combustion turbines, and combined cycle), nuclear and biomass units can be retired during a model run if
their retention is deemed uneconomic.
Nuclear Retirement at Age 80: EPA Platform v6 assumes that commercial nuclear reactors will be
retired upon license expiration, which includes two 20-year operating extensions that are assumed to be
3-15
-------
granted for each reactor by the Nuclear Regulatory Commission (NRC). EPA Platform v6 assumes an
80-year life. EPA Platform v6 incorporates life extension costs to enable these operating life extensions.
(See Sections 4.2.8 and 4.5)
a:
5
o
9 T
* 3 -
o
(0
Q.
«
O
Figure 3-2 Scheduled Retirements of Existing Nuclear Capacity
Under 80-Year Life Assumption
Impact of 80-Year Lifetime on Existing Nuclear Fleet
T 120
2026 2031 2036 2041 2050 2055 2060 2065 2070
Year
I Capacity Retired Capacity Remaining
3.8 Heat Rates
Heat rates, expressed in British thermal units (Btus) per kilowatt-hour (kW-hr), are a measure of an
Electric Generating Unit's (EGU's) generating efficiency. As in previous versions of NEEDS, it is
assumed in NEEDS v6 that, with the exception of deploying the heat rate improvement option described
below, heat rates of existing EGUs remain constant overtime. This assumption reflects two offsetting
factors:
1. Plant efficiencies tend to degrade over time, and
2. Increased maintenance and component replacement act to maintain, or improve, an
EGU's generating efficiency.
The heat rates for the model plants in EPA Platform v6 are based on values from Annual Energy Outlook
2017 (AEO 2017) informed by fuel use and net generation data reported on Form EIA-923. These values
were screened and adjusted using a procedure developed by EPA (as described below) to ensure that
the heat rates used in EPA Platform v6 are within the engineering capabilities of the various EGU types.
Based on engineering analysis, the upper and lower heat rate limits shown in Table 3-10 were applied to
coal steam, oil/gas steam, combined cycle, combustion turbine, and Internal Combustion (IC) engines. If
the reported heat rate for such a unit was below the applicable lower limit or above the upper limit, the
upper or lower limit was substituted for the reported value.
3-16
-------
Table 3-10 Lower and Upper Limits Applied to Heat Rate Data in EPA Platform v6
Plant Type
Heat Rate (Btu/kWh)
Lower Limit Upper Limit
Coal Steam
8
300
14
500
Oil/Gas Steam
8
300
14
500
Combined Cycle - Natural Gas
5
500
15
000
Combined Cycle - Oil
6
000
15
000
Combustion Turbine - Natural Gas - 80 MW and above
8
700
18
700
Combustion Turbine - Natural Gas < 80 MW
8
700
36
800
Combustion Turbine - Oil and Oil/Gas - 80 MW and above
6
000
25
000
Combustion Turbine - Oil and Oil/Gas < 80 MW
6
000
36
800
IC Engine - Natural Gas
8
700
18
000
IC Engine - Oil and Oil/Gas - 5 MW and above
8
700
20
500
IC Engine - Oil and Oil/Gas < 5 MW
8
700
42
000
EPA Platform v6 is capable of offering to coal steam model plants a heat rate improvement option that is
fully integrated into the Integrated Planning Model (IPM) framework. This capability enables IPM to
determine economic uptake of heat rate improvements at each model plant, and it can be activated or
deactivated as an investment option in any given scenario analyzed in IPM. Note that the heat rate
improvement option is deactivated in EPA Platform v6, and is assumed to remain deactivated unless
otherwise noted in EPA analyses using EPA Platform v6.
As an EGU's heat rate improves, less fuel is needed to produce the same amount of electricity. Because
less fuel is combusted to produce the same amount of electricity, pollutant emissions are reduced per
kW-hr of electricity produced. Furthermore, heat rate improvement has accompanying economic benefits,
such as reducing fuel costs associated with generating the same amount of electricity. EPA is aware that
a variety of technical approaches has been applied at existing coal steam EGUs to reduce auxiliary power
consumption and fuel consumption and thereby increase net electrical output per unit of heat input. Heat
rate improvement studies have examined opportunities for efficiency improvements as a means of
reducing heat rate and regulating air pollutant emissions from coal-fired power plants. EPA is also aware
that a diverse range of factors affects site-specific EGU heat rate improvements. Heat rate improvement
cost and performance assumptions will be documented for any scenario analysis that activates the heat
rate improvement option, and EPA welcomes further technical engagement on that option accordingly.
3.9 Existing Environmental Regulations
This section describes the existing federal, regional, and state SO2, NOx, mercury, HCI and CO2
emissions regulations that are represented in the EPA Platform v6. EPA Platform v6 also includes three
non-air federal rules affecting EGUs: Cooling Water Intakes (316(b)) Rule and Coal Combustion
Residuals from Electric Utilities (CCR), both promulgated in 2014, and the Effluent Limitations and
Guidelines Rule finalized in 2015. The first four subsections discuss national and regional regulations.
The next four subsections describe state level environmental regulations, a variety of legal settlements,
emission assumptions for potential units and renewable portfolio standards. Finally, the NY minimum oil
rule follows the subsection presenting the Canadian regulations for CO2 and renewables.
3.9.1 SO2 Regulations
Unit-level Regulatory SO2 Emission Rates and Coal Assignments: Before discussing the national and
regional regulations affecting SO2, it is important to note that unit-level SO2 permit rates including SO2
regulations arising out of State Implementation Plan (SIP) requirements, which are not only state-specific
but also county-specific, are captured at model set-up in the coal choices given to coal fired existing units
3-17
-------
in EPA Platform v6. Since SO2 emissions are dependent on the sulfur content of the fuel used, the SO2
permit rates are used in IPM to define fuel capabilities.
For instance, a unit with a SO2 permit rate of 3.0 Ibs/MMBtu would be provided only with those
combinations of fuel choices and SO2 emission control options that would allow the unit to achieve an out-
of-stack rate of 3.0 Ibs/MMBtu or less. If the unit finds it economical, it may elect to burn a fuel that would
achieve a lower SO2 rate than its specified permit limit. In EPA Platform v6, there are six different sulfur
grades of bituminous coal, four different grades of subbituminous coal, four different grades of lignite, and
one sulfur grade of residual fuel oil. There are two different SO2 scrubber options and one DSI option for
coal units. Further discussion of fuel types and sulfur content is contained in Chapter 7. Further
discussion of SO2 control technologies is contained in Chapter 5.
National and Regional SO2 Regulations: The national program affecting SO2 emissions in EPA Platform
v6 is the Acid Rain Program established under Title IV of the Clean Air Act Amendments (CAAA) of 1990,
which set a goal of reducing annual SO2 emissions by 10 million tons below 1980 levels. The program,
which became operational in year 2000, affects all SO2 emitting electric generating units greater than 25
MWs. The program provides trading and banking of allowances overtime across all affected electric
generation sources.
The annual SO2 caps over the modeling time horizon in EPA Platform v6 reflect the provisions in Title IV.
For allowance trading programs like the Acid Rain Program that allow banking of unused allowances over
time, we usually estimate an allowance bank that is assumed to be available by the first year of the
modeling horizon (which is 2021 in EPA Platform v6). However, the Acid Rain Program has
demonstrated a substantial oversupply of allowances that continues to grow overtime, and we anticipate
projecting that the program's emission caps will not bind the model's determination of SO2 emissions
regardless of any level of initial allowance bank assumed. Therefore, EPA Platform v6 does not assume
any Title IV SO2 allowance bank amount for the November 2018 Reference Case year of 2021
(notwithstanding that a large allowance bank will exist in that year in practice), because such an
assumption would have no material impact on projections given the nonbinding nature of that program.
Calculating the available 2021 allowances involved deducting allowance surrenders due to NSR
settlements and state regulations from the 2021 SO2 cap of 8.95 million tons. The surrenders totaled 977
thousand tons in allowances, leaving 7.973 million of 2021 allowances remaining. Specifics of the
allowance surrender requirements under state regulations and NSR settlements can be found in Table
3-23 and Table 3-24.
EPA Platform v6 also includes a representation of the Western Regional Air Partnership (WRAP)
Program, a regional initiative involving New Mexico, Utah, and Wyoming directed toward addressing
visibility issues in the Grand Canyon and affecting SO2 emissions starting in 2018. The WRAP
specifications for SO2 are presented in Table 3-15.
3.9.2 NOx Regulations
Much like SO2 regulations, existing NOx regulations are represented in EPA Platform v6 through a
combination of system level NOx programs and generation unit-level NOx limits. In EPA Platform v6, the
NOx SIP Call trading program, Cross State Air Pollution Rule (CSAPR), and the CSAPR Update Rule are
represented. Table 3-15 shows the specification for the entire modeling time horizon.
By assigning unit-specific NOx rates based on 2017 data, EPA Platform v6 is implicitly representing Title
IV unit-specific rate limits and Clean Air Act Reasonably Available Control Technology (RACT)
requirements for controlling NOx emissions from electric generating units in ozone non-attainment areas
or in the Ozone Transport Region (OTR).19 Unlike SO2 emission rates, NOx rates are calculated off
19 The OTR consists of the following states: Maine, New Hampshire, Vermont, Massachusetts, Rhode Island,
Connecticut, New York, New Jersey, Pennsylvania, Delaware, Maryland, District of Columbia, and northern Virginia.
3-18
-------
historical data and reflect the fuel mix for that particular year and burn at the unit. NEEDs represents up
to four scenario NOx rates based on historical data to capture seasonal and existing control
variability. These rates are constant and do not change independent of fuel mix assumed in the model. If
the unit undertakes a post-combustion control retrofit or a coal-to-gas retrofit, then these rates would
change in the model projections.
NOx Emission Rates
Future emission projections for NOx are a product of a unit's utilization (heat input) and emission rate
(Ibs/MMBtu). A unit's NOx emission rate can vary significantly depending on the NOx reduction
requirements to which it is subject. For example, a unit may have a post-combustion control installed
(i.e., SCR or SNCR), but only operate it during the particular time of the year in which it is subject to NOx
reduction requirements (e.g., the unit only operates its post-combustion control during the ozone season).
Therefore, its ozone-season NOx emission rate would be lower than its non-ozone-season NOx emission
rate. Because the same individual unit can have such large variation in its emission rate, the model
needs a suite of emission rate "modes" from which it can select the value most appropriate to the
conditions in any given model scenario. The different emission rates reflect the different operational
conditions a unit may experience regarding upgrades to its combustion controls and operation of its
existing post-combustion controls. Four modes of operation are developed for each unit, with each mode
carrying a potentially different NOx emission rate for that unit under those operational conditions.
The emission rates assigned to each mode are derived from historical data (where available) and
presented in NEEDS v6. When the model is run, IPM selects one of these four modes through a decision
process depicted in Figure 3-4 below. The four modes address whether or not units upgrade combustion
controls and/or operate existing post-combustion controls; the modes themselves do not address what
happens to the unit's NOx rate if it is projected to add a new post-combustion NOx control. If a unit is
projected to add a new post-combustion control, then after the model selects the appropriate mode it
adjusts downward its emission rate to reflect the retrofit of SCR or SNCR; the adjusted rate will reflect the
greater of a percentage removal from the mode's emission rate or an emission rate floor. The full process
for determining the NOx rate of units in EPA Platform v6 model projections is summarized in Figure 3-3
below.
Figure 3-3 Modeling Process for Obtaining Projected NOx Emission Rates
*
NEEDS
Assignment of emission rates
(derived from historic data) to
each of four NOx Modes. Modes
reflect different potential
operational conditions at a unit.
Historical NOx
Emission Rate Data
(e.g., 2017)
Model Projections
Assignment of NOx emission rate
based on one of four NEEDS
"modes" rates with potential
adjustment if the unit is projected to
add post-combustion retrofit control
technology.
NOx Emission Rates in NEEDS v6 Database
The NOx rates were derived, wherever possible, directly from actual monitored NOx emission rate data
reported to EPA under the Acid Rain and Cross-State Air Pollution Rule in 2017.20 The emission rates
20 By assigning unit-specific NOx rates based on 2017 data, EPA Platform v6 is implicitly representing Title IV unit-
specific rate limits and Clean Air Act Reasonably Available Control Technology (RACT) requirements for controlling
NOx emissions from electric generating units in ozone non-attainment areas or in the Ozone Transport Region (OTR).
Unlike SO2 emission rates, NOx emission rates are assumed not to vary with coal type, but are dependent on the
combustion properties of the generating unit. Under the EPA Platform v6, the NOx emission rate of a unit can only
change if the unit is retrofitted with NOx post-combustion control equipment or if it is assumed to install state-of-the-art
3-19
-------
themselves reflect the impact of applicable NOx regulations21. For coal-fired units, NOx rates were used in
combination with empirical assessments of NOx combustion control performance to prepare a set of four
possible starting NOx rates to assign to a unit, depending on the specific NOx reduction requirements
affecting that unit in a model run.
The reason for having a framework of four potential NOx rate "modes" applicable to each unit in NEEDS is
to enable the model to select from a range of NOx rates possible at a unit, given its configuration of NOx
combustion controls and its assumed operation of existing post-combustion controls. There are up to four
basic operating states for a given unit that significantly impact its NOx rate, and thus there are four NOx
rate "modes".
Mode 1 and mode 2 reflect a unit's emission rates with its existing configuration of combustion and post-
combustion (i.e., SCR or SNCR) controls.
• For a unit with an existing post-combustion control, mode 1 reflects the existing post-combustion
control not operating and mode 2 the existing post-combustion control operating. However:
o If a unit has operated its post-combustion control year round during 2017, 2016, 2015,
2014, 2011, 2009, and 2007 years then mode 1 = mode 2, which reflects that the control
will likely continue to operate year round,
o If a unit has not operated its post-combustion control during 2017, 2016, 2015, 2014,
2011, 2009, and 2007 years, mode 1 will be based on historic data and mode 2 will be
calculated using the method described under Question 3 in Attachment 3-1.
o If a unit has operated its post-combustion control seasonally in recent years (i.e., either
only in the summer or winter, but not both), mode 1 will be based on historic data from
when the control was not operating, and mode 2 will be based on historic data from when
the SCR was operating.
• For a unit without an existing post-combustion control, mode 1 = mode 2 which reflects the unit's
historic NOx rates from a recent year.
Mode 3 and mode 4 emission rates parallel modes 1 and 2 emission rates, but are modified to reflect
installation of state-of-the-art combustion controls on a unit if it does not already have them.
• For units that already have state-of-the-art combustion controls: Mode 3 = mode 1 and mode 4 =
mode 2.
Emission rates derived for each unit operating under each of these four modes are presented in NEEDS
v6. Note that not every unit has a different emission rate for each mode, because certain units cannot in
practice change their NOx rates to conform to all potential operational states described above.
NOx combustion controls. In instances where a coal steam unit converts to natural gas, the NOx rate is assumed to
reduce by 50%.
21 Because 2017 NOx rates reflect CSAPR, we no longer apply any incremental CSAPR related NOx rate adjustments
exogenously for CSAPR affected units in EPA Platform v6.
3-20
-------
Figure 3-4 How One of the Four NOx Modes Is Ultimately Selected for a Unit
Did the source operate a
post-combustion control
in 2017?
Non-ozone Season
Ozone Season
Is it a seasonal or
annual requirement?
Annual
Mode 1: Existing combustion controls, no post-
combustion control operating
Mode 2: Existing combustion controls, post-
combustion control operating (where applicable)
Mode 3:
If SNCR - SOA combustion controls, no post
combustion control operating
If SCR - Mode 3 = Mode 1
Existing combustion controls, no post-combustion
controls operating (where applicable)
Mode 4:
If SNCR - SOA combustion controls, post-
combustion controls operating
If SCR - Mode 4 = Mode 2
Existing combustion controls, post-combustion
controls operating (where applicable)
State-of-the-art combustion controls (SOA combustion controls')
The definition of "state-of-the-art" varies depending on the unit type and configuration. Table 3-11
indicates the incremental combustion controls that are required to achieve a "state-of-the-art" combustion
control configuration for each unit. For instance, if a wall-fired, dry bottom boiler (highlighted below)
currently has LNB but no overfire air (OFA), the "state-of-the-art" rate calculated for such a unit would
assume a NOx emission rate reflective of overfire air being added at the unit. As described in the
attachment of this chapter, the "state-of-the-art" combustion controls reflected in the modes are only
assigned to a unit if it is subject to a new (post-2017) NOx reduction requirement (i.e., a NOx reduction
requirement that did not apply to the unit during its 2017 operation that forms the historic basis for
deriving NOx rates for units in EPA Platform v6). Existing reduction requirements as of 2017 under which
units have already made combustion control decisions would not trigger the assignment of the "state-of-
the-art" modes that reflect additional combustion controls.
Table 3-11 State-of-the-Art Combustion Control Configurations by Boiler Type
Boiler Type
Existing NOx
Combustion Control
Incremental Combustion Control
Necessary to Achieve "State-of-the-Art"
Tangential Firing
Does not Include LNC1 and LNC2
Includes LNC1, but not LNC2
Includes LNC2, but not LNC3
Includes LNC1 and LNC2 or LNC3
LNC3
CONVERSION FROM LNC1 TO LNC3
CONVERSION FROM LNC2 TO LNC3
Wall Firing, Dry Bottom
Does not Include LNB and OFA
Includes LNB, but not OFA
Includes OFA, but not LNB
Includes both LNB and OFA
LNB + OFA
OFA
LNB
Note:
LNB = Low NOx Burner Technology, LNC1 = Low NOx coal-and air nozzles with close-coupled overfire air, LNC2 =
3-21
-------
Low NOx Coal-and-Air Nozzles with Separated Overfire Air, LNC3 = Low NOx Coal-and-Air Nozzles with Close-
Coupled and Separated Overfire Air, OFA = Overfire Air
The emission rates for each generating unit under each mode are included in the NEEDS v6 database,
described in Chapter 4. Attachment 3-1 gives further information on the procedures employed to derive
the four NOx mode rates.
Because of the complexity of the fleet and the completeness/incompleteness of historic data, there are
instances where the derivation of a unit's modeled NOx emission rate is more detailed than the
description provided above. For a more complete step-by-step description of the decision rules used to
develop the NOx rates, please see Attachment 3-1.
3.9.3 Multi-Pollutant Environmental Regulations
CSAPR
EPA Platform v6 includes the Cross-State Air Pollution Rule (CSAPR) Rule and CSAPR Update Rule,
federal regulatory measures affecting 23 states to address transport under the 1997, 2006, and 2008
National Ambient Air Quality Standards (NAAQS) for fine particle pollution and ozone. CSAPR requires
fossil-fired EGUs greater than 25 MW in a total of 22 states to reduce annual SO2 emissions, annual NOx
emissions, and/or ozone season NOx emissions to assist in attaining the 1997 ozone and fine particle and
2006 fine particle National Ambient Air Quality Standards (NAAQS). The CSAPR Phase 2 combined
annual emissions budgets are 1,372.631 thousand tons SO2 for CSAPR SO2 Group 1 ;22 597.579
thousand tons SO2 for CSAPR SO2 Group 2;23 and 1,069.256 thousand tons for annual NOx.24 As the
budgets are significantly above current emission levels, i.e. they are not binding, the EPA did not include
a starting bank of allowances for these programs for simplicity.
The original Phase 2 combined ozone season NOx emissions budget was 0.59 million tons; however,
several of the state budgets were remanded. As the CSAPR Update Rule addresses the D.C. Circuit's
remand, the remanded budgets were not included in the EPA Platform v6. The programs' assurance
provisions, which restrict the maximum amount of exceedance of an individual state's emissions budget
in a given year through the use of banked or traded allowances to 18% or 21 % of the state's budget are
also included. For more information on CSAPR, go to https://www.epa.gov/csapr/overview-cross-state-
air-pollution-rule-csapr.
The state budgets for Ozone Season NOx for the CSAPR Update Rule are shown in Table 3-12.
Additionally, Georgia was modeled as a separate region, with Georgia units unable to trade allowances
with units in other states, and received its CSAPR Phase 2 budget and assurance level, as shown in the
table below. This is because Georgia, unlike the other states covered by the CSAPR Update Rule, did
not significantly contribute to a downwind nonattainment or maintenance receptor for the 2008 NAAQS
and, furthermore, did not have a remanded Ozone Season NOx budget related to a D.C. Circuit Court
decision on the original Cross-State Air Pollution Rule.
The programs' assurance provisions, which restrict the maximum amount of exceedance of an individual
state's emissions budget in each year through the use of banked or traded allowances to 18% or 21 % of
the state's budget, are also implemented. The starting allowance bank in 2021 is 98,670 tons, which is
equal to the number of banked allowances at the start of the CSAPR Update program after old CSAPR
allowances were converted. This is equal to one-and-a-half times the sum of the states' 21% variability
22 Illinois, Indiana, Iowa, Kentucky, Maryland, Michigan, Missouri, New Jersey, New York, North Carolina, Ohio,
Pennsylvania, Tennessee, Virginia, West Virginia, Wisconsin
23 Alabama, Georgia, Kansas, Minnesota, Nebraska, South Carolina
24 Alabama, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Maryland, Michigan, Minnesota, Missouri, Nebraska,
New Jersey, New York, North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, Virginia, West Virginia,
Wisconsin
3-22
-------
limits. For more information on CSAPR, go to https://www.epa.gov/csapr. For more information on the
CSAPR Update, go to https://www.epa.gov/airmarkets/final-cross-state-air-pollution-rule-update.
Table 3-12 CSAPR Update State Budgets, Variability Limits, and Assurance Levels for Ozone-
Season NOx(Tons)
Budget
Variability Limit
Assurance Level
Alabama
13,211
2,774
15,985
Arkansas
9,210
1,934
11,144
Iowa
11,272
2,367
13,639
Illinois
14,601
3,066
17,667
Indiana
23,303
4,894
28,197
Kansas
8,027
1,686
9,713
Kentucky
21,115
4,434
25,549
Louisiana
18,639
3,914
22,553
Maryland
3,828
804
4,632
Michigan
17,023
3,575
20,598
Missouri
15,780
3,314
19,094
Mississippi
6,315
1,326
7,641
New Jersey
2,062
433
2,495
New York
5,135
1,078
6,213
Ohio
19,522
4,100
23,622
Oklahoma
11,641
2,445
14,086
Pennsylvania
17,952
3,770
21,722
Tennessee
7,736
1,625
9,361
Texas
52,301
10,983
63,284
Virginia
9,223
1,937
11,160
Wisconsin
7,915
1,662
9,577
West Virginia
17,815
3,741
21,556
CSAPR Update Region Total
313,626
N/A
N/A
Georgia Budget, Variability Limit, and Assurance Level
for Ozone-Season NOx
Georgia
24,041
5,049
29,090
MATS
Finalized in 2011, the Mercury and Air Toxics Rule (MATS) establishes National Emissions Standards for
Hazardous Air Pollutants (NESHAPS) for the "electric utility steam generating unit" source category,
which includes those units that combust coal or oil for the purpose of generating electricity for sale and
distribution through the electric grid to the public. EPA Platform v6 applies the input-based (Ibs/MMBtu)
MATS control requirements for mercury and hydrogen chloride to covered units.
EPA Platform v6 assumes that all active coal-fired generating units with a capacity greater than 25 MW
have complied with the MATS filterable PM requirements through the operation of either electrostatic
3-23
-------
precipitator (ESP) or fabric filter (FF) particulate controls. No additional PM controls beyond those in
NEEDS v6 are modeled in EPA Platform v6.
EPA Platform v6 does not model the alternative SO2 standard offered under MATS for units to
demonstrate compliance with the rule's HCI control requirements. Coal steam units with access to lignite
in the modeling are required to meet the "existing coal-fired unit low Btu virgin coal" standard. For more
information on MATS, go to http://www.epa.gov/mats/.
Regional Haze
The Clean Air Act establishes a national goal for returning visibility to natural conditions through the
"prevention of any future, and the remedying of any existing impairment of visibility in Class I areas [156
national parks and wilderness areas], where impairment results from manmade air pollution." On July 1,
1999, EPA established a comprehensive visibility protection program with the issuance of the regional
haze rule (64 FR 35714). The rule implements the requirements of section 169B of the CAAA and
requires states to submit State Implementation Plans (SIPs) establishing goals and long-term strategies
for reducing emissions of air pollutants (including SO2 and NOx) that cause or contribute to visibility
impairment. The requirement to submit a regional haze SIP applies to all 50 states, the District of
Columbia, and the Virgin Islands. Among the components of a long-term strategy is the requirement for
states to establish emission limits for visibility-impairing pollutants emitted by certain source types
(including EGUs) that were placed in operation between 1962 and 1977. These emission limits are to
reflect Best Available Retrofit Technology (BART). States may perform individual point source BART
determinations, or meet the requirements of the rule with an approved BART alternative. An alternative
regional SO2 cap for EGUs under Section 309 of the regional haze rule is available to certain western
states whose emission sources affect Class 1 areas on the Colorado Plateau.
Since 2010, EPA has approved regional haze State Implementation Plans (SIPs) or, in a few cases, put
in place regional haze Federal Implementation Plans for several states. The BART limits approved in
these plans (as of August 2017) that will be in place for EGUs are represented in the EPA Platform v6 as
follows.
• Source-specific NOx or SO2 BART emission limits, minimum SO2 removal efficiency requirements for
FGDs, limits on sulfur content in fuel oil, constraints on fuel type (e.g., natural gas only or prohibition
of certain fuels such as petroleum coke), or commitments to retire units are applied to the relevant
EGUs.
• EGUs in states that rely on CSAPR trading programs to satisfy BART must meet the requirements of
CSAPR.
• EGUs in states that rely on state power plant rules to satisfy BART must meet the emission limits
imposed by those state rules.
• For the three western states (New Mexico, Wyoming, and Utah) with approved Section 309 SIPs for
SO2 BART, emission constraints were not applied as current and projected emissions are well under
the regional SO2 cap.
Table 3-28 lists the NOx and SO2 limits applied to specific EGUs and other implementations applied in
IPM. For more information on the Regional Haze Rule, go to https://www.epa.gov/visibilitv.
3.9.4 CO2 Regulations
The Regional Greenhouse Gas Initiative (RGGI) is a CO2 cap and trade program affecting fossil fired
electric power plants 25 MW or larger in Connecticut, Delaware, Maine, Maryland, Massachusetts, New
3-24
-------
Hampshire, New York, Rhode Island, and Vermont.25 Table 3-15 shows the specifications for RGGI that
are implemented in EPA Platform v6.
As part of California's Assembly Bill 32 (AB32), the Global Warming Solutions Act, a multi-sector GHG
cap-and-trade program was established that targets 1990 emission levels by 2020.26 The cap begins in
2013 for electric utilities and large industrial facilities, with distributors of transportation, natural gas, and
other fuels joining the capped sectors in 2015. In addition to in-state sources, the cap-and-trade program
also covers the emissions associated with qualifying, out-of-state EGUs that sell power into
California. Due to the inherent complexity in modeling a multi-sector cap-and-trade program where the
participation of out-of-state EGUs is determined based on endogenous behavior (i.e., IPM determines
whether qualifying out-of-state EGUs are projected to sell power into California), EPA has developed a
simplified methodology to model California's economy-wide cap-and-trade program as follows.
• Adopt the AB32 cap-and-trade allowance price from ElA's AEO2017 Reference Case, which fully
represents the non-power sectors. All qualifying fossil-fired EGUs in California are subject to this
price signal, which is applied through the end of the modeled time horizon since the underlying
legislation requires those emission levels to be maintained.
• Assume the marginal CO2 emission rate for each IPM region that exports power to California to be
0.428 MT/MWh.
• For each IPM region that exports power to California, convert the $/ton CO2 allowance price
projection into a mills/kWh transmission wheeling charge using the marginal emission rate from the
previous step. The additional wheeling charge for qualifying out-of-state EGUs is equal to the
allowance price imposed on affected in-state EGUs. Applying the charge to the transmission link
ensures that power imported into California from out-of-state EGUs must account for the cost of CO2
emissions represented by its generation, such that the model may clear the California market in a
manner consistent with AB32 policy treatment of CO2 emissions.
Federal CO2 standards for existing sources are not modeled, given ongoing litigation and regulatory
review of the Clean Power Plan.27 For new fossil fuel-fired sources, EPA Platform v6 continues to include
the Standards of Performance for Greenhouse Gas Emissions from New, Modified, and Reconstructed
Stationary Sources: Electric Generating Units (New Source Rule).28 Although this rule is also being
reviewed,29 the standards of performance are legally in effect until such review is completed and/or
revised (unlike the Clean Power Plan, which has been stayed by the Supreme Court).
3.9.5 Non-Air Regulations Impacting EGUs
Cooling Water Intakes (316(b)) Rule
Section 316(b) of the Clean Water Act requires that National Pollutant Discharge Elimination System
(NPDES) permits for facilities with cooling water intake structures ensure that the location, design,
construction, and capacity of the structures reflect the best technology available to minimize harmful
25 As of this publication, the states of New Jersey and Virginia have expressed intent to join RGGI but have not yet
concluded state regulatory proceedings to do so. If/when RGGI's composition and/or policy details change through
applicable final rules by participating states, we will adjust that program's representation in our modeling platform and
issue updated documentation accordingly.
26 In July of 2017, AB 398 was signed into law. AB 398 extends the timeframe for cap-and-trade program through
2030 and further lowered the cap to at least 40% below the 1990 levels. This new regulation will be considered in
future updates to IPM.
27 80 FR 64662 (Clean Power Plan, which has been stayed by the Supreme Court) and 82 FR 16329 (Clean Power
Plan Review).
28 80 FR 64510
29 82 FR 16330
3-25
-------
impacts on the environment. Under a 1995 consent decree with environmental organizations, EPA
divided the section 316(b) rulemaking into three phases. All new facilities except offshore oil and gas
exploration facilities were addressed in Phase I in December 2001; all new offshore oil and gas
exploration facilities were later addressed in June 2006 as part of Phase III. This final rule also removes
a portion of the Phase I rule to comply with court rulings. Existing large electric-generating facilities were
addressed in Phase II in February 2004. Existing small electric-generating and all manufacturing facilities
were addressed in Phase III (June 2006). However, Phase II and the existing facility portion of Phase III
were remanded to EPA for reconsideration because of legal proceedings. This final rule combines these
remands into one rule, and provides a holistic approach to protecting aquatic life impacted by cooling
water intakes. The rule covers roughly 1,065 existing facilities that are designed to withdraw at least 2
million gallons per day of cooling water. EPA estimates that 544 power plants are affected by this rule.
The final regulation has three components for affected facilities: 1) reduce fish impingement through a
technology option that meets best technology available requirements, 2) conduct site-specific studies to
help determine whether additional controls are necessary to reduce entrainment, and 3) meet
entrainment standards for new units at existing facilities when additional capacity is added. EPA Platform
v6 includes cost of complying with this rule. The cost assumptions and analysis for 316(b) can be found
in Chapter 8.7 of the Rule's Technical Development Document for the Final Section 316(b) Existing
Facilities Rule at https://www.epa.qov/sites/production/files/2015-04/documents/coolinq-water phase-
4 tdd 2014.pdf.
For more information on 316(b), go to http://water.epa.gov/lawsreqs/lawsquidance/cwa/316b/index.cfm.
Combustion Residuals from Electric Utilities (CCR)
In December of 2014, EPA finalized national regulations to provide a comprehensive set of requirements
for the safe disposal of coal combustion residuals (CCRs), commonly known as coal ash, from coal-fired
power plants. The final rule is the culmination of extensive study on the effects of coal ash on the
environment and public health. The rule establishes technical requirements for CCR landfills and surface
impoundments under Subtitle D of the Resource Conservation and Recovery Act.
EPA Platform v6 includes cost of complying with this rule's requirements by taking the estimated plant-
level compliance cost identified in the 2014 Regulatory Impact Analysis (RIA) for the CCR final rule and
apportioning them into unit-level cost. Three categories of unit-level cost were quantified; capital cost,
fixed operating and maintenance cost (FOM), and variable operating and maintenance (VOM) cost. The
method for apportioning these costs to the unit-level for inclusion in EPA Platform is discussed in the
Addendum to the RIA for EPA's 2015 Coal combustion Residuals (CCR) Final Rule. The initial plant-level
cost estimates are discussed in the Rule's Regulatory Impact Analysis.
In September of 2017, EPA granted petitions to reconsider some provisions of the rule. In granting the
petitions, EPA determined that it was appropriate, and in the public's interest to reconsider specific
provisions of the final CCR rule based in part on the authority provided through the Water Infrastructure
for Improvements to the Nation (WIIN) Act. At time of this modeling update, EPA had not committed to
changing any part of the rule, or agreeing with the merits of the petition - the Agency is simply granting
petitions to reconsider specific provisions. Should EPA decide to revise specific provisions of the final
CCR rule, it will go through notice and comment period, and the rules corresponding model specification
would be subsequently changed in future base case platforms.
For more information on CCR, go to http://www2.epa.gov/coalash/coal-ash-rule.
3-26
-------
Effluent Limitation and Guidelines (ELG)
In September of 2015, EPA finalized a rule revising the regulations for Steam Electric Power Generating
category (40 CFR Part 423).30 The rule established federal limits on the levels of toxic metals in
wastewater that can be discharged from power plants. The rule established or updated standards for
wastewater streams from flue gas desulfurization, fly ash, bottom ash, flue gas mercury control, and
gasification of fuels. EPA estimated that approximately 12% of steam electric power plants would incur
some compliance cost. EPA reflects this rule in this base case by apportioning the estimated total capital
and FOM costs to likely affected units based on controls and capacity. The cost adders are reflected in
the model inputs and were applied starting in 2023, by which point the requirements were expected to be
fully implemented.
In August of 2017, EPA noted that it would conduct a rulemaking to potentially revise the limitations and
standards for bottom ash transport water and flue gas desulfurization wastewater. EPA noted that, given
the typical timeline to propose and finalize a rulemaking, it would postpone earliest compliance dates by 2
years. Therefore, in EPA Platform v6, EPA has postponed the full implementation by 2 years, but has not
made any capital or FOM adjustments reflecting new limitations and standards as no new standards have
been finalized at the time of model update.
3.9.6 State-Specific Environmental Regulations
EPA Platform v6 represents enacted laws and regulations in 27 states affecting emissions from the
electricity sector. Table 3-23 summarizes the provisions of state laws and regulations that are
represented in EPA Platform v6.
The NY minimum oil burn rule was implemented for the following units through facility level minimum
generation constraints in the 2021, 2023, and 2025 run years. The minimum generation limits are
calculated using the capacity factors shown in Table 3-13.
Table 3-13 NY Minimum Oil Burn Rule Plant Level Oil Capacity Factor Requirements
Oil Capacity Factor
(%)
Winter
Winter Shoulder
Summer
Steam Facilities (Heavy Oil)
Astoria
2.10%
0.20%
0.50%
East River
3.00%
0.40%
0.60%
Northport
5.20%
0.50%
2.00%
Ravenswood
0.70%
0.20%
0.60%
Combined Cycle (Light Oil)
Astoria Energy
2.90%
0.00%
0.00%
Charles Polletti Power Plant
3.00%
0.40%
0.00%
Ravenswood
1.00%
0.10%
0.00%
3.9.7 New Source Review (NSR) Settlements
New Source Review (NSR) settlements refer to legal agreements with companies resulting from the
permitting process under the CAAA which requires industry to undergo an EPA pre-construction review of
proposed environmental controls either on new facilities or as modifications to existing facilities where
there would result a "significant increase" in a regulated pollutant. EPA Platform v6 includes NSR
30 https://www.epa.aov/ea/steam-electric-power-aeneratina-effluent-auidelines-2015-final-rule
3-27
-------
settlements with 34 electric power companies. A summary of the units affected and how the settlements
were modeled can be found in Table 3-24.
Nine state settlements and ten citizen settlements are also represented in EPA Platform v6. These are
summarized in Table 3-25 and Table 3-26 respectively.
3.9.8 Emission Assumptions for Potential (New) Units
Emissions from existing and planned/committed units vary from installation to installation based on the
performance of the generating unit and the emissions regulations that are in place. In contrast, there are
no location-specific variations in the emission and removal rate capabilities of potential new units. In IPM,
potential new units are modeled as additional capacity and generation that may come online in each
model region. Across all model regions, the emission and removal rate capabilities of potential new units
are the same, and they reflect applicable federal emission limitations on new sources. The specific
assumptions regarding the emission and removal rates of potential new units in EPA Platform v6 are
presented in Table 3-17. (Note: Nuclear, wind, solar, and fuel cell technologies are not included in Table
3-17 because they do not emit any of the listed pollutants.) For additional details on the modeling of
potential new units, see Chapter 4.
3.9.9 Energy Efficiency and Renewable Portfolio Standards
Renewable Portfolio Standards (RPS) generally refers to various state-level policies that require the
addition of renewable generation to meet a specified share of statewide generation. In EPA Platform v6,
the state RPS requirements are represented at a state level based on requirements. Table 3-19 shows
the state level RPS requirements. In addition, state level solar carve-out requirements have been
implemented in EPA Platform v6.
3.9.10 Canada CO2 and Renewable Regulations
Several CO2 regulations in Canada are represented in EPA Platform v6. Under the Reduction of Carbon
Dioxide Emissions from Coal-fired Generation of Electricity Regulations, the CO2 standard of 420 tonne
/GWh of electricity produced apply to both new coal-fired electricity generating units commissioned after
July 1, 2015, and existing coal units that have reached their end-of-life date as defined by the regulation.
EPA Platform v6 also models the British Columbia's carbon tax, Manitoba's Emissions Tax on Coal and
Petroleum Coke Act, and the Ontario and Quebec's participation in Western Climate Initiative (WCI) cap-
and-trade program. British Columbia's carbon tax sets a tax rate of $35 per tonne of CO2 equivalent
emissions beginning April 1, 2018 and increases it each year by $5 per tonne until it reaches $50 per
tonne in 2021. Coming into force on January 1, 2012, Manitoba's Emissions Tax on Coal and Petroleum
Coke Act requires a tax rate of $10 per tonne of CO2 equivalent emissions on coal-fired and petroleum
coke-fired units. Ontario and Quebec's participation in WCI is modeled through the application of the CO2
allowance price from CA AB32. EPA Platform v6 also models the province level renewable electricity
programs in Canada. Table 3-14 shows the province level renewable electricity requirements as a
percentage of electricity sales.
Table 3-14 Canada Renewable Electricity Requirements (%) in EPA Platform v6
Province
2021
2023
2025
2030
2035
2040
2045
2050
British Columbia
93
93
93
93
93
93
93
93
Alberta
30
30
30
30
30
Saskatchewan
50
50
50
50
50
New Brunswick
40
40
40
40
40
40
40
40
Nova Scotia
40
40
40
40
40
40
40
40
Prince Edward Island
30
30
30
30
30
30
30
30
3-28
-------
3.10 Emissions Trading and Banking
Several environmental air regulations included in EPA Platform v6 involve regional trading and banking of
emission allowances: This includes the five programs of the Cross-State Air Pollution Rule (CSAPR) -
SO2 Region 1, SO2 Region 2, Annual NOx, CSAPR Update Rule Ozone Season NOx Region 1, and
CSAPR Update Rule Ozone Season NOx Region 2; the Regional Greenhouse Gas Initiative (RGGI) for
CO2; the SIP Call Ozone Season NOx; and the West Region Air Partnership's (WRAP) program regulating
SO2 (adopted in response to the federal Regional Haze Rule).
Table 3-15 and Table 3-16 summarize the key parameters of these trading and banking programs as
incorporated in EPA Platform v6. EPA Platform v6 does not include any explicit assumptions on the
allocation of emission allowances among model plants under any of the programs.
3.10.1 Intertemporal Allowance Price Calculation
Under a perfectly competitive cap-and-trade program that allows banking (with a single, fixed future cap
and full "banking" allowed), the allowance price always increases by the discount rate between periods if
affected sources have allowances banked between those two periods. This is a standard economic result
for cap-and-trade programs and is consistent with producing a least-cost solution.
EPA Platform v6 uses the same discount rate assumption that governs all intertemporal economic
decision-making in the model to compute the increase in allowance price for cap-and-trade programs
when banking is engaged as a compliance strategy. The approach assumes that allowance trading is a
standard activity engaged in by generation asset owners and that their intertemporal investment decisions
as related to allowance trading will not fundamentally differ from other investment decisions. For more
information on how this discount rate was calculated, please see Section 10.4.
Table 3-15 Trading and Banking Rules in EPA Platform v6 - Part 1
SIP Call - Ozone Season NOx
WRAP- SO2
RGGI - CO2
Coverage
All fossil units > 25 MW1
All fossil units > 25 MW2
All fossil units > 25 MW3
Timing
Ozone Season (May - September)
Annual
Annual
Size of Initial Bank
(MTons)
The bank starting in 2021 is assumed
to be zero
The bank starting in 2021 is
assumed to be zero
2021: 49,442
Total Allowances
(MTons)
2016-2054: 72.845
2018 -2054: 89.6
2021: 75,148
2022: 72,873
2023: 70,598
2024: 68,323
2025: 66,048
2026: 63,773
2027: 61,498
2028: 59,223
2029: 56,948
2030 - 2054: 54,673
Notes:
1 Rhode Island, Connecticut, Delaware, District of Columbia, Massachusetts, North Carolina, and South Carolina are the NOx SIP
Call states not covered by the CSAPR Ozone Season program.
2 New Mexico, Utah, Wyoming
3 Connecticut, Delaware, Maine, New Hampshire, New York, Vermont, Rhode Island, Massachusetts, Maryland
3-29
-------
Table 3-16 CASPR Trading and Banking Rules in EPA Platform v6 - Part 2
CSAPR - SO2 - Region 1
CSAPR - SO2 -
Region 2
CSAPR-
Annual NOx
CSAPR Update
Rule - Ozone
Season NOx -
Region 1
CSAPR Update
Rule - Ozone
Season NOx -
Region 2
Coverage
All fossil units > 25 MW1
All fossil units >
25 MW2
All fossil units >
25 MW3
All fossil units >
25 MW4
All fossil units >
25 MW5
Timing
Annual
Annual
Annual
Ozone Season
(May -
September)
Ozone Season
(May -
September)
Size of Initial
Bank
(MTons)
The bank starting in 2021
is assumed to be zero
The bank
starting in 2021
is assumed to be
zero
The bank starting
in 2021 is
assumed to be
zero
The cap in 2021
includes 21% of
banking
The bank starting
in 2021 is
assumed to be
zero
Total
Allowances
(MTons)
2021 -2054: 1372.631
2021 -2054:
597.579
2021 -2054:
1069.256
2021: 411.9106
2022 - 2054:
313.24
2021 - 2054:
24.041
Notes:
1 Illinois, Indiana, Iowa, Kentucky, Maryland, Michigan, Missouri, New Jersey, New York, North Carolina, Ohio, Pennsylvania, Tennessee,
Virginia, West Virginia, Wisconsin
2 Alabama, Georgia, Kansas, Minnesota, Nebraska, South Carolina
3 Alabama, Georgia, Illinois, Indiana, Iowa, Kansas, Kentucky, Maryland, Michigan, Minnesota, Missouri, Nebraska, New Jersey, New York,
North Carolina, Ohio, Pennsylvania, South Carolina, Tennessee, Virginia, West Virginia, Wisconsin
4 Alabama, Arkansas, Iowa, Illinois, Indiana, Kansas, Kentucky, Louisiana, Maryland, Michigan, Missouri, Mississippi, New Jersey, New
York, Ohio, Oklahoma, Pennsylvania, Tennessee, Texas, Virginia, Wisconsin, West Virginia
5 Georgia
3-30
-------
Table 3-17 Emission and Removal Rate Assumptions for Potential (New) Units in EPA Platform v6
Controls,
Removal, and
Emissions
Rates
Ultra
Supercritical
Pulverized Coal
Ultra
Supercritical
Pulverized Coal
with 30% CCS
Ultra
Supercritical
Pulverized Coal
with 90% CCS
Advanced
Combined
Cycle
Advanced
Combined Cycle
with Carbon
Sequestration
Advanced
Combustion
Turbine
Biomass-
Bubbling
Fluidized Bed
(BFB)
Geothermal
Landfill
Gas
CM
o
CO
Removal /
Emissions
Rate
98% with a floor of
0.06 Ibs/MMBtu
98% with a floor of
0.06 Ibs/MMBtu
98% with a floor of
0.06 Ibs/MMBtu
None
None
None
0.08
Ibs/MMBtu
None
None
NO„
Emission Rate
0.07 Ibs/MMBtu
0.07 Ibs/MMBtu
0.07 Ibs/MMBtu
0.011
Ibs/MMBtu
0.011 Ibs/MMBtu
0.011 Ibs/MMBtu
0.02
Ibs/MMBtu
None
0.09
Ibs/MMBtu
Hg
Removal /
Emissions
Rate
90%
90%
90%
Natural Gas:
0.000138
Ibs/MMBtu
Oil:
0.483
Ibs/MMBtu
Natural Gas:
0.000138 Ibs/MMBtu
Oil:
0.483 Ibs/MMBtu
Natural Gas:
0.000138
Ibs/MMBtu
Oil:
0.483 Ibs/MMBtu
0.57
Ibs/MMBtu
3.70
None
co2
Removal /
Emissions
Rate
202.8-215.8
Ibs/MMBtu
30%
90%
Natural Gas:
117.08
Ibs/MMBtu
Oil:
161.39
Ibs/MMBtu
90%
Natural Gas:
117.08 Ibs/MMBtu
Oil:
161.39 Ibs/MMBtu
None
None
None
HCL
Removal /
Emissions
Rate
99% 0.001
Ibs/MMBtu
99% 0.001
Ibs/MMBtu
99% 0.001
Ibs/MMBtu
3-31
-------
Table 3-18 Recalculated NOx Emission Rates for SCR Equipped Units Sharing Common Stacks
with Non-SCR Units
Plant Name
UniquelD_
Capacity
NOx Post-
SCR Online
Mode 1
Mode 2
Mode 3
Mode 4
Final
(MW)
Comb Control
_Year
NOx Rate
NOx Rate
NOx Rate
NOx Rate
Ghent
1356_B_2
484
0.340
0.253
0.340
0.253
Ghent
1356_B_3
480
SCR
2004
0.075
0.075
0.075
0.075
Chalk Point LLC
1571 B 1
331
SCR
2009
0.075
0.075
0.075
0.075
Chalk Point LLC
1571_B_2
336
SNCR
0.270
0.237
0.270
0.237
FirstEnergy W H
Sammis
2866_B_5
300
SNCR
0.283
0.258
0.283
0.258
FirstEnergy W H
Sammis
2866_B_6
600
SCR
2010
0.075
0.075
0.075
0.075
FirstEnergy W H
Sammis
2866_B_7
600
SCR
2010
0.075
0.075
0.075
0.075
Charles R Lowman
56_B_1
80
0.252
0.723
0.155
0.155
Charles R Lowman
56_B_2
235
SCR
2008
0.302
0.075
0.302
0.075
Crist
641_B_4
75
SNCR
0.285
0.285
0.139
0.139
Crist
641_B_5
75
SNCR
0.285
0.285
0.139
0.139
Crist
641_B_6
291
SCR
2012
0.075
0.075
0.075
0.075
Crist
641_B_7
465
SCR
2004
0.075
0.075
0.075
0.075
Gorgas
8_B_10
703
SCR
2002
0.100
0.100
0.100
0.100
Gorgas
8_B_8
161
0.355
0.296
0.355
0.296
Gorgas
8_B_9
170
0.355
0.296
0.355
0.296
Clifty Creek
983_B_4
196
SCR
2003
0.260
0.075
0.260
0.075
Clifty Creek
983_B_5
196
SCR
2003
0.258
0.075
0.258
0.075
Clifty Creek
983_B_6
196
0.325
0.309
0.325
0.309
3-32
-------
Table 3-19 Renewable Portfolio Standards in EPA Platform v6
State Renewable Portfolio Standards in % - AEO 2018
State
2021
2023
2025
2030
2035
2040
2045
2050
Arizona
6.3%
7.4%
8.5%
8.5%
8.5%
8.5%
8.5%
8.5%
California
34.8%
38.3%
41.7%
50.0%
50.0%
50.0%
50.0%
50.0%
Colorado
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
21.2%
Connecticut
26.5%
30.0%
34.0%
44.0%
44.0%
44.0%
44.0%
44.0%
District of Columbia
20.0%
20.0%
26.0%
42.0%
50.0%
50.0%
50.0%
50.0%
Delaware
15.2%
16.6%
18.1%
18.1%
18.1%
18.1%
18.1%
18.1%
Iowa
0.6%
0.6%
0.6%
0.6%
0.6%
0.6%
0.5%
0.5%
Illinois
9.8%
11.5%
13.1%
14.0%
14.0%
14.0%
14.0%
14.0%
Massachusetts
21.5%
23.5%
25.5%
30.5%
35.5%
40.5%
45.5%
50.5%
Maryland
25.0%
25.0%
25.0%
25.0%
25.0%
25.0%
25.0%
25.0%
Maine
40.0%
40.0%
40.0%
40.0%
40.0%
40.0%
40.0%
40.0%
Michigan
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
15.0%
Minnesota
25.7%
25.7%
28.4%
28.4%
28.4%
28.4%
28.4%
28.4%
Missouri
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
10.6%
Montana
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
10.4%
North Carolina
7.0%
7.0%
7.0%
7.0%
7.0%
7.0%
7.0%
7.0%
New Hampshire
19.8%
21.2%
23.0%
23.0%
23.0%
23.0%
23.0%
23.0%
New Jersey
28.6%
35.6%
42.3%
54.7%
53.6%
53.6%
53.6%
53.6%
New Mexico
15.8%
15.8%
15.8%
15.8%
15.8%
15.8%
15.8%
15.8%
Nevada
17.3%
17.3%
21.9%
21.9%
21.9%
21.9%
21.9%
21.9%
New York
25.3%
28.9%
32.5%
41.4%
41.4%
41.4%
41.4%
41.4%
Ohio
6.7%
8.5%
10.2%
11.1%
11.1%
11.1%
11.1%
11.1%
Oregon
14.1%
14.1%
21.0%
27.6%
36.1%
41.1%
42.6%
42.6%
Pennsylvania
8.0%
8.0%
8.0%
8.0%
8.0%
8.0%
8.0%
8.0%
Rhode Island
17.5%
20.5%
23.5%
31.0%
38.5%
38.5%
38.5%
38.5%
Texas
4.3%
4.2%
4.1%
3.9%
3.7%
3.5%
3.4%
3.2%
Vermont
62.4%
67.6%
68.8%
79.8%
85.0%
85.0%
85.0%
85.0%
Washington
11.8%
11.8%
11.8%
11.8%
11.8%
11.8%
11.8%
11.8%
Wisconsin
9.6%
9.6%
9.6%
9.6%
9.6%
9.6%
9.6%
9.65%
State
RPS Solar
Carve-outs
State
2021
2023
2025
2030
2035
2040
2045
2050
District of Columbia
1.9%
2.5%
2.9%
4.5%
5.0%
5.0%
5.0%
5.0%
Delaware
1.8%
2.2%
2.5%
2.5%
2.5%
2.5%
2.5%
2.5%
Illinois
1.05%
1.23%
1.41%
1.50%
1.50%
1.50%
1.50%
1.50%
Massachusetts
0.17%
0.18%
0.20%
0.24%
0.28%
0.32%
0.36%
0.40%
Maryland
2.50%
2.50%
2.50%
2.50%
2.50%
2.50%
2.50%
2.50%
Minnesota
1.19%
1.19%
1.19%
1.19%
1.19%
1.19%
1.19%
1.19%
Missouri
0.21%
0.21%
0.21%
0.21%
0.21%
0.21%
0.21%
0.21%
North Carolina
0.11%
0.11%
0.11%
0.11%
0.11%
0.11%
0.11%
0.11%
New Hampshire
0.70%
0.70%
0.70%
0.70%
0.70%
0.70%
0.70%
0.70%
New Jersey
5.10%
5.10%
4.80%
2.21%
1.10%
1.10%
1.10%
1.10%
New Mexico
3.17%
3.17%
3.17%
3.17%
3.17%
3.17%
3.17%
3.17%
Nevada
1.04%
1.04%
1.31%
1.31%
1.31%
1.31%
1.31%
1.31%
Ohio
0.27%
0.34%
0.41%
0.45%
0.45%
0.45%
0.45%
0.45%
Pennsylvania
0.50%
0.50%
0.50%
0.50%
0.50%
0.50%
0.50%
0.50%
Note 1: The Renewable Portfolio Standard percentages are applied to modeled electricity sale projections.
Note 2: North Carolina standards are adjusted to account for swine waste and poultry waste set-asides.
3-33
-------
List of tables and attachments that are uploaded directly to the web:
Table 3-20 Regional Net Internal Demand in EPA Platform v6
Table 3-21 Annual Transmission Capabilities of U.S. Model Regions in EPA Platform v6 - 2021
Table 3-22 Turndown Assumptions for Coal Steam Units in EPA Platform v6
Table 3-23 State Power Sector Regulations included in EPA Platform v6
Table 3-24 New Source Review (NSR) Settlements in EPA Platform v6
Table 3-25 State Settlements in EPA Platform v6
Table 3-26 Citizen Settlements in EPA Platform v6
Table 3-27 Complete Availability Assumptions in EPA Platform v6
Table 3-28 BART Regulations included in EPA Platform v6
Attachment 3-1 NOx Rate Development in EPA Platform v6
3-34
-------
4. Generating Resources
Existing, planned-committed, and potential are the three types of generating units modeled in EPA
Platform v6. Electric generating units currently in operation are termed as existing units. Units that are
anticipated to be in operation in the near future, for having broken ground or secured financing, are
planned-committed units. Potential units refer to new generating options that IPM builds to meet industry
capacity expansion projections. Existing and planned-committed units enter IPM as exogenous inputs,
whereas potential units are endogenous to IPM in that the model determines the location and size of the
potential units to build.
This chapter is organized as follows.
(1) Section 4.1 provides background information on the National Electric Energy Data System (NEEDS),
the database that serves as the repository for information on existing and planned-committed electric
generating units modeled,
(2) Section 4.2 provides detailed information on existing non-nuclear generating units,
(3) Section 4.3 provides detailed information on planned-committed units,
(4) Section 4.4 provides detailed information on potential units, and
(5) Section 4.5 describes assumptions pertaining to existing and potential nuclear units.
4.1 National Electric Energy Data System (NEEDS)
EPA Platform v6 uses the NEEDS v6 database as its source for data on all existing and planned-
committed units. Section 4.2 discusses the sources used in developing data on existing units. The
population of existing units in the NEEDS v6 represents electric generating units that were in operation
through the end of 2017. Section 4.3 discusses the sources used in developing data on planned-
committed units. The population of planned-committed includes units online or scheduled to come online
from 2018 through June 30, 2021, with the exception of Vogtle nuclear units 3 and 4 that are scheduled to
come online after 2021.
4.2 Existing Units
The sections below describe the procedures for determining the population of existing units in NEEDS v6,
as well as the capacity, location, and configuration information of each unit in the population. Details are
also given on the model plant aggregation scheme and associated cost and performance characteristics
of the units.
4.2.1 Population of Existing Units
The October 2017 EIA Form 860M is the primary data source on existing units. Table 4-2 specifies the
screening rules applied to the data source to ensure data consistency and adaptability for use in EPA
Platform v6. Table 4-50 lists all units that are excluded from the NEEDS v6 population based on
application of the screening rules.
4-1
-------
Table 4-1 Data Sources for NEEDS v6 for EPA Platform v6
Data Source1
Data Source Documentation
EIA Form 860
EIA Form 860 is an annual survey of utility and non-utility power plants at the generator
level. It contains data such as summer, winter, and nameplate capacity, location (state
and county), operating status, prime mover, energy sources and in-service date of
existing and proposed generators. NEEDS v6 uses the annual 2015 EIA Form 860,
annual 2016 Early Release EIA Form 860, 2017 Early Release EIA Form 860, May 2017
EIA Form 860M, October 2017 EIA Form 860M and the July 2018 EIA Form 860M as the
primary generator data inputs.
EIA Form 860 also collects data of steam boilers such as energy sources, boiler
identification, location, operating status and design information; and associated
environmental equipment such as NOx combustion and post-combustion controls, FGD
scrubber, mercury control and particulate collector device information. Note that boilers in
plants with less than 10 MW do not report all data elements. The association between
boilers and generators is also provided. Note that boilers and generators are not
necessarily in a one-to-one correspondence. NEEDS v6 uses 2015 EIA Form 860 and
2016 Early Release EIA Form 860 as the primary boiler data inputs.
ElA's Annual Energy
Outlook (AEO)
The Energy Information Administration (EIA) Annual Energy Outlook presents annually
updated forecasts of energy supply, demand and prices covering a 30-year time horizon.
The projections are based on results from ElA's National Energy Modeling System
(NEMS). Information from AEO 2017 such as heat rates and planned-committed units
were used in NEEDS v6.
EPA's Emission
Tracking System
The Emission Tracking System (ETS) database is updated quarterly. It contains
information including primary fuel, heat input, SO2, NOx, Mercury, and HCI controls, and
SO2 and NOx emissions. NEEDS v6 uses annual and seasonal ETS (2017) data as one
of the primary data inputs for NOx rate development and environmental equipment
assignment.
Utility and Regional
EPA Office Comments
Comments from utilities, regional EPA offices and other stakeholders regarding the prior
versions of NEEDS.
Note:
1 Shown in Table 4-1 are the primary issue dates of the indicated data sources used. Other vintages of these data
sources were also used in instances where data were not available for the indicated issued date, or where there were
methodological reasons for using other vintages of the data.
Table 4-2 Rules Used in Populating NEEDS v6 for EPA Platform v6
Scope
Rule
Capacity
Excluded units with reported summer capacity, winter capacity and nameplate capacity of zero or
blank.
Status
Excluded units that were out of service for three consecutive years (i.e., generators or boilers with
status codes "OS" or "OA" in the latest three reporting years) and units that were no longer in
service and not expected to be returned to service (i.e., generators or boilers with status codes of
"RE"). Status of boiler(s) and associated generator(s) were taken into account for determining
operation status
Planned or
Committed
Units
Included planned units that had broken ground or secured financing and were expected to be
online by June 30, 2021; two nuclear units that are scheduled to come online after 2021 were also
included
Firm/Non-firm
Electric Sales
Excluded non-utility onsite generators that do not produce electricity for sale to the grid on a net
basis
Excluded all mobile and distributed generators
Note:
The two nuclear units are Vogtle, units 3&4
The NEEDS v6 includes steam units at the boiler level and non-steam units at the generator level
(nuclear units are also at the generator level). A unit in NEEDS v6, therefore refers to a boiler in the case
of a steam unit and a generator in the case of a non-steam unit. Table 4-3 provides a summary of the
population and capacity of the existing units included in NEEDS v6 through 2017. The final population of
4-2
-------
existing units is supplemented based on information from other sources, including comments from
utilities, submissions to EPA's Emission Tracking System, Annual Energy Outlook and other research.
EPA Platform v6 removes units from the NEEDS inventory based on public announcements of future
closures. The removal of such units pre-empts IPM from making any further decisions regarding the
operational status or configuration of the units. The units considered for removal from NEEDS are
identified from reviewing several data sources including:
1. EIA Electric Generator Capacity data (EIA Form 860M), July 2018 release
2. PJM Future Deactivation Requests and PJM Generator Deactivations, July 2018 (updated frequently)
3. ERCOT Generator Interconnection Status Report, July 2018 (updated frequently)
4. MISO Generation Interconnection Queue, July 2018 (updated frequently)
5. Research by EPA and ICF staff
Units are removed from the NEEDS inventory only if a high degree of certainty could be assigned to
future implementation of the announced action. The available retirement-related information was
reviewed for each unit, and the following rules are applied to remove:
1. Units that are listed as retired in the July 2018 EIA Form 860M
2. Units with a planned retirement year prior to June 30, 2021 in July 2018 EIA Form 860M
3. Units that have been cleared by a regional transmission operator (RTO) or independent system
operator (ISO) to retire before 2021, or whose RTO/ISO clearance to retire is contingent on actions
that can be completed before 2021
4. Units that have committed specifically to retire before 2021 under federal or state enforcement
actions or regulatory requirements
5. And finally, units for which a retirement announcement can be corroborated by other available
information.
Units required to retire pursuant to enforcement actions or state rules in 2022 or later are retained in
NEEDS v6. Such 2022-or-later retirements are captured as constraints on those units in IPM modeling,
and the units are retired in future year projections per the terms of the related requirements. Table 4-50
and Table 4-51 list all units that are removed from the NEEDS v6 inventory.
Table 4-3 Summary Population (through 2017) of Existing Units in NEEDS v6
Plant Type Number of Units Capacity (MW)
Biomass
186
3,876
Coal Steam
593
226,339
Combined Cycle
1,837
246,866
Combustion Turbine
5,381
143,285
Energy Storage
81
659
Fossil Waste
81
1,049
Fuel Cell
72
130
Geothermal
164
2,396
Hydro
3,805
79,186
IGCC
5
815
Landfill Gas
1,576
1,913
Municipal Solid Waste
165
2,123
Non-Fossil Waste
216
2,027
Nuclear
90
92,260
4-3
-------
Plant Type
Number of Units
Capacity (MW)
O/G Steam
443
74,999
Offshore Wind
1
29
Onshore Wind
1,185
87,185
Pumped Storage
148
22,196
Solar PV
2,452
24,144
Solar Thermal
16
1,754
Tires
2
52
US Total
18,499
1,013,283
4.2.2 Capacity
The unit capacity data implemented in NEEDS v6 reflects net summer dependable capacity31. Table 4-4
summarizes the hierarchy of data sources used in compiling capacity data. In other words, capacity
values are taken from a particular source only if the sources listed above it do not provide adequate data
for the unit in question.32
Table 4-4 Hierarchy of Data Sources for Capacity in NEEDS v6
Sources Presented in Hierarchy
Net Summer Capacity from Comments / ICF Research
July 2018 EIA Form 860M Net Summer Capacity
October 2017 EIA Form 860M Net Summer Capacity
May 2017 EIA Form 860M Net Summer Capacity
2015 EIA Form 860 Net Summer Capacity
Notes:
If the capacity of a unit is zero MW, the unit is excluded from NEEDS population.
As noted earlier, NEEDS v6 includes boiler-level data for steam units and generator-level data for non-
steam units. Capacity data in EIA Form 860 are generator-specific, not boiler-specific. Therefore, it was
necessary to develop an algorithm for parsing generator-level capacity to the boiler level for steam
producing units.
The capacity-parsing algorithm used for steam units in NEEDS v6 took into account boiler-generator
mapping. Fossil steam electric units have boilers attached to generators that produce electricity. There
are generally four types of links between boilers and generators: one boiler to one generator, one boiler to
many generators, many boilers to one generator and many boilers to many generators.
The capacity-parsing algorithm used for steam units in NEEDS v6 utilizes steam flow data with the boiler-
generator mapping. Under EIA Form 860, steam units report the maximum steam flow from the boiler to
the generator. There is, however, no further data on the steam flow of each boiler-generator link.
Instead, EIA Form 860 contains only the maximum steam flow for each boiler. Table 4-5 summarizes the
algorithm used for parsing capacity with data on maximum steam flow and boiler-generator mapping. In
Table 4-5, MFB, refers to the maximum steam flow of boiler /' and MWg; refers to the capacity of generator
j. The algorithm uses the available data to derive the capacity of a boiler, referred to as MWb; in Table
4-5.
31 As used here, net summer dependable capacity is the net capability of a generating unit in megawatts (MW) for
daily planning and operation purposes during the summer peak season, after accounting for station or auxiliary
services.
32 EIA Form 860M (July, 2018 release) was the most recent data available at the time when NEEDS v6 was finalized.
4-4
-------
Table 4-5 Capacity-Parsing Algorithm for Steam Units in NEEDS v6
Type of Boiler-Generator Links
For Boiler B1 to BN linked
to Generators G1 to GN
One-to-One
One-to-Many
Many-to-One
Many-to-Many
MWBi =
MWGj
MWBi =
IjMWGj
MWBi =
(MFBi / ZiMFBi) * MWGj
MWBi =
(MFBi / ZiMFBi) * ZjMWGj
Notes:
MFb, = maximum steam flow of boiler /'
MWq; = electric generation capacity of generator j
Since EPA Platform v6 uses net energy for load as demand, the NEEDS includes only generators that
sell the majority of their power to the electric grid. The approach is intended to be broadly consistent with
the generating capacity used in the AEO projections where demand is net energy for load. The
generators that should be in NEEDS v6 by this qualification are determined from the 2014 EIA Form 923
non-utility source and disposition data set.
4.2.3 Plant Location
The physical location of each unit in NEEDS is represented by the unit's model region, state, and county
data.
State and County
NEEDS v6 uses the state and county data from October 2017 EIA Form 860M.
Model Region
For each unit, the associated model region was derived based on NERC assessment regions reported in
EIA Form 860 and ISO/RTO reports. For units with no NERC assessment region data, state and county
data were used to derive associated model regions. Table 3-1 in Chapter 3 provides a summary of the
mapping between NERC assessment regions and EPA Platform v6 model regions.
4.2.4 Online Year
The EPA Platform v6 uses online year to capture when a unit entered service. NEEDS includes online
years for all units in the population. Online years for boilers were from 2015 EIA Form 860, and online
years for generators were derived primarily from reported in-service dates in May 2017 version of EIA
Form 860M.
EPA Platform v6 includes constraints to set the retirement year for generating units that are firmly
committed to retiring after June 30, 2021 based on state or federal regulations and enforcement actions.
In addition, existing nuclear units must retire when they reach age 80. (See section 3.7 for a discussion
of the nuclear lifetime assumption.) Economic retirement options are also provided to coal, oil and gas
steam, combined cycle, combustion turbines, biomass, and nuclear units to allow the model the option to
retire a unit if it finds economical to do so. In IPM, a retired unit ceases to incur FOM and VOM costs.
The unit, however, continues to make annualized capital cost payment on any previously incurred capital
cost for model-installed retrofits projected prior to retirement.
4.2.5 Unit Configuration
Unit configuration refers to the physical specification of a unit's design. Unit configuration in EPA
Platform v6 drives model plant aggregation and modeling of pollution control options and mercury
emission modification factors. NEEDS v6 contains for each unit, data on the firing and bottom type, as
well as existing and committed emission controls the unit has. Table 4-6 shows the hierarchy of data
sources used in determining a unit configuration. The sources listed below are also supplemented by
4-5
-------
recent ICF and EPA research to ensure the unit configuration data in NEEDS is the most comprehensive
and up-to-date possible.
Table 4-6 Data Sources for Unit Configuration in NEEDS v6
Unit
Component
Primary Data
Source
Secondary Data Source
Tertiary Data
Source
Other
Sources
Default
Firing Type
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
-
-
-
Bottom Type
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
-
-
Dry
SO2 Pollution
Control
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
NSR Settlement
or Comments
-
No
Control
NOx Pollution
Control
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
NSR Settlement
or Comments
-
No
Control
Particulate
Matter Control
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
NSR Settlement
or Comments
-
-
Mercury Control
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
NSR Settlement
or Comments
-
-
HCL Control
2015/2016 EIA 860
EPA's Emission Tracking
System (ETS) - 2015/2017
NSR Settlement
or Comments
-
-
4.2.6 Model Plant Aggregation
While EPA Platform v6 using IPM is comprehensive in representing all the units contained in NEEDS v6,
an aggregation scheme is used to combine existing units with similar characteristics into model plants.
The aggregation scheme serves to reduce the size of the model, making the model manageable while
capturing the essential characteristics of the generating units. The aggregation scheme is designed so
that each model plant represents only generating units from a single state. The design makes it possible
to obtain state-level results directly from IPM outputs. In addition, the aggregation scheme supports the
modeling of plant-level emission limits on fossil generation.
The aggregation scheme encompasses a variety of different classification categories including location,
size, technology, heat rate, fuel choices, unit configuration, SO2 emission rates, and environmental
regulations among others. Units are aggregated together only if they match on all the different categories
specified for the aggregation. The 11 major categories used for the aggregation scheme in EPA Platform
v6 are the following.
(1) Model Region
(2) Unit Technology Type
(3) Cogen
(4) Fuel Demand Region
(5) Applicable Environmental Regulations
(6) State
(7) Facility (ORIS) for fossil units
(8) Unit Configuration
(9) Heat Rates
(10) Fuel
(11) Size
4-6
-------
Table 4-7 shows the number of actual units by generation technology type and the related number of
aggregated model plants in the EPA Platform v6. For each plant type, the table shows the number of
generating units and the number of model plants representing the generating units.33
Table 4-7 Aggregation Profile of Model Plants as Provided at Set up of EPA Platform v6
Existing and Planned/Committed Units
Plant Type
Number of Units
Number of IPM Model Plants
Biomass
300
165
Coal Steam
675
527
Combined Cycle
2,032
891
Combustion Turbine
5,988
2,535
Energy Storage
85
41
Fossil Waste
86
25
Fuel Cell
72
35
Geothermal
174
31
Hydro
5,455
252
IGCC
5
2
IMPORT
1
1
Landfill Gas
1,643
307
Municipal Solid Waste
166
60
Non-Fossil Waste
267
140
Nuclear
115
115
O/G Steam
590
399
Offshore Wind
1
1
Onshore Wind
1,570
89
Pumped Storage
155
27
Solar PV
2,532
98
Solar Thermal
17
5
Tires
2
1
Total
21,931
5,747
New Units
Plant Type
Number of IPM Model Plants
New Battery Storage
168
New Biomass
134
33 (1) The "Number of IPM Model Plants" shown for many of the "Plant Types" in the "Retrofits" block in Table 4-7
exceeds the "Number of IPM Model Plants" shown for "Plant Type" "Coal Steam" in the block labeled "Existing and
Planned - Committed Units", because a particular retrofit "Plant Type" can include multiple technology options and
multiple timing options (e.g., Technology A in Stage 1 + Technology B in Stage 2 + Technology C in Stage 3, the
reverse timing, or multiple technologies simultaneously in Stage 1).
(2) Since only a subset of coal plants is eligible for certain retrofits, many of the "Plant Types" in the "Retrofits" block
that represent only a single retrofit technology (e.g., "Retrofit Coal with SNCR") have a "Number of IPM Model Plants"
that is a smaller than the "Number of IPM Model Plants" shown for "Plant Type" "Coal Steam".
(3) The total number of model plants representing different types of new units often exceeds the 67 U.S. model
regions and varies from technology to technology for several reasons. First, some technologies have multiple
vintages (i.e., different cost and/or performance parameters depending on which run-year in which the unit is
created), which must be represented by separate model plants in each IPM region. Second, some technologies are
not available in particular regions (e.g., geothermal is geographically restricted to certain regions).
4-7
-------
New Combined Cycle
456
New Combined Cycle with Carbon Capture
228
New Combustion Turbine
456
New Fuel Cell
150
New Geothermal
93
New Hydro
153
New Landfill Gas
379
New Nuclear
132
New Offshore Wind
894
New Onshore Wind
5,358
New Solar PV
1,373
New Solar Thermal
261
New Ultrasupercritical Coal with 30% CCS
266
New Ultrasupercritical Coal with 90% CCS
266
New Ultrasupercritical Coal without CCS
138
Total
10,905
Retrofits
Plant Type
Number of IPM Model Plants
Retrofi
Coal
with ACI
74
Retrofi
Coal
with ACI +
CCS
92
Retrofi
Coal
with ACI +
CCS + HRI
92
Retrofi
Coal
with ACI +
CCS + HRI + SCR
20
Retrofi
Coal
with ACI +
CCS + HRI + SNCR
29
Retrofi
Coal
with ACI +
CCS + SCR
20
Retrofi
Coal
with ACI +
DS
20
Retrofi
Coal
with ACI +
DS
+ HRI
20
Retrofi
Coal
with ACI +
DS
+ HRI + SCR
31
Retrofi
Coal
with ACI +
DS
+ HRI + SCR + Scrubber
22
Retrofi
Coal
with ACI +
DS
+ HRI + Scrubber
18
Retrofi
Coal
with ACI +
DS
+ HRI + Scrubber + SNCR
14
Retrofi
Coal
with ACI +
DS
+ HRI + SNCR
27
Retrofi
Coal
with ACI +
DS
+ SCR
31
Retrofi
Coal
with ACI +
DS
+ SCR + Scrubber
22
Retrofi
Coal
with ACI +
DS
+ Scrubber
18
Retrofi
Coal
with ACI +
DS
+ Scrubber + SNCR
14
Retrofi
Coal
with ACI +
DS
+ SNCR
31
Retrofi
Coal
with ACI +
HR
74
Retrofi
Coal
with ACI +
HR
+ SCR
62
Retrofi
Coal
with ACI +
HR
+ SCR + Scrubber
62
Retrofi
Coal
with ACI +
HR
+ Scrubber
53
Retrofi
Coal
with ACI +
HR
+ Scrubber + SNCR
74
Retrofi
Coal
with ACI +
HR
+ SNCR
61
Retrofi
Coal
with ACI +
SCR
62
Retrofi
Coal
with ACI +
SCR + Scrubber
62
Retrofi
Coal
with ACI +
Scrubber
52
4-8
-------
Retrofi
Coa
with ACI + Scrubber + SNCR
75
Retrofi
Coa
with ACI + SNCR
62
Retrofi
Coa
with C2G
454
Retrofi
Coa
with C2G + SCR
454
Retrofi
Coa
with CCS
791
Retrofi
Coa
with CCS + HRI
788
Retrofi
Coa
with CCS + HRI + SCR
252
Retrofi
Coa
with CCS + HRI + SCR + Scrubber
208
Retrofi
Coa
with CCS + HRI + Scrubber
232
Retrofi
Coa
with CCS + HRI + Scrubber + SNCR
152
Retrofi
Coa
with CCS + HRI + SNCR
180
Retrofi
Coa
with CCS + SCR
255
Retrofi
Coa
with CCS + SCR + Scrubber
212
Retrofi
Coa
with CCS + Scrubber
240
Retrofi
Coa
with CCS + Scrubber + SNCR
156
Retrofi
Coa
with CCS + SNCR
183
Retrofi
Coa
with DSI
21
Retrofi
Coa
with DSI + HRI
70
Retrofi
Coa
with DSI + HRI + SCR
75
Retrofi
Coa
with DSI + HRI + SCR + Scrubber
21
Retrofi
Coa
with DSI + HRI + Scrubber
26
Retrofi
Coa
with DSI + HRI + SNCR
69
Retrofi
Coa
with DSI + SCR
109
Retrofi
Coa
with DSI + SCR + Scrubber
33
Retrofi
Coa
with DSI + Scrubber
38
Retrofi
Coa
with DSI + SNCR
103
Retrofi
Coa
with HRI
482
Retrofi
Coa
with HRI + SCR
432
Retrofi
Coa
with HRI + SCR + Scrubber
450
Retrofi
Coa
with HRI + Scrubber
357
Retrofi
Coa
with HRI + Scrubber + SNCR
408
Retrofi
Coa
with HRI + SNCR
342
Retrofi
Coa
with SCR
242
Retrofi
Coa
with SCR + Scrubber
582
Retrofi
Coa
with Scrubber
224
Retrofi
Coa
with Scrubber + SNCR
544
Retrofi
Coa
with SNCR
203
Retrofi
Combined Cycle with CCS
2787
Retrofi
Oil/Gas steam with SCR
222
Total
13,691
Retirements
Plant Type
Number of IPM Model Plants
Biomass Retirement
165
CC Retirement
891
Coal Retirement
5,394
4-9
-------
CT Retirement
2,535
Geothermal Retirement
31
Hydro Retirement
252
IGCC Retirement
2
Landfill Gas Retirement
307
Nuke Retirement
115
Oil/Gas steam Retirement
1,075
Total
10,767
Grand Total (Existing and Planned/Committed + New + Retrofits + Retirements):41,110
4.2.7 Cost and Performance Characteristics of Existing Units34
In EPA Platform v6, the cost and performance characteristics of an existing unit are determined by the
unit's heat rates, emission rates, variable operation and maintenance cost (VOM), and fixed operation
and maintenance costs (FOM). For existing units, only the cost of maintaining (FOM) and running (VOM)
the unit are modeled because capital costs and all related carrying capital charges are sunk, and hence,
economically irrelevant for projecting least-cost investment and operational decisions going forward. The
section below discusses the cost and performance assumptions for existing units used in the EPA
Platform v6.
Variable Operating and Maintenance Cost (VOM)
VOM represents the non-fuel variable cost associated with producing electricity. If the generating unit
contains pollution control equipment, VOM includes the cost of operating the control equipment. Table
4-8 below summarizes VOM assumptions used in EPA Platform v6. The following further discusses the
components of VOM costs and the VOM modeling methodology.
Variable O&M Approach: EPA Platform v6 uses a modeling construct termed as Segmental VOM to
capture the variability in operation and maintenance costs that are treated as a function of the unit's
dispatch pattern. Generally speaking the construct captures costs associated with major maintenance
and consumables. The VOM for combined cycles and combustion turbine units includes the costs of both
major maintenance and consumables while for coal steam and oil/gas steam units includes only the cost
of consumables. The VOM cost of various emission control technologies is also incorporated.
Major maintenance: Major maintenance costs are those required to maintain a unit at its delivered
performance specifications and whose terms are usually dictated through its long term service agreement
(LTSA). The three main areas of maintenance for gas turbines include combustion inspection, hot gas
path inspection, and major inspections. All of these costs are driven by the hours of operation and the
number of starts that are incurred within that time period of operation. In a cycling or mid-merit type mode
of operation, there are many starts, accelerating the approach of an inspection. As more starts are
incurred compared to the generation produced, cost per generation increase. For base load operation
there are fewer starts spread over more generation, lowering the cost per generation. While this
nomenclature is for gas-turbine based systems, steam turbine based systems have a parallel construct.
Consumables: The model captures consumable costs, as purely a function of output and does not varies
across the segmented time-period. In other words, the consumables cost component is held constant
over both peak and off-peak segments. Consumables include chemicals, lube oils, make-up water, waste
water disposal, reagents, and purchased electricity.
34 All units excluding nuclear units.
4-10
-------
Data Sources for Gas-Turbine Based Prime Movers:
ICF has engaged its deep expertise in operation & maintenance costs for these types of prime movers to
develop generic variable O&M costs as a function of technology.
As mentioned above the variable O&M for gas-turbine based systems tracks Long Term Service
Agreement costs, start-up and consumables.
Data Sources for Stand-Alone Steam Turbine Based Prime Movers:
The value levels of non-fuel variable O&M data for stand-alone steam turbine plants is based on ICF
experience. The VOM cost adders of various emission control technologies are based on cost functions
described in Chapter 5.
Table 4-8 VOM Assumptions in EPA Platform v6
Capacity Type
SO2 Control
NOx Control
Hg Control
Variable O&M
(2016$/mills/kWh)
Biomass
-
-
-
7.29
No NOx Control
No Hg Control
1.43
AC I
2.90
No SO2 Control
SCR
No Hg Control
2.39
AC I
3.86
SNCR
No Hg Control
2.36
AC I
3.83
No NOx Control
No Hg Control
3.5
AC I
4.97
Dry FGD
SCR
No Hg Control
4.46
AC I
5.93
SNCR
No Hg Control
4.43
Coal Steam
AC I
5.9
No NOx Control
No Hg Control
3.95
AC I
5.43
Wet FGD
SCR
No Hg Control
4.91
AC I
6.39
SNCR
No Hg Control
4.88
AC I
6.35
No NOx Control
No Hg Control
8.21
AC I
9.68
DSI
SCR
No Hg Control
9.17
AC I
10.64
SNCR
No Hg Control
9.14
AC I
10.61
No NOx Control
1.98-3.78
Combined Cycle
No SO2 Control
SCR
No Hg Control
2.12-3.92
SNCR
2.61 -4.41
No NOx Control
3.31 -15.7
Combustion Turbine
No SO2 Control
SCR
No Hg Control
3.45 -15.84
SNCR
3.94 -16.33
4-11
-------
Capacity Type
SO2 Control
NOx Control
Hg Control
Variable O&M
(2016$/mills/kWh)
Fuel Cell
-
-
-
44.91
Geothermal
-
-
-
5.49
Hydro
-
-
-
2.66
IGCC
-
-
-
2.28-4.04
Landfill Gas / Municipal
Solid Waste
-
-
-
6.54
Oil/gas Steam
No SO2 Control
No NOx Control
No Hg Control
0.83
SCR
0.97
SNCR
1.46
Pumped Storage
-
-
-
10.17
Solar PV
-
-
-
0
Solar Thermal
-
-
-
0
Wind
-
-
-
0
Fixed Operation and Maintenance Cost (FOM)
FOM represents the annual fixed cost of maintaining a unit. FOM costs are incurred independent of
generation levels and signify the fixed cost of operating and maintaining the unit's availability to provide
generation. Table 4-9 summarizes the FOM assumptions35. Note that FOM varies by the age of the unit,
and the total FOM cost incurred by a unit depends on its capacity size. The values appearing in this table
include the cost of maintaining any associated pollution control equipment. The values in Table 4-9 are
based on FERC (Federal Energy Regulatory Commission) Form 1 data maintained by SNL and ICF
research. The following further discusses the procedure for developing the FOM costs.
Stand Alone - Steam Turbines Based Prime Movers
O&M cost data for existing coal and oil/gas steam units were developed starting with FERC Form 1 data
sets from the years 2011 to 2016. The FERC Form-1 database does not explicitly report separate fixed
and variable O&M expenses. In deriving Fixed O&M costs, generic variable O&M costs are assigned to
each individual power plant. Next, the assumed variable O&M cost is subtracted from the total O&M
reported by FERC Form-1 to calculate a starting point for fixed O&M. Thereafter, other cost items which
are not reported by FERC Form-1 are added to the raw FOM starting point. These unreported cost items
are selling, general, and administrative expenses (SG&A), property taxes, insurance, and routine capex.
A detailed description of the fixed O&M derivation methodology is provided below.
35 Cogen units whose primary purpose is to provide process heat are called as bottoming cycle units and are
identified based on Form EIA 860. Such units are provided a FOM of zero in EPA Platform v6. This is to
acknowledge the fact that the economics of such a unit cannot be comprehensively modeled in a power sector
focused model.
4-12
-------
Figure 4-1 Derivation of Plant Fixed O&M Data
Get FERC
FORM -1
O&M data
Calculate
FOM by
subtracting
non-fuel
VOM from
O&M
Add SG&A,
routine
Cap Ex,
property taxes
and insurance
i) Assign generic VOM cost to each unit in FERC Form 1 based on the control configuration.
Subtract this VOM from the total O&M cost from FERC Form 1 to calculate raw FOM cost.
The FOM cost of operating the existing controls is estimated based on cost functions in
Chapter 5 and deducted from the raw FOM cost. Aggregate this unit level raw FOM cost
data into age based categories. The weighted average raw FOM costs for uncontrolled units
by age group is the output of this step and is used as the starting point for subsequent steps.
ii) An owner/operator fee for SG&A services in the range of 20-30% is added to raw fixed O&M
figures in step 1.
iii) Property tax and insurance cost estimates in $/kW-year are also added. These figures vary
by plant type.
iv) A generic percentage value to cover routine capex is added to raw fixed O&M figures in step
1. The percentage varies by prime mover and is based on a review of FERC Form 1 data
v) Finally, generic FOM cost adders for various emission control technologies are estimated
using cost functions described in Chapter 5. Based on the emission control configuration of
each unit in NEEDS, the appropriate emission control cost adder is added to the FOM cost of
an uncontrolled unit from step iv.
The fixed O&M derivation approach relies on top-down derivation of fixed costs based on FERC Form-1
data and ICF's own non-fuel variable O&M, SG&A, routine capex, property tax, and insurance.
Gas-Turbine Based Prime Movers
Similar to the stand-alone steam turbine based prime movers, the Fixed O&M for gas-turbine based
systems tracks: labor, routine maintenance, property taxes, insurance, owner/operator SG&A, and routine
capital expenditures. These generic Fixed O&M costs as a function of technology are based on ICF's
deep expertise in fixed O&M costs for these types of prime movers
Table 4-9 FOM Assumptions in EPA Platform v6
Plant Type
SO2 Control
NOx Control
Hg Control
Age of Unit
FOM (2016$ /kW-
Yr)
Biomass
—
—
—
All Years
134.52
No Hg
Control
0 to 40 Years
28.34
40 to 50 Years
32.4
No NOx
Greater than 50 Years
41.63
Control
0 to 40 Years
28.42
No SO2
Control
AC I
40 to 50 Years
32.49
Coal Steam
Greater than 50 Years
41.72
No Hg
Control
0 to 40 Years
29.12
40 to 50 Years
33.18
SCR
Greater than 50 Years
42.41
AC I
0 to 40 Years
29.2
40 to 50 Years
33.27
4-13
-------
Plant Type
SO2 Control
NOx Control
Hg Control
Age of Unit
FOM (2016$ /kW-
Yr)
Greater than 50 Years
42.5
No Hg
Control
0 to 40 Years
28.62
40 to 50 Years
32.69
SNCR
Greater than 50 Years
41.92
0 to 40 Years
28.71
AC I
40 to 50 Years
32.77
Greater than 50 Years
42
No Hg
Control
0 to 40 Years
38
40 to 50 Years
42.06
No NOx
Greater than 50 Years
51.29
Control
0 to 40 Years
38.08
AC I
40 to 50 Years
42.15
Greater than 50 Years
51.38
No Hg
Control
0 to 40 Years
38.78
40 to 50 Years
42.84
Dry FGD
SCR
Greater than 50 Years
52.07
0 to 40 Years
38.86
AC I
40 to 50 Years
42.93
Greater than 50 Years
52.16
No Hg
Control
0 to 40 Years
38.28
40 to 50 Years
42.35
SNCR
Greater than 50 Years
51.58
0 to 40 Years
38.36
AC I
40 to 50 Years
42.43
Greater than 50 Years
51.66
No Hg
Control
0 to 40 Years
37.59
40 to 50 Years
41.66
No NOx
Greater than 50 Years
50.89
Control
0 to 40 Years
37.68
AC I
40 to 50 Years
41.75
Greater than 50 Years
50.97
No Hg
Control
0 to 40 Years
38.37
40 to 50 Years
42.44
Wet FGD
SCR
Greater than 50 Years
51.67
0 to 40 Years
38.46
AC I
40 to 50 Years
42.53
Greater than 50 Years
51.75
No Hg
Control
0 to 40 Years
37.88
40 to 50 Years
41.95
SNCR
Greater than 50 Years
51.17
0 to 40 Years
37.96
AC I
40 to 50 Years
42.03
Greater than 50 Years
51.26
No Hg
Control
0 to 40 Years
29.7
40 to 50 Years
33.77
DSI
No NOx
Greater than 50 Years
43
Control
0 to 40 Years
29.78
AC I
40 to 50 Years
33.85
Greater than 50 Years
43.08
4-14
-------
Plant Type
SO2 Control
NOx Control
Hg Control
Age of Unit
FOM (2016$ /kW-
Yr)
SCR
No Hg
Control
0 to 40 Years
30.48
40 to 50 Years
34.55
Greater than 50 Years
43.78
AC I
0 to 40 Years
30.57
40 to 50 Years
34.63
Greater than 50 Years
43.86
SNCR
No Hg
Control
0 to 40 Years
29.98
40 to 50 Years
34.05
Greater than 50 Years
43.28
AC I
0 to 40 Years
30.07
40 to 50 Years
34.14
Greater than 50 Years
43.37
Combined Cycle
No SO2
Control
No NOx
Control
No Hg
Control
-
29.19
SCR
No Hg
Control
-
30.54
SNCR
No Hg
Control
-
29.89
Combustion Turbine
No SO2
Control
No NOx
Control
No Hg
Control
-
18.7
SCR
No Hg
Control
-
20.72
SNCR
No Hg
Control
-
19.23
Fuel Cell
—
—
—
All Years
0
Geothermal
—
—
—
All Years
93.51
Hydro
—
—
—
All Years
14.89
IGCC
No SO2
Control
No NOx
Control
-
All Years
102.34
Landfill Gas /
Municipal Solid
Waste
-
-
-
All Years
234.69
Oil/gas Steam
No SO2
Control
No NOx
Control
No Hg
Control
0 to 40 Years
16.94
40 to 50 Years
25.72
Greater than 50 Years
33.51
SCR
No Hg
Control
0 to 40 Years
18.05
40 to 50 Years
26.84
Greater than 50 Years
34.62
SNCR
No Hg
Control
0 to 40 Years
17.15
40 to 50 Years
25.93
Greater than 50 Years
33.72
Pumped Storage
-
-
-
All Years
17.27
Solar PV
-
-
-
All Years
27.99
Solar Thermal
-
-
-
All Years
77.93
Wind
-
-
-
All Years
28.18
4-15
-------
Heat Rates
Heat Rates describe the efficiency of the unit expressed as BTUs per kWh. The treatment of heat rates is
discussed in Section 3.8.
Lifetimes
Unit lifetime assumptions are detailed in Sections 3.7 and 4.2.8.
SO? Rates
Section 3.9.1 contains a detailed discussion of SO2 rates for existing units.
NOy Rates
Section 3.9.2 contains a detailed discussion of NOx rates for existing units.
Mercury Emission Modification Factors (EMF)
Mercury EMF refers to the ratio of mercury emissions (mercury outlet) to the mercury content of the fuel
(mercury inlet). Section 5.4.2 contains a detailed discussion of the EMF assumptions in EPA Platform v6.
Coqeneration Units
For cogeneration units, the dispatch decisions in IPM are only based on the benefits obtained from the
electric portion of a cogeneration unit. In IPM, a cogeneration unit uses a net heat rate, which is
calculated by dividing heat content of fuel consumed for power generation by electricity generated from
this fuel. To capture the total emissions from the cogeneration unit, a multiplier is applied to the power
only emissions. The multiplier is calculated as a ratio between the total heat rate and the net heat rate
where the total heat rate is calculated by dividing the heat content of fuel consumed for power and steam
generation by electricity generated from this fuel.
Coal Switching
Recognizing that boiler modifications and fuel handling enhancements may be required for unrestricted
switching from bituminous to subbituminous coal, and vice versa, the following procedure applies in EPA
Platform v6 to coal units that have the option to burn both bituminous and subbituminous coals.
(i) An examination of the EIA Form 923 coal delivery data for the period 2008-2016 is conducted for each
unit to determine the unit's historical maximum share of bituminous coal and that of subbituminous coal.
For example, if in at least one year during the period 2008-2016 a unit burned 90% or less subbituminous
coal, its historical maximum share of subbituminous coal is set at 90%.
(ii) The following rules then apply.
Blending Subbituminous Coal:
If a unit's historical maximum share of subbituminous coal is greater than 90%, the unit incurs no fuel
switching cost adder to increase its subbituminous coal burn. The unit is assumed to have already made
the fuel handling and boiler investments needed to burn up to 100% subbituminous coal. It would
therefore face no additional cost. In addition, the unit's heat rate is assumed to reflect the impact of
burning the corresponding proportion of subbituminous coal.
If a unit's historical maximum share of subbituminous coal is less than 90%, the unit incurs a heat rate
penalty of 5% and a fuel switching cost adder. The heat rate penalty reflects the impact of the higher
4-16
-------
moisture content subbituminous coal on the unit's heat rate. And the cost adder is designed to cover
boiler modifications, or alternative power purchases in lieu of capacity deratings that would otherwise be
associated with burning subbituminous coal with its lower heating value relative to bituminous coal. The
cost adder is determined as follows:
• If the unit's historical maximum share of subbituminous coal is less than 20%, the unit can burn
up to 20% subbituminous coal at no cost adder. Burning beyond 20% subbituminous coal, the
unit incurs a cost adder of 270 (2016$ per kW).
• If the unit's historical maximum share of subbituminous coal is greater than 20% but less than
90%, the unit can burn up to its historical maximum share of subbituminous coal at no cost adder.
Burning beyond its historical maximum share of subbituminous coal, the unit incurs a cost adder
calculated by the following equation:
Fuel Switching Cost Adder (2016$ per kW) =
270 x
(100 — Historical Maximun Share of Subbituminous)~|
(100 -20) J
Blending Bituminous Coal:
If a unit's historical maximum share of bituminous coal is greater than 90%, the unit incurs no fuel
switching cost adder.
If a unit's historical maximum share of bituminous coal is less than 90%, the unit incurs a fuel switching
cost adder determined as follows:
If the unit's historical maximum share of bituminous coal is less than 20%, the unit can burn up to
20% bituminous coal at no cost adder. Burning beyond 20% bituminous coal, the unit incurs a
cost adder of 54 (2016$ per kW).
If the unit's historical maximum share of bituminous coal is greater than 20% but less than 90%,
the unit can burn up to its historical maximum share of bituminous coal at no cost adder. Burning
beyond its historical maximum share of bituminous coal, the unit incurs a cost adder calculated by
the following equation:
Fuel Switching Cost Adder (2016$ per kW) =
((100 — Historical Maximun Share of Bituminous))
54 x [ — j
4.2.8 Life Extension Costs for Existing Units
The modeling time horizon in EPA Platform v6 extends to 2050 and covers a period of almost 30 years.
This time horizon requires consideration of investments, beyond routine maintenance, necessary to
extend the life of existing units. The life extension costs for different unit types are summarized in Table
4-10 below. Each unit has the option to retire or incorporate the life extension costs. These costs were
based on a review of 2007-2016 FERC Form 1 data maintained by SNL regarding reported annual capital
expenditures made by older units. The life extension costs were added once the unit reaches its
assumed lifespan. However, if the unit reaches its lifespan before the first run year, then the life
extension cost was applied when the unit reaches twice its lifespan age. The assumption implies if the
unit has reached its lifespan before the first run year, it has already incurred the necessary life extension
4-17
-------
related investment costs and is considered sunk. Life extension costs for nuclear units are discussed in
Section 4.5.1.
Table 4-10 Life Extension Cost Assumptions Used in EPA Platform v6
Lifespan without Life
Life Extension
Capital Cost of
Life Extension Cost
Plant Type
Extension
Cost
New Unit
as Proportion of New
Expenditures
(2016$/kW)
(2016$/kW)
Unit Capital Cost (%)
Biomass
40
291
4,429
6.6
Coal Steam
40
212
3,639
5.84
Combined Cycle
30
89
978
9.06
Combustion Turbine
30
246
678
36.3
IC Engine
30
177
1,342
13.2
Oil/Gas Steam
40
182
3,311
5.5
IGCC
40
241
3,254
7.4
Landfill Gas
20
823
9,023
9.1
Notes:
Life extension expenditures double the lifespan of the unit.
4.3 Planned-Committed Units
EPA Platform v6 includes all planned-committed units that are likely to come online because ground has
been broken, financing obtained, or other demonstrable factors indicate a high probability that the unit will
be built before June 30, 2021.
4.3.1 Population and Model Plant Aggregation
Table 4-11 summarizes the extent of inventory of planned-committed units represented by unit types and
generating capacity.
Table 4-11 Summary of Planned-Committed Units in NEEDS v6 for EPA Platform v6
Plant Type
Capacity (MW)
Year Range Described
Renewables/Non -conventional
Biomass
12
2019 -2019
Energy Storage
22
2018 -2019
Hydro
244
2018-2020
Non-Fossil Waste
44
2018-2020
Onshore Wind
3,483
2018 -2019
Solar PV
431
2018-2020
Subtotal
4,237
Fossil/Conventional
Combined Cycle
18,195
2018-2020
Combustion Turbine
2,302
2018-2021
Nuclear
2,200
2022 - 2023
O/G Steam
23
2018 -2018
Subtotal
22,720
Grand Total
26,957
Table 4-12 gives a breakdown of planned-committed units by IPM region, plant type, and capacity.
4-18
-------
Table 4-12 Planned-Committed Units by Model Region in NEEDS v6 for EPA Platform v6
IPM Region
Plant Type
Capacity (MW)
ERC_PHDL
Onshore Wind
588
Combustion Turbine
1,061
ERC_REST
Non-Fossil Waste
23
Onshore Wind
160
ERC_WEST
Onshore Wind
660
Biomass
12
FRCC
Combined Cycle
1,640
Solar PV
149
MIS_AMSO
Combined Cycle
1,000
MIS_IA
Onshore Wind
66
MIS INKY
Combined Cycle
644
MIS MAPP
Combustion Turbine
218
MIS_MNWI
Combustion Turbine
Onshore Wind
215
40
MIS_WUMS
Combined Cycle
Solar PV
700
2
NENG_CT
Combined Cycle
Combustion Turbine
1,230
90
NENG_ME
O/G Steam
23
NY_Z_C&E
Solar PV
4
NY_Z_G-I
Combined Cycle
Non-Fossil Waste
705
19
PJM_ATSI
Combined Cycle
273
PJM_Dom
Combined Cycle
Combustion Turbine
1,585
300
PJM EMAC
Combined Cycle
1,368
PJM_PENE
Combined Cycle
Combustion Turbine
926
13
PJM_SMAC
Combined Cycle
755
PJM West
Combined Cycle
1,187
PJM WMAC
Combined Cycle
3,472
S_C_TVA
Combined Cycle
1,052
S_SOU
Nuclear
2,200
S_VACA
Combined Cycle
1,072
SPP N
Combustion Turbine
6
SPP SPS
Onshore Wind
800
SPP_WAUE
Onshore Wind
98
SPP_WEST
Combustion Turbine
Onshore Wind
399
200
Combined Cycle
586
WEC_CALN
Non-Fossil Waste
2
Solar PV
200
WECC_CO
Onshore Wind
30
WECC_NM
Onshore Wind
580
WECC_PNW
Hydro
Onshore Wind
244
60
4-19
-------
IPM Region
Plant Type
Capacity (MW)
Solar PV
56
Energy Storage
22
WECC_SCE
Onshore Wind
171
Solar PV
20
WECC_WY
Onshore Wind
30
Note:
Any unit in NEEDS v6 that has an online year of 2018 or later was considered a
Planned/Committed Unit.
4.3.2 Capacity
The capacity data of planned-committed units in NEEDS v6 was obtained from the sources reported in
Table 4-1.
4.3.3 State and Model Region
State location data for the planned-committed units in NEEDS v6 came from the information sources
noted in Section 4.3.1. The state-county information was then used to assign planned-committed units to
their respective model regions.
4.3.4 Online and Retirement Year
As noted above, planned-committed units included in NEEDS v6 are only those likely to come on-line
before June 2021, as 2021 is the first analysis year in the EPA Platform v6. All planned-committed units
were assigned an online year and given a default retirement year of 9999.
4.3.5 Unit Configuration, Cost, and Performance
All planned-committed units in NEEDS v6 assume the cost, performance, and unit configuration
characteristics of potential units that are available in 2021. A detailed description of potential unit
assumptions is provided below in Section 4.4.
4.4 Potential Units
The EPA Platform v6 includes options for developing a variety of potential units that may be built at a
future date in response to electricity demand and the constraints represented in the model. Defined by
region, technology, and the year available, potential units with an initial capacity of zero MW are inputs
into IPM. When the model is run, the capacity of certain potential units is raised from zero to meet
demand and other system and operating constraints. This results in the model's projection of new
capacity.
In Table 4-7, the block labeled "New Units" provides the type and number of potential units available in
EPA Platform v6. The following sections describe the cost and performance assumptions for the potential
units represented in the EPA Platform v6.
4-20
-------
4.4.1 Methodology Used to Derive the Cost and Performance Characteristics of Conventional
Potential Units
The cost and performance characteristics of conventional potential units in EPA Platform v6 are derived
primarily from assumptions used in the Annual Energy Outlook (AEO) 2017 published by the U.S.
Department of Energy's Energy Information Administration.
4.4.2 Cost and Performance for Potential Conventional Units
Table 4-13 shows the cost and performance assumptions for potential conventional units. The cost and
performance assumptions are based on the size (i.e., net electrical generating capacity in MW) indicated
in the table. However, the total new capacity that is added in each model run for these technologies is
not restricted to these capacity levels.
The table includes several components of cost. The total installed cost of developing and building a new
plant is captured through capital cost. It includes expenditures on pollution control equipment that new
units are assumed to install to satisfy air regulatory requirements. The capital costs shown are typically
referred to as overnight capital costs. They include engineering, procurement, construction, startup, and
owner's costs (for such items as land, cooling infrastructure, administration and associated buildings, site
works, switchyards, project management, and licenses). The capital costs of new non-wind and non-
solar units are increased to account for the cost of maintaining and expanding the transmission network.
This cost based on AEO 2017 is equal to 97 $/kW outside of WECC and NY regions and 145 $/kW within
these regions. The capital costs do not include interest during construction (IDC). IDC is added to the
capital costs during the set-up of an IPM run. Calculation of IDC is based on the construction profile of
the build option and the discount rate. Details on the discount rates used in the EPA Platform v6 are
provided in Chapter 10 of this documentation.
Table 4-13 also shows fixed operating and maintenance (FOM) and variable operating and maintenance
(VOM) components of cost. FOM is the annual cost of maintaining a generating unit. It represents
expenses incurred regardless of the extent that the unit is run. It is expressed in units of $ per kW per
year. VOM represents the non-fuel variable costs incurred in running an electric generating unit. It is
proportional to the electrical energy produced and is expressed in units of $ per MWh.
In addition to the three components of cost, Table 4-13 indicates the first run year available, lead time,
vintage periods, heat rate, and availability for each type of unit. Lead time represents the construction
time needed for a unit to come online. Vintage periods are used to capture the cost and performance
improvements resulting from technological advancement and learning-by-doing. Mature technologies and
technologies whose first year available is not at the start of the modeling time horizon may have only one
vintage period, whereas newer technologies may have several vintage periods. Heat rate indicates the
efficiency of the unit and is expressed in units of energy consumed (Btus) per unit of electricity generated
(kWh). Availability indicates the percentage of time that a generating unit is available to provide electricity
to the grid once it is online. Availability takes into account estimates of the time consumed by planned
maintenance and forced outages. The emission characteristics of the potential units can be found in
Table 3-17.
4.4.3 Short-Term Capital Cost Adder
In addition to the capital costs shown in Table 4-13 and Table 4-16, EPA Platform v6 includes a short-
term capital cost adder that kicks in if the new capacity deployed in a specific model run year exceeds
certain upper bounds. This adder is meant to reflect the added cost incurred due to short-term
competition for scarce labor and materials. Table 4-14 shows the cost adders for each type of potential
unit for model run years through 2035. The adder is not imposed after 2035, assuming markets for labor
and materials have sufficient time to respond to changes in demand.
The column labeled "Step 1" in Table 4-14 indicates the total amount of capacity of a particular plant type
that can be built in a given model run year without incurring a cost adder. However, if the Step 1 upper
4-21
-------
bound is exceeded, then either the Step 2 or Step 3 cost adder is incurred by the entire amount of
capacity deployed, where the level of the cost adder depends upon the total amount of new capacity
added in that run year. For example, the Step 1 upper bound in 2021 for landfill gas potential units is 625
MW. If no more than this total new landfill gas capacity is built in 2021, only the capital cost shown in
Table 4-16 is incurred. If the model builds between 625 and 1,088 MW, the Step 2 cost adder of
$3,979/kW applies to the entire capacity deployed. If the total new landfill gas capacity exceeds the Step
2 upper bound of 1,088 MW, then the Step 3 capacity adder of $12,639/kW is incurred by the entire
capacity deployed in that run year. The short-term capital cost adders shown in Table 4-14 were derived
from AEO assumptions.
4.4.4 Regional Cost Adjustment
The capital costs reported in Table 4-13 are generic. Before implemented, the capital cost values are
converted to region-specific costs by applying regional cost adjustment factors that capture regional
differences in labor, material, and construction costs and ambient conditions. These factors are
calculated by multiplying the regional cost and ambient condition multipliers. The regional cost multipliers
are based on county level estimates developed by the Energy Institute at University of Texas at Austin36.
The ambient condition multipliers are from AEO 2017. Table 4-15 summarizes the regional cost
adjustment factors at the IPM region and technology level. The factors are applied to both conventional
technologies shown in Table 4-13 and renewable and nonconventional technologies shown in Table 4-16.
However, they are not applied to hydro and geothermal technologies as site-specific costs are used for
these two technologies.
36 New U.S. Power Costs: by County, with Environmental Externalities, University of Texas at Austin, Energy Institute.
July 2016
4-22
-------
Table 4-13 Performance and Unit Cost Assumptions for Potential (New) Capacity from Conventional Technologies in EPA Platform v6
Advanced
Combined
Cycle
Advanced
Combined Cycle
with CCS
Advanced
Combustion
Turbine
Nuclear
Ultrasupercritical
Coal with 30% CCS
Ultrasupercritical
Coal with 90% CCS
Ultrasupercritical
Coal without CCS
Size (MW)
429
429
237
2234
650
650
650
First Year Available
2021
2021
2021
2023
2021
2021
2021
Lead Time (Years)
3
3
2
6
4
4
4
Availability
87%
87%
93%
90%
85%
85%
85%
Vintage #1 (2021)
Heat Rate (Btu/kWh)
6,267
7,514
9,264
10,459
9,644
11,171
8,704
Capital (2016$/kW)
1,081
2,104
662
5,644
4,953
5,477
3,580
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #2 (2023)
Heat Rate (Btu/kWh)
6,233
7,504
8,907
10,459
9,433
10,214
8,514
Capital (2016$/kW)
1,064
2,059
651
5,300
4,863
5,378
3,516
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #3 (2025)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
Capital (2016$/kW)
1,041
2,003
636
5,164
4,746
5,249
3,431
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #4 (2030)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
Capital (2016$/kW)
963
1,833
580
4,804
4,434
4,904
3,205
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #5 (2035)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
Capital (2016$/kW)
902
1,698
536
4,527
4,198
4,642
3,035
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #6 (2040)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
4-23
-------
Advanced
Combined
Cycle
Advanced
Combined Cycle
with CCS
Advanced
Combustion
Turbine
Nuclear
Ultrasupercritical
Coal with 30% CCS
Ultrasupercritical
Coal with 90% CCS
Ultrasupercritical
Coal without CCS
Capital (2016$/kW)
857
1,589
505
4,283
3,991
4,413
2,885
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #7 (2045)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
Capital (2016$/kW)
816
1,487
477
4,049
3,792
4,193
2,741
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Vintage #8 (2050)
Heat Rate (Btu/kWh)
6,200
7,493
8,550
10,459
9,221
9,257
8,323
Capital (2016$/kW)
778
1,390
454
3,810
3,585
3,965
2,592
Fixed O&M (2016$/kW/yr)
9.9
33.2
6.8
99.7
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2.0
7.1
10.6
2.3
7.1
9.5
4.6
Notes:
a Capital cost represents overnight capital cost.
4-24
-------
Table 4-14 Short-Term Capital Cost Adders for New Power Plants in EPA Platform v6 (2016$)
Plant Type
2021
2023
2025
2030
2035
Step 1
Step 2
Step 3
Step 1
Step 2
Step 3
Step 1
Step 2
Step 3
Step 1
Step 2
Step 3
Step 1
Step 2
Step 3
Biomass
Upper Bound (MW)
1,904
3,312
No limit
1,270
2,208
No limit
1,270
2,208
No limit
3,174
5,520
No limit
3,174
5,520
No limit
Adder ($/kW)
-
1,714
5,443
-
1,685
5,352
-
1,646
5,230
-
1,543
4,903
-
1,466
4,658
Coal Steam - UPC
Upper Bound (MW)
18,361
31,932
No limit
12,241
21,288
No limit
12,241
21,288
No limit
30,602
53,220
No limit
30,602
53,220
No limit
Adder ($/kW)
-
1,640
5,209
-
1,610
5,115
-
1,572
4,992
-
1,468
4,664
-
1,390
4,415
Coal Steam - UPC30
Upper Bound (MW)
18,361
31,932
No limit
12,241
21,288
No limit
12,241
21,288
No limit
30,602
53,220
No limit
30,602
53,220
No limit
Adder ($/kW)
-
2,269
7,206
-
2,228
7,076
-
2,174
6,906
-
2,031
6,452
-
1,923
6,108
Coal Steam - UPC90
Upper Bound (MW)
18,361
31,932
No limit
12,241
21,288
No limit
12,241
21,288
No limit
30,602
53,220
No limit
30,602
53,220
No limit
Adder ($/kW)
-
2,509
7,969
-
2,463
7,825
-
2,404
7,636
-
2,246
7,134
-
2,126
6,754
Combined Cycle
Upper Bound (MW)
132,125
229,782
No limit
88,083
153,188
No limit
88,083
153,188
No limit
220,208
382,970
No limit
220,208
382,970
No limit
Adder ($/kW)
-
490
1,555
-
481
1,528
-
469
1,491
-
433
1,376
-
406
1,290
Combustion Turbine
Upper Bound (MW)
66,275
115,260
No limit
44,183
76,840
No limit
44,183
76,840
No limit
110,458
192,100
No limit
110,458
192,100
No limit
Adder ($/kW)
-
298
945
-
291
924
-
281
893
-
255
809
-
235
747
Fuel Cell
Upper Bound (MW)
1,725
3,000
No limit
1,150
2,000
No limit
1,150
2,000
No limit
2,875
5,000
No limit
2,875
5,000
No limit
Adder ($/kW)
-
3,101
9,850
-
3,007
9,551
-
2,896
9,200
-
2,615
8,305
-
2,386
7,578
Geothermal
Upper Bound (MW)
883
1,536
No limit
589
1,024
No limit
589
1,024
No limit
1,472
2,560
No limit
1,472
2,560
No limit
Adder ($/kW)
-
3,772
11,983
-
3,763
11,954
-
3,744
11,892
-
3,700
11,754
-
3,636
11,549
Landfill Gas
Upper Bound (MW)
625
1,088
No limit
417
725
No limit
417
725
No limit
1,042
1,813
No limit
1,042
1,813
No limit
Adder ($/kW)
-
3,979
12,639
-
3,915
12,437
-
3,822
12,140
-
3,577
11,361
-
3,379
10,733
Nuclear
Upper Bound (MW)
32,327
56,220
No limit
21,551
37,480
No limit
21,551
37,480
No limit
53,878
93,700
No limit
53,878
93,700
No limit
Adder ($/kW)
-
2,499
7,939
-
2,347
7,456
-
2,287
7,264
-
2,127
6,757
-
2,005
6,368
Solar Thermal
Upper Bound (MW)
2,830
4,921
No limit
1,886
3,281
No limit
1,886
3,281
No limit
4,716
8,202
No limit
4,716
8,202
No limit
Adder ($/kW)
-
2,327
7,390
-
2,736
8,691
-
2,640
8,385
-
2,430
7,719
-
2,286
7,262
Solar PV
Upper Bound (MW)
25,858
46,265
No limit
18,406
32,011
No limit
18,406
32,011
No limit
46,016
80,027
No limit
46,016
80,027
No limit
Adder ($/kW)
-
366
1,163
-
398
1,263
-
384
1,218
-
359
1,141
-
339
1,077
Onshore Wind
Upper Bound (MW)
33,941
67,466
No limit
30,238
52,588
No limit
30,238
52,588
No limit
75,595
131,470
No limit
75,595
131,470
No limit
Adder ($/kW)
-
716
2,275
-
693
2,200
-
667
2,120
-
602
1,911
-
575
1,827
Offshore Wind
Upper Bound (MW)
1,725
3,000
No limit
1,150
2,000
No limit
1,150
2,000
No limit
2,875
5,000
No limit
2,875
5,000
No limit
Adder ($/kW)
-
2,143
6,808
-
1,933
6,139
-
1,893
6,012
-
1,798
5,712
-
1,752
5,565
Hydro
Upper Bound (MW)
10,360
18,018
No limit
6,907
12,012
No limit
6,907
12,012
No limit
17,267
30,030
No limit
17,267
30,030
No limit
Adder ($/kW)
-
1,043
3,313
-
1,043
3,313
-
1,043
3,313
-
1,043
3,313
-
1,043
3,313
4-25
-------
Table 4-15 Regional Cost Adjustment Factors for Conventional and Renewable Generating Technologies in EPA Platform v6
Model
Region
ERC_PHDL
ERC_REST
ERC_WEST
FRCC
MIS_AMSO
MIS_AR
MIS_D_MS
MISJA
MISJL
MISJNKY
MIS_LA
MIS_LMI
MIS_MAPP
MIS_MIDA
MIS_MNWI
MIS_MO
MIS_WOTA
MIS_WUMS
NENG_CT
NENG_ME
NENGREST
NY_Z_A
NY_Z_B
NY_Z_C&E
NY_Z_D
NY_Z_F
NY Z G-l
Combined
Cycle
1.006
0.977
0.999
0.983
0.955
0.977
0.958
1.001
1
0.987
0.958
1.009
0.97
0.996
1.006
0.995
0.956
1.028
1.181
1.064
1.115
1.061
1.076
1.11
1.076
1.129
1.195
Combined
Cycle
with
Carbon
Capture
1.006
0.977
0.999
0.983
0.955
0.977
0.958
1.001
1
0.987
0.958
1.009
0.97
0.996
1.006
0.995
0.956
1.028
1.181
1.064
1.115
1.061
1.076
1.11
1.076
1.129
1.195
Combustion
Turbine
1.042
1.027
1.038
1.033
1.015
1.022
1.013
1.017
1.016
1.007
1.013
1.015
1.003
1.015
1.02
1.015
1.01
1.032
1.146
1.074
1.105
1.072
1.081
1.111
1.092
1.122
1.161
Nuclear Biomass
0.979
0.969
0.976
0.976
0.963
0.977
0.968
0.999
0.999
0.998
0.967
1.016
0.986
0.997
1
0.995
0.966
1.013
1.068
1.042
1.053
1.039
1.043
1.056
1.045
1.055
1.068
0.922
0.922
0.922
0.948
0.93
0.93
0.93
0.968
1.017
1.01
0.93
0.995
0.968
0.968
0.968
1.017
0.93
1.01
1.03
1.03
1.03
1.034
1.034
1.034
1.034
1.034
1.034
Landfill
Gas
0.92
0.92
0.92
0.949
0.933
0.933
0.933
0.968
1.019
0.994
0.933
0.997
0.968
0.968
0.968
1.019
0.933
0.994
1.009
1.009
1.009
0.999
0.999
0.999
0.999
0.999
0.999
Offshore
Wind
1.002
0.968
0.989
0.961
0.949
0.977
0.958
1.041
1.014
1.003
0.957
1.024
1.035
1.04
1.05
1.016
0.956
1.045
1.081
1.065
1.068
1.021
1.027
1.038
1.043
1.06
1.079
Onshore
Wind
1.002
0.968
0.989
0.961
0.949
0.977
0.958
1.041
1.014
1.003
0.957
1.024
1.035
1.04
1.05
1.016
0.956
1.045
1.081
1.065
1.068
1.021
1.027
1.038
1.043
1.06
1.079
Solar Solar
PV Thermal
0.96
0.94
0.95
0.94
0.92
0.95
0.93
1.01
1
0.99
0.93
1.01
0.99
1.01
1.02
1
0.92
1.03
1.08
1.02
1.04
1
1
1.02
1.01
1.04
1.09
0.916
0.889
0.909
0.899
0.865
0.914
0.884
0.993
0.99
0.972
0.879
1.002
0.945
0.984
1.008
0.981
0.875
1.029
1.103
0.993
1.034
0.988
0.992
1.005
0.986
1.04
1.13
Fuel
Cell
0.9
0.9
0.9
1
0.9
0.9
0.9
1
1
1
0.9
1
1
1
1
1
0.9
1
1
1
1
1
1
1
1
1
1
Ultra
supercritical
Coal
without CCS
1.005
0.981
0.997
1.001
0.958
0.995
0.972
1.013
1.021
1.009
0.968
1.025
0.976
1.007
1.015
1.013
0.964
1.046
1.112
1.048
1.075
1.05
1.058
1.08
1.056
1.085
1.119
Ultra
supercritical
Coal with
30% CCS
1.005
0.981
0.997
1.001
0.958
0.995
0.972
1.013
1.021
1.009
0.968
1.025
0.976
1.007
1.015
1.013
0.964
1.046
1.112
1.048
1.075
1.05
1.058
1.08
1.056
1.085
1.119
Ultra
supercritical
Coal with
90% CCS
0.992
0.969
0.985
0.991
0.947
0.987
0.962
1.008
1.02
1.008
0.956
1.022
0.967
1
1.01
1.009
0.952
1.044
1.116
1.047
1.075
1.046
1.054
1.078
1.053
1.085
1.122
4-26
-------
Combined
Ultra
Ultra
Ultra
Model
Combined
Cycle
with
Combustion
MnMoar
Riomacc
Landfill
Offshore
Onshore
Solar
Solar Fuel supercritical
supercritical
supercritical
Region
Cycle
Willi
Turbine
IMUlflCCII
DIUI Id99
Gas
Wind
Wind
PV
Thermal Cell Coal
Coal with
Coal with
i^cirpon
Capture
without CCS
30% CCS
90% CCS
NY_Z_J
1.257
1.257
1.205
1.074
1.227
1.26
1.093
1.093
1.12
1.216 1.2 1.157
1.157
1.162
NY_Z_K
1.241
1.241
1.196
1.073
1.227
1.26
1.092
1.092
1.1
1.163 1.2 1.153
1.153
1.158
PJM_AP
1.073
1.073
1.088
1.034
1.01
0.994
1.008
1.008
0.98
0.961 1
1.072
1.072
1.069
PJM_ATSI
1.031
1.031
1.046
1.018
1.01
0.994
1.007
1.007
0.99
0.974 1
1.043
1.043
1.039
PJM_COMD
1.022
1.022
1.026
1.009
1.01
0.994
1.04
1.04
1.03
1.042 1
1.039
1.039
1.039
PJM_Dom
1.144
1.144
1.153
1.046
0.913
0.911
1.018
1.018
0.99
0.964 0.9 1.13
1.13
1.127
PJM_EMAC
1.209
1.209
1.179
1.073
1.065
1.033
1.066
1.066
1.06
1.09 1
1.144
1.144
1.148
PJM_PENE
1.097
1.097
1.105
1.047
1.065
1.033
1.024
1.024
1
0.988 1
1.083
1.083
1.081
PJM_SMAC
1.155
1.155
1.144
1.063
1.065
1.033
1.036
1.036
1.01
0.99 1
1.118
1.118
1.118
PJM_West
0.991
0.991
1.019
1.004
1.01
0.994
0.989
0.989
0.97
0.939 1
1.012
1.012
1.008
PJM_WMAC
1.151
1.151
1.144
1.06
1.065
1.033
1.043
1.043
1.02
1.018 1
1.113
1.113
1.113
S_C_KY
0.981
0.981
1.015
0.99
0.934
0.933
0.979
0.979
0.95
0.919 0.9 1.006
1.006
1.004
S_C_TVA
0.957
0.957
1.003
0.979
0.934
0.933
0.968
0.968
0.94
0.899 0.9 0.981
0.981
0.975
S_D_AECI
0.989
0.989
1.014
0.992
1.017
1.019
1.013
1.013
0.99
0.971 1
1.005
1.005
0.999
S_SOU
0.963
0.963
1.02
0.969
0.925
0.925
0.953
0.953
0.92
0.873 0.9 0.982
0.982
0.972
S_VACA
1.015
1.015
1.059
1.003
0.913
0.911
0.975
0.975
0.94
0.896 0.9 1.033
1.033
1.025
SPP_N
1
1
1.032
0.986
0.973
0.975
1.016
1.016
0.98
0.948 1
1.009
1.009
0.998
SPP_NEBR
0.976
0.976
1.009
0.988
0.968
0.968
1.029
1.029
0.98
0.945 1
0.982
0.982
0.971
SPP_SPS
0.992
0.992
1.028
0.98
0.956
0.952
1.005
1.005
0.96
0.92 1
0.991
0.991
0.979
SPP_WAUE
0.974
0.974
1.006
0.987
0.968
0.968
1.034
1.034
0.99
0.947 1
0.979
0.979
0.97
SPP_WEST
0.978
0.978
1.02
0.978
0.956
0.952
0.991
0.991
0.96
0.918 1
0.989
0.989
0.978
WEC_BANC
1.232
1.232
1.173
1.072
1.076
1.055
1.124
1.124
1.1
1.112 1
1.208
1.208
1.203
WEC_CALN
1.23
1.23
1.172
1.071
1.076
1.055
1.123
1.123
1.1
1.109 1
1.207
1.207
1.201
WEC_LADW
1.183
1.183
1.141
1.055
1.076
1.055
1.104
1.104
1.07
1.076 1
1.167
1.167
1.151
WEC_SDGE
1.154
1.154
1.12
1.046
1.076
1.055
1.084
1.084
1.05
1.049 1
1.141
1.141
1.123
WECC_AZ
1.187
1.187
1.19
1.011
1
0.982
1.035
1.035
1
0.97 1
1.181
1.181
1.166
WECC_CO
1.157
1.157
1.194
0.988
0.936
0.947
1.027
1.027
0.98
0.932 1
1.156
1.156
1.142
WECCJD
1.045
1.045
1.07
1.004
1.002
0.982
1.048
1.048
1
0.965 1
1.066
1.066
1.058
WECC IID
1.262
1.262
1.236
1.036
1
0.982
1.069
1.069
1.04
1.028 1
1.252
1.252
1.233
4-27
-------
Model
Region
Combined
Cycle
Combined
Cycle
with
Carbon
Capture
Combustion
Turbine
Nuclear
Biomass
Landfill
Gas
Offshore
Wind
Onshore
Wind
Solar
PV
Solar
Thermal
Fuel
Cell
Ultra
supercritical
Coal
without CCS
Ultra
supercritical
Coal with
30% CCS
Ultra
supercritical
Coal with
90% CCS
WECC_MT
1.021
1.021
1.054
0.992
1.002
0.982
1.039
1.039
0.99
0.953
1
1.037
1.037
1.03
WECC_NM
1.131
1.131
1.161
0.99
1
0.982
1.018
1.018
0.98
0.938
1
1.129
1.129
1.115
WECC_NNV
1.157
1.157
1.137
1.04
1.002
0.982
1.087
1.087
1.05
1.045
1
1.157
1.157
1.147
WECC_PNW
1.123
1.123
1.109
1.035
1.002
0.982
1.074
1.074
1.04
1.032
1
1.145
1.145
1.144
WECC_SCE
1.18
1.18
1.139
1.054
1.076
1.055
1.1
1.1
1.07
1.071
1
1.163
1.163
1.144
WECC_SNV
1.23
1.23
1.22
1.03
1
0.982
1.071
1.071
1.04
1.042
1
1.237
1.237
1.219
WECC_UT
1.05
1.05
1.075
1.002
1.002
0.982
1.043
1.043
1
0.962
1
1.063
1.063
1.051
WECC_WY
1.016
1.016
1.055
0.987
1.002
0.982
1.031
1.031
0.98
0.927
1
1.024
1.024
1.012
Table 4-16 Performance and Unit Cost Assumptions for Potential (New) Renewable and Non-Conventional Technology Capacity in EPA
Platform v6
Biomass-
Bubbling
Fluidized Bed
(BFB)
Landfill Gas
Solar
Photovoltaic
Solar
Thermal
Onshore
Wind
Offshore
Wind
Geothermal
LGHI
LGLo
LGVLo
Fuel Cells
Size (MW)
50
50
50
10
100
100
100
600
First Year Available
2021
2021
2021
2021
2021
2021
2021
2021
Lead Time (Years)
4
4
3
3
1
3
3
3
Availability
83%
90% - 95%
90%
87%
90%
90%
95%
95%
Generation Capability
Economic
Dispatch
Economic
Dispatch
Economic Dispatch
Economic
Dispatch
Generation
Profile
Economic
Dispatch
Generation
Profile
Generation
Profile
Vintage #1 (2021-2054)
Vintage #1 (2021
Heat Rate (Btu/kWh)
13,500
30,000
18,000
18,000
18,000
8,653
0
0
0
0
Capital (2016$/kW)
3,733
3,072-21,106
8,556
10,780
16,598
6,889
1034
6,717
1,404
4,529
Fixed O&M (2016$/kW/yr)
110.34
105-542
410.32
410.32
410.32
0.00
11.35
62.69
49.46
116.64
Variable O&M (2016$/MWh)
5.49
0.00
9.14
9.14
9.14
44.9
0
3.5
0
0
Vintage #2 (2023
Heat Rate (Btu/kWh)
7,807
0
0
0
0
Capital (2016$/kW)
6680
1009
6,555
1,372
4,169
Fixed O&M (2016$/kW/yr)
0.0
10.74
59.9
48.72
111.15
Variable O&M (2016$/MWh)
44.9
0
3.5
0
0
Vintage #3 (2025
Heat Rate (Btu/kWh) | | III
6,960
0
0
0
0
4-28
-------
Biomass-
Bubbling
Fluidized Bed
(BFB)
Landfill Gas
Solar
Photovoltaic
Solar
Thermal
Onshore
Wind
Offshore
Wind
Geothermal
LGHI
LGLo
LGVLo
Fuel Cells
Capital (2016$/kW)
6434
984
6,396
1,337
4,122
Fixed O&M (2016$/kW/yr)
0.0
10.13
57.12
47.98
109.58
Variable O&M (2016$/MWh)
44.9
0
3.5
0
0
Vintage #4 (2030
Heat Rate (Btu/kWh)
0
0
0
0
0
Capital (2016$/kW)
921
6,047
1,242
4,006
921
Fixed O&M (2016$/kW/yr)
10.13
50.15
46.13
105.66
10.13
Variable O&M (2016$/MWh)
0
3.5
0
0
0
Vintage #5 (2035
Heat Rate (Btu/kWh)
0
0
0
0
0
Capital (2016$/kW)
870
5,762
1,234
3,952
870
Fixed O&M (2016$/kW/yr)
10.13
50.15
44.29
104.98
10.13
Variable O&M (2016$/MWh)
0
3.5
0
0
0
Vintage #6 (2040
Heat Rate (Btu/kWh)
0
0
0
0
0
Capital (2016$/kW)
819
5,527
1,218
3,898
819
Fixed O&M (2016$/kW/yr)
10.13
50.15
42.44
104.29
10.13
Variable O&M (2016$/MWh)
0
3.5
0
0
0
Vintage #7 (2045
Heat Rate (Btu/kWh)
0
0
0
0
0
Capital (2016$/kW)
772
5,354
1,195
3,837
772
Fixed O&M (2016$/kW/yr)
10.13
50.15
40.6
103.54
10.13
Variable O&M (2016$/MWh)
0
3.5
0
0
0
Vintage #8 (2050
Heat Rate (Btu/kWh)
0
0
0
0
0
Capital (2016$/kW)
726
5,243
1,165
3,775
726
Fixed O&M (2016$/kW/yr)
10.13
50.15
38.75
102.8
10.13
Variable O&M (2016$/MWh)
0
3.5
0
0
0
4-29
-------
4.4.5 Cost and Performance for Potential Renewable Generating and Non-Conventional
Technologies
Table 4-16 summarizes the cost and performance assumptions in EPA Platform v6 for potential
renewable and non-conventional technology generating units. The parameters shown in the table are
based on AEO 2017 for biomass, landfill gas, and fuel cell. For onshore wind, solar PV, and solar
thermal technologies, the parameters shown are based on the National Renewable Energy Laboratory's
(NREL's) 2017 Annual Technology Baseline (ATB) mid-case. For offshore wind, the parameters shown
are based on the NREL's 2016 ATB mid-case. The size (MW) shown in Table 4-16 represents the
capacity on which unit cost estimates were developed and does not indicate the total potential capacity
that the model can build of a given technology. Due to the distinctive nature of generation from
renewable resources, some of the values shown are averages or ranges that are discussed in further
detail in the following subsections. The short-term capital cost adder in Table 4-14 and the regional cost
adjustment factors in Table 4-15 apply equally to the renewable and non-conventional generation
technologies as to the conventional generation technologies.
Wind Generation
EPA Platform v6 includes onshore wind, offshore-shallow, offshore-mid depth, and offshore-deep wind
generation technologies. The following sections describe key aspects of the representation of wind
generation: wind quality and resource potential, distance to transmission, generation profiles, reserve
margin contribution, and capital cost calculation.
Wind Quality and Resource Potential: The NREL resource base for onshore wind is represented by ten
techno-resource groups (TRG). Based on a review of levelized cost of electricity, EPA Platform v6 only
models the resource categories TRG1-TRG8. The NREL resource base for offshore wind is represented
by shallow (TRG1-TRG4), mid-depth (TRG5-TRG7), and deep (TRG8-TRG10) categories. In EPA
Platform v6, the resource categories TRG1, TRG2, TRG3, TRG5, TRG6, and TRG8 are modeled. Table
4-38, Table 4-17, Table 4-18, and Table 4-19 present the onshore, offshore shallow, offshore mid-depth,
and offshore deep wind resource assumptions.
Table 4-17 Offshore Shallow Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class
in EPA Platform v6
IPM Region
State
TRG
Cost Class
1
2
3
CN_BC
BC
CM CO
143
1,000
991
1,760
CN MB
MB
3
997
997
13,978
CN_NB
NB
CM CO
994
999
862
997
1,389
1
982
1,017
10,824
CN_NF
NF
2
997
985
15,445
3
952
1,014
11,688
1
985
1,007
109,060
CN_NL
NL
2
980
1,017
102,486
3
984
1,006
32,049
1
727
CN_NS
NS
2
985
997
16,158
3
999
960
34,831
CN_ON
ON
CM CO
999
995
370
992
46,890
CN_PE
PE
2
650
4-30
-------
IPM Region
State
TRG
Cost Class
1
2
3
3
986
959
13,816
1
989
970
46,105
CN_PQ
PQ
2
968
996
17,275
3
959
984
53,478
ERC_REST
TX
CM CO
2,990
2,962
2,992
3,035
5,030
13,893
MIS INKY
IN
3
385
MIS_LMI
Ml
CM CO
2,499
2,482
306
2,512
8,878
MIS_MNWI
Ml
3
53
Wl
3
184
MIS_WOTA
LA
3
983
108
TX
3
12
1
302
Ml
2
489
MIS_WUMS
3
1,484
1,502
6,397
Wl
CM CO
743
1,498
1,499
2,031
NENG CT
CT
3
259
1
76
NENG_ME
ME
2
469
412
3
495
498
646
1
2,474
2,459
6,487
MA
CM CO
2,497
2,403
2,409
4,104
NENGREST
NH
3
181
1
0
Rl
CM CO
707
416
NY Z A
NY
3
389
544
1,203
NY Z B
NY
3
492
470
NY Z C&E
NY
3
475
520
293
NY Z J
NY
3
355
NY_Z_K
NY
CM CO
930
998
1,064
980
4,102
1,495
PJM_ATSI
OH
CM CO
189
1,496
1,423
8,263
PJM_COMD
IL
CM CO
973
971
NC
2
2,449
2,510
2,953
PJM_Dom
3
2,374
2,603
8,061
VA
CM CO
1,471
2,462
2,444
DE
3
2,989
879
PJM_EMAC
MD
3
2,897
3,009
NJ
2
2,950
3,042
1,786
4-31
-------
IPM Region
State
TRG
Cost Class
1
2
3
3
2,905
3,028
5,414
VA
CM CO
948
2,944
2,903
7,832
PJM_PENE
PA
CM CO
155
492
447
1,917
PJM West
Ml
3
1,134
S SOU
GA
3
2,892
2,958
3,740
NC
2
2,932
2,022
S_VACA
3
2,929
3,046
34,677
SC
CM CO
1,261
2,956
2,520
31,482
WEC_CALN
CA
CM CO
42
147
CA
CM CO
39
134
WECC_PNW
OR
^ CM CO
46
281
469
WA
3
1,018
WECC_SCE
CA
CM CO
75
151
Table 4-18 Offshore Mid-Depth Regional Potential Wind Capacity (MW) by Wind TRG and Cost
Class in EPA Platform v6
IPM Region
State
TRG
Cost Class
1
2
3
CN_BC
BC
6
987
1,012
2,526
CN_NB
NB
6
995
1,000
3,159
CN_NF
NF
LO CD
989
991
1,008
994
7,419
2,148
CN_NL
NL
LO CD
996
992
993
997
28,647
6,691
CN_NS
NS
LO CD
994
955
962
1,034
8,245
45,843
CN ON
ON
6
986
998
3,149
CN_PE
PE
LO CD
376
982
1,002
13,613
CN_PQ
PQ
LO CD
975
993
946
1,003
89,535
34,451
ERC REST
TX
6
2,983
2,864
9,713
MIS LMI
Ml
6
2,487
2,511
1,480
Ml
5
619
MIS_WUMS
6
1,169
Wl
6
1,498
987
NENG_ME
ME
LO CD
500
489
111
501
643
4-32
-------
IPM Region
State
TRG
Cost Class
1
2
3
MA
5
2,494
2,149
48,461
6
2,469
2,365
4,730
NENGREST
NH
6
5
Rl
5
2,492
779
6
2,472
131
NY_Z_K
NY
LO CD
962
878
924
1,013
659
20,641
PJM COMD
IL
6
1,357
PJM_Dom
NC
6
2,482
2,443
5,735
VA
6
2,041
DE
6
342
PJM_EMAC
MD
6
623
NJ
6
2,742
3,028
18,787
VA
6
2,972
3,001
1,472
PJM PENE
PA
6
37
S_VACA
NC
6
2,887
2,999
14,190
SC
6
2,825
587
WEC_CALN
CA
LO CD
48
308
CA
6
18
WECC_PNW
OR
LO CD
317
481
WECC SCE
CA
6
49
Table 4-19 Offshore Deep Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class in
EPA Platform v6
IPM Region
State
TRG
Cost Class
1
2
3
CN NF
NF
8
939
976
145,825
CN NL
NL
8
992
991
448,905
CN NS
NS
8
990
1,005
87,606
CN PQ
PQ
8
989
1,010
198,807
MIS WUMS
Ml
8
1,489
1,348
27,571
NENG ME
ME
8
422
560
75,668
NENGREST
MA
8
951
2,091
149,968
Rl
8
2,477
2,437
745
NY Z K
NY
8
725
1,087
20,795
WEC CALN
CA
8
2,480
1,797
WECC_PNW
CA
8
2,113
OR
8
2,973
3,008
118
WECC SCE
CA
8
2,047
4-33
-------
Generation Profiles: Unlike other generation technologies, which dispatch on an economic basis subject
to their availability constraint, wind and solar technologies can dispatch only when the wind blows and the
sun shines. To represent intermittent renewable generating sources such as wind and solar, EPA
Platform v6 uses hourly generation profiles. Each eligible wind and solar photovoltaic plant is provided
with 8760 hourly generation profiles. These profiles are customized for each wind TRG within an IPM
region and state combination.
The generation profile indicates the amount of generation (kWh) per MW of available capacity. The wind
generation profiles were prepared with data from NREL. Table 4-39 shows the generation profiles for
onshore and offshore wind plants in all model region, state, and TRG combinations for vintage 2021.
Improvements in onshore wind and offshore wind capacity factors overtime are modeled through three
vintages (2021, 2030, and 2040) of new wind units.
To obtain the seasonal generation for the units in a particular wind class in a specific region, the installed
capacity is multiplied by the number of hours in the season and the seasonal capacity factor. Capacity
factor is the average "kWh of generation per MW" from the applicable generation profile. The annual
capacity factors for wind generation that are used in EPA Platform v6 were obtained from NREL and are
shown in Table 4-20, Table 4-22, Table 4-24, and Table 4-26.
Reserve Margin Contribution (also referred to as capacity credit): EPA Platform v6 uses reserve margins,
discussed in detail in Section 3.6, to model reliability. Each region has a reserve margin requirement
which is used to determine the total capacity needed to reliably meet peak demand. The ability of a unit
to assist a region in meeting its reliability requirements is modeled through the unit's contribution to
reserve margin. If the unit has 100 percent contribution towards reserve margin, then the entire capacity
of the unit is counted towards meeting the region's reserve margin requirement. However, if any unit has
less than a 100 percent contribution towards reserve margin, then only the designated share of the unit's
capacity counts towards the reserve margin requirement.
All units except those that depend on intermittent resources have 100% contributions toward reserve
margin. This means that wind and solar have limited (less than 100 percent) contributions toward reserve
margins.
Capacity credit assumptions for onshore wind, offshore wind, and solar PV units are estimated as the
function of penetration of solar and wind in the EPA Platform v6. A two-step approach is developed to
estimate the capacity credit at a unit level. In the first step, the method estimates the sequence of solar
and wind units to build in each IPM region. To do so each solar and wind unit in an IPM Region is sorted
from cheapest to most expensive in-terms of cost and potential revenue generation. Unit level capital
costs, FOM costs, capital charge rate, and average energy price in each IPM region are used in this
analysis. In the second step, capacity credit is calculated for each unit in the sequence as the ratio
between the MW of peak reduced and the capacity of the unit. Unit level hourly generation profiles and
regional hourly load curves are used in this analysis. These initial regional capacity credit curves are
scaled at the NEMS region level to approximately result in capacity credits equal to those projected in
AEO 2017 at the same level of penetration. This approach allows the EPA Platform v6 to endogenously
account for the decline of capacity credit for intermittent resources with their rising penetration. Table
4-21, Table 4-23, Table 4-25 and Table 4-27 present the reserve margin contributions apportioned to new
wind plants in the EPA Platform v6.
4-34
-------
1
2
3
4
5
6
7
8
9
1C
TR(
1
2
3
4
5
6
7
8
9
10
TR<
1
2
3
RG
1
2
3
RG
5
6
Table 4-20 Onshore Average Capacity Factor by Wind TRG
Capacity Factor
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
50.16%
52.30%
54.05%
49.05%
51.24%
53.01%
48.23%
50.71%
52.70%
46.92%
49.52%
51.63%
44.71%
47.76%
50.25%
41.12%
44.74%
47.67%
35.74%
39.48%
42.49%
28.93%
32.25%
34.95%
22.71%
26.13%
28.93%
14.32%
16.77%
18.79%
Table 4-21 Onshore Reserve Margin Contribution by Wind TRG
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
0% - 49%
0% - 51 %
0% - 53%
0% - 84%
0% - 87%
0% - 90%
0% - 83%
0% - 87%
0% - 90%
0% - 82%
0% - 87%
0% - 90%
0% - 81%
0% - 86%
0% - 90%
0% - 78%
0% - 85%
0% - 90%
0% - 76%
0% - 84%
0% - 90%
0% - 75%
0% - 83%
0% - 90%
0%-1%
0%-1%
0%-1%
0%-1%
0%-1%
0%-1%
Table 4-22 Offshore Shallow Average Capacity Factor by Wind TRG
Capacity Factor
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
51% 52% 53%
47% 48% 48%
43% 44% 44%
Table 4-23 Offshore Shallow Reserve Margin Contribution by Wind TRG
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
0% - 88% 0% - 89% 0% - 90%
0% - 88% 0% - 89% 0% - 90%
0% - 88% 0% - 89% 0%_- 90%
Table 4-24 Offshore Mid Depth Average Capacity Factor by Wind TRG
Capacity Factor
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
51% 52% 52%
48% 49% 49%
4-35
-------
Table 4-25 Offshore Mid Depth Reserve Margin Contribution by Wind TRG
TRG
Vintage #1 (2021 -2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
5
6
0% - 88% 0% - 89% 0% - 90%
0% - 88% 0% - 89% 0% - 90%
Table 4-26 Offshore Deep Average Capacity Factor by Wind TRG
TRG
Capacity Factor
Vintage #1 (2021-2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
8
53% 54% 55%
Table 4-27 Offshore Deep Reserve Margin Contribution by Wind TRG
TRG
Vintage #1 (2021 -2054) Vintage #2 (2030-2054) Vintage #3 (2040-2054)
8
0% - 87% 0% - 89% 0% - 90%
Capital cost calculation: Capital costs for wind units include spur-line transmission costs. The resources
for wind and solar are highly sensitive to location. These spur-line costs represent the cost of needed
spur lines, and are based on an estimated distance to transmission infrastructure. NREL develops these
supply curves based on a geographic-information-system analysis, which estimates the resource
accessibility costs in terms of supply curves based on the expected cost of linking renewable resource
sites to the high-voltage, long-distance transmission network. For IPM modeling purposes, the NREL
spur line cost curves are aggregated into a piecewise step curve for each resource class within each
model region and state combination. The sizes of the initial steps are based on the model region load,
while the last step holds the residual resource. The TRG level spur line cost curves for each model
region and state combination are aggregated into a six-step cost curve for onshore wind and into a three-
step cost curve for offshore wind. To obtain the capital cost for a particular new wind model plant, the
capital cost adder applicable to the new plant by resource and cost class shown in Table 4-28, Table
4-29, Table 4-30, and Table 4-40 is added to the base capital cost shown in Table 4-16.
The tax credit extensions for new wind units as prescribed in H.R. 2029, the Consolidated Appropriations
Act of 2016, are implemented through reductions in capital costs. As the credits are based on
construction start date, the 2019 production tax credit (40% of initial value) is assigned to the 2021 run-
year builds for wind units.
Table 4-28 Capital Cost Adder (2016$/kW) for New Offshore Shallow Wind Plants in EPA Platform
v6
IPM Region
State
TRG
Cost Class
1
2
3
CN_BC
BC
2
3
1,010
1,004
1,040
1,091
CN MB
MB
3
2,062
2,090
2,183
CN_NB
NB
2
3
746
242
802
597
741
1
783
790
870
CN_NF
NF
2
791
800
958
3
863
919
1,261
1
228
289
839
CN_NL
NL
2
202
236
688
3
148
230
785
4-36
-------
IPM Region
State
TRG
Cost Class
1
2
3
1
867
CN_NS
NS
2
89
187
570
3
51
60
574
CN_ON
ON
2
3
797
172
912
202
1,116
CN_PE
PE
2
3
768
380
394
544
1
715
722
842
CN_PQ
PQ
2
649
671
1,540
3
623
628
1,535
ERC_REST
TX
2
3
17
3
43
8
82
42
MIS INKY
IN
3
3
MIS_LMI
Ml
2
3
54
5
99
15
50
MIS_MNWI
Ml
3
74
Wl
3
94
MIS_WOTA
LA
3
39
82
TX
3
26
1
76
Ml
2
116
MIS_WUMS
3
8
23
69
Wl
2
3
57
4
12
52
NENG CT
CT
3
10
1
63
NENG_ME
ME
2
43
74
3
18
38
73
1
12
45
86
MA
2
3
9
24
30
71
NENGREST
NH
3
10
1
75
Rl
2
3
44
25
NY Z A
NY
3
5
7
12
NY Z B
NY
3
9
34
NY Z C&E
NY
3
22
61
75
NY Z J
NY
3
2
NY_Z_K
NY
2
3
1
2
3
16
34
56
PJM_ATSI
OH
2
3
6
1
2
8
PJM_COMD
IL
2
3
7
5
PJM_Dom
NC
2
15
56
119
4-37
-------
IPM Region
State
TRG
Cost Class
1
2
3
3
2
6
41
VA
2
3
27
22
30
DE
3
6
24
MD
3
10
35
PJM_EMAC
NJ
2
3
4
3
21
9
^ CM
CM CM
VA
2
3
24
17
30
44
PJM_PENE
PA
2
3
9
4
6
12
PJM West
Ml
3
6
S SOU
GA
3
6
13
25
S_VACA
NC
2
3
39
4
99
9
45
SC
2
3
37
2
3
24
WEC_CALN
CA
2
3
63
73
CA
2
3
19
17
WECC_PNW
OR
1
2
3
8
10
15
WA
3
36
WECC_SCE
CA
2
3
96
171
Table 4-29 Capital Cost Adder (2016$/kW) for New Offshore Mid Depth Wind Plants in EPA
Platform v6
IPM Region
State
TRG
Cost Class
1
2
3
CN BC
BC
6
1,037
1,116
1,151
CN NB
NB
6
428
717
841
CN_NF
NF
5
6
766
820
775
1,046
851
1,329
CN_NL
NL
5
6
262
253
430
555
955
863
CN_NS
NS
5
6
585
73
760
88
847
659
CN ON
ON
6
318
365
678
CN_PE
PE
5
6
659
562
577
741
CN_PQ
PQ
5
6
681
646
689
665
872
1,366
ERC REST
TX
6
2
9
50
4-38
-------
IPM Region
State
TRG
Cost Class
1
2
3
MIS LMI
Ml
6
15
44
90
MIS_WUMS
Ml
5
6
77
95
Wl
6
35
106
NENG_ME
ME
5
6
53
32
95
50
87
MA
5
6
6
20
8
56
45
78
NENGREST
NH
6
36
Rl
5
6
62
34
74
65
NY_Z_K
NY
5
6
16
1
43
4
57
20
PJM COMD
IL
6
5
PJM_Dom
NC
6
2
9
57
VA
6
27
DE
6
1
PJM_EMAC
MD
6
12
NJ
6
2
3
14
VA
6
19
25
34
PJM PENE
PA
6
8
S_VACA
NC
6
6
19
49
SC
6
38
38
WEC_CALN
CA
5
6
74
64
CA
6
15
WECC_PNW
OR
5
6
5
11
WECC SCE
CA
6
47
Table 4-30 Capital Cost Adder (2016$/kW) for New Offshore Deep Wind Plants in EPA Platform v6
IPM Region
State
TRG
Cost Class
1
2
3
CN NF
NF
8
759
761
1,158
CN NL
NL
8
321
560
1,131
CN NS
NS
8
566
588
913
CN PQ
PQ
8
695
737
1,071
MIS WUMS
Ml
8
14
31
90
NENG ME
ME
8
2
3
72
NENGREST
MA
8
2
3
42
Rl
8
66
72
75
NY Z K
NY
8
8 8 25
WEC CALN
CA
8
71
95
WECC_PNW
CA
8
12
OR
8
5
12
18
WECC SCE
CA
8
51
4-39
-------
As an illustrative example, Table 4-31 shows the calculations that would be performed to derive the
potential electric generation, reserve margin contribution, and cost of potential (new) onshore capacity in
wind class 3, cost class 1 in the WECC_CO model region in run year 2021.
Table 4-31 Example Calculations of Wind Generation Potential, Reserve Margin Contribution, and
Capital Cost for Onshore Wind in WECC_CO at Wind Class 3, Cost Class 1.
Required Data
Table 4-30
Table 4-31
Table 4-31
Table 4-31
Potential wind capacity (C) =
Winter average generation (Gw) per available MW =
Winter Shoulder average generation ( Gvvs) per available MW =
Summer average generation (Gs) per available MW =
951 MW
558 kWh/MW
569 kWh/MW
477 kWh/MW
Hours in Winter (Hw) season (December- February) =
Hours in Winter Shoulder (Hws) season (Mar, Apr, Oct. Nov.) =
Hours in Summer (Hs)season (May - September) =
2160 hours
2928 hours
3672 hours
Table 4-20b
Reserve Margin Contribution (RM) WECC_CO, Wind Class 3 =
19 percent
Table 4-16
Table 4-32
Table 4-15
Capital Cost (Cap. . ,) in vintage range for year 2050 =
Capital CostAdder(CC40«,cf)for onstore cost class 1 =
Regional Factor (RF)
$1165/kW
$342/kW
1.027
Calculations
Generation Potential = C xC^x Hw + C x Gws x Hws +C x Gsx Hs
= 951 MW X 558 kWh/MW X 2160 hours +
951 MW X 569 kWh/MW X 2928 hours
+
951 MW X 477 kWh/MW X 3672 hours
= 4.395 GWh
Reserve Margin Contribution = RM xC
= 19% 951 MW
= 185 MW
Capital Cost
~ (c«p2050 X ^ ' CCj1ojv c1) X C
($1,165/kW XI. 027+ $342) X 951 MW
= $1,463,507
Solar Generation
EPA Platform v6 includes solar PV and solar thermal generation technologies. The following sections
describe four key aspects of the representation of solar generation: solar resource potential, generation
profiles, reserve margin contribution, and capital cost calculation.
Solar Resource Potential: The resource potential estimates for solar PV and solar thermal technologies
were developed by NREL by model region, state, and resource class. The NREL resource base for solar
PV is represented by eight resource classes. In EPA Platform v6, the higher capacity factor resource
classes of 3-8 are modeled for solar PV. The NREL resource base for solar thermal is represented by
five resource classes. The solar thermal technology has a ten hour thermal energy storage (TES) and is
considered a dispatchable resource for modeling purposes. These are summarized in Table 4-41 and
Table 4-42.
4-40
-------
Generation Profiles: Table 4-43 shows the generation profiles for solar PV plants in all model region,
state, and resource combinations. The capacity factors for solar generation that are used in EPA
Platform v6 were obtained from NREL and are shown in Table 4-46 and Table 4-47.
Reserve margin contribution (also referred to as capacity credit): The reserve margin contribution section
for wind units summarizes the approach followed for calculating the reserve margin contribution for solar
PV units. Table 4-32 presents the reserve margin contributions apportioned to new solar PV units in the
EPA Platform v6. The solar thermal units are assumed to have 10 hourTES and are assigned 100%
reserve margin contribution.
Table 4-32 Solar Photovoltaic Reserve Margin Contribution by Resource Class
Resource Class
2
3
4
5 6
7
8
Reserve Margin
0%-
0%-
0%-
0s
O
0s
O
0%-
Contribution
6%
71%
90%
90% 90%
90%
0% - 36%
Capital Costs: Similar to wind, capital costs for solar units include transmission spur line cost adders. The
resource class level spur line cost curves for each model region and state combination are aggregated
into a six-step cost curve. Table 4-44 and Table 4-45 illustrate the capital cost adder by resource and
cost class for new solar plants.
The solar PV tariffs are incorporated through capital cost adders in 2021 run year. The tariffs are
calculated as an average of the tariffs for 2018-2020 ((30% + 25% + 20%)/3 = 25%). The solar PV
module cost in 2021 is assumed to be 350 2017$/kW based on an analysis performed by NREL.
The tax credit extensions for new solar units as prescribed in H.R. 2029, the Consolidated Appropriations
Act of 2016, are implemented through reductions in capital costs. As the credits are based on
construction start date, the 2020 Investment tax credit (ITC) of 26% is assigned to the 2021 run-year
builds for solar PV units.
Geothermal Generation
Geothermal Resource Potential: Thirteen model regions in EPA Platform v6 have geothermal potential.
The potential resource in each of these regions is shown in Table 4-33 and is based on NREL ATB 2016.
GEO-Hydro Flash37, GEO-Hydro Binary, GEO-NF EGS Flash and GEO-NF EGS Binary are the included
technologies.
37 In dual flash systems, high temperature water (above 400°F) is sprayed into a tank held at a much lower pressure
than the fluid. This causes some of the fluid to "flash," i.e., rapidly vaporize to steam. The steam is used to drive a
turbine, which, in turn, drives a generator. In the binary cycle technology, moderate temperature water (less than
400°F) vaporizes a secondary, working fluid, which drives a turbine and generator. Due to its use of more plentiful,
lower temperature geothermal fluids, these systems tend to be most cost effective and are expected to be the most
prevalent future geothermal technology.
4-41
-------
Table 4-33 Regional Assumptions on Potential Geothermal Electric Capacity
IPM Model Region
Capacity (MW)
WEC_CALN
530
WEC_LADW
93
WECC_AZ
33
WECC_CO
26
WECCJD
277
WECCJID
3,203
WECC_MT
36
WECC_NM
178
WECC_NNV
1,900
WECC_PNW
1,272
WECC_SCE
561
WECCJJT
225
WECC WY
48
Grand Total
8,382
Cost Calculation: EPA Platform v6 does not contain a single capital cost, but multiple geographically
dependent capital costs for geothermal generation. The assumptions for geothermal were developed
using NREL 2016 ATB cost and performance estimates for 152 sites. Both dual flash and binary cycle
technologies were represented. The 152 sites were aggregated into 93 different options based on
geographic location and cost and performance characteristics of geothermal sites in each of the 13
eligible IPM regions where geothermal generation opportunities exist. Table 4-34 shows the potential
geothermal capacity and cost characteristics for applicable model regions.
Table 4-34 Potential Geothermal Capacity and Cost Characteristics by Model Region
IPM Region
Capacity
(MW)
Capital Cost
(2016$)
FO&M
(2016$/kW-yr)
7
13,379
417
10
19,535
518
14
12,347
341
15
20,164
535
WEC_CALN
37
4,247
125
68
4,988
128
70
6,020
138
111
8,767
259
199
6,228
168
WEC_LADW
34
59
9,006
5,976
269
169
WECC_AZ
33
19,005
501
WECC_CO
10
15
19,550
13,945
518
379
8
19,997
531
10
21,106
542
12
17,579
457
WECCJD
13
16,325
439
20
12,987
344
23
18,113
497
26
9,563
267
4-42
-------
IPM Region
Capacity
(MW)
Capital Cost
(2016$)
FO&M
(2016$/kW-yr)
34
8,564
234
46
11,742
331
86
11,455
285
6
7,898
236
23
7,297
224
25
8,885
267
66
6,085
163
WECCJID
79
9,470
278
93
3,202
118
119
4,630
143
203
5,803
145
2,589
4,050
107
9
19,797
525
WECC_MT
11
16,457
443
16
16,068
430
6
17,611
408
WECC_NM
11
34
19,491
6,047
517
169
127
4,341
129
11
9,991
294
12
15,737
375
13
17,289
481
14
20,232
536
16
10,693
337
16
13,199
351
19
16,757
451
30
11,792
342
44
14,311
390
WECC_NNV
50
12,296
344
97
3,072
106
103
6,335
195
131
5,437
167
139
9,420
280
154
7,285
211
171
8,617
259
241
7,978
246
262
4,542
153
377
4,016
146
8
15,294
437
10
9,883
263
10
17,327
488
WECC_PNW
11
12
14,223
14,225
413
388
12
15,323
403
12
12,498
345
13
12,648
366
4-43
-------
IPM Region
Capacity
(MW)
Capital Cost
(2016$)
FO&M
(2016$/kW-yr)
16
11,821
318
18
10,850
316
19
19,668
522
21
16,483
462
28
15,623
413
40
12,170
291
50
8,752
258
51
5,115
161
126
7,132
212
155
3,240
114
202
7,023
203
457
4,146
131
9
16,223
445
25
7,936
260
WECC_SCE
25
45
14,164
6,671
386
156
110
5,628
143
347
3,072
105
2
15,681
400
7
11,025
343
12
6,723
224
WECCJJT
15
16,101
462
16
10,080
320
64
3,072
115
108
7,389
228
WECC_WY
48
13,175
348
Landfill Gas Electricity Generation
Landfill Gas Resource Potential: Estimates of potential electric capacity from landfill gas are based on
the AEO 2014 inventory. EPA Platform v6 represents the "high", "low", and "very low" categories of
potential landfill gas units. The categories refer to the amount and rate of methane production from the
existing landfill site. Table 4-48 summarizes potential electric capacity from landfill gas.
There are several things to note about Table 4-48. The AEO 2014 NEMS region level estimates of the
potential electric capacity from new landfill gas units are disaggregated to IPM regions based on
electricity demand. The limits listed in Table 4-48 apply to the IPM regions indicated in column 1. In EPA
Platform v6, the new landfill gas electric capacity in the corresponding IPM regions shown in column 1
cannot exceed the limits shown in columns 3-5. As noted earlier, the capacity limits for three categories
of potential landfill gas units are distinguished in this table based on the rate of methane production at
three categories of landfill sites: LGHI = high rate of landfill gas production, LGLo = low rate of landfill gas
production, and LGLVo = very low rate of landfill gas production. The values shown in Table 4-48
represent an upper bound on the amount of new landfill capacity that can be added in each of the
indicated model regions and states for each of the three landfill categories. The cost and performance
assumptions for adding new capacity in each of the three landfill categories are presented in Table 4-16.
4-44
-------
Energy Storage
Energy storage is the capture of energy produced at one time for use at a later time. Presently, the most
common energy storage technologies are pumped storage and lithium-ion battery storage. EPA Platform
v6 now includes battery storage by IPM region and state.
Table 4-35 summarizes the key cost and performance assumptions for new battery storage as
implemented in November 2018 Reference Case. These assumptions are based on Annual Energy
Outlook (AEO) 2018 inputs.
Table 4-35 Performance and Unit Cost Assumptions for Potential (New) Battery Storage
Battery Storage
Size (MW)
30
First Year Available
2021
Lead Time (Years)
1
Availability (%)
96.4
Reserve Margin Contribution (%)
100
Generation Capability
Economic Dispatch
Storage System Efficiency (%)
85
Charge Capacity (Hours)
4
Fixed O&M (2016$/kW-Yr)
35
Variable O&M (2016$/MWh)
7.1
Capital Cost without IDC (2016$/kW)
2021
2,048
2023
1,977
2025
1,906
2030
1,740
2035
1,574
2040
1,418
2045
1,271
2050
1,131
Multiple U.S. states have instituted standalone targets and mandates for energy storage procurement.
Table 4-36 summarizes the state-specific energy storage mandates that are included in the November
2018 Reference Case. Under Assembly Bill No. 2514 and Assembly Bill No. 2868, the California Public
Utilities Commission (CPUC) established energy storage targets for the state's three investor-owned
utilities (lOUs) Pacific Gas and Electric Company, Southern California Edison, and San Diego Gas &
Electric. Hence, the California state mandates are modeled at the utility level in the November 2018
Reference Case.
4-45
-------
Table 4-36 Energy Storage Mandates in the November 2018 Reference Case
State/Region
Bill
Mandate Type
Mandate Specifications
Implementation
Status
California
Assembly Bill No.
2514
Target in MW
Energy storage target of 1,325 megawatts for
Pacific Gas and Electric Company, Southern
California Edison, and San Diego Gas & Electric
by 2020, with installations required no later than
the end of 2024.
2020
Assembly Bill No.
2868
Target in MW
500 MW of distributed energy storage systems
(166.66 MW for each of PG&E, SCE, and
SDG&E).
2020
Senate Bill No.
801
Target in MW
SB 801 directs LADWP to work with the city
council of Los Angeles regarding potential
deployment of 100 MW of energy storage
solutions. SCE is procuring 20 MW.
2019
New York
New York State
Energy Storage
Target
Target in MW
1,500 Megawatts by 2025.
2025
New Jersey
Assembly Bill No.
3723
Target in MW
600 megawatts of energy storage by 2021 and
2,000 megawatts of energy storage by 2030.
2021
Oregon
House Bill 2193
Target in MWh
per electric
company
An electric company shall procure one or more
qualifying energy storage systems that have the
capacity to store at least five-megawatt hours of
energy on or before January 1, 2020.
2020
Massachusetts
Chapter 188
Target in MWh
200 Megawatt hour (MWh) energy storage target
for electric distribution companies to procure
viable and cost-effective energy storage systems
to be achieved by January 1, 2020.
2020
4-46
-------
4.5 Nuclear Units
4.5.1 Existing Nuclear Units
Population. Plant Location, and Unit Configuration: To provide maximum granularity in forecasting the
behavior of existing nuclear units, all 96 nuclear units in EPA Platform v6 are represented by separate
model plants. As noted in Table 4-7, the 96 nuclear units include 94 currently operating units plus Vogtle
Units 3 and 4, which are scheduled to come online post 2021. All are listed in Table 4-49. The
population characteristics, plant location, and unit configuration data in NEEDS v6 were obtained
primarily from EIA Form 860 and AEO 2018.
Capacity: Nuclear units are baseload power plants with high fixed (capital and fixed O&M) costs and
relatively low variable (fuel and variable O&M) costs. Due to their low variable costs, nuclear units are run
to the maximum extent possible, i.e., up to their availability. Consequently, a nuclear unit's capacity factor
is equivalent to its availability. Thus, EPA Platform v6 uses capacity factor assumptions to define the
upper bound on generation from nuclear units. Nuclear capacity factor assumptions in EPA Platform v6
are based on an Annual Energy Outlook projection algorithm. The nuclear capacity factor projection
algorithm is described below:
• For each reactor, the capacity factor over time is dependent on the age of the reactor.
• Capacity factors increase initially due to learning, and decrease in the later years due to aging.
• For individual reactors, vintage classifications (older and newer) are used.
• For the older vintage (start before 1982) nuclear power plants, the performance peaks at 25 years:
o Before 25 years: Performance increases by 0.5 percentage point per year;
o 25-80 years: Performance remains flat; and
• For the newer vintage (start in or after 1982) nuclear power plants, the performance peaks at 30
years:
o Before 30 years: Performance increases by 0.7 percentage points per year;
o 30-80 years: Performance remains flat; and
• The maximum capacity factor is assumed to be 90 percent. Hence, a unit is not allowed to grow to a
capacity factor higher than 90 percent. However, if a unit began with a capacity factor above 90
percent, it is allowed to retain that capacity factor. Given historical capacity factors are above 90
percent, the projected capacity factors range from 60 percent to 96 percent.
Cost and Performance: Unlike non-nuclear existing conventional units discussed in section 4.2.7,
emission rates are not needed for nuclear units, since there are no SO2, NOx, CO2, or mercury emissions
from nuclear units.
As with other generating resources, EPA Platform v6 uses heat rate, variable O&M costs (VOM) and fixed
O&M costs (FOM) to characterize the cost of operating existing nuclear units. The data are from AEO
2018 and are shown in Table 4-49.
EPA Platform v6 also uses the nuclear capacity uprates from AEO 2017 and ICF research. These are
shown in Table 4-37.
Table 4-37 Nuclear Uprates (MW) as Incorporated in EPA Platform v6
Name
Plant ID
Unit ID
Year
Change in MWs
Columbia
371
1
2017
19.3
Browns Ferry
46
1
2017
164.7
Browns Ferry
46
2
2017
164.7
Browns Ferry
46
3
2017
164.7
Peach Bottom
3166
2
2017
21.7
Peach Bottom
3166
3
2017
21.7
4-47
-------
EPA Platform v6 imposes lifetime extension costs for nuclear units (See Section 4.2.8) and a maximum
lifetime of 80 years (See Section 3.7).
As nuclear units have aged, some units have been retired from service, or are planning to retire over the
modeled time horizon. For a list of remaining nuclear units, see the NEEDS v6 database. Furthermore,
IPM provides nuclear units with the choice to retire, based on the economics.
Zero Emission Credit (ZEC) Programs: New York and Illinois passed legislation in 2017 to provide
support to selected existing nuclear units that could be at risk of early closure due to declining profitability.
The New York Clean Energy Standard for a 12-year period creates ZECs that are currently applicable for
Fitzpatrick, Ginna, and Nine Mile Point nuclear power plants. The New York load-serving entities (LSEs)
are responsible for purchasing ZECs equal to their share of the statewide load, providing an additional
revenue stream to the nuclear power plants holding the ZECs. Similar to the New York program, the
Illinois Future Energy Jobs Bill creates a ZEC program covering a 10-year term for Clinton and Quad
Cities nuclear power plants.
EPA Platform v6 implicitly models the effect of ZECs by disabling the retirement options for Fitzpatrick,
Ginna, Nine Mile Point, Clinton, and Quad Cities nuclear power plants in the 2021, 2023, and 2025 run
years.
Nuclear Retirement Limits: In EPA Platform v6, endogenous retirements in 2021 of nuclear units are
limited to 8,000 MW38. It is assumed that nuclear units will retire at a pace of 2000 MW per year during
the 2018-2021 period. This 2000 MW per year rate is estimated based on a review of nuclear retirements
in recent years.
Life Extension Costs: Attachment 4-1 summarizes the approach to estimate unit level life extension costs
for existing nuclear units. Nuclear units are assumed to have a maximum lifetime of 80 years (see
Section 3.7). Unlike other plant types, life extension costs for nuclear units are calculated as a function of
age and are applied starting 2021 run year and continue through age 80.
4.5.2 Potential Nuclear Units
The cost and performance assumptions for nuclear potential units that the model has the option to build in
EPA Platform v6 are shown in Table 4-13. The cost assumptions are from AEO 2017.
38 The 8,000 MW limit includes the scheduled retirements of Oyster Creek, Three Mile Island, Pilgrim, and Indian
Point nuclear units 2 and 3.
4-48
-------
List of tables that are uploaded directly to the web:
Table 4-38 Onshore Regional Potential Wind Capacity (MW) by Wind TRG and Cost Class
Table 4-39 Wind Generation Profiles
Table 4-40 Capital Cost Adder (2016$/kW) for New Onshore Wind Plants
Table 4-41 Solar Photovoltaic Regional Potential Capacity (MW) by Resource and Cost Class
Table 4-42 Solar Thermal Regional Potential Capacity (MW) by Resource and Cost Class
Table 4-43 Hourly Solar Generation Profiles
Table 4-44 Capital Cost Adder (2016$/kW) for New Solar PV Plants
Table 4-45 Capital Cost Adder (2016$/kW) for New Solar Thermal Plants
Table 4-46 Solar Photovoltaic Average Capacity Factor by Resource class
Table 4-47 Solar Thermal Capacity Factor by Resource Class and Season
Table 4-48 Potential Electric Capacity from New Landfill Gas Units (MW)
Table 4-49 Characteristics of Existing Nuclear Units
Table 4-50 Generating Units from EIA Form 860 Not Included
Table 4-51 Generating Units Not Included Due to Recent Announcements
Attachment 4-1 Nuclear Power Plant Life Extension Cost Development Methodology
4-49
-------
5. Emission Control Technologies
EPA Platform v6 includes an update of emission control technology assumptions. EPA contracted with
engineering firm Sargent & Lundy to update and add to the retrofit emission control cost models originally
developed for EPA and used in EPA Base Case v.4.10 and updated in EPA Base Case v.5.13. EPA
Platform v6 includes updated assumptions regarding control options for sulfur dioxide (SO2), nitrogen
oxides (NOx), mercury (Hg), carbon dioxide (CO2), and acid gases (HCI). These emission control options
are listed in Table 5-1. They are available in EPA Platform v6 for meeting existing and potential federal,
regional, and state emission limits. Besides the emission control options shown in Table 5-1 and
described in this chapter, EPA Platform v6 offers other compliance options for meeting emission limits.
These include fuel switching, adjustments in the dispatching of electric generating units, and the option to
retire a unit.
Table 5-1 Summary of Emission Control Technology Retrofit Options in EPA Platform v6
SO2 Control
Technology
Options
NOx Control
Technology Options
Mercury Control
Technology Options
CO2 Control
Technology Options
HCI Control
Technology
Options
Limestone Forced
Oxidation (LSFO)
Scrubber
Selective Catalytic
Reduction (SCR)
System
Activated Carbon
Injection (ACI) System
CO2 Capture and
Sequestration
Limestone Forced
Oxidation (LSFO)
Scrubber
Lime Spray Dryer
(LSD) Scrubber
Selective Non-
Catalytic Reduction
(SNCR) System
SO2 and NOx Control
Technology Removal
Cobenefits
Coal-to-gas
Lime Spray Dryer
(LSD) Scrubber
Dry Sorbent
Injection (DSI)
Heat Rate
Improvement
Dry Sorbent
Injection (DSI)
Detailed reports and example calculation worksheets for Sargent & Lundy retrofit emission control models
used by EPA are available in Attachment 5-1 through Attachment 5-7.
5.1 Sulfur Dioxide Control Technologies - Scrubbers
Two commercially available Flue Gas Desulfurization (FGD) scrubber technology options for removing the
SO2 produced by coal-fired power plants are offered in EPA Platform v6: Limestone Forced Oxidation
(LSFO) — a wet FGD technology and Lime Spray Dryer (LSD) — a semi-dry FGD technology which
employs a spray dryer absorber (SDA). In wet FGD systems the polluted gas stream is brought into
contact with a liquid alkaline sorbent (typically limestone) by forcing it through a pool of the liquid slurry or
by spraying it with the liquid. In dry FGD systems the polluted gas stream is brought into contact with the
alkaline sorbent in a semi-dry state through use of a spray dryer. The removal efficiency for SDA drops
steadily for coals whose SO2 content exceeds 3 lbs S02/MMBtu, the technology is therefore provided to
only plants which have the option to burn coals with sulfur content no greater than 3 lbs S02/MMBtu. In
EPA Platform v6 when a unit retrofits with an LSD SO2 scrubber, it loses the option of burning certain high
sulfur content coals (see Table 5-2).
The LSFO and LSD SO2 emission control technologies are available to existing unscrubbed units. They
are also available to existing scrubbed units with reported removal efficiencies of less than 50%. Such
units are considered to have an injection technology and are classified as unscrubbed for modeling
purposes in the NEEDS v6 database. The scrubber retrofit costs for these units are the same as those
for regular unscrubbed units retrofitting with a scrubber.
Default SO2 removal rates for wet and dry FGD were based on data reported in EIA 860 (2015). These
default removal rates were the average of all SO2 removal rates for a dry or wet FGD as reported in EIA
860 (2015) for the FGD installation year.
To reduce the incidence of implausibly high, outlier removal rates, the following adjustment is made.
Units for which reported EIA Form 860 (2015) SO2 removal rates are higher than the average of the upper
5-1
-------
quartile of SO2 removal rates across all scrubbed units are assigned the upper quartile average. The
adjustment is not made, however, if a unit's reported removal rate was recently confirmed by utility
comments. Furthermore, one upper quartile removal rate is calculated across all installation years and
replaces any reported removal rate that exceeds it no matter the installation year.
Existing units not reporting FGD removal rates in EIA Form 860 (2015) will be assigned the default SO2
removal rate for a dry or wet FGD for that installation year.
As shown in Table 5-2, for FGD retrofits installed by the model, the assumed SO2 removal rates will be
98% for wet FGD and 95% for dry FGD.
The procedures used to derive the cost of each scrubber type are discussed in detail in the following
sections.
Table 5-2 Summary of Retrofit SO2 Emission Control Performance
Assumptions in EPA Platform v6
Performance Assumptions
Limestone Forced Oxidation (LSFO)
Lime Spray Dryer (LSD)
Percent Removal*
98%
with a floor of 0.06 Ibs/MMBtu
95%
with a floor of 0.08 Ibs/MMBtu
Capacity Penalty
Calculated based on characteristics of
the unit:
See Table 5-3
Calculated based on characteristics of
the unit:
See Table 5-3
Heat Rate Penalty
Cost (2016$)
Applicability
Units >25 MW
Units >25 MW
Sulfur Content Applicability
Coals < 3 lbs SCh/MMBtu
Applicable Coal Types
BA, BB, BD, BE, BG, BH, SA, SB, SD,
SE, LD, LE, LG, LH, PK, and WC
BA, BB, BD, BE, SA, SB, SD, SE, LD,
and LE
* If the SO2 permit rate of the unit is lower than the floor rate then the SO2 permit rate is considered as the floor rate.
Potential (new) coal-fired units built by IPM are also assumed to be constructed with a wet scrubber
achieving a removal efficiency of 98%. In EPA Platform v6 the costs of potential new coal units include
the cost of scrubbers.
5.1.1 Methodology for Obtaining SO2 Controls Costs
Sargent & Lundy's updated performance/cost models for wet and dry SO2 scrubbers are implemented in
EPA Platform v6 to develop the capital, fixed O&M (FOM), and variable O&M (VOM) components of cost.
For details of Sargent & Lundy Wet FGD and SDA FGD cost models, see Attachment 5-1 and Attachment
5-2.
Capacity and Heat Rate Penalty: In IPM the amount of electrical power required to operate a retrofit
emission control device is represented through a reduction in the amount of electricity that is available for
sale to the grid. For example, if 1.6% of the unit's electrical generation is needed to operate the scrubber,
the generating unit's capacity is reduced by 1.6%. This is the "capacity penalty." At the same time, to
capture the total fuel used in generation both for sale to the grid and for internal load (i.e., for operating
the control device), the unit's heat rate is scaled up such that a comparable reduction (1.6% in the
previous example) in the new higher heat rate yields the original heat rate39. The factor used to scale up
39 Mathematically, the relationship of the heat rate and capacity penalties (both expressed as positive percentage
values) can be represented as follows:
Heat Rate Penalty =
1
Cap acity P enalty
100
-1 xlOO
5-2
-------
the original heat rate is called "heat rate penalty." It is a modeling procedure only and does not represent
an increase in the unit's actual heat rate (i.e., a decrease in the unit's generation efficiency)40. In EPA
Platform v6 specific LSFO and LSD heat rate and capacity penalties are calculated for each installation
based on equations from the Sargent & Lundy models that take into account the rank of coal burned, its
uncontrolled SO2 rate, and the heat rate of the model plant.
Table 5-3 presents the capital, VOM, and FOM costs as well as the capacity and heat rate penalty for two
SO2 emission control technologies (LSFO and LSD) included in EPA Platform v6 for an illustrative set of
generating units with a representative range of capacities and heat rates.
40 The NEEDS heat rate is an unmodified, "original" heat rate to which this retrofit-based heat rate penalty procedure
is applied. This procedure is limited to units at which IPM adds a retrofit in the model.
5-3
-------
Table 5-3 Illustrative Scrubber Costs (2016$) for Representative Sizes and Heat Rates under the Assumptions in EPA Platform v6
Capacity (MW)
Heat Rate
(Btu/kWh)
Capacity
Penalty
(%)
Heat
Variable
100
300
500
700
1000
Scrubber Type
Rate
Penalty
(%)
O&M
(mills/kWh)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
LSFO
Minimum Cutoff: > 25
9,000
-1.60
1.63
2.24
879
24.3
638
11.6
550
8.6
499
8.0
450
6.6
MW
Maximum Cutoff: None
10,000
-1.78
1.82
2.47
919
24.7
667
11.9
575
8.9
522
8.2
470
6.8
Assuming 3 Ib/MMBtu
S02 Content Bituminous
11,000
-1.96
2.00
2.70
956
25.1
694
12.2
599
9.1
543
8.4
490
7.0
Coal
LSD
Minimum Cutoff: > 25
9,000
-1.18
1.20
2.60
745
17.8
546
8.9
472
6.8
424
5.8
424
5.3
MW
Maximum Cutoff: None
10,000
-1.32
1.33
2.89
779
18.2
571
9.2
494
7.0
443
5.9
443
5.5
Assuming 2 Ib/MMBtu
S02 Content Bituminous
11,000
-1.45
1.47
3.17
812
18.5
594
9.4
514
7.3
461
6.1
461
5.7
Coal
Note 1: The above cost estimates assume a boiler burning 3 Ib/MMBtu S02 Content Bituminous Coal for LSFO and 2 Ib/MMBtu S02 Content Bituminous Coal for LSD.
Note 2: The Variable O&M costs in this table do not include the cost of additional auxiliary power (VOMP) component in the Sargent & Lundy tool as for modeling purposes, IPM reflects the auxiliary power
consumption through capacity penalty.
5-4
-------
5.2 Nitrogen Oxides Control Technology
There are two main categories of NOx reduction technologies: combustion and post-combustion controls.
Combustion controls reduce NOx emissions during the combustion process by regulating flame
characteristics such as temperature and fuel-air mixing. Post-combustion controls operate downstream of
the combustion process and remove NOx emissions from the flue gas. All the technologies included in
EPA Platform v6 are commercially available and currently in use in numerous power plants.
5.2.1 Combustion Controls
EPA Platform v6 does not model combustion control upgrades as a retrofit option. The decision was
based on two considerations, the relatively low cost of combustion controls compared with that of post
combustion NOx controls and the possible impact on model size. EPA identified units in NEEDS that
have not employed state-of-the-art combustion controls. EPA then estimated the NOx rates for such units
based on an analysis of historical rates of units with state-of-the-art NOx combustion controls. Emission
rates provided by State-of-the-Art combustion controls are presented in Attachment 3-1.
5.2.2 Post-combustion NOx Controls
EPA Platform v6 provides two post-combustion retrofit NOx control technologies for existing coal units:
Selective Catalytic Reduction (SCR) and Selective Non-Catalytic Reduction (SNCR). In EPA Platform v6,
oil/gas steam units are provided with only SCR retrofits. NOx reduction in a SCR system takes place by
injecting ammonia (Nhb) vapor into the flue gas stream where the NOx is reduced to nitrogen (N2) and
water (H2O) abetted by passing over a catalyst bed typically containing titanium, vanadium oxides,
molybdenum, and/or tungsten. As its name implies, SNCR operates without a catalyst. In SNCR a
nitrogenous reducing agent (reagent), typically urea or ammonia, is injected into, and mixed with, hot flue
gas where it reacts with the NOx in the gas stream reducing it to nitrogen gas and water vapor. Due to
the presence of a catalyst, SCR can achieve greater NOx reductions than SNCR. However, SCR costs
are higher than SNCR costs.
Table 5-4 summarizes the performance and applicability assumptions for each post-combustion NOx
control technology and provides a cross-reference to information on cost assumptions.
Table 5-4 Summary of Retrofit NOx Emission Control Performance Assumptions in EPA Platform
v6
Control Performance
Assumptions
Selective Catalytic Reduction
(SCR)
Selective Non-Catalytic Reduction
(SNCR)
Unit Type
Coal
Oil/Gas
Coal
Percent Removal
90%
80%
Pulverized Coal: 25% (25-200 MW), 20%
(200-400 MW), 15% (>400 MW)
Fluidized Bed: 50%
Rate Floor
Bituminous: 0.07 Ib/MMBtu
Subbituminous and Lignite:
0.05 Ib/MMBtu
-
Pulverized Coal: 0.1 Ib/MMBtu
Fluidized Bed: 0.08 Ib/MMBtu
Size Applicability
Units >25 MW
Units >
25 MW
Units > 25 MW
Costs (2016$)
See Table 5-5
See
Table 5-6
See Table 5-5
5-5
-------
5.2.3 Methodology for Obtaining SCR Costs for Coal
The updated performance and cost models for SCR and SNCR technologies developed for EPA by
Sargent & Lundy are implemented in EPA Platform v6 to develop the capital, fixed O&M (FOM), and
variable O&M (VOM) components of cost. For details of Sargent & Lundy SCR and SNCR cost models,
see Attachment 5-3 and Attachment 5-4.
Table 5-5 presents the SCR and SNCR capital, VOM, and FOM costs and capacity and heat rate
penalties for an illustrative set of coal generating units with a representative range of capacities and heat
rates.
5-6
-------
Table 5-5 Illustrative Post-combustion NOx Control Costs (2016$) for Coal Plants for Representative Sizes and Heat Rates under the
Assumptions in EPA Platform v6
Capacity (MW)
Control Type
Heat Rate
Capacity
Penalty
(%)
Heat Rate
Variable
O&M
100
300
500
700
1000
(Btu/kWh)
Penalty (%)
(mills/kWh)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
SCR
9,000
-0.54
0.54
1.35
373
1.95
304
0.85
282
0.72
269
0.66
257
0.60
Minimum Cutoff: > 25 MW
10,000
-0.56
0.56
1.45
405
2.06
333
0.91
309
0.78
295
0.71
282
0.66
11,000
-0.58
0.59
1.56
437
2.17
361
0.97
335
0.83
321
0.76
307
0.71
SNCR - Tangential, 25%
Removal Efficiency
Minimum Cutoff: > 25 MW
9,000
10,000
-0.05
0.78
1.17
1.29
55
56
0.49
0.51
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Maximum Cutoff: 200 MW
11,000
1.42
58
0.52
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
SNCR - Tangential, 20%
Removal Efficiency
Minimum Cutoff: > 200 MW
9,000
10,000
-0.05
0.63
0.93
1.04
N/A
N/A
N/A
N/A
29
30
0.26
0.26
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
Maximum Cutoff: 400 MW
11,000
1.14
N/A
N/A
31
0.27
N/A
N/A
N/A
N/A
N/A
N/A
SNCR - Tangential, 15%
Removal Efficiency
Minimum Cutoff: > 400 MW
9,000
10,000
-0.05
0.49
0.70
0.78
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
21
22
0.19
0.19
18
18
0.16
0.16
15
15
0.13
0.13
Maximum Cutoff: None
11,000
0.85
N/A
N/A
N/A
N/A
22
0.20
19
0.16
15
0.13
SNCR - Fluidized Bed
9,000
-0.05
1.51
2.33
44
0.38
24
0.21
18
0.16
15
0.13
12
0.11
Minimum Cutoff: > 25 MW
10,000
2.59
45
0.39
24
0.21
18
0.16
15
0.13
12
0.11
Maximum Cutoff: None
11,000
2.85
46
0.40
25
0.22
19
0.16
15
0.13
13
0.11
Note 1: *Assumes Bituminous Coal, NOx rate: 0.5 Ib/MMBtu, and SO2 rate: 2.0 Ib/MMBtu
Note 2: The Variable O&M costs in this table do not include the cost of additional auxiliary power (VOMP) component in the S&L tool as for modeling purposes, IPM reflects the auxiliary power consumption
through capacity penalty.
Note 3: Heat rate penalty includes the effect of capacity penalty.
5-7
-------
5.2.4 Methodology for Obtaining SCR Costs for Oil/Gas Steam Units
The cost calculations for SCR described in section 5.2.3 apply to coal units. For SCR on oil/gas steam
units, the cost calculation procedure shown in Table 5-6 is used. The scaling factor for capital and fixed
O&M costs, described in footnote a, applies to all size units from 25 MW and up.
Table 5-6 Post-Combustion NOx Controls for Oil/Gas Steam Units in EPA Platform v6
Post-Combustion
Capital
Fixed O&M
Variable O&M
Percent
Control Technology
($/kW)
($/kW-yr)
($/MWh)
Removal
SCRa
86.38
1.25
0.14
80%
Notes:
The "Coefficients" in the table above are multiplied by the terms below to determine costs.
"MW" in the terms below is the unit's capacity in megawatts.
Cost data are adjusted to 2016$ by EPA.
a SCR Cost Equations:
SCR Capital Cost and Fixed O&M: (200/MW)0 35
The scaling factors shown above apply up to 500 MW. The cost obtained for a 500 MW unit applies for units larger
than 500 MW.
Example for 275 MW unit:
SCR Capital Cost ($/kW) = 86.38 * (200/275)0 35 = 77.27 $/kW
SCR FOM Cost ($/kW-yr) = 1.25 * (200/275)°35 = 1.12 $/kW-yr
SCR VOM Cost ($/MWh) = 0.14 $/MWh
5.2.5 Methodology for Obtaining SNCR Costs for Coal
In the Sargent & Lundy's cost update for SNCR, the NOx removal efficiency varies by unit size and burner
type as summarized in Table 5-4. Additionally, the capital, fixed, and variable operating and maintenance
costs of SNCR on circulating fluidized bed (CFB) units are distinguished from the corresponding costs for
other boiler types (e.g., cyclone and wall fired). As with SCR, an air heater modification cost applies for
plants that burn bituminous coal whose SO2 content is 3 Ibs/MMBtu or greater.
5.2.6 SO2 Controls for Units with Capacities from 25 MW to 100 MW (25 MW < capacity <100
MW)
In EPA Platform v6, coal units with capacities between 25 MW and 100 MW are offered the same SO2
control options as larger units. However, for purposes of modeling, the costs of controls for these units
are assumed to be equivalent to that of a 50 MW for Dry FGD and 100 MW for Wet FGD. These
assumptions are based on several considerations. First, to achieve economies of scale, several units in
this size range are likely to be ducted to share a single common control, so the minimum capacity cost
equivalency assumption, though generic, would be technically plausible. Second, single units in this size
range that are not grouped to achieve economies of scale are likely to switch to a lower sulfur coal,
repower or convert to natural gas firing, use dry sorbent injection, and/or reduce operating hours.
Illustrative scrubber costs for 25-100 MW coal units with a range of heat rates can be found by referring to
the LSFO 100 MW and LSD 100MW "Capital Costs ($/kW)" and "Fixed O&M" columns in Table 5-3. The
Variable O&M cost component, which applies to units regardless of size, can be found in the fifth column
in this table.
5.3 Biomass Co-firing
Biomass co-firing is provided as an option for those coal-fired units in EPA Platform v6 that per EIA Form
923 had co-fired biomass during the 2012-2016 period. Table 5-7 lists the units provided with the co-
firing option and the limit on share of the biomass co-firing. The remaining coal power plants are not
5-8
-------
provided this choice as logistics and boiler engineering considerations place limits on the extent of
biomass that can be fired. The logistical considerations arise primarily because biomass is only economic
to transport a limited distance from where it is grown due to its relatively low energy density. In addition,
the extent of storage that can be devoted at a power plant to such a fuel is another limiting factor. Boiler
efficiency and other engineering considerations, largely driven by the relatively higher moisture content
and lower heat content of biomass compared to fossil fuel, also plays a role in limiting the potential
adoption of co-firing.
Table 5-7 Coal Units with Biomass Co-firing Option in EPA Platform v6
Plant Name
Unit ID
Biomass Co-Firing Share Limit (%)41
Virginia City Hybrid Energy Center
1
8
Virginia City Hybrid Energy Center
2
8
University of Iowa Main Power Plant
BLR11
10
University of Iowa Main Power Plant
BLR10
10
Northampton Generating Company LP
BLR1
10
TES Filer City Station
2
10
TES Filer City Station
1
10
Manitowoc
9
10
Schiller
6
10
Schiller
4
10
T B Simon Power Plant
BLR4
10
5.4 Mercury Control Technologies
For any power plant, mercury emissions depend on the mercury content of the fuel used, the combustion
and physical characteristics of the unit, and the emission control technologies deployed. In the absence
of activated carbon injection (ACI), mercury emission reductions below the mercury content of the fuel are
strictly due to characteristics of the combustion process and incidental removal resulting from other
pollution control technologies, e.g., the SO2, NOx, and particulate matter controls. The following
discussion is divided into three parts. Sections 5.4.1 and 5.4.2 explain the two factors that determine
mercury emissions that result from unit configurations lacking ACI under EPA Platform v6. Section 5.4.1
discusses how mercury content of fuel is modeled in EPA Platform v6. Section 5.4.2 looks at the
procedure used in EPA Platform v6 to capture the mercury reductions resulting from different unit and
(non-mercury) control configurations. Section 5.4.3 explains the mercury emission control options that
are available under EPA Platform v6. Each section indicates the data sources and methodology used.
5.4.1 Mercury Content of Fuels
Coal
Assumptions pertaining to the mercury content of coal (and the majority of emission modification factors
discussed below in Section 5.4.2) are derived from EPA's "Information Collection Request for Electric
Utility Steam Generating Unit Mercury Emissions Information Collection Effort" (ICR).42 A two-year effort
41 In EPA Platform v6, the limit on biomass co-firing is expressed as the percentage of the facility (ORIS code) level
power output that is produced from biomass. Based on analysis by EPA's power sector engineering staff, a
maximum of 10% of the facility level power output (not to exceed 50 MW) can be fired by biomass in modeling
projections.
42 Data from the ICR can be found at http://www.epa.qov/ttn/atw/combust/utiItox/mercurv.html. In 2009, EPA
collected some additional information regarding mercury through the Collection Effort for New and Existing Coal- and
Oil-Fired Electricity Utility Steam Generating Units (EPA ICR No.2362.01 (OMB Control Number 2060-0631),
however the information collected was not similarly comprehensive and was thus not used to update mercury
assumptions in this EPA Platform.
5-9
-------
initiated in 1998 and completed in 2000, the ICR had three main components: (1) identifying all coal-fired
units owned and operated by publicly-owned utility companies, Federal power agencies, rural electric
cooperatives, and investor-owned utility generating companies, (2) obtaining "accurate information on the
amount of mercury contained in the as-fired coal used by each electric utility steam generating unit with a
capacity greater than 25 megawatts electric [MWe]), as well as accurate information on the total amount
of coal burned by each such unit," and (3) obtaining data by coal sampling and stack testing at selected
units to characterize mercury reductions from representative unit configurations.
The ICR resulted in more than 40,000 data points indicating the coal type, sulfur content, mercury content
and other characteristics of coal burned at coal-fired utility units greater than 25 MW. To make this data
usable, these data points were first grouped by IPM coal types and IPM coal supply regions. IPM coal
types divide bituminous, subbituminous, and lignite coal into different grades based on sulfur content.
Oil, natural gas, and waste fuels
Assumptions pertaining to the mercury content for oil, gas, and waste fuels are based on data derived
from previous EPA analysis of mercury emissions from power plants.43 Table 5-8 provides a summary of
the assumptions on the mercury content for oil, gas, and waste fuels.
Table 5-8 Assumptions on Mercury Concentration in Non-Coal Fuel in EPA Platform v6
Fuel Type
Mercury Concentration (Ibs/TBtu)
Oil
0.48
Natural Gas
0.00a
Petroleum Coke
2.66
Biomass
0.57
Municipal Solid Waste
71.85
Geothermal Resource
2.97-3.7
Note:
a The values appearing in this table are rounded to two decimal places. The zero value shown for natural gas is
based on an EPA study that found a mercury content of 0.000138 Ibs/TBtu. Values for geothermal resources
represent a range.
5.4.2 Mercury Emission Modification Factors
Emission Modification Factors (EMFs) represent the mercury reductions attributable to the specific burner
type and configuration of SO2, NOx, and particulate matter control devices at an electric generating unit.
An EMF is the ratio of outlet mercury concentration to inlet mercury concentration, and depends on the
unit's burner type, particulate control device, post-combustion NOx control and SO2 scrubber control. In
other words, the mercury reduction achieved (relative to the inlet) during combustion and flue-gas
treatment process is (1-EMF), such that the lower the EMF, the greater the mercury reduction. If the EMF
is 0.25, then only 25% of the inlet mercury concentration is emitted as outlet mercury concentration, and
therefore the unit has achieved a 75% reduction in mercury that would otherwise be emitted without the
properties influencing the EMF. The EMF varies by the type of coal (bituminous, subbituminous, and
lignite) used during the combustion process.
Deriving EMFs involves obtaining mercury inlet data by coal sampling and mercury emission data by
stack testing at a representative set of coal units. As noted above, EPA's EMFs were initially based on
1999 mercury ICR emission test data. More recent testing conducted by the EPA, DOE, and industry
43 Analysis of Emission Reduction Options for the Electric Power Industry," Office of Air and Radiation, U.S. EPA,
March 1999.
5-10
-------
participants44 has provided a better understanding of mercury emissions from electric generating units
and mercury capture in pollution control devices. Overall the 1999 ICR data revealed higher levels of
mercury capture for bituminous coal-fired plants than for subbituminous and lignite coal-fired plants, and
significant capture of ionic Hg in wet-FGD scrubbers. Additional mercury testing indicates that for
bituminous coals, SCR systems have the ability to convert elemental Hg into ionic Hg and thus allow
easier capture in a downstream wet-FGD scrubber. This understanding of mercury capture with SCRs is
incorporated in EPA Platform v6 mercury EMFs for unit configurations with SCR and wet scrubbers.
Table 5-9 below provides a summary of EMFs used in EPA Platform v6. Table 5-10 provides definitions
of acronyms for existing controls that appear in Table 5-9. Table 5-11 provides a key to the burner type
designations appearing in Table 5-9.
Table 5-9 Mercury Emission Modification Factors Used in EPA Platform v6
Burner
Type
Particulate Control
Post-
combustion
Control - NOx
Post-
combustion
Control - S02
Bituminous
EMF
Subbituminous
EMF*
Lignite
EMF
FBC
Cold Side ESP
No SCR
None
0.65
0.1
0.62
FBC
Cold Side ESP
No SCR
Dry FGD
0.64
0.1
1
FBC
Cold Side ESP + FF
No SCR
None
0.05
0.1
0.43
FBC
Cold Side ESP + FF
No SCR
Dry FGD
0.05
0.1
1
FBC
Fabric Filter
No SCR
None
0.05
0.1
0.43
FBC
Fabric Filter
No SCR
Dry FGD
0.05
0.1
0.43
FBC
Hot Side ESP + FGC
No SCR
None
1
0.1
1
FBC
Hot Side ESP + FGC
No SCR
Dry FGD
0.6
0.1
1
FBC
No Control
No SCR
None
1
0.1
1
Non
FBC
Cold Side ESP
SCR
None
0.64
0.1
1
Non
FBC
Cold Side ESP
SCR
Wet FGD
0.1
0.1
0.56
Non
FBC
Cold Side ESP
SCR
Dry FGD
0.64
0.1
1
Non
FBC
Cold Side ESP
No SCR
None
0.64
0.1
1
Non
FBC
Cold Side ESP
No SCR
Wet FGD
0.05
0.1
0.56
Non
FBC
Cold Side ESP
No SCR
Dry FGD
0.64
0.1
1
Non
FBC
Cold Side ESP + FF
SCR
None
0.2
0.1
1
Non
FBC
Cold Side ESP + FF
SCR
Wet FGD
0.1
0.1
0.56
Non
FBC
Cold Side ESP + FF
SCR
Dry FGD
0.05
0.1
1
Non
FBC
Cold Side ESP + FF
No SCR
None
0.2
0.1
1
Non
FBC
Cold Side ESP + FF
No SCR
Wet FGD
0.05
0.1
0.56
Non
FBC
Cold Side ESP + FF
No SCR
Dry FGD
0.05
0.1
1
Non
FBC
Cold Side ESP + FGC
SCR
None
0.64
0.1
1
Non
FBC
Cold Side ESP + FGC
SCR
Wet FGD
0.1
0.1
0.56
Non
FBC
Cold Side ESP + FGC
SCR
Dry FGD
0.64
0.1
1
Non
FBC
Cold Side ESP + FGC
No SCR
None
0.64
0.1
1
Non
FBC
Cold Side ESP + FGC
No SCR
Wet FGD
0.05
0.1
0.56
Non
FBC
Cold Side ESP + FGC
No SCR
Dry FGD
0.64
0.1
1
Non
FBC
Cold Side ESP + FGC + FF
SCR
None
0.2
0.1
1
Non
FBC
Cold Side ESP + FGC + FF
SCR
Wet FGD
0.1
0.1
0.56
Non
FBC
Cold Side ESP + FGC + FF
SCR
Dry FGD
0.05
0.1
1
Non
FBC
Cold Side ESP + FGC + FF
No SCR
None
0.2
0.1
1
Non
FBC
Cold Side ESP + FGC + FF
No SCR
Wet FGD
0.05
0.1
0.56
Non
FBC
Cold Side ESP + FGC + FF
No SCR
Dry FGD
0.05
0.1
1
Non
FBC
Fabric Filter
SCR
None
0.11
0.1
1
Non
FBC
Fabric Filter
SCR
Wet FGD
0.1
0.1
0.56
Non
FBC
Fabric Filter
SCR
Dry FGD
0.05
0.1
1
44 For a detailed summary of emissions test data see Control of Emissions from Coal-Fired Electric Utility Boilers: An
Update, EPA/Office of Research and Development, February 2005. This report can be found at
www.epa.aov/ttnatw01/utilitv/hawhitepaperfinal.pdf.
5-11
-------
Burner
Type
Particulate Control
Post-
combustion
Control - NOx
Post-
combustion
Control - S02
Bituminous
EMF
Subbituminous
EMF*
Lignite
EMF
Non FBC
Fabric Filter
No SCR
None
0.11
0.1
1
Non FBC
Fabric Filter
No SCR
Wet FGD
0.1
0.1
0.56
Non FBC
Fabric Filter
No SCR
Dry FGD
0.05
0.1
1
Non FBC
Hot Side ESP
SCR
None
0.9
0.1
1
Non FBC
Hot Side ESP
SCR
Wet FGD
0.1
0.1
1
Non FBC
Hot Side ESP
SCR
Dry FGD
0.6
0.1
1
Non FBC
Hot Side ESP
No SCR
None
0.9
0.1
1
Non FBC
Hot Side ESP
No SCR
Wet FGD
0.05
0.1
1
Non FBC
Hot Side ESP
No SCR
Dry FGD
0.6
0.1
1
Non FBC
Hot Side ESP + FF
SCR
None
0.11
0.1
1
Non FBC
Hot Side ESP + FF
SCR
Wet FGD
0.1
0.1
0.56
Non FBC
Hot Side ESP + FF
SCR
Dry FGD
0.05
0.1
1
Non FBC
Hot Side ESP + FF
No SCR
None
0.11
0.1
1
Non FBC
Hot Side ESP + FF
No SCR
Wet FGD
0.03
0.1
0.56
Non FBC
Hot Side ESP + FF
No SCR
Dry FGD
0.05
0.1
1
Non FBC
Hot Side ESP + FGC
SCR
None
0.9
0.1
1
Non FBC
Hot Side ESP + FGC
SCR
Wet FGD
0.1
0.1
1
Non FBC
Hot Side ESP + FGC
SCR
Dry FGD
0.6
0.1
1
Non FBC
Hot Side ESP + FGC
No SCR
None
0.9
0.1
1
Non FBC
Hot Side ESP + FGC
No SCR
Wet FGD
0.05
0.1
1
Non FBC
Hot Side ESP + FGC
No SCR
Dry FGD
0.6
0.1
1
Non FBC
Hot Side ESP + FGC + FF
SCR
Dry FGD
0.05
0.1
1
Non FBC
Hot Side ESP + FGC + FF
No SCR
None
0.11
0.1
1
Non FBC
Hot Side ESP + FGC + FF
No SCR
Dry FGD
0.05
0.1
1
Non FBC
No Control
SCR
None
1
0.1
1
Non FBC
No Control
SCR
Wet FGD
0.1
0.1
1
Non FBC
No Control
SCR
Dry FGD
0.6
0.1
1
Non FBC
No Control
No SCR
None
1
0.1
1
Non FBC
No Control
No SCR
Wet FGD
0.58
0.1
1
Non FBC
No Control
No SCR
Dry FGD
0.6
0.1
1
Non FBC
PM Scrubber
SCR
None
0.9
0.1
1
Non FBC
PM Scrubber
SCR
Wet FGD
0.1
0.1
1
Non FBC
PM Scrubber
SCR
Dry FGD
0.6
0.1
1
Non FBC
PM Scrubber
No SCR
None
0.9
0.1
1
Non FBC
PM Scrubber
No SCR
Wet FGD
0.05
0.1
1
Non FBC
PM Scrubber
No SCR
Dry FGD
0.6
0.1
1
Note: 2017 annual emissions data suggests that, with subbituminous coal, many configurations are now achieving at least 90%
removal of mercury. This table was updated from previous versions to reflect this recent observation. For 2017 emissions data,
see: https://ampd.epa.gov.
Table 5-10 Definition of Acronyms for Existing Controls
Acronym
Description
ESP
Electrostatic Precipitator - Cold Side
HESP
Electrostatic Precipitator - Hot Side
ESP/O
Electrostatic Precipitator - Other
FF
Fabric Filter
FGD
Flue Gas Desulfurization - Wet
DS
Flue Gas Desulfurization - Dry
SCR
Selective Catalytic Reduction
PMSCRUB
Particulate Matter Scrubber
5-12
-------
Table 5-11 Key to Burner Type Designations in Table 5-9
"PC" refers to conventional pulverized coal boilers. Typical configurations include wall-fired and tangentially fired
boilers (also called T-fired boilers). In wall-fired boilers the burner's coal and air nozzles are mounted on a single
wall or opposing walls. In tangentially fired boilers the burner's coal and air nozzles are mounted in each corner of
the boiler.
"Cyclone" refers to cyclone boilers where air and crushed coal are injected tangentially into the boiler through a
"cyclone burner" and "cyclone barrel" which create a swirling motion allowing smaller coal particles to be burned in
suspension and larger coal particles to be captured on the cyclone barrel wall where they are burned in molten slag.
"Stoker" refers to stoker boilers where lump coal is fed continuously onto a moving grate or chain, which moves the
coal into the combustion zone in which air is drawn through the grate and ignition takes place. The carbon
gradually burns off, leaving ash which drops off at the end into a receptacle, from which it is removed for disposal.
"FBC" refers to "fluidized bed combustion" where solid fuels are suspended on upward-blowing jets of air, resulting
in a turbulent mixing of gas and solids and a tumbling action which provides especially effective chemical reactions
and heat transfer during the combustion process.
"Other" refers to miscellaneous burner types including cell burners and arch-, roof-, and vertically-fired burner
configurations.
5.4.3 Mercury Control Capabilities
EPA Platform v6 offers two options for mercury pollution control: (1) combinations of SO2, NOx, and
particulate controls which deliver mercury reductions as a co-benefit; and (2) Activated Carbon Injection
(ACI), a retrofit option specifically designed for mercury control. These two options are discussed below.
Mercury Control through SO2 and NOx Retrofits
Units that install SO2, NOx, and particulate controls reduce mercury emissions as a byproduct of these
retrofits. Section 5.4.2 described how EMFs are used to capture mercury emissions depending on the rank
of coal burned, the generating unit's combustion characteristics, and the specific configuration of SO2,
NOx, and particulate controls (i.e., hot and cold-side electrostatic precipitators (ESPs), fabric filters (also
called "baghouses"), and particulate matter (PM) scrubbers).
Activated Carbon Injection (ACQ
The technology used for mercury control in EPA Platform v6 is Activated Carbon Injection (ACI) downstream
of the combustion process in coal fired units. Sargent & Lundy's updated cost and performance
assumptions for ACI are used (and are described further below).
Three alternative ACI options are represented as capable of providing 90% mercury removal for all possible
configurations of boiler, emission controls, and coal types used in the U.S. electric power sector. The
three ACI options differ, based on whether they are used in conjunction with an electrostatic precipitator
(ESP) or a fabric filter (also called a "baghouse"). The three ACI options are:
• ACI with Existing ESP
• ACI with Existing Baghouse
• ACI with an Additional Baghouse (also referred to as Toxecon)
In the third option listed above the additional baghouse is installed downstream of the pre-existing
particulate matter device and the activated carbon is injected after the existing controls. This
configuration allows the fly ash to be removed before it is contaminated by the mercury.
For modeling purposes, EPA assumes that all three configurations use brominated ACI, where a small
amount of bromine is chemically bonded to the powdered carbon, which is injected into the flue gas
stream. EPA recognizes that amended silicates and possibly other non-carbon, non-brominated
5-13
-------
substances are in development and may become available as alternatives to brominated carbon as a
mercury sorbent.
The applicable ACI option depends on the coal type burned, its SO2 content, the boiler and particulate
control type, and in some instances consideration of whether an SO2 scrubber (FGD) system and SCR
NOx post-combustion control are present. Table 5-12 shows the ACI assignment scheme used to achieve
90% mercury removal. EPA Platform v6 does not explicitly model ACI retrofit options.
5-14
-------
Table 5-12 Assignment Scheme for Mercury Emissions Control Using Activated Carbon Injection (ACI) in EPA Platform v6
Air pollution controls
Bituminous Coal
Subbituminous Coal
Lignite Coal
Sorbent Inj
Sorbent Inj
Sorbent Inj
Burner Type
Particulate Control Type
SCR
FGD
ACI
Toxecon
Rate
ACI
Toxecon
Rate
ACI
Toxecon
Rate
System
System
Required?
Required?
(lb/million
acfm)
Required?
Required?
(lb/million
acfm)
Required?
Required?
(lb/million
acfm)
FBC
Cold Side ESP + Fabric Filter without FGC
-
-
Yes
No
2
Yes
No
2
Yes
No
2
FBC
Cold Side ESP without FGC
-
-
Yes
No
5
Yes
No
5
Yes
No
5
FBC
Fabric Filter
-
Dry FGD
No
No
0
Yes
No
2
Yes
No
2
FBC
Fabric Filter
-
-
Yes
No
2
Yes
No
2
Yes
No
2
FBC
Hot Side ESP with FGC
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
-
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
-
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
-
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
SCR
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
SCR
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter with FGC
SCR
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
-
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
-
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
-
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
SCR
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
SCR
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP + Fabric Filter without FGC
SCR
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Cold Side ESP with FGC
-
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP with FGC
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP with FGC
-
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP with FGC
SCR
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP with FGC
SCR
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP with FGC
SCR
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Cold Side ESP without FGC
-
-
Yes
No
5
Yes
No
5
Yes
No
5
Non-FBC
Cold Side ESP without FGC
-
Wet FGD
Yes
No
5
Yes
No
5
Yes
No
5
Non-FBC
Cold Side ESP without FGC
SCR
-
Yes
No
5
Yes
No
5
Yes
No
5
Non-FBC
Cold Side ESP without FGC
SCR
Dry FGD
Yes
No
5
Yes
No
5
Yes
No
5
Non-FBC
Cold Side ESP without FGC
SCR
Wet FGD
Yes
No
5
Yes
No
5
Yes
No
5
Non-FBC
Fabric Filter
-
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Fabric Filter
-
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Fabric Filter
-
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Fabric Filter
SCR
Dry FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Fabric Filter
SCR
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Fabric Filter
SCR
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
-
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
-
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
-
Dry FGD
No
No
0
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
SCR
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
SCR
Dry FGD
No
No
0
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter with FGC
SCR
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
-
Dry FGD
No
No
0
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
-
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
-
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
SCR
Dry FGD
No
No
0
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
SCR
-
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP + Fabric Filter without FGC
SCR
Wet FGD
Yes
No
2
Yes
No
2
Yes
No
2
Non-FBC
Hot Side ESP with FGC
-
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP with FGC
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP with FGC
-
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
5-15
-------
Air pollution controls
Bituminous Coal
Subbituminous Coal
Lignite Coal
Sorbent Inj
Sorbent Inj
Sorbent Inj
Burner Type
Particulate Control Type
SCR
System
FGD
System
ACI
Required?
Toxecon
Required?
Rate
(lb/million
acfm)
ACI
Required?
Toxecon
Required?
Rate
(lb/million
acfm)
ACI
Required?
Toxecon
Required?
Rate
(lb/million
acfm)
Non-FBC
Hot Side ESP with FGC
SCR
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP with FGC
SCR
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP with FGC
SCR
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
-
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
-
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
SCR
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
SCR
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
Hot Side ESP without FGC
SCR
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
-
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
-
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
SCR
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
SCR
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
No Control
SCR
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
-
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
-
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
-
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
SCR
Dry FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
SCR
-
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Non-FBC
PM Scrubber
SCR
Wet FGD
Yes
Yes
2
Yes
Yes
2
Yes
Yes
2
Note: In the table above "Toxecon" refers to the option described as "ACI System with an Additional Baghouse" and "ACI + Full Baghouse with a Sorbent Injection (Inj) Rate of 2 lbs/million acfm" elsewhere
in this chapter.
5-16
-------
5.4.4 Methodology for Obtaining ACI Control Costs
The updated ACI model developed by Sargent & Lundy assumes that the carbon feed rate dictates the
size of the equipment and resulting costs. The feed rate in turn is a function of the required removal (in
this case 90%) and the type of particulate control device. The model assumes a carbon feed rate of 5
pounds of carbon injected for every 1,000,000 actual cubic feet per minute (acfm) of flue gas would
provide the stipulated 90% mercury removal rate for units shown in Table 5-13 as qualifying for ACI
systems with existing ESP. For generating units with fabric filters a lower injection rate of 2 pound per
million acfm is required. Alternative sets of costs were developed for each of the three ACI options: ACI
systems for units with existing ESPs, ACI for units with existing fabric filters (baghouses), and the
combined cost of ACI plus an additional baghouse for units that either have no existing particulate control
or that require ACI plus a baghouse in addition to their existing particulate control. There are various
reasons that a combined ACI plus additional baghouse would be required. These include situations
where the existing ESP cannot handle the additional particulate load associate with the ACI or where S03
injection is currently in use to condition the flue gas for the ESP. Another cause for combined ACI and
baghouse is use of PRB coal whose combustion produces mostly elemental mercury, not ionic mercury,
due to this coal's low chlorine content.
For the combined ACI and fabric filter option a full size baghouse with an air-to-cloth (A/C) ratio of 4.0 is
assumed, as opposed to a polishing baghouse with a 6.0 A/C ratio45.
Table 5-13 presents the capital, VOM, and FOM costs as well as the capacity and heat rate penalties for
the three ACI options represented in EPA Platform v6. For each ACI option, values are shown for an
illustrative set of generating units with a representative range of capacities and heat rates. For details of
Sargent & Lundy ACI cost model, see Attachment 5-6.
5.5 Hydrogen Chloride (HCI) Control Technologies
The following sub-sections describe how HCI emissions from coal are represented, the emission control
technologies available for HCI removal, and the cost and performance characteristics of these
technologies in EPA Platform v6.
5.5.1 Chlorine Content of Fuels
HCI emissions from the power sector result from the chlorine content of the coal that is combusted by
electric generating units. Data on chlorine content of coals had been collected as part of EPA's 1999
"Information Collection Request for Electric Utility Steam Generating Unit Mercury Emissions Information
Collection Effort" (ICR 1999) described above in section 5.4.1. This data is incorporated into the model in
order to provide the capability for EPA Platform v6 to project HCI emissions. The procedures used for this
are presented below.
Western subbituminous coal (such as that mined in the Powder River Basin) and lignite coal contain
natural alkalinity in the form of non-glassy calcium oxide (CaO) and other alkaline and alkaline earth
oxides. This fly ash (classified as 'Class C' fly ash) has a natural pH of 9 and higher and the natural
alkalinity can effectively neutralize much of the HCI in the flue gas stream prior to the primary control
device.
45 The "air-to-cloth" (A/C) ratio is the volumetric flow, (typically expressed in Actual Cubic Feet per Minute, ACFM) of
flue gas entering the baghouse divided by the areas (typically in square feet) of fabric filter cloth in the baghouse.
The lower the A/C ratio, e.g., A/C = 4.0 compared to A/C = 6.0, the greater area of the cloth required and the higher
the cost for a given volumetric flow
5-17
-------
Table 5-13 Illustrative Activated Carbon Injection (ACI) Costs (2016$) for Representative Sizes and Heat Rates
under the Assumptions in EPA Platform v6
Heat
Rate
Penalty
(%)
Capacity (MW)
Heat Rate
(Btu/kWh)
Capacity
Variable
100
300
500
700
1000
Control Type
Penalty
(%)
O&M cost
(mllls/kWh)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
ACI System with an
Existing ESP ACI with a
Sorbent Injection Rate of
9,000
-0.02
0.02
2.22
40.08
0.32
15.76
0.13
10.21
0.08
7.67
0.06
5.66
0.05
5 lbs/million acfm
assuming Bituminous
Coal
10,000
-0.02
0.02
2.47
40.73
0.33
16.01
0.13
10.37
0.08
7.79
0.06
5.75
0.05
11,000
-0.02
0.02
2.71
41.32
0.33
16.24
0.13
10.52
0.08
7.90
0.06
5.84
0.05
ACI System with an
Existing Baghouse ACI
with a Sorbent Injection
9,000
-0.02
0.02
1.59
34.94
0.28
13.74
0.11
8.90
0.07
6.68
0.05
4.94
0.04
Rate of 2 lbs/million
acfm Assuming
Bituminous Coal
10,000
-0.02
0.02
1.77
35.50
0.29
13.95
0.11
9.04
0.07
6.79
0.05
5.01
0.04
11,000
-0.02
0.02
1.95
36.02
0.29
14.16
0.11
9.17
0.07
6.89
0.06
5.09
0.04
ACI System with an
Additional Baghouse
ACI + Full Baghouse
9,000
-0.62
0.62
0.47
293.68
1.03
221.56
0.77
196.97
0.69
182.87
0.64
169.38
0.59
with a Sorbent Injection
Rate of 2 lbs/million
acfm Assuming
Bituminous Coal
10,000
-0.62
0.62
0.52
316.91
1.11
240.14
0.84
213.77
0.75
198.60
0.69
184.06
0.64
11,000
-0.62
0.62
0.58
339.68
1.19
258.37
0.90
230.25
0.81
214.02
0.75
198.45
0.69
Note 1: The above cost estimates assume bituminous coal consumption.
Note 2: The Variable O&M costs in this table do not include the cost of additional auxiliary power (VOMP) component in the S&L tool as for modeling purposes, IPM reflects the
auxiliary power consumption through capacity penalty.
5-18
-------
Eastern bituminous coals, by contrast, tend to produce fly ash with lower natural alkalinity. Though
bituminous fly ash (classified as 'Class F' fly ash) may contain calcium, it tends to be present in a glassy
matrix and unavailable for acid-base neutralization reactions.
In order to assess the extent of expected natural neutralization, resulting in large part from the alkalinity of
the fly ash, the 2010 ICR46 data was examined. According to that data, units burning some of the
subbituminous coals without operating acid gas control technology emitted substantially lower HCI
emissions than would otherwise be expected if the emissions were based solely on the chlorine content
of those coals. Comparing the assumed CI content of the subbituminous coals modeled in EPA Platform
v6 with the estimated values based on responses to the 2010 ICR supports the EPA Platform v6
assumption that combustion of subbituminous and lignite coals results in a 95% reduction in HCI
emissions relative to the assumed chlorine content of the coal.
5.5.2 HCI Removal Rate Assumptions for Existing and Potential Units
SO2 emission controls on existing and new (potential) units provide the HCI reductions indicated in Table
5-14. New supercritical pulverized coal units (column 3) that the model builds include FGD (wet or dry)
which is assumed to provide a 99% removal rate for HCI. For existing conventional pulverized coal units
with pre-existing FGD (column 5), the HCI removal rate is assumed to be 5% higher than the reported SO2
removal rate up to a maximum of 99% removal. In addition, for fluidized bed combustion units (column 4)
with no FGD and no fabric filter, the HCI removal rate is assumed to be the same as the SO2 removal rate
up to a maximum of 95%. FBCs with fabric filters are assumed to have an HCI removal rate of 95%.
Table 5-14 HCI Removal Rate Assumptions for Potential (New) and Existing Units
in EPA Platform v6
Potential (New)
Existing Units with FGD
Gas
Controls ==>
Ultra-Supercritical
Pulverized Coal with
30%/90% CCS
Fluidized Bed Combustion (FBC)
Conventional Pulverized Coal
(CPC) with Wet or Dry FGD
HCI
Removal
Rate
99%
Without fabric filter:
Same as reported SO2 removal rate
up to a maximum of 95%
With fabric filter: 95%
Reported SO2 removal rate +
5% up to a maximum of 99%
5.5.3 HCI Retrofit Emission Control Options
The retrofit options for HCI emission control are discussed in detail in the following sub-sections and
summarized in Table 5-15.
Wet and Dry FGD
In addition to providing SO2 reductions, wet scrubbers (Limestone Forced Oxidation, LSFO) and dry
scrubbers (Lime Spray Dryer, LSD) reduce HCI as well. For both LSFO and LSD the HCI removal rate is
assumed to be 99% with a floor of 0.001 Ibs/MMBtu. This is summarized in columns 2-5 of Table 5-15.
46 Collection Effort for New and Existing Coal- and Oil-Fired Electricity Utility Steam Generating Units (EPA ICR
No.2362.01 (OMB Control Number 2060-0631)
5-19
-------
Table 5-15 Summary of Retrofit HCI (and SO2) Emission Control Performance
Assumptions in EPA Platform v6
Performance
Assumptions
Limestone Forced Oxidation
(LSFO)
Lime Spray Dryer (LSD)
Dry Sorbent Injection (DSI)
SO2
HCI
SO2
HCI
SO2
HCI
Percent
Removal
98%
with a floor of
0.06
Ibs/MMBtu
99%
with a floor of
0.001
Ibs/MMBtu
95%
with a floor
of
0.08
Ibs/MMBtu
99%
with a floor
of
0.001
Ibs/MMBtu
50%
98%
with a floor of
0.002 Ibs/MMBtu
Capacity
Penalty
Calculated based on
characteristics of the unit:
See Table 5-3
Calculated based on
characteristics of the unit:
See Table 5-3
Calculated based on
characteristics of the unit: See
Excerpt from Table 5-17
Heat Rate
Penalty
Cost (2011$)
Applicability
Units > 25 MW
Units > 25 MW
Units >25 MW
Sulfur Content
Applicability
Coals < 3.0 lbs of
S02/MMBtu
Coals < 2.0 lbs of S02/MMBtu
Applicable
Coal
Types
BA, BB, BD, BE, BG, BH, SA, SB,
SD, SE, LD, LE, LG, LH, PK, and
WC
BA, BB, BD, BE, SA, SB,
SD, SE, LD, and LE
BA, BB, BD, SA, SB, SD, and
LD
Dry Sorbent Injection
EPA Platform v6 includes dry sorbent injection (DSI) as a retrofit option for achieving (in combination with a
particulate control device) both SO2 and HCI removal. In DSI for HCI reduction, a dry sorbent is injected
into the flue gas duct where it reacts with the HCI and SO2 in the flue gas to form compounds that are then
captured in a downstream fabric filter or electrostatic precipitator (ESP) and disposed of as waste. A
sorbent is a material that takes up another substance by either adsorption on its surface or absorption
internally or in solution. A sorbent may also chemically react with another substance. The sorbent
assumed in the cost and performance characterization discussed in this section is Trona (sodium
sesquicarbonate), a sodium-rich material with major underground deposits found in Sweetwater County,
Wyoming. Trona is typically delivered with an average particle size of 30 jjm diameter, but can be
reduced to about 15 jjm through onsite in-line milling to increase its surface area and capture capability.
While the Sargent & Lundy description of the DSI technology includes references to the hydrated lime
option, only the Trona option is implemented in EPA Platform v6.
Removal rate assumptions: The removal rate assumptions for DSI are summarized in Table 5-15. The
assumptions shown in the last two columns of Table 5-15 were derived from assessments by EPA
engineering staff in consultation with Sargent & Lundy. As indicated in this table, the assumed SO2
removal rate for DSI + fabric filter is 50%. The retrofit DSI option on an existing unit with existing ESP is
always provided in combination with a fabric filter (Toxecon configuration).
Methodology for Obtaining DSI Control Costs: The cost and performance model for DSI was updated by
Sargent & Lundy. The model is used to derive the cost of DSI retrofits with two alternative, associated
particulate control devices, i.e., ESP and fabric filter "baghouse". The cost model notes that the cost
drivers of DSI are quite different from those of wet or dry FGD. Whereas plant size and coal sulfur rates
are key underlying determinants of FGD cost, sorbent feed rate and fly ash waste handling are the main
drivers of the capital cost of DSI with plant size and coal sulfur rates playing a secondary role.
Furthermore, the DSI sorbent feed rate and variable O&M costs are based on assumptions that a fabric
filter and in-line Trona milling are used, and that the SO2 removal rate is 50%. The corresponding HCI
removal effect is estimated to be 98% for units with fabric filter.
5-20
-------
The cost of fly ash waste handling, the other key contributorto DSI cost, is a function of the type of
particulate capture device and the flue gas SO2.
Total waste production involves the production of both reacted and unreacted sorbent and fly ash.
Sorbent waste is a function of the sorbent feed rate with an adjustment for excess sorbent feed. Use of
sodium-based DSI may make the fly ash unsalable, which would mean that any fly ash produced must be
landfilled along with the reacted and unreacted sorbent waste. Typical ash contents for each fuel are
used to calculate a total fly ash production rate. The fly ash production is added to the sorbent waste to
account for the total waste stream for the VOM analysis.
For purposes of modeling, the total VOM includes the first two component costs noted in the previous
paragraph, i.e., the costs for sorbent usage and the costs associated with waste production and disposal.
Table 5-16 presents the capital, VOM, and FOM costs as well as the capacity and heat rate penalties of a
DSI retrofit for an illustrative and representative set of generating units with the capacities and heat rates
indicated. For details of Sargent & Lundy DSI cost model, see Attachment 5-5.
5.6 Fabric Filter (Baghouse) Cost Development
Fabric filters are not endogenously modeled as a separate retrofit option. In EPA Platform v6, an existing
or new fabric filter particulate control device is a pre-condition for installing a DSI retrofit, and the cost of
these retrofits at plants without an existing fabric filter include the cost of installing a new fabric filter. This
cost was added to the DSI costs discussed in section 5.5. The costs associated with a new fabric filter
retrofit are derived from the cost and performance updated by Sargent & Lundy. Similarly, dry scrubber
retrofit costs also include the cost of a fabric filter.
The engineering cost analysis is based on a pulse-jet fabric filter which collects particulate matter on a
fabric bag and uses air pulses to dislodge the particulate from the bag surface and collect it in hoppers for
removal via an ash handling system to a silo. This is a mature technology that has been operating
commercially for more than 25 years. "Baghouse" and "fabric filters" are used interchangeably to refer to
such installations.
Capital Cost: The major driver of fabric filter capital cost is the "air-to-cloth" (A/C) ratio. The A/C ratio is
defined as the volumetric flow, (typically expressed in Actual Cubic Feet per Minute, ACFM) of flue gas
entering the baghouse divided by the areas (typically in square feet) of fabric filter cloth in the baghouse.
The lower the A/C ratio, e.g., A/C = 4.0 compared to A/C = 6.0, the greater the area of the cloth required
and the higher the cost for a given volumetric flow. An air-to-cloth ratio of 4.0 is used in EPA Platform v6,
and it is assumed that the existing ESP remains in place and active.
Table 5-17 presents the capital, VOM, and FOM costs for fabric filters as represented in EPA Platform v6
for an illustrative set of generating units with a representative range of capacities and heat rates. See
Attachment 5-7 for details of the Sargent & Lundy fabric filter PM control cost model.
5-21
-------
Table 5-16 Illustrative Dry Sorbent Injection (DSI) Costs (2016$) for Representative Sizes and Heat Rates
under Assumptions in EPA Platform v6
Control
Type
Heat Rate
(Btu/kWh)
so2
Rate
(lb/
MMBtu)
Capacity
Penalty
(%)
Heat
Rate
Penalty
(%)
Variable
O&M
(mills/kWh)
Capacity (MW)
100
300
500
700
1000
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
DSI
9,000
2.0
-0.37
0.37
5.86
124.2
3.57
56.6
1.31
39.2
0.83
30.8
0.62
23.9
0.45
Assuming
10,000
2.0
-0.41
0.41
6.51
128.0
3.60
58.3
1.33
40.4
0.84
31.8
0.62
24.6
0.46
Bituminous
Coal
11,000
2.0
-0.45
0.45
7.16
131.5
3.63
59.9
1.34
41.5
0.85
32.6
0.63
25.3
0.46
Note 1: A S02 removal efficiency of 50% is assumed in the above calculations.
Note: The Variable O&M costs in this table do not include the cost of additional auxiliary power (VOMP) component in the S&L tool as for modeling purposes, IPM reflects the auxiliary
power consumption through capacity penalty.
Table 5-17 Illustrative Particulate Controls Costs (2016$) for Representative Sizes and Heat Rates
under the Assumptions in EPA Platform v6
Coal Type
Heat Rate
(Btu/kWh)
Capacity
Penalty
(%)
Heat
Rate
Penalty
(%)
Variable
O&M
(mills/kWh)
Capacity (MW)
100
300
500
700
1000
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-
yr)
9,000
0.05
254
0.9
206
0.7
187
0.7
175
0.6
164
0.6
Bituminous
10,000
-0.60
0.60
0.06
276
1.0
224
0.8
203
0.7
191
0.7
178
0.6
11,000
0.07
298
1.0
242
0.8
220
0.8
206
0.7
193
0.7
Note: The Variable O&M costs in this table do not include the cost of additional auxiliary power (VOMP) component in the S&L tool as for modeling purposes, IPM reflects the auxiliary
power consumption through capacity penalty.
5-22
-------
5.7 Coal-to-Gas Conversions47
In EPA Platform v6, existing coal plants are given the option to burn natural gas in addition to coal by
investing in a coal-to-gas retrofit. There are two components of cost in this option: boiler modification
costs and the cost of extending natural gas lateral pipeline spurs from the boiler to a natural gas main
pipeline. These two components of cost and their associated performance implications are discussed in
the following sections.
5.7.1 Boiler Modifications for Coal-To-Gas Conversions
Enabling natural gas firing in a coal boiler typically involves installation of new gas burners and
modifications to the ducting, windbox (i.e., the chamber surrounding a burner through which pressurized
air is supplied for fuel combustion), and possibly to the heating surfaces used to transfer energy from the
exiting hot flue gas to steam (referred to as the "convection pass"). It may also involve modification of
environmental equipment. Engineering studies are performed to assess operating characteristics like
furnace heat absorption and exit gas temperature; material changes affecting piping and components like
superheaters, reheaters, economizers, and recirculating fans; and operational changes to sootblowers,
spray flows, air heaters, and emission controls.
The following table summarizes the cost and performance assumptions for coal-to-gas boiler modifications
as incorporated in EPA Platform v6. The values in the table were developed by EPA's engineering staff
based on technical papers48 and discussions with industry engineers familiar with such projects. They
were designed to be broadly applicable across the existing coal fleet (with the exceptions noted in the
table). Coal-to-gas retrofit options in EPA Platform v6 force a permanent change in fuel type from coal to
natural gas. Coal therefore can no longer be fired.
Table 5-18 Cost and Performance Assumptions for Coal-to-Gas Retrofits in EPA Platform v6
Factor
Description
Notes
Applicability:
Existing pulverized coal (PC) fired and
cyclone boiler units of a size greater
than 25 MW:
Not applicable forfluidized bed
combustion (FBC) and stoker boilers.
Capacity Penalty:
None
The furnace of a boiler designed to burn coal
is oversized for natural gas, and coal boilers
include equipment, such as coal mills, that are
not needed for gas. As a result, burning gas
should have no impact on net power output.
Heat Rate
Penalty:
+ 5%
When gas is combusted instead of coal, the
stack temperature is lower and the moisture
loss to stack is higher. This reduces
efficiency, which is reflected in an increase in
the heat rate.
Incremental
Capital Cost:
PC units: 2016$/kW =
288*(75/MW)a0.35
Cyclone units: 2016$/kW =
403*(75/MW)a0.35
The cost function covers new gas burners and
piping, windbox modifications, air heater
upgrades, gas recirculating fans, and control
system modifications.
Example for 50 MW PC unit:
$/kW = 288*(75/50)a0.35 = 332
47 As discussed here coal-to-gas conversion refers to the modification of an existing boiler to allow it to fire natural gas.
It does not refer to the addition of a gas turbine to an existing boiler cycle, the replacement of a coal boiler with a new
natural gas combined cycle plant, or to the gasification of coal for use in a natural gas combustion turbine
48 For an example see Babcock and Wilcox's White Paper MS-14 "Natural Gas Conversions of Exiting Coal-Fired
Boilers" 2010 (https://slidex.tips/download/natural-aas-conversions-of-existina-coal-fired-boilers).
5-23
-------
Factor
Description
Notes
Incremental
Fixed O&M:
-33% FOM cost of the existing coal unit
Due to reduced needs for operators,
maintenance materials, and maintenance staff
when natural gas combusted, FOM costs
decrease by 33%.
Incremental
Variable O&M:
-25% VOM cost of the existing coal unit
Due to reduced waste disposal and
miscellaneous other costs, VOM costs
decrease by 25%.
Fuel Cost:
Natural Gas
To obtain natural gas the unit incurs the cost
of extending lateral pipeline spurs from the
boiler to the local transmission mainline. See
section 5.7.2
NOx emission rate:
50% of existing coal unit NOx emission
rate, with a floor of
0.05 Ibs/MMBtu
The 0.05 Ibs/MMBtu floor is the same as the
NOx rate floor for new retrofit SCR on units
burning subbituminous coal
SO2 emissions:
Zero
5.7.2 Natural Gas Pipeline Requirements for Coal-To-Gas Conversions
For every individual coal boiler in the U.S., EPA had ICF to determine the miles and associated cost of
extending pipeline laterals from each boiler to the interstate natural gas pipeline system. This work was
performed for EPA Base Case v5.13 and has not been updated for EPA Platform v6. For further detail,
please see EPA Base Case v5.13 documentation.
Table 5-21 shows the pipeline costing results for each qualifying existing coal fired unit represented in
EPA Platform v6.
5.8 Retrofit Assignments
In IPM, model plants that represent existing generating units have the option of maintaining their current
system configuration, retrofitting with pollution controls, or retiring. The decision to retrofit or retire is
endogenous to IPM and based on the least cost approach to meeting demand subject to modeled system
and operational constraints. IPM is capable of modeling retrofits and retirements at each applicable
model unit at three different points in time, referred to as three stages. At each stage a retrofit set may
consist of a single retrofit (e.g. LSFO Scrubber) or pre-specified combinations of retrofits (e.g., ACI +
LSFO Scrubber +SCR). In EPA Platform v6, first stage retrofit options are provided to existing coal-
steam and oil/gas steam plants. These plants - as well as combined cycle plants, combustion turbines,
biomass and nuclear plants - are also given retirement as an option in stage one. Third stage retrofit
options are offered to coal-steam plants only. Table 5-19 presents the first stage retrofit options available
by plant type; Table 5-20 presents the second and third stage retrofit options available to coal-steam
plants. The cost of multiple retrofits on the same model plant, whether installed in one or multiple stages,
are additive. In linear programming models such as IPM, projections of pollution control equipment
capacity and retirements are limited to the pre-specified combinations listed in Table 5-19 and Table 5-20
below.
5-24
-------
Table 5-19 First Stage Retrofit Assignment Scheme in EPA Platform v6
Plant Type
Retrofit Option 1st Stage
Criteria
Coal Steam
Coal Retirement
All coal steam boilers
Coal Steam SCR
All coal steam boilers that are 25 MW or larger and do not possess an
existing SCR control option
Coal Steam SNCR - Non FBC
Boilers
All non FBC coal steam boilers that are 25 MW or larger and do not possess
an existing post combustion NOx control option
Coal Steam SNCR - FBC
Boilers
All coal FBC units that are 25 MW or larger and do not possess an existing
post combustion NOx control option
LSD Scrubber
All unscrubbed coal steam boilers 25 MW or larger and burning less than 3
Ibs/MMBtu SO2 coal
LSFO Scrubber
All unscrubbed and non FBC coal steam boilers 25
MW or larger
CO2 Capture and Storage
All scrubbed coal steam boilers 400 MW or larger
ACI - Hg Control Option
(with and without Toxecon)
All coal steam boilers larger than 25 MW that do not have an ACI and have
an Hg EMF greaterthan 0.1. Actual ACI technology type will be based on
the boilers fuel and technology configuration. See discussion in Chapter 5.
LSD Scrubber + SCR
Combination options - Individual technology level restrictions apply
LSD Scrubber + SNCR
LSFO Scrubber + SCR
LSFO Scrubber + SNCR
ACI + SCR
ACI + SNCR
ACI + LSD Scrubber
ACI + LSFO Scrubber
ACI + LSD Scrubber + SCR
ACI + LSFO Scrubber + SCR
ACI + LSD Scrubber + SNCR
ACI + LSFO Scrubber + SNCR
DSI
All unscrubbed and non FBC coal steam boilers 25 MW or larger with Fabric
Filter and burning less than 2 Ibs/MMBtu SO2 coal.
DSI + Fabric Filter
All unscrubbed and non FBC-coal steam boilers 25 MW or larger without
Fabric Filter, with CESP or HESP, and burning less than 2 Ibs/MMBtu SO2
coal.
DSI + SCR
Combination options - Individual technology level restrictions apply
DSI + SNCR
ACI + DSI
ACI + DSI + SCR
ACI + DSI + SNCR
Heat Rate Improvement
All coal steam boilers with a heat rate larger than 9,500 Btu/kWh
Coal-to-Gas
All coal steam boilers that are 25 MW or larger
Integrated Gasification Combined Cycle
IGCC Retirement
All integrated gasification combined cycle units
5-25
-------
Plant Type
Retrofit Option 1st Stage
Criteria
Combined Cycle
CC Retirement
All combined cycle units
Combustion Turbine
CT Retirement
All combustion turbine units
Nuclear
Nuclear Retirement
All nuclear power units
Oil and Gas Steam
Oil/Gas Retirement
All oil/gas steam boilers
Oil/Gas Steam SCR
All oil/gas steam boilers 25 MW or larger that do not possess an existing post
combustion NOx control option
Table 5-20 Second and Third Stage Retrofit Assignment Scheme in EPA Platform v6
Plant Type
Retrofit Option 1st Stage
Retrofit Option 2nd Stage
Retrofit Option 3rd Stage
Coal Steam
SO2 Control Option
Heat Rate Improvement
HCI Control Option
Heat Rate Improvement
NOx Control Option1
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
NOx Control Option
Heat Rate Improvement
SO2 Control Option2
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
NOx Control Option
Heat Rate Improvement
SO2 Control Option
Heat Rate Improvement
Hg Control Option3
HCI Control Option
Heat Rate Improvement
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
CO2 Control Option4
None
None
NOx Control Option1 + SO2
Control Option2
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
SO2 Control Option
Heat Rate Improvement
HCI Control Option
Heat Rate Improvement
NOx Control Option1 + Hg
Control Option3
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
SO2 Control Option2 + Hg
NOx Control Option
Heat Rate Improvement
5-26
-------
Plant Type
Retrofit Option 1st Stage
Retrofit Option 2nd Stage
Retrofit Option 3rd Stage
Control Option3
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
NOx Control Option1 + SO2
Control Option2 + Hg Control
Option3
CO2 Control Option
None
Heat Rate Improvement
CO2 Control Option
Coal Retirement
None
NOx Control Option
Heat Rate Improvement
HCI Control Option5
SO2 Control Option
Heat Rate Improvement
Heat Rate Improvement
None
Coal Retirement
None
NOx Control Option1 + HCI
Control Option5
SO2 Control Option
Heat Rate Improvement
Heat Rate Improvement
None
Coal Retirement
None
NOx Control Option
Heat Rate Improvement
Hg Control Option3 + HCI
SO2 Control Option
Heat Rate Improvement
Control Option5
Heat Rate Improvement
None
Coal Retirement
None
NOx Control Option1 + HCI
SO2 Control Option
Heat Rate Improvement
Control Option5 + Hg Control
Heat Rate Improvement
None
Option3
Coal Retirement
None
NOx Control Option
None
SO2 Control Option
None
Heat Rate Improvement
HCI Control Option
None
CO2 Control Option
None
Coal Retirement
None
Coal-to-Gas
NOx Control Option
None
Oil/Gas Retirement
None
Coal Retirement
None
None
Oil and Gas Steam
NOx Control Option1
Oil/Gas Retirement
None
Oil/Gas Retirement
None
None
Notes:
1"NOx Control Option" implies that a model plant may be retrofitted with one of the following NOx control technologies: SCR,
SNCR - non-FBC, or SNCR - FBC
2"S02 Control Option" implies that a model plant may be retrofitted with one of the following SO2 control technologies: LSFO
scrubber or LSD scrubber
3"Hg Control Option" implies that a model plant may be retrofitted with one of the following activated carbon injection
technology options for reduction of mercury emissions: ACI or ACI + Toxecon
4"C02 Control Option" implies that a model plant may be retrofitted with carbon capture
and storage technology
5"HCI Control Option" implies that a model plant may be retrofitted with a DSI (with milled
Trona)
5-27
-------
List of tables and attachments that are directly uploaded to the web:
Attachment 5-1 Wet FGD Cost Methodology
Attachment 5-2 SDA FGD Cost Methodology
Attachment 5-3 SCR Cost Methodology
Attachment 5-4 SNCR Cost Methodology
Attachment 5-5 DSI Cost Methodology
Attachment 5-6 Hg Cost Methodology
Attachment 5-7 PM Cost Methodology
Table 5-21 Cost of Building Pipelines to Coal Plants in EPA Platform v6
5-28
-------
6. C02 Capture, Storage, and Transport
6.1 CO2 Capture
The EPA Platform v6 using IPM can build Ultra-Supercritical (USC) coal and Natural Gas Combined
Cycle (NGCC) Electric Generating Units (EGUs) with Carbon Capture49 and Storage (CCS) technology.
In addition, IPM includes a retrofit option to add CCS technology to existing coal steam and NGCC
EGUs.
6.1.1 CO2 Capture for Potential EGUs
Carbon capture for potential USC EGUs is represented as two model plant options with different CO2
capture efficiencies of 30 percent and 90 percent. EPA Platform v6 can offer CCS with a CO2 capture
efficiency of 90 percent for new NGCC units.50 The USC with CCS and NGCC with CCS model plant
options are configured assuming construction at greenfield sites. The cost and performance data
provided in Table 6-1 is based on the Annual Energy Outlook 2017 (AEO 2017). The basis for these
costs are studies prepared for the U.S. Department of Energy's (DoE's) Energy Information
Administration (EIA).5152
The USC costs were developed for a generic 650-megawatt (MW) net output USC EGU with a nominal
heat rate of 8,609 British Thermal Units (Btus) per kilowatt-hour (kW-hr) in 2021. The USC EGU uses a
"one-on-one" configuration. That is, the EGU is comprised of one pulverized coal (PC) steam generator
and one steam turbine (ST). The steam generator is fired with Illinois No. 6 (Herrin seam, Old Ben
Mine) bituminous coal and operates at steam conditions of 3,800 pounds per square inch-absolute
(psia) and 1,112 degrees Fahrenheit (°F). USC with a CCS is equipped with an amine-based, post-
combustion CO2 capture system. Mercury (Hg), sulfur oxides (SOx), nitrogen oxides (NOx), and
particulate matter (PM) emissions from the USC EGU are controlled with state-of-the-art air pollution
control equipment including Dry Sorbent Injection (DSI), Activated Carbon Injection (ACI), Wet Flue
Gas Desulfurization (WFGD) scrubber; low NOx burners (LNBs), Selective Catalytic Reduction (SCR),
and a fabric filter baghouse.
49 The term "carbon capture" refers primarily to removing carbon dioxide (CO2) from the flue gases emitted by fossil
fuel-fired EGUs.
50 Note that the NGCC with CCS option is disabled in the EPA Platform v6 November 2018 Reference Case.
51 Energy Information Administration (EIA). "Capital Cost Estimates for Utility Scale Electricity Generating Plants"
(November 2016). "Cost and Performance Characteristics of New Generating Technologies, Annual Energy Outlook
2077"(January 2017). "Addendum: Capital Cost Estimates for Utility Scale Electricity Generating Plants" (April 2017).
52 Note that the science of thermodynamics only refers to subcritical and supercritical states. "Ultra-Supercritical" is
an industry term that refers to operating at higher temperatures and/or pressures within the supercritical regime.
Distinct liquid and gas phases do not exist in a substance at a temperature and pressure above its critical point.
6-1
-------
Table 6-1 Cost and Performance Assumptions for Potential USC and NGCC with and without
Carbon Capture53
Advanced
Combined
Cycle
Advanced
Combined
Cycle with
CCS
Ultrasupercritical
Coal with 30%
CCS
Ultrasupercritical
Coal with 90%
CCS
Ultrasupercritical
Coal without CCS
Vintage #1 (2021)
Heat Rate (Btu/kWh)
6,267
7,514
9,644
11,171
8,609
Capital (2016$/kW)
1,081
2,104
4,953
5,477
3,580
Fixed O&M (2016$/kW/yr)
9.9
33.2
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2
7.1
7.1
9.5
4.6
Vintage #2 (2023)
Heat Rate (Btu/kWh)
6,233
7,504
9,433
10,214
8,514
Capital (2016$/kW)
1,064
2,059
4,863
5,378
3,516
Fixed O&M (2016$/kW/yr)
9.9
33.2
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2
7.1
7.1
9.5
4.6
Vintage
#3 (2025-2054)
Heat Rate (Btu/kWh)
6,200
7,493
9,221
9,257
8,323
Capital (2016$/kW)
1,041
2,003
4,746
5,249
3,431
Fixed O&M (2016$/kW/yr)
9.9
33.2
69.6
80.8
42.1
Variable O&M (2016$/MWh)
2
7.1
7.1
9.5
4.6
The NGCC costs were developed for a generic 702-MW net output NGCC EGU with a nominal heat
rate of 6,267 Btus per kW-hr in 2021. The USC EGU uses a "two-on-two-on-one" configuration. That
is, the combined cycle technology EGU is comprised of two natural gas-fired F5-class combustion
turbines (CTs), two supplementary Heat Recovery Steam Generators (HRSGs), and one ST. The
NGCC facility is fueled with pipeline-quality natural gas with a higher heating value (HHV) of 1,040 Btus
per standard cubic foot (scf). Steam is produced at 3,800 psia and 1,112 oF. The two HRSGs extract
heat from the two CTs to power the one ST. NGCC with a CCS is equipped with an amine-based,
post-combustion CO2 capture system. NOx emissions from the NGCC EGU are controlled with LNBs
and a SCR system.
6.1.2 CO2 Capture via Retrofitting Existing EGUs
EPA Platform v6 offers the option of retrofitting CCS to existing coal-fired power plants and NGCCs at a
CO2 capture efficiency of 90 percent.54 The CO2 capture process is modeled assuming the use of an
amine-based, post-combustion CO2 capture system.
The cost and performance data provided in Table 6-2 is based on the Sargent & Lundy55 cost algorithm
(Attachment 6-1 summarizes this study) and a DoE/National Environmental Technology Laboratory (NETL)
53 The cost and performance characteristics for these new units are also shown in Table 4-13 and discussed further in
Chapter 4.
54 Note that the NGCC with CCS option is disabled in the EPA Platform v6 November 2018 Reference Case.
55 Sargent & Lundy. "IPM Model - Updates to Cost and Performance forAPC Technologies - CO2 Reduction Cost
Development Methodology." Project 13527-001; February 2017.
6-2
-------
study.56 As part of developing documentation for EPA Platform v6, the capital costs were converted to
2016 dollars from the 2011 dollar basis used in the referenced DoE/NETL study. Note that one of the
carbon capture information resources is the Shell Cansolv® technology, which was installed on Unit 357
at SaskPower's Boundary Dam Power Station near Estevan, Saskatchewan, Canada in October 2014.58
One issue that must be addressed when installing an amine-based, post-combustion CO2 capture
system is that sulfur oxides (e.g., SO2 and sulfur trioxide (SO3)) in the EGU flue gas can degrade the
amine-based solvent used to absorb the CO2 from the EGU flue gas. Since the amine will preferentially
absorb SO2 before CO2, it will be necessary to treat the EGU flue gas to lower the sulfur oxide
concentration to 10 parts per million by volume (ppmv) or less. Meeting this constraint will require
installing supplemental WFGD technology (e.g., the SO2 "polishing" scrubber referenced in footnote 58),
or retrofitting existing FGD.
Table 6-2 Performance and Unit Cost Assumptions for Carbon Capture Retrofits on Coal Plants
Capacity (MW)
Heat Rate
(Btu/kWh)
Capital
Cost
($/kW)
Fixed
O&M
($/kW-yr)
Variable
O&M
(mills/kWh)2
Capacity
Penalty
(%)
Heat Rate
Penalty (%)
9,000
2,595
36.9
3.15
33.6
50.6
400
10,000
2,960
41.2
3.71
37.3
59.5
11,000
3,373
46.1
4.32
41.0
69.6
9,000
1,852
23.7
2.57
19.2
23.7
700
10,000
2,071
26.1
2.93
21.3
27.0
11,000
2,302
28.6
3.31
23.4
30.6
9,000
1,625
19.7
2.40
13.4
15.5
1,000
10,000
1,810
21.6
2.71
14.9
17.5
11,000
2,001
23.6
3.03
16.4
19.6
Note:
11ncremental costs are applied to the derated (after retrofit) MW size.
2The CO2 Transportation, Storage, and Monitoring portion of the variable O&M has been
removed from Sargent & Lundy cost method and modeled separately.
The capacity-derating penalty and associated heat rate penalty are an output of the Sargent & Lundy model
(see section 5.1.1 for further details in regards to these penalties.)
6.2 CO2 Storage
The capacity and cost assumptions for CO2 storage in EPA Platform v6 are based on the
Geosequestration Cost Analysis Tool (GeoCAT); a spreadsheet model developed for the U.S. EPA by ICF,
Inc. (ICF) in support of the U.S. EPA's Underground Injection Control (UIC) Program forC02 Geologic
Storage Wells.59 For EPA Platform v6, ICF updated the major cost components in the GeoCAT model,
including revising onshore and offshore injection and monitoring costs to reflect 2016 industry drilling
56 DoE/NETL. "Cost and Performance Baseline for Fossil Energy Plants Volume 1a: Bituminous Coal (PC) and
Natural Gas to Electricity. Revision 3." DoE/NETL-2015/1723. July 6, 2015. (See
https://vwvw.netl.doe.aov/proiects/files/CostandPerformanceBaselineforFossilEneravPlantsVolume1aBitCoalPCandN
aturalGastoElectRev3 070615.pdf)
57 At the time of project execution, Sask Power's Boundary Dam Unit 3 was a 43-year old lignite-fired 139 MW net
generating unit. Upon completion, Boundary Dam Unit 3 became the first utility-scale power plant retrofitted with
CCS technology. Sask Power estimates that the $1.2 billion project extended Unit 3's life by 30 years. Note that the
associated energy penalty for installing the CCS technology derated Unit 3 from 139 to 110 MWs.
58 The Shell Cansolv® carbon capture system at Boundary Dam Unit 3 uses a proprietary amine solvent to absorb
SO2 and CO2 from the EGU flue gases. The carbon capture process requires very low SO2 levels in the flue gases
prior to CO2 capture because, if present, the amine would preferentially absorb SO2 before CO2. The Shell Cansolv®
SO2 capture process was installed upstream of the CO2 scrubber to "polish" the feed to the CO2 scrubber.
59 Federal Requirements Under the UIC Program for CO2 Geologic Sequestration Wells, Federal Register, December
10, 2010 (Volume 75, Number 237), pages 77229-77303.
6-3
-------
costs.60 All cost components in the model were also converted to a 2016 dollar basis. In addition to
updating costs in the model, ICF updated storage capacity, well injectivity, and other assumptions by
state and offshore area primarily using data from the research program conducted at DoE/NETL.
Assumptions for the amount of carbon dioxide injected for EOR was updated using the past several years
of performance data.
The GeoCAT model combines detailed characteristics of sequestration capacity by state and geologic
setting for the U.S. with costing algorithms for individual components of CO2 geologic sequestration.
The model outputs are regional sequestration cost curves that indicate how much potential storage
capacity is available at different CO2 storage cost points.
The GeoCAT model includes three modules:
1. A unit cost specification module,
2. A project scenario costing module, and
3. A geologic and regional cost curve module.
The unit cost specification module includes data and assumptions for 120 cost elements falling within the
following categories:
1. Geologic site characterization
2. Area of review and corrective action (including fluid flow and reservoir modeling during and
after injection and identification, evaluation, and remediation of existing wells within the area of
review)
3. Injection well and other facilities construction
4. Well operation
5. Monitoring the movement of CO2 in the subsurface
6. Mechanical integrity testing
7. Financial responsibility (to maintain sufficient resources for activities related to closing and
remediation of the site)
8. Post injection site care
9. Site closure
10. General and administrative
Of the ten cost categories for geologic CO2 sequestration listed above, the largest cost drivers (in
roughly descending order of magnitude) are well operation, injection well and other facilities construction,
and monitoring the movement of CO2 in the subsurface. The cost estimates are consistent with the
requirements for geologic storage facilities under the UIC Class VI rule61 and Greenhouse Gas (GhG)
Reporting Program Subpart RR62. The price of oil assumed for the calculation of EOR economics is
$75/barrel.
60 The major data sources for updating costs was the Bureau of Labor Statistics (BLS) Producers Price Index (PPI)
for various products and services related to oil and gas well drilling (https://www.bls.gov/ppi/). the "Joint Association
Survey of Drilling Costs" published by the American Petroleum Institute (http://www.api.org/products-and-
services/statistics#tab overview), and the "Well Cost Study" published by the Petroleum Services Association of
Canada (https://www.psac.ca/resources/well-cost-studv-overview/).
61 Supra Note 59.
62 Title 40 of the Code of Federal Regulations (CFR), Part 98 (Mandatory GhG Reporting), Subpart RR (Geologic
Sequestration of CO2). See https://ecfr.io/Title-40/sp40.23.98.rr.
6-4
-------
The costs derived in the unit cost specification module are used in the GeoCAT project scenario
costing module to develop commercial scale costs for eight sequestration scenarios compliant with UIC
Class VI standards and GhG Reporting Program Subpart RR:
1.
Deep saline formations
2.
Depleted gas fields
3.
Depleted oil fields
4.
Enhanced oil recovery
5.
Enhanced coal bed methane recovery
6.
Enhanced shale gas
7.
Basalt storage
8.
Unmineable coal seams
EPA's GeoCAT application for CO2 sequestration includes only storage capacity for the first four
sequestration scenarios. The last four reservoir types are not included because they are not considered
technically mature enough to allow CO2 storage in the foreseeable future.
The current GeoCAT model includes the most recent DoE analysis of the lower-48 states CO2
sequestration capacities from the "Carbon Sequestration Atlas of the United States and Canada
Version 5."63 ICF enhanced these assessments to include additional details needed for economic
modeling such as the distribution of capacity by state, drilling depth, injectivity, etc. The geologic and
regional cost curve module applies regionalized unit cost factors to these geologic characterizations to
develop regional geologic storage cost curves.64 The analysis of storage volumes is carried out by
regional carbon sequestration partnerships as overseen by NETL in Morgantown, West Virginia. State
level onshore and offshore capacity volumes are reported for storage in oil and gas reservoirs and
deep saline formations. The great majority of storage volume is in deep saline formations, which are
present in many states and in most states with oil and gas production. In the most recent version of the
Atlas, offshore storage volumes have also been broken out by DoE into the Gulf of Mexico, Atlantic,
and Pacific Outer Continental Shelf (OCS) regions. ICF carried out a separate analysis to break out
CO2 EOR storage potential from the total potential in oil and gas reservoirs reported in NATCARB.
Efficiency Assumptions for EOR Uses of CO2
Relying on recent performance data, the geologic storage cost curve for EOR is based on an average
EOR efficiency of 10 thousand cubic feet (Mcf) of CO2 per incremental barrel of crude oil. The NETL
"CO2 EOR Primer"65 shows that from the start of CO2 floods in 1972 to 2008 the average efficiency was
7.66 Mcf per bbl. Data for the most recent seven year has shown a lower average efficiency of over 10.32
Mcf/bbl. This creates an average of 8.62 Mcf/bbl for all years from 1972 to 2016.
63 Carbon Sequestration Atlas of the United States and Canada - Version 5 (2015), U.S. Department of Energy,
National Energy Technology Laboratory, Morgantown, WV https://www.netl.doe.gov/research/coal/carbon-
storaqe/atlasv. Accessed mid-October 2016 with data updates through 2015.
64 Detailed discussions of the GeoCAT model and its application for EPA can be found in U.S. Environmental
Protection Agency, Office of Water, "Geologic CO2 Sequestration Technology and Cost Analysis, Technical Support
Document" (EPA 816-B-08-009) June 2008, https://www.epa.gov/sites/production/files/2015-
07/documents/support uic co2 technoloavandcostanalvsis.pdf and Harry Vidas, Robert Hugman and Christa Clapp,
"Analysis of Geologic Sequestration Costs for the United States and Implications for Climate Change Mitigation,"
Science Digest, Energy Procedia, Volume 1, Issue 1, February 2009, Pages 4281-4288. Available online at
https://www.sciencedirect.com/science/article/pii/S18766102090Q8832.
65 National Energy Technology Laboratory, "Carbon Dioxide Enhanced Oil Recovery", 2010,
https://www.netl.doe.gov/file%20library/research/oil-gas/C02_EOR_Primer.pdf
6-5
-------
Historical CO2 EOR: 1972-2008
Billion cubic feet of CO2
11,000
Million barrels of crude oil
1,437
Mcf/barrel
7.66
Source: NETL, "Carbon Dioxide Enhanced Oil
Recovery", 2010
Historical CO2 EOR: 2009-2016
Billion cubic feet of CO2
8,339
Million barrels of crude oil
808
Mcf/barrel
10.32
Source: ICF estimates based on EPA GHG Inventory
and Oil & Gas Journal Annual EOR Survey
Historical CO2 EOR: 1972-2016
Billion cubic feet of CO2
19,339
Million barrels of crude oil
2,244
Mcf/barrel
8.62
Source: Sum of prior two tables
The average of all historical and ongoing EOR projects through the end of their lifetimes is likely to
exceed 9.0 Mcf/bbl as they continue to operate at ratios above 10 Mcf/bbl.66 ICF has chosen a calibration
point of 10 Mcf/bbl for the average of potential future CO2 EOR under the belief that the quality of future
projects would likely be worse (i.e., require more CO2 per unit of incremental oil production) than historical
projects. The historical projects were presumably chosen, in part, because of their favorable
characteristics, as well as their proximity to sources of CO2. Future projects are likely to have poorer
characteristics and would be expected to require more CO2. Also, because of the economic incentive to
get the 45Q tax credits, it is possible that project designs might change to use more CO2, particularly in
the years the credit is available. The revised average efficiency value of 10 Mcf/bbl is approximately 15
percent higher than the original version of GeoCAT, which was calibrated to the older historical data.
There is considerable variation in CO2 requirements among projects initiated in the US. The most efficient
projects use less than 3 Mcf of net CO2 per barrel produced, while the least efficient can use up to 20
Mcf/bbl.67 This wide variation is represented in GeoCAT where assumed efficiencies across all state and
geologic reservoir qualities is approximately 5 to 25 Mcf/bbl. There are 19 Mcf of CO2 per metric ton, so
in terms of weight, GeoCAT has a range of 0.26 to 1.32 tons of carbon dioxide per barrel of increment
crude oil produced by CO2 flooding. The average is 0.53 metric tons CO2 per incremental barrel.
66 For example, assuming an average of 10 years of future operation at the 2016 ratios leads to a lifetime average for
all historical and ongoing CO2 EOR project of 9.09 Mcf/bbl.
67 During the EOR operations, some CO2 is produced with the oil and then separated and recycled back into the
reservoir. "Net volume" of CO2 is the original amount of CO2 delivered to the EOR project, while "gross injected
volume" includes both original volumes and reinjected volumes. The gross volumes are typically about twice the net
volumes.
6-6
-------
The results of the project scenario costing module are taken as inputs into the geologic and regional
cost curve module of GeoCAT, which generates national and regional cost curves indicating the volume
of sequestration capacity in each region and state in the U.S. as a function of total cost per ton of CO2
including all capital and operating costs. The result is a database of sequestration capacity by state,
geologic reservoir type, and cost step.
Table 6-3 shows the NATCARB V storage volumes for the U.S. Lower-48 as allocated to GeoCAT
categories. Total Lower-48 capacity is assessed at 8,216 gigatonnes. There are no volumes in the
current model for potential storage in depleted gas field reservoirs because these are not reported in
NATCARB.
For EPA Platform v6, GeoCAT represents storage opportunities in 37 of the lower 48 continental
states.68 Louisiana and Texas have both onshore and offshore state-level storage cost curves. In
addition, because NATCARB does not provide state-level data, there are multi-state Atlantic offshore and
Pacific offshore storage cost curves. The result is 41 storage cost curves shown in Table 6-4.
Table 6-3 Lower-48 CO2 Sequestration Capacity by Region
Offshore Allocation in GeoCAT
Onshore
Offshore
Total
Louisiana
Texas
GOM Total
Pacific
Atlantic
Total
C02 Enhanced Oil Recovery
Low
11.2
1.1
12.3
Mid
15.0
1.5
16.4
1.5
0.0
1.5
0.0
0.0
1.5
High
22.5
2.2
24.7
Depleted Oil
Low
128.0
11.8
139.8
Mid
170.7
15.7
186.4
12.7
3.0
15.7
0.1
0.0
15.7
High
256.0
23.6
279.6
Unmineable Coal
Low
47.8
2.0
49.8
Mid
63.7
2.6
66.4
0.0
0.0
0.0
2.6
0.0
2.6
High
95.6
4.0
99.5
Saline
Low
4,252
1,708
5,960
Mid
5,669
2,277
7,947
1,240
798
2,038
37
202
2,277
High
12,477
3,416
15,893
Totals
Oil Subtotal
(EOR plus Depleted Oil Fids.)
Low
4,439
1,723
6,162
Mid
5,919
2,297
8,216
1,254 801
2,055 40
202
2,297
High
12,851
3,446
16,297
Low
139.2
12.9
152.1
Mid
185.6
17.2
202.8
14.16 2.97
17.13 0.05
0.00
17.18
High
278.5
25.8
304.2
may not sum to reported totals due to rounding.
The cost curves in Table 6-4 are in the form of step functions. In any given year within the IPM model, a
specified amount of storage is available at a particular step price until either the annual storage limit or
the total storage capacity is reached. In determining whether the total storage capacity has been
reached, the model tracks the cumulative storage used up through the current year. Once the
cumulative storage used equals the total storage capacity at that price step, no more storage is available
going forward at that particular step price and, so, higher priced steps must be used.
68 The states without identified storage opportunities in EPA Reference Case v6 are Connecticut, Iowa, Maine,
Massachusetts, Minnesota, Nevada, New Hampshire, New Jersey, Rhode Island, Vermont, and Wisconsin. These
states were either not assessed or were found to not have storage opportunities in NATCARB for the four
sequestration scenarios included in EPA's inventory, (i.e., deep saline formations, depleted gas fields, depleted oil
fields, and enhanced oil recovery).
6-7
-------
C02 storage opportunities are relevant not just to power sector sources, but also to sources in other
industrial sectors. Therefore, before being incorporated as a supply representation into EPA Platform
v6, the original C02 storage capacity in each storage region was reduced by an estimate of the
storage that would be occupied by C02 generated by other industrial sector sources at the relevant
level of cost effectiveness (represented by $/ton C02 storage cost).
To do this, ICF first estimated the level of industrial demand for C02 storage in each C02 storage
region in a scenario where the value of abating C02 emissions is assumed to be $50 per ton (this
abatement value is relevant not only to willingness to pay for storage but also for the cost of capture
and transportation of the abated CO2).69 The quantity of industrial sequestration economic at $50/ton
represent the "high quality" industrial sources that have high CO2 purity and would be easiest to
capture, rehydrate, and compress. They are made up of ethanol plants, hydrogen production at
refineries and merchant plants and gas processing plants where CO2 is removed from the natural gas.
This amount was calculated as 128 million tons per year.
Then, for each region, ICF calculated the ratio of the industrial demand to total storage capacity
available for a storage price of less than zero dollars per ton. An upper limit of $0.00 per ton was
chosen under the belief that the earliest uses of CO2 from industrial sources would continue the current
practice of targeting EOR opportunities. Converting this quantity of capacity reserved for industrial CCS to
a percent value and subtracting from 100 percent, ICF obtained the percent of storage capacity
available to the electricity sector at less than zero dollars per ton. Finally, the "Annual Step Bound
(MMTons)" and "Total Storage Capacity (MMTons)" was multiplied by this percentage value for each
step below zero dollars70 in the cost curves for the region to obtain the reduced storage capacity that
went into the storage cost curves for the electric sector in EPA Platform v6 November 2018 Reference
Case. Thus, the values shown in Table 6-4 represent the storage available specifically to the electric
sector.
The price steps in the Table 6-4 are the same from region to region. (That is, STEP9 [column 2] has a
step cost value of $9.07/Ton [column 3] across all storage regions [column 1], This across-region
price equivalency holds for every step.) However, the amount of storage available in any given year
(labeled "Annual Step Bound (MMTons)" in column 4) and the total storage available over all years
(labeled "Total Storage Capacity (MMTons)" in column 5) vary from region to region. In any given
region, the cost curves are the same for every run year, indicating that over the modeling time horizon
no new storage is being identified to augment the current storage capacity estimates. This
assumption is not meant to imply that no additional storage could be added. Such additional capacity
could be represented in the model if model runs exhaust key components of the currently estimated
storage capacity.
6.3 CO2 Transport
Each of the 64 IPM model regions can send C02 to the 41 regions represented by the storage cost curves
in Table 6-4. The associated transport costs (in 2016$/Ton) are shown in Table 6-5. For the model, ICF
has also updated assumptions about the costs of CO2 pipelines. These costs were derived by first
calculating the pipeline distance from each of the C02 Production Regions to each of the C02 Storage
Regions listed in Table 6-4. Since there are large economies of scale for pipelines, C02 transportation
costs depend on how many power plants and industrial C02 sources could share a pipeline over a given
69 The approach that ICF employed to estimate industrial demand forC02 storage is described in ICF International,
"Methodology and Results for Initial Forecast of Industrial CCS Volumes," January 2009.
70 Zero and negative cost steps represent storage available from enhanced oil recovery (EOR) where oil
producers either pay or offer free storage for CO2 that is injected into mature oil wells to enhance the amount of oil
recovered. The value of the CO2 for EOR is calculated using the average price of crude oil of $75/bbl taken from
ICF's Base Case forecast of mid-2018. There is also a market for CO2 injection in enhanced coal bed methane
(ECBM) production. ECBM is excluded from EPA's inventory as discussed earlier.
6-8
-------
distance. Consequently, the method assumes that the longer the distance from the source of the C02 to
the sink for the C02, the greater the chance for other sources to share in the transportation costs,
including pipeline costs (in $/inch-mile) and cost of service (in $/ton per 75 miles). These cost
components are functions of the required diameter and thickness of the pipeline and the flow capacity of
the pipeline, which themselves are functions of the assumed number of power plants using the pipeline.
CO2 transportation cost are based on a pipeline cost of $183,000 per inch-mile, which is consistent with
ICF's natural gas supply curve and basis differential assumptions from GMM. This pipeline cost estimate
is based on recent pipeline cost trends as represented in ICF's 2017 report to API entitled "U.S. Oil and
Gas Infrastructure Investment Through 2030."71
List of tables that are uploaded directly to the web:
Table 6-4 CO2 Storage Cost Curves in EPA Platform v6
Table 6-5 C02 Transportation Matrix in EPA Platform v6
Attachment 6-1 CO2 Reduction Cost Development Methodology
71 See: https://www.api.org/news-policy-and-issues/energy-infrastructure/oil-gas-infrastructure-study-2017
6-9
-------
7. Coal
The next three chapters cover the representation and underlying assumptions for fuels in EPA Platform
v6. The current chapter focuses on coal, chapter 8 on natural gas, and chapter 9 on other fuels (fuel oil,
biomass, nuclear fuel, and waste fuels) represented in EPA Platform v6.
This chapter presents four main topics. The first is a description of how the coal market is represented in
EPA Platform v6. This includes a discussion of coal supply and demand regions, coal quality
characteristics, and the assignment of coals to power plants.
The second topic is the coal supply curves which were developed for EPA Platform v6 and the bottom-up,
mine-based approach used to develop curves that would depict the coal choices and associated prices
that power plants will face over the modeling time horizon. Included are discussions of the methods and
data used to quantify the economically recoverable coal reserves, characterize their cost, and build the 81
coal supply curves that are implemented in EPA Platform v6. Illustrative examples are included of the
step-by-step approach employed in developing the supply curves.
The third topic is coal transportation. It includes a description of the transport network, the methodology
used to assign costs to the links in the network, and a discussion of the geographic, infrastructure, and
regulatory considerations that come into play in developing specific rail, barge and truck transport rates.
Finally, EPA addresses competing sources of supply and competing sources of demand. On the supply
side, this includes imported coal that arrives from non-U.S. or non-Canadian basins. On the demand
side, EPA addresses power plants competition for demand in the form of international thermal exports, as
well as domestic industrial/residential/commercial demand for thermal coal. These assumptions are
discussed in Section 7.4.
The assumptions for the coal supply curves and coal transportation were finalized in December 2016, and
were developed through a collaborative process with EPA supported by the following independent team
of coal experts (with key areas of responsibility noted in parenthesis): ICF (IPM model integration and
team coordination), Wood Mackenzie (coal supply curve development), and Hellerworx (coal
transportation).
7.1 Coal Market Representation in EPA Platform v6
Coal supply, coal demand, coal quality, and the assignment of specific types of coals to individual coal
fired generating units are the four key components of the endogenous coal market modeling framework in
EPA Platform v6. The modeling representation attempts to realistically reflect the actual options available
to each existing coal fired power plant while aggregating data sufficiently to keep the model size and
solution time within acceptable bounds.
Each coal-fired power plant modeled is reflected as its own coal demand region. The demand regions
are defined to reflect the coal transportation options (rail, barge, truck, conveyer belt) that are available to
the plant. These demand regions are interconnected by a transportation network to at least one of the 36
geographically dispersed coal supply regions. The model's supply-demand region links reflect actual on-
the-ground transportation pathways. Every coal supply region can produce and each coal demand region
can demand at least one grade of coal. Based on historical and engineering data (as described in
Section 7.1.5 below), each coal fired plant is also assigned several coal grades which it may use if that
coal type is available within its demand region.
In EPA Platform v6 the endogenous demand for coal is generated by coal fired power plants interacting
with a set of exogenous supply curves (see Table 7-26 for coal supply curve data) for each coal grade in
each supply region. The curves show the supply of coal (by coal supply region and coal grade) that is
available to meet the demand at a given price. The supply and demand for each grade of coal is linked to
and affected by the supply and demand for every other coal grade across supply and demand regions.
7-1
-------
The transportation network or matrix in Table 7-25 provides delivery cost to move coal from a free-on-
board point of sale in the coal basin to the end-use power plant. The transportation cost combined with
the free-on-board supply cost reflects the delivered cost a plant sees when making its coal selection. IPM
derives the equilibrium coal consumption and prices that result when the entire electric system is
operating at least cost while meeting emission constraints and other operating requirements over the
modeling time horizon.
7.1.1 Coal Supply Regions
There are 36 coal supply regions in EPA Platform v6, each representing geographic aggregations of coal-
mining areas that supply one or more coal grades. Coal supply regions may differ from one another in
the types and quality of coal they can supply. Table 7-1 lists the coal supply regions included in EPA
Platform v6.
Figure 7-1 provides a map showing the location of both the coal supply regions listed in Table 7-1 and the
broader supply basins commonly used when referring to U.S. coal reserves.
Table 7-1 Coal Supply Regions in EPA Platform v6
Region
State
Supply Region
Central Appalachia
Central Appalachia
Central Appalachia
Central Appalachia
Dakota Lignite
Dakota Lignite
East Interior
East Interior
East Interior
Gulf Lignite
Gulf Lignite
Gulf Lignite
Northern Appalachia
Northern Appalachia
Northern Appalachia
Northern Appalachia
Northern Appalachia
Rocky Mountains
Rocky Mountains
Rocky Mountains
Rocky Mountains
Southern Appalachia
Southwest
Southwest
West Interior
West Interior
West Interior
West Interior
Western Montana
Kentucky, East
Tennessee
Virginia
West Virginia, South
Montana, East
North Dakota
Indiana
Kentucky, West
Illinois
Texas
Louisiana
Mississippi
Maryland
Ohio
Pennsylvania, Central
Pennsylvania, West
West Virginia, North
Utah
Colorado, Green River
Colorado, Raton
Colorado, Uinta
Alabama
Arizona
New Mexico, San Juan
Arkansas, North
Kansas
Missouri
Oklahoma
Montana, Bull Mountains
WN
UT
KE
TN
VA
WS
ME
ND
IN
KW
IL
TX
LA
KS
MO
OK
OH
PC
PW
MS
MD
CG
CR
CU
NS
AN
AL
AZ
MT
7-2
-------
Region
State
Supply Region
Western Montana
Montana, Powder River
MP
Western Wyoming
Wyoming, Green River
WG
Wyoming Northern PRB
Wyoming, Powder River Basin (8800)
WH
Wyoming Southern PRB
Wyoming, Powder River Basin (8400)
WL
Alberta
Alberta
AB
British Columbia
British Columbia
BC
Saskatchewan
Saskatchewan
SK
Figure 7-1 Map of the Coal Supply Regions in EPA Platform v6
Coal Supply Region
Coal Supply Regions
7.1.2 Coal Demand Regions
Coal demand regions are designed to reflect coal transportation options available to power plants. Each
existing coal plant is reflected as its own individual demand region. The transportation infrastructure (i.e.,
rail, barge, or truck/conveyor belt), proximity to mine (i.e., mine mouth or not mine mouth), and
transportation competitiveness levels (i.e., non-competitive, low-cost competitive, or high-cost
competitive) are developed specific to each coal plant (demand region).
7-3
-------
When IPM is run, it determines the amount and type of new generation capacity to add within each of
IPM's 67 U.S. model regions. These model regions reflect the administrative, operational, and
transmission geographic structure of the electricity grid. Since these new plants could be located at
various locations within the region, a generic transportation cost for different coal types is developed for
these new plants and the methodology for deriving that cost is described in the transportation section of
this chapter. See Table 7-27 for the list of coal plant demand regions reflected in the transportation
matrix.
7.1.3 Coal Quality Characteristics
Coal varies by heat content, SO2 content, HCI content, and mercury content among other characteristics.
To capture differences in the sulfur and heat content of coal, a two letter "coal grade" nomenclature is
used. The first letter indicates the "coal rank" (bituminous, subbituminous, or lignite) with their associated
heat content ranges (as shown in Table 7-2). The second letter indicates their "sulfur grade," i.e., the SO2
ranges associated with a given type of coal. (The sulfur grades and associated SO2 ranges are shown in
Table 7-3).
Table 7-2 Coal Rank Heat Content Ranges
Coal Type
Heat Content (Btu/lb)
Classification
Bituminous
>10,260- 13,000
B
Subbituminous
> 7,500- 10,260
S
Lignite
less than 7,500
L
Table 7-3 Coal Grade SO2 Content Ranges
SO2 Grade
SO2 Content Range (Ibs/MMBtu)
A
0.00-0.80
B
0.81 - 1.20
D
1.21 - 1.66
E
1.67-3.34
G
3.35-5.00
H
> 5.00
The assumptions in EPA Platform v6 on the heat, HCI, mercury, SO2, and ash content of coal are derived
from EPA's "Information Collection Request for Electric Utility Steam Generating Unit Mercury Emissions
Information Collection Effort" (ICR)72.
A two-year effort initiated in 1998 and completed in 2000, the ICR had three main components: (1)
identifying all coal-fired units owned and operated by publicly-owned utility companies, Federal power
agencies, rural electric cooperatives, and investor-owned utility generating companies, (2) obtaining
"accurate information on the amount of mercury contained in the as-fired coal used by each electric utility
steam generating unit with a capacity greater than 25 megawatts electric, as well as accurate information
on the total amount of coal burned by each such unit,", and (3) obtaining data by coal sampling and stack
testing at selected units to characterize mercury reductions from representative unit configurations. Data
regarding the SO2, chlorine, and ash content of the coal used was obtained along with mercury content.
The ICR captured the origin of the coal burned, and thus provided a pathway for linking emission
properties to coal basins.
72 Data from the ICR can be found at http://www.epa.aov/ttn/atw/combust/utiltox/mercurv.html
7-4
-------
The 1998-2000 ICR resulted in more than 40,000 data points indicating the coal type, sulfur content,
mercury content, ash content, chlorine content, and other characteristics of coal burned at coal-fired utility
units greater than 25 MW.
Annual fuel characteristic and delivery data reported on EIA Form 923 also provide continual data points
on coal heat content, sulfur content, and geographic origin, which are used as a check against
characteristics initially identified through the ICR.
7.1.4 Coal Emission Factors
To make this data usable in EPA Platform v6, the ICR data points were first grouped by IPM coal grades
and IPM coal supply regions. Using the grouped ICR data, the average heat, SO2, mercury, HCI, and ash
content were calculated for each coal grade/supply region combination. In instances where no data were
available for a particular coal grade in a specific supply region, the national average SO2 and mercury
values for the coal grade were used as the region's values. The coal characteristics of Canadian coal
supply regions are based on the coal characteristics of the adjacent U.S. coal supply regions. The
resulting values are shown in Table 7-4. The CO2 values were derived from data in the Energy
Information Administration's Annual Energy Outlook 2016.
Table 7-4 Coal Quality Characteristics by Supply Region and Coal Grade
Coal
Coal
SO2 Content
Mercury
Ash Content
HCI Content
CO2 Content
Cluster
ouppiy
Region
Grade
(Ibs/MMBtu)
vonicni
(Ibs/Tbtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
Number
SA
0.59
5.29
5.47
0.009
215.5
1
AB
SB
0.94
6.06
6.94
0.013
215.5
5
SD
1.43
5.35
11.60
0.008
215.5
1
BB
1.09
4.18
9.76
0.012
204.7
4
AL
BD
1.35
7.28
10.83
0.029
204.7
1
BE
2.68
12.58
10.70
0.028
204.7
1
AN
BG
4.23
9.36
7.83
0.079
202.8
1
AZ
BB
1.05
5.27
7.86
0.067
207.1
2
BC
BD
1.40
6.98
8.34
0.096
216.1
4
BB
0.90
4.09
8.42
0.021
209.6
4
SB
0.93
2.03
7.06
0.007
212.8
1
CR
BB
1.05
5.27
7.86
0.067
209.6
2
CU
BB
0.86
4.01
7.83
0.009
209.6
4
BE
2.25
6.52
6.61
0.214
203.1
2
IL
BG
4.56
6.53
8.09
0.113
203.1
3
BH
5.58
5.43
9.06
0.103
203.1
1
BE
2.31
5.21
7.97
0.036
203.1
3
IN
BG
4.27
7.20
8.22
0.028
203.1
3
BH
6.15
7.11
8.63
0.019
203.1
3
BB
1.04
4.79
6.41
0.112
206.4
5
BD
1.44
5.97
7.45
0.087
206.4
2
r\b
BE
2.12
7.93
7.71
0.076
206.4
4
BG
3.79
11.99
10.21
0.041
206.4
4
KS
BG
4.84
4.09
8.47
0.133
202.8
5
BG
4.46
6.90
8.01
0.097
203.1
3
BH
5.73
8.16
10.21
0.053
203.1
3
LA
LE
2.49
7.32
17.15
0.014
212.6
1
7-5
-------
Coal
Supply
Region
Coal
SO2 Content
Mercury
Content
(Ibs/Tbtu)
Ash Content
HCI Content
CO2 Content
Cluster
Grade
(Ibs/MMBtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
Number
MD
BE
BG
2.78
3.58
15.62
16.64
11.70
16.60
0.072
0.018
204.7
204.7
5
7
ME
LE
1.83
11.33
11.69
0.019
219.3
2
MO
BG
4.54
5.91
9.46
0.023
215.5
3
SA
0.62
4.24
3.98
0.007
215.5
1
MP
SB
0.98
6.25
5.81
0.023
215.5
2
SD
1.49
4.53
10.13
0.006
215.5
1
MS
LE
2.76
12.44
21.51
0.018
216.5
3
MT
BB
1.05
5.27
7.86
0.067
215.5
2
ND
LE
2.27
8.30
12.85
0.014
219.3
1
SB
0.89
4.60
14.51
0.014
209.2
3
NS
SD
1.55
7.54
23.09
0.007
209.2
2
SE
1.90
8.65
23.97
0.008
209.2
1
BE
3.08
18.70
7.08
0.075
204.7
6
OH
BG
3.99
18.54
8.00
0.071
204.7
7
BH
6.43
13.93
9.13
0.058
204.7
4
OK
BG
4.65
26.07
13.54
0.051
202.8
6
BB
1.06
23.03
58.98
0.032
204.7
6
BD
1.42
21.67
49.31
0.066
204.7
3
PC
BE
2.57
17.95
9.23
0.096
204.7
6
BG
3.79
21.54
9.59
0.092
204.7
2
BH
6.29
34.71
13.89
0.148
204.7
5
BE
2.51
8.35
5.37
0.090
204.7
4
PW
BG
3.69
8.56
6.48
0.059
204.7
1
BH
7.78
16.46
11.56
0.046
204.7
2
SK
LD
1.51
7.53
11.57
0.014
219.3
1
LE
2.76
12.44
21.51
0.018
219.3
3
TN
BE
2.13
8.43
6.47
0.043
206.4
4
LE
3.00
14.65
25.65
0.020
212.6
4
TX
LG
3.91
14.88
25.51
0.036
212.6
1
LH
5.67
30.23
23.95
0.011
212.6
1
BA
0.67
4.37
7.39
0.015
209.6
1
UT
BB
0.94
3.93
8.58
0.016
209.6
4
BD
1.37
4.38
10.50
0.026
209.6
4
BE
2.34
9.22
7.41
0.095
209.6
4
BB
1.05
4.61
6.97
0.054
206.4
5
VA
BD
1.44
5.67
7.97
0.028
206.4
2
BE
2.09
8.40
8.05
0.028
206.4
4
BB
1.13
1.82
5.58
0.005
214.3
3
WG
SB
1.06
4.22
8.72
0.009
214.3
4
SD
1.33
4.33
10.02
0.008
214.3
1
SE
2.22
4.41
5.71
0.008
214.3
2
WH
SA
0.52
5.61
5.51
0.010
214.3
2
WL
SA
0.71
5.61
7.09
0.010
214.3
3
SB
0.93
6.44
7.92
0.012
214.3
5
7-6
-------
Coal
Supply
Region
Coal
SO2 Content
Mercury
Content
(Ibs/Tbtu)
Ash Content
HCI Content
CO2 Content
Cluster
Grade
(Ibs/MMBtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
(Ibs/MMBtu)
Number
BD
1.46
10.27
9.18
0.099
204.7
6
WN
BE
2.55
10.28
7.89
0.092
204.7
7
BG
4.00
9.27
6.92
0.074
204.7
1
BH
6.09
8.82
9.62
0.045
204.7
3
BB
1.09
5.75
9.15
0.091
206.4
1
WS
BD
BE
1.32
1.94
8.09
8.83
9.25
9.89
0.098
0.102
206.4
206.4
5
4
BG
4.67
7.13
6.39
0.051
206.4
3
Next, a clustering algorithm was used to further aggregate the data in EPA Platform v6 for model size
management purposes. The clustering analysis was performed on the SO2, mercury, and HCI data
shown in Table 7-4 using the SAS statistical software package. Clustering analysis places objects into
groups or clusters, such that data in a given cluster tend to be similar to each other and dissimilar to data
in other clusters. The clustering analysis involved two steps. First, the number of clusters of SO2,
mercury, and HCI concentrations for each IPM coal type was determined based on the range in SO2,
mercury, and HCI concentrations across all coal supply regions for a specific coal grade. Each coal type
used either one to seven clusters. The total number of clusters for each coal grade was limited to keep
the model size and run time within feasible limits. Second, for each coal grade the clustering procedure
was applied to all the regional SO2, mercury, and HCI values shown in Table 7-4 for that coal grade.
Using the SAS cluster procedure, each of the constituent regional values was assigned to a cluster and
the cluster average SO2, mercury, and HCI were estimated. The resulting values are shown in Table 7-5
through Table 7-9.
7-7
-------
Table 7-5 Coal Clustering by Coal Grade - S02 Emission Factors (Ibs/MMBtu)
Coal Type by Sulfur Grade
S02 Emission Factors (Ibs/MMBtu)
Cluster #1
Cluster #2
Cluster #3
Cluster #4
Cluster# 5
Cluster #6
Cluster# 7
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low Sulfur Bituminous (BA)
0.67
0.67
0.67
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Bituminous (BB)
1.09
1.09
1.09
1.05
1.05
1.05
1.13
1.13
1.13
0.95
0.86
1.09
1.04
1.04
1.05
1.06
1.06
1.06
-
-
-
Low Medium Sulfur Bituminous (BD)
1.35
1.35
1.35
1.44
1.44
1.44
1.42
1.42
1.42
1.39
1.37
1.40
1.32
1.32
1.32
1.46
1.46
1.46
-
-
-
Medium Sulfur Bituminous (BE)
2.68
2.68
2.68
2.25
2.25
2.25
2.31
2.31
2.31
2.19
1.94
2.51
2.78
2.78
2.78
2.82
2.57
3.08
2.55
2.55
2.55
High Sulfur Bituminous (BG)
3.98
3.69
4.23
3.79
3.79
3.79
4.50
4.27
4.67
3.79
3.79
3.79
4.84
4.84
4.84
4.65
4.65
4.65
3.78
3.58
3.99
High Sulfur Bituminous (BH)
5.58
5.58
5.58
7.78
7.78
7.78
5.99
5.73
6.15
6.43
6.43
6.43
6.29
6.29
6.29
-
-
-
-
-
-
Low Sulfur Subbituminous (SA)
0.60
0.59
0.62
0.52
0.52
0.52
0.71
0.71
0.71
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Subbituminous (SB)
0.93
0.93
0.93
0.98
0.98
0.98
0.89
0.89
0.89
1.06
1.06
1.06
0.94
0.93
0.94
-
-
-
-
-
-
Low Medium Sulfur Subbituminous (SD)
1.42
1.33
1.49
1.55
1.55
1.55
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Subbituminous (SE)
1.90
1.90
1.90
2.22
2.22
2.22
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Medium Sulfur Lignite (LD)
1.51
1.51
1.51
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Lignite (LE)
2.38
2.27
2.49
1.83
1.83
1.83
2.76
2.76
2.76
3.00
3.00
3.00
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LG)
3.91
3.91
3.91
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LH)
5.67
5.67
5.67
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Table 7-6 Coal Clustering by Coal Grade - Mercury Emission Factors (Ibs/TBtu)
Mercur
/ Emission Factors (Ibs/TBtu)
Coal Type by Sulfur Grade
Cluster #1
Cluster #2
Cluster #3
Cluster #4
Cluster #5
Cluster #6
Cluster# 7
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Low Sulfur Bituminous (BA)
4.37
4.37
4.37
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Bituminous (BB)
5.75
5.75
5.75
5.27
5.27
5.27
1.82
1.82
1.82
4.05
3.93
4.18
4.70
4.61
4.79
23.03
23.03
23.03
-
-
-
Low Medium Sulfur Bituminous (BD)
7.28
7.28
7.28
5.82
5.67
5.97
21.67
21.67
21.67
5.68
4.38
6.98
8.09
8.09
8.09
10.27
10.27
10.27
-
-
-
Medium Sulfur Bituminous (BE)
12.58
12.58
12.58
6.52
6.52
6.52
5.21
5.21
5.21
8.53
7.93
9.22
15.62
15.62
15.62
18.33
17.95
18.70
10.28
10.28
10.28
High Sulfur Bituminous (BG)
9.06
8.56
9.36
21.54
21.54
21.54
6.73
5.91
7.20
11.99
11.99
11.99
4.09
4.09
4.09
26.07
26.07
26.07
17.59
16.64
18.54
High Sulfur Bituminous (BH)
5.43
5.43
5.43
16.46
16.46
16.46
8.03
7.11
8.82
13.93
13.93
13.93
34.71
34.71
34.71
-
-
-
-
-
-
Low Sulfur Subbituminous (SA)
4.94
4.24
5.29
5.61
5.61
5.61
5.61
5.61
5.61
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Subbituminous (SB)
2.03
2.03
2.03
6.25
6.25
6.25
4.60
4.60
4.60
4.22
4.22
4.22
6.25
6.06
6.44
-
-
-
-
-
-
Low Medium Sulfur Subbituminous (SD)
4.74
4.33
5.35
7.54
7.54
7.54
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Subbituminous (SE)
8.65
8.65
8.65
4.41
4.41
4.41
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Medium Sulfur Lignite (LD)
Medium Sulfur Lignite (LE)
7.53
7.81
7.53
7.32
7.53
8.30
11.33
11.33
11.33
12.44
12.44
12.44
14.65
14.65
14.65
—
—
—
—
—
—
—
—
—
High Sulfur Lignite (LG)
High Sulfur Lignite (LH)
14.88
30.23
14.88
30.23
14.88
30.23
--
—
--
--
—
--
--
—
--
--
—
--
—
—
—
—
—
—
7-8
-------
Table 7-7 Coal Clustering by Coal Grade - Ash Emission Factors (Ibs/MMBtu)
Ash Emission Factors (Ibs/MMBtu)
Coal Type by Sulfur Grade
Cluster #1
Cluster #2
Cluster #3
Cluster #4
Cluster #5
Cluster #6
Cluster# 7
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Value
Low
High
Low Sulfur Bituminous (BA)
7.39
7.39
7.39
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Bituminous (BB)
9.15
9.15
9.15
7.86
7.86
7.86
5.58
5.58
5.58
8.65
7.83
9.76
6.69
6.41
6.97
58.98
58.98
58.98
-
-
-
Low Medium Sulfur Bituminous (BD)
10.83
10.83
10.83
7.71
7.45
7.97
49.31
49.31
49.31
9.42
8.34
10.50
9.25
9.25
9.25
9.18
9.18
9.18
-
-
-
Medium Sulfur Bituminous (BE)
10.70
10.70
10.70
6.61
6.61
6.61
7.97
7.97
7.97
7.48
5.37
9.89
11.70
11.70
11.70
8.16
7.08
9.23
7.89
7.89
7.89
High Sulfur Bituminous (BG)
7.08
6.48
7.83
9.59
9.59
9.59
8.03
6.39
9.46
10.21
10.21
10.21
8.47
8.47
8.47
13.54
13.54
13.54
12.30
8.00
16.60
High Sulfur Bituminous (BH)
9.06
9.06
9.06
11.56
11.56
11.56
9.49
8.63
10.21
9.13
9.13
9.13
13.89
13.89
13.89
-
-
-
-
-
-
Low Sulfur Subbituminous (SA)
4.97
3.98
5.47
5.51
5.51
5.51
7.09
7.09
7.09
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Subbituminous (SB)
7.06
7.06
7.06
5.81
5.81
5.81
14.51
14.51
14.51
8.72
8.72
8.72
7.43
6.94
7.92
-
-
-
-
-
-
Low Medium Sulfur Subbituminous (SD)
10.58
10.02
11.60
23.09
23.09
23.09
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Subbituminous (SE)
23.97
23.97
23.97
5.71
5.71
5.71
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Medium Sulfur Lignite (LD)
Medium Sulfur Lignite (LE)
11.57
15.00
11.57
12.85
11.57
17.15
11.69
11.69
11.69
21.51
21.51
21.51
25.65
25.65
25.65
—
—
—
—
—
—
—
—
—
High Sulfur Lignite (LG)
High Sulfur Lignite (LH)
25.51
23.95
25.51
23.95
25.51
23.95
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
Table 7-8 Coal Clustering by Coal Grade - HCI Emission Factors (Ibs/MMBtu)
HCI Emission Factors (Ibs/MMBtu)
Cluster #1
Cluster #2
Cluster #3
Cluster #4
Cluster# 5
Cluster #6
Cluster #7
Coal Type by Sulfur Grade
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Cluster
Data Range
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low Sulfur Bituminous (BA)
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Bituminous (BB)
0.09
0.09
0.09
0.07
0.07
0.07
0.01
0.01
0.01
0.01
0.01
0.02
0.08
0.05
0.11
0.03
0.03
0.03
-
-
-
Low Medium Sulfur Bituminous (BD)
0.03
0.03
0.03
0.06
0.03
0.09
0.07
0.07
0.07
0.06
0.03
0.10
0.10
0.10
0.10
0.10
0.10
0.10
-
-
-
Medium Sulfur Bituminous (BE)
0.03
0.03
0.03
0.21
0.21
0.21
0.04
0.04
0.04
0.07
0.03
0.10
0.07
0.07
0.07
0.09
0.07
0.10
0.09
0.09
0.09
High Sulfur Bituminous (BG)
0.07
0.06
0.08
0.09
0.09
0.09
0.06
0.02
0.11
0.04
0.04
0.04
0.13
0.13
0.13
0.05
0.05
0.05
0.04
0.02
0.07
High Sulfur Bituminous (BH)
0.10
0.10
0.10
0.05
0.05
0.05
0.04
0.02
0.05
0.06
0.06
0.06
0.15
0.15
0.15
-
-
-
-
-
-
Low Sulfur Subbituminous (SA)
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Subbituminous (SB)
0.01
0.01
0.01
0.02
0.02
0.02
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
-
-
-
-
-
-
Low Medium Sulfur Subbituminous (SD)
0.01
0.01
0.01
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Subbituminous (SE)
0.01
0.01
0.01
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Medium Sulfur Lignite (LD)
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Lignite (LE)
0.01
0.01
0.01
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
0.02
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LG)
0.04
0.04
0.04
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LH)
0.01
0.01
0.01
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
7-9
-------
Table 7-9 Coal Clustering by Coal Grade - C02 Emission Factors (Ibs/MMBtu)
Coal Type by Sulfur Grade
C02 Emission Factors (Ibs/MMBtu)
Cluster #1
Cluster #2
Cluster #3
Cluster #4
Cluster #5
Cluster #6
Cluster# 7
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Cluster
Value
Data Range
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low
High
Low Sulfur Bituminous (BA)
209.6
209.6
209.6
—
—
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Bituminous (BB)
206.4
206.4
206.4
210.7
207.1
215.5
214.3
214.3
214.3
208.4
204.7
209.6
206.4
206.4
206.4
204.7
204.7
204.7
-
-
-
Low Medium Sulfur Bituminous
(BD)
Medium Sulfur Bituminous (BE)
204.7
204.7
204.7
206.4
206.4
206.4
204.7
204.7
204.7
212.9
209.6
216.1
206.4
206.4
206.4
204.7
204.7
204.7
-
-
-
204.7
204.7
204.7
203.1
203.1
203.1
203.1
203.1
203.1
206.7
204.7
209.6
204.7
204.7
204.7
204.7
204.7
204.7
204.7
204.7
204.7
High Sulfur Bituminous (BG)
204.1
202.8
204.7
204.7
204.7
204.7
206.2
203.1
215.5
206.4
206.4
206.4
202.8
202.8
202.8
202.8
202.8
202.8
204.7
204.7
204.7
High Sulfur Bituminous (BH)
203.1
203.1
203.1
204.7
204.7
204.7
203.6
203.1
204.7
204.7
204.7
204.7
204.7
204.7
204.7
-
-
-
-
-
-
Low Sulfur Subbituminous (SA)
215.7
215.5
216.1
214.3
214.3
214.3
214.3
214.3
214.3
-
-
-
-
-
-
-
-
-
-
-
-
Low Sulfur Subbituminous (SB)
212.8
212.8
212.8
215.5
215.5
215.5
209.2
209.2
209.2
214.3
214.3
214.3
214.9
214.3
215.5
-
-
-
-
-
-
Low Medium Sulfur
Subbituminous (SD)
215.1
214.3
215.5
209.2
209.2
209.2
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Medium Sulfur Subbituminous
(SE)
Low Medium Sulfur Lignite (LD)
209.2
209.2
209.2
214.3
214.3
214.3
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
219.3
219.3
219.3
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
Medium Sulfur Lignite (LE)
216.0
212.6
219.3
219.3
219.3
219.3
217.9
216.5
219.3
212.6
212.6
212.6
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LG)
212.6
212.6
212.6
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
High Sulfur Lignite (LH)
212.6
212.6
212.6
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
7-10
-------
7.1.5 Coal Grade Assignments
The grades of coal that may be used by specific generating units were determined by an expert
assessment of the ranks of coal that a unit had used in the past, the removal efficiency of the installed
FGD, and the SO2 permit rate of the unit. Examples of the coal grade assignments made for individual
plants in EPA Platform v6 are shown in Table 7-10. Not all of the coal grades allowed to a plant by the
coal grade assignment are necessarily available in the plant's assigned coal demand region (due to
transportation limitations). IPM endogenously selects the coal burned by a plant by taking into account
both the constraint of the plant's coal grade assignment and the constraint of the coals actually available
within a plant's coal demand region.
Table 7-10 Example of Coal Assignments Made in EPA Platform v6
Plant Name
Unit
Permit Rate
(Ibs/MMBtu)
Scrubber?
Fuels Allowed
Mt Storm
3
0.15
Yes
BA,BB,BD
Mitchell
1
1.2
Yes
BA,BB,BD,BE,BG,BH
Scherer
1
1.2
Yes
SA,SB, SD,SE
Newton
1
0.5
No
SA, SB, SD, SE
R M Heskett
B2
1.97
Yes
LD, LE, LG, LH, SA, SB, SD, SE
San Miguel
SM-1
1.2
Yes
LD, LE, LG, LH
7.2 Coal Supply Curves
7.2.1 Nature of Supply Curves Developed for EPA Platform v6
In keeping with IPM's data-driven bottom-up modeling framework, a bottom-up approach (relying heavily
on detailed economic and resource geology data and assessments) was used to prepare the thermal coal
supply curves for EPA Platform v6. EPA utilized Wood Mackenzie to develop the curves based on their
extensive experience in preparing mine-by-mine estimates of cash operating costs for operating mines in
the U.S., their access to both public and proprietary data sources, and their active updating of the data
through both research and interviews.
In order to establish consistent nomenclature, Wood Mackenzie first mapped its internal list of coal
regions and qualities to EPA's 36 coal supply regions (described above in sections 7.1.1) and 14 coal
grades (described above in section 7.1.3). The combined code list is shown in Table 7-11 below with the
IPM coal supply regions appearing in the rows and the coal grades in the columns. Wood Mackenzie
then created supply curves for each region and coal-grade combination (indicated by the "x" in Table
7-11) for forecast years 2021, 2023, 2025, 2030, 2035, 2040, 2045, and 2050.
Table 7-11 Basin-Level Groupings Used in Preparing v6 Coal Supply Curves
Bituminous
Lignite
Subbituminous
Coal Supply
Region
Geo Region
Geo. Sub-Region
BA
BB
BD
BE
BG
BH
LD
LE
LG
LH
SA
SB
SD
SE
AB
Canada
Alberta, Canada
X
X
X
AL
Appalachia
Southern Appalachia
X
X
X
AN
Interior
West Interior
X
AZ
West
Southwest
X
BC
Canada
British Columbia
X
CG
West
Rocky Mountain
X
X
CR
West
Rocky Mountain
X
CU
West
Rocky Mountain
X
7-11
-------
Bituminous
Lignite
Subbituminous
Coal Supply
Region
Geo Region
Geo. Sub-Region
BA
BB
BD
BE
BG
BH
LD
LE
LG
LH
SA
SB
SD
SE
IL
Interior
East Interior (Illinois
Basin)
X
X
X
IN
Interior
East Interior (Illinois
Basin)
X
X
X
KE
Appalachia
Central Appalachia
X
X
X
X
KS
Interior
West Interior
X
KW
Interior
East Interior (Illinois
Basin)
X
X
LA
Interior
Gulf Lignite
X
MD
Appalachia
Northern Appalachia
X
X
ME
West
Dakota Lignite
X
MO
Interior
West Interior
X
MP
West
Powder River Basin
X
X
X
MS
Gulf
Gulf Lignite Coast
X
MT
West
Western Montana
X
ND
West
Dakota Lignite
X
NS
West
Southwest
X
X
X
OH
Appalachia
Northern Appalachia
X
X
X
OK
West
West Interior
X
PC
Appalachia
Northern Appalachia
X
X
X
X
X
PW
Appalachia
Northern Appalachia
X
X
X
SK
Canada
Saskatchewan
X
X
TN
Appalachia
Central Appalachia
X
TX
Interior
Gulf Lignite
X
X
X
UT
West
Rocky Mountain
X
X
X
X
VA
Appalachia
Central Appalachia
X
X
X
WG
West
Western Wyoming
X
X
X
X
WH
West
Powder River Basin
X
WL
West
Powder River Basin
X
X
WN
Appalachia
Northern Appalachia
X
X
X
X
WS
Appalachia
Central Appalachia
X
X
X
X
7.2.2 Cost Components in the Supply Curves
Costs are represented as total cash costs, which is a combination of a mine's operating cash costs plus
royalty & levies. These costs are estimated on a Free on Board (FOB) basis at the point of sale. Capital
costs (either expansionary or sustaining) are not included in the cash cost estimate for existing mines.
For projects, the expansionary capital is spread across the mine life and included into the costs. We
believe that total cash cost is the best metric for the supply curves as coal prices tend to be ultimately
determined by the incremental cost of production (i.e. total cash cost).
Operating cash cost
These are the direct operating cash costs and includes, where appropriate, mining, coal preparation,
product transport, and overheads. No capital cost component or depreciation & amortization charge is
included for operating mines. Expansionary capital is included for new greenfield projects. Operating
cash costs consist of the following elements:
7-12
-------
• Mining costs - Mining costs are the direct cost of mining coal and associated waste material for
surface and underground operations. It includes any other mine site costs, such as ongoing
rehabilitation / reclamation, security, community development costs. It also includes the cost of
transporting raw coal from the mining location to the raw coal stockpile at the coal preparation plant.
• Coal preparation - The cost of coal preparation includes raw coal stockpile reclaim, crushing and
screening, washing and marketable coal product stockpiling (if applicable).
• Transport - This covers all transport costs of product coal to point of sale. Transport routes with
multiple modes (e.g. truck and rail) are shown as total cost per marketable ton for all stages of the
transport route. Loading charges are included in this cost if relevant.
• Overheads - This is any non-production related general and administration overheads that are
essential to the production and sale of a mine's coal product. Examples would be mine site staff not
related to mining, essential corporate management or a sales and marketing charge.
It is important to note that although the formula for calculating mine costs is consistent across regions,
some tax rates and fees vary by state and mine type. In general, there are two mine types: underground
(deep) or surface mines. Underground mining is categorized as being either a longwall (LW) or a
continuous room-and-pillar mine (CM). Geologic conditions and characteristics of the coal seams
determine which method will be used. Surface mines are typically categorized by the type of mining
equipment used in their operation such as draglines (DL), or truck & shovels (TS). These distinctions are
important because the equipment used by the mine affects productivity measures and ultimately mine
costs. Further information on operating cost methodology and assumptions can be found in Attachment
7-1.
Royalties and Levies
These include, where appropriate, coal royalties, mine safety levies, health levies, industry research
levies and other production taxes.
7.2.3 Procedures Employed in Determining Mining Costs
The total cash costs of mines have been estimated in current year terms using public domain information
including; geological reports, reported statistics on production, labor and input costs, and company
reports. The estimates have been validated by reference to information gained by visits to operations,
and discussions with industry participants.
Because the estimates are based only on public information and analysis, and do not represent private
knowledge of an operation's actual costs, there may be deviations from actual costs. In instances where
confidential information is held by Wood Mackenzie, it has not been used to produce the published
estimates. Several methods are employed for cost estimation depending on the availability of information
and the diversity of mining operations. When possible, Wood Mackenzie analysts developed detailed
lists of mine related costs. Costs such as employee wages & benefits, diesel fuel, spare parts, roof bolts,
and explosives among a host of others are summed to form a mine's operating cash costs.
Where information is incomplete, cost items are grouped into categories that can be compared with
industry averages by mine type and location. These averages can be adjusted up or down based on new
information or added assumptions. The adjustments take the form of cost multipliers or parameter
values. Specific cost multipliers are developed with the aid of industry experts and proprietary formulas.
This method is at times used to convert materials and supplies, on-site trucking costs and mine and
division overhead categories into unit removal costs by equipment type. To check the accuracy of these
cost estimates, cash flow analysis of publicly traded companies is used. Mine cash-costs are extracted
from corporate cash flows and compared with the initial estimates. Adjustments for discrepancies are
made on a case-by-case basis.
7-13
-------
Many of the cost assumptions associated with labor and productivity were taken from the Mine Safety
Health Administration (MSHA) database. All active mines report information specific to production levels,
number of employees and employee hours worked. Wood Mackenzie supplements the basic MSHA data
with information obtained from mine personnel interviews and industry contacts. Phone conversations
and conferences with industry professionals provide additional non-reported information such as work
schedules, equipment types, percentages of washed coal, and trucking distances from the mine to wash-
plants and load-out terminals.
For each active or proposed mine, Wood Mackenzie reports the estimated cost to take coal from the mine
to a logical point-of-sale. The logical point-of-sale may be a truck or railcar load-out or even a barge
facility. This is done to produce a consistent cost comparison between mines. Any transport costs
beyond the point-of-sale terminal are not part of this analysis and are not reflected in the supply curves
themselves.
7.2.4 Procedure Used In Determining Mine Productivity
Projected production and stripping ratios are the key determinants of surface mine productivity. Wood
Mackenzie assumes mining costs increase as stripping ratios increase. The stripping ratio is the quantity
of overburden removed relative to the quantity of coal recovered. Assuming that reserves are developed
where they are easiest to mine and deliver to market, general theory suggests that as the easy reserves
are depleted, greater amounts of overburden must be handled for the same amount of coal production;
thus causing a decrease in mining productivity. However, some productivity loss may be offset by
technology improvements in labor saving equipment.
In order to calculate the amount of employee hours, and therefore the labor cost, of future production
Wood Mackenzie uses a multi-step process. First, employee hours associated with coal production for
each mine are obtained from MSHA. Total production is then divided by these hours to calculate
productivity, measured in short tons per employee hour. Future production levels are divided by this
productivity measurement to obtain future employee hours needed to produce that volume of coal. From
there, the total staffing level can be determined and the associated cost calculated.
A similar approach is used for underground mines. First, as background, the specific factors affecting
productivity at such mines are identified. For example, underground mines do not have stripping ratios.
Productivity estimates for these mines largely depend on the type of mining technique used (which is a
function of the region's geology). For instance, longwall-mines can produce a high volume of low cost
coal but geologic constraints like small reserve blocks and the occurrence of faulting tends to limit this
technique to certain regions. In addition to geologic constraints, there are variables that can impact
underground-mine productivity that are often difficult to quantify and forecast.
7.2.5 Procedure to Determine Total Recoverable Reserves by Region and Type
Before mine operators are allowed to mine coal, they must request various permits, conduct
environmental impact studies (EIS) and, in many cases, notify corporate shareholders. In each of these
instances, mine operators are asked to estimate annual production and total recoverable reserves. Wood
Mackenzie uses the mine operators' statements as the starting point for production and reserves
forecasts. If no other material is available, interviews with company personnel will provide an estimate.
Region and coal type determinations for unlisted reserves are based on public information reported for
similarly located mines. Classifying reserves this way means considering not only a mine's geographic
location but also its geologic conditions such as depth and type of overburden and the specific identity of
the coal seam(s) being mined. For areas where public information is not available or is incomplete, Wood
Mackenzie engineers and geologists estimate reserve amounts based on land surveys and reports of
coal depth and seam thickness provided by the U.S. Geologic Service (USGS). This information is then
used to extrapolate reserve estimates from known coal sources to unknown sources. Coal quality
determinations for unknown reserves are assigned in much the same way.
7-14
-------
Once a mine becomes active, actual production numbers reported in corporate SEC filings and MSHA
reports are subtracted from the total reserve number to arrive at current reserve amounts. Wood
Mackenzie consistently updates the reserves database when announcements of new or amended
reserves are made public. As a final check, the Wood Mackenzie supply estimates are balanced against
the Demonstrated Reserve Base (DRB)73 estimates to ensure that they do not exceed the DRB
estimates.
7.2.6 New Mine Assumptions
New mines have been included based on information that Wood Mackenzie maintains on each supply
region. They include announced projects, coal lease applications and unassigned reserves reported by
mining companies. Where additional reserves are known to exist, additional incremental steps have been
added and designated with the letter "N" in the "Step Name" field of the supply curves. These
incremental steps were added based on characteristics of the specific region, typical mine size, and cost
trends. They do not necessarily imply a specific mine or mine type.
Wood Mackenzie has also identified technical coal reserves that may be commercial in the longer-term,
but would most likely not be developed until after the completion of mine development already underway
or announced. These reserves are often the "last step" in a coal supply curve due to the more difficult
geologic conditions and have been designated using the above methodology.
In addition to new mines, Wood Mackenzie also identifies extension mines. These are denoted with the
letter "A" at the end of an existing mine step name (e.g., E2A). These mine steps reflect the extension of
a particular mine operating through a new lease covering tracts not previously recoverable under the
existing mine operation. These mine expansions, like new mines, include the capital expansionary
component in their cost of production.
7.2.7 Other Notable Procedures
Currency Assumptions
For consistency with the cost basis used in EPA Platform v6, costs are converted to real 2016$.
Future Cost Adjustments
Changes in mine productivity are a key factor impacting the evolution of costs overtime. In general, mine
productivity is expected to continue to decline - in large part due to worsening geology and more difficult
to mine reserves. Productivity has declined at a -1.03% compound annual growth rate (CAGR) from
2000-2015 as shown in Figure 7-2.
73 Posted by the Energy Information Administration (EIA) in its Coal Production Report.
7-15
-------
9
8
E 7
-------
Supply Growth Limitations
To the maximum extent possible, the IPM model is set up to determine the optimal volume of coal supply,
which can be profitably supplied. For two of the lower cost basins (Powder River and Illinois basins),
maximum production capacities are included as constraints (production ceilings) to more accurately
reflect the upper bound of what could be produced in a given year. Those limits, represented in millions
of tons per year, are shown in Figure 7-4 below. These ceilings are necessary to guard against modeling
excess annual production capacity in certain basins. For instance, in the PRB, several of the "new" mines
reflect expansion mines that would not be developed until the initial mine is further depleted. In this case,
the production ceiling helps safeguard against a modeling scenario that would simultaneously produce
from both of these mines.
Figure 7-4 Maximum Annual Coal Production Capacity per Year (Million Short Tons)
2021
2023
2025
2030
2035
2040
2045
2050
ILB
180
200
220
240
240
240
240
240
PRB
480
500
520
560
600
600
600
600
7.2.8 Supply Curve Development
The description below describes the development of the coal supply curves. Table 7-26 shows the actual
coal supply curves.
Once costs are estimated for all new or existing mines, they are sorted by cash cost, lowest to highest,
and plotted cumulatively by production to form a supply curve. The supply curve then represents all
mines - new or existing as well as both underground and surface mines- irrespective of market demand.
Mines located toward the bottom of the curve have the lowest cost and are most likely to be developed
while the mines at the top of the curve are higher cost and will likely wait to be developed. The process
for developing a cumulative supply curve is illustrated in Figure 7-5 and Figure 7-6 below.
Figure 7-5 Illustration of Preliminary Step in Developing a Cumulative Coal Supply Curve
E = EXISTING MINE
N = NEW MINE
U = UNDERGROUND MINE
S = SURFACE MINE
New or
Existinq?
Mine
Type
Cost
Production
N
A
S
$ 30
2
E
B
u
1 20
4
N
C
s
$ 32
1
N
D
s
$ 36
0.5
E
E
s
$ 29
2
N
F
s
$ 28
2.5
E
G
u
$ 25
5
E
H
u
$ 23
4
E
I
u
$ 27
3
N
J
s
$ 35
0.25
Mine Cost and Production Amts
~ Production
In the table and graph above, mine costs and production are sorted alphabetically by mine name. To
develop a supply curve from the above table the values must be sorted by mine costs from lowest to
highest. A new column for cumulative production is added, and then a supply curve graph is created
which shows the costs on the 'Y' axis and the cumulative production on the 'X' axis. Notice below that the
curve contains all mines - new or existing as well as both underground and surface mines. The resulting
7-17
-------
curve is a continuous supply curve but can be modified to show costs as a stepped supply curve. (Supply
curves in stepped format are used in linear programming models like IPM.) See Figure 7-7 for a stepped
version of the supply curve example shown in Figure 7-6. Here each step represents an individual mine,
the width of the step reflects the mine's production, and its height shows the cost of production.
Figure 7-6 Illustration of Final Step in Developing a Cumulative Coal Supply Curve
New or
Existing? [»
Minep
Tvi>e|»
Cost[-*
Production P
Cum
Production
E
B
U
t 20
4
4
E
H
U
I 23
4
a
E
G
u
t 25
5
13
E
1
u
I 27
3
16
N
F
s
t 28
2.5
18.5
E
E
s
I 29
2
20.5
N
A
s
t 30
2
22.5
N
C
s
I 32
1
23.5
N
J
s
t 35
0.25
23.75
N
D
s
I 36
0.5
24.25
Smooth Supply Curve
$- -I 1 1 1 1 r—
0 5 10 15 20 25
Cumulative Production (Tons)
Figure 7-7 Example Coal Supply Curve in Stepped Format
Stepped Supply Cuive
Cumulative Production
MINE NAME
PRODUCTION AMOUNT
BHGI FEACJD
4 8 13 16 18.5 20.5 22.5 24 25 25.5
E
1
E
2
E
3
E
4
E
5
E
mm
6
E
mm
7
E
E
8
9
u
E
c
10
E
t
11
E
E
0
12
13
E
E
n
14
15
N
16
N
N
S
c
17
18
E
19
E
¦¦
20
N
21
N
Bl
22
N
23
N
24
N
25
$
20
-
-
-
-
-
-
$
20
-
-
-
-
-
-
$
20
-
-
-
-
-
-
-
$
23
-
-
-
-
-
-
$
23
-
-
-
-
-
-
$
23
-
-
-
-
-
-
%
23
-
-
-
-
-
-
-
$
25
-
-
-
-
-
-
$
25
-
-
-
-
-
-
$
25
-
-
-
-
-
-
$
25
-
-
-
-
-
-
$
25
-
-
-
-
-
-
-
$
27
-
-
-
-
-
-
$
27
-
-
-
-
-
-
$
27
-
-
-
-
-
-
-
$
28
-
-
-
-
-
-
$
28
-
-
-
-
-
-
$
28
-
-
-
-
-
-
- $
29
-
-
-
-
-
- $
29
-
-
-
-
-
-
- $
30
-
-
-
-
-
- $
30
-
-
-
-
-
-
- $
32
-
-
-
-
-
-
- $
35
-
-
-
-
-
-
- $
36
7-18
-------
7.2.9 EPA Platform v6 Assumptions and Outlooks for Major Supply Basins
Powder River Basin CPRB')
The PRB is somewhat unique to other U.S. coal basins in that producers have the ability to add
significant production volumes relatively easily and at a profit. That said, the decisions on production
volumes are largely based on the market conditions, namely the price. For instance, in a low price
environment producers tend to moderate production volumes to maintain attractive prices, and choose to
ramp up production when prices are higher. The evolution of costs in the PRB will be strongly correlated
to the rate at which producers ramp up production at existing mines, which as indicated will depend on
market conditions.
Wood Mackenzie anticipates productivity at most existing PRB mining operations to decline at very
modest rates over the forecast horizon, with increasing strip ratios at least partly offset by improved usage
of labor and capital. As most PRB mines are progressing downward, the ratios of overburden to coal
(strip ratios) will increase in the future. The productivity of new mines will be quite low during the early
stages of their life span.
Mining at several locations is steadily proceeding westward toward the Joint Line railroad and, at current
and forecasted levels of production, around 2023 several mines are expected to eventually reach the line.
This event will result in a costly movement across the railroad, requiring significant capital investment and
reduced production as the transition is made. During the move across the Joint Line railroad, strip ratios
will spike and productivity will fall as new box cuts are created.
Illinois Basin ALB)
Production costs in the Illinois basin have been steadily decreasing in recent years as new low cost mines
are opened using more efficient longwall mining techniques. Development of these longwalls has slowed
as natural gas prices fell significantly. Many developments have been delayed until prices, and demand,
recover. In the long-term, the shape of the ILB supply curve is expected to increase in production
capacity and decrease in costs. However, this is not due to a lowering of costs at existing mines. Rather
it is caused by new mines being coming online that have lower operating costs than existing mines.
Given its large scale growth potential, investments in rail infrastructure development will have to keep
pace. While Wood Mackenzie expect there to be some bottlenecks in expanding transportation in the
basin early on, they project that once utilities begin committing to taking ILB coal, railroads will make the
necessary changes to accommodate the change. However, there is a risk that rail infrastructure in the
basin will not be able to keep up with the rate of growth in ILB which could limit the region's otherwise
strong growth potential.
Central Appalachia (CAPP)
Geologic conditions in the CAPP region are challenging, with thin seams and few underground reserves
amenable to more efficient longwall mining techniques. Costs of production in CAPP have risen
substantially in recent years as the region has struggled with mining thinner seams as reserves deplete,
mining accidents have led to increased inspections, and mine permitting has become increasingly difficult
as opposition to surface mining intensifies - with the revocation of some section 404 permits that regulate
the discharge into U.S. waterways. Since surface mining is the lowest cost form of production in CAPP,
reduced growth in surface mining operations is adding to increasing cost in the region.
In the years leading up to 2017, producers have cut back production significantly as coal prices
plummeted. Many companies went bankrupt and closed a large proportion of mines. As a result,
average costs have fallen substantially as high cost, low productivity mines were closed. In an effort to
retain margins, producers implemented a variety of tactics at continuing operations to try to keep
7-19
-------
production costs from continuing to increase, including shifting more production to lower cost operations
and selling lesser quality raw coal to save on coal preparation/washing costs.
Northern Appalachia (NAPP)
Mining cost escalation in NAPP has slowed considerably recently. Future cost for the basin as a whole
will depend largely on the development of new reserve areas.
Out of the possible 17 billion short tons (Bst) of reserves, only 2.2 Bst has been identified - meaning
located at an existing mine or a named project. The remainder are reserves that are available for
development in the region but no engineering or permitting work has begun.
7.3 Coal Transportation
Table 7-25 presents the coal transportation matrix.
Within the United States, steam coal for use in coal-fired power plants is shipped via a variety of
transportation modes, including barge, conveyor belt, rail, truck, and lake/ocean vessel. A given coal-
fired plant typically only has access to a few of these transportation options and, in some cases, only has
access to a single type. The number of transportation options that a plant has when soliciting coal
deliveries influences transportation rate levels that plant owners are able to negotiate with transportation
providers.
Within the Eastern United States, rail service is provided predominately by two major rail carriers in the
region, Norfolk Southern (NS) and CSX Transportation (CSX). Within the Western United States, rail
service is also provided predominately by two major rail carriers, Burlington Northern Santa Fe (BNSF)
and Union Pacific (UP). Plants in the Midwestern United States may have access to rail service from
BNSF, CSX, NS, UP, the Canadian National (CN), Canadian Pacific (CP), or short-line railroads. Barge,
truck, and vessel service is provided by multiple firms, and conveyor service is only applicable to coal-
fired plants directly located next to mining operations (e.g., mine-mouth plants).
Between 2012 (when the coal transportation rate assumptions for EPA Base Case v.5.13 were finalized),
and 2016, coal production in the United States declined by 288 million tons/year, or 28% (from 1.016
billion tons in 2012 to 728 million tons in 2016.)74 Approximately 46 gigawatts of coal-fired generating
capacity (or about 14% of the total coal-fired generating capacity in the United States) retired in the period
between 2012 and 2016.75
Despite the large decline in coal production, transportation rate levels for most coal movements declined
relatively little in real terms between 2012 and 2016, as most providers of coal transportation elected to
accept declines in coal volume rather than making large reductions in rates in an attempt to compete
more aggressively with natural gas-fired generation.76 Between 2016 and 2020, rates for all modes of
coal transportation are expected to increase in real terms due to increases in fuel prices from the very low
2016 levels. Over the longer term, however, rates for most modes of coal transportation are expected to
be flat to declining in real dollars from the 2020 levels, reflecting relatively low levels of expected coal
demand throughout the 2021-2050 forecast period used in EPA Platform v6.
74 The coal production data cited here is U.S. Energy Information Administration (EIA) data. The data is from the
quarterly coal report released October 2017, and is available at https://www.eia.aov/coal/production/auarterlv/.
75 Data available at https://www.eia.aov/electricitv/data.php#aencapacitv.
76 As will be discussed in more detail later in this section, both BNSF and CSX did introduce some innovative rail
contracting structures in an attempt to make the dispatch of selected coal-fired generating plants more competitive
with natural gas-fired generation, and the coal transportation rate assumptions in EPA Platform v6 have been
modified to account for the effects of these programs. However, these programs only apply to a small number of
coal-fired generating units.
7-20
-------
The transportation methodology and rates presented below reflect expected long-run equilibrium
transportation rates as of December 2016, when the coal transportation rate assumptions for EPA
Platform v6 were finalized. The forecasted changes in transportation rates during the 2021-2050 forecast
period reflect expected changes in long-term equilibrium transportation rate levels, including the long-term
market dynamics that will drive these pricing levels.
All of the transportation rates discussed in this document are expected 2020 rates and are shown in 2016
real dollars.
7.3.1 Coal Transportation Matrix Overview
Description
The general structure of the coal transportation matrix in EPA Platform v6 is similar to the structure used
in EPA Base Case 5.13. Each of the U.S. and Canadian coal-fired generating plants included in EPA
Platform v6 is individually represented in the coal transportation matrix. This allows the coal
transportation routings, coal transportation distances, and coal transportation rates associated with each
individual coal-fired generating plant to be estimated on a plant-specific basis. The coal transportation
matrix shows the total rate to transport coal from selected coal supply regions to each individual coal-fired
generating plant.
The coal supply regions associated with each coal-fired generating plant in EPA Platform v6 are largely
unchanged from the previous version of IPM. The coal supply regions associated with each coal-fired
generating plant are the coal supply regions which were supplying each plant as of early 2016, have
supplied each plant in previous years, or are considered economically and operationally feasible sources
of additional coal supply during the forecast period in EPA Platform v6 (2021-2050.) A more detailed
discussion of the coal supply regions can be found in previous sections.
Methodology
Each coal supply region and coal-fired generating plant is connected via a transportation link, which can
include multiple transportation modes. For each transportation link, cost estimates, in terms of $/ton,
were calculated utilizing mode-based transportation cost factors, analysis of the competitive nature of the
moves, and overall distance that the coal type must move over each applicable mode. An example of the
calculation methodology for movements including multiple transportation modes is shown in Figure 7-8.
Figure 7-8 Calculation of Multi-Mode Transportation Costs (Example)
Rail Cost ($/ton) =
I Rate (mills/ton-mile) x Rail
Mileage
Transloading
Cost ($/ton)
Barge Cost ($/ton) =
Loading Cost ($/ton) + Barge Mill Rate
(mills/ton-mile) x Barge Mileage
7.3.2 Calculation of Coal Transportation Distances
Definition of applicable supply/demand regions
Coal-fired generating plants are linked to coal supply regions based on historical coal deliveries, as well
as based on the potential for new coal supplies to serve each coal-fired generating plant going forward. A
generating plant will usually have transportation links with more than one supply region, depending on the
various coal types that can be physically delivered and burned at that particular plant. On average, each
coal-fired generating plant represented in IPM is linked with about eight coal supply regions. Some plants
may have more than the average number of transportation links and some may have fewer, depending on
7-21
-------
the location of each plant, the transportation modes available to deliver coal to each plant, the boiler
design and emissions control technologies associated with each plant, and other factors that affect the
types of coal that can be burned at each plant.
For "mine-mouth" plants (plants for which the current coal supply is delivered from a single nearby mine,
generally by conveyor belt or using truck transportation) that are 200 MW or larger, Hellerworx has
estimated the cost of constructing facilities that would allow rail delivery of alternative coal supplies, and
the transportation rates associated with the delivery of alternative coal supplies. This includes the
construction of rail spurs (between one and nine miles in length depending on the proximity of each plant
to existing railroad lines) to connect each plant with existing railroad lines.
Transportation Links for Existing Coal-Fired Plants
Transportation routings from particular coal supply regions to particular coal-fired generating plants were
developed based on third-party software77 and other industry knowledge available to Hellerworx. Origins
for each coal supply region were based on significant mines or other significant delivery points within the
supply region, and the destination points were plant-specific for each coal-fired generating plant
represented in IPM. For routes utilizing multiple modes (e.g. rail-to-barge, truck-to-rail, etc.), distances
were developed separately for each transportation mode.
Transportation Links for New Coal-Fired Plants
Transportation links for new coal-fired plants that were under construction as of December 2016 were
developed using the same methodology as for existing plants, and these committed new plants were
included in IPM as of their expected date of commercial operation.
Coal transportation costs for new coal-fired plants not yet under construction (i.e., coal transportation
costs for new coal plants modeled by IPM) were estimated by selecting an existing coal plant within each
IPM Region whose coal supply alternatives, and coal transportation costs, were considered
representative of the coal supply alternatives and coal transportation costs that would likely be faced by
new coal plants within that same IPM Region. In cases where there are no existing coal plants within a
particular IPM Region, the coal supply alternatives and coal transportation costs applicable to that IPM
Region were estimated using a methodology similar to that used for the existing coal plants.78 Using this
consistent methodology across all of the IPM regions helps ensure that coal transportation costs for new
coal plants are properly integrated with and assessed fairly vis-a-vis existing coal-fired assets within the
IPM modeling structure.
7.3.3 Overview of Rail Rates
Competition within the railroad industry is limited. Two major railroads in the Western U.S. (BNSF and
UP) and two major railroads in the Eastern U.S. (CSX and NS) currently originate most of the U.S. coal
traffic that moves by rail.
As noted earlier in this section, rail rates have declined relatively little in recent years, despite a significant
decline in coal demand. However, continued strong competition from natural gas-fired generation and
renewables over the duration of the forecast period used in EPA Platform v6 (2021-2050) is expected to
limit future coal demand, and to lead to further real declines in rail rates over the long term.
77 Rail routing and mileage calculations utilize ALK Technologies PC*Miler software.
78 Since the Canadian government has phased out coal-fired generation in Ontario, and in late 2016 announced plans
to phase out coal-fired generation in Alberta by 2030, coal-fired generation was not modeled in the Canadian
provinces where it is not currently used.
7-22
-------
As of December 2016, the differential between rates at captive plants and rates at competitively served
plants was relatively narrow. The current relatively small differentials between captive and competitive
rates are expected to persist over the long-term.
All of the rail rates discussed below include railcar costs, and include fuel surcharges at expected 2020
fuel price levels.
Overview of Rail Competition Definitions
Within the transportation matrix, rail rates are classified as being either captive or competitive (see Table
7-12) depending on the ability of a given coal demand region to solicit supplies from multiple suppliers.
Competitive rail rates are further subdivided into high- and low-cost competitive subcategories.
Competition levels are affected both by the ability to take delivery of coal supplies from multiple rail
carriers, the use of multiple rail carriers to deliver coal from a single source (e.g., BNSF/UP transfer to
NS/CSX for PRB coal moving east), or the option to take delivery of coal via alternative transportation
modes (e.g., barge, truck or vessel).
Table 7-12 Rail Competition Definitions
Competition Type
Definition
Captive
Demand source can only access coal supplies through a single provider; demand source
has limited power when negotiating rates with railroads.
High-Cost Competitive
Demand source has some, albeit still limited, negotiating power with rail providers;
definition typically applies to demand sources that have the option of taking delivery from
either of the two major railroads in the region.
Low-Cost Competitive
Demand source has a strong position when negotiating with railroads; typically, these
demand sources also have the option of taking coal supplies via modes other than rail
(e.g., barge, truck, or lake/ocean vessel).
Rail Rates
As previously discussed, rail rates are subdivided into three competitive categories: captive, high-cost
competitive, and low-cost competitive. Moves are further subdivided based on the distance that the coal
supply must move over rail lines: <200 miles, 200-299 miles, 300-399 miles, 400-649 miles, and 650+
miles. Within the Western U.S., mileages are only subdivided into two categories (<300 miles and 300+
miles), given the longer distances that these coal supplies typically move.
Initial rate level assumptions were determined based on an analysis of recent rate movements, current
rate levels in relation to maximum limits prescribed by the STB, expected coal demand, diesel prices,
recent capital expenditures by railroads, and projected productivity improvements. In general, shorter
moves result in higher applicable rail rates due to the lesser distance over which fixed costs can be
spread. As previously discussed, rail rates reflect anticipated 2020 costs in 2016 real dollars.
Rates Applicable to Eastern Moves
Rail movements within the Eastern U.S. are handled predominately by the region's two major carriers, NS
and CSX. Some short movements are handled by a variety of short-line railroads. Most plants in the
Eastern U.S. are served solely by a single railroad (i.e., they are captive plants). The practical effect of
this is that CSX and NS do not compete aggressively at the limited number of plants that have access to
both major railroads, and the rates for high-cost competitive plants tend to be similar to the rates for
captive plants. Table 7-13 presents the 2020 eastern rail rates.
7-23
-------
Table 7-13 Assumed Eastern Rail Rates for 2020
(2016 mills/ton-mile)
Mileage Block
Captive
High-Cost Competitive
Low-Cost Competitive
< 200
115
115
90
200-299
77
77
66
300-399
67
67
57
400-649
57
57
49
650+
37
37
35
In an attempt to help coal-fired generating plants located on its system compete more effectively with
natural gas-fired generation, CSX recently introduced a new structure for some of its rail contracts that
includes both fixed and variable components. Under this contracting structure, about 70% of the total rail
rate is a variable component that is charged on a $/ton basis for each ton of coal shipped, and the
remaining 30% of the total rail rate is a fixed dollar amount that is paid on a monthly basis. The goal of
this contract structure was to reduce dispatch costs (thus improving the utilization of the generating plants
using this contract structure), while leaving unchanged or increasing the total amount of revenue CSX
earns.
However, many larger generators (whose systems included both CSX-served plants, and plants served
by NS or other transportation providers) felt that this contracting structure might tend to favor CSX-served
plants at the expense of other plants on their own systems, and/or unnecessarily complicate dispatching.
As a result, the contracting structure that includes fixed and variable rail rate components is currently only
used by a limited number of smaller generators, which have only CSX-served plants. This rail contracting
structure is modeled on a plant-specific basis within EPA Platform v6.
Rates Applicable to Midwestern Moves
Plants in the Midwestern U.S. may be served by BNSF, CN, CP, CSX, NS, UP or short-line railroads.
However, the rail network in the Midwestern U.S. is very complex, and most plants are served by only one
of these railroads. The Midwestern U.S. also includes a higher proportion of barge-served and truck-
served plants than is the case in the Eastern or Western U.S. Table 7-14 depicts 2020 rail rates in the
Midwest.
Table 7-14 Assumed Midwestern Rail Rates for 2020
(2016 mills/ton-mile)
Mileage Block
Captive
High-Cost Competitive
Low-Cost Competitive
<200
115
115
90
200-299
77
77
66
300-399
57
57
49
400-649
56
56
48
650+
41
41
35
Rates Applicable to Western Moves
Rail moves within the Western U.S. are handled predominately by BNSF and UP. Rates for Western coal
shipments from the PRB are forecast separately from rates for Western coal shipments from regions
other than the PRB. This reflects the fact that in many cases coal shipments from the PRB are subject to
competition between BNSF and UP, while rail movements of Western coal from regions other than the
PRB consist primarily of Colorado and Utah coal shipments that originate on UP, and New Mexico coal
shipments that originate on BNSF. PRB coal shipments also typically involve longer trains moving over
longer average distances than coal shipments from the other Western U.S. coal supply regions, which
means these shipments typically have lower costs per ton-mile than non-PRB coal shipments. In the
west, there are enough plants that have access to both BNSF and UP or a neutral carrier that the western
7-24
-------
railroads are concerned with losing coal volume to the competing railroad, and do offer more of a rate
discount to plants that can access both railroads (e.g., high-cost competitive).
During periods of unusually low natural gas prices, BNSF offered temporary spot rail rate discounts to a
few selected generating plants using PRB coal in order to improve the utilization of these plants.
Hellerworx believes that these discounts applied only to selected captive generating plants using PRB
coal in the Gulf Coast region, were implemented only when natural gas prices reached very low levels,
and were implemented primarily in the form of allowing rail rates at selected captive plants to temporarily
fall to the rate level applicable to competitively served plants. Since it is Hellerworx's belief that these rate
discounts would only apply at very low natural gas prices (likely below $3.00/MMBtu, in 2016$), these
rate discounts are not modeled in EPA Platform v6, but could be included in sensitivity analyses involving
very low natural gas prices.
Non-PRB Coal Moves
The assumed non-PRB western rail rates for 2020 are shown in Table 7-15.
Table 7-15 Assumed Non-PRB Western Rail Rates for 2020
(2016 mills/ton-mile)
Mileage Block
Captive
High-Cost Competitive
Low-Cost Competitive
< 300
52
36
36
300+
34
23
23
The assumed PRB western rail rates for 2020 are available in Table 7-16.
PRB Moves Confined to BNSF/UP Rail Lines
Table 7-16 Assumed PRB Western Rail Rates for 2020
(2016 mills/ton-mile)
Mileage Block
Captive
High-Cost Competitive
Low-Cost Competitive
< 300
30.8
21.2
21.2
300+
24.4
18.0
18.0
PRB Moves Transferring to Eastern Railroads
For PRB coal moving west-to-east, the coal transportation matrix assumes that the applicable low-cost
competitive assumption is applied to the BNSF/UP portion of the rail mileage, and an assumption of either
$2.20 per ton or 35 mills per ton-mile (whichever is higher) is applied to the portion of the movement that
occurs on railroads other than BNSF and UP. (The $2.20 per ton assumption is a minimum rate for short-
distance movements of PRB coal on Eastern railroads.)
7.3.4 Truck Rates
Truck rates include loading and transport components, and all trucking flows are considered competitive
because highway access is open to any trucking firm. The truck rates shown in Table 7-17 are expected
2020 rate levels, reflective of current rates as of December 2016, and expected changes in labor costs,
fuel prices, and equipment costs.
Table 7-17 Assumed Truck Rates for 2020
(2016 Real Dollars)
Market
Loading Cost ($/ton)
Transport (mills/ton-mile)
All Markets
1.20
110
7-25
-------
7.3.5 Barge and Lake Vessel Rates
As with truck rates, barge rates include loading and transport components, and all flows are considered
competitive because river access is open to all barge firms. The transportation matrix subdivides barge
moves into three categories, which are based on the direction of the movement (upstream vs.
downstream) and the size of barges that can be utilized on a given river. As with the other types of
transportation rates forecast in this analysis, the barge rate levels shown in Table 7-18 are expected 2020
rate levels reflective of current rates as of December 2016, and expected changes in labor costs, fuel
prices, and equipment costs.
Table 7-18 Assumed Barge Rates for 2020
(2016 Real Dollars)
Type of Barge Movement
Loading Cost
($/ton)
Transport
(mills/ton-mile)
Upper Mississippi River, and Downstream on the Ohio River System
4.10
13.5
Upstream on the Ohio River System
3.90
13.0
Lower Mississippi River
3.00
10.1
Notes:
1. The Upper Mississippi River is the portion of the Mississippi River north of St. Louis.
2. The Ohio River System includes the Ohio, Big Sandy, Kanawha, Allegheny, and Monongahela Rivers.
3. The Lower Mississippi River is the portion of the Mississippi River south of St. Louis.
Rates for transportation of coal by lake vessel on the Great Lakes were forecast on a plant-specific basis,
taking into account the lake vessel distances applicable to each movement, the expected backhaul
economics applicable to each movement (if any), and the expected changes in labor costs and fuel and
steel prices over the long-term.
7.3.6 Transportation Rates for Imported Coal
Transportation rates for imported coal reflect expectations regarding the long-term equilibrium level for
ocean vessel rates, taking into account expected long-run equilibrium levels for labor, fuel, and equipment
costs.
In EPA Platform v6, it is assumed that imported coal is likely to be used only at plants that can receive
this coal by direct water delivery (i.e., via ocean vessel or barge delivery to the plant). This is based on
an assessment of recent transportation market dynamics, which suggests that railroads are unlikely to
quote rail rates that will allow imported coal to be cost-competitive at rail-served plants. Moreover, import
rates are higher for the Alabama and Florida plants than for New England plants because many of the
Alabama and Florida plants are barge-served (which requires the coal to be transloaded from ocean
vessel to barge at an ocean terminal, and then moved by barge to the plant), whereas most of the New
England plants can take imported coal directly by vessel. The assumed costs are summarized in Table
7-25.
7.3.7 Other Transportation Costs
In addition to the transportation rates already discussed, the transportation matrix assumes various other
rates that are applied on a case-by-case basis, depending on the logistical nature of a move. These
charges apply when coal must be moved between different transportation modes (e.g., rail-to-barge or
truck-to-barge) - see Table 7-19.
7-26
-------
Table 7-19 Assumed Other Transportation Rates for 2020
(2016 Real Dollars)
Type of Transportation
Rate ($/ton)
Rail-to-Barge Transfer
1.50
Rail-to-Vessel Transfer
2.00
Truck-to-Barge Transfer
2.00
Rail Switching Charge for Short line
2.10
Conveyor
1.00
7.3.8 Long-Term Escalation of Transportation Rates
Overview of Market Drivers
According to data published by the Association of American Railroads (AAR), labor costs accounted for
about 32% of the rail industry's operating costs in 2014, and fuel accounted for an additional 21%. The
remaining 47% of the rail industry's costs relate primarily to locomotive and railcar ownership and
maintenance, and track construction and maintenance.
The performance of various cost indices for the railroad industry over the past five years (2Q2011 -
2Q2016) is summarized in Figure 7-9. As shown in Figure 7-9, the RCAF79 Unadjusted for Productivity
(RCAF-U), which tracks operating expenses for the rail industry, decreased at an annualized rate of
2.9%/year between the second quarter of 2011 and the second quarter of 2016, largely as a result of the
steep decline in fuel prices during 2015 and 2016.
Excluding fuel, the railroad industry's overall input costs (e.g., equipment) decreased by .3% in real terms
during the 2Q2011-2Q2016 period. The railroad industry's labor costs increased by an average of 0.4%
per year during the same period. During the 2021-2050 forecast period used in EPA Platform v6,
Hellerworx expects that labor costs for the railroad industry will continue to increase by approximately
0.5% per year in real terms. The rail industry's equipment and other costs are expected to remain flat.
79 The Rail Cost Adjustment Factor (RCAF) refers to several indices created for regulatory purposes by the STB,
calculated by the AAR, and submitted to the STB for approval. The indices are intended to serve as measures of the
rate of inflation in rail inputs. The meaning of various RCAF acronyms that appear in this section can be found in the
insert in Figure 7-9.
7-27
-------
Figure 7-9 Rail Cost Indices Performance (2Q2011-2Q2016)
1.400 •
1.300 •
1.200 •
Performance of Rail Cost Indices, 2Q2011-2Q2016
(annualized nom inal % change)
RCAF
Unadjusted
for
Productivity
(RCAF-U)
RCAF
Adjusted for
Productivity
(RCAF-U)
All-inclusive
Index Less
Fuel (All-LF)
Labor
Component
of RCAF-U
GDP-IPD
-2.9%
-3.8%
1.2%
1.9%
1.5%
1.100 •
1.000
0.900
0.800 •
0.700 •
ro
a.
CM
CM
CM
co
CO
CO
CO
O
CM
O
CM
O
CM
O
CO
O
CM
O
o
CM
o
o
CM
o
CM
O
CM
O
CO
o
CM
o
lO
LO
LO
LO
o
CM
o
CM
o
CM
o
CO
o
CM
o
o
CM
o
o
CM
o
CM
O
CM
O
CO
o
CM
o
— RCAF-U — —-RCAF-A All-LF GDP-IPD
~ RCAF-U Labor Component
The other major transportation modes used to ship coal (barge and truck) have cost drivers broadly
similar to those for rail transportation (labor costs, fuel costs, and equipment costs). However, a
significant difference in cost drivers between the transportation modes relates to the relative weighting of
fuel costs for the different transportation modes. Estimates as shown in Figure 7-10 show that, at 201480
fuel prices, fuel costs accounted for about 20% of long-run marginal costs for the rail industry, 35% of
long-run marginal costs for barges, and 50% of long-run marginal costs for trucks.
80 2014 was used as the reference point for fuel prices in this analysis because a) at the time the coal transportation
rate assumptions were finalized, the latest analysis of railroad operating expenses available from the AAR contained
2014 data, and b) the average fuel price forecast by EIA for the 2020-2040 period (in real dollars) was close to the
2014 fuel price level.
7-28
-------
Figure 7-10 Long-Run Marginal Cost Breakdown by Transportation Mode
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0% J ^ T ^ T
Rail Barge Truck
¦ Fuel ~ Other
7.3.9 Market Drivers Moving Forward
Diesel Fuel Prices
The Energy Information Administration's (ElA's) Annual Energy Outlook (AEO)81 forecast of long-term
equilibrium prices fordiesel fuel used in the transportation sector (see Table 7-20) shows expected prices
ranging from about $3.22/gallon in 2020 to about $4.74/gallon in 2040 (2016 real dollars). This
represents an average annual real increase in diesel fuel prices of about 2.0%/year during 2020-2040.
The coal transportation rate forecast for EPA Platform v6 assumes that this average rate of increase in
diesel fuel prices will apply over EPA's entire forecast period (2021-2050).
Table 7-20 EIA AEO Diesel Fuel Forecast, 2020-2040
(2016 Real Dollars)
Year
Rate ($/gallon)
2020
3.22
2025
3.60
2030
3.90
2035
4.31
2040
4.74
Annualized % Change, 2020-2040
2.0%
Source: EIA
81 As noted at the beginning of this section, the coal transportation rate assumptions for EPA Platform v6 were
finalized in December 2016. At that time, the Annual Energy Outlook 2016 forecast was the latest available.
7-29
-------
Labor Costs
As noted earlier, labor costs for the rail industry are expected to increase approximately 0.5% faster than
overall inflation, on average over the forecast period. Because competition is stronger in the barge and
trucking industries than in the rail industry, labor costs in the barge and truck industries are expected to
increase at approximately the same rate as overall inflation, on average over the forecast period.
Productivity Gains
The most recent data which was available from the AAR at the time the coal transportation rate
assumptions used in EPA Platform v6 were finalized in December 2016 (covering 2010-2014) show that
rail industry productivity increased at an annualized rate of approximately 1.4% per year during this
period. Since coal-fired generation is expected to continue to face strong competition from natural gas-
fired generation and renewables during the entire 2021-2050 forecast period used in EPA Platform v6
(which will significantly limit coal demand), approximately half of the railroad industry's expected
productivity gains (0.7% per year) are forecast to be passed through to coal shippers.
The potential for significant productivity gains in the trucking industry is relatively limited since truckload
sizes, operating speeds, and truck driver hours are all regulated by law. Although increased lock outages
and the associated congestion on the inland waterway system as the river infrastructure ages may reduce
the rate of future productivity gains in the barge industry, limited productivity gains are expected to occur,
and these productivity gains are expected to be largely passed through to shippers since the barge
industry is highly competitive.
Long-Term Escalation of Coal Transportation Rates
Based on the foregoing discussion, rail rates are expected to decline at an average rate of 0.1% per year
in real terms during 2020-2050. Over the same period, barge and lake vessel rates are expected to
increase at an average rate of 0.2% per year, which includes some pass-through of productivity gains in
those highly competitive industries. Truck rates are expected to escalate at an average rate of
1.0%%/year during 2020-2050, largely due to increases in fuel costs. Rates for conveyor transportation
and transloading services are expected to be flat in real terms, on average over the forecast period.
The basis for these forecasts is summarized in Table 7-21.
7-30
-------
Table 7-21 Summary of Expected Escalation for Coal Transportation Rates, 2020-2050
Mode
Component
Component
Weighting
Real Escalation
Before Productivity
Adjustment
(%/year)
Productivity Gains
Passed Through to
Shippers (%/year)
Real Escalation
After Productivity
Adjustment
(%/year)
Rail
Fuel
Labor
Equipment
21%
32%
47%
2.0%
0.5%
0.0%
Total
100%
0.6%
0.7%
-0.1%
Barge &
Vessel
Fuel
Labor &
Equip.
35%
65%
2.0%
0.0%
Total
100%
0.7%
0.5%
0.2%
Truck
Fuel
Labor &
Equip.
50%
50%
2.0%
0.0%
Total
100%
1.0%
0.0%
1.0%
Conveyor
Total
0.0%
0.0%
0.0%
Transloading
Terminals
Total
0.0%
0.0%
0.0%
7.3.10 Other Considerations
Estimated Construction Costs for Railcar Unloaders and Rail Spurs at Mine-Mouth Plants
In order to allow mine-mouth generating plants (i.e., coal-fired generating plants which take all of their
current coal supply from a single nearby mine) to access additional types of coal, the costs of
constructing facilities that would allow rail delivery of coal was estimated for almost all82 of the mine-
mouth generating plants with total capacity of 200 MW or more.
The facilities needed for rail delivery of coal to generating plants of this relatively large size were assumed
to be: a) a rotary dump railcar unloader capable of handling unit train coal shipments, which is estimated
to cost about $25 million installed (in 2016$). b) at least three miles of loop track, which would allow for
one trainload of coal to be unloaded, and a second trainload of coal to simultaneously be parked on the
plant site preparatory to unloading, and c) at least one mile of additional rail spur track to connect the
trackage on the plant site with the nearest railroad main line. Since construction costs for rail trackage
capable of handling coal trains is estimated at about $3 million per mile (in 2016$), the minimum
investment required to construct the facilities needed for rail delivery of coal was estimated at $37 million.
In some cases, the length of the rail spur required to reach the nearest main line (which was estimated on
a plant-specific basis) is considerably longer than one mile. In cases where a rail spur longer than one
mile was required to reach the main line, the cost of the additional trackage was estimated using the
same construction cost of $3 million per mile (2016$) referenced earlier.
The total cost of the facilities required for rail delivery of coal was converted to an annualized basis based
on each plant's historical average coal burn from 2012-2015, and a capital recovery factor of 11.29%.
82 The costs of rail coal delivery were not estimated for mine-mouth plants located in the Powder River Basin or
Illinois Basin coal fields, since the coal reserves in these coal fields are among the largest, and among the cheapest
to mine, anywhere in the United States.
7-31
-------
The cost of transporting additional types of coal to each mine-mouth generating plant was then calculated
using the same methodology described earlier in this section, and added to the annualized cost for the
rail delivery facilities, to arrive at an estimated "all-in" cost for delivering additional types of coal to the
mine-mouth plants.
7.4 Coal Exports, Imports, and Non-Electric Sectors Demand
The coal supply curves used in EPA Platform v6 represent the total steam coal supply in the United
States. While the U.S. power sector is the largest consumer of native coal - roughly 91 % of mined U.S.
coal in 2017 was used in electricity generation - non-electricity demand must also be taken into
consideration in IPM modeling in order to determine the market-clearing price. Furthermore, some coal
mined within the U.S. is exported out of the domestic market, and some foreign coal is imported for use in
electricity generation, and these changes in the coal supply must be detailed in the modeling of the coal
supply available to coal power plants. The projections for imports, exports, and non-electric sector coal
demand are based on ElA's AEO 2017.
In EPA Platform v6, coal exports and coal-serving residential, commercial and industrial demand are
designed to correspond as closely as possible to the projections in AEO 2017 both in terms of the coal
supply regions and coal grades that meet this demand. The projections used exclude exports to Canada,
as the Canadian market is modeled endogenously within IPM. First, the subset of coal supply regions
and coal grades in EPA Platform v6 are identified that are contained in or overlap geographically with
those in EIA Coal Market Module (CMM) supply regions and coal grades that are projected as serving
exports and non-electric sector demand in AEO 2017. Next, coal for exports and non-electricity demand
are constrained by CMM supply region and coal grade to meet the levels projected in AEO 2017. These
levels are shown in Table 7-22, Table 7-23 and Table 7-24.
Table 7-22 Coal Exports (Million Short Tons)
Name
2021
2023
2025
2030
2035
2040
2045
2050
Central Appalachia - Bituminous Medium Sulfur
4.46
4.92
5.43
6.93
8.00
10.21
10.69
10.65
East Interior - Bituminous High Sulfur
0.00
0.00
0.00
0.89
1.57
3.32
3.41
2.92
East Interior - Bituminous Medium Sulfur
4.49
4.95
5.46
6.07
6.47
6.94
7.33
7.82
Northern Appalachia - Bituminous High Sulfur
2.39
2.63
2.90
3.71
4.28
5.46
0.00
0.00
Northern Appalachia - Bituminous Medium Sulfur
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Rocky Mountain - Bituminous Low Sulfur
2.78
2.56
2.25
1.79
2.07
2.65
2.77
2.77
Western Montana - Bituminous Low Sulfur
0.00
0.00
0.00
0.00
1.50
1.91
2.00
2.00
Western Montana - Subbituminous Low Sulfur
0.90
0.99
1.10
1.40
0.00
0.00
0.00
0.00
Western Montana - Subbituminous Medium Sulfur
5.26
5.80
6.40
8.17
1.04
0.00
0.00
0.00
Wyoming PRB - Subbituminous Low Sulfur
0.06
0.05
0.03
0.00
8.10
11.62
12.17
12.17
IPM then endogenously determines which IPM coal supply region(s) and coal grade(s) will be selected to
meet the required export or non-electric sector coal demand as part of the cost-minimization coal market
equilibrium. Since there are more coal supply regions and coal grades in EPA Platform v6 than in AEO
2017, the specific regions and coal grades that serve export and non-electric sector demand are not pre-
specified but modeled.
7-32
-------
Table 7-23 Residential, Commercial, and Industrial Demand (Million Short Tons)
Name
2021
2023
2025
2030
2035
2040
2045
2050
Central Appalachia - Bituminous Low Sulfur
2.60
2.69
2.68
2.55
2.45
2.44
2.41
2.38
Central Appalachia - Bituminous Medium Sulfur
7.83
8.12
8.12
7.73
7.45
7.42
7.37
7.33
East Interior - Bituminous High Sulfur
5.32
5.40
5.33
5.01
4.72
4.55
4.57
4.63
East Interior - Bituminous Medium Sulfur
0.79
0.81
0.81
0.79
0.78
0.78
0.77
0.77
Northern Appalachia - Bituminous High Sulfur
0.47
0.48
0.47
0.44
0.41
0.40
0.40
0.40
Northern Appalachia - Bituminous Medium Sulfur
2.14
2.13
2.06
1.94
1.84
1.81
1.76
1.72
Rocky Mountain - Bituminous Low Sulfur
4.95
5.08
5.07
4.84
4.62
4.56
4.62
4.68
Southern Appalachia - Bituminous Low Sulfur
0.19
0.20
0.20
0.19
0.18
0.18
0.18
0.18
Southern Appalachia - Bituminous Medium Sulfur
1.11
1.16
1.16
1.11
1.09
1.09
1.08
1.07
Wyoming PRB - Subbituminous Low Sulfur
4.28
4.40
4.37
4.12
3.92
3.84
3.83
3.85
Dakota Lignite - Lignite Medium Sulfur
4.87
4.95
4.88
4.57
4.29
4.12
4.14
4.20
West Interior - Bituminous High Sulfur
0.38
0.40
0.40
0.38
0.37
0.37
0.37
0.37
Arizona/New Mexico - Bituminous Low Sulfur
0.14
0.14
0.14
0.14
0.13
0.13
0.13
0.14
Arizona/New Mexico - Subbituminous Medium Sulfur
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
Western Montana - Subbituminous Low Sulfur
0.07
0.00
0.00
0.00
0.00
0.00
0.00
0.07
Western Wyoming - Subbituminous Low Sulfur
2.41
2.46
2.45
2.32
2.20
2.15
2.17
2.20
Western Wyoming - Subbituminous Medium Sulfur
1.12
1.16
1.16
1.12
1.08
1.08
1.10
1.11
Western Montana - Subbituminous Medium Sulfur
0.00
0.08
0.08
0.08
0.08
0.08
0.08
0.00
Imported coal is only available to 25 coal facilities, which are eligible to receive imported coal. These
facilities, which may receive imported coal, along with the cost of transporting this coal to the demand
regions, are in Table 7-25. The total U.S. imports of steam coal are limited to AEO 2017 projections as
shown in Table 7-24.
Table 7-24 Coal Import Limits (Million Short Tons)
2021
2023
2025
2030
2035
2040
2045
2050
Annual Coal Imports Cap
5.76
4.67
3.78
2.23
1.4
1.4
5
5.01
7-33
-------
Attachment 7-1 Mining Cost Estimation Methodology and Assumptions
Labor Costs
Productivity and labor cost rates are utilized to estimate the total labor cost associated with the mining
operation. This excludes labor involved in any coal processing / preparation plant.
Labor productivity is used to calculate mine labor and salaries by applying an average cost per employee
hour to the labor productivity figure reported by MSHA or estimated based on comparable mines.
Labor cost rates are estimated based on employment data reported to MSHA. MSHA data provides
employment numbers, employee hours worked and tons of coal produced. These data are combined
with labor rate estimates from various sources such as union contracts, census data and other sources
such as state employment websites to determine a cost per ton for mine labor. Hourly labor costs vary
between United Mine Workers of America (UMWA) and non-union mines, and include benefits and
payroll taxes. Employees assigned to preparation plants, surface activities, and offices are excluded from
this category and are accounted for under coal washing costs and mine overhead.
Surface Mining
The prime (raw coal) strip ratio and overburden volume is estimated on a year by year basis. Estimates
are entered of the amount of overburden83 moved each year, split by method to allow for different unit
mining costs. The unit rate cost for each method excludes any drill and blast costs, and labor costs, as
these are accounted for separately. Drill and blast costs are estimated as an average cost per volume of
prime overburden. If applicable, dragline re-handle is estimated separately and a summation gives the
total overburden moved.
The different overburden removal methods are:
• Dragline - the estimated volume of prime overburden moved
• Dragline re-handle - the estimated volume of any re-handled overburden
• Truck and shovel - including excavators.
• Other - examples would be dozer push, front end loader, or cast blasting. If overburden is moved by
cast blasting the unit rate is taken to be zero as the cost is already included in the drill and blast
estimate.
Surface mining costs also include the cost of coal mining estimated on a raw ton basis.
Underground Mining
Raw coal production is split by type into either continuous miner or longwall. Cost estimates can be input
either on a unit rate or a fixed dollar amount, as the cost structure of underground mining generally has a
large fixed component from year to year. Costs are divided into:
• Longwall
• Continuous miner
• Underground services
Underground services costs cover categories such as ventilation, conveyor transport, gas drainage,
secondary roof support etc.
83 Overburden refers to the surface soil and rock that must be removed to uncover the coal.
7-34
-------
Mine Site Other
This covers any mine site costs that are outside the direct production process. Examples are ongoing
rehabilitation/reclamation, security, community development costs.
Raw Haul
Costs for transporting raw coal from the mining location to the raw coal stockpile at the coal preparation
plant or rail load out. A distance and a unit rate allows for an increasing cost over time if required.
List of tables that are uploaded directly to the web:
Table 7-25 Coal Transportation Matrix in EPA Platform v6
Table 7-26 Coal Supply Curves in EPA Platform v6
Table 7-27 Coal Demand Regions in EPA Platform v6
7-35
-------
8. Development of Natural Gas Supply Curves for EPA Platform v6
8.1 Introduction
Natural gas supply curves and regional basis are key inputs to IPM and are developed using ICF's Gas
Market Model (GMM). ICF develops and maintains the GMM for use by both private and public sector
clients. The model has been used to examine strategic issues relating to natural gas supply, pipeline
infrastructure, pricing, and demand characteristics.
Like IPM, GMM is a large-scale linear programming model that incorporates a detailed representation of
gas supply characteristics, demand characteristics, and an integrating pipeline transportation model to
develop forecasts of gas supply, demand, prices, and flows. GMM is a full supply/demand equilibrium
model of the North American gas market. The model solves for monthly natural gas prices throughout
North America, given different supply/demand conditions, the assumptions for which are specified by
each scenario. Overall, the model solves for monthly market clearing prices by considering the
interaction between supply and demand at each node in the model.
On the supply side, prices are determined by production and storage price curves that reflect prices as a
function of production and storage utilization. Prices are also influenced by "pipeline discount" curves,
which reflect the change in basis or the marginal value of gas transmission as a function of load factor.
On the demand side, prices are represented by a curve that captures the fuel-switching behavior of end-
users at different price levels. The model balances supply and demand at all nodes in the model at the
market clearing prices. ICF updates GMM model inputs and performs calibration of the model on a
monthly basis to ensure the model reliably reflects historical gas market behavior. Figure 8-1 shows the
supply side of the calculation in GMM, and Figure 8-2 shows the interaction of IPM and GMM.
Figure 8-1 GMM Gas Quantity and Price Response
Production And
Storage Gas Price
Pipeline
Value
Deliverability
Production
Gas
Price
Production
And Storage
Only Includes Storage
During The Withdrawal Season
Pipeline Load Factor
Gas
Transmission
(Inelastic #
Demand
Distillate \
Switching \
Residual Oil Switching
Quantity Consumed
Gas
Demand
Includes Storage
During The Injection Season
8-1
-------
Figure 8-2 IPM/GMM Interaction
Natural Gas Modeling Power Sector Modeling
Natural Gas Supply Curves
Integrated
Planning
Model (IPM)
Gas Market
Model (GMM)
Power Sector Gas Demand
Activities
Activities
Assumption Updates
Scenario Analysis
Assumption Updates
Scenario Analysis
Products Products
Natural Gas Projections Environmental Policy Assessments
Natural Gas Supply Curves Inputs for Air Quality Modeling
To establish gas supply curves in EPA's Platform v6, both GMM and IPM are operated in tandem and are
iterated to develop a consistent Henry Hub gas price and total gas demand forecast that informs the
derivation of those supply curves. In subsequent analyses using IPM, EPA's Platform v6 will continue to
use these originally derived gas supply curves (without necessitating re-modeling in GMM), unless
otherwise documented with a given scenario analysis. EPA's Platform v6 uses natural gas market
assumptions in power market modeling as follows:
• IPM takes the natural gas supply curves, which are developed based on GMM outputs and
specified as a function of Henry Hub prices.
• For each year, three sets of seasonal natural gas transportation differentials (summer, winter, and
winter shoulder) are added to the supply curves to generate the final delivered curves by IPM
region.
• IPM projects the power sector's demand for natural gas. The natural gas demand is compared to
the supply curve to find the market-clearing price for natural gas.
• IPM's linear programming formulation takes into consideration the gas supply curves as well as
coal supply curves and detailed power plant modeling in determining electric market equilibrium
conditions. Oil usage is modeled as a function of price, which is exogenously supplied to IPM.
This chapter is divided in the following sections. The chapter starts with a brief synopsis of GMM, the
primary tool used for generating the natural gas supply curves. This is followed by detailed discussions of
modeling methodologies and data used in GMM. The methodologies and data description are grouped in
the following five sections:
i) Resources data and reservoir description
ii) Treatment of frontier resources and exports
iii) Exploration and production technology characterization
iv) Oil prices
v) Demand assumptions
This is followed by discussion of natural gas assumptions and supply curves used for EPA Platform v6.
8-2
-------
8.2 Brief Synopsis of GMM
ICF's integrated natural gas model, GMM, is designed to perform comprehensive assessments of the
entire North American gas flow pattern. It is a large-scale, dynamic linear program that models economic
decision-making to minimize the overall cost of meeting natural gas demand. GMM is reliable and
efficient in analyzing the broad range of natural gas market issues.
Figure 8-3 Geographic Coverage of GMM
Northern
Nodes
Everett
rLNG
Atlantic
Offshore
California
Oflstmre
Elba
Island
LNG
^ Lake
Charles LNG
Copyright 2010. ICF International
Important features of GMM are described below.
Natural Gas Market Prices in GMM are determined by the marginal (or incremental) value of natural gas
at 121 regional market centers. Prices are "at the margin", not "average". Marginal prices do not
translate directly into pipeline or utility revenues. Prices represent "market center" prices as opposed to
delivered prices. Gas prices are determined by the balance of supply and demand in a regional
marketplace. Supply is determined considering both availability of natural gas deliverability at the
wellhead, the transportation capacity and cost to deliver gas to market centers.
Natural gas production prices are determined from spot gas price curves that yield price as a function of
deliverability utilization: Curves reflect price for gas delivered into the transmission system (including
gathering cost). Gas storage withdrawal price curves are added to the production price curves during the
withdrawal season. Pipeline value curves are then added to yield a total supply curve for a node. The
intersection of the supply curve and the demand curve (including net storage injections) yields the
marginal price at a node. Price is set by the demand curve when all available supply is utilized.
Demand is modeled for residential, commercial, industrial, and power sectors for each of the 121 nodes.
GMM solves for gas demand across different sectors, given economic growth, weather, and the level of
price competition between gas and oil. Econometric equations define demand by sector. The industrial
8-3
-------
and power sectors incorporate fuel competition, dispatch decisions, new power plant builds, economic
growth, and weather. GMM solves the power generation dispatch on a regional basis to determine the
amount of gas used in power generation, which is allocated along with end-use gas demand to model
nodes. GMM iterates with IPM to better capture electric sector demand for natural gas.
Electric generation is modeled regionally with plant dispatch based upon operating cost. Competing
power generation technologies are evaluated on a full-cost basis to determine lowest cost capacity
additions.
Transportation is modeled by over 423 transportation links between the nodes, balancing seasonal,
sectoral, and regional demand and prices, including pipeline tariffs and capacity allocation. Node
structure developed to reflect points of change or influence on the pipeline system. These points include:
major demand and supply centers, pipeline hubs and market centers, and points of divergence in pipeline
corridors. The pipeline network is largely represented as bundles of pipes, though in some regions
individual pipes are represented. Gas moves over the network at variable cost. The variable cost as a
function of pipeline throughput (i.e., pipeline discount curve) is used to determine the market value of
capacity (i.e., the transportation basis) for each period in the forecast for each pipeline link.
Figure 8-4 Example Pipeline Discount Curve
N LA Hub (61) to East KY/TN (18)
0.80
0.70
0.60
0.50
~
£
2
>
M
-------
OR
Second Criteria: The project has the necessary approvals to proceed from FERC and other
relevant regulatory proceedings;
OR
Third Criteria: The project has been filed with FERC and has the necessary firm shipper
commitments;
OR
Fourth Criteria: The project has been filed with FERC and does not have the necessary shipper
commitments, but does appear to have sufficient market support;
OR
Fifth Criteria: The project has NOT yet been filed with FERC, but appears to have sufficient
market support.
For the fourth and fifth criteria, ICF typically considers supply growth directly upstream of the project,
market growth for markets that are relevant to the project's delivery point/s, and basis differentials that
exceed the per unit cost of pipeline expansion as indicators of market support. If the indicators are all
positive, ICF will add the project as a "generic" project and size it based on the level of market support. In
the case in which there are multiple generic projects for a single GMM link, the generic projects will be
sized in aggregate based on the total level of market support for expansion of the link. Generic projects
are classified as such until one of the first three criteria are satisfied.
For certain markets like New York, New Jersey and New England, ICF looks closely at regulatory support
for the project which could override the criteria above in determining the pipeline additions in GMM. For
example, if a project like Constitution pipeline has been denied water permits even though it has broad
market support, ICF does not include it in its base case.
Pipeline cost assumptions used in GMM have been derived by considering data from Oil and Gas Journal
(OGJ) surveys of pipeline projects. Using regression analysis of the OGJ data across years, we
determined an average U.S. pipeline cost of $183,000 per inch-mile for 2017 (in 2016 dollars) for large
gas transmission pipelines. The pipeline cost for future years is kept flat in real terms post 2017.
Regional cost multipliers have also been derived from OGJ data as the pipeline costs vary by region.
Cost multipliers can be different across regions; for example, costs are relatively high in the Northeast
where projects have been very difficult and time consuming to construct.
Supply is modeled by using node-level natural gas deliverability or supply capability, including import and
export levels while accounting for gas storage injections and withdrawals at different gas prices. Total
supply in the United States comes from three sources: production from natural gas fields located in the
lower 48 states, Canadian imports, Alaska, and LNG imports. Natural gas production activity is
represented in 82 of the 121 Model nodes where historical production has occurred, or where future
production appears likely. The "base trend" for deliverability and gas price are developed from ICF's
resource assessment using the Hydrocarbon Supply Model and a long-term "marker price". If the monthly
solution price deviates from the "marker price", future exploration and production (E&P) activity is
adjusted. Deliverability responds to prices, with a lag of 2 to 18 months.
The components of supply (i.e., gas deliverability, storage withdrawals, supplemental gas, LNG imports,
and Mexican imports) are balanced against demand (i.e., end-use demand, power generation gas
demand, LNG exports, and Mexican exports) at each of the nodes and gas prices are solved for in the
market simulation module.
Natural Gas Storage activity is represented for 24 U.S. and two Canadian storage regions, with activity
allocated to individual nodes based on historical field level storage capacity. Regional differences in the
physical and market characteristics of storage are captured in the storage injection and withdrawal
relationships separately estimated for each region:
Differences between market area storage and supply area storage.
8-5
-------
Differences between regions with primarily depleted field storage and regions with primarily
aquifer storage.
Net monthly withdrawals are calculated from a "storage supply curve" that reflects the level of withdrawals
relative to gas prices. The curve has been fit to actual historical data. Net monthly injections are
calculated from econometrically fit relationships that consider working gas levels, gas prices, and weather
(i.e.. cooling degree days). The level of gas storage withdrawals and injections are calculated within the
supply and demand balance algorithm based on working gas levels, gas prices, and extraction/injection
rates and costs.
Figure 8-5 GMM Natural Gas Storage Regions
Csi(£ada East
Montana/North
Oregon/Washington
Minnesota
Iowa/South
Pennsylvania
>ming/Utah
Nebraska
Indiana
Colorado
Californ \al)
Arkansas
i Alabama
Storage levels have an impact on GMM's seasonal basis differentials, which are an important component
in constructing the gas supply curves and/or basis differentials that are then input into IPM. The arbitrage
value of storage is driven by the seasonal difference in the supply-area gas prices plus the seasonal
difference in pipeline transportation value. Storage expansions (or increased utilization of existing
storage) decreases seasonal basis differentials in the region surrounding the storage facilities.
8.3 Resources Data and Reservoir Description
This section describes the approach used in GMM and documents the changes to the resource data and
reservoir characterization work conducted for EPA Platform v6.
8.3.1 U.S. Resources and Reserves
This section describes the U.S. resource data sources and methodology used in GMM for EPA Platform
v6.
Current U.S. and Canada gas production is from over 400 trillion cubic feet (Tcf) of proven gas reserves.
ICF assumes that the U.S. and Canada natural gas resource base totals roughly 3,500 Tcf of unproved
8-6
-------
plus discovered but undeveloped gas resource. This can supply the U.S. and Canada gas markets for
over 100 years (at current consumption levels). Shale gas accounts for over 50 percent of remaining
recoverable gas resources. No significant restrictions on well permitting and fracturing are assumed
beyond restrictions that are currently in place.
Data sources: Conventional resource base assessment is based on data from the U.S. Geological Survey
(USGS), Minerals Management Service (MMS), and Canadian Gas Potential Committee (CGPC) using
ICF's Hydrocarbon Supply Model (HSM).
In the area of unconventional gas, ICF has worked for many years with the Gas Research Institute
(GRI)/Gas Technology Institute (GTI) to develop a database of tight gas, coalbed methane, and Devonian
Shale reservoirs in the U.S. and Canada. Along with USGS assessments of continuous plays, the
database was used to help develop the HSM's "cells", which represent resources in a specific geographic
area, characterizing the unconventional resource in each basin, historical unconventional reserves
estimates and typical decline curves. ICF has recently revised the unconventional gas resource
assessments based on new gas industry information on the geology, well production characteristics, and
costs. The new assessments include major shale units such as the Fort Worth Barnett Shale, the
Marcellus Shale, the Haynesville Shale, and Western Canada shale plays. ICF has built up a database
on gas compositions in the United States and has merged that data with production data to allow the
analysis of net versus raw gas production.
Resources are divided into three general categories: new fields/new pools, field appreciation, and
unconventional gas. The methodology for resource characterization and economic evaluation differs for
each.
New Fields
Conventional new discoveries are characterized by size class. For the United States, the number of
fields within a size class is broken down into oil fields, high permeability gas fields, and low permeability
gas fields based on the expected occurrence of each type of field within the region and interval being
modeled. The fields are characterized further as having a hydrocarbon make-up containing a certain
percent each of crude oil, dry natural gas, and natural gas liquids. In Canada, fields are oil, sweet
nonassociated gas, or sour nonassociated gas.
The methodology uses a modified "Arps-Roberts" equation to estimate the rate at which new fields are
discovered. The fundamental theory behind the find-rate methodology is that the probability of finding a
field is proportional to the field's size as measured by its areal extent, which is highly correlated to the
field's level of reserves. For this reason, larger fields tend to be found earlier in the discovery process
than smaller fields. The new equation developed by ICF accurately tracks discovery rates for mid- to
small-size fields. Since these are the only fields left to be discovered in many mature areas, the more
accurate find-rate representation is an important component in analyzing the economics of exploration
activity in these areas.
An economic evaluation is made in the model each year for potential new field exploration programs
using a standard discounted after-tax discounted cash flow (DCF) analysis. This DCF analysis takes into
account how many fields of each type are expected to be found and economics of developing each. The
economic decision to develop a field is made using "sunk cost" economics where the discovery cost is
ignored and only time-forward development costs and production revenues are considered. However, the
model's decision to begin an exploration program includes all exploration and development costs.
Field Appreciation
Field appreciation refers to potential resources that can be proved from already discovered fields. These
inventories are referred to as appreciation, growth-to-known or "probables." The inventories of probables
are increased due to expected future appreciation due to many factors that include higher recovery
percentages of the gas in-place resulting from infill drilling and application of improved technology and
experience gained in the course of developing and operating the field.
8-7
-------
Unconventional Gas
The ICF assessment method for shale gas is a "bottom-up" approach that first generates estimates of
unrisked and risked gas-in-place (GIP) from maps of depth, thickness, organic content, and thermal
maturity. Then, ICF uses a different model to estimate well recoveries and production profiles. Unrisked
GIP is the amount of original gas-in-place determined to be present based upon geological factors—
without risk reductions. "Risked GIP" includes a factor to reduce the total gas volume based on proximity
to existing production and geologic factors such as net thickness (e.g., remote areas, thinner areas, and
areas of high thermal maturity have higher risk). ICF calibrates expected well recoveries with specific
geological settings to actual well recoveries by using a rigorous method of analysis of historical well data.
To estimate the contributions of changing technologies ICF employs the "learning curve" concept used in
several industries. The "learning curve" describes the aggregate influence of learning and new
technologies as having a certain percent effect on a key productivity measure (for example cost per unit
of output or feet drilled per rig per day) for each doubling of cumulative output volume or other measure of
industry/technology maturity. The learning curve shows that advances are rapid (measured as percent
improvement per period of time) in the early stages when industries or technologies are immature and
that those advances decline through time as the industry or technology matures. Generally speaking, we
find the learning curve effect is roughly 20 percent per doubling of cumulative wells.
Major Unconventional Natural Gas Categories
Definition of Unconventional Gas: Quantities of natural gas that occur in continuous,
widespread accumulations in low quality reservoir rocks (including low permeability or tight
gas, coalbed methane, and shale gas), that are produced through wellbores but require
advanced technologies or procedures for economic production.
Tight Gas is defined as natural gas from gas-bearing sandstones or carbonates with an in situ
permeability (flow rate capability) to gas of less than 0.1 millidarcy. Many tight gas sands have
in situ permeability as low as 0.001 millidarcy. Wells are typically vertical or directional and
require artificial stimulation.
Coalbed Methane is defined as natural gas produced from coal seams. The coal acts as both
the source and reservoir for the methane. Wells are typically vertical but can be horizontal.
Some coals are wet and require water removal to produce the gas, while others are dry.
Shale Gas is defined as natural gas from shale formations. The shale acts as both the source
and reservoir for the methane. Older shale gas wells were vertical while more recent wells are
primarily horizontal with artificial stimulation. Only shale formations with certain characteristics
will produce gas.
Shale Oil with Associated Gas is defined as associated gas from oil shale in horizontal
drilling plays such as the Bakken in the Williston Basin. The gas is produced through
boreholes along with the oil.
8.3.2 Upstream Cost and Technology Factors
In ICF's methodology, supply technology advancements effects are represented in three categories:
• Improved exploratory success rates
• Cost reductions of platform, drilling, and other components
• Improved recovery per well
8-8
-------
These factors are included in the model by region and type of gas and represent several dozen actual
model parameters. ICF's database contains base year cost for wells, platforms, operations and
maintenance, and other relevant cost items.
Figure 8-6 GMM U.S. and Canada Projected Gas Production by Source
160
140
120
p
o
CD
&
/ / / /
¦ Conventional Onshore
¦ Coalbed Methane aTight
¦ Offshore ¦ Shale
8.3.3 Historical Gas Production
ICF consistently updates the production and resource data that it used for EPA Platform v6. The
historical production data for the model comes from PointLogic. An example comparison of production
forecasts for the San Juan and Raton basins is shown in Figure 8-7.
8-9
-------
Figure 8-7 Production Comparison for San Juan and Raton Basins
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
CO
LO
h-
CD
T—
CO
LO
N-
CD
T—
CO
LO
h-
O)
T—
CO
LO
h-
O)
T—
T—
T—
t—
CM
CM
CM
CM
CM
CO
CO
CO
CO
CO
'=t
o
o
o
O
o
O
O
O
O
o
o
o
o
o
o
o
o
o
o
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
CM
c
C
C
c
C
C
C
£=
£=
C
C
C
C
C
C
C
C
C
c
03
03
03
cc
CO
03
03
03
03
03
OS
03
03
03
03
03
03
03
03
-3
—>
—>
—>
~5
—>
—>
—)
~5
~3
~3
—>
-PointLogic Historical Data EPA Base Case 2017
8.3.4 Treatment of Frontier Resources and Exports
Arctic Projects
GMM does not have resources located in frontier regions. Arctic projects (specifically Alaska and
Mackenzie Valley gas pipelines) are not included in our projection.
Existing and Potential Liquefied Natural Gas (LNG) Terminals
LNG is natural gas that has been transformed to a liquid by super-cooling it to minus 260 degrees
Fahrenheit, reducing its volume by a factor of 600. LNG is then shipped on board special carriers, and
the process is reversed at a receiving facility with the re-gasified product delivered via pipeline. Based on
current global LNG market conditions, ICF assumes that the six U.S. LNG terminals currently under
construction are completed and expanded in future. Those terminals are Sabine Pass, Freeport, Cove
Point, Cameron, Corpus Christi, and Elba Island. By 2020, ICF projects U.S. LNG export capacity will be
10.3 billion cubic feet per day (Bcfd). Given the near-term low oil price expectations, we project that
North American export terminal capacity utilization will average about 63% through 2020. U.S. export
volumes are projected to approach six Bcfd by 2020. ICF projects that the capacity utilization will
increase to over 80% by 2024. In addition to the U.S. trains currently under construction, ICF assumes
an additional 5.8 Bcfd of export capacity will come online in North America between 2020 and 2035. ICF
assumes that only one export facility will be built in British Columbia: Woodfibre (0.3 Bcfd).
8-10
-------
Figure 8-8 Existing and Proposed Marine LNG Terminals as of May 2017
US Jurisdiction
© FERC
£ MARAD/USCG
North American LNG Import/Export Terminals
Approved
Import Terminals
U.S.
APPROVED ¦ UNDER CONSTRUCTION - FERC
1. Corpus Christi, TX: 0.4 Bcfd (Cheniere - Corpus Christi LNG) (CP12-507)
APPROVED - NOT UNDER CONSTRUCTION - FERC
2. Salinas. PR: 0.6 Bcfd (Aguirre Offshore GasPort. LLC) (CP13-193)
APPROVED • NOT UNDER CONSTRUCTION - MARAD/Coast Guard
3. Gulf of Mexico: 1.0 Bcfd (Main Pass McMoRan Exp.)
4. Gulf of Mexico: 1.4 Bcfd (TORP Technology-Bienville LNG)
Export Terminals
U.S.
APPROVED ¦ UNDER CONSTRUCTION - FERC
5. Sabine, LA: 0.7 Bcfd (Cheniere/Sabine Pass LNG) (CP11-72 & CP14-12)
6. Hackberry, LA: 2.1 Bcfd (Sempra-Cameron LNG) (CP13-25)
7. Freeport, TX: 2.14 Bcfd (Freeport LNG Dev/Freeport LNG Expansion/FLNG
Liquefaction) (CP12-509) (CP15-518)
8. Cove Point. MD: 0.82 Bcfd (Dominion-Cove Point LNG) (CP13-113)
9. Corpus Christi, TX: 2.14 Bcfd (Cheniere - Corpus Christi LNG) (CP12-507)
10. Sabine Pass. LA: 1.40 Bcfd (Sabine Pass Liquefaction) (CP13-552) ~
11. Elba Island, GA: 0.35 Bcfd (Southern LNG Company) (CP14-103)
APPROVED - NOT UNDER CONSTRUCTION - FERC
12. Lake Charles. LA: 2.2 Bcfd (Southern Union - Lake Charles LNG) (CP14-120)
13. Lake Charles. LA: 1.08 Bcfd (Magnolia LNG) (CP14-347)
14. Hackberry, LA: 141 Bcfd (Sempra - Cameron LNG) (CP15-560)
15. Sabine Pass. TX: 2.1 Bcfd (ExxonMobil - Golden Pass) (CP14-517)
Canada
APPROVED - NOT UNDER CONSTRUCTION
16. Port Hawkesbury. NS: 0.5 Bcfd (Bear Head LNG)
17. Kitimat, BC: 3.23 Bcfd (LNG Canada)
18. Squamish, BC: 0.29 Bcfd (Woodfibre LNG Ltd)
19. Prince Rupert Island. BC: 2.74 Bcfd (Pacific Northwest LNG)
* Trains 5 & 6 with Train 5 under construction
As of May 1, 2017
Figure 8-9 LNG Export Volumes versus Capacity
18
16
14
12
1 10
-------
Pipeline Exports to Mexico
Mexico's demand for natural gas continues to rise, while its domestic production has been declining
Since 2010, Mexico's imports of U.S. gas have gone up over 300%, reaching 3.9 Bcfd in 2016. As
Mexico continues to add gas-fired generation and sponsor new pipelines from the U.S., exports will
continue to grow. ICF projects that exports will reach 8 Bcfd by 2035.
Figure 8-10 U.S. Pipeline Exports to Mexico
¦ California "WestTexas/New Mexico "Arizona BSouthTexas
8.4 Oil Prices
Natural gas prices and LNG export levels are forecasted by taking into account oil prices. The following
section contains discussions about the crude oil price assumptions used for EPA Platform v6.
ICF uses the Refiner Acquisition Cost of Crude Oil (RACC) price as an oil price input to GMM. The
RACC price is a term commonly used in discussing crude oil. It is the cost of crude oil to the refiner,
including transportation and fees. For the long-term, ICF projects a slow recovery in oil prices to an
equilibrium marginal production cost of $75/bbl (in $2016). The residual oil price averages between 70
and 100 percent of the RACC price on a dollar per Btu basis. This is the price used to determine
switching in the industrial sector. Figure 8-11 shows the ICF RACC price projection.
8-12
-------
Figure 8-11 Refiners' Acquisition Cost of Crude (RACC)
8.5 Demand Assumptions
Gas demand is calculated by sets of algorithms and equations for each sector and region. The model
calculates monthly "real-time consumption", not "billed volumes." Demand reported by DOE/EIA
represents billed volumes, which is time-lagged. Recent DOE/EIA and Statistics Canada data have been
considered in the calibration of the model. However, the historical data represents ICF's backcast of the
market. ICF performs market reconnaissance and data analysis each month to support the GMM
calibration. GMM models natural gas demand in four end-use sectors: residential, commercial, industrial,
and power generation.
Residential/Commercial gas demand calculated from regional equations fit econometrically to weather,
economic growth, and price elasticity. The 41 regions for which residential/commercial demand are
calculated are shown in Figure 8-12. Regression analysis was separately completed for each sector for
34 Lower-48 and 7 Canadian regions. The seventh Canadian region, Atlantic Provinces, has no historical
gas use. An Alaska region is included in the Model, but all Alaska end use gas demand is input
exogenously.
8-13
-------
Figure 8-12 GMM Residential/Commercial Gas Demand Regions
low Englani
Industrial gas demand is based on a detailed breakout of industrial activity by census region. The U.S.
is divided into 11 regions based on census region boundaries. The model includes ten industry sectors,
focusing on gas-intensive industries. Those 10 industries are:
• Food
• Pulp and Paper
• Petroleum Refining
• Chemicals
• Stone, Clay, and Glass
• Iron and Steel
• Primary Aluminum
• Other Primary Metals
• Other Manufacturing
• Non-Manufacturing
Three end-use categories (Process Heat, Boilers, and Other End Uses) are modeled separately for each
sector.
• Process heat: This includes all uses of gas for direct heating as opposed to indirect heating (e.g.,
steam production). The GMM econometric modeling indicated that forecasts for process heat for
each industrial sector are a function of growth in output, the energy intensity trend, and the price
elasticity. Growth in output overtime for most industries is controlled by industrial production
indices. Energy intensity is a measure of the amount of gas consumed per unit of output. Energy
intensity tends to decrease overtime as industries become more efficient.
8-14
-------
• Boilers: This category includes natural gas-fired boilers whose purpose is to meet industrial
steam demand. GMM econometric models indicated that gas demand for boilers is a function of
the growth in industrial output and the amount of gas-to-oil switching. Industry steam
requirements grow based on industrial production growth. A large percentage of the nominally
"dual-fired" boilers cannot switch due to environmental and technical constraints.
• Other end uses: This category includes ail other uses for gas, including non-boiler cogeneration,
on-site electricity generation, and space heating. Like the forecasts for process heat, the GMM
econometric modeling showed "other end uses" for each industrial sector to be a function of
growth in output, the energy intensity trend, and the price elasticity.
The chemicals sector also includes feedstock demands for ammonia, methanol, and non-refinery
hydrogen. Canada is divided into 6 regions based on provincial boundaries. The approach for Canada is
a regression fit of historic data similar to that used in the residential/commercial sectors - sub-sectors of
Canadian industrial demand are not modeled separately. The Canadian industrial sector also includes
power generation gas demand. The Atlantic Provinces in Eastern Canada have no historical industrial
gas demand.
Energy intensity and price elasticity inputs are based on Industrial Sector Technology Use Model (ISTUM-
2). Boiler switching curves are defined from work for GRI (now GTI). The GMM captures both near-term
price-induced switching and "demand destruction" effects of high gas prices.
Figure 8-13 GMM Industrial Gas Demand Regions
Power generation demand in the GMM is modeled for 13 dispatch regions for the contiguous U.S. All of
the power sector inputs in GMM are changed to be consistent with IPM results overtime. Most
importantly, the total gas use regionally is bench-marked against IPM's gas use.
8-15
-------
Figure 8-14 GMM Power Generation Gas Demand Regions
New /
England
WECC
Northwest
MAPP
MAAC
ECAR
WECC
CA/NV
WECC
Rockies
SPP
SERC
ERCOT
Pipeline fuel consumption is the gas consumed in the operation of pipelines, primarily in compressors, as
well as pipeline losses. Pipeline gas-use is a function of the fuel rate and the volume of gas moved on
each pipeline corridor. Pipeline gas-use is estimated as a percent of natural gas throughput for each link
in the pipeline network. Pipeline gas-use is allocated evenly between the upstream and downstream
nodes for each link. Historical pipeline gas-use is derived from EIA data on a state by state basis, and
then mapped to each link in the pipeline network.
Lease & Plant gas use represents natural gas used in well, field, and lease operations (such as gas used
in drilling operations, heaters, dehydrators, and field compressors) and used as fuel in natural gas
processing plants. The lease and plant gas-use is forecast based on historical percentages of the dry
gas produced at each node. Regional factors determine the share of lease & plant gas use for each
supply region.
8-16
-------
Figure 8-15 GMM U.S. and Canada Gas Demand Projection
140
120
¦ Other ¦ Residential ¦ Commercial ¦ Power ¦ Industrial
Note: "Other" includes pipeline fuel and lease & plant
There are four key drivers for natural gas demand in GMM. They are:
i) Macroeconomic parameters: From 2018 forward, ICF assumes U.S. GDP grows at 2.1% per
year, and Canada GDP grows at 2.0% per year.84
ii) Electric Demand Growth: Electric demand growth rate is assumed to be 0.68% per year
consistent with EPA Platform v6.
iii) Demographics: Projected demographic trends are consistent with trends over the past 20 years.
U.S. population growth averages about 1% per year throughout our projection.
iv) Weather: Future weather is assumed consistent with regional and monthly average heating and
cooling degree days (HDD/CDD) over the past 20 years (1997 through 2016).
8.6 Discussion of GMM Results Underlying the Natural Gas Supply Curves85
In this section, we describe GMM results underlying the natural gas supply curves for EPA Platform v6. A
typical GMM run generates the following outputs:
• Natural gas prices
• Natural gas production by region
• Natural gas consumption by region and sector
84 The U.S. Congressional Budget Office assumes an average annual GDP growth rate of 1.9% between 2018 and
2028, while the 2018 U.S. Energy Information Administration Annual Energy Outlook used an average annual GDP
growth rate of 2.0% between 2018 and 2050.
85 The GMM results presented in this section are illustrative and consistent with a draft version of the EPA Platform v6
November 2018 Reference Case. GMM was not rerun for a final calibration with EPA Platform v6 November 2018
Reference Case using IPM.
8-17
-------
Table 8-1 summarizes the supply/demand balance and Henry Hub price for a GMM run underlying the
natural gas supply curves. The regional breakout in the supply/demand data is by census region and the
mapping to the state and GMM nodes is provided in Figure 8-16 and Figure 8-17. Table 8-3 provides
additional results.
Table 8-1 Supply/Demand Balance and Henry Hub Price for a GMM Run Underlying the Natural
Gas Supply Curves in EPA Platform v6
Demand (Bcf per year)
2017
2021
2023
2025
2030
2035
2040
2045
2050
New England
849
939
952
987
1,009
1,034
1,063
1,066
1,092
Mid-Atlantic
3,747
4,471
4,776
4,883
5,286
5,437
5,457
5,645
5,849
East North Central
3,852
4,413
4,442
4,498
4,747
5,160
5,385
5,508
5,723
West North Central
1,782
1,935
1,937
1,948
1,959
2,027
2,034
2,007
2,015
South Atlantic
3,694
4,522
4,536
4,528
4,917
5,294
5,630
5,905
6,192
East South Central
1,705
2,076
2,077
2,115
2,296
2,353
2,427
2,448
2,507
West South Central
6,322
6,974
7,053
7,047
7,334
7,502
7,604
7,641
7,774
Mountain
1,778
1,887
1,947
2,021
2,003
2,167
2,307
2,378
2,451
Pacific (contiguous)
2,888
2,815
2,794
2,758
2,534
2,512
2,559
2,645
2,679
Alaska
347
304
303
301
296
295
295
295
295
Total L-48
26,619
30,033
30,514
30,785
32,084
33,486
34,465
35,242
36,281
Total United States
26,965
30,337
30,817
31,086
32,380
33,780
34,760
35,537
36,576
Exports/Imports (Bcf per year)
Net LNG Exports from US
593
2,698
3,343
3,990
5,065
5,064
5,164
5,160
5,159
Net Pipeline Exports to Mexico
1,589
1,967
2,142
2,316
2,632
2,945
2,928
2,858
2,911
Net Pipeline Imports from Canada
1,913
1,470
1,481
1,694
2,135
2,371
2,285
2,278
1,942
Supply (Bcf per year)
New England
0
0
0
0
0
0
0
0
0
Mid-Atlantic
5,229
8,845
9,762
10,225
11,388
12,128
12,608
13,284
13,915
East North Central
1,595
2,780
3,108
3,341
3,812
4,133
4,328
4,520
4,683
West North Central
1,140
1,097
1,047
1,005
942
874
798
758
753
South Atlantic
1,559
2,288
2,534
2,751
3,173
3,433
3,591
3,755
3,890
East South Central
795
861
816
779
753
802
806
866
927
West South Central
11,869
12,406
12,189
12,178
12,536
13,067
13,168
13,732
14,423
Mountain
4,402
4,500
4,389
4,315
4,301
4,649
5,119
5,645
5,838
Pacific (contiguous)
196
174
175
178
181
183
178
170
163
Alaska
311
303
302
307
309
313
303
293
285
Total L-48
26,788
32,951
34,020
34,772
37,087
39,269
40,596
42,731
44,593
Total United States
27,099
33,254
34,322
35,079
37,395
39,582
40,899
43,024
44,878
2017
2021
2023
2025
2030
2035
2040
2045
2050
Henry Hub, Nom$/MMBtu
2.82
3.14
3.61
3.77
4.38
4.60
5.76
7.33
8.18
8-18
-------
Figure 8-16 Demand Region Definition
New
England
Mountain
West North
Central
East South
Central
West South
Central
Cove pi
' LNG
Atlantic
Offshore
Chartts LNG
Reynosa
South
Atlantic
Copyright 2018, ICF
West Coast
Rocky Mountain
Midcontinent
Northeast
Southwest
Everett
r LNG
Atlanta
Offshore
California
Offshore
Copyright 2018, ICF
8-19
-------
8.6.1 Supply Curves for EPA Platform v6
Henry Hub is a pipeline interchange hub in Louisiana Gulf Coast near Erath, LA, where eight interstate
and three intrastate pipelines interconnect. Liquidity at this point is very high and it serves as the primary
point of exchange for the New York Mercantile Exchange (NYMEX) active natural gas futures markets.
Henry Hub prices are considered as a proxy for U.S. natural gas prices. Natural gas from the Gulf moves
through the Henry Hub onto long-haul interstate pipelines serving demand centers. Due to the
importance and significance of Henry Hub, GMM generated supply curves are specified at Henry Hub
prices.
For IPM modeling, GMM generates a price forecast over a time horizon and a set of time dependent
price/supply curves based on that price path for each year in the forecast. For each year, the mid-point
price of the supply curve is set equal to the solved Henry Hub price from GMM and the mid-point volume
is set equal to the solved gas consumption for the power sector from GMM. Each supply curve's elasticity
is set equal to the effective price-elasticity for gas supply in that year. In this manner, even while GMM
has itself projected particular levels of gas supply and consumption (and corresponding market-clearing
prices) overtime, the information included in those projections is input into IPM in the form of gas supply
curves that enable IPM to solve for levels of power sector gas consumption and resulting gas prices that
respect a least-cost power production future.
The final resulting supply curves developed for years 2021, 2023, 2025, 2030, 2035, 2040, 2045, and
2050 are shown in Figure 8-18 and Table 8-5. In the very short-term, gas supply is price inelastic
because there are few years to respond to the market changes. Overtime, gas supply becomes more
price elastic because producers have more time to respond to the market changes. Thus, the supply
curves are much more price elastic by 2025. In the longer term, resource depletion tends to offset
elasticity making the curves slightly less elastic than they are between 2025 and 2030.
Figure 8-18 Supply Curves for 2021, 2023, 2025, 2030, 2035, 2040, 2045, and 2050
$7.00
$6.00
$5.00
$4.00
3
£
to
$3.00
o
CM
0)
o
CL
$2.00
$1.00
$0.00
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Quantity (Quads)
8-20
-------
8.6.2 Basis
Basis is the difference in gas price in a given market from the widely used Henry Hub reference price.
Basis reflects the price in a given market based on demand, available supply, and the cost of transporting
gas to that location. A negative basis value represents that the gas price in that area is lower than the
Henry Hub price. Basis between two nodes in GMM is the difference in prices between the two nodes.
The GMM utilizes its network of 121 nodes that comprises 423 gas pipeline corridors to assess the basis
between two desired nodes. The pipeline corridors between nodes are represented by pipeline links and
can be characterized by their maximum capacity. Each of the links has an associated discount curve
(derived from GMM natural gas transportation module), which represents the marginal value of gas
transmission on that pipeline segment as a function of the pipeline's load factor. The basis value is
calculated by using the supply/demand balance in two nodes along with the resulting prices in each node
and the cost of transporting gas between the two nodes as determined by the discount curve on that link.
The discount curve is a function of the pipeline tariffs and the load factor. The discount curves are
continuously calibrated to accurately reflect historical basis values. Their parameters can be adjusted to
account for regulatory changes that can affect pipeline values.
The GMM solves for basis monthly. Basis pressure (i.e., spiking basis) will generally occur when average
monthly load factors rise to above 80%. Since many U.S. markets are winter peaking, the higher basis
typically occurs in the winter months when gas use and load factors are highest. The IPM relies on
seasonal basis that reflects averages of the monthly basis values solved for in the GMM.
GMM is not only used to estimate the gas supply curves, but also used to estimate the relationship of gas
price at Henry Hub to gas prices elsewhere in the country. IPM uses these gas supply curves and
regional price relationships (differentials) overtime as inputs, based on GMM-projected future of gas
supply/demand. While EPA's Platform v6 has the flexibility to re-determine the relationship of power
sector gas demand to supply and to accordingly find different gas price futures, EPA's Platform v6 will
maintain the future (basis differential) price relationship between Henry Hub and each regional location in
a national supply picture as originally determined by these GMM projections. Table 8-4 provides the full
set of seasonal basis differentials at the IPM region level.
8.6.3 Delivered Price Adders
As stated in section 8.2, GMM prices are market center prices and not delivered prices. In order to
estimate delivered prices at a power plant, an adder is applied to the seasonal basis from GMM. ICF
calculated this delivered price adder for each state by comparing its GMM historical prices with historical
delivered gas prices to electric power plants based on EIA-176 data. The delivered price adders
implemented in EPA Platform v6 are shown in Table 8-2.
8-21
-------
Table 8-2 Delivered Price Adders
State
Adder (2016$/MMBtu)
State
Adder (2016$/MMBtu)
Alabama
0.01
Nebraska
0.46
Alaska
0.95
Nevada
0.21
Arizona
0.03
New Hampshire
0.01
Arkansas
0.13
New Jersey
0.26
California
0.16
New Mexico
0.02
Colorado
0.18
New York
0.17
Connecticut
0.05
North Carolina
0.24
Delaware
0.01
North Dakota
0.09
Florida
0.02
Ohio
0.03
Georgia
0.00
Oklahoma
0.03
Idaho
0.05
Oregon
0.01
Illinois
0.16
Pennsylvania
0.04
Indiana
0.14
Rhode Island
0.00
Iowa
0.26
South Carolina
0.17
Kansas
0.13
South Dakota
0.01
Kentucky
0.23
Tennessee
0.05
Louisiana
0.04
Texas
0.19
Maine
0.03
Utah
0.08
Maryland
0.13
Virginia
0.06
Massachusetts
0.03
Washington
0.10
Michigan
0.16
West Virginia
0.13
Minnesota
0.35
Wisconsin
0.09
Mississippi
0.03
Wyoming
0.06
Missouri
0.12
US
0.13
Montana
0.44
Canada
0.13
List of tables that are uploaded directly to the web:
Table 8-3 El A Style Gas Report for EPA Platform v6
Table 8-4 Natural Gas Basis for EPA Platform v6
Table 8-5 Natural Gas Supply Curves for EPA Platform v6
8-22
-------
9. Other Fuels and Fuel Emission Factor Assumptions
Besides coal (Chapter 7) and natural gas (Chapter 8), EPA Platform v6 also includes assumptions for
residual fuel oil, distillate fuel oil, biomass, nuclear, and waste fuels. This chapter describes the
assumptions pertaining to characteristics, market structures, and prices of these other fuels. As reported
in previous chapters, natural gas is represented by an exogenous supply curve along with a basis
differential approach informed by a resource fundamentals model. Coal is represented by a robust set of
supply curves and a detailed representation of the associated coal transport network. Together they are
designed to capture the intricacies of the resource base and market for these fuels which accounted for
about 64% of U.S. electric generation in 201686. As with coal, the price and quantity of biomass
combusted is determined by balancing supply and demand using a set of geographically differentiated
supply curves. In contrast, fuel oil, nuclear, and waste fuel prices are exogenously determined and
entered into IPM during model set-up as constant price points that apply to all levels of supply. The
following treats each of these remaining fuels in turn and concludes with a discussion of the emission
factors for all the fuels represented in EPA Platform v6.
9.1 Fuel Oil
Two petroleum derived fuels are included in EPA Platform v6. Distillate fuel oil is distilled from crude oil,
and residual fuel oil is a residue of the distillation process. The fuel oil prices are from AEO 2017 and are
shown in Table 9-1. They are regionally differentiated according to the NEMS (National Energy Modeling
System) regions used in AEO 2017. These prices are mapped to their corresponding IPM regions for use
in EPA Platform v6.
Table 9-1 Fuel Oil Prices by NEMS Region in EPA Platform v6
Residual Fuel Oil Prices (2016$/MMBtu)
AEO NEMS Region
2021
2023
2025
2030
2035
2040
2045
2050
ERCT
12.86
13.56
14.22
15.34
16.50
17.69
18.02
18.93
FRCC
11.35
12.05
11.26
12.40
13.45
14.48
14.81
15.71
MROE
11.32
11.94
12.60
13.70
14.79
15.79
16.12
16.85
MROW
4.30
5.00
5.65
6.77
7.94
9.13
9.46
10.37
NEWE
11.75
11.94
12.60
13.72
14.88
16.08
16.40
17.31
NYCW
14.19
14.89
15.54
16.66
17.82
19.02
19.35
20.25
NYLI
11.03
11.01
11.67
12.77
13.86
14.86
15.19
15.92
NYUP
9.60
10.30
10.95
12.07
13.23
14.43
14.76
15.66
RFCE
10.47
10.60
11.26
12.36
13.45
14.45
14.87
15.78
RFCM
9.55
10.25
10.90
12.02
13.18
14.38
14.71
15.61
RFCW
10.67
11.37
12.03
13.14
14.31
15.50
15.83
16.74
SRDA
12.16
12.86
13.51
14.63
15.80
16.99
17.32
18.23
SRGW
8.51
9.20
9.86
10.98
12.14
13.34
13.66
14.57
SRSE
9.59
10.28
10.94
12.06
13.22
14.42
14.74
15.65
SRCE
8.58
9.28
9.94
11.05
12.22
13.41
13.74
14.65
SRVC
10.47
10.60
11.26
12.37
13.54
14.73
15.06
15.97
SPNO
8.58
9.28
9.94
11.05
12.22
13.41
13.74
14.65
SPSO
10.83
11.52
12.18
13.30
14.46
15.66
15.99
16.89
AZNM
11.48
12.18
12.84
13.95
15.12
16.31
16.64
17.55
86 EIA. Detailed EIA-923 monthly and annual survey data back to 1990. Available at
https://vwvw.eia.gOv/electricitv/data.php#aeneration
9-1
-------
Residual Fuel Oil Prices (2016$/MMBtu)
AEO NEMS Region
2021
2023
2025
2030
2035
2040
2045
2050
CAMX
11.26
11.95
12.61
13.73
14.89
16.09
16.41
17.32
NWPP
10.43
11.94
12.60
13.70
14.79
15.79
16.12
16.85
RMPA
7.91
8.61
9.27
10.38
11.55
12.74
13.07
13.98
Distillate Fuel Oil Prices (2016$/MMBtu)
NEMS Region
2021
2023
2025
2030
2035
2040
2045
2050
ERCT
16.57
16.85
17.65
18.96
20.26
21.58
21.95
22.89
FRCC
18.58
19.39
20.17
21.46
22.73
24.12
24.45
25.39
MROE
17.61
18.08
18.88
20.19
21.49
22.81
23.18
24.11
MROW
17.00
17.27
18.07
19.38
20.68
22.00
22.37
23.31
NEWE
16.98
17.38
18.16
19.46
20.72
22.11
22.44
23.38
NYCW
19.92
21.02
21.80
23.09
24.35
25.75
26.07
27.02
NYLI
19.92
21.02
21.80
23.09
24.35
25.75
26.07
27.02
NYUP
19.92
21.02
21.80
23.09
24.35
25.75
26.07
27.02
RFCE
19.57
20.55
21.33
22.59
23.86
25.23
25.56
26.49
RFCM
17.61
18.08
18.88
20.19
21.49
22.81
23.18
24.11
RFCW
17.96
18.54
19.33
20.63
21.92
23.26
23.62
24.56
SRDA
16.57
16.85
17.65
18.96
20.26
21.58
21.95
22.89
SRGW
17.31
17.68
18.48
19.80
21.11
22.48
22.86
23.80
SRSE
17.77
18.37
19.17
20.24
21.53
22.93
23.31
24.23
SRCE
16.65
16.89
17.69
19.00
20.31
21.63
22.00
22.93
SRVC
18.58
19.39
20.17
21.46
22.73
24.12
24.45
25.39
SPNO
16.98
17.26
18.05
19.36
20.66
21.99
22.36
23.29
SPSO
16.61
16.88
17.68
18.99
20.29
21.62
21.99
22.92
AZNM
19.25
20.11
20.90
22.21
23.52
24.84
25.27
26.21
CAMX
19.18
19.95
19.83
21.15
22.45
23.77
24.20
25.14
NWPP
19.18
20.01
20.81
22.22
23.51
24.84
25.27
26.21
RMPA
19.25
20.11
20.90
22.22
23.52
24.84
25.27
26.21
9.2 Biomass Fuel
Biomass is offered as a fuel for existing dedicated biomass power plants and potential (new) biomass
direct fired boilers. In addition to its use as the prime mover fuel for these plants, it is also offered for co-
firing to those coal fired plants that have co-fired biomass in the recent past. Section 5.3 provides further
details of these selected coal plants.
EPA Platform v6 uses biomass supply curves based on those in the Department of Energy's 2016 Billion-
Ton Report (DOE Report). Biomass supply curves at the IPM region and state level are generated by
aggregating county level supply curves from the DOE Report. Power plants demand biomass from the
supply curve corresponding to the IPM region and state in which they are located. No inter-region trading
of biomass is allowed. Each biomass supply curve depicts the price-quantity relationship for biomass and
varies overtime. There is a separate curve for each model run year. The supply component of the curve
represents the aggregate supply in each region of agricultural residues, forestry residues, energy crops,
waste, and trees. The price component of the curve includes transportation costs of $15 per dry ton.
The supply curves represent the IPM region and state-specific delivered biomass fuel cost at the plant
9-2
-------
gate. A storage cost of $20 per dry ton is added to each step of the agricultural residue supply curves to
reflect the limited agricultural growing season87. The biomass supply curves are summarized in Table
9-4. The biomass prices are derived endogenously based on the aggregate power sector demand for
biomass in each IPM region and state. The results are unique market-clearing prices for each IPM region
and state. All plants using biomass from that IPM region and state face the same market-clearing price.
9.3 Nuclear Fuel
The AEO 2018 price for nuclear fuel is used as the nuclear fuel price assumption for 2021 -2050 in EPA
Platform v6. The 2021, 2023, 2025, 2030, 2035, 2040, 2045, and 2050 prices are 0.64, 0.64, 0.65, 0.65,
0.66, 0.67, 0.68, and 0.69 2016 $/MMBtu, respectively.
9.4 Waste Fuels
The waste fuels include waste coal, petroleum coke, fossil waste, non-fossil waste, tires, and municipal
solid waste (MSW). Table 9-2 describes the characteristics of these fuels, the extent to which they are
represented in NEEDS, and the assumptions pertaining to their use and pricing. Furthermore, the fuels
are provided to only existing and planned committed units. Potential (new) generating units that the
model "builds" are not given the option to burn these fuels. In IPM model output, tires, MSW, and non-
fossil waste are included under existing non-fossil other, while waste coal and petroleum coke are
included under coal.
Table 9-2 Waste Fuels in NEEDS v6 and EPA Platform v6
Modeled
Fuel in
NEEDS
Number
of Units
in
NEEDS
Total
Capacity
in
NEEDS
Description
Supply and Cost
Modeled
By
Assumed
Price
Waste
Coal
22
1,435
MW
"Usable material that is a byproduct of previous coal
processing operations. Waste coal is usually
composed of mixed coal, soil, and rock (mine
waste). Most waste coal is burned as-is in
unconventional fluidized-bed combustors. For
some uses, waste coal may be partially cleaned by
removing some extraneous noncombustible
constituents. Examples of waste coal include fine
coal, coal obtained from a refuse bank or slurry
dam, anthracite culm, bituminous gob, and lignite
waste."
https://www.eia. qov/tools/qlossarv/index.php?id=W
Supply
Curve
Based on
AEO
2017
AEO 2017
Petroleum
Coke
14
1,213
MW
A residual product, high in carbon content and low
in hydrogen, from the cracking process used in
crude oil refining.
Price
Point
$42.60/Ton
Fossil
Waste
81
1,049
MW
Waste products of petroleum or natural gas
including blast furnace and coke oven gas. They do
not include petroleum coke or waste coal which are
specified separately among the "Modeled Fuels".
Price
Point
0
87 http://www.extension.iastate.edu/aadm/crops/pdf/a1-22.pdf,
http://www.rand.org/content/dam/rand/pubs/technical reports/2011/RAND TR876.pdf
9-3
-------
Modeled
Fuel in
NEEDS
Number
of Units
in
NEEDS
Total
Capacity
in
NEEDS
Description
Supply and Cost
Modeled
By
Assumed
Price
Non-Fossil
Waste
221
2,071
MW
Non-fossil waste products that do not qualify as
biomass. These include waste products of liquid
and gaseous renewable fuels (e.g., red and black
liquor from pulping processes, digester gases from
waste water treatment). They do not include urban
wood waste which is included in biomass.
Price
Point
0
Tires
2
52 MW
Discarded vehicle tires.
Price
Point
0
Municipal
Solid
Waste
165
2,123
MW
"Residential solid waste and some nonhazardous
commercial, institutional, and industrial wastes."
httDs://www.eia.aov/tools/alossarv/index.DhD?id=M
Price
Point
0
9.5 Fuel Emission Factors
Table 9-3 brings together all the fuel emission factor assumptions implemented in EPA Platform v6. For
sulfur dioxide, chlorine, and mercury in coal, where emission factors vary widely based on the rank,
grade, and supply source of the coal, cross references are given to tables that provide more detailed
treatment of the topic. Nitrogen oxides (NOx) are not included in Table 9-3 because NOx emissions are a
factor of the combustion process, and are not primarily fuel based.
Table 9-3 Fuel Emission Factor Assumptions in EPA Platform v6
Fuel Type
Carbon Dioxide
(Ibs/MMBtu)
Sulfur Dioxide
(Ibs/MMBtu)
Mercury
(Ibs/TBtu)
HCI (Ibs/MMBtu)
Coal
Bituminous
Subbituminous
Lignite
202.8-216.1
209.2-216.1
212.6-219.3
0.67-7.78
0.52-2.22
1.51 -5.67
1.82 -34.71
2.03-8.65
7.32 -30.23
0.005-0.214
0.006-0.023
0.011 -0.036
Natural Gas
117.08
0
0.00014
0
Fuel Oil
Distillate
Residual
161.39
173.91
0-2.65
1.04
0.48
0.48
o o
Biomass
195
0.08
0.57
0
Waste Fuels
Waste Coal
Petroleum Coke
Fossil Waste
Non-Fossil Waste
Tires
Municipal Solid Waste
204.7
225.1
321.1
0
189.6
91.9
8.22
7.27
0.08
0
1.65
0.35
63.9
2.66
0
0
3.58
71.85
0.0921
0.0213
0
0
0
0
Note: Table 7-4 has coal emission factor on a coal supply region level.
List of tables that are uploaded directly to the web:
Table 9-4 Biomass Supply Curves in EPA Platform v6
9-4
-------
10. Financial Assumptions
10.1 Introduction and Summary
This chapter presents the financial assumptions used in the EPA Platform v6. EPA Platform v6 models a
diverse set of generation and emission control technologies, each of which requires financing88, and
incorporates updates to reflect The Tax Cuts and Jobs Act of 2017.89 The capital charge rate converts the
capital cost for each investment into a stream of levelized annual payments that ensures recovery of all
costs associated with a capital investment including recovery of and return on invested capital and
income taxes. The discount rate is used to convert all dollars to present values and IPM minimizes the
present value of annual system costs. The discount rate is set equal to the weighted average costs of
capital. Describing the methodological approach to quantifying the discount and capital charge rates in
the EPA Platform v6 is the primary purpose of this chapter.
10.2 Introduction to Risk
The cost of capital is the level of return investors expect to receive for alternative investments of
comparable risk. Investors will only provide capital if the return on the investment is equal to or greater
than the return available to them for alternative investments of comparable risk. Accordingly, the long-run
average return required to secure investment resources is proportional to risk. There are several
dimensions to risk that are relevant to power sector operations, including:
• Market Structure -The risk of an investment in the power sector is heavily dependent on
whether the wholesale power market is regulated or deregulated. The risks are higher in a
deregulated market compared to a traditionally regulated utility market. Slightly more than half of
U.S. generation capacity is deregulated (operated by Independent Power Producers (IPPs), or
'merchants').90 IPPs often sell power into spot markets supplemented by near-term hedges. In
contrast, regulated plants sell primarily to franchised customers at regulated rates, an
arrangement that significantly mitigates uncertainty, and therefore risk.91
• Technology - The selection of new technology investment options is partially driven by the risk
profile of these technology investments. For instance, in a deregulated merchant market an
investment in a peaking combustion turbine is likely to be much riskier than an investment in a
combined cycle unit. This is because a combustion turbine operates as a peaking unit and can
generate revenues only in times of high demand, or via capacity payments, while a combined
cycle unit is able to generate revenues over a much larger number of hours in a year from the
energy markets as well as via capacity payments. An investor in a combined cycle unit,
therefore, would require a lower return due to a more diversified stream of revenue, and receive a
lower risk premium than an investor in a combustion turbine, all else equal.
88 The capital charge rates discussed here apply to new (potential) units and environmental retrofits that IPM selects.
The capital cost of existing and planned/committed generating units (also referred to as 'firm'), and the emission
controls already on these units are considered sunk costs and are not represented in the model.
89 The Tax Cuts and Jobs Act of 2017, Pub.L. 115-97.
90 SNL classifies power plants as merchant and unregulated if a plant in question was not part of any rate case.
Based on this classification criterion, in 2016 about 52% of all operating capacity is merchant and unregulated
capacity.
91 There is a potential third category of risk, where IPPs enter into long-term (e.g., ten years or longer), known-price
contracts with credit worthy counterparties (e.g., traditionally regulated utilities). With a guaranteed, longer-term
price, the risk profile of this segment of the IPP fleet is similar enough to be treated as regulated plants.
10-1
-------
• Leverage - There are financial risks related to the extent of leverage. Reliance on debt over
equity in financing a project increases the risk of insolvency. This dynamic applies to all
industries, power included.92
• Financing Structure - Lastly, there are also financing structure risks (e.g., corporate vs. project
financing), also referred to as non-recourse financing. There is no clear risk implications from the
structure alone, but rather this element interacts with other dimensions of risks making
considerations of leverage, technology, and market structure more important.
• Systemic - Systemic risk is when financial performance correlates with overall market and
macro-economic conditions such that investment returns are poor when market and economic
conditions are poor, and vice versa. For example, if investors are less likely to earn recovery of
and on investments during recessions, then these risks are systemic, and increase required
expected rates of return. This emphasis on correlated market risk is based on the Capital Asset
Pricing Model (CAPM), which is used to produce key financial assumptions for EPA Platform v6.
Other risks are handled in the cash flows and are treated as non-correlated with the market.
10.2.1 Deregulation - Market Structure Risks
As noted, the power sector in North America can be divided into the traditional regulated sector (also
known as "cost of service" or "utility" sector) and deregulated merchant sector (also known as
"competitive", "merchant", "deregulated"93 or"IPP" sector).
Traditional Regulated
The traditional regulated market structure is typical of the vertically integrated utilities whose investments
are approved through a regulatory process and the investment is provided a regulated rate of return,
provided the utility's investments are deemed prudent. In this form of market structure, returns include
the return of the original investment plus a return on invested capital that are administratively determined.
Returns are affected by market conditions due to regulatory lag and other imperfections in the process,
but overall regulated investments are less exposed to the market than deregulated investments, all else
equal.
Deregulated Merchant
In a deregulated merchant market structure, investments bear a greater degree of market risk, as the
price at which they can sell electricity is dependent on what the short-term commodity and financial hedge
markets will bear. Return on investment in this form of market structure is not only dependent on the
state of the economy, but also on commodity prices, capital investment cycles, and remaining price-
related regulation (e.g., FERC price caps on capacity prices). The capital investment cycle can create a
"boom and bust" cycle, which imparts risk or uncertainty in the sector that can be highly correlated with
overall macro-economic trends. The operating cash flows from investments in this sector are more
volatile as compared to the traditional regulated sector, and hence, carry more business or market risk.94
Overall, there is ample supporting evidence for the theoretical claim that deregulated investments are
more risky than utility investments. For example:
92 We use the terms debt and leverage interchangeably.
93 Wholesale generators cannot be economically unregulated; they can be Exempt Wholesale Generator ("EWG")
subject to FERC jurisdiction. The moniker of deregulated is used to convey greater market risk relative to regulated
utility plants.
94 In this documentation, the terms "merchant financing, "deregulated"," I PR", "non-utility" and "merchant" refer to this
type of market structure.
10-2
-------
All three large publicly traded IPPs95 are rated as sub-investment grade96 while all utilities are investment
grade.
• All major IPPs have gone bankrupt over the last 15 years97.
• Estimates of beta, a measure of risk using CAPM, leverage, debt costs, and weighted average
cost of capital, consistently produce higher risk for deregulated power plants.
10.3 Federal Income Tax Law Changes
EPA Platform v6 incorporates updates to reflect The Tax Cuts and Jobs Act of 2017. The four most
significant changes in the federal corporate income tax code are:
• Rate - The corporate tax rate is lowered 14 percentage points from 35%98 to 21%; the 21% rate
is in place starting in 2018 and remains in place indefinitely; the lower tax rate decreases capital
charges in all periods and all sectors, all else held equal. When state income taxes are included,
the average rate decreases 13.1 percentage points, from 39.2% to 26.1%. This applies to all
sectors, utility and IPP.
• Depreciation - The new tax law expands near-term bonus depreciation (also referred to as
expensing) for the IPP sector only until 2027; the utility sector is unaffected.
• Interest Expense - The new law lowers tax deductibility of interest expense for the IPP sector,
which continues indefinitely; the utility sector is unaffected.
• Net Operating Losses - The new law limits the use of Net Operating Losses (NOL) to offset
taxable income. This applies to all sectors, utility and IPP.
Other important features of the new tax law include:
• Annual Variation of Provisions - The legislation specifies permanent changes (tax rate and
NOL usage limit) applying to both sectors, utility and IPP. The legislation also applies temporary
changes that vary year-by-year through to 2027 (depreciation and tax deductibility of interest)
(See Table 10-1) applying to the IPP sector only. This creates different capital charge rates for
each year through 2027. We calculate these parameters for IPM run years 2021, 2023, 2025,
and 2030 and thereafter. This set covers a wide range of financing conditions even though we do
not estimate every year.
95 Dynegy Inc. Calpine Corp. and NRG Energy Inc are the three IPP's whose ratings were B2, Ba3 and Ba3 in 2016.
96 Below minimum investment grade.
97 Dynegy, Calpine, and NRG were bankrupt - i.e. the three large public IPPs were bankrupt. Also, Mirant (major
IPP), Boston Generating (IPP), EFH (utility with large IPP component), and FES (utility with large IPP component)
have been or are bankrupt.
98 The average state income tax rate is 6.45 percent. State income tax is deductible, and hence, the combined rate is
26.1% (26.1 =21 +(1-0.21 )*6.45).
10-3
-------
Table 10-1 Summary Tax Changes
Parameter
Previous
2021"
2023
2025
2030 and Later
Marginal Tax
35
21
21
21
21
Rate - Federal
Maximum NOL
No limit. All
Carry Forward
Carry Forward
Carry Forward
Carry Forward
(Net Operating
losses in excess
cannot
cannot
cannot
cannot exceed
Loss) Carry
of income are
exceed 80%
exceed 80%
exceed 80%
80% of Taxable
Forward Usage
carried forward
of Taxable
of Taxable
of Taxable
Income
and usable
Income
Income
Income
immediately.
Tax Deductibility
100%100
IPP 30% of
30% of EBIT;
30% of EBIT;
30% of EBIT;
of Interest
EBITDA;
Utilities
Utilities
Utilities MACRS
Expense
Utilities
MACRS
MACRS
MACRS
Bonus
CM
O
O
IPP 100%;
IPP 80%103;
IPP 40%104;
0
Depreciation101
Utilities 0%
Utilities 0%
Utilities 0%
• Renewables - The legislation has minor direct potential impacts on the renewable sector's tax
credits via the Base Erosion Anti-Abuse Tax (BEAT). The maximum effect of BEAT could
decrease the value of PTC and ITC by up to 20%105; estimates of the expected impact are not yet
available, but are expected to be less. In addition, the total decrease in corporate income taxes
may decrease tax credit appetite accordingly. Nevertheless, as we lack requisite data at this time
we do not apply any additional changes to renewable financing beyond the above-mentioned
changes, which affect all capacity types.
• Utilities Versus IPPs - As noted, the legislation treats utilities and IPPs differently. The new tax
code exempts utilities from changes in tax deductibility of interest and accelerated depreciation.
The financing assumptions used in IPM modeling are a blend (weighted average) of the utility and
IPP average. The weighting is 70% utility and 30% IPP, and hence, the greatest weight is on the
least affected sector. This partly mitigates the impacts of the changes.
• Capital Charge Rates - In the past, we calculated the blended capital charge rates by taking the
weighted average of each input and calculating a single capital charge rate by technology and
location. As a result of the legislation, combined with the IPM model's ability to vary capital
charge rates by run year, the blended average is calculated for specific run years. In addition, we
99 IPM run years in the near term are 2021, 2023, 2025, and 2030.
100 No limit except losses in excess of income can be carried forward. The losses were limited to first few years.
101 Referred to as expensing. If depreciation exceeds income in first year, it can be carried forward to succeeding
years up to 80% of EBITDA.
102 Bonus depreciation was available but only in the period before IPM runs, and only for new equipment.
103 For thermal power plants coming on line in 2023, the 100% would apply only to costs incurred through end of
2022. We are hence assuming practically all capital costs are incurred prior to 2023.
104 Remaining basis depreciated at MACRS schedule.
105 https://www.conaress.aov/115/bills/hr1/BILLS-115hr1enr.xml. "Part VII - Base Erosion and Anti-Abuse Tax, Sec
59A, Tax in Base Erosion Payments of Taxpayers with Substantial Gross Receipts, (b), (1), (B), (ii), (II) the portion of
the applicable section 38 credits not in excess of 80 percent of the lesser of the amount..."
See also https://www.mwe.com/en/thouqht-leadership/publications/2017/12/renewable-enerqy-tax-bill-update-no-
chanqe-ptc-itc. A company's regular tax liability reflects certain credits that make it more likely that such a company
is subject to the BEAT. However, the Bill provides that only 20 percent of the PTC and ITC be taken into account.
Thus, 20 percent of the PTC and ITC might be denied depending on a company's BEAT status and relevant
computations in a given year.
10-4
-------
have changed the calculation for a given run year. We calculate the capital charge rates for
utilities and IPPs, and then take the weighted average of the resulting capital charge rates rather
than calculating one blended capital charge rate based on the weighted average inputs. This is
because the functional relationship between the inputs and the capital charge rates is now
different and it is less accurate to use the prior approach.
• Discount Rates - The discount rate equals the weighted average after tax cost of capital
(WACC) and is affected by the change in the corporate income tax rate only. The discount rate is
invariant overtime, sectors, and technologies. Therefore, the calculation methodology for
discount rate used in IPM is unchanged.
10.4 Calculation of the Financial Discount Rate
10.4.1 Introduction to Discount Rate Calculations
A discount rate is used to translate future cash flows into current dollars by considering factors such as
expected inflation and the ability to earn interest, which make one dollar tomorrow worth less than one
dollar today. The discount rate allows intertemporal trade-offs and represents the risk adjusted time value
of money.106
The discount rate adopted for modeling investment behavior should reflect the time preference of money
or the rate at which investors are willing to sacrifice present consumption for future consumption. The
return on private investment represents the opportunity cost of money and is commonly used as an
appropriate approximation of a discount rate.107
The real discount rate for all expenditures (capital, fuel, variable operations and maintenance, and fixed
operations and maintenance costs) in the EPA Platform v6 is 4.25%.108
10.4.2 Summary of Results
The tables below present a summary of the key financial assumption for the EPA Platform v6. A
description of these values and the attendant methodological approaches follow throughout the chapter.
106 The discount rate is the inverse of compound interest or return rate; the existence of interest, especially compound
interest creates an opportunity cost for not having dollars immediately available. Thus, future dollars need to be
discounted to be comparable to immediately available dollars.
107 For a perspective on the legal basis for utilities having the right to have the opportunity to earn such returns under
certain conditions such as prudent operations, see Bluefield Water Works and Improvement Co. v Public Service
Comm'n 262 US 679, 692 (1923). See also Federal Power Comm'n versus Hope Natural Gas Co., 320 US 591, 603
(1944).
108 This rate is equivalent to the real discount rate for a combined cycle plant under hybrid 70:30 utility to merchant
ratio assumption. It represents the most common type of thermal generation investment. This is also the hybrid real
weighted average after tax cost of capital.
10-5
-------
Table 10-2 Financial Assumptions for Utility and Merchant Cases
EPA Platform v6 - Utility WACC using daily beta for 2012-2015
Parameters
Value
Risk-free rate
3.45%109
Market premium
6.30%110
Equity size premium
0.46%111
Levered beta112
0.53
Debt/total value113
0.51
Cost of debt
4.33%114
Debt beta
0.00
Un levered beta115
0.33
Target debt/total value116
0.50
Relevered beta
0.52
Cost of equity (with size premium)117
7.20%
WACC
5.2%
EPA Platform v6 - Merchant WACC using 55% Target Debt
Parameters
Value
Risk-free rate
3.45%
Market premium
6.30%
Equity size premium
1.21 %118
Levered beta119
1.35
Debt/total value120
0.68
Cost of debt121
7.20%
109 Represents 10-year historical average (2007- June 2016) on a 20-year treasury bond. See discussion of risk-free
rate and market premium. The five year average (2012- June 2016) on a 20-year treasury bond is 2.70%. The five-
(2012- June 2016) and ten-year (2007-June 2016) averages for the 30-year bond are 3.04% and 3.65% respectively.
110 Represents the 10-year risk premium as of October 1, 2016 (A. Damodaran)
111 Size premiums according to size groupings taken from Duff & Phelps 2016 Valuation Handbook. Equity size
premium is based on weighted average of each company's equity size premium, weighted by each company's equity
capitalization level.
112 Levered betas were calculated using four years (2012-2015).
113 Debt/total value ratio is the simple average of net debt to equity ratio for the past 5 years.
114 Cost of debt is based on 5-year weighted average of debt yields for 17 utilities. The weights assigned are the
equity share of each utility. The cost of debt using the approach described in the next footnote is 4.45%.
115 Calculated using Hamada equation.
116 Target debt/total value for utility case is based on historical 5 years of average D/E for utilities
117 Cost of equity represents the simple average cost of equity. In the case of utility and merchant ROE, the decrease
reflects primarily the lower beta.
118 Size premiums according to size groupings taken from Duff & Phelps 2016 Valuation Handbook. Equity size
premium is based on weighted average of each company's equity size premium, weighted by each company's equity
capitalization level.
119 Levered betas were calculated using five years (2012- June 2016) of historical stock price data. Weekly returns
were used in the analysis.
120 Debt/total value for merchant case is calculated as simple average of the 5-year total debt to total value for each
IPP.
121 Cost of debt is based on historical 5-year weighted average of yields to maturity on outstanding debt. Analyzed
merchant companies did not issue long-term debt of 20 year or greater duration in the last five years in this analysis
(2012-2016).
10-6
-------
Debt beta122
0.18
Un levered beta123
0.69
Target debt/ total value124
0.55
Relevered beta
1.19
Cost of equity (with size premium)125
12.16%
WACC
8.40%
Table 10-3 Weighted Average Cost of Capital
Utility
Share
Utility
WACC
Merchant
Share
Merchant
WACC
Weighted
Average
Nominal
Inflation
Weighted
Average Real
WACC
WACC
70%
5.2%
30%
8.40%
6.16%
1.83%
4.25%
10.5 Discount Rate Components
The discount rate is a function of the following parameters:
• Capital structure (share of equity and debt)
• Post-tax cost of debt
• Post-tax cost of eq u ity
The WACC is used as the discount rate and is calculated as follows:126
WACC = [Share of Equity * Cost of Equity]
+ [Share of Preferred Stock * Cost of Preferred Stock]
+ [Share of Debt *After Tax Cost of Debt]
The methodology relies on debt and equity (common stock) because preferred stock is generally a small
share of capital structures, especially in the IPP sector. Its intermediate status between debt and equity
in terms of access to cash flow also tends not to change the weighted average.127 Typically, net cash
flows are used to fund senior debt before subordinated debt, and all debt before equity. Therefore, the
risk of equity is higher than debt, and the rates of return reflect this relationship. Notwithstanding,
consistent with our use of utility debt that has recourse to the corporation rather than individual assets, we
use IPP debt that has recourse to the corporation rather than individual assets because the data are more
robust.
10.6 Market Structure: Utility-Merchant Financing Ratio
With two distinct market structures, EPA Platform v6 establishes appropriate weights for regulated and
deregulated financial assumptions to produce a single, hybrid set of utility capital charge rates for new
122 Debt beta for DYN, CPN, and NRG calculated using the Merton model.
123 Calculated using Hamada equation. In merchant case, it was modified slightly to include the riskiness of debt.
124 The capitalization structure (debt to equity (D/E)) for merchant financings is assumed to be 55/45.
125 Cost of equity represents the simple average cost of equity. In both the utility and merchant cases, the decrease
primarily reflects the lower beta.
126 Sometimes abbreviated as ATWACC. The pretax WACC is higher due to the inclusion of income taxes. Income
taxes are included in the capital charges. All references are to the after-tax WACC unless indicated.
127 Debt generally has first call on cash flows and equity has a residual access.
10-7
-------
units. The EPA Platform v6, uses a weighting of 70:30, regulated to deregulated, based on recent
capacity addition shares by market type (See Table 10-4).128
Table 10-4 Share of Annual Thermal Capacity Additions by Market
Entity
2012
2013
2014
2015
2016
Average
Regulated
70%
88%
60%
58%
64%
68%
Merchant
30%
12%
40%
42%
36%
32%
10.7 Capital Structure: Debt-Equity Share
10.7.1 Introduction and Shares for Utilities and IPPs
The second step in calculating the discount rate is the determination of the capital structure, specifically
the debt to equity (D/E) or debt to value (D/V) ratio for utility and merchant investments.129 This is
calculated by determining the total market value of the company, and the market value of its debt and
equity. The market value of the company is the sum of the market value of its debt and equity. We also
determined the capital structure for the various technology types.
The target capitalization structure for utilities was assumed to be 50:50. This was based on the
capitalization over the 2012 to 2016 period. The capitalization structure for merchant financings is
assumed to be 55/45, reflecting the greater risk inherent to this market.130
10.7.2 Utility and Merchant
For utility financing, the empirical evidence suggests that utility rate of return is based on an average
return to the entire rate base. Thus, EPA Platform v6 assumes that the required returns for regulated
utilities are independent of technology. In contrast, the merchant debt capacity is based on market risk
and varies by technology.
10.7.3 Merchant by Technology
Assigning merchant technology risk is difficult because there is a lack of publicly traded securities that
provide an empirical basis for differentiating between the risks, and hence, financing parameters for
different activities.131 Nevertheless, we assigned merchant technology market risk as follows:
• Combined Cycles - The capitalization structure for merchant financing of combined cycles is
assumed to be 55/45.
• Peaking Units - A peaking unit such as a combustion turbine is estimated to have a capital
structure of 40/60. Peaking units have a less diverse, and therefore, more risky revenue stream.
128 In contrast to new units, existing coal units can be classified as belonging to a merchant or regulated market
structure. Hence, for retrofit investments, the EPA Platform v6 assumption is that coal plants owned by a utility get
purely utility financing parameters coal plants owned by merchant companies get purely merchant financing
parameters.
129 A project's capital structure is the appropriate debt capacity given a certain level of equity, commonly represented
as "D/E." The debt is the sum of all interest bearing short- and long-term liabilities, while equity is the amount that the
project sponsors inject as equity capital.
130 The U.S.-wide average authorized equity ratio during the last 5 years (2012-2016) for 146 utility companies was
50.22%. Debt/total value for merchant case is calculated as simple average of the 5-year total debt to total value for
each IPP.
131 There were only three major IPP companies with traded equity. This is insufficient to conduct statistical analysis.
10-8
-------
• Coal Units - A new coal unit is estimated to have a capital structure of 40/60, reflecting higher
risk than a combined cycle unit. This is reflected in observed higher financing costs at the two
IPP companies with coal, NRG and Dynegy, as compared to Calpine, which has no coal, only gas
and geothermal. While statistical analysis cannot be performed with such a small sample size, it
is supported qualitatively.
• Fossil Units - New, non-peaking fossil fuel-fired plants face additional risks associated with a
potential cost on future CO2 emissions, which the EIA handles by increasing the cost of debt and
equity for new coal plants.132 EPA Platform v6 extends this treatment of risk to new combined
cycle plants.
• Nuclear Units — A new nuclear unit is estimated to have a capital structure of 40/60. There is
high risk associated with a new IPP nuclear unit. This is supported by: (1) the financial
challenges facing existing nuclear units, (2) the very limited recent new nuclear construction, (3)
statements by financial institutions, and (4) the lack of ownership of nuclear power plants by pure
play IPP companies. Of the three pure play companies only one has partial ownership of a single
nuclear power plant. With this one exception, only utilities and affiliates of utilities own nuclear
units.
• Renewable Units — A new merchant renewable unit is estimated to have a capital structure of
55/45. This is the highest debt share among the major classes of generation options, and
therefore, the lowest cost of capital. This is in part because renewables have access to a third
source of financing in tax equity. Tax equity receives the tax benefits such as ITC, PTC, losses
available to defray income tax, overtime by making a payment upfront. These benefits are not
transferable to other companies. There is a risk that the tax credits may become less valuable
overtime (e.g., the company providing the tax equity does not have sufficient taxable income), or
the project may not perform and have inadequate operations to generate expected PTC volumes.
This risk is less than typical equity, since the tax credits value is not subject to as much variation
as regular equity. These projects are also easier to hedge because they have zero variable
costs, and hence, the annual volume of output is less uncertain, all else equal, and often receive
support via power purchase agreements and renewable energy credits. Limits of relying on even
greater debt include the scheduled lowering of the PTC and ITC overtime, and the potential for
performance problems.
Table 10-5 Capital Structure Assumptions
Technology
Utility
Merchant
Combustion Turbine
50/50
40/60
Combined Cycle
50/50
55/45
Coal & Nuclear
50/50
40/60
Renewables
50/50
55/45
Retrofits
50/50
40/60
10.8 Cost of Debt
The third step in calculating the discount rate is to assess the cost of debt.133 The utility and merchant
cost of debt is assumed the same across all technologies.
132 ElA's Annual Energy Outlook 2013; the capital charge rates shown for Supercritical Pulverized Coal and
Integrated Gasification Combined Cycle (IGCC) without Carbon Capture include a 3% adder to the cost of debt and
equity. See Levelized Cost of New Generation Resources in the Annual Energy Outlook 2013 (p.2),
http://www.eia.gov/forecasts/aeo/er/pdf/electricitv qeneration.pdf
133 Measured as yield to maturity.
10-9
-------
Table 10-6 Nominal Debt Rates
Technology
Utility
Merchant
Combustion Turbine
4.33%
7.20%
Combined Cycle
4.33%
7.20%
Coal & Nuclear
4.33%
7.20%
Renewables
4.33%
7.20%
Retrofits
4.33%
7.20%
10.8.1 Merchant Cost of Debt
The cost of debt for the merchant sector was estimated to be 7.2%. It is calculated by taking a 5-year
(2012-2016) weighted average of debt yields from existing company debt with eight or more years to
maturity. The weights assigned to each company debt yields were based on that company's market
capitalization. During the most recent 5 years (2012-2016), none of the existing long-term debt exceeded
twelve years to maturity, hence above average yields are based on debt with maturity between eight and
twelve years.
10.8.2 Utility Cost of Debt
The cost of debt for the utility sector was estimated to be 4.33%. It is calculated based on the 5-year
(2012-2016) average of a set of 17 utilities weighted by enterprise value (See Table 10-7).
Table 10-7 Utilities Used to Calculate Cost of Debt
Name
Ameren Corp
American Electric Power Co Inc.
Cleco Corp
CMS Energy Corp
Empire District Electric Co/The
Great Plains Energy Inc.
MGE Energy Inc.
Westar Energy Inc.
WEC Energy Group
Consolidated Edison Inc.
Southern Co/The
UIL Holdings Corp
Avista Corp
~ IDACORP Inc.
PG&E Corp
Pinnacle West Capital Corp
Xcel Energy Inc.
10.9 Return on Equity (ROE)
10.9.1 Introduction and Beta
The final step in calculating the discount rate is the calculation of the required rate of return on equity
(ROE). The ROE is calculated using the formula:
ROE = risk free rate + beta x equity risk premium + size premium
10-10
-------
The formula is the key finding of the CAPM and reflects that a premium on return is required as
investment risk increases, and that premium is proportional to the systemic risk of the investment.134
Systemic risk is measured by the impact of market returns on the investment's returns and is measured
by beta.135
There are several additional aspects of estimating beta:
• Time Period - The most common practice is to use five years of historical returns to estimate
beta.
• Returns - Daily returns are commonly used to estimate beta except for illiquidly traded stocks
when weekly returns are used to avoid under estimating beta. The utility estimates presented
use daily data and the IPP estimates used weekly estimates.
• Unlevered Betas - It is useful to estimate unlevered betas that eliminate the effects of leverage.
This facilitates comparison across investments with different leverage levels, and allows
recalculation to account forgoing forward changes in leverage levels. This recalculation involves
a technique known as the Hamada136 equation.
• Debt Betas - When a company is facing financial distress, the debt can become the new equity
as part of corporate reorganization under the federal bankruptcy code. Hence, during the
bankruptcy period, the debt trades like equity. There is a technique to adjust the beta by
calculating a debt beta. This technique is employed because one of the three IPP companies,
Dynegy, was having significant financial distress especially early in the 2012-2016 period.
10.9.2 Risk-Free Rate and Equity Risk Premium
The risk-free rate of return and equity risk premium are market parameters, and are not company-specific.
They also determine the average market-wide level of returns on equity. Therefore, the average return of
the market equals the sum of the risk-free rate of return and equity risk premium.
In this analysis, we use the Duff and Phelps 2016 Valuation Handbook, Industry Cost of Capital. Duff and
Phelps recommends an estimate of 5.5% for the market risk premium137. At the same time, Duff and
Phelps recommends a 4% risk-free rate. Thus the total is 9.5%.
The EPA estimates are based on the approach of using long-term averages for both the risk-free rate and
the market risk premium. Specifically, EPA estimates the risk-free rate of return and the market risk
premium based on 10-year averages. The risk free rate assumption is 3.45% which is the 10-year (2007-
2016) average of U.S. Treasury 20 year bond rates. The market risk premium is the ten year average
provided by Professor Damodaran of 6.3%138. The sum of the two is 9.75%, and is close to Duff and
Phelps recommendation of 9.5%.
134 The financial literature on CAPM originally did not emphasize the size premium (also referred to as the liquidity
premium). It emerged from later findings that the estimated required return was too low for small stocks (i.e., with low
equity value).
135 Beta is the covariance of market and the stock's returns divided by the variance of the market's return.
136 In corporate finance, Hamada's equation is used to separate the financial risk of a levered firm from its business
risk.
137 Duff and Phelps, 2016 Valuation Handbook, March 2016, see also Client Alert, Duff and Phelps Increases U.S.
Equity Risk Premium Recommendation to 5.5% Effective January 31, 2016.
138 As of October 1, 2016.
10-11
-------
10.9.3 Beta
Utility betas average 0.53 during the 2012 to 2015 period on a levered basis (see Table 10-8). This
estimate is based on daily returns. This estimate was chosen because it was intermediate between the
ten-year average and the 2012-2016 estimate when using partial year 2016 data. For example, the ten-
year beta (2007-2016) is even higher at 0.60 daily, and the 2012-2016 partial year estimate is 0.5
because the partial year 2016 data is much lower than the 2012-2015 average.139
Table 10-8 Estimated Annual Levered Beta for S15ELUT Utility Index Based on Daily Returns140
Year
Levered Beta
2012
0.35
2013
0.70
2014
0.44
2015
0.62
2016 (through June)
0.25
Average (2012-15)
0.53
IPP betas average 1.35 based on weekly returns from 2012-June 2016. We did not observe issues with
partial year 2016 data. After decreasing leverage from 68% to 55%, and adjusting the beta estimate, the
beta decreases to 1.19. Even after correcting for the greater financial risk of IPPs due to higher leverage,
the betas of IPPs are higher than utilities. The unlevered betas of utilities is 0.33 and of the IPPs is
0.69141.
10.9.4 Equity Size Premium
It is observed that long-run returns of smaller, less liquidly traded companies have higher returns than
predicted using the market risk premium. Therefore, an equity size of liquidity premium is added. Based
on the 2016 Duff and Phelps Valuation Handbook there was a significant equity size premium for IPPs of
1.21% and a smaller premium for utilities at 0.46%.
10.9.5 Nominal ROEs
Utility
The utility ROE is 7.20% in nominal terms. The utility ROE is the single most influential parameter in the
estimate of the discount rate because of the 70% weight given to utilities compared to IPPs, and the
decrease in interest rates due to the tax shield on debt (debt interest payments are tax deductible).
The estimated utility ROE in EPA Platform v6 is lower than what state and federal commissions have
awarded the shareholder-owned electric utilities recently.142 In some cases, commissions use a different
139 One-half weight to 2016.
140 S15ELUT Index comprises of 22 utilities: American Electric Power Co. Inc., Great Plains Energy Inc., Westar
Energy Inc., IDACORP Inc., PG&E Corp., Pinnacle West Capital Corp., Xcel Energy Inc., NextEra Energy Inc, Duke
Energy Corp, Southern Co, Exelon Corp., Edison International, PPL Corp., Eversource Energy, First Energy Corp.,
Entergy Corp., Alliant Energy Corp., OGE Energy Corp., Hawaiian Electric Industries Inc., ALETTE Inc., PNM
Resources Inc., and El Paso Electric Co.
141 Unlevered betas are lower than levered betas. Levered beta is directly measured from the company's stock
returns with no adjustment made for the debt financing undertaken by the company. The leveraged beta of the
market equals one.
142 SNL-based rate case statistics for 2012-2016 suggest nationwide average ROE rate of 9.93%.The Edison Electric
Institute's Financial Update, Rate Case Summary, Q4 2015 reported average approved returns on equity of 9.6% the
second lowest in its three decades of data.
10-12
-------
approach or assumptions143. If it were shown that the existence of higher returns at other utilities
prevented utilities receiving the estimated return here while still attracting sufficient capital, this could
mean that the estimate here is too low. However, ICF's experience notes that the trend is to lower
returns and this is a long-term analysis focused on cost of capital for future investments that can occur 25
years or more in the future.
IPP
The nominal ROE for IPPs is 12.16%. The IPP required ROE is sensitive to the amount of debt and the
analysis assumes future delevering. Specifically, the IPP ROE assumes 55% debt rather than 68% debt,
which is the 2012-2016 average.
10.9.6 WACC/Discount Rate
The WACCs are 5.2% in nominal terms for utilities and 8.40% in nominal terms for IPPs (see Table 10-3).
Using a 70:30 utility/merchant weighting, the weighted average WACC under utility financing and
merchant financing is a 6.16% WACC. The real hybrid WACC is 4.25%.
10.10 Calculation of Capital Charge Rate
10.10.1 Introduction to Capital Charge Rate Calculations
The capital charge rate is used to convert the capital cost into a stream of levelized annual payments that
ensures capital recovery of an investment. The number of payments is equal to book life of the unit or the
years of its book life included in the planning horizon (whichever is shorter). Table 10-9 to Table 10-11
presents the capital charge rates by technology type used in EPA Platform v6. As discussed in section
10.3, the changes to the Tax Code have caused capital charge rates to vary by run year, therefore the
tables below show the rates for the individual run years through 2030. Capital charge rates are a function
of underlying discount rate, book and debt life, taxes and insurance costs, and depreciation schedule.
Table 10-9 Real Capital Charge Rate - Blended (%)144
New Investment Technology Capital
Hybrid (70/30 Utility/Merchant)
2021
2023
2025
2030 and
Beyond
Environmental Retrofits - Utility Owned
10.77%
10.77%
10.77%
10.77%
Environmental Retrofits - Merchant
Owned
14.05%
14.05%
14.12%
14.54%
Advanced Combined Cycle
8.64%
8.64%
8.66%
8.77%
Advanced Combustion Turbine
9.02%
9.02%
9.02%
9.10%
143 Some regulatory commissions use what is known as the dividend growth model. This model assumes that the
current market price of a company's stock is equal to the discounted value of all expected future cash flows. In this
approach, the time period is assumed to be infinite, and the discount rate is a function of the share price, earnings per
share and estimated future growth in dividends. The challenge with using this approach is estimating future growth in
earnings. Commissions rely on stock analyst forecasts of future growth rates for dividends. In other cases,
commissions may allow for other parameters such as flotation costs (costs of issuing stock). We did not use this
approach because it is less commonly used. There also appears to be a tendency of allowed rates of return as a
group to be too low during periods with high financial costs and too high during periods of low financing costs. This
may be to ensure comparability with similar utility companies. There is also a literature that indicates that as betas
deviate from 1, the CAPM returns are too low and too high. We did not address these issues directly in part because
the results were comparable to other results, with the exception of being lower than allowed returns.
144 Capital charge rates were adjusted for expected inflation and represent real rates. The expected inflation rate
used to convert future nominal to constant real dollars is 1.83%. The future inflation rate of 1.83% is based on an
assessment of implied inflation from an analysis of yields on 10-year U.S. Treasury securities and U.S. Treasury
Inflation Protected Securities (TIPS) over a period of 5 years (2012-2016).
10-13
-------
New Investment Technology Capital
Hybrid (70/30 Utility/Merchant)
2021
2023
2025
2030 and
Beyond
Ultra Supercritical Pulverized Coal
without Carbon Capture145
10.96%
10.96%
11.01%
11.18%
Ultra Supercritical Pulverized Coal with
Carbon Capture
8.31%
8.31%
8.32%
8.43%
Nuclear
8.31%
8.31%
8.33%
8.43%
Nuclear without Production Tax Credit
8.31%
8.31%
8.33%
8.43%
Nuclear with Production Tax Credit146
7.10%
7.09%
7.10%
7.19%
Biomass
8.14%
8.12%
8.12%
8.12%
Wind, Landfill Gas, Solar and
Geothermal
9.79%
9.78%
9.77%
9.77%
Hydro
8.09%
8.09%
8.11%
8.21%
Table 10-10 Real Capital Charge Rate - IPP (%)
New Investment Technology
Capital (IPP)
2021
2023
2025
2030 and
Beyond
Environmental Retrofits - Merchant
Owned
14.05%
14.05%
14.12%
14.54%
Advanced Combined Cycle
10.89%
10.89%
10.97%
11.33%
Advanced Combustion Turbine
11.83%
11.81%
11.81%
12.07%
Ultra Supercritical Pulverized Coal
without Carbon Capture
14.05%
14.06%
14.23%
14.78%
Ultra Supercritical Pulverized Coal
with Carbon Capture
11.22%
11.22%
11.27%
11.62%
Nuclear without Production Tax
Credit
11.22%
11.22%
11.29%
11.62%
Nuclear with Production Tax Credit
9.71%
9.69%
9.71%
10.00%
Biomass
10.60%
10.56%
10.53%
10.53%
Wind, Landfill Gas, Solar and
Geothermal
11.77%
11.73%
11.70%
11.70%
Hydro
10.61%
10.61%
10.67%
11.01%
Table 10-11 Real Capital Charge Rate - Utility (%)
New Investment Technology
Capital Utility
2021
2023
2025
2030 and
Beyond
Environmental Retrofits - Utility
Owned
10.77%
10.77%
10.77%
10.77%
Advanced Combined Cycle
7.67%
7.67%
7.67%
7.67%
Advanced Combustion Turbine
7.82%
7.82%
7.82%
7.82%
145 EPA has adopted the procedure followed in ElA's Annual Energy Outlook 2013; the capital charge rates shown for
Supercritical Pulverized Coal and Integrated Gasification Combined Cycle (IGCC) without Carbon Capture include a
3% adder to the cost of debt and equity. See Levelized Cost of New Generation Resources in the Annual Energy
Outlook 2013 (p.2), http://www.eia.gov/forecasts/aeo/er/pdf/electricitv qeneration.pdf
146 The Energy Policy Act of 2005 (Sections 1301, 1306, and 1307) provides a production tax credit (PTC) of 18
mills/kWh for 8 years up to 6,000 MW of new nuclear capacity. The financial impact of the credit is reflected in the
capital charge rate shown in for "Nuclear with Production Tax Credit (PTC)." NEEDS v6 integrates 2,200 MW of new
nuclear capacity at Vogtle nuclear power plant. Therefore, in EPA Platform v6 only 3,800 MW of incremental new
nuclear capacity will be provided with this tax credit.
10-14
-------
New Investment Technology
Capital Utility
2021
2023
2025
2030 and
Beyond
Ultra Supercritical Pulverized Coal
without Carbon Capture
9.63%
9.63%
9.63%
9.63%
Ultra Supercritical Pulverized Coal
with Carbon Capture
7.06%
7.06%
7.06%
7.06%
Nuclear without Production Tax
Credit
7.06%
7.06%
7.06%
7.06%
Nuclear with Production Tax Credit
5.98%
5.98%
5.98%
5.98%
Biomass
7.08%
7.08%
7.08%
7.08%
Wind, Landfill Gas, Solar and
Geothermal
8.94%
8.94%
8.94%
8.94%
Hydro
7.01%
7.01%
7.01%
7.01%
10.10.2 Capital Charge Rate Components
The capital charge rate is a function of the following parameters:
• Capital structure (debt/equity shares of an investment)
• Pre-tax debt rate
• Debt life
• Post-tax return on equity
• Other costs such as property taxes and insurance
• State and federal corporate income taxes
• Depreciation schedule
• Book life
Table 10-12 presents a summary of various assumed book lives, debt lives and the years over which the
investment is fully depreciated. The book life or useful life of a plant was estimated based on publicly
available financial statements of utility and merchant generation companies.147
Table 10-12 Book Life, Debt Life and Depreciation Schedules for EPA Platform v6
Technology
Book Life
(Years)
Debt Life
(Years)
U.S. MACRS Depreciation
Schedule (Years)
Combined Cycle
30
20
20
Combustion Turbine
30
15
15
Coal Steam and IGCC
40
20
20
Nuclear
40
20
15
Solar, Geothermal, Wind, and Landfill Gas
20
20
5
Biomass
40
20
7
Environmental Retrofits
15
15
15
Depreciation Schedule
For the utility sector, the U.S. MACRS depreciation schedules were obtained from IRS Publication 946
that lists the schedules based on asset classes.148 149 The document specifies a 5-year depreciation
147 SEC 10K filings of electric utilities and merchant companies. For example, Calpine's 10K lists 35 years of useful
life for base load plants, DTE energy uses 40 years for generation equipment; Dynegy uses 30 years for power
generation facilities.
10-15
-------
schedule for wind energy projects and 20 years for electric utility steam production plants. These exclude
combustion turbines and nuclear power plants, which each have a separate listing of 15 years. As a result
of the tax code changes, the merchant sector is allowed to depreciate assets on an accelerated schedule
through 2027. Accelerated depreciation is allowed starting in 2018 with 100% depreciation and phases
out at 20% annual between 2023 and 2027.
Taxation and Insurance Costs
The maximum U.S. corporate income tax rate is 21%.150 State taxes vary but the weighted average state
corporate marginal income tax rate is 6.45%. This yields a net effective corporate income tax rate of
26.1%.
U.S. state property taxes are approximately 0.9%, based on a national average basis. This is based on
extensive primary and secondary research conducted by EPA using property tax rates obtained from
various state agencies.
Insurance costs are approximately 0.3% on a national average basis.
148 MACRS refers to the Modified Accelerated Cost Recovery System, issued after the release of the Tax Reform Act
of 1986.
149 IRS Publication 946, "How to Depreciate Property," Table B-2, Class Lives and Recovery Periods.
150 Internal Revenue Service, Publication 542.
10-16
------- |