UPPER MIDWEST
INTEGRATED
RESOURCE PLAN
2020-2034

Northern States Power Company
Docket No. E002/RP-19-368

xcelenergy.com   © 2019 Xcel Energy Inc.   Xcel Energy is a registered trademark of Xcel Energy Inc.   19-06-130

 Xcel Energy

Table of Contents

TABLE OF CONTENTS
Cover Letter
Table of Contents
The Resource Plan
Chapter 1: Executive Summary
Chapter 2: Planning Landscape
Chapter 3: Minimum System Needs
Chapter 4: Preferred Plan
Chapter 5: Economic Modeling Framework
Chapter 6: Customer Rate and Cost Impacts
Certificate of Service & Service Lists
Appendices

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
The Resource Plan – i

PREFERRED PLAN
CHAPTER 1:

EXECUTIVE SUMMARY

1

I.
II.

1
4
7
14
18
19

INTRODUCTION
EXECUTIVE SUMMARY
A. Proposed Resource Mix
B. Plan Priorities
III. CONCLUSION
CHAPTER 2: PLANNING LANDSCAPE
I.
II.

INTRODUCTION
PLANNING LANDSCAPE
A. Carbon Reduction Goals
B. Regional Reliability and Market Constructs
C. Distributed Energy Resources
D. Community and Employee Considerations
E. Customer Preferences
F. Supply and Technology Trends
G. Jurisdictional Updates
H. Conclusion

CHAPTER 3: MINIMUM SYSTEM NEEDS
I.
II.

III.
IV.
V.

VI.
VII.

July 1, 2019

MEETING CUSTOMER NEEDS
A. Customer Needs Forecast
MISO RESOURCE ADEQUACY REQUIREMENTS
A. MISO Footprint Capacity Obligation
B. Zonal Capacity Obligation
C. Capacity Obligations Derived From Forecasted
Demands
D. Capacity Accreditation of Resources
DEMAND SIDE MANAGEMENT
RELIABILITY REQUIREMENT
EXISTING RESOURCES
A. Renewable Resources
B. Nuclear
C. Coal
D. Natural Gas (and Oil-Fired) Fleet
NET RESOURCE SURPLUS/DEFICIT
MEETING RENEWABLE ENERGY
REQUIREMENTS AND GOALS

19
23
24
25
34
36
37
39
40
44
45
45
46
50
50
51
51
52
54
55
56
56
57
57
57
58
59

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
The Resource Plan – ii

A. Minimum Compliance Requirements
B. RES Compliance
C. SES Compliance
VIII. ENERGY POLICY AND COMPANY GOALS
IX.
REFERENCE CASE
CHAPTER 4: THE PREFERRED PLAN
I.

II.

III.
IV.

V.

VI.

PLANNING OBJECTIVES
A. Environmental and Innovation
B. Reliability
C. Cost
D. Risk and Flexibility
E. Our Employees and Communities
THE PREFERRED PLAN
A. Transforming Our Energy Mix
B. Fleet Transition
C. Keeping Rates Affordable for Customers
ACTION PLANS
A. Five-Year Plan
B. Long-Term Plan
NORTH DAKOTA PLAN
A. Plan Components
B. Resource Planning Framework Status
C. North Dakota Combustion Turbine
PUBLIC INTEREST ANALYSIS
A. Reliability
B. Impact to Customer Bills
C. Environmental Effects
D. Socioeconomic Impacts
E. Flexibility to Respond to Change
F. Limiting Risks
CONCLUSION

CHAPTER 5: ECONOMIC MODELING FRAMEWORK
I.
II.

III.
July 1, 2019

ASSUMPTIONS
SCENARIOS
A. Reference Case Scenario
B. Early Coal Family
C. Early Nuclear Family
D. Extend Nuclear Family
FUTURES SCENARIOS AND SENSITIVITIES

59
60
61
62
62
65
66
67
68
68
69
70
70
71
72
77
79
79
81
83
84
86
86
87
88
88
89
89
90
90
91
92
92
97
97
97
98
98
99

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
The Resource Plan – iii

IV.

V.

VI.

A. Futures Scenarios
B. Traditional Sensitivities
STRATEGIST ANALYSIS AND RESULTS
A. Initial Full Optimization
B. Revised Targeted Optimization
C. Energy Efficiency and Demand Response Analysis
D. Final Scenario Analysis
E. Modeling Results and Conclusions
PREFERRED PLAN SELECTION AND
ASSESSMENT
A. Baseload Study Analysis Results
B. Early Retirement of Sherco 3 by 2030
C. Demand Response Adjustment to Scenario 9
D. Futures Scenarios Results
E. Preferred Plan Benefits
CONCLUSION

CHAPTER 6: CUSTOMER RATE AND COST IMPACTS
I.
II.

III.

IV.
V.

July 1, 2019

REFERENCE CASE REVENUE REQUIREMENTS
FORECAST METHODOLOGY
ESTIMATED RATE IMPACTS AND
METHODOLOGY
A. Methodology
B. Estimated Overall Rate Impacts
C. Key Drivers
D. These Estimates are not Directly Comparable to Rate
Impact Analysis in a Rate Case
ESTIMATED RATE IMPACTS BY CLASS PER YEAR
A. Methodology and Calculations
B. Fuel Costs and Purchased Energy
C. Production Expense, Property Taxes, Deferred
Income Taxes, Tax Depreciation and Removal
Expense and Decommissioning Accrual
D. Generation Rate Base Costs Including Plant in
Service, Depreciation, CWIP and Accumulated
Deferred Income Taxes
E. Bulk Transmission Costs
DETERMINING CLASS RATE IMPACTS
FACTORS IMPACTING NEAR-AND LONG-TERM
RATE ESTIMATES

99
104
106
106
108
109
110
111
115
116
117
118
118
122
123
124
126
127
127
128
129
130
131
133
134
134
136
136
136
138

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Appendices – i

APPENDICES
Appendix A –

Compliance Matrix

Appendix B –

Acronyms & Terms

Appendix C –

About Xcel Energy

Appendix D –

Non-Technical Summary

Appendix E –

Xcel Energy Carbon Report: Building a Carbon-Free Future

Appendix F1 –

Load and Distributed Energy Resource Forecasting

Appendix F2 –

Strategist Modeling Assumptions & Inputs

Appendix F3 –

Scenario Sensitivity Analysis: PVRR & PVSC Summary

Appendix F4 –

High Electrification Scenario Description

Appendix F5 –

Generation Interconnection Costs Methodology

Appendix F6 –

Resource Options

Appendix F7 –

Minnesota Energy Storage Systems Assessment

Appendix G1 –

Demand Side Management

Appendix G2 –

Study: Potential for Load Flexibility at NSP (Brattle)

Appendix G3 –

DR Cost Effectiveness at NSP (Brattle)

Appendix G4 –

DR Stakeholder Engagement Summary (GPI)

Appendix H –

Environmental Regulations Review

Appendix I –

Supporting Infrastructure: Transmission & Distribution

Appendix J1 –

Baseload Study

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Appendices – ii

Appendix J2 –

Reliability Requirement

Appendix J3 –

Y2 Study: King & Sherco Unit 3 Coal

Appendix J4 –

Y2 Study: Monticello, Prairie Island, & All Coal & Nuclear

Appendix K –

Xcel Energy Resources: Nuclear

Appendix L –

Xcel Energy Resources: Sherco CC

Appendix M –

Xcel Energy Corporate Responsibility Report

Appendix N1 –

Annual Report: Minnesota Electric

Appendix N2 –

10-Year Plan: North Dakota Electric

Appendix N3 –

10-Year Plan: South Dakota

Appendix N4 –

Renewable Energy Compliance Positions

Appendix N5 –
Appendix N6 –

Biennial Report: Renewable Energy Obligation-Renewable Energy
Standard Compliance Report
Renewable Energy Standard: Rate Impact Report:

Appendix N7 –

Annual Report: Solar Energy Standard

Appendix N8 –

Monthly Report: Solar Gardens June 2019

Appendix N9 –

Triennial Filing: Nuclear Decommissioning

Appendix N10 – Nuclear Worker Transition Plan
Appendix O1 –

Summary of IRP Stakeholder Engagement

Appendix O2 –

Scope & Outline: Center for Energy & Environment Study

Appendix O3 –

Impact of Nuclear Fleet on the Minnesota Economy (NEI)

Appendix P1 –

E3 Summary

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Appendices – iii

Appendix P2 –

RESOLVE & RECAP Low Carbon Scenario Analysis (E3)

Appendix P3 –

Minnesota PATHWAYS Report June 2019 (E3)

Appendix Q –

NSP Wind & Solar Integration Study (Enernex)

Appendix R –

Study: Interconnection Cost Estimates – CC (Excel Engineering)

Appendix S –

Combined Heat & Power Study (EPRI)

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Tables and Figures: The Resource Plan – i

THE RESOURCE PLAN LIST OF TABLES AND FIGURES
Number
Figure 1-1
Figure 1-2
Figure 1-3
Figure 1-4
Figure 1-5
Figure 2-1
Figure 2-2
Figure 3-1
Figure 3-2
Figure 3-3
Figure 3-4
Table 3-1
Figure 3-5
Table 3-2
Figure 3-6
Figure 3-7
Figure 3-8
Table 3-3
Table 3-4
Figure 3-9
Figure 4-1
Figure 4-2
Table 4-1
Figure 4-3
Table 4-2
Figure 5-1
Table 5-1
Figure 5-2
July 1, 2019

Title
Projected Carbon Emissions Through 2030
Preferred Plan Highlights
Preferred Plan Energy Mix through 2034
NSP System Reliability Requirement Calculation
Preferred Plan Average Rate Impact for the NSP System
Xcel Energy Integrated Resource Plan Objectives
Modeled Wind and Solar ELCC as Penetration Increases
Net Resource Need/Surplus Calculation
Forecasted Peak Load, After Energy Efficiency Adjustments
(MW)
Forecasted Net Energy Requirements, After Energy Efficiency
Adjustments (GWh)
MISO Planning Reserve Margin Calculation – NSP System
Planning Year June 1, 2018 to May 31, 2019
Capacity Obligation Calculation – 2020 Example
NSP System Reliability Requirement Calculation – 2020
Example
Reference Case Load and Resources, 2020-2034 Planning
Period
Renewable Energy Requirements and Objectives – NSP
System
REC Production and Retirement Obligations for NSP System
– Existing Resources Only
NSP System SREC Production and Minnesota Annual
Requirements
Reference Case Annual Expansion Plan (UCAP)
Reference Case Annual Expansion Plan (ICAP)
Reference Case Energy Mix in 2020 and 2034
Xcel Energy Integrated Resource Plan Objectives
Preferred Plan Energy Mix
Preferred Plan Resource Additions (MW)
Preferred Plan Average Nominal Cost Comparison (NSP
System)
Expansion Plan Comparisons Preferred Plan – North Dakota
Plan – Summary of Differences
Effective Reserve Margin Used in Strategist Modeling
2019 Resource Plan Futures Scenarios
Peak Demand, Net of EE Impacts, by Futures Scenario (MW)

Page #
1
2
8
16
17
19
29
46
47
48
52
52
55
58
59
61
62
63
63
64
67
71
72
78
85
93
100
101

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Figure 5-3
Figure 5-4
Figure 5-5
Table 5-2
Figure 5-6
Figure 5-7
Figure 5-8
Figure 5-9
Table 5-3
Figure 5-10
Table 5-4
Figure 5-11
Table 5-5
Table 5-6
Table 5-7
Table 5-8
Figure 6-1
Figure 6-2
Figure 6-3
Table 6-1
Figure 6-4
Table 6-2
Table 6-3
Table 6-4
Figure 6-5
Table 6-5

July 1, 2019

Table of Contents
Tables and Figures: The Resource Plan – ii

Fuel Price Assumptions, by Futures Scenario
New Resource Cost Assumptions, by Futures Scenario
($/MWh; $/kW-mo)
Carbon Emissions Cost Assumptions ($/Short Ton)
Scenario 9 (Preferred Plan) DR and EE Cost Effectiveness
Analyses ($2019 millions)
Expansion Plans by Scenario (GW, Cumulative Nameplate
Capacity Resource Additions by Fuel Type)
Scenario PVSC Deltas from Reference Case ($2019 millions)
Scenario PVRR Deltas from Reference Case ($2019 millions)
Cumulative PVRR Cost or Savings Deltas by Scenario,
Compared to the Reference Plan ($2019 million)
Scenario Modeling Portfolio Risk Metrics
Scenario PVSC Deltas from Reference Case, PI Extension
Cases Eliminated ($2019 millions)
Preferred Plan NPV Savings under Different Futures Scenarios
($2019 millions)
Preferred Plan Annual Costs or Savings Compared to the
Reference Case, by Scenario ($2019 millions)
Preferred Plan Base Expansion Plan (MW)
High Electrification Scenario Expansion Plan (MW)
High Distributed Solar Scenario Expansion Plan (MW)
Cumulative 2020-2034 Additions by Resource Type and
Scenario (MW)
Preferred Plan Average Nominal Cost Comparison (NSP
System)
Preferred Plan Average Nominal Cost Comparison (State of
Minnesota)
Annual Percent Change in Revenue Requirements (2020-2040)
Preferred Plan above Reference Case – State of Minnesota
Estimated Incremental Impact of Preferred Plan State of
Minnesota – All Customers
Incremental Rate Impact of Preferred Plan State of Minnesota
– All Customers
E8760 Energy Allocator
Stratification Analysis by Plant Type
D10S Capacity Allocator
Incremental Rate Impact of Preferred Plan by Customer Class
– State of Minnesota
Preferred Plan Estimated Rate Impacts by Class per Year

102
103
104
110
111
112
113
114
115
116
118
120
120
121
121
121
125
126
129
132
133
134
135
135
137
138

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Tables and Figures: Appendices – i

APPENDICES LIST OF TABLES AND FIGURES
Appendix C: About Xcel Energy
Number
Figure 1
Figure 2

Title
Xcel Energy Territory
NSP System Generation Resources

Page #
1
2

Appendix D: Non-Technical Summary
Number
Figure 1
Figure 2
Figure 3
Table 1
Table 2
Table 3
Figure 4
Figure 5
Figure 6

Title
Xcel Energy Integrated Resource Plan Objectives
Preferred Plan Highlights
Net Resource Need/Surplus Calculation
Reference Case Load and Resources 2020-2034 Planning
Period
Reference Case Annual Expansion Plan (MW)
Preferred Plan Resource Additions (MW)
Preferred Plan Energy Mix
Preferred Plan Average Nominal Cost Comparison NSP
System
Incremental Rate Impact of Preferred Resource Plan
State of Minnesota – All Customers

Page #
2
3
11
12
13
15
15
18
19

Appendix F1: Load and Distributed Energy Resources Forecasting
Number
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7

July 1, 2019

Title
NSP System Total Median Net Energy (GWh) (Includes 1.5
Percent EE Adjustment)
Gross Energy Requirements Forecast Compared to Net
Energy Requirements Forecast
NSP System Median Base Summer Peak Demand (MW)
(Includes 1.5 Percent EE Adjustment)
Gross Peak Demand Forecast Compared to Net Peak Demand
Forecast
Illustration of EE Adjustment – NSP System Demand (MW)
NSP System Total Net Energy (MWh) (Includes 1.5 Percent
EE Adjustment)
NSP System Total Base Summer Peak Demand (MW)
(Includes 1.5 Percent EE Adjustment)

Page #
2
3
4
5
8
11
12

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Figure 8
Table 1
Table 2
Table 3
Figure 9
Figure 10
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9
Table 10
Table 11
Figure 11
Figure 12
Figure 13
Figure 14

Table of Contents
Tables and Figures: Appendices – ii

NSP System Total Net Summer Peak Demand (MW) (Includes
1.5 Percent EE Adjustment)
Annual Net Energy (MWh) (Including 1.5 Percent EE
Adjustment)
Annual Base Summer Peak Demand (MW) (Includes 1.5
Percent EE Adjustment)
Annual Net Peak Demand (MW) (Includes 1.5 Percent EE
Adjustment)
Net Energy Requirements (MWh) – Comparison of Current
and Previous Energy Forecast Median (50th Percentile)
Forecast
Base Peak Demand (MW) – Comparison of Current and
Previous Demand Forecast Median (50th Percentile) Forecast
Jurisdiction Abbreviations
Monthly MWh Sales Series
Monthly Customer Series
Monthly Economic and Demographic Series
Monthly Data Variables used in Demand Model
Monthly Weather Variables used in Sales Models
Other Monthly Variables
Monthly Binary Variables
Reference Case – NSP System Distributed Solar PV Forecast
(Nameplate MW/AC)
High Forecast Case – NSP System Distributed Solar PV
Forecast
(Nameplate MW/AC)
NSP System Distributed Energy Storage Systems Forecast
NSP System Base Level Compared to High Electrification
Sensitivity – 2020-2034

12
13
13
14
15
16
20
20
20
20
21
21
21
21
28
29
30
32

Appendix F2: Strategist Modeling Assumptions and Inputs
Number
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7

July 1, 2019

Title
Discount Rate and Capital Structure
CO2 Costs
Externality Costs
Strategist Demand and Energy Forecast
High Load Sensitivity
Low Load Sensitivity
Energy Efficiency Bundles

Page #
1
3
4
6
7
8
10

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table 8
Table 9
Table 10
Table 11
Table 12
Table 13
Table 14
Table 15
Table 16
Table 17
Table 18
Table 19
Table 20
Table 21
Table 22
Table 23

Table of Contents
Tables and Figures: Appendices – iii

Demand Response Forecast
Fuel and Market Price Forecasts
Surplus Capacity Credit
Transmission Delivery Costs
Integration and Congestion Costs
Distributed Solar Forecast
Thermal Generic Information (Costs in 2018 Dollars)
Renewable Generic Information (Costs in 2018 Dollars)
Storage Generic Information (Costs in 2018 Dollars)
Levelized Capacity Costs by In-Service Year
Base Renewable Levelized Costs by In-Service Year
Low Renewable Levelized Costs by In-Service Year
High Renewable Levelized Costs by In-Service Year
2018 FERC Heat Rates
Heat Rate Tests – 2015-2018
Load and Resources Tables, 2020-2034 Planning Period

11
13
14
15
16
17
21
21
22
23
24
25
26
27
28
30

Appendix F4: High Electrification Scenario
Number
Figure 1
Figure 2
Figure 3

Title
Incremental Annual Load under a High Electrification Scenario
Incremental Peak Demand under a High Electrification
Scenario
Electricity Demand on a Peak Day in 2034 (Left) and 2050
(Right) under a High Electrification Sensitivity

Page #
3
4
4

Appendix F6: Resource Options
Number
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7

July 1, 2019

Title
Existing Coal Resources
Existing Nuclear Resources
Existing Natural Gas and Oil Resources
Existing Biomass Resources
Existing Hydroelectric Resources
Existing and Near Term Wind Resources
Existing Solar Resources

Page #
2
3
4
5
6
7
8

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Tables and Figures: Appendices – iv

Appendix F7: Minnesota Energy Storage Systems Assessment
Number
Figure 1
Figure 2

Title
Clean Air Task Force – Smoothed Daily Load and Renewable
Energy Generation, Mixed Renewable Scenario
Battery Storage for Peak Demand Reduction

Page #
9
11

Appendix G1: Demand Side Management
Number
Figure 1
Figure 2
Figure 3
Table 1
Figure 4
Figure 5
Table 2
Table 3
Table 4
Table 5

Title

Demand Response Avoided Cost Benefits (2030)
Cost-effective Demand Response Potential
Controllable Demand (Gen. MW)
Demand Response Five-Year Action Plan
Future Demand Response Growth in 5-Year Planning Period
Historical Electric CIP Achievements 2004-2018
Energy Efficiency Scenarios
Percent of Portfolio Energy
Present-Value of Revenue Requirements (PVRR) Energy Efficiency
Scenarios (in Millions)
Present-Value of Societal Costs (PVSC) Energy Efficiency Scenarios (in
Millions)

Page #
8
10
15
17
18
27
29
31
32
33

Appendix H: Environmental Regulations Review
Number
Table 1
Table 2

Title
Xcel Energy Upper Midwest CO2 Emission Reductions
Xcel Energy Upper Midwest System Status – NAAQS
Attainment

Page #
10
17

Appendix I: Supporting Infrastructure: Transmission & Distribution
Number
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
July 1, 2019

Title
Modeled wind and solar ELCC as penetration increases
NSP System Reliability Requirement Calculation –
2020 Example
Xcel Energy Strategic Priorities – Applied to Distribution
Illustrative Electrical Grid
Annual Distribution Planning Process
Staged Approach to Enhanced Planning Analyses
Platform Considerations

Page #
8
12
21
22
25
34
35

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Figure 8
Figure 9

Table of Contents
Tables and Figures: Appendices – v

Timing and Pace Considerations
Potential Evolution in Planning Practices

36
37

Appendix JI: Baseload Study
Number
Table 1
Figure 1
Table 2
Table 3
Table 4
Figure 2
Figure 3
Figure 4
Table 5
Table 6
Table 7
Table 8
Table 9
Table 10
Figure 5
Figure 6
Figure 7
Figure 8
Table 11

Title
Abbreviated List of NERC Event Category Definitions
Net Capacity Additions, Excluding Wind and Solar
Stability Study – Renewable Generation Additions Modeled
Generation Turned Off In Analysis
Major Faults Analyzed
Stability Plot – Importance of Resource Siting on Stability
Stability Plot – Importance of Resource Siting on System
Response
Incremental Steps to Result in 3-Unit Metro Equivalent
Services
Dispatch Assumptions
Generation Turned-Off in Analysis
Additional Generation Assumptions
Faults Analyzed
Thermal Overload Results
Voltage Violation Results
Stability Plot – Insufficient System Damping Twin Cities Metro
Area Voltage
Stability Plot – Insufficient System Damping
Outside Twin Cities Metro Area Voltage
System Collapse – Summer 2023 P01 Case
System Collapse – Summer 2023 P02 Case
Economic Analysis – Baseload Unit Retirement Date
Assumptions

Page #
7
40
42
43
43
44
45
46
47
48
48
49
51
51
52
53
54
55
57

Appendix J2: Reliability Requirement
Number
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5

July 1, 2019

Title
MISO Local Resource Zones
MISO RIIA Study Finding – Declining ELCC Value for Wind
and Solar Resources
E3 Marginal ELCC (%) – Four-Hour Demand Response
Renewable Output and Load January 26 – February 8, 2019
Xcel Energy Wind Generation Actuals vs. Forecast January 28-

Page #
3
8
9
10
12

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Figure 6
Figure 7
Table 1
Table 2
Figure 8

Table of Contents
Tables and Figures: Appendices – vi

31, 2019
Xcel Energy Solar Generation Actuals vs. Forecast
January 28-31, 2019
Snow Impacting a Large Solar Installation
Hourly Wind and Solar Capacity Factors February 5, 2019
Planning Parameters to Derive NSP System Level of Market
Reliance – High Stress Scenarios
NSP System Reliability Requirement Calculation – 2020
Example

13
14
16
22
23

Appendix K: Xcel Energy Resources: Nuclear
Number
Figure 1

Title
Scenario PVRR Deltas from Reference Case ($2019 millions)

Page #
9

Appendix L: Xcel Energy Resources: Sherco CC
Number
Figure 1

Title
Projected Firm Capacity as Compared To Summer Peak and
Reliability Requirement

Page #
7

Appendix N4: Renewable Energy Compliance Positions
Number
Figure 1
Figure 2

Title
Annual RES & REO REC NSP System Obligation &
Production
State of Minnesota Annual SES REC Obligation & Production

Page #
2
3

Appendix N6: Renewable Energy Standard: Rate Impact Report
Number
Table 1
Table 2
Table 3
Table 4
Table 5
Table 6
Table 7
Table 8
Table 9

July 1, 2019

Title
Annualized Historic RES Rate Impact (2005-2012)
Annualized Future RES Rate Impact (2013-2019)
Annualized Future RES Rate Impact (2020-2026)
Annualized Future RES Rate Impact (2027-2034)
Annualized SES Rate Impact (2017-2026)
Annualized SES Rate Impact (2027-2034)
Levelized RES Rate Impact
Levelized SES Rate Impact
Transmission Projects Attributable to the RES

Page #
8
9
10
11
12
13
14
15
16

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Table of Contents
Tables and Figures: Appendices – vii

Table 10

Transmission Project Costs Attributable to the RES (70% of
MPV Project Annual Revenue Requirements Assigned to RES)

17

Table 11

MISO Data Used in RES Avoided Energy and Capacity
Calculations
MISO Data Used in SES Avoided Energy and Capacity
Calculations

18

Table 12

19

Appendix N10: Nuclear Worker Transition Plan
Number
Table 1

Title
Nuclear Employee Breakdown

Page #
3

Appendix O1: Summary of IRP Stakeholder Engagement
Number
Table 1

July 1, 2019

Title
Stakeholder Workshops

Page #
2

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 1 of 139

2020-2034 UPPER MIDWEST INTEGRATED RESOURCE PLAN
CHAPTER 1
EXECUTIVE SUMMARY
I.

INTRODUCTION

This 2020-2034 Upper Midwest Integrated Resource Plan charts the path toward
achieving some of the most ambitious carbon reduction goals of any utility in the U.S.
Specifically – we aim to reduce carbon emissions 80 percent by 2030, and provide 100
percent carbon-free energy by 2050. This Resource Plan not only reaches the 2030
goal through retirement of our coal fleet, extension of nuclear, aggressive renewable
additions, and demand-side management including both energy efficiency (EE) and
demand response (DR), and a mix of load supporting, firm dispatchable resources – it
embraces technology and innovation and is well-grounded in reliability and
affordability. And while the last stretch of total carbon reduction – from 80 to 100
percent – will require technologies that have not yet been developed or deployed
economically, we are confident that we can work with regulators, policymakers, and
stakeholders to position ourselves so we are prepared to take advantage of the costeffective solutions that emerge over the course of the next 30 years.
Figure 1-1: Projected Carbon Emissions Through 2030

Our Preferred Plan is the product of an unprecedented stakeholder process that
included 13 public workshops, independent expert analysis, and months of
information sharing as we developed a Preferred Plan. As a result of those efforts,
our Preferred Plan is the product of an unusual amount of consensus this early in the
Resource Plan process. That consensus is represented by an agreement signed by the
Company, the Clean Energy Organizations,1 Center for Energy and Environment,
The Clean Energy Organizations include Clean Grid Alliance, Minnesota Center for Environmental Advocacy, Fresh
Energy, and Union of Concerned Scientists.

1

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 2 of 139

Sierra Club, and LIUNA Minnesota and North Dakota that resolves (among those
parties) many of fundamental building blocks of our plan.
Those building blocks include the elimination of coal-fired generation from our
system by 2030, as well as the reduced, seasonal dispatch of Sherco 2 until its
retirement in 2023. The agreement also includes the acquisition of at least 3,000
megawatts (MW) of utility-scale solar by 2030, and a substantial increase in EE
programs, representing an average annual savings of over 780 gigawatt hours (GWh).
Finally, the agreement includes support for the Company’s proposal to take
ownership of the Mankato Energy Center (MEC) combined cycle (CC), which will be
central to our reliability strategy as we retire 2,400 MW of coal and integrate several
gigawatts (GW) of new renewable resources. The Company’s Preferred Plan builds
upon this agreement and adds proposals to operate our carbon-free Monticello
nuclear plant for an additional 10 years beyond its current license, add a significant
amount of DR resources, and construct a new CC at our Sherco site. In total, we
have an ambitious plan that supports the Company’s goal of reducing carbon
emissions 80 percent by 2030, and it moves us toward our ultimate vision of 100
percent carbon-free energy by 2050.
Figure 1-2: Preferred Plan Highlights

Throughout this process, we have taken steps to ensure that we can meet these
progressive carbon reduction goals while preserving the reliability our customers have
enjoyed for decades. To that end, the Company’s engineering and operations teams
have conducted extensive analyses to ensure that we can continue to serve customers
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 3 of 139

every hour of every day, even as we progress toward relying on intermittent resources
for a majority of our generation. In this work, we have aimed to embrace change
while addressing the physical realities of our system and the responsibility that comes
with providing a genuinely essential service.
The addition of several gigawatts of renewable resources requires that we consider not
only our traditional summer peak, but also whether we have sufficient dispatchable
resources to meet other peaks, including in winter when solar energy is typically
unavailable and wind resources may not be available for long periods of time. Our
Preferred Plan addresses these reliability issues in three ways. First, the extension of
Monticello by an additional 10 years and the continued operation of Prairie Island will
anchor our grid in around-the-clock, carbon-free energy. Second, we are proposing to
take ownership of the Mankato Energy Center and build a new CC plant at our
Sherco site in 2026. These dispatchable resources will be critical as we retire 2,400
MW of coal-fired baseload and transition to a system that is nearly 60 percent
renewable and intermittent generation. Finally, we propose several firm dispatchable,
load-supporting resources – but defer these additions until the latter part of the
decade, in anticipation of technological advancements that will improve the
functionality and drive down the cost of resources, like storage, that can take the place
of traditional gas peaking units.
We also recognize that the achievement of our carbon reduction goals will depend on
our ability to keep rates affordable. We believe that our Preferred Plan accomplishes
this by keeping annual cost growth below the rate of inflation. The modest cost of
our plan is facilitated by our strategy of deferring resource additions until later in the
plan and making use of existing assets on our system. Additionally, we believe
technological improvements will continue to drive the costs of renewables down,
which is a key element in our strategy of proposing significant solar additions in the
latter half of the next decade.
We also know that our proposed plan includes impacts both on the communities we
serve and our employees. We appreciate not only the challenge – but the stakes for
those impacted – and we plan to build on our successful track record of working with
our communities, policymakers, stakeholders and employees to successfully manage
this clean energy transition.
We further recognize that the agreement underlying our Preferred Plan is simply the
beginning of a process. And although elements of our Preferred Plan are captured by
the Settlement, the parties to the agreement have not endorsed the entire plan and the
Commission has not yet approved the plan. As a result, we look forward to a healthy
discussion on the best way forward. That said, we view the agreement – which
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 4 of 139

promises the elimination of coal and the new prominence of solar on our system – as
a great foundation from which to work. We believe both the process and outcome of
this collaborative effort are a testament to the regulatory landscape in the states we
serve, and we look forward to continuing the discussion around this transformational
plan and our collective energy future.
II.

EXECUTIVE SUMMARY

In our last Resource Plan (Docket No. E002/RP-15-21), we discussed the rapid
evolution of our industry due to changing technology, enhanced customer
expectations, and the increasing consensus around the importance of carbon
reduction. We also noted that partnership among our stakeholders, communities, and
the Company would become even more important to navigating these changes. In
approving our prior plan, the Commission likewise noted that resource planning is a
collaborative and iterative process and that a full understanding of the relevant facts
requires exposure to the views of engaged and knowledgeable stakeholders.
We are filing this 2020-2034 Upper Midwest Integrated Resource Plan following an
unprecedented stakeholder process that included 13 public workshops with topics
from the evolving resource planning process, to more technical considerations, such
as transmission and system reliability. We also engaged a third-party consultant—
Energy and Environmental Economics, Inc. (E3) to conduct independent, parallel
analysis to inform the Company’s future resource strategy. E3 presented its findings
to a diverse group of stakeholders at a workshop in April 2019. We then presented
our own preliminary Preferred Plan at our final stakeholder workshop in May 2019.
We believe this combination of a significant internal effort, extensive collaboration,
independent expert analysis, and transparency has improved not only the process that
led to the development of our Preferred Plan but also the plan itself. In fact, it was
through this stakeholder engagement that the Company, the Clean Energy
Organizations,2 Center for Energy and Environment, Sierra Club, and LIUNA
Minnesota and North Dakota were able to reach an agreement that addressed many
of the cornerstones of our Preferred Plan, including: (1) retirement of our last two
coal units by 2030; (2) seasonal dispatch of Sherco 2 until its retirement in 2023; (2)
acquisition of the MEC CC; (3) acquisition of at least 3,000 MW of utility-scale solar
by 2030; and (4) a substantial EE goal.
We acknowledge that this agreement is just the start of the process – a process that
The Clean Energy Organizations include Clean Grid Alliance, Minnesota Center for Environmental Advocacy, Fresh
Energy, and Union of Concerned Scientists.

2

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 5 of 139

began with the Commission and its request to conduct a holistic review of our
baseload resources. As we return to the Commission and begin to engage with the
Commission directly in this Integrated Resource Plan docket, we look forward to the
opportunity to demonstrate the substantial benefits of our Preferred Plan. It is also
true that the terms of the agreement outlined above do not cover all components of
our Preferred Plan, and we recognize that stakeholders continue to have wide-ranging
perspectives on our collective energy future. We welcome those perspectives as part
of this process, and we look forward to more collaboration and iteration as this
docket moves forward. That said, we view the agreement as a very good start and a
positive outcome from our stakeholder process; we appreciate the Commission
setting us on the path; and, we believe the agreement demonstrates that stakeholders
and the Company can find common ground and build consensus around key building
blocks of a plan that satisfies the needs of our five-state Upper Midwest region – and
meets individual state goals as well. Indeed, meeting the varied interests of our
integrated system was an important foundation of our planning process.
Both the agreement and our overall Preferred Plan are consistent with the Company’s
environmental goals. For more than a decade, Xcel Energy has been a leading wind
energy provider in the nation and has pursued a successful strategy to transition to
clean energy. We have surpassed both national and international goals, including the
U.S. commitment under the Paris Climate Accord of 26-28 percent reduction in
carbon emissions by 2025. To-date, we have reduced carbon emissions 38 percent
companywide from 2005 levels. We are proud of these achievements and grateful to
our many stakeholders who have played a role in our journey.
In December 2018, the Company expanded on its commitment to clean energy by
announcing industry-leading goals to reduce carbon emissions 80 percent Companywide by 2030,3 and to provide 100 percent carbon-free electricity across our service
territory by 2050. This 2020-2034 Upper Midwest Integrated Resource Plan charts a
path to accomplishing these goals through the elimination of all coal generation on
our system by 2030, the addition of over 5,000 MW of renewables, and the expansion
of our industry leading EE and DR programs. It accomplishes these environmental
milestones while not sacrificing operational reliability or affordability. Specifically, we
propose to do the following:


3

Coal Resources - Retire our last two units early: King in 2028 (nine years

early) and Sherco 3 in 2030 (ten years early). Additionally, continue our plan to
retire Sherco 1 and 2 in 2026 and 2023, respectively, and commit to offering
Sherco Unit 2 into Midcontinent Independent System Operator (MISO) on a
seasonal basis until its retirement.

From 2005 levels.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 6 of 139

 Nuclear Resources- Operate our Monticello unit through 2040 (10 years
longer than its current license) and operate both Prairie Island units through
the end of their current licenses (PI Unit 1 to 2033 and PI Unit 2 to 2034).4
 Renewable Resources – While the exact wind and solar mix could vary based
on a variety of reasons, at this time we propose to add 4,000 MW of cumulative
utility scale resources by 2034 (the first being in 2025) and approximately 1,200
MW of cumulative wind by 2034 to replace wind that is set to retire from our
system during that period.
 Combined Cycle Resources – Acquire and operate MEC and build, own and
operate the Sherco CC to satisfy significant capacity and operational need
created by coal closures.
 Firm Load Supporting Resources – Starting in 2031, add approximately
1,700 MW of cumulative firm dispatchable, load-supporting resources by 2034.
 Demand Side Management (DSM) - Include EE programs representing
approximately 780 GWh of savings annually through 2034 (compared to
average annual energy savings of 444 GWh in our last Resource Plan) and the
addition of 400 MW of incremental DR by 2023 with a total of over 1,500 MW
DR by 2034.
This plan demonstrates that we can achieve our 2030 goal with existing technologies
and resources while maintaining both reliability and affordability.5 However, it also
creates opportunities to introduce emerging technologies as part of the solution. We
see opportunities for innovation in our ongoing EE and DR programs. Likewise, we
believe the industry will deliver new and improved technologies that will support our
long-term need for firm, load supporting resources. The plan also advances a
framework that achieves these goals in manageable steps as opposed to transitioning
the entire system and grid all at once. By doing so, we can continue to ensure the
reliability of our system and maintain flexibility to respond to future market trends,
technology advancements, and changing regulatory policies
Below, we discuss our proposed resource mix further, as well as the priorities and
considerations that drove the development of our plan.

Given that our operating licenses for Prairie Island run until 2033 and 2034, we believe there is sufficient time to
address the future of that plant in upcoming resource plans.
5 As we explained our December 2018 announcement, we recognize that serving customers with 100 percent carbonfree electricity will likely require technologies not yet commercial available, and we look forward to discussing these
technological developments in future resource plans.
4

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

A.

Resource Plan – Chapter 1
Page 7 of 139

Proposed Resource Mix

Our Preferred Plan reflects a significant transformation of our resources. We have
more than 1,300 MW of energy resources subject to power purchase contracts that
are expiring. Our plan is also informed by an extensive study of all of our baseload
resources, completed in response to the Commission’s last Resource Plan Order.
That study included seven Attachment Y2 studies by MISO and a more traditional
NERC-based analysis of our fleet by an external consultant. All of these potential
retirements were then studied in conjunction with the addition of significant
renewable resources needed to meet our 80-by-30 goal, which identified reliability and
stability issues that will need to be resolved as we move through the planning period.
As a result of this work, our Preferred Plan takes a measured approach to adding and
retiring resources, and it prioritizes reliability and long-term system planning – as it
must. In the first five years, we have no incremental capacity needs and propose only
minimal additions.6 In fact, there are no significant resource additions until 2025
when our first utility-scale solar is proposed. By relying on our existing resources in
the near term, we preserve flexibility to respond to changing customer needs and
regulatory policies, and we can monitor technological change to ensure we make
future resource investments at the speed of value when they are in the best interest of
our customers. We will continue our aggressive support of EE and DR and are
looking to emerging resources to be part of that solution.

6 Our actions in the next five years will address previously approved or pending resource additions and retirements, wind
repowering and procurement to meet specific customer or program needs, community solar garden growth, and DSM
programs.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 8 of 139

Figure 1-3: Preferred Plan Energy Mix through 2034

That said, in light of the potential baseload retirements and expiring power contracts,
we must address nearly 75 percent of the energy-producing resources on the NSP
System during the 15-year planning horizon. We developed the Preferred Plan with
an eye toward maximizing cost-effective renewable resources, backed by natural gas to
support renewable integration and system reliability, in an effort to minimize market
and commodity exposure. By doing so, our system will not be overly reliant on any
one fuel source, and we will retain our trademark reliability – along with the flexibility
to consider the economics of new resources as our baseload plants retire.
We discuss the components of our proposed resource mix in greater detail below.
1.

Coal

With respect to coal-fired generation, our 2020-2034 Resource Plan represents a
monumental step forward in transitioning our fleet. Today, as a result of our
agreement with the Clean Energy Organizations, Center for Energy and
Environment, Sierra Club, and LIUNA, we are proposing to retire our King plant in
2028 and Sherco 3 in 2030 – meaning that Xcel Energy will complete its transition
away from coal-fired generation in 2030 – a full decade earlier than previously
anticipated. In total, we plan retire approximately 2,400 MW of coal-fired generation
in the next decade.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 9 of 139

The early retirement of these plants allows us to reduce and ultimately eliminate our
reliance on coal, enable additional cost-effective renewable resources, and save
customers money. In addition to these retirements and the early retirements of
Sherco Units 1 and 2 approved in our 2015 Resource Plan, we are also proposing to
offer Sherco Unit 2 into MISO on a seasonal basis until its retirement in 2023, which
we expect will reduce its carbon emissions in the near term.
This accelerated transition away from coal requires the Company to plan for the
retirement of 2,400 MW of coal-fired generation in the next decade, which represents
almost one-fourth of the total capacity in our current generation fleet. This will be an
unprecedented period of transition for our system that necessitates a prudent
replacement strategy. Our strategy for replacing these MWs includes a combination
of natural gas CC resources, continued reliance on nuclear generation, large renewable
additions during the planning period, and a continued commitment to both EE and
DR, all of which will be critical to maintaining reliability throughout this baseload
transition. We discuss each in turn below.
2.

Nuclear

Carbon-free nuclear generation has been a cornerstone of our generation fleet for
nearly half a century. Today, our nuclear plants generate about half the carbon-free
energy for our Upper Midwest customers – amounting to the avoidance of about 7
million metric tons of carbon dioxide annually. This is equivalent to removing 1.5
million cars from the road. Our nuclear fleet is therefore critical to meeting our “80by-30” goal and maintaining that level into the future.
Our nuclear units enable the Company to achieve and maintain our carbon reduction
goals while incorporating incremental renewables at a reasonable pace and
maintaining reliability. Nuclear is also an important system resource during the winter
months, as it does not experience fuel supply issues and has a great track record
during cold weather events – making it a critical piece of our reliability strategy, which
we discuss below.
In light of these considerations and others discussed later in this filing, our Preferred
Plan includes operating our Monticello nuclear plant until 2040, along with the
continued operation of Prairie Island through its current operating licenses (which
expire outside the planning period of this Resource Plan, in 2033 and 2034). By
continuing the operation of these plants and extending our Monticello license, we can
continue to enjoy the substantial carbon-free benefits these baseload units provide
while saving our customers money by leveraging existing assets on our system.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 10 of 139

Absent a Monticello operating extension, based on the reliability needs of the system,
any suitable replacement resource would add carbon to our portfolio. We simply
could not maintain our system reliably, or affordably, given the massive renewable
additions and corresponding transmission infrastructure that would be required to
replace our Monticello nuclear plant, if it were even possible by 2030, given MISO’s
current transmission expansion issues.
The recommendation to extend the Monticello unit is supported by its operational
performance, which has achieved an average capacity factor of 96.5 percent over the
past three years (including a record-setting 99.3 percent in 2018). Moreover, we
achieved this performance all while reducing production costs by more than 20
percent since 2015. We believe this performance demonstrates that we can achieve
deep carbon reduction along with industry-leading safety and reliability at an
affordable cost. For all of these reasons, our nuclear strategy is sound and is in our
customers’ best interest and consistent with the public interest.
Procedurally, we intend to bring a petition for a Certificate of Need (CON) to address
the Monticello license extension request to the Commission in the coming years. In
that filing, we will provide detailed capital budgets and O&M forecasts, as well as
economic modeling to justify our request. Given that the Prairie Island Units’ licenses
do not expire until 2033 and 2034, we believe we have time to address the future of
these units in our next Resource Plan. We look forward to engaging with the Prairie
Island Indian Community, Monticello, and Red Wing as we begin a discussion about
the role of nuclear in our energy future.
3.

Renewable Resources

Substantial renewable additions are a central component of our energy future and thus
a cornerstone of this Preferred Plan – which proposes to add 4,000 MW of costeffective, utility-scale solar generation and approximately 1,200 MW of cumulative
wind resource additions. While the exact mix of wind and solar added to our system
may vary (in concert with a variety of factors including technology advancements and
price changes), our commitment to renewable energy will not.
In total, our Preferred Plan envisions a system that is approximately 60 percent
renewable energy – a level that puts us among those leading the nation. And, while
we are confident in our ability to deliver on our reliability commitment at this high
level of renewable penetration, we are somewhat cautious at the same time about
going much beyond those levels in light of our own experience, as well as recent
industry studies regarding the complexity and complications of an exceedingly high

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 11 of 139

renewable grid.7 That said, some of our customers and municipalities have
environmental goals that include the achievement of 100 percent renewable energy to
meet their needs, and we are confident we can meet those needs given the substantial
renewable additions proposed in this Resource Plan.
The capacity value of renewables combines with our cost-effective gas and nuclear
generation to deliver safe and reliable service that will withstand the summer and
winter peaks of the Upper Midwest. Significantly, with these additions, there would be
enough solar generation to power more than 650,000 homes each year.
Wind generation also continues to play a prominent role in this Resource Plan. Xcel
Energy has long been one of the nation’s leading providers of wind energy, and we
are currently engaged in the largest build-out of new wind resources in our Company’s
history – thanks in large part to the Commission’s approval of our last Resource Plan
and our 1,850 MW wind portfolio. By 2024, wind will provide approximately 35
percent of the electricity for our customers in this region, making it the largest
component of our overall generation portfolio.
4.

Combined Cycle Resources

In addition to our carbon-free nuclear baseload resources, the continuation of
dispatchable generation on our system will be vital to our ability to manage the
retirement of approximately 2,400 MW of coal-fired generation over the next decade
while maintaining reliability. It will also facilitate our ability to successfully integrate
large amounts of renewables; we can ramp the output of these resources up or down
in response to our system’s changing needs throughout the day, as renewable
resources generate more or less energy due to their variable nature. Finally,
dispatchable generation will also help us plan for the expected marginal decline in load
carrying capability from renewables as their penetration increases, which we believe
could result in additional capacity needs.
To that end, our Preferred Plan includes our acquisition of MEC (a 760 MW two-unit
CC), as proposed in Docket No. IP6949,E002/PA-18-702,8 as well as our plan to
build the approximately 800 MW Sherco CC located in Becker, Minnesota in the mid2020s. As discussed in the pending MEC docket, that plant is already an integral part
of our system, as its output is committed to the Company through two Commissionapproved PPAs. By securing ownership of the plant, we can mitigate the risk
See https://twin-cities.umn.edu/news-events/research-brief-planning-future-energy-demand-renewable-energy and
MISO’s Renewable Integration Impact Assessment (RIIA), which we discuss in Appendix J2: Reliability Requirement.
8 We will incorporate any Commission decision from that docket into our modeling and supplement the record as
necessary.
7

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 12 of 139

associated with expiration of the first PPA in 2026, thereby achieving additional
certainty with respect to capacity and dispatchable energy.
As discussed in our last Resource Plan, we propose to locate a CC at the existing
Sherco site because it will allow us to cost-effectively address significant transmission
issues identified by the MISO Attachment Y2 study, ensure the stability and reliability
of the transmission system, mitigate impacts to the local community and our
employees, and potentially provide improved access to natural gas supplies for
communities in Central Minnesota.
Together, our MEC acquisition and constructing the Sherco CC will not materially
impact the amount of gas generation on our system. As already discussed, MEC is
already an existing resource on our system, and the Sherco CC will primarily offset the
retirement of other gas generation on our system, including the Cottage Grove facility
(approximately 250 MW in 2027) and Black Dog 5 (approximately 300 MW in 2032).
This additional gas generation is not only reasonable, but an operational necessity in
light of the much larger coal retirements planned – and the large amounts of variable
renewable additions we anticipate in the same period.
5.

Load Supporting Resources

Reliability is the bedrock of any resource plan. We are particularly focused on the
reliability of our system in this plan, however, as we plan for such a large turnover of
our baseload fleet and transition to a portfolio that is approximately 60 percent
renewable and intermittent generation. We recognize that our transition to cleaner
energy will only be successful if we can execute our vision without disrupting our
customers’ lives and businesses, so we are steadfastly committed to maintaining our
performance when it comes to this core tenant of our business.
Based on the results of extensive reliability studies that we discuss further below, we
are proposing approximately 1,700 MW of cumulative additions of firm dispatchable,
load supporting resources from 2031-2034. The need for these dispatchable
resources emerges in this later timeframe due to the major plant retirements already
discussed, as well as the expiration of several PPAs. Our reliability analysis
demonstrates that these additions are necessary to continue to support grid reliability
and resiliency in light of the increased renewables being added to the system and the
baseload units being retired. That said, because these units are not needed until the
out years of our current plan, we have not identified a specific resource type to meet
this need. However, with the expected price declines and technology development,
between now and the 2030s, we fully expect utility-scale storage will be an integral
resource used to meet this need. Likewise, we believe the deployment of advanced
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 13 of 139

grid investments could position DR to better compete with traditional generation to
fill some of this firm dispatchable need. We are committed to pursuing all of these
options not only in the longer term, but in the near term as well in order to position
ourselves to leverage this technology as it matures.
In addition, as discussed in our last Resource Plan, system retirements will impact our
current blackstart plans and we are currently analyzing our blackstart path to
determine the best fit for our system needs. While we do not propose any action
related to the system blackstart at this time, we anticipate addressing this in our next
Resource Plan or earlier, if system needs dictate the need to do so.
By keeping options open and remaining technology agnostic, we can acknowledge the
need for a firm resource at the tail end of our plan but allow the market to advance as
we file future resource plans and continue to collaborate with our stakeholders and
the Commission as the need for these resources begins to materialize.
In the meantime, we are analyzing potential locations and sizing of storage solutions
as well as the potential values storage assets might provide to the system.
6.

Energy Efficiency

Our Preferred Plan also proposes to add significant amounts of EE based on the
December 2018 Minnesota Energy Efficiency Potential Study: 2020-2029. In fact, our
proposal includes an annual average of over 780 GWh of savings for 2020-2034. Our
last Resource Plan included 1.5 percent annual EE savings assumption, but our
current proposal achieves much higher levels of savings – ranging from approximately
2 to 2.5 percent annually. Relative to a 1.5 percent assumed savings level, our
proposal achieves more than 200 MW of additional demand savings by 2023, and
more than 800 MW by 2034.
7.

Demand Response

Finally, consistent with the Commission’s Order in our last Resource Plan, our
Preferred Plan proposes to add 400 MW of incremental DR by 2023 (with a total of
over 1,500 MW of DR by 2034). When it comes to DR, the Company leads the way
in MISO, with 830 MW registered in the current planning year. In the last Resource
Plan, the Commission ordered the addition of 400 MW of incremental DR by 2023.
As we understood the Commission’s reasoning, it sought to add incremental, cost
effective DR to avoid near-term reliance on additional combustion turbines. As can
be seen in our analysis, however, no combustion turbines or other firm, dispatchable
resource additions are required until the 2031 timeframe as the model instead prefers
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 14 of 139

solar additions as the most attractive resource in the 2025-2030 timeframe.
That said, we decided to include the DR in our Preferred Plan for several reasons: (1)
to be consistent with the Commission’s Order in our last Resource Plan, (2) to fill
gaps if/when the solar capacity credit declines, (3) to help meet firm dispatchable
resource needs in the 2030s, (4) to help support customer programs, and (5) to
integrate new and emerging technology and tools. We note that for purposes of our
modeling, we have included all of the DR identified in the Brattle study as costeffective, including expansions to conventional DR programs (i.e., Savers Switch,
smart thermostats, and interruptible rates) and a non-conventional smart electric
water heater program. Additionally, we included the addition of Auto DR, another
non-conventional DR program that automates control of various end-uses like HVAC
and lighting. We believe the advancement of our grid and technology generally may
take the form of less traditional DR, so we are requesting the flexibility to evaluate
and pursue the required incremental DR through a variety of means and technologies
over the coming years.
In this filing, our objective is to bring forward information on all of the viable options
so the Commission, stakeholders, and the Company can engage in an informed
exchange.
B.

Plan Priorities
1.

Reliability

The foundation of our business is providing safe and reliable electric service, and the
purpose of a Resource Plan is to identify the appropriate resources to continue
providing that service to our customers. Building on the reliability and stability issues
identified as part of our Baseload Study and renewable integration work, and
recognizing that many other utilities within the MISO planning area are also planning
to retire their baseload units, we made reliability and resilience a primary consideration
of this Resource Plan.
To that end, we have conducted a detailed analysis of what resources will be necessary
over the full planning period – once many of our baseload units are retired and the
renewable resources have taken their place as our primary source of generation. As
part of that work, we have paid increased attention to analyses around our winter
peak, when solar is diminished and wind facilities can also drop off as a result of
extreme temperatures. That analysis points to a baseline operational level of firm
resources needed to continue to support a reliable and resilient grid at all hours of the
day, on all days of the year. This operational guidance was then used in our modeling
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 15 of 139

tool, Strategist, to inform the resource decisions and ensure that all resource mixes we
considered would be operationally feasible and reliable to meet our ongoing need to
serve our customers. Below, we summarize how we determined the appropriate
operational requirement.
Within a large pool of generation resources and an established wholesale energy
market like MISO, there is a tendency for market members to project reliance on
market resources based on the size of that pool, rather than the specific performance
of those resources and the capabilities of the overall system to deliver additional
resources. As we move further into a future that relies less on centralized and
dispatchable generation resources, these operational considerations around system
and resource capabilities become exponentially more important. In other words, as
renewable penetration increases throughout the MISO footprint, it becomes
increasingly important to consider the variable nature of these resources and their
effect on the overall pool when considering reliability and market reliance. Thus, while
we can, and do, still rely on the market, that reliance should be tempered during
extreme events, because the nature of these events is such that they tend to impact a
geographical footprint that is broader than a single plant or transmission line outage.
While MISO is working to address these transmission needs, there is a clear need for
more collaboration to enable transmission capability to help support the market’s
ability to facilitate carbon-free objectives going forward.
Due to the variability of renewable generation, the current generation fleet encounters
times in which Net Load (defined as the difference between gross demand and
renewable generation supply at a given point in time) is near, or even equal to, the
gross demand on the system. This is evident in extreme cases, such as the 2019 polar
vortex (when MISO used an average of 6,500 MW resource “reserves9” to remain
operational), but also during normal winter operations like February 5, 2019, which
was representative of conditions we typically experience throughout the winter
season. For instance, on February 5th, the system encountered 16 hours of demand
greater than 5,500 MW (60 percent of annual peak demand). During this same time,
the Net Load was above 5,400 MW, with wind and solar together producing only 6
percent of their installed nameplate capacity (dipping at certain hours to 3 percent).
Another example, on July 29, 2018, the entire MISO wind portfolio (over 17,000 MW
at that time) had a combined output of minus 11 MW – meaning the wind turbines
that were online, were taking more power than they were producing. This hour was
part of an approximately 110 hour sustained stretch in which the combined output of
all wind resources in the MISO footprint fell well below the accredited values used in
These reserves consisted of non-firm resources offered by neighboring regional transmission organizations into the
MISO market
9

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 16 of 139

present planning processes.
This real-world experience reveals several operational truths:
 First, variable resources cannot meet demand in all hours of the year; firm
dispatchable resources are necessary.
 Second, simply increasing the level of renewables on the system cannot address
resource shortfalls. With an increased level of renewables, we see some
improved ability to meet demand but still encounter several hours in which the
net load is very close to the gross load. In fact, the amount of additional
renewable generation that would be required to meet customer demand in the
above scenarios and without other resources could be in excess of 180,000
MW.
 Third, our ability to rely on the MISO market during winter peaking events is
limited by periods of extremely low renewable generation across the MISO
footprint and a shortfall of these resources compared to their accredited
capacity.
 Finally, the current state of battery storage technology does not have the ability
to match the duration of such events without significant (and very expensive)
over-build of those resources, and DSM programs also lack the scale to
significantly impact the analysis.
In light of these issues, we have determined that sufficient firm, dispatchable
resources are required to meet the approximate 6,400 MW winter peak load
obligation, and we have imposed this requirement in our Strategist modeling as part of
this Resource Plan. Figure 4 below demonstrates the calculation of the firm resources
used meet this need.
Figure 1-4: NSP System Reliability Requirement Calculation

Our analysis shows that these resources will help us match the net load gaps discussed
above by ensuring that we maintain a stable and reliable energy system for customers,
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 1
Page 17 of 139

while moving through our baseload transition and achieving our nation-leading
carbon goals. We discuss our reliability and operational analysis in greater detail later
in this filing.
2.

Affordability

Another priority for Xcel Energy, and our Resource Plan, is energy affordability.
Currently, the average monthly Minnesota Xcel Energy residential customer’s
electricity bill is below the national average. Our goal is to keep bill increases at or
below the rate of inflation – and this Resource Plan positions us well for success. In
fact, our Preferred Plan achieves over 80 percent carbon reductions (from 2005 levels)
for a nominal customer cost of just over one percent Compound Annual Growth
Rate (CAGR) over the plan period. The opportunity to achieve such significant
reductions in our carbon emissions for a nominal increase in cost is one of the
principal benefits of our Preferred Plan. The following graph shows the relative cost
growth of our Preferred Plan in comparison to the national average:
Figure 1-5: Preferred Plan Average Rate Impact for the NSP System

Nominal $ cents/kWh

18
16
14
12
10
8

2020

2022

2024

NSP System Base Case

2026

2028

2030

2032

NSP System Preferred Plan

2034

2036

2038

2040

EIA Forecasted National Average

To be clear, the resources the Company needs to add over the next 15 years to
continue providing safe and reliable service, to comply with state energy requirements,
and to address plant retirements and PPA expirations come at some cost. But we
believe that cost – which keeps rates at or below the level of inflation – is both
modest and appropriate compared to the substantial benefits we have described here.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

III.

Resource Plan – Chapter 1
Page 18 of 139

CONCLUSION

Our Preferred Plan – which accounts for more variables and changes than any other
previous Xcel Energy resource plan – proposes to eliminate coal, add even more
renewables, and continue our industry-leading EE and DR programs, all while
preserving reliability and affordability for our customers. It also meets the varied
interests of our five-state Upper Midwest region. And by planning ahead and charting
an orderly, gradual transition of our generation fleet, we believe we can achieve all of
these goals while managing the impacts to our host communities and employees,
preserving the reliability and stability of our system, and maintaining affordability for
our customers. For these reasons, and those discussed throughout this filing, we
believe our Preferred Plan is in the public interest and merits Commission approval.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 19 of 139

CHAPTER 2
PLANNING LANDSCAPE
I.

INTRODUCTION

Northern States Power Company, doing business as Xcel Energy, submits to the
Minnesota Public Utilities Commission this 2020–2034 Upper Midwest Integrated
Resource Plan, which eliminates coal-based generation from our system by 2030,
proposes to add thousands of megawatts of renewable resources – and charts the path
toward achieving some of the most ambitious carbon reduction goals of any utility in
the United States.
Northern States Power Company-Minnesota is a wholly-owned operating subsidiary
of Xcel Energy, Inc. that owns and operates, in conjunction with its affiliate Northern
States Power Company-Wisconsin, the integrated NSP System of generation and
transmission assets that serves more than 1.8 million customers in Michigan,
Minnesota, North Dakota, South Dakota, and Wisconsin. This Resource Plan builds
on our strong foundation of cost-effective environmental performance and the
generating fleet transition we began in our last Resource Plan.
Our plan is founded on unprecedented levels of stakeholder engagement and
technical analyses that examined an orderly retirement of our baseload generating
units. We analyzed numerous assumptions and sensitivities to identify the plan that
best meets customer needs, achieves our obligations and goals, and ensures we
maintain a resilient and reliable grid. The Preferred Plan we propose emerged as the
best suite of resources that balances our planning objectives, as follows:
Figure 2-1: Xcel Energy Integrated Resource Plan Objectives

To understand our Preferred Plan, we first present a Reference Case. The Reference
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 20 of 139

Case is the baseline scenario identifying the resources necessary to continue meeting
our customers’ needs, comply with renewable energy requirements, achieve our 80
percent CO2 reduction from 2005 levels objective, add 400 MW of incremental
Demand Response (DR) consistent with the Commission’s Order in our last plan, and
achieve the significant EE targets identified in the Minnesota Energy Efficiency Potential
Study.1
The Commission’s Rules provide the factors to consider in issuing its findings of fact
and conclusions.2 In addition to considering the characteristics of the available
resource options and of the Preferred Plan as a whole, resource options and plans
must be evaluated on their ability to:
A. Maintain or improve the adequacy and reliability of utility service,
B. Keep the customers' bills and the utility's rates as low as practicable, given
regulatory and other constraints,
C. Minimize adverse socioeconomic effects and adverse effects upon the
environment,
D. Enhance the utility's ability to respond to changes in the financial, social, and
technological factors affecting its operations, and
E. Limit the risk of adverse effects on the utility and its customers from financial,
social, and technological factors that the utility cannot control.
Our Preferred Plan meets these criteria, and provides the flexibility to address the
evolving planning landscape, including the changes we expect to the NSP System to
achieve our ambitious vision of a 100 percent carbon-free energy mix by 2050. That
said, we respectfully request the Commission to approve our Preferred Plan, as
follows:
 Coal Resources. Retire our last two units early: King in 2028 and Sherco Unit
3 in 2030. Additionally, continue our plan to retire Sherco 1 and 2 in 2026 and
2023, respectively.
o Our plan also commits to offer Sherco Unit 2 into MISO on a seasonal

basis until its retirement, and working with our employees at, and the
communities around, the Sherco and King plants to support them
through the transition of our remaining coal fleet.

 Nuclear Resources. Operate our Monticello unit through 2040 (10 years
longer than its current license) and operate both Prairie Island units through

1

2

Available at: http://mn.gov/commerce-stat/pdfs/mn-energy-efficiency-potential-study.pdf

Minn. R. 7843.0500, subp. 3

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 21 of 139

the end of their current licenses (PI Unit 1 to 2033 and PI Unit 2 to 2034).3
o We expect to initiate a Certificate of Need proceeding with the
Commission within the next five years and begin working toward license
extension with the Nuclear Regulatory Commission during this
timeframe.
 Renewable Resources. While the exact wind and solar mix could vary based
on a variety of reasons, at this time we propose to add 4,000 MW of cumulative
utility scale resources by 2034 (the first being in 2025) and approximately 1,200
MW of cumulative wind by 2034 to replace wind that is set to retire from our
system during that period. We specifically request flexibility in the timing and
amounts of renewable additions to take advantage of market and other
conditions that we believe will provide value to our customers.
o Wind. We are committed to pursuing repowering and/or contract
extension opportunities for this resource and we intend to pursue
incremental wind resources as needed to meet customer needs in
growing customer programs like Renewable*Connect.
o Solar. Our Preferred Plan proposes an initial planned addition of 500
MW in 2025. Our plan includes forecasted growth of distributed solar.
If actual distributed solar capacity exceeds our expectations, we
anticipate this will displace a portion of our proposed grid-scale solar
resources.
 Combined Cycle Resources. Acquire and operate the Mankato Energy
Center and build, own and operate (Sherco CC) to satisfy significant capacity
and operational need created by coal closures.4
 Firm Load Supporting Resources. Extend the life of Blue Lake Units 1-4
through 2020-2023.5 Add approximately 1,700 MW of cumulative firm
dispatchable, load-supporting resources between 2031-2034. Because these
units are not needed until the out-years of our current plan, we have not
identified a specific resource type to meet this need. With the expected price
declines and technology development between now and the 2030s, we believe
utility-scale storage will be an integral resource used to meet this need.
 Demand Side Management. EE programs representing 2-2.5 percent of
savings annually (over 780 GWh for each of 2020-2034), compared to average
Given that our operating licenses for Prairie Island run until 2033 and 2034, we believe there is sufficient time to
address the future of that plant in upcoming resource plans.
4
MEC is currently pending Commission consideration in Docket No. IP6949, E002/PA-18-702. For the Sherco CC, we
expect to submit our plans in a separate proceeding.
5
Pending decision in Docket E,G002/D-19-161.
3

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 22 of 139

annual energy savings of 444 GWh in our last Resource Plan, and the addition
of 400 MW of DR by 2023. We are requesting the flexibility to evaluate and
pursue the required incremental DR through a variety of means and
technologies over the coming years.
 Storage and Other Emerging Technologies. Pursue storage and other
emerging technologies, on both a large and small scale.
Finally, as we have previously discussed, system retirements will impact our current
blackstart plans, or our ability to restart the system in the event of a catastrophic
failure. While we do not propose any action related to the system blackstart at this
time, we anticipate addressing this in our next Resource Plan or earlier, if system
needs dictate the need to do so.
The balance of our Resource Plan discusses the evolving planning landscape, presents
our Preferred Plan and the economic modeling framework from which it emerged,
and estimated customer cost and rate impacts. We additionally provide Appendices
that include the results of our Baseload Study, Load and Distributed Energy Resource
(DER) forecasts, and discussion about our Supporting Transmission and Distribution
Infrastructure, Environmental Regulations and Compliance, and numerous others.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

II.

Resource Plan – Chapter 2
Page 23 of 139

PLANNING LANDSCAPE

Every business that conducts long term planning needs to account for both its
internal goals and the external environment within which it operates. In this Chapter,
we discuss some of the key internal and external market contexts that affect how we
have developed, and plan to execute on, our Preferred Plan – which supports our
ambitious carbon reduction vision, to reduce our carbon emissions to 80 percent
below 2005 levels by 2030.
Specifically in this section we examine:
 Xcel Energy’s Carbon Reduction Goals
 Regional Reliability and Market Constructs
 Distributed Energy Resources (DER)
 Community and Employee Considerations
 Customer Preferences
 Supply and Technology Trends
 Jurisdictional Updates
While all of these factors affect how we develop our plan, a few stand out above
others as being particularly influential in this Integrated Resource Plan cycle – chief
among them, regional market constructs and renewable integration. While the
regional system operator that designs many of our market and planning requirements
continues to examine the effects of high renewable adoption on the grid, it has not yet
developed robust and forward-looking capacity accreditation constructs to account
for how renewables’ contributions to peak demand will change over time. This
introduces complexity to designing a plan far into the future, and how we carry out
those plans.
Likewise, while we are committed to substantially increasing renewables on our
system to achieve our carbon reduction goals, we also anticipate facing challenges to
integrating this new clean generation, given the delayed interconnection studies and
current limited state of open transmission availability. Our ability to connect these
new renewables in a cost-effective manner depends materially on constructs that
enable careful management of our interconnection rights in the near-to-medium term
as well as new transmission in the long term.
These and other factors, such as DER adoption rates, community and employee
impacts, and satisfying the needs of five different states, all affect how we developed
the Preferred Plan presented in this filing – and the issues we anticipate encountering
as we pursue our goals to lead the energy transition while keeping our grid services
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 24 of 139

reliable and affordable.
A.

Carbon Reduction Goals

In December 2018, the Company announced its goals to reduce carbon dioxide (CO2)
emissions 80 percent by 2030 below 2005 levels companywide, and to serve
customers with 100 percent carbon-free electricity by 2050. We believe our 2030 goal
is achievable with the clean generation and energy storage technologies available
today. We believe our 2050 vision, however, will be achievable only with
advancements in new technologies such as: carbon-free dispatchable generation
technologies and longer-duration storage that are not currently available at the
necessary scale and cost, or carbon capture and sequestration. Until these or other
technologies are further developed and commercialized, we will require a certain
amount of conventional flexible and dispatchable generation to integrate increasing
levels of renewables on the grid.
To achieve our 80 percent reduction by 2030 goal, we anticipate the following
elements will be essential parts of future plans, across one or more of our service
areas:
 Adding thousands of megawatts of additional renewable resources to our
system and incrementally retiring emitting baseload generation, while also
incorporating flexible, dispatchable generation to enable grid reliability
throughout this transition;
 Operating our carbon-free nuclear units at least through the remainder of their
licenses, with the potential for license extensions;
 Supporting the strategic electrification of certain end uses and enabling flexible
demand, which will help to reduce carbon emissions in other sectors while also
providing flexible loads to help integrate more renewables;
 Investing in critical infrastructure, such as transmission and advanced grid
technology on our distribution system, to integrate the DER our customers
choose, as well as improve reliability and the customer experience.
These goals, the science behind them, and the path we will take to achieving them, are
all detailed further in Appendix E: Xcel Energy Carbon Report:Building a CarbonFree Future.
With these aggressive carbon goals in mind as one of the main tenets of our Preferred
Plan, below we discuss the key forces that affect how we have developed, and plan to
execute on, our Preferred Plan.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

B.

Resource Plan – Chapter 2
Page 25 of 139

Regional Reliability and Market Constructs

The Company’s Upper Midwest system is part of the Midcontinent Independent
System Operator (MISO) market. MISO is charged with several responsibilities, chief
of which are overseeing wholesale energy markets in the member region and planning
for bulk system reliability (i.e. transmission planning, generator interconnection, and
ensuring sufficient reserve margins). Many aspects of MISO’s operations affect how
we conduct resource planning, but here we focus primarily on system reliability
constructs that will be increasingly tested as we and others transition to a fuel mix
relying on more variable renewable resources.
1.

Reserve Margin

One of MISO’s core responsibilities includes administering resource adequacy
requirements to enable utilities like us, and other Load Serving Entities (LSEs), across
the region to fulfill their obligation to serve customers reliably. Trends are emerging,
however, that raise questions regarding how planning constructs will adapt to ensure
the system remains reliable as emitting, but stable, baseload generation continues to
retire and be replaced by clean, but variable, renewable energy.
MISO and its system reliability oversight organization, the North American Electric
Reliability Corporation (NERC) undertake studies to determine the appropriate level
of reserve capacity that should be maintained, what affect a resource retirement has
on the broader system, and how increasing renewable adoption will change how they
analyze and ensure grid reliability. All of these studies point toward an increasingly
complex grid that will have to be carefully managed through the transition to a lowercarbon future.
MISO’s Planning Reserve Margin (PRM) analysis is one important piece of the
current reliability planning paradigm. The PRM is an estimation of how much
generating capacity, over and above expected customer load, needs to be present on
the system to ensure reliability in all but the most extreme circumstances (called a 1in-10 year Loss of Load Expectation or LOLE). In the 2018 report, MISO
established a reference planning reserve margin of 17.1 percent for the 2018-2019
planning year; in other words, they determined that the total installed capacity
available on the system should be 17.1 percent higher than the system’s peak load.6
This reference threshold is provided to the NERC, which sets standards and studies
6 MISO’s PRM for the 2018-2019 Planning Year indicates a PRM for both installed capacity (ICAP) and a rating that
derates capacity to account for potential outages (called UCAP). The UCAP PRM for the 2018 planning year is 8.4
percent These two measures of PRM are discussed further in the next section on Minimum System Needs.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 26 of 139

reliability across the continent, and of which regional system and transmission
operators like MISO are a part.
As part of its oversight and governance activities, NERC conducts a reserve margin
analysis across all system operators in North America, in a report called the Long
Term Reliability Assessment (LTRA). The 2018 LTRA indicated that MISO is one of
three regions that are projected to drop below their reference reserve margin levels by
the year 2023, unless certain measures are taken.7 This report indicates that inclusion
of Tier 2 resources (those that are in more advanced stages of planning but not yet
under construction) would likely allow for the MISO footprint to preserve system
reliability. However, the unprecedented rate of announced, but not yet evaluated,
baseload generation retirements and uncertainty in future firm capacity additions
creates a tension between maintaining reliability and transitioning away from baseload
generation.
It is important to note that retiring some baseload generation and transitioning to a
cleaner grid with more wind and solar does not present an insurmountable challenge;
indeed we are proposing to retire all of our coal by 2030. Rather, the transition will
need to be actively and carefully managed, likely with incremental retirements and
supporting transmission upgrades that are carefully studied.
2.

Renewable Integration Challenges

In addition to challenges around baseload retirement issues, we also see planning
issues developing around how renewable additions are evaluated for their reliability
impacts. In the aggregate, when MISO has studied high levels of renewable
penetration on the grid, grid instability increases and capacity values of variable
resources decline, sometimes significantly. Retaining firm dispatchable generating
units helps ensure the system will continue to operate reliably.
MISO has also recognized that its capacity accreditation framework – the manner by
which it assesses variable renewables’ ability to contribute to peak demand needs –
will likely change as these resources become more prevalent on the grid. However,
MISO has not yet developed sufficiently robust forward guidance for resource
planning processes to account for how those values might change in the future, which
creates uncertainty in the resource planning process. We discuss these renewable
integration studies and specific capacity accreditation issues in more detail below.

See “NERC Long Term Reliability Assessment 2018” at 14. Available at:
https://www.nerc.com/pa/RAPA/ra/Reliability%20Assessments%20DL/NERC_LTRA_2018_12202018.pdf

7

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 27 of 139

a.

Renewable Integration Impact Assessment

In preparation for an expected future grid with high levels of non-dispatchable
renewable penetration and declining baseload generation, MISO is undertaking
additional studies with respect to its system’s reliability and resource adequacy of its
system. The Renewable Integration Impact Assessment (RIIA) seeks to inform future longterm planning by understanding what the power system will need to operate reliably
with these high levels of variable resources, specifically by examining operational
adequacy, transmission adequacy, system stability and resource adequacy limitations.
In Phase I, the study examined a scenario in which variable generation achieves a 40
percent share of the total capacity on the MISO system. It found that the complexity
of operating such a system reliably is significantly higher than that of even a system
with 30 percent variable resources. Under the circumstances studied, the system
experienced more dynamic stability issues and other operational stressors, and
resource adequacy requirements increased. For example, the modeled system
exhibited high levels of energy curtailment and very high ramping rates in the hours
when variable resources were not always available to meet demand. In this scenario,
loss of load projections were narrowed to fewer likely hours during the year, but the
probability of occurrence increased significantly over the current state. This points to
the value that flexible, dispatchable resources supporting grid stability continue to
provide in these scenarios; while they run for fewer hours than the current market
constructs would warrant, they need to be able to respond quickly, moving from
minimum generation levels to higher levels of output to meet these fluctuations in net
load quickly.
Further, at high levels of wind and/or solar adoption, the RIIA study found that
appropriate resource adequacy values to assign these resources degraded, sometimes
significantly from current levels. As a key piece of planning our future system, these
resource adequacy capacity accreditation values are discussed in more depth below.
b.

Capacity Accreditation Values

Variable renewable resources such as wind and solar are becoming more cost
competitive and utilities across the region, including the Company, are increasingly
adopting these technologies as important components of their resource mixes. This
generation is largely displacing more traditional, thermal dispatchable units. As
variable resources are dependent on natural resource availability in a given moment
(i.e. wind blowing or sun shining), their capacity does not replace retiring dispatchable
units one-for-one in terms of the amount of energy it produces, or assurance that this
capacity will produce energy when needed. To account for this, MISO applies a
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 28 of 139

certain accreditation discount to these resources to get a more appropriate
probabilistic view of how much capacity can be counted on to contribute to peak
demand across the year. This is captured in a measure called the Effective Load
Carrying Capability (ELCC).
These administratively set values have a significant impact on how we achieve our
carbon reduction goals while maintaining affordable and reliable service. Currently,
MISO assigns wind generation an average ELCC value of 15.7 percent; meaning that
for every 100 MW of wind installed, only 15.7 MW can be counted as capacity toward
the planning reserve margin. For new solar resources, in the absence of an observed
historical value, MISO assigns the current initial year default ELCC of 50 percent.
The appropriateness of these values in reflecting actual grid conditions is dependent
on the pace at which wind and solar penetration increases on the grid, and
subsequently how MISO conducts review and adjusts the values. The ELCC is
currently evaluated as an annual average, and forward values are not projected.
In reality, however, the capacity value these intermittent resources can provide are
subject to diminishing marginal returns. When a single variable resource type
increases its penetration level on the grid, each incremental unit of capacity inherently
provides a little less capacity benefit to the system than the last unit. For example,
MISO’s RIIA study estimates that solar in particular would experience steep ELCC
reductions within the first 10 gigawatts installed, and this value continues to drop off
at higher levels of adoption. Further, in particular for these variable assets, the realized
capacity value may change throughout the year, as the capacity value a wind or solar
plant can provide reasonably changes in accordance with seasonally variable
environmental conditions.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 29 of 139

Figure 2-2: Modeled Wind and Solar ELCC as Penetration Increases8

The operational realities surrounding future variable resource additions and their
seasonal aspects aside, we continue to use the administratively-set annual average
ELCC levels in our planning that MISO has established for today’s market.
While we recognize that it is difficult for MISO to accurately project future wind and
solar penetration levels and load shapes (two key variables in determining future
ELCC values), this presents a key challenge as we plan our future system. As the
ELCC construct does not currently provide forward-looking values, we have to apply
current values to our resource modeling process, even though we are modeling 15
years into the future. However, we know in reality that these values will degrade as
we and others add variable renewables to the MISO system, and so what appears to
be a net capacity surplus today may look rather different in future assessments.
It is worth noting here that we may encounter changing assumed resource adequacy
contributions for use-limited resources in the future as well. In general, resources
such as DR and energy storage would be subject to declining ELCC values as they
become more prevalent on the system, in the same way wind or solar ELCCs
realistically decline.9 Notably, MISO is also considering changes to how it accounts
MISO. “Renewable Integration Impact Assessment” Workshop presentation June 5, 2018. Available at:
https://cdn.misoenergy.org/20180605%20RIIA%20Workshop%20Presentation213125.pdf
9 See the E3 Study in Appendix P2 for further discussion on how marginal ELCC for DR and energy storage resources
may decline as adoption increases.
8

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 30 of 139

for DR’s capacity accreditation overall, such as enforcing more stringent testing
requirements. MISO is also following up on actual performance during DR events,
which may result in accredited value reductions going forward. Both these factors
mean that the DR we currently depend on as a baseline resource in our portfolio, in
addition to that which we may select in this or future resource plans, may not yield the
same benefits in future years as we have historically expected.
c.

Interconnection Queue

The current state of grid interconnection processes and transmission capabilities in
MISO introduce complexity not only to our planning processes, but also how we
execute on the plan.
The MISO generator interconnection process is designed to allow generators reliable,
non-discriminatory access to the electric transmission system, in a timely manner,
while maintaining transmission system reliability. Recently, as the number of
proposed projects in MISO has expanded significantly, this process has been mired in
delays. Delay impacts are particularly evident in the Definitive Planning Process
(DPP) phases, where MISO undertakes generation interconnection studies. Current
studies are a number of months behind, due to the large volume (including speculative
requests) and a generator interconnection process that allows late withdrawals from
the queue with limited consequences. Despite some recent process reforms, MISO
has not been able to keep pace with the expanding queue. And where projects do
make it through the DPP, they are sometimes assigned high transmission system
upgrade costs that challenge the project’s economic viability. MISO’s interconnection
challenge is multi-faceted.
First, there is a substantial volume of capacity currently in the queue requesting study
and interconnection approval. As of early June 2019 there were over 100 GW of new
capacity in the active MISO queue (although this number has fluctuated substantially),
the vast majority of which is comprised of wind and solar projects.10 Each cycle of
the DPP is handling expanding levels of requested capacity; for example, the recently
completed cycle for the MISO West region (started in August 2016) started out with
31 projects totaling just over 5,600 MW.11 The April 2019 DPP study cycle,

10 MISO “Generator Interconnection: Overview.” Updated as of June 1, 2019, at:
https://cdn.misoenergy.org/GIQ%20Web%20Overview272899.pdf
11
See “MISO DPP 2016 August West Area Phase 1 Study.” Siemens (August 20, 2018). Available at:
https://cdn.misoenergy.org/GI_DPP_2016_Aug_West_Phase1_SIS_Report277263.pdf

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 31 of 139

scheduled to begin in March 2020, includes 58 projects totaling 8,800 GW in the same
area.12 While the level of proposed new renewable project is a positive indication of
aspirational renewable development in the region, MISO has also indicated that a
substantial amount of this capacity is speculative, in early stages of project
development or representing duplicative requests.
Further, the existing transmission system’s capability to interconnect new projects
without substantial infrastructure upgrades is limited, and thus, the generation
interconnection planning studies indicate there will likely be costly upgrades assigned
to the prospective generators. In the past, initiatives such as CapX2020 and MultiValue Projects (MVPs) were able to integrate large quantities of new renewable power
and socialize transmission infrastructure costs across a larger swath of benefitting
MISO customers. However, wind power in particular expanded on the MISO grid
faster than expected, and the interconnection capacity afforded by these projects has
been largely used. Since then, few new transmission lines have been proposed or
approved for the purposes of renewable integration.
Generally speaking, this means that, if new generation projects in the queue want to
interconnect, the generation interconnection study process identifies substantial
additional transmission system upgrade costs and assigns them to the generation
owner(s). In the aforementioned MISO West DPP cycle that recently completed, for
example, the approximately 5,600 MW of proposed projects were expected to incur
approximately $3.2 billion in transmission upgrades, if all were to interconnect to the
system.13 These assigned high-cost transmission system upgrade requirements can
sometimes render projects uneconomic, forcing a queue withdrawal and additional
MISO study on the remaining projects.
d.

Regional Seam Issues

Limitations on transmission infrastructure and coordination, both within MISO and
between MISO and the Southwest Power Pool (SPP), illustrate further challenges.
Within MISO, the transmission system is showing constraints and the resulting
curtailment slows our progress toward a cleaner energy future across the Upper
Midwest system. Currently, wind generation from the western part of MISO flows
toward the load centers in the east such as the Minneapolis–St. Paul area and load
centers across the transmission interconnection between Minnesota and Wisconsin.
However, existing west-to-east transmission capacity is, at times, operating at its limit.
See MISO “Definitive Planning Phase Estimated Schedule.” Updated as of June 1 2019. Available at:
https://cdn.misoenergy.org/Definitive%20Planning%20Phase%20Estimated%20Schedule106547.pdf
13 ”MISO DPP 2016 August West Area Phase 1 Study.” Siemens, September 2018, at xvii.
12

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 32 of 139

The transmission interface across the Minnesota-Wisconsin border in particular is
currently stability-limited, and trying to force additional renewable energy through
these lines could result in voltage collapses in Northern Wisconsin that would
destabilize the grid. Curtailing this energy at its source in the west is operationally and
economically inefficient, keeping us from fully utilizing the inexpensive and clean
energy to which we have access; but, without additional transmission capacity, we will
more frequently encounter this problem as we add more renewable generation to our
system.
Further, coordination (or historical lack thereof) between MISO and SPP introduces
challenges to bringing onto the system, and utilizing, more clean energy. First, for
projects that can be considered interregional in nature, a project must currently meet
economic benefit hurdles in a joint review, as well as separate MISO and SPP regional
evaluations. This slows the process significantly, and may overestimate the amount of
interconnection upgrades required, adding to project uncertainty and cost.
Second, although our load and generation sit within MISO, the nature of power flows
inevitably results in some of our energy entering the SPP system. In turn, both MISO
and SPP may charge to transmit that energy from the point of generation to the load,
challenging a project’s economic viability or raising customer costs for projects
already online.
Finally, MISO and SPP disagree on what should happen when one region or the other
has to “lean” more on the system than its contracted delivery amounts for a certain
time. Where SPP would levy penalties in this scenario, MISO views this situation as a
normal and acceptable result of an integrated grid. All of these issues increase
transaction costs and uncertainty for a given generation project coming online, and
represents a potential barrier to efficiently bringing additional renewable generation to
the grid.
3.

Mitigation Efforts

In response to direction from FERC and in recognition of the challenges described
above, MISO is undertaking several actions that could serve to mitigate challenges to
bringing new, clean resources online. In essence, they allow generation owners to
leverage existing interconnection agreements to maximize utilization and fit renewable
additions into the relatively few remaining open spaces on the grid. While we expect
these processes to mitigate some of the near term challenges to additional renewable
capacity, they do not address all challenges (in particular our ability to depend on
neighboring regions for renewables and maintaining reliability) and we expect that
longer term solutions will eventually need to be developed.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 33 of 139

a.

Generator Replacement Process

Interconnection study delays and speculative queueing are challenges not only to
projects that are actually commercially viable, but also to generation owners that are
looking to retire aging assets. Companies that are required to meet a certain level of
reserve capacity, like Xcel Energy, face potential compliance and commercial risk if
we retire existing assets without the ability to re-utilize that interconnection capacity.
Recognizing these issues, MISO filed and received approval for a proposed
Replacement Generator Process that would allow current generation owners to retain
and re-utilize these interconnection rights where a resource plans to retire, within
certain technical limitations on the new generator’s attributes. The new generating
units could be developed on the same site, or on a site in close proximity that uses the
same grid interconnection point. Per the new MISO tariff provisions, the new
generation resource would need to have an in-service date not later than three years
after the existing generator ceased operation. Importantly, these projects would be
studied outside the traditional DPP timeline, with the intention of avoiding the
significant delays associated with that process, as described above.
b.

FERC 845

In 2018, FERC issued Order 845, Reform of Generator Interconnection Procedures and
Agreements14 that also opens additional opportunities for generation owners to add
resources to the system outside the normal interconnection queue process. The
Order directs all transmission providers to develop a procedure to allow
interconnection customers to use surplus availability at an existing point of
interconnection without that new project entering the full MISO queue and planning
process, within certain technical limitations. MISO has referred to surplus
interconnection availability as “Net Zero” interconnection, as the addition of this new
project would not result in an overall increase to the interconnection capacity
requirements of the site; rather, it would be expected to increase the overall utilization
of the interconnection site.
While MISO allowed Net Zero resources prior to FERC 845, the new Order also
allows existing interconnection rights owners the first right to utilize the surplus
availability on that interconnection. It also revises the definition of a generating facility
to explicitly include energy storage resources. These actions work to support
generation owners increasing renewable utilization on existing interconnections, and
could support future project hybridization (e.g. solar and storage or wind and storage).
We expect that generator replacement, Net Zero, and other FERC Order
14

See Reform of Generator Interconnection Procedures and Agreements, 163 FERC ¶ 61,043 (2018) (Order No. 845).

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 34 of 139

implementation efforts will alleviate some of the barriers to planning and executing on
a future with substantial renewable additions. However, these do not address the
underlying challenges around queue length and timeline, intra-MISO and interregional
seams congestion challenges, and integrating high levels of renewables reliably and
affordably. MISO has recently attempted to mitigate the queue volume challenge by
proposing process reforms that increase the stringency of entering this phase of
interconnection process; however, while recognizing the challenges MISO faces,
FERC recently rejected the proposal.15 While the Company and others have begun
contemplating new MVP-like projects, the lack of alignment across MISO and long
lead-times required for such projects mean that these challenges are unlikely to be
resolved in the near term.
C.

Distributed Energy Resources

At the same time as we work to clean our power supply, we also recognize that
customers are now exercising more choice around how and from where they consume
energy. This is a key consideration as we plan our resource mix for the next 15 years.
Some customers are choosing DER that can reduce customer consumption and even
provide energy back to our system from decentralized locations on the grid. Examples
of DER include, but are not limited to: rooftop solar panels, energy storage,
community solar gardens, or the EE enabled by a smart thermostat or time of use
electric rate. To-date, community solar gardens makes up the clear majority of the
DER on our system in the Upper Midwest.
Our customers’ adoption of DER and new types of load mean that consumption
patterns from our centralized power system are changing. This can represent an
opportunity: if we can harness the benefits of these resources to make demand more
flexible, we can use this to better match demand to energy production from our large,
variable renewable resources. For example, we could utilize managed or “smart”
charging of electric vehicles (EVs), to delay charging to off-peak hours or to times
when renewable output is the highest. We could also use advanced metering
technology alongside customer programs and tariffs to enable load shifting away from
peak hours.
There are also DER coming onto our grid, in the form of electric transportation
options – enabling not only flexible load opportunities but also broader economywide emissions reduction – and we have developed several programs and rate options
See FERC “Order Rejecting Tariff Revisions re: Midcontinent Independent System Operator, Inc. under ER19-637.”
Available at: https://elibrary.ferc.gov/idmws/file_list.asp?accession_num=20190319-3076
15

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 35 of 139

to encourage that adoption.
The transportation electrification initiatives we have implemented and continue to
develop are not only enabling customer choice, but also to the broader economy in
helping to meet greenhouse gas (GHG) emissions reductions goals. The
transportation sector is now the leading contributor to Minnesota’s overall
emissions,16 and as our system (and the state’s electric sector more broadly) continues
to transition to a cleaner energy mix, transportation’s share of GHG emissions will
continue to expand. This shift highlights an opportunity for the electric sector to
facilitate GHG reductions in the transportation sector, as electricity is increasingly
used for transportation fuel.
While the opportunities are exciting, it is also important to recognize that customer
adoption of DER and new types of load behind the meter introduce uncertainties in
our planning processes, particularly if we do not have adequate visibility into how and
when that new DER or demand is coming onto our system.
One tool we have to mitigate this DER and electrification uncertainty is our
Integrated Distribution Plan (IDP) and grid modernization efforts. Our IDP process
and proposed investments will help us leverage DER and new load to enable more
flexible demand management, improve reliability and, we anticipate, enable better
decision-making about large-scale investments as well. In our last IDP we discussed
these enabling technologies, which include: Advanced Distribution Management
Systems (ADMS), which will allow us to better integrate DER onto our grid and
maintain reliability; the Field Area Network (FAN), which enables two-way
communication from field devices and Company back office operations; Advanced
Metering Infrastructure (AMI), which enables time of use rates by metering
consumption at smaller time intervals; and automated, remote reliability-sensing and
enhancing technologies such as and Fault Location, Isolation, and Service Restoration
(FLISR). Many of these investments will help us develop capabilities to use load
more flexibly. We also hope to enhance the customer experience by creating new
programs and offerings that fit their needs and preferences based on the capabilities
these investments provide.
Our Preferred Plan is substantially dependent on anticipated customer load, which
incorporates the best estimates we have about customer adoption of DER, as well as
robust statistical forecasting methods. We have also modeled sensitivities to address
some of this uncertainty. But we still often do not have visibility into which
See the Minnesota Pollution Control Agency’s statewide GHG inventory data, available at
https://www.pca.state.mn.us/air/greenhouse-gas-emissions-data.
16

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 36 of 139

technologies, and at what pace, customers will adopt and thus, how that changing load
will affect our grid needs in the future.
D.

Community and Employee Considerations

As we move forward with our carbon reduction goals, we are cognizant that phasing
out some of our legacy generation assets has a significant impact not only on our
energy mix, but on the economies of communities where those plants are located and
the employees who work in those plants. This is particularly true of our coal facilities,
where the plants are prominent places of employment and contributors to the
property tax base in the community. This is why we make efforts to spur economic
development in locations where our current units will eventually be phased out.
For example, since our most recent Resource Plan, where we proposed to retire the
Sherco 1 and 2 coal units in Becker, we have worked extensively with local units of
government, community stakeholders, and the State to draw new development to
support the local economy. This includes a planned CC generating unit at the Sherco
site, the relocated Northern Metal Recycling facility, and, prospectively a new Google
data center with energy matched by a wind facility. Some of that activity (e.g. the
Google data center) is also anticipated to spur new renewable energy development on
our system.
Related, we are participating in a study overseen by Center for Energy and
Environment (CEE) that will examine the impacts of the large baseload generation
plants in Minnesota on the host communities. The other participants in the study
include the Coalition of Utility Cities, Minnesota Power, and the Prairie Island Indian
Community. The study will consist of a quantitative and qualitative component. The
quantitative component of the study is similar to the study we conducted for Sherco 1
and 2 in our last Resource Plan. For the qualitative component, CEE will engage with
host community residents and business to gauge awareness, opinions and concerns
around potential power plant closures. Efforts on both components are underway
and we will supplement this Resource Plan filing when each component is completed.
As this docket progresses, we expect to be able to incorporate further findings and
hold additional discussions incorporating the finalized report outcomes. Further
discussion of the scope and status of this study is included as Appendix O2.
In addition to community impacts, we are also aware that these plant closures impact
our employees and their families. With this in mind, and consistent with our past
practices, we will work with these impacted employees to transition them to other
Xcel Energy plants or areas of the company. In the past, when plants have been
closed or converted (and impacted headcount) we have provided résumé writing
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 37 of 139

services, support for interview practice, job training, and job shadowing opportunities.
Through natural attrition and job relocations, we have been able to successfully “rehome” nearly all impacted employees from plant closures and conversions to-date.
Moving forward we will work with local unions and set a course to negotiate
multiskilling for our impacted sites. This skill set will position our employees for other
job opportunities within Xcel Energy. As we get closer to closure dates, temporary
workforce will be utilized to back fill benefit employees who have relocated to other
positions within the company. This strategy lessens the burden and stress for benefit
employees to find positions, as plants near closure dates. In addition, plant
management, Work Force Relations and Human Resources will work together with
other business organizations within the company to help coordinate interviews for
affected employees.
And, as we continue toward achievement of our aggressive carbon goals, we will
continue to make significant investments in clean energy in the states we serve. As we
do so, we will look for opportunities to create fair access to clean energy programs,
jobs and economic development opportunities. Going forward, we continue to be
dedicated to working with employees, communities, and stakeholders to manage
community impacts throughout our clean energy transition.
E.

Customer Preferences

Our Upper Midwest system continues to serve a diverse mix of customers with varied
interests and preferences. While most customers continue to prioritize affordability,
we have seen increasing interest in sustainability, carbon reduction, and clean energy
objectives. Again, these are important considerations to keep in mind while planning
our resource mix for the future.
1.

Municipal

Cities and municipalities are increasingly setting and developing strategies around
sustainability and climate goals. In fact, there are 11 cities in our Upper Midwest
jurisdiction that have set carbon reduction or renewable energy goals. Minneapolis is
the most prominent example, as evidenced by the Clean Energy Partnership that had
just started when we filed our last Resource Plan. Since then, the partnership has
flourished and advanced, helping to achieve progress toward the city’s sustainability
goals. Other municipalities and communities are also developing goals and action
plans around renewable energy and climate goals. We work with many of these
communities through our Partners in Energy program to support achievement of
these goals.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

2.

Resource Plan – Chapter 2
Page 38 of 139

Commercial and Industrial Customers

Our commercial and industrial customers place a high priority on keeping costs low to
remain competitive in their own markets. This is particularly true of large industrial
customers, where energy costs can make up a substantial portion of their operating
expenses. However, corporate efforts to achieve sustainability goals are also
increasing, both in the US broadly and within our system. And as the cost of
renewable energy declines, affordability and sustainability goals increasingly go hand
in hand. Within our system, several of our corporate customers are co-members of
the Minnesota Sustainable Growth Coalition, which is a business-led public-private
partnership working to advance clean energy and other sustainability and circular
economy objectives. We hear from these and other corporate customers across our
Upper Midwest system that sustainability and clean energy are important to them, and
they want us to offer products that meet these needs, and Renewable*Connect is one
such product.
In 2015 we worked with customers to develop Renewable*Connect. The program
achieved full subscription in its first year, and in January 2019 we filed for an
expansion of this program, and included an option for high load factor customers (i.e.
those that operate continuously during the day) to be served primarily with
competitively priced wind and a smaller portion of solar. Significantly, this program
advances the sustainability goals of the participating companies without creating
additional costs that must be carried by other, non-participating customers. In 2019,
we also developed a new program called Certified Renewable Percentage (CRP). The
CRP is a new Renewable Energy Certificate (REC)-based accounting methodology
that clarifies the percentage of our system energy delivered to customers that is
renewable. The CRP is not a subscription service or program customers need to
enroll in. Instead, the Company will retire sufficient RECs on behalf of all our retail
customers such that the total RECs retired annually reflects the portion of delivered
energy that is renewable. This will allow all retail customers to claim the percentage of
renewable energy on the system as the starting point towards their sustainability goals.
Our willingness to work with customers to balance clean energy objectives and
affordability needs, while facilitating economic development opportunities, has also
attracted new customers to the service area. The new Google Data Center slated for
development in Becker is a clear example. Google has an objective to match 100
percent of its energy consumption needs with renewable energy purchased from
incremental projects, and has future plans to go even further by sourcing carbon-free
energy for its operations on a 24/7 basis. We were able to work together with Google
to develop a proposal to help it achieve its renewable energy goals.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

3.

Resource Plan – Chapter 2
Page 39 of 139

Residential Customers

Residential customers likewise tell us that they value choices and clean, affordable,
reliable energy. In response, we have developed programs that offer more convenient
payment options, rebates for EE upgrades, and the chance to reduce the
environmental impact of their consumption by choosing renewable energy.
Customers are taking advantage of these programs in large numbers – and they have
expressed strong satisfaction with their ability to select programs that best meet their
individual energy needs.
F.

Supply and Technology Trends

Trends around the supply of generation and energy storage equipment we need to
fulfill our resource plan have a significant impact on the mix and timing of our
resource proposals.
In the years since our last Resource Plan, wind and solar technology costs have
continued to improve overall; solar in particular has experienced significant cost
declines, with installed costs falling over 35 percent on average since 2015.17
Consistent with past years, we generally expect wind and solar capital costs will
continue to decline, although at perhaps a slower pace as these technologies advance
on their respective maturity curves. We also expect technology advancements
improve capacity factors. These two factors continue to improve the cost
competitiveness of wind and solar resources in real terms, changes to incentive
policies notwithstanding, relative to the other resource options we may consider. For
example, the National Renewable Energy Laboratory projects that large-scale solar
prices could decline 17 percent in real terms over the next 10 years.18 For modeling
purposes, however, our generic representations of these resources are static in terms
of capacity factor and accredited capacity, and we have represented the combined
future cost and performance trends through the levelized energy cost forecast for
these technologies.
We also continue to examine the role energy storage can play in meeting our system
needs. The Company has been developing our experience around the type of services
energy storage can provide to our system, through the operation of an existing
17 Bolinger, Mark and Seel, Joachim. Utility Scale Solar 2018 Edition. Lawrence Berkeley National Laboratory September
2018. Available at: https://emp.lbl.gov/utility-scale-solar
18 This projection reflects data provided by the NREL’s Annual Technology Baseline report, which we use in our modeling.
This trend reflects changes in projected solar levelized costs in real terms (2016$) and are not adjusted for the potential
impact of tax credits.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 40 of 139

pumped hydro facility in Colorado and several pilot battery energy storage
installations across Xcel Energy’s service areas. Technologically, we expect grid-scale
energy storage will support our clean energy goals in the future, by helping us
maintain grid stability and supporting peak management while integrating the higher
quantities of intermittent renewable generation we envision on our system. We are
committed to pursuing this technology although challenges remain, in particular for
battery energy storage, to managing seasonal renewable energy variability and longer
duration demand-shifting needs.
From a cost perspective, battery energy storage has experienced significant
improvements over the last few years and we would expect battery energy storage
costs to further decline going forward. There are also other battery chemistries and
different types of storage that may emerge as technology research and development
continues. While we are confident utility-scale battery storage will be a part of our
long-term resource mix, we are also evaluating the potential for near-term battery
storage around our service territories to fulfill distribution system or other needs.
Finally, as we have noted, achieving our 100 percent carbon-free electricity by 2050
goal will require further development of technologies that have not yet been identified
and commercialized. While not included in our resource planning optimization, we
continue to monitor industry activity around other emerging technologies that may
contribute to achievement of our goals. In addition to potential new battery
chemistries mentioned above, potential emerging clean energy technologies include
advanced nuclear reactors, carbon capture and storage applications, hybridized gashydrogen generators, other types of energy storage technologies beyond batteries, and
others. As new technologies achieve commercialization, we will remain technology
agnostic as we consider including them in our future resource planning analyses.
G.

Jurisdictional Updates

Our integrated Upper Midwest system provides service on a multi-jurisdictional basis
to 1.8 million customers across five states. Through this integration, we have
historically leveraged economies of scale to support needed investments. Each
resource on the Upper Midwest system – whether generation or transmission – was
developed in consideration of the whole system, to take advantage of the economies
of scale available through integrated system planning. Below we provide a brief
overview of key prevailing and emerging energy policy in Minnesota, North Dakota,
South Dakota, Wisconsin, and Michigan, where these objectives may affect our future
planning for the Upper Midwest system as a whole.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

1.

Resource Plan – Chapter 2
Page 41 of 139

Minnesota

In 2007, Minnesota passed the Next Generation Energy Act, which set successive,
economy-wide greenhouse gas reduction goals relative to 2005 levels; 15 percent
below reduction by 2015, 30 percent below by 2025, and 80 percent below by 2050.
Since that time, the electric sector has outpaced emissions reductions in all other
sectors, achieving a 29 percent reduction below 2005 levels by 2016 (the most recent
year reported). However, the state has missed its economy-wide goal, achieving
approximately 12 percent reduction over the same time period.19 This data drives our
view that the electric sector can, and likely must, facilitate reductions in other sectors,
such as transportation and building energy use, if Minnesota is to meet its economywide carbon reduction goals.
Specific to the electric sector, the Minnesota Renewable Portfolio Standard, passed in
2007, remains our prevailing clean energy requirement. The 2019 session included
passage, however, of energy storage provisions that (1) enable utility-owned energy
storage pilot projects, and clarifying cost recovery for such pilots, (2) requiring
resource plans to include an Energy Storage Systems Assessment, including how
storage may contribute to generation and capacity needs and ancillary services, and (3)
requiring a cost-benefit study of energy storage systems by the Department of
Commerce.20 We provide our Energy Storage Systems Assessment in Appendix F7.
2.

North Dakota

Since our last Resource Plan, we submitted a Resource Treatment Framework (RTF)
simultaneously to the Minnesota Public Utilities Commission (MPUC) and the North
Dakota Public Service Commission (NDPSC). The RTF filing was submitted to the
NDPSC in compliance with a Commission Order adopting the Negotiated Agreement
in the Company’s last rate case,21 with the purpose of establishing a framework to
address the costs and benefits of the Company’s Upper Midwest system resources in a
way that was both fair to all customers and aligned with North Dakota’s policy
objectives.
In the RTF, we developed a proposal that evaluated four potential implementation
structures: (1) legal separation of the North Dakota portion of the system; (2) pseudo
See Minnesota Pollution Control Agency “Greenhouse Gas Emissions Data.” Available at:
https://www.pca.state.mn.us/air/greenhouse-gas-emissions-data
20 “19-5227 – Omnibus Jobs, Economic Development, Comerce and Energy.” Available at:
19

https://www.senate.mn/committees/2019-2020/3098_Committee_on_Jobs_and_Economic_Growth_Finance_and_Policy/195227.pdf

See North Dakota Case No. PU-12-813. In Minnesota, the Company submitted the RTF consistent with our
commitments made in MPUC Docket No. E002/M-16-223

21

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 42 of 139

separation, applying generator-specific cost/benefit allocations; (3) proxy-pricing,
which would replace the cost of a disputed generation resource with a proxy price
deemed acceptable; or (4) gaining regulatory alignment in the selection of resources,
so that all states could fully participate in the integrated NSP system. Our intent was
to seek a framework solution that can not only address current resource disputes, but
also those that may arise in the future.
Our initial proposal would have legally separated the North Dakota portion of our
service area into a separate operating company, with pseudo-separation identified as a
second best option. Under the legal separation framework, we felt confident we
could retain the benefits of economies of scale in our resource planning, yet be able to
customize resource portfolios for certain states, to the extent necessary. Pseudoseparation could also prove feasible, if each state agreed to the method by which we
would propose to conduct these cost allocations.
In response to this proposal, the North Dakota Staff recommended against legal
separation, instead preferring a solution that used a form of proxy pricing to
determine appropriate resource cost allocations for resources in the Upper Midwest
system. Further, Staff was interested in pursuing the Company’s proposal to institute
a formal resource planning process in the state, in order to provide the North Dakota
Commission with more information about planned resource additions earlier in the
process, and to provide the Company more certainty regarding the state’s policy
objectives and deemed customer needs.
While we do not see proxy pricing as a viable forward-looking solution for reconciling
resource treatment, we continue to believe that instituting a formal resource planning
process is beneficial and will provide more clarity to the Commission and the
Company on what resources may work as system resources. In response to the
Commission’s request for a more detailed proposal, we submitted a framework
outlining essential pieces of a North Dakota Resource Plan, including a default
presumption that the system would continue to be planned in an integrated fashion,
and a proposed timeline for filing that is consistent with the Minnesota process. We
discussed the proposal at an informal hearing with North Dakota Commissioners in
March of this year, where the Commission confirmed that it is interested in a more
formalized resource planning process.
Currently, we do not recover the full PPA costs of the Aurora, Marshall, and North
Star solar PPAs, the Community Based Energy Development (C-BED) PPAs, or the
Renewable Development Fund (RDF) PPAs from our North Dakota customers.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

3.

Resource Plan – Chapter 2
Page 43 of 139

South Dakota

We have also faced challenges in recovering the costs of certain resources in South
Dakota. In December 2016, the South Dakota Public Utilities Commission (SDPUC)
suspended the fuel clause adjustment in order to investigate the costs of certain
disputed resources recovered through the fuel clause.22 The Company worked with
South Dakota Commission Staff to development a Settlement to address these
concerns. In September 2017, the Commission approved a Settlement Stipulation
that addressed the costs recovery for the Aurora solar resource and several biomass
resources. The Settlement also required the development of proxy prices for several
remaining disputed resources, which include the Marshall and North Star solar PPAs,
C-BED PPAs, and Renewable Development Fund PPAs. The Company submitted a
proxy price proposal to the South Dakota Commission which provided eight potential
proxy pricing options for the Commission’s consideration and recommends a
combination of market-based and index-based pricing, depending on the particular
resource.23 We are engaging in ongoing discussions with South Dakota Staff as we
work toward a resolution of these outstanding issues. While we are working to
develop proxy prices as a resolution for past resources, when disputes arise in the
future we would work to develop a cost allocation mechanism to allocate the costs
and benefits of new resources to the participating jurisdictions.
Currently, we do not recover the full PPA costs of the Aurora solar project from our
South Dakota customers and we are recovering the costs of the PPAs for the disputed
resource subject to refund based on the resolution of the proxy price proceeding.
4.

Wisconsin

In Wisconsin, the Company is subject to a Renewable Portfolio Standard (RPS) equal
to 12.89 percent of its three-year average in-state retail energy sales. In 2018,
excluding renewable energy used for voluntary renewable programs, NSPW provided
24.3 percent of its retail energy sales from RPS-eligible renewable-based energy
sources, and the Company is in compliance with its 2018 RPS requirements.24 This
requirement has not changed recently. However, the state’s newly elected governor
has identified climate objectives as one focus of his administration. In line with these
goals, the administration has proposed carbon reduction goals for the state’s electric
sector that are broadly consistent with our objectives to reduce emissions by 80
percent from 2005 levels by 2030, and 100 percent by 2050. It is not yet clear
See South Dakota Docket No. EL16-037.
See South Dakota Docket No. EL18-004.
24 See Docket 5-RF-2018 Renewable Portfolio Compliance Plan for CY 2018. Northern States Power Company, a Wisconsin
Corporation.
22
23

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 2
Page 44 of 139

whether these proposed goals will result in additional mandates for the electric sector.
5.

Michigan

In Michigan, the Company is subject to a Renewable Energy Standard (RES) equal to
15 percent of retail sales by 2021, with a goal of 35 percent renewables and 25 percent
energy waste reduction by 2025. We are ahead of schedule on both these goals,
already exceeding the 2021 RES requirements and expecting to meet the 2025 goals
by 2020.
This year will also be the first time we are required to submit a Resource Plan to the
Michigan Public Service Commission, in accordance with Public Act 341 of 2016.
Michigan’s new Resource Plan process allows us to file the same resource plan in
Michigan as we are filing with the Minnesota Commission, and on a similar
timeframe, with the understanding that the Michigan Commission may ask for
additional supplemental information to help them evaluate the plan as it relates to
Michigan.25 We plan to file this multi-state Resource Plan in Michigan on July 31.
H.

Conclusion

We believe our progressive carbon reduction goals are the right path forward for the
our customers, our employees, our Company and the State. We are confident in our
ability to achieve these goals while maintaining the reliability and affordability our
customers count on. This transition, while exciting, is not without challenges. We
will continue to navigate significant market and regulatory uncertainty, jurisdictional
differences, technology changes and more. That said, we are optimistic that these
challenges also present opportunities to engage with customers, regulators, market
operators, communities, and employees on our goals in a way that meets multiple
stakeholders’ objectives.

25

See U-15896/U-18461 at 11.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 45 of 139

CHAPTER 3
MINIMUM SYSTEM NEEDS
Our resource planning process focuses on deep carbon reductions while serving our
Upper Midwest customers reliably and affordably. In this Chapter, we describe in
more detail how we arrived at the minimum amount of resources our system will need
through the planning period. The system needs and existing resources evaluated here
formulate the baseline upon which we have developed the Reference Case, our
modeling scenarios, and ultimately our Preferred Plan.
We have made the following changes to aspects of our Minimum System Needs
approach with this Resource Plan:
 Supply-side Resource Treatment for DSM. In this Resource Plan we are treating
both EE and DR as supply-side resources, rather our previous treatment of EE
as an adjustments to future load. Supply-side resources available to the model
now include incremental EE and DR in “bundles,” or amounts of achievement
that we formed from the Minnesota Energy Efficiency Potential Study and Brattle
Demand Response Potential Study, respectively.
 Reliability Requirement. We have developed and applied a threshold requirement
for firm dispatchable resources. This Requirement is needed to ensure system
reliability and resilience until MISO evolves its capacity accreditation construct
to better recognize the variability and declining incremental electric load
carrying capability of wind and solar resources.
I.

MEETING CUSTOMER NEEDS

Forecasting customers’ needs for electricity is a key component of any resource plan,
and provides the foundation for determining the type and amount of resources that
will be needed over the 15 year planning period. The first step is forecasting the
amount of electricity our customers will need over the planning period. To this, we
add a reserve margin that is prescribed by MISO. We then subtract the resources we
already have or expect to have (with some adjustments), to determine our net surplus
or need.
We illustrate this concept and discuss each of the components below.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 46 of 139

Figure 3-1: Net Resource Need/Surplus Calculation
Customer Needs Forecast
Plus MISO Reserve Margin
Equals Total Capacity Obligation
Minus Demand Response Capability
Minus Generation Capacity (measured by UCAP)
Minus Generation Adjustments
Equals Net Resource Need/Surplus

A.

Customer Needs Forecast

Forecasting our customers’ energy needs starts with a capacity, or peak demand,
assessment, which informs the total amount of generating capacity (in megawatts, or
MW) needed to meet our customers’ needs in the highest demand hour (i.e. peakhour) in each year of the planning period. We also assess the amount of total energy
(measured in megawatt hours or MWh) we expect customers to consume in each year
of the planning period. Together, the peak demand and total energy needs inform the
type of generating resources that will best meet customer needs.
1.

Peak Demand Requirements

We use econometric analysis and historical actual coincident net peak demand data to
determine system capacity requirements for each year. We provide a detailed
discussion about our peak demand forecasting methodology in Appendix F1.
Our current forecast shows essentially flat load relative to current levels, with an
average annual growth rate of less than 0.2 percent, after accounting for EE. Figure
3-2 below shows the current forecast in relation to the forecast from our last
Resource Plan.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 47 of 139

Figure 3-2: Forecasted Peak Load, After Energy Efficiency Adjustments (MW)1
11,000
10,500
10,000

Load (MW)

9,500
9,000
8,500
8,000
7,500

2016 IRP Forecast

7,000

2019 IRP Forecast

6,500
6,000

We have changed our approach to how we present our peak forecast from our last
Resource Plan based on our treatment of EE as a supply-side resource in this Plan.
The peak forecast in our previous Resource Plan included both a demand reduction
associated with historic EE, as well as from the impact of future incremental EE as
approved in the previous Resource Plan.
For this Resource Plan, we have changed our approach to addressing EE in two ways.
First, the load forecast used in our Strategist modeling no longer has embedded
incremental EE, although it is shown in that manner in the charts above for
comparison purposes with previous Resource Plans. Instead, we treat EE as a supplyside option that the Strategist model can select in its resource optimization. To do so,
we developed EE “Bundles,” which we describe in Part III below and in more detail
in the Strategist Assumptions Appendix F2. As a result, rather than adjust our peak
forecast based on an assumed level of future EE adoption (this would have been 1.5
percent per the level approved in our last Resource Plan), we are reflecting our
commitment to higher levels of EE achievement (approximately 2.5 percent reduction)
in our Reference Case. This level of EE reflected in the Reference Case is
representative of two of the EE Bundles available for Strategist to select.

Although we modeled EE bundles as supply-side resources in this Resource Plan, we show the estimated
resulting EE as a load reduction from gross load for purposes of the chart above.
1

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

2.

Resource Plan – Chapter 3
Page 48 of 139

Energy Requirements

We forecast declining energy needs of approximately 0.4 percent over the 2020-2034
planning period, after accounting for EE included in the Reference Case. As
discussed above, the inclusion of two incremental EE Bundles reflects achievement of
approximately 2.5 percent EE, which leaves our Net Demand substantially lower than
forecast in our last Resource Plan. Figure 3-3 below compares our estimated net
energy demand adjusted by the two EE Bundles, to the energy forecast in our last
Resource Plan.
Figure 3-3: Forecasted Net Energy Requirements, After Energy Efficiency
Adjustments2 (GWh)
55,000

Energy Demand (GWh)

50,000

45,000

40,000

35,000

2016 IRP Forecast
2019 IRP Forecast

30,000

3.

Forecast Adjustments

After determining the base peak capacity and energy demand forecasts, we make
certain forecast adjustments to account for the impact of events or trends we
reasonably expect to occur in the planning period. We summarize our key
adjustments below:
DSM. In past Resource Plans, the load forecasts used by Strategist were adjusted for
Although we modeled EE bundles as supply-side resources in this Resource Plan, we show the estimated
resulting EE as a demand reduction from gross demand for purposes of the chart above.
2

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 49 of 139

the expected effects of existing DSM programs. In this Resource Plan, based on
feedback from stakeholders, incremental EE is no longer embedded in the load
forecast, rather EE is treated as potential supply-side resource in our modeling, like
DR. Both EE and DR are shown as separate line items in our Loads and Resources
table below, though in an effort to maintain consistency with Load and Resource
reporting between this and previous Resource Plans, we show EE “above the line” as
a subtraction from gross load. We further discuss the EE and DR (collectively, DSM)
in the context of our resource planning process in Appendix G1.
Distributed and Small Scale Customer Solar Generation. We have historically considered
customer adoption of distributed solar (i.e. DG solar as well as CSG) installations as a
modification to load in the resource planning process. In this Resource Plan, we have
accounted for DG solar including CSG resources as a supply-side resource with
assumed adoption levels, as shown in the Loads and Resources calculation below.
Reference Case assumptions currently take into account interconnection requests and
expectations based on policy-driven programs. However, we also conduct sensitivity
testing around potential increased levels of adoption and are working to develop new
tools that improve our understanding of how key market drivers will affect customer
distributed solar adoption going forward. We note that our methods for projecting
distributed solar installations are currently evolving. As our tools and methods
mature, we will increasingly incorporate them into both our Resource Plan and IDP
processes.
Expected Customer Changes. We also make adjustments to account for known changes
in load on our system. These typically reflect expected changes in specific large
customers’ electricity usage, either as a result of increased behind the meter energy
generation (decreasing demand) or increased production activities (increasing
demand).
Light Duty Electric Vehicle Adoption. We adjust our residential energy and peak demand
forecasts to account for increasing use of plug-in electric vehicle charging. These
forecasts are based on expectations around current stock and future adoption
(including the effect of financial incentives to facilitate adoption), and the expected
electricity consumption per vehicle.
We use standard statistical modeling techniques to reflect these and other potential
sources of variation around our expected forecasts. We discuss our forecasting
process, inputs, assumptions, adjustments and results in more detail in Appendix F1:
Load and Distributed Energy Resource Forecasting.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

II.

Resource Plan – Chapter 3
Page 50 of 139

MISO RESOURCE ADEQUACY REQUIREMENTS

MISO prescribes Resource Adequacy (RA) requirements that are intended to help
ensure adequate reliability of the bulk electric supply system. MISO’s RA process
requires load serving entities (LSE) like the Company to maintain resources that
exceed their level of demand by a specific margin (planning reserve margin or PRM)
to cover potential uncertainty in the availability of resources or level of demand.3 The
RA requirements are fundamental to the resource planning process, and inform the
level of capacity we need in our portfolio to adequately serve customers over a longterm planning process. We describe the various aspects of the calculation below, and
note that our effective reserve margin is 2.98 percent.
MISO’s RA requirements are set based on an annual planning period; the 2018/2019
planning period covers June 1, 2018 through May 31, 2019. Prior to each planning
year, MISO determines two different capacity obligations for each LSE; one for the
entire MISO footprint as a whole, and one for the Local Resource Zone (LRZ or
Zone) where the LSE resides. 4 For any particular planning year, an LSE’s PRM is the
greater of the LSE’s capacity obligation for the MISO footprint or its capacity
obligation for its LRZ.
A.

MISO Footprint Capacity Obligation

By November 1 prior to a planning period, MISO issues the finalized PRM applicable
to all LSEs within its footprint. MISO determines the PRM by performing a technical
probabilistic analysis to determine the minimum PRM needed to achieve a Loss of
Load Expectation (LOLE) of 0.1 day per year, expressed as a percentage. For
example, for the planning year covering June 1, 2018 through May 31, 2019 the
overall MISO PRM was 17.1 percent on an installed capacity (ICAP)5 basis and 8.4
percent on an unforced capacity rating (UCAP) basis.6 The study also provides
The factors affecting availability and demand include: Planned maintenance, Unplanned or forced outages
of generating facilities, Deratings in resource capabilities, Variations in weather, and Load forecasting
uncertainty.
4 Almost all of the NSP system load is located within LRZ 1, which includes almost all of Minnesota, western
Wisconsin, and the Dakotas. Approximately 7 MW of load along the Minnesota-Iowa border is located in
LRZ 3.
5 ICAP refers to units’ Installed Capacity Rating, which is a capacity accreditation measure based on annual or
historical tested generating. The ICAP is the lesser of the generator verification testing capacity or the
interconnection service capacity.
6 UCAP refers to units’ Unforced Capacity Rating, which is a function of the unit’s installed capacity and its
anticipated forced outage rate. A generator’s anticipated forced outage rate is typically based on the individual
unit’s historical performance. UCAP = ICAP x (1 – Forced Outage Rate). See “Planning Year 2018-2019
Loss of Load Expectation Study Report” at 5. Available at:
https://www.misoenergy.org/api/documents/getbymediaid/80578
3

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 51 of 139

forward-looking PRM values, through 2028. Over the planning period MISO
examined in the 2018-2019 LOLE study, the UCAP PRM remained relatively
constant between 8.3-8.4 percent.
Each LSE is required to have resources sufficient to meet the forecasted demand at
the time of MISO’s peak demand, plus its PRM margin. MISO’s tariff acknowledges
a state regulatory body’s authority to establish a PRM for LSEs within its jurisdiction,
which would override the PRM otherwise determined by MISO. None of the NSP
System states have established a PRM separate from MISO.
B.

Zonal Capacity Obligation

Additionally, MISO makes an annual determination regarding the amount of capacity
required within each of MISO’s Zones, called the Local Clearing Requirement (LCR).
The LCR is determined as a function of each Zone’s Local Reliability Requirement
(LRR) and its Capacity Import Limit (CIL). The LRR represents the necessary
resource requirement in order for a Zone to achieve an LOLE of 0.1 day per year,
without relying on resources outside of the Zone. Each Zone, having a smaller
footprint than the overall MISO footprint does not benefit from the same level of
peak load diversity as does the larger, more diverse MISO footprint. Thus, it can be
expected that a Zone’s LLR is greater than the sum of its LSEs’ MISO footprint
obligations. If a Zone within which an LSE operates has import capacity, however,
the resulting LCR is reduced. As a result, LSEs usually plan their minimum system
needs based on the MISO-wide PRM rather than the zonal requirement.
For the 2018-2019 planning year, Zone 1 was assigned an LRR of 114.8 percent,
which, in capacity terms equates to an LRR of 20.2 GW. However, when accounting
for Zone 1’s CIL of 4.4 GW, Zone 1’s LCR is 15.8 GW. This is less than the MISO
footprint PRM of 18.4 GW. Thus, Zone 1’s import capabilities allow LSEs within
Zone 1, including the Company, to plan to the MISO-wide UCAP PRM of 8.4
percent rather than the higher LRR value.
C.

Capacity Obligations Derived From Forecasted Demands

After determining the relevant PRM, each LSE can derive its MISO-wide and zonal
capacity obligation from its forecast of peak demand (peak load). While LSEs typically
forecast the peak demand for their individual system, the resource adequacy process
requires the LSE to also forecast:
 The LSE's demand at the time of the MISO footprint’s peak demand (MISO
Coincident Peak Demand, or MISO CPD), and
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 52 of 139

 The LSE's demand at the time of the LRZ’s peak demand (Zonal Coincident
Peak Demand, or Zonal CPD).
Again, because each LRZ footprint is smaller than the MISO footprint, the LRZ’s
load diversity is lower than the load diversity of the MISO system, and an LSE’s
Zonal CPD is equal to or greater than its MISO CPD.
The NSP System CPD factor measures how closely our system peak matches the
MISO system peak. A coincidence factor of 95 percent indicates that we expect to
experience load levels that are approximately 95 percent of our peak load during times
when the total MISO system load is peaking. In other words, the timing of our peak
and the MISO peak does not match exactly, so we are able to reduce the amount of
reserves we carry as a result. After accounting for the coincidence factor, our
effective reserve margin drops from 8.4 percent to 2.98 percent. We illustrate this
calculation in Figure 3-4 below.
Figure 3-4: MISO Planning Reserve Margin Calculation – NSP System
Planning Year June 1, 2018 to May 31, 2019
95 

 

 
 2.98 

  1
 

8.4 
 

 

 1 
 
 

 

Putting these pieces together, we used our effective reserve margin, in combination
with our annual load forecasts over the planning period, to determine our overall
capacity obligation for the same period. For 2020, this calculation results in the
following approximate obligation:
Table 3-1: Capacity Obligation Calculation – 2020 Example
Total Capacity Obligation Component
Forecasted load
NSP Effective Reserve Margin
NSP Obligation

Value
9.1 GW
x (1+ 2.98%)
= 9.4 GW

Our estimated obligation for all planning period years can be found in the Load and
Resources table in Section VI. below.
D.

Capacity Accreditation of Resources

After these obligation levels have been determined, we consider the type of resources
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 53 of 139

suitable to meet that requirement. MISO’s tariff and business practices set forth
procedures to enable various types of resources to be used to achieve our RA
requirements. While there are different requirements among the various types of
resources, common characteristics require resources participate in the annual
registration process, requiring annual testing and reporting of capability or reporting
of historical output. Each resource must have firm delivery to load, and resources
must be available throughout the entire planning period.
Resources used to achieve MISO’s RA requirements are referred to as “Planning
Resources.” Planning Resources include the following sub-types:
 Capacity Resources: Physical Generation Resources (i.e. physical assets and
purchase agreements), External Resources if located outside of MISO’s
footprint, and DR Resources participating in MISO’s energy and operating
reserves market, available during emergencies.
 Load Modifying Resources: Behind-the-Meter Generation and DR available during
emergencies, which reduces the demand for energy supplies coming from the
LSE.
 Energy Efficiency Resources: Installed measures on retail customer facilities
designed and tested to achieve a permanent reduction in electric energy usage
while maintaining a comparable quality of service.
MISO’s resource accreditation represents a measure of a resource’s reliable
contribution to the system’s resource adequacy needs. A generator’s operation,
maintenance, and utilization directly impact the portion of nameplate capacity rating
recognized as an accredited resource. Therefore, instead of using installed or
nameplate capacity (i.e. ICAP), MISO calculates the unforced capacity value (i.e.
UCAP) for each resource to determine its expected contribution to RA. These are
calculated differently depending on the resource’s dispatchability or variability:
 Dispatchable generation resources, DR and EE – MISO assigns a UCAP value for
dispatchable generation resources by discounting their installed capacity by an
anticipated forced outage rate. Resources where availability depends on other
factors are measured differently; for example, MISO has a process to determine
the UCAP for DR resources using a documented process of assessing the
resource’s observed responsiveness and load reduction effectiveness.
 Variable resources – MISO assigns variable resources, such as grid-scale solar and
wind, a UCAP value that is a function of the individual unit’s historical
performance during the peak hours of the planning period. Currently, these
units are measured on historical performance during the operating hours of
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 54 of 139

1500 to 1700 in the months of June-August over the three most recent
summers. Each site must have one complete historic period of data prior to
unit accreditation. If sufficient operating history is not available, MISO assigns
a proxy value.
Our modeling selects resources based on their UCAP values, to ensure we maintain
adequate capacity on our system over the planning period. Additionally, as discussed
below in Section IV, we included a further Reliability Requirement in our planning
process to address MISO’s evolving processes.
III.

DEMAND SIDE MANAGEMENT

DSM programs offer our customers opportunities to lower their energy use and
manage their peak demand, in particular through our Conservation Improvement
Programs. As noted previously, these programs include both EE and DR. We base
our forecasts and potential incremental additions on historic achievements through
our programs, as well as external studies about expected and potentially achievable
adoption rates.
As previously discussed, we adjusted the customer capacity and energy forecasts in
this Resource Plan to distinguish incremental EE from the load forecast. We
modeled incremental DR and EE achievements as “Bundles” to be evaluated
alongside other resource options. Each Bundle represents a combination of program
achievements expected to lead to a certain amount of avoided load or energy per year,
at an estimated blended cost.
For EE, these Bundles include measures that work to reduce a customer’s overall
energy usage throughout the year. The DR Bundles, on the other hand, reflect a
customer’s commitment to discrete reductions in demand (e.g. on a day when peak
load is expected to be high otherwise). These actions are expected to reduce the
anticipated annual system peak demand, as well as smooth demand on specific days
when weather or other conditions lead to high demand at a certain point in time. In
the Order approving our last Resource Plan,7 the Commission directed that the
Company “shall acquire no less than 400 MW of additional DR by 2023.” In this
Resource Plan, we included one DR Bundle in our modeling for both the Reference
Case and Preferred Plan.
We discuss the studies that informed our expected EE and DR levels, our analysis,
See E002/RP-15-21 Order Approving Plan with Modifications and Establishing Requirements for Future
Resource Plan Filings (January 11, 2017), Order Point 10.
7

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 55 of 139

and the changing DSM landscape in more detail in Appendix G1.
IV.

RELIABILITY REQUIREMENT

A new planning element we included in this Resource Plan is a Reliability
Requirement which, in short, will ensure that we can serve customers with reliable
energy every hour of every day.
The need for the Reliability Requirement stems from the inherently variable nature of
renewable resources and availability of any time limited resource. Further, as the
penetration of these resources increases, their value to meet peak customer needs
decreases. Although MISO is beginning to recognize these challenges, its current
planning constructs do not yet incorporate any measures to address them. Until
MISO determines how best to address these gaps, we believe it is incumbent on us as
the utility to take steps to ensure that our system is resilient and that our customers
will be reliably served. Fundamentally, we have a responsibility to ensure we have
access to a sufficient level of firm dispatchable resources in all grid conditions that can
flexibly adapt to variable renewable resource performance to meet our customers’
needs.
To develop the Requirement, we analyzed industry insights and data from case
studies, including the 2019 polar vortex and normal winter and summer days. From
this information, we developed a method determine a threshold level of firm
dispatchable resources needed to serve customer loads, that reflects a reasonable level
of reliance on MISO resources and DR to meet a portion of the need.
Figure 3-5 below demonstrates the calculation of the Reliability Requirement we
applied in our modeling for this Plan.
Figure 3-5: NSP System Reliability Requirement Calculation –
2020 Example
Peak Demand Proxy – 6.4 GW
Minus Firm DR (Winter) Proxy – (0.2) GW
Minus Firm Market Supply Proxy – (0.5) GW
Reliability Requirement - 5.7 GW
(Firm dispatchable resources)

To implement this Requirement, we applied it as a threshold in our Strategist
modeling to ensure that our firm dispatchable resources do not fall below this level –
even while our Preferred Plan achieves an 80 percent carbon reduction by 2030. In
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 56 of 139

the 2020-2034 planning period, the Requirement ranges from approximately 5.6 GW
early in the planning period, to about 6.0 GW by the end of the planning period. We
clarify here, however, that while this concept is essential to include until MISO
evolves its capacity accreditation constructs, the Requirement as applied in our
modeling has little effect for this Resource Plan. The model does not select any firm
dispatchable additions as a direct result of the Reliability Requirement until near the
end of the planning period, in 2031. This long runway leaves ample time for MISO
and its stakeholders to address this aspect of its planning and provide additional
direction.
V.

EXISTING RESOURCES

Our current generating resources8 comprise a diverse portfolio including nuclear, coal,
wind, biomass, solar, hydro, natural gas and oil-fueled facilities. Physical generating
assets owned by the Company have a net maximum capacity of over 9,500 MW,
including 850 MW of wind.9 In addition to these assets, we maintain PPAs
representing a net maximum capacity of over 3,700 MW.10 Together, these provide
over 13,200 MW of generation resources, of which over 4,300 MW11 is supplied by
renewables. A total of over 6,000 MW12 is supplied by carbon-free resources.
A.

Renewable Resources

In total, we currently have over 4,300 MW of renewable capacity serving the NSP
System, including:13
 Over 2,600 MW of wind
 840 MW of solar, including community solar programs and grid-scale solar14
 680 MW of hydroelectric power15
 160 MW of biomass and landfill gas
As of July 2019; excludes some resources included in modeling that are expected to be online by the end of
2019.
9 Maximum capacity represented here reflects capacities included in Strategist modeling. It approximates Net
Maximum Capacity, which is defined as the units Gross Maximum Capacity, less any capacity that is used for
that unit’s station service or auxiliary load.
10 This total excludes 425 MW of renewable diversity capacity credit from contracts with Manitoba Hydro.
11 Id.
12 Id.
13 Note: these values are approximate.
14 Per Docket No. E002/RP-15-21 Order Point 4a (January 11, 2017), our solar acquisitions will exceed the
650 MW through CSG resources or other cost-effective acquisitions. The CSG program is on track to
exceed the ordered 650 MW by year ending 2019, per the most recent forecast CSG Monthly Report (filed
June 14, 2019) and included in this filing as Appendix N8.
15 Excluding capacity associated with diversity agreement contracts with Manitoba Hydro.
8

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 57 of 139

 

B.

Nuclear

Our Monticello and Prairie Island nuclear plants provide nearly 1,740 MW of clean
energy and capacity to our customers and play an important role in achieving our goal
of an 80 percent reduction in system carbon emissions by 2030, while maintaining
reliability and affordability. The monthly capacity factors of our nuclear facilities are
historically 90 percent or higher. Together, our nuclear plants currently provide nearly
30 percent of our energy mix. In terms of production costs (fuel plus O&M), both
plants have achieved reductions of more than 20 percent since 2015, with average
costs now below $30/MWh.
C.

Coal

Our coal fleet includes our Sherco Units 1, 2, and 3 in Becker, Minnesota and the
Allen S. King plant in Oak Park Heights, Minnesota. This coal fleet provides almost
2,400 MW of baseload and cycling generating capacity, and supports system reliability.
In our last Resource Plan, the Commission approved our proposal to retire Sherco
Units 1 and 2 in 2026 and 2023, respectively. These retirements are reflected in our
Reference Case discussed below. Our Preferred Plan further proposes to retire the
King plant in 2028 and Sherco 3 in 2030, after which coal would no longer be part of
our energy mix.
D.

Natural Gas (and Oil-Fired) Fleet

Our natural gas fleet consists of both intermediate and peaking generation. We have
five owned or contracted intermediate-type generating assets that provide just over
2,400 MW of capacity. We have peaking-type resources located at seven sites,
providing another 2,350 MW of capacity. Combined, these facilities provide valuable
load following capabilities for our system, cycling as necessary to provide important
flexibility to our generation operations and support to our growing renewable
resources. Our Reference Case also includes pending and proposed capacity resource
additions including MEC as proposed in Docket No. IP6949,E002/PA-18-702, and
the Sherco CC that the Commission acknowledged in its Order in our last Resource
Plan.16 These pending resources appear in separate line items in our net resource
calculation below.  
The Commission approved our proposed schedule to retire Sherco Units 1 and 2, and found that more
likely than not there will be a need for approximately 750 MW of intermediate capacity coinciding with the
retirement of Sherco Unit 1 in 2026. See In the Matter of Xcel Energy’s 2016-2030 Integrated Resource Plan,
ORDER APPROVING PLAN WITH MODIFICATIONS AND ESTABLISHING REQUIREMENTS FOR FUTURE
RESOURCE PLAN FILINGS, Ordering Point Nos. 7 and 8, Docket No. E002/RP-15-21 (January 11, 2017).
16

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

VI.

Resource Plan – Chapter 3
Page 58 of 139

NET RESOURCE SURPLUS/DEFICIT

As described above, our forecast of customers’ peak demand and MISO Resource
Adequacy requirements are used to determine our overall total generating capacity
obligation. From this, we deduct our expected load management achievements and
UCAP generating capacity of the various resources we have included in our Reference
Case to determine our net generation capacity surplus or deficit . We anticipate a net
surplus through 2026 and a deficit thereafter.
20F

Table 3-2: Reference Case Load and Resources,17 2020-2034 Planning Period
2020

2021

2022

2023

2024

Forecasted gross load

10,499

10,559

10,621

10,684

10,755

Forecasted EE18
(reduction to load)*
Forecasted net load
MISO System
Coincident

(1,386)

(1,472)

(1,517)

(1,609)

9,112
95%

9,087
95%

9,103
95%

Coincident Load
MISO PRM
NSP Obligation

8,657
8.40%
9,384

8,633
8.40%
9,358

940

2025

2026

System needs

2027

2028

2029

2030

2031

2032

2033

2034

10,820

10,886

10,954

11,140

11,232

11,320

11,418

11,518

11,619

11,717

(1,707)

(1,822)

(1,921)

(1,992)

(2,125)

(2,215)

(2,278)

(2,366)

(2,352)

(2,324)

(2,415)

9,075
95%

9,048
95%

8,998
95%

8,965
95%

8,963
95%

9,014
95%

9,016
95%

9,042
95%

9,052
95%

9,166
95%

9,295
95%

9,301
95%

8,648
8.40%
9,374

8,621
8.40%
9,345

8,595
8.40%
9,317

8,548
8.40%
9,266

8,517
8.40%
9,232

8,514
8.40%
9,230

8,564
8.40%
9,283

8,565
8.40%
9,285

8,590
8.40%
9,312

8,599
8.40%
9,321

8,708
8.40%
9,439

8,831
8.40%
9,572

8,836
8.40%
9,579

955

970

989

1,007

1,023

1,038

1,053

1,066

1,054

1,043

1,032

1,021

1,010

1,000

270

290

312

322

339

380

392

406

421

438

456

476

497

527

550

2,390
1,603
3,295
627
0
110
877
596
182
335

2,390
1,603
3,295
627
0
110
997
650
182
339

2,390
1,603
3,295
627
0
110
989
696
181
344

2,390
1,603
3,295
627
0
86
989
670
180
348

1,699
1,603
3,141
627
0
86
989
659
179
352

1,699
1,603
2,829
627
0
63
162
642
178
356

1,699
1,603
2,624
627
0
63
162
637
177
360

1,017
1,603
2,136
627
727
63
162
622
176
365

1,017
1,603
2,018
627
727
22
162
616
175
369

1,017
1,603
2,018
627
727
22
162
594
174
373

1,017
1,603
2,018
627
727
22
162
593
174
377

1,017
992
2,018
627
727
22
162
578
173
381

1,017
992
1,765
627
727
22
156
575
172
385

1,017
992
1,765
627
727
22
152
511
171
389

1,017
484
1,765
627
727
22
152
492
170
393

42
11,267
1,884

48
11,486
2,128

55
11,571
2,196

60
11,559
2,213

66
10,746
1,429

72
9,634
368

78
9,460
228

83
9,040
(190)

89
8,913
(370)

95
8,905
(380)

100
8,920
(392)

105
8,311
(1,010)

111
8,066
(1,373)

116
8,026
(1,546)

121
7,521
(2,058)

Reference Case resources (UCAP)
Load Management
(existing)
Load Management*
(potential study)
Coal
Nuclear
Natural Gas/Oil
MEC*
Sherco CC*
Biomass/RDF
Hydro
Wind
Grid-scale solar
Solar*Rewards
Community Solar
Distributed Solar
Existing Resources
Net Resource
(Need)/Surplus

While the Order also addressed next steps for the replacement generation at Sherco, legislation was passed as
part of the 2017 Legislative Session that in summary, allows the Company to proceed with the construction
of the replacement unit at Sherco in accordance with the parameters specified in the legislation, and without a
Certificate of Need. See Laws of Minnesota 2017, chapter 5 – H.F. No. 113, section 1.
17 In addition to existing and approved resources, those indicated with a * include pending or proposed
resources that we have included across all Scenarios, including the Reference Case.
18 Includes EE savings from historically installed measures, as well as future EE from bundles modeled in this
Resource Plan, achieving 2-3% savings levels. Also includes minimal EV and coincidence adjustments.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 59 of 139

VII. MEETING RENEWABLE ENERGY REQUIREMENTS AND
GOALS
A.

Minimum Compliance Requirements

Each of the states in the NSP System has a different public policy with respect to
renewable energy requirements or objectives. Figure 3-6 below illustrates each state’s
renewable energy standard (RES).
Figure 3-6: Renewable Energy Requirements and Objectives – NSP System

Three of our states have renewable standards expressed as a percentage of electric
retail sales from qualifying resources by a certain date. Minnesota’s RES is the
highest, requiring that 30 percent of the Company’s energy come from renewables,
with at least 24 percent of the electricity we provide to retail customers coming from
wind energy by 2020 19 Legislation passed in the 2013 session also established a Solar
Energy Standard (SES) for Minnesota that requires that investor-owned utilities in the
state generate 1.5 percent of 2020 retail sales, net of customer exclusions, from solar
energy resources. Of that 1.5 percent, 10 percent must come from systems with
2F

This requirement is included in the total 30 percent RES, and we are authorized to count a limited amount
of solar energy towards an overall 25 percent wind and solar requirement (amounting to 1% of total sales).
The SES is assessed separately. Large hydro does not count as a renewable energy source for purpose of the
Minnesota RES. Minn. Stat. § 216B.1691.

19

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 60 of 139

capacity less than 40 kW.20 The legislation also established a goal of 10 percent of
energy sales from solar by 2030.
North Dakota and South Dakota each have a voluntary objective that includes
renewable or recycled energy. 21 Further, our North Dakota regulators have indicated
that compliance with the North Dakota Renewable Energy Objective should be
accomplished with competitively-priced energy.
21F

To-date we have implemented a strategy to have the entire NSP System comply with,
at the very least, the highest of renewable energy requirements across our
jurisdictions; in this case, the Minnesota RES. This strategy also places us in
compliance with the specific requirements in each of our other jurisdictions. As a
result, we have been planning for renewable energy additions, and allocating their
benefits, to all of our jurisdictions (with certain exceptions as discussed in the
Planning Landscape). As state energy policies continue to evolve, however, we will
continue to examine whether this requires a strategy change going forward, and
engage our Commissions as needed on that topic.
B.

RES Compliance

Given existing and previously approved resources, we project continued compliance
with the renewable energy goals and standards in each of our NSP states. The
Company currently maintains a set of banked Renewable Energy Credits (RECs) for
future compliance. 22 In the past, we have leveraged our REC bank to manage the
size, type, and timing of renewable energy additions on our system, to ensure that we
identify and acquire the renewable generation resources that provide our customers
with the greatest value at the lowest cost. Given our recent focus on adding
renewable capacity, however, we now generate RECs in excess of our baseline
obligations each year. We currently generate sufficient RECs annually from eligible
renewable resources to account for over 40 percent of the energy we provide our NSP
System customers, which outpaces our annual requirement
3F

24F

The original legislation set a threshold of 20 kW, but was increased to 40 kW in 2018, per HF3232. See
“Minnesota Renewable energy Standard: Utility Compliance.” Minnesota Department of Commerce (January
2019) at 7. Available at: https://www.leg.state.mn.us/docs/2019/mandated/190330.pdf
21 As defined in North Dakota Century Code, 49-02-25, recycled energy means “systems producing electricity
from currently unused waste heat resulting from combustion or other processes into electricity and which do
not use an additional combustion process. The term does not include any system whose primary purpose is
the generation of electricity unless the generation system consumes wellhead gas that would otherwise be
flared, vented, or wasted.” South Dakota Codified Law 49-34A-94 contains a similar definition.
22 A REC is an accounting device designed to reflect the renewable energy attributes of a particular MWh of
renewable energy generation. RECs are the currency for compliance with state renewable targets.
20

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 61 of 139

Figure 3-7 below illustrates annually generated RECs across the NSP System in the
Reference Case scenario. In this scenario, we will have sufficient RECs to comply
with the current renewable energy goals and standards of all of our NSP jurisdictions
through 2034, even without securing additional renewable resources.
Figure 3-7: REC Production and Retirement Obligations for NSP System –
Existing Resources Only

RECs (Million MWh)

25
20
15
10
5
-

2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034
REC Retirement Obligation

C.

REC Production

SES Compliance

As previously mentioned, Minnesota law requires us to provide our customers with
solar-generated energy equal to at least 1.5 percent of our annual customer demand by
2020, and a goal of 10 percent by 2030. We have developed a portfolio of programs
to provide solar options to residential and commercial customers, and have also
grown our utility-scale solar profile. As a result, we expect to meet the SES
requirements through the planning period, per our Reference Case. We also expect
the solar capacity additions included in our Reference Case to provide sufficient
energy to meet the 10 percent goal by 2030. Figure 3-8 below demonstrates our
annual estimated SREC production relative to Minnesota requirements and goals, for
the Reference Case scenario.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 62 of 139

Figure 3-8: NSP System SREC Production and Minnesota Annual
Requirements
9

SRECs (Million MWh)

8
7
6
5
4
3
2
1
-

2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034
Production – Existing Resources
Production – Reference Case Additions

MN SES Obligation (1.5% of sales)
MN SES Target (10% of sales by
2030)

We discuss our renewable energy standard compliance further in Appendix N4:
Renewable Energy Compliance Positions.
VIII. ENERGY POLICY AND COMPANY GOALS
As demonstrated, we believe that we are well positioned to meet minimum system
needs. At least through 2024, we expect that we will be able to meet those needs with
existing and already-approved resources. However, in 2018, we committed to an
ambitious carbon reduction vision, to achieve 80 percent below 2005 carbon
emissions levels by 2030 and 100 percent carbon-free energy by 2050. We are
committed to achieving this goal, and as such, have modeled our Reference Case,
Preferred Plan, and all other scenarios using our 80 percent carbon reduction target as
a guidepost.
IX.

REFERENCE CASE

We incorporate all of the aforementioned elements into the Strategist modeling tool,
which allows us to explore how we best meet our customer and policy requirements
under a variety of conditions and at a reasonable cost. We work with internal and
external subject matter experts to develop starting assumptions that reflect their
expert opinion of likely future conditions. We then test the robustness of the plans
through sensitivity analysis by individually changing key assumptions and re-running
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 63 of 139

the plans under these changed assumptions. Our analysis resulted in the following
Reference Case Expansion Plan, depicted in Tables 3-3 and 3-4 below:
Table 3-3: Reference Case Annual Expansion Plan (UCAP)23
Grid-Scale
Solar
Battery
CC
Firm
Dispatchable
DR
EE
Wind
Distributed
Solar
Total

2020
0

2021
0

2022
0

2023
0

2024
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

270
115
0
154

20
130
0
22

21
116
0
22

10
133
20
21

540

172

159

184

2025
250

2026
0

2027
500

0
0
0

0
0
0

0
727
0

17
143
7
21

41
145
11
21

12
154
10
21

188

468

196

2028
250

2029
250

2030
0

2031
500

2032
250

2033
0

2034
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
206

0
0
330

0
0
330

14
157
11
20

15
155
2
20

17
140
17
20

19
138
2
20

20
136
9
20

21
129
5
20

22
126
82
20

23
126
247
19

1,453

442

443

179

685

630

579

745

Table 3-4: Reference Case Annual Expansion Plan (ICAP)24
Grid-Scale
Solar
Battery
CC
Firm
Dispatchable
DR
EE
Wind
Distributed
Solar
Total

2020
0

2021
0

2022
0

2023
0

2024
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

270
115
0
154

20
130
0
22

21
116
0
22

10
133
126
21

540

172

159

290

2025
500

2026
0

2027
1000

0
0
0

0
0
0

0
835
0

17
143
45
21

41
145
70
21

12
154
66
21

226

777

252

2028
500

2029
500

2030
0

2031
1000

2032
500

2033
0

2034
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
232

0
0
374

0
0
374

14
157
72
20

15
155
10
20

17
140
107
20

19
138
16
20

20
136
56
20

21
129
31
20

22
126
523
20

23
126
1581
19

2,098

700

784

193

1,232

932

1,065

2,123

The Reference Case presented here would result in the following energy mix:

Note: This table includes CC, EE, DR, and Distributed Solar resources that are also reflected in the Load
and Resources Table.
24 Note: This table shows ICAP values of the resources indicated in Table 3-4 above.
23

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 3
Page 64 of 139

Figure 3-9: Reference Case Energy Mix in 2020 and 2034

We outline and discuss the starting assumptions, scenarios, and sensitivities that
formed our Strategist analysis, and resulted in our Preferred Plan, below and in
Appendices F2 and F3.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 65 of 139

CHAPTER 4
THE PREFERRED PLAN
The Preferred Plan we propose in this 2020-2034 Upper Midwest Resource Plan
reflects extensive collaboration with stakeholders as well as independent expert
analysis. It supports our states’ clean energy goals and the Company’s goal of
reducing carbon emissions 80 percent by 2030 – and our ultimate vision of 100
percent carbon-free energy by 2050.
Key components of our Preferred Plan include:
 Retiring nearly 2,400 MW of remaining coal-fired capacity by 2030, including
previously approved Sherco 1 and 2 retirements and newly proposed
accelerated retirement timelines for Sherco 3 and King. We also plan to
implement seasonal dispatch at Sherco 2 prior to its retirement.
 Adding thousands of megawatts of new renewable resources, including
substantial solar additions and replacement of expiring wind PPAs.
 Continuing to operate our nuclear plants at least until the end of their licenses,
and extending operation of Monticello to 2040, as these resources anchor the
grid in around the clock carbon-free energy.
 Significantly increasing our EE and DR resources, which will reduce our overall
system demand.
Maintaining grid reliability and resilience through this transformation – as we must –
will require firm and dispatchable load supporting resources and potentially significant
transmission development. Accordingly, in order to meet reliability needs and
support renewable integration as we retire legacy coal units, our Plan includes
continued operation of our nuclear units (including a proposed 10-year extension of
Monticello), acquisition of the MEC CC, and construction of a CC at Sherco, which
we proposed in our last Resource Plan.
At the same time, we believe there may are exciting opportunities to pilot batteries,
DER, and other clean, innovative technologies in the Upper Midwest. With respect
to DR, in particular, we are seeking the flexibility to evaluate and pursue the required
incremental DR through a variety of means and technologies that may go beyond
conventional DR.
We have also sought to retain strategic flexibility by deferring decisions on certain
generating units such as Prairie Island, which can be addressed in the next planning
cycle. Doing so leaves room for innovation and allows for reassessment of
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 66 of 139

technologies, costs, and capabilities before making substantial investments. It is also
consistent with our longstanding belief that a deliberate and well-thought-out fleet
transition is critical to facilitating successful community and employee transitions.
Finally, our Preferred Plan comes at a reasonable cost to our customers – with
estimated rate impacts that are at, or below, the rate of inflation. In other words, we
can achieve industry-leading reduction to CO2 emissions at a cost that is consistent
with the expected national average increase in electricity prices.
In summary, the course we have charted in this Preferred Plan drives toward our goal
of achieving significant carbon reductions by 2030 and positions us to deliver on our
longer term vision of a carbon-free electricity mix by 2050 – all without sacrificing our
ability to deliver the reliable and affordable power that our customers count on every
day. In this section we discuss: (1) our primary planning objectives and how they are
reflected in our Preferred Plan, (2) the key components of the Plan and the actions we
intend to take to achieve it, (3) the estimated customer cost impacts of our Preferred
Plan, and (4) how this Plan meets the Commission’s public interest objectives. We
take each in turn.
I.

PLANNING OBJECTIVES

When we began this Resource Plan process more than a year ago, we framed key
planning objectives that would set the framework for development of our plan. The
objectives are complex. They sometimes overlap and conflict, but each played a
critical role in guiding our thinking and analysis which ultimately culminated in a plan
that achieves substantial environmental benefits, maintains reliability, keeps costs low,
and minimizes risks to our customers.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 67 of 139

Figure 4-1: Xcel Energy Integrated Resource Plan Objectives

A.

Environmental and Innovation

Environmental benefits and the technological innovations that will help us achieve
them are front and center in this Resource Plan process. We have made a bold
commitment to achieve 80 percent carbon reduction from 2005 levels by 2030, and
have considered this target a modeling pillar for all of our potential scenarios. Our
Preferred Plan achieves this goal in several ways. First, our Preferred Plan eliminates
coal from our system by 2030, extends our carbon-free Monticello nuclear plant to
2040, adds at least 4,000 MW of new renewable resources, including substantial new
solar capacity additions, maintains the wind levels committed to in our previous
resource plan, and replaces renewables with renewables when they reach the end of
their life.
It is important to note that, because many of these resource additions are not needed
for a number of years, maintaining flexibility in how we achieve our carbon goals is
essential. We have watched the planning landscape evolve at a remarkable rate over
the last decade and we expect the rapid pace of innovation to continue. In fact, we
expect technological advancements and innovations will create opportunities that we
can seize upon in future procurement processes and integrated resource planning
cycles if we retain the flexibility to do so. For example, future technology costs and
transmission considerations may influence our mix of wind, solar and other nonemitting resources. Likewise, the need for firm and dispatchable load supporting
capacity additions beyond 2030 may be better filled by battery storage and other
advanced technology solutions. Where appropriate, we aim to be technology agnostic
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 68 of 139

and open to what is coming next.
B.

Reliability

Our responsibility to ensure a reliable electricity supply for our customers is a
fundamental underpinning of our Preferred Plan. We therefore developed a
Reliability Requirement that establishes a minimum level of firm dispatchable
resources that is required to serve our customers’ needs in every hour of every day.
The Reliability Requirement was developed through analysis of industry trends and
careful study of our system’s performance (and the broader MISO system’s
performance) during both winter and summer days when renewables were
unavailable, sometimes for lengthy durations. We discuss the development of the
Reliability Requirement in greater detail in Appendix J2.
This Requirement does not drive any resource additions in our Preferred Plan until
after 2030. Prior to 2030, our Preferred Plan relies on two primary sources to ensure
reliability: (1) the MEC and Sherco CCs and (2) our nuclear units. With respect to the
CC units, intermediate gas resources efficiently address reliability challenges because
they can vary output to adapt as demand for electricity changes over the course of the
day and year. The CC units are large rotating machines, so also provide important
grid stability benefits and can also play an important role in our blackstart plans.1
With respect to nuclear generation, our proposed Monticello extension not only
represents a carbon-free workhorse of a resource, it also enhances fuel diversity and
provides a generation resource that is not subject to seasonal fuel supply limitations.
C.

Cost

Along with leading the clean energy transition and enhancing the customer
experience, keeping customer costs low is one of Xcel Energy’s central, guiding
objectives. Since our last Resource Plan, renewable technology costs – and in
particular, solar costs – have continued to decline; we expect this trend to continue
going forward. Taking advantage of technological advancements is one reason that
we can deliver a Preferred Plan that delivers deep carbon reductions for a nominal
customer cost of just over one percent Compound Annual Growth Rate (CAGR)
over the planning period. And over the long run, our Preferred Plan is expected to
yield net present value savings. In comparison to the Reference Case, which does not
As previously discussed, upcoming generation retirements will impact our current blackstart plans (i.e., our
ability to restart the system in the event of a catastrophic failure). While we do not propose any specific action
related to the system blackstart at this time, we anticipate addressing this in our next Resource Plan or earlier,
if system needs dictate the need to do so.

1

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 69 of 139

include accelerated coal retirements or an extension of the Monticello nuclear unit,
the Preferred Plan yields $203 million of benefits on present value revenue
requirements (PVRR) basis and $461 million of benefits on a present value societal
costs (PVSC) basis.
D.

Risk and Flexibility

Finally, while holding environmental, reliability, and cost objectives in balance, we also
seek to mitigate customer risk by ensuring fuel diversity, maintaining appropriate
capacity length in our portfolio, and maintaining flexibility in our plans. Portfolio fuel
diversity is essential to risk mitigation – especially so, as we transition away from coal.
Incorporating a mix of nuclear, load management, intermediate and peaking natural
gas capacity, and renewables into our long-term plans ensures that our portfolio is
adequately diverse and mitigates the risk associated with overdependence on any one
fuel source. Further, the proposed resource additions identified in our Preferred Plan
result in a capacity position that is between 500 to 1,000 MW long in any given year.
We believe this modest length is prudent, particularly as we propose to substantially
increase renewable resources – adding more than 4,000 MW of incremental new
renewable capacity, much of which we anticipate will be grid-scale solar – in addition
to our already large wind fleet.
Both MISO and independent analyses suggest that capacity accreditation for solar in
particular will decline substantially as more capacity is added. We expect MISO will
ultimately recognize this conclusion from its ongoing study of issues associated with
integration of high levels of renewables in its planning construct.2 Therefore, what we
believe today to be a long capacity position may actually erode over time.
Maintaining a significant amount of flexibility in our future plans is essential to reliably
and affordably navigating the transition of our fleet. To that end, we are deferring a
decision on pursuing a license extension at the Prairie Island nuclear plant to
subsequent resource plans, thereby preserving flexibility to respond to market
conditions at that time. We are also optimistic that the firm dispatchable, load
supporting resources needed in the post-2030 timeframe could be provided by new
non-emitting technologies rather than traditional gas CTs. In addition, as we look to
add solar resources to meet capacity needs in the mid to late 2020s, we are also open
to allowing other resource types to compete, to ensure that we secure the most costeffective resource solutions for our customers. As the industry and technology
We discuss MISO’s Renewable Integration Impact Assessment (RIIA) in more detail in our Baseload Study,
provided as Appendix J1.
2

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 70 of 139

continue to rapidly evolve we will evaluate opportunities to bring these potential
alternative solutions onto our grid.
E.

Our Employees and Communities

Underscoring all four of our objectives is our commitment to our employees and the
communities within which we operate. We do not make plant closure decisions
lightly, and we are committed to supporting our employees at the Sherco and King
plants as we prepare to retire these facilities. In the past we have provided career
support services to our employees facing plant closures, and we expect to continue
providing this support in the future. We also know that the Company is a major
presence in terms of employment and local tax revenues in Becker and Oak Park
Heights and the surrounding areas. We also have partners at our Sherco site with
Liberty Paper and SMMPA (Southern Minnesota Municipal Power Agency). We are
currently participating, alongside Minnesota Power, in a Host Community Impact
study, to better understand the potential impact of power plant retirements on host
communities. A similar study helped inform our work with the communities
surrounding our Sherco 1 and 2 Units, and their planned closure as approved in our
last Resource Plan. Since that time, we have worked with Becker and Sherburne
County, as well as existing and prospective customers to spur economic development
in the area, which also includes our plans to build, own, and operate the Sherco CC.
As discussed in Appendix O1: Summary of IRP Stakeholder Engagement, we are
committed to continue to work with our employees and communities to navigate this
transition together.
II.

THE PREFERRED PLAN

Our Preferred Plan is the product of an unprecedented stakeholder process that
included 13 public workshops, independent expert analysis, and months of analysis
and information sharing. As a result of those efforts – as well as the significant
engagement of our stakeholder community over the past year, our Preferred Plan is
the product of an unusual amount of consensus this early in the Resource Plan
process.
Significant consensus has emerged around the following components of our Preferred
Plan:
 Elimination of coal-fired generation from our system by 2030,
 Reduced, seasonal dispatch of Sherco Unit 2 until its retirement in 2023,
 Acquisition of at least 3,000 MW of utility-scale solar by 2030,
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 71 of 139

 A substantial increase in EE savings, and
 Support for the Company’s proposal to take ownership of the MEC.
Our Preferred Plan builds upon this foundation and includes even more renewable
resources, additional DR resources, continued operation of our carbon-free
Monticello nuclear plant for an additional 10 years, and a new CC at our Sherco site.
In the balance of this section, we present the change in our Energy Mix that will result
from our Preferred Plan, and discuss key aspects of the transition below.
A.

Transforming Our Energy Mix

From an energy mix perspective, the Preferred Plan eliminates the coal energy
contribution and increases the renewable energy contribution by over 20 percent by
2034, rising to approximately 56 percent.
Figure 4-2: Preferred Plan Energy Mix

The fleet transformation underlying our Preferred Plan achieves a nearly 84 percent
reduction in CO2 emissions from 2005 levels by 2030, and maintains at least an 80
percent CO2 reduction through 2034.
Table 4-1 below presents the amount and timing of the resource additions that
comprise our Preferred Plan.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 72 of 139

Table 4-1: Preferred Plan Resource Additions (MW)
Grid-Scale
Solar
Battery
CC
Firm
Dispatchable
DR
EE
Wind
Distributed
Solar
Total

2020
0

2021
0

2022
0

2023
0

2024
0

2025
500

2026
500

2027
1000

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
0
0

0
835
0

270
115
0
154

20
130
0
22

21
116
0
22

10
133
126
21

17
143
45
21

41
145
70
21

12
154
66
21

540

172

159

290

226

777

752

2028
500

2029
500

2030
500

2031
0

2032
500

2033
0

2034
0

0
0
0

0
0
0

0
0
0

0
0
606

0
0
0

0
0
374

0
0
748

14
157
72
20

15
155
10
20

17
140
107
20

19
138
16
20

20
136
56
20

21
129
31
20

22
126
523
20

23
126
81
19

2,098

700

784

693

838

700

1,065

997

In the next section, we discuss key aspects of our Preferred Plan and our fleet
transition in more detail.
B.

Fleet Transition

Our 2020-2034 Resource Plan represents a progressive step forward in transitioning
our fleet – meaning that the Company will complete its transition away from coalfired generation in 2030 – a full decade earlier than previously anticipated. In total,
we plan to retire approximately 2,400 MW of coal-fired generation in the next decade.
This will be an unprecedented period of transition for our system that necessitates a
prudent replacement strategy. Our strategy for replacing this capacity includes a
significant amount of additional renewable generation supported by natural gas CC
resources, continued reliance on nuclear generation, and large EE and DR additions
during the planning period.
1.

Renewable Resources

Substantial renewable additions are a central component of our energy future and a
cornerstone of this Preferred Plan. Although the quantities of future wind and solar
additions may shift somewhat in concert with technology and market fluctuations, our
commitment to renewable energy will not. In total, our Preferred Plan envisions a
system that is approaching 60 percent renewable in 2034. High levels of renewables
combined with cost-effective gas and nuclear generation will combine to create a safe
and reliable system that will withstand the summer and winter extremes of the Upper
Midwest. Our Preferred Plan proposes to add at least 4,000 MW of cumulative
utility-scale resources by 2034 (the first being in 2025) and approximately 1,200 MW
of wind by 2034, to replace wind that would otherwise retire from our system during
that period.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 73 of 139

With these additions, there would be enough solar generation to power more than
650,000 homes each year. Wind generation also continues to play a prominent role in
this Preferred Plan. Xcel Energy has long been one of the nation’s leading providers
of wind energy, and we are currently engaged in the largest build-out of new wind
resources in our Company’s history – thanks in large part to the Commission’s
approval of our last resource plan and our 1,850 MW wind portfolio. By 2034, wind
will provide 37 percent of the electricity for our customers in this region, making it
the largest component of our overall generation portfolio.
2.

Coal Resources

Large coal-based generating units have been an important baseload resource to stable
grid operations for many decades. As we and other utilities move away from coal for
economic, environmental and public policy reasons, we must do so with the
maintenance of a resilient and reliable grid at the forefront of our minds.
In our most recent Resource Plan, the Commission approved our proposal to retire
our Sherco Units 1 and 2 in 2026 and 2023, respectively. Our proposal to retire these
units early was informed by technical analyses that also determined the Sherco CC
included in this Preferred Plan is essential to mitigate grid issues. Similarly, our
proposal in this Resource Plan to retire our remaining two coals units early – King in
2028 (nine years early), and Sherco 3 in 2030 (ten years early) – is informed by
technical and other analyses discussed in our Baseload Study, provided as Appendix
J1. As also described in the Economic Modeling Framework Chapter of this
Resource Plan, we have included estimated grid reliability mitigation costs into our
Strategist modeling underlying the Preferred Plan.
Finally, our Preferred Plan also includes a commitment to offer Sherco Unit 2 into
MISO on a seasonal basis until its retirement in 2023 and Commission consideration.
3.

Nuclear Resources

Our Preferred Plan proposes to operate our Monticello nuclear unit through 2040 (10
years longer than its current license), and continued operation of both of our Prairie
Island units through the end of their current licenses (PI Unit 1 to 2033 and PI Unit 2
to 2034).3 Nuclear is central to achieving our carbon reduction goals while
incorporating incremental renewables at a reasonable pace and maintaining reliability.
Given that our operating licenses for Prairie Island run until 2033 and 2034, we believe there is sufficient
time to address the future of that plant in upcoming resource plans.
3

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 74 of 139

Nuclear is also an important system resource during the winter months, as it does not
experience fuel supply issues and has a great track record during cold weather events
– making it a critical component of our reliability strategy. Finally, the continued
operation of Monticello contributes to the affordability of our plan by leveraging an
existing, high-performing asset on our system. We discuss the benefits of nuclear, as
well as the performance of our nuclear fleet, in greater detail in Appendix K.
4.

Combined Cycle Resources

In addition to our carbon-free nuclear baseload, the continuation of dispatchable
generation on our system will be vital to our ability to manage the retirement of
approximately 2,400 MW of coal-fired generation over the next decade while
maintaining reliability. It will also facilitate our ability to successfully integrate large
amounts of renewables, because we can ramp the output of these resources up or
down in response to our system’s changing needs throughout the day as renewable
resources generate more or less due to their variable nature. To that end, our
Preferred Plan includes our proposed acquisition of MEC,4 which is a 760 MW twounit CC, as well as our plan to build the approximately 800 MW Sherco CC located in
Becker, Minnesota in the mid-2020s.
As discussed in the pending MEC docket, that plant is already an integral part of our
system, as its output is committed to the Company through two Commissionapproved PPAs. By securing ownership of the plant, we can mitigate the risk
associated with expiration of the first PPA in 2026, thereby achieving additional
certainty with respect to capacity and dispatchable energy. As discussed in our last
Resource Plan, siting a CC at the existing Sherco site will cost-effectively address grid
issues identified by the MISO Attachment Y2 study of the Sherco Unit 1 and 2
retirements. Additionally, the Sherco CC will primarily offset the retirement of other
gas units on our system, including the Cottage Grove facility (approximately 250 MW
in 2027) and Black Dog 5 (approximately 300 MW in 2032). Replacing this capacity is
not only reasonable but operationally necessary in light of the much larger coal
retirements planned and the large amounts of variable renewable additions we
anticipate in the same period. The Sherco CC will also mitigate impacts to the local
community and our employees, and potentially provide improved access to natural gas
supplies for communities in Central Minnesota. We discuss the Sherco CC further in
Appendix L.

As proposed in Docket No. IP6949,E002/PA-18-702. We will incorporate any Commission decision from
that docket into our modeling and supplement the record as necessary.

4

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

5.

Resource Plan – Chapter 4
Page 75 of 139

Firm and Dispatchable Load Supporting Resources

Reliability is central to resource planning. We are particularly focused on the
reliability of our system in this Resource Plan, however, as we embark on a complete
transformation of our baseload fleet, and transition to a portfolio of variable
renewables that approaches 60 percent of our overall generation. Our transition to
cleaner energy will only be successful if we can execute our vision without disrupting
our customers’ lives and businesses by ensuring a resilient grid that enables us to meet
our obligation to provide reliable service.
To this end, our Preferred Plan proposes to begin adding approximately 1,700 MW of
cumulative firm dispatchable, load-supporting resources in the 2031 to 2034
timeframe. The need for these dispatchable resources emerges in this later timeframe
due to the major plant retirements already discussed, as well as the expiration of
several PPAs. Our reliability analysis underlying this Resource Plan demonstrates that
these additions are necessary to continue to support grid reliability and resiliency in
light of the increased renewables being added to the system and the baseload units
being retired.
That said, because these units are not needed until the out-years of our current plan,
we have not identified a specific resource type to meet this need. With the expected
price declines and technology development between now and the 2030s, we believe
storage will be an integral resource used to meet this need. Likewise, we believe the
deployment of advanced grid investments could position DR to better compete with
traditional generation for this purpose. We are committed to pursuing all of these
options not only in the longer term, but also in the near-term in order to leverage this
technology as it matures. Generally, by keeping options open and remaining
technology agnostic, we can acknowledge the need for a firm resource at the tail end
of our plan, but allow the market to advance as we submit future Resource Plans,
continue to collaborate with our stakeholders, and engage with the Commission as the
need for these resources begins to materialize.
In the meantime, we are analyzing potential locations and sizing of storage solutions
as well as the potential values storage assets might provide to the system.
6.

Energy Efficiency and Demand Response

Demand Side Management (DSM, which collectively is EE and DR) resources
empower our customers to control their energy usage and their monthly electric bills.
Load control DR programs are an important part of our resource mix as they can be
used during periods of peak demand, helping maintain system reliability. EE reduces
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 76 of 139

the consumption of energy all together, which has both system and environmental
benefits. Taken together, DSM resources are an important part of maintaining system
reliability as well reaching our environmental goals.
The DSM aspects of our Preferred Plan includes average annual energy savings of
over 780 GWh in each of 2020-2034, compared to average annual energy savings of
444 GWh in our last Resource Plan. In addition, our Preferred Plan also incorporates
an incremental 400 MW of DR by 2023 and grows to over 1,500 MW total by the end
of the planning period. Importantly, this Resource Plan also signals a change in how
we approach EE. In previous plans, we have treated EE as a reduction to customer
load. In this Resource Plan, EE is considered a supply-side resource that the
economic modeling considers alongside other resource types.
In our last Resource Plan, the Commission approved 1.5 percent annual EE savings
on a go-forward basis. The level of EE we propose in this Plan is based on the 2018
Minnesota Energy Efficiency Potential Study, and proposes to achieve savings levels ranging
from approximately two percent to 2.5 percent annually. This level of EE achieves
more than 800 MW of additional demand savings by 2034 compared to the 1.5
percent level approved in our last Resource Plan.
Finally, consistent with the Commission’s Order in our last Resource Plan, our
Preferred Plan proposes to add 400 MW of incremental DR by 2023, and grows our
total portfolio to over 1,500 MW total by the end of the planning period. When it
comes to DR, the Company leads the way in MISO, with 830 MW registered in the
current planning year. In our last Resource Plan, the Commission ordered the
addition of 400 MW of incremental DR by 2023. As we understood the Commission’s
reasoning, it sought to add incremental, cost effective DR to avoid near-term reliance
on additional combustion turbines (CTs). As can be seen in our analysis, however, no
CTs or other firm, dispatchable resource additions are required until the 2031
timeframe as the model instead prefers solar additions in the 2025-2030 timeframe.
That said, we decided to include the DR in our Preferred Plan for several reasons: (1)
to be consistent with the Commission’s Order in our last Resource Plan, (2) to fill
gaps if/when the solar capacity credit declines, (3) to help meet firm dispatchable
resource needs in the 2030s, (4) to help support customer programs, and (5) to
integrate new and emerging technology and tools. We note that for purposes of our
modeling, we have included all of the DR identified in the Brattle study as costeffective, including expansions to conventional DR programs (i.e., Savers Switch,
smart thermostats, and interruptible rates) and a non-conventional smart electric
water heater program. Additionally, we included the addition of Auto DR, another

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 77 of 139

non-conventional DR program that automates control of various end-uses like HVAC
and lighting.
We believe the advancement of our grid, and technology generally, may take the form
of less traditional DR, so with this Resource Plan we are requesting the flexibility to
evaluate and pursue the required incremental DR through a variety of means and
technologies over the coming years. Our objective with this resource type is to bring
forward information on several viable options so the Commission, stakeholders, and
the Company can engage in an informed exchange. We provide an analysis and
detailed discussion of EE and DR in Appendix G1. We also discuss how we applied
EE and DR as supply-side resources in our Strategist modeling in Chapter 5.
Economic Modeling Framework.
C.

Keeping Rates Affordable for Customers

Our Preferred Plan keeps annual cost growth at or below the rate of inflation. In
other words, we can achieve significant CO2 emissions reductions, with cost impacts
that are roughly consistent with the expected national average increase in electricity
prices.
To show the cost impact of our Preferred Plan over the long-term, we provide as
Figure 4-3 below, a CAGR comparison to the national average nominal cost CAGR.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 78 of 139

Figure 4-3: Preferred Plan Average Nominal Cost Comparison
(NSP System)
18
17
Nominal $ cents/kWh

16
15

Average Nominal Rate of Increase (2020-2040)
EIA Forecasted National Average
NSP System Base Case
NSP System Preferred Plan

2.4%
1.2%
1.1%

14
13
12
11
10
9
8

2020

2022

2024

NSP System Base Case

2026

2028

2030

2032

NSP System Preferred Plan

2034

2036

2038

2040

EIA Forecasted National Average

* Notes: National energy cost forecast from Energy Information Administration (EIA) Annual Energy Outlook 2019, Table Energy
Supply, Disposition, Prices and Emissions – Reference Case. End use prices, all sector average.5 The Preferred Plan and Reference
Plan lines include the costs of Solar Rewards*Community.

We derived this long-term projection using a combination of a shorter-range financial
forecast and the Strategist model. The modest cost increase associated with our plan
is attributable, in large part, to our strategy of deferring resource additions until later
in the plan and making use of existing assets on our system. We believe technological
improvements will continue to drive the costs of renewables down, which is a key
driver in our strategy of proposing significant solar additions in the latter half of the
next decade.
We provide our full analysis and discussion regarding customer cost and rate impacts
in Chapter 6: Customer Cost and Rate Impacts.

See https://www.eia.gov/outlooks/aeo/data/browser/#/?id=8-AEO2019&region=00&cases=ref2019&start=2017&end=2050&f=A&linechart=~ref2019-d111618a.70-8AEO2019&ctype=linechart&sid=ref2015-d021915a.70-8-AEO2015~ref2019-d111618a.70-8AEO2019&sourcekey=0 The EIA’s Annual Energy Outlook was published in January 2019. The report is
available at https://www.eia.gov/outlooks/aeo/pdf/aeo2019.pdf.
5

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 79 of 139

In the next section, we discuss the actions we plan to take over the next five years,
and in the longer term, to achieve our Preferred Plan and deliver on our goals to
achieve deep carbon reductions in our electricity mix.
III.

ACTION PLANS

A.

Five-Year Plan

Our Preferred Plan does not identify any incremental capacity needs through 2024.
Thus, our actions in the next five years primarily address previously approved or
pending resource additions and retirements, wind repowering and procurement to
meet specific customer needs, and continuing to achieve reductions in energy demand
and load through ambitious DSM programs. We also plan to make targeted
investments in supporting infrastructure to accommodate increased renewable energy
and DER on the grid, and to gain operational experience with technologies that may
play a larger role on our grid in the future.
Wind. We expect that the 1,850 MW of wind generation resulting from our recent
acquisitions and RFPs will achieve commercial operation by 2022. These additions
were assumed across all of our scenario analyses. Further, we expect to replace the
approximately 170 MW of wind capacity that will expire in the next five years. We are
committed to pursuing repowering and/or contract extension opportunities for this
capacity, as part of our “no going back” renewables strategy. Further, we intend to
pursue incremental renewable resources as needed to meet customer needs in growing
customer programs like Renewable*Connect.
Solar. Our Preferred Plan includes significant amounts of large scale solar resources.
However, the initial planned addition of 500 MW does not occur until 2025, which is
just outside of our five-year Action Plan window. In order to procure this initial
tranche of solar, we expect to implement a competitive acquisition process in the
2023 to 2024 timeframe. We expect this timeline will allow us sufficient lead time to
acquire these solar resources and bring them online by the end of 2025.
On the distributed solar side, we have included forecasted growth in our plan. If
actual distributed solar capacity additions exceed our expectations, we anticipate this
will simply displace a portion of our proposed utility-scale solar resources.
Hydro. We anticipate adding 125 MW of energy and capacity through a PPA with
Manitoba Hydro in 2021. This incremental contract with Manitoba Hydro is in
addition to our existing PPA and diversity exchange and was executed in 2010.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 80 of 139

Nuclear. Our Preferred Plan includes a request to operate our Monticello nuclear unit
for an additional 10 years beyond its current license. While the license does not end
until 2030, we expect to begin a Certificate of Need proceeding with the Commission
within the next five years. We also expect to begin working toward license extension
with the NRC during this timeframe.
Natural Gas/Oil Peaking. We anticipate extending the life of Blue Lake Units 1-4
through 2020-2023,6 which provides 153 MW of peaking capacity to the NSP System.
Our Preferred Plan further includes our acquisition of MEC, which is currently
pending Commission consideration. Finally, we plan to continue development
activities associated with the Sherco CC during the next five years.
In addition, as discussed in our last Resource Plan, system retirements will impact our
current blackstart plans and we are currently analyzing our blackstart path to
determine the best fit for our system needs. While we do not propose any action
related to the system blackstart at this time, we anticipate addressing this in our next
Resource Plan or earlier, if system needs dictate the need to do so.
Coal. As approved in our last Resource Plan, we will take action with MISO and retire
Sherco Unit 2 in 2023, and intend to offer it into MISO on a seasonal basis until that
time. Though outside the five-year action window, we are proposing to retire the
remainder of our coal units (Sherco 1, Sherco 3 and King) before 2030. As with our
previous plant retirements, we plan to begin working with our employees and host
communities to prepare for this transition.
Demand Response. Our Preferred Plan proposes to acquire 400 MW of DR resources
by 2023, which we intend to evaluate and pursue through a variety of means and
technologies over the coming years.
Supporting infrastructure. Aside from the grid-scale and DER additions included in our
Plan, sufficient supporting infrastructure is essential to facilitate our fleet
transformation, ensure grid resilience and reliability, and to enable greater DER and
DR resources on our system. For example, we anticipate completing transmission
investments, such as the Huntly-Wilmarth project, in late 2021. We expect further
and substantial transmission infrastructure development will be necessary over the
long-term, which will involve planning in the near-term. We also are continuing to
refine our advanced grid strategy and intend to propose implementation of
foundational grid modernization investments – and continue our work to integrate
6

Pending decision in Docket E,G002/D-19-161

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 81 of 139

planning processes at all levels of the grid.
Gaining technology experience. As discussed above, we know that energy storage will be
essential for our future grid, in order to integrate variable renewable energy without
the use of traditional firm dispatchable generation. In the near term, we plan to take
steps to gain additional experience with energy storage in the NSP system. For
instance, we are co-investing in a microgrid project with Fort McCoy in the NSPW
system that will pair solar with storage as a resiliency solution, supplementing
traditional diesel backup generators. This project, slated to come online in 2021, will
not only support resiliency at Fort McCoy, but will also help us gain valuable
experience in maintaining and operating an energy storage facility, especially in the
context of new market guidelines in MISO. We anticipate that it will produce income
streams to the benefit of all customers through energy price arbitrage, ancillary
services, and using the battery as a capacity resource.
Resource treatment across states. We continue to explore options with the North Dakota
Public Service Commission to create a resource planning process that can more
formally accommodate generation portfolio preferences. We believe additional
discussions with all of our state Commissions will be necessary during the five-year
action planning period to address differing energy policies and changes in cost
allocations that may result.
B.

Long-Term Plan

By 2025 we expect we will have achieved approximately 60 percent CO2 reduction
from 2005 levels, per the measures highlighted in our Preferred Plan. In the 2025 and
beyond timeframe, there are several key aspects of our system that we will need to
address to ensure we can achieve both our 2030 carbon reduction goals, and
ultimately, our longer term goals to achieve 100 percent carbon-free electricity. For
instance, we anticipate that increasing levels of variable renewable energy and
additional baseload unit retirements will necessarily affect the way MISO plans for the
broader grid in the future. Notwithstanding the rapid pace of change occurring in our
industry, there are several action items on our long-term planning horizon.
New Transmission Infrastructure. Increasing renewable energy on the broader MISO grid
is nearly a certainty; wind and solar projects make up over 85 percent of proposed
capacity currently in the MISO generator interconnection queue.7 However, as noted
MISO “Generator Interconnection: Overview.” Updated as of June 1, 2019, at:
https://cdn.misoenergy.org/GIQ%20Web%20Overview272899.pdf
7

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 82 of 139

previously, the queue process is mired with delays and high interconnection cost
estimates, in large part due to a lack of available transmission capacity. While
generator replacement processes will help us place renewables on our system using
existing interconnection rights to an extent, we will need more transmission capacity
to come online to effectively achieve the level of renewables we envision on our
system – and more broadly within MISO – by 2030 and beyond.
Increasing energy storage on our system. As previously noted we expect that battery energy
storage system costs have room to decline over the next several years. We also know
that, as variable renewable adoption on the grid increases and more baseload capacity
comes offline, we will need a mix of low and non-emitting technologies that can
address renewable variability and ensure reliability cost effectively for customers.
Battery energy storage holds substantial promise for meeting these needs, and given
expected cost declines, we anticipate that we will be able to install cost effective
energy storage on the NSP System in the future.
Prairie Island. Our Preferred Plan continues the operation of both Units to the end of
their current operating licenses – 2033 for Unit 1 and 2034 for Unit 2. Our baseload
scenario modeling results presented in the next Chapter show that there may be value
in extending the life of this plant. However, given our operating licenses extend
nearly to the end of the planning period, we do not yet need to begin pursuing
relicensing. Therefore, we plan to continue working with our stakeholders and
evaluate Prairie Island in the context of our future system in subsequent resource
planning cycles, rather than locking into a decision at this time. In the meantime,
however, Prairie Island continues to serve an important function on the grid and in
providing cost-effective and carbon-free baseload power through the planning period.
North Dakota CT. As discussed further in Part IV below, the Company agreed to take
steps to locate a natural gas CT in the state of North Dakota, to be operational by
December 31, 2025. We remain committed to locating more generation in North
Dakota in the future, and we expect to address this resource in our next Resource
Plan.
Meeting Statewide Statutory Environmental Goals. Minnesota’s Next Generation Energy
Act contains a goal for statewide carbon reductions of 80 percent (from 2005 levels)
by 2050. While the statewide goal is for all sectors, the Preferred Plan we propose in
this Resource Plan achieves over 80 percent reduction by 2030. We know that the
electric sector has a unique role to play in achieving the statewide goals. E3’s
Minnesota PATHWAYS Report, included as Appendix P3, indicates that reducing
carbon emissions in the electricity sector enables beneficial electrification to further
mitigate emissions in other sectors – in particular the transportation and building
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 83 of 139

sectors. As we discuss in this Resource Plan, our path to achieving 80 percent carbon
reduction from 2005 levels in our system relies on retiring baseload coal units,
increasing renewable energy, extending nuclear operating licenses, and implementing
incremental DSM, and including energy storage. To be sure, all electric utility systems
are different and will encounter different challenges on their path to carbon reduction
– and there remain challenges to maintaining that level of carbon reduction if current
barriers (e.g. transmission constraints or lack of cost effective long-duration storage
options) are not mitigated. However, the analysis findings that led us to our current
Preferred Plan give us confidence that the electric sector as a whole can help achieve
Minnesota’s 80 percent carbon reduction by 2050 goal. We further discuss our
outlook regarding this statewide goal and potential barriers in Appendix H.
Meeting Company Goals to Achieve 100 Percent Carbon Free Electricity by 2050. As noted
throughout this Resource Plan, our Preferred Plan charts a path toward achieving our
2030 carbon reduction goals, and positions us to address our longer term vision of
achieving 100 percent carbon-free electricity by 2050. This goal, however, is not one
we can achieve cost-effectively or reliably without substantial technological
breakthroughs. This includes, in particular, carbon-free dispatchable energy resources
that will help us balance variable renewables’ output relative to customer demand.
Significant research and development is required to bring potential new technologies
to commercialization stage; these could include solutions such as longer-duration
battery energy storage, other types of energy storage, hydrogen-fired generation,
advanced nuclear technologies, carbon capture and storage, and others.8 We continue
to monitor potential emerging technologies and are excited to see what applications
will emerge to help make our vision a reality in the long term.
IV.

NORTH DAKOTA PLAN

As discussed in the Planning Landscape, we plan and operate a single Upper Midwest
system that serves customers in five states. Consistent with the terms of the
Settlement in Case No. PU-07-776, since 2008 we have filed our Upper Midwest
Resource Plans with the North Dakota Commission, and included in each of them an
analysis of a Resource Plan scenario compliant with Federal and North Dakota laws
only. As with the previous 2016-2030 Upper Midwest Resource Plan, the current
Plan refers to this scenario as simply the “North Dakota Plan.”

8

Selected emerging technologies are discussed further in Appendix F6.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

A.

Resource Plan – Chapter 4
Page 84 of 139

Plan Components

Our Preferred Plan for our Upper Midwest system is designed to support the
Company’s goal of an 80 percent reduction in carbon emissions by 2030. Our 2030
goal is not driven by a particular policy directive from one of our jurisdictions. We
believe planning to meet this goal is in the best interest of all our customers.
Therefore, this objective is reflected in our North Dakota Plan. The North Dakota
Plan differs from the Preferred Plan in the following ways:
1. All CO2 costs have been removed;
2. Incremental Demand Response (DR) was removed; and
3. Community Solar Garden (CSG) program costs are excluded.
When we developed our Preferred Plan we included the externality and regulatory
costs of CO2 approved by the Minnesota Public Utilities Commission. Removing the
CO2 in the North Dakota Plan had only minor impacts as shown below:

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 85 of 139

Table 4-2: Expansion Plan Comparisons
Preferred Plan – North Dakota Plan – Summary of Differences
Preferred Plan
2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

Large Scale Solar

0

0

0

0

0

500

500

1000

500

500

Battery
CC

0
0

0
0

0
0

0
0

0
0

0
0

0
0

0
835

0
0

0
0

Firm Dispatchable

2031

2032

500

0

500

0
0

0
0

0
0

2033

2034

Total

0

0

4,000

0
0

0
0

0
835

0

0

0

0

0

0

0

0

0

0

0

606

0

374

748

1,728

DR

270

20

21

10

17

41

12

14

15

17

19

20

21

22

23

542

EE

115

130

116

133

143

145

154

157

155

140

138

136

129

126

126

2,041

0

0

0

126

45

70

66

72

10

107

16

56

31

523

81

1,202

154

22

22

21

21

21

21

20

20

20

20

20

20

20

19

442

2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

Total

Large Scale Solar

0

0

0

0

0

500

500

500

0

1000

1000

0

500

0

0

4,000

Battery

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

Wind
Distributed Solar

North Dakota Plan

CC

0

0

0

0

0

0

0

835

0

0

0

0

0

0

0

835

Firm Dispatchable

0

0

0

0

0

0

0

0

0

0

0

606

0

374

748

1,728

DR

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

EE

115

130

116

133

143

145

154

157

155

140

138

136

129

126

126

2,041

0

0

0

126

45

70

66

72

10

107

16

56

31

523

81

1,202

154

22

22

21

21

21

21

20

20

20

20

20

20

20

19

442

Wind
Distributed Solar

Difference - North Dakota Plan Compared to Preferred Plan
2020

2021

2022

2023

2024

2025

2026

2027

2028

2029

2030

2031

2032

2033

2034

Total

Large Scale Solar

0

0

0

0

0

0

0

-500

-500

500

500

0

0

0

0

0

Battery

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

CC

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

Firm Dispatchable

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0

DR

-270

-20

-21

-10

-17

-41

-12

-14

-15

-17

-19

-20

-21

-22

‐23 

-542

EE

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0 

0

Wind

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0 

0

Distributed Solar

0

0

0

0

0

0

0

0

0

0

0

0

0

0

0 

0

In the modeling for the North Dakota Plan, solar additions for 2027-2029 in the
Preferred Plan are delayed to 2029-2030. As discussed previously, our Preferred Plan
also includes a large incremental addition of DR in recognition of the Minnesota
Commission’s Order in our last Resource Plan requiring 400 MW of additional DR by
2023. While we expect most of these DR programs to be implemented in Minnesota,
as we continue to develop additional DR programs for our system we would consider
proposing to add cost-effective DR programs for our North Dakota customers as
well.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 86 of 139

The exclusion of the costs of CSG does not impact the resources additions for the
North Dakota Plan. Instead, the costs of CSG are allocated so that North Dakota
customers pay a market rate for the energy from the CSG resources. The allocation
of the costs to North Dakota will also reflect previous cost-recovery decisions that
exclude costs related to the disputed resources identified in the rate case Settlement of
Case No. PU-12-813 and subsequent cases.
B.

Resource Planning Framework Status

On December 21, 2018 we proposed a framework that outlined essential pieces of a
North Dakota Resource Plan, including a default presumption that the system would
continue to be planned in an integrated fashion.9 We discussed the proposal at an
informal hearing with North Dakota Commissioners in March of this year. The
North Dakota Commission confirmed that they are interested in a more formalized
resource planning process. We look forward to working with the North Dakota
Commission and Staff to further develop a North Dakota planning process.
C.

North Dakota Combustion Turbine

Pursuant to the Settlement in Case No. PU-12-813, the Company agreed to take steps
to locate a system natural gas CT in the state of North Dakota, to be operational by
December 31, 2025. Specifically, Xcel Energy agreed to:
…develop, own, and operate (or alternatively, cause to be developed and operated on
its behalf through a power purchase agreement or other contractual arrangement) a
combustion turbine with a capacity of at least 200 MW in eastern North Dakota,
no later than December 31, 2025. The costs of the generating facility will be allocated
to all state jurisdictions served by the Company in a manner consistent with other
NSP System resources. Attainment of this commitment is contingent on the
Company's receipt of all necessary and appropriate permits and regulatory approvals.
Further, except as modified by this Section II, all provisions of the 2036
Commitment remain in place, including without limitation, the requirements that the
combustion turbine agreed to in this paragraph reasonably 1) addresses a system
capacity need and 2) represents a least-cost resource when also considering the local
reliability and system benefits of developing thermal generation in North Dakota.
The five-year Action Plan associated with this 2020-2034 Resource Plan runs through
9

See North Dakota Case No. PU-12-813.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 87 of 139

2024. Thus, the Commission will not find specific mention of a North Dakota
natural gas CT addition in the current short-term Action Plan; rather, proposed
resource additions in 2025 will be within the Action Plan developed in the next
Resource Planning cycle and addressed directly in that filing.
While the longer-term plan for resource additions do not reflect a firm peaking
addition until 2031, we acknowledge the above-stated Settlement commitment and
will continue to assess NSP System capacity needs over the next couple of years, the
likelihood of gaining the necessary approvals in all NSP System states, and the
operational feasibility and cost-effectiveness of a peaking plant located in eastern
North Dakota. Given the long planning horizons, many things can change in the next
5 to 10 years in terms of energy policy, technology, and economic conditions. We
remain committed to locating more generation in the state and a more timely and
beneficial option may become evident over time.
V.

PUBLIC INTEREST ANALYSIS

Based on our detailed analysis, we conclude that the Preferred Plan is in the public
interest. We believe it best balances our goals to ensure reliability, achieve significant
carbon reduction, and maintain reasonable costs to customers.
The Commission’s Rules identify the factors that the Commission is to consider when
determining if the Resource Plan selected is in the public interest.10 Specifically, these
Rules require that resource options and resource plans are to be evaluated on their
ability to:
 Maintain or improve the adequacy and reliability of utility service,
 Keep the customers’ bills and the utility rates as low as practicable, given
regulatory and other constraints,
 Minimize adverse socioeconomic effects and adverse effects upon the
environment,
 Enhance the utility’s ability to respond to changes in the financial, social, and
technological factors affecting its operations, and
 Limit the risk of adverse effects on the utility and its customers from financial,
social, and technological factors that the utility cannot control.

10

Minn. R. 7843.0500, subp. 3.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 88 of 139

Our Preferred Plan is best able to meet these criteria, especially when analyzed on a
comprehensive basis in light of the planning landscape the Company and the broader
industry operate within.
A.

Reliability

Our Preferred Plan is designed to allow us to continue to provide safe and reliable
service to our customers as we take steps to achieve deep carbon reductions in our
electricity mix. We know that as we, and other utilities in Minnesota, the Upper
Midwest, and the broader MISO market retire baseload units and continue to add
substantial variable renewable generation capacity to the grid, planning processes will
need to adapt to ensure a reliable and stable grid. We have done the work to
understand potential grid impacts from the orderly fleet transition we propose. We
have also developed a Reliability Requirement that will ensure we continue to serve
our customers’ energy needs every hour of every day in the interim until MISO
updates its planning construct to recognize the changes underway. We discuss the
work we have done to we meet our obligation to provide reliable service in our
Baseload Study, provided as Appendix J1.
Related, in connection with our pursuit of vast new quantities of renewables, we have
included the proposed MEC acquisition and planned Sherco CC as resources in our
Preferred Plan in order to ensure new renewable resources can be adequately
integrated. We also leave open the possibility that future firm, dispatchable needs
identified in the Plan could be met with non-emitting alternatives such as DR or
energy storage. Finally, we discuss other supporting infrastructure that will support
our goals to integrate additional variable renewable energy in Appendix I. By
including these various elements, our Preferred Plan positions us to ensure the
continued adequacy and reliability of the NSP system throughout the planning period
and beyond.
B.

Impact to Customer Bills

Affordability is one of the key objectives that framed our analysis. Our Preferred Plan
achieves significant carbon reductions while ensuring reliability at a cost in line with
expected inflation rates, or an annual cost increase of just over one percent, on
average, over the planning period. The fact that we do not need to take any actions in
the near-term supports the flexibility that we seek with this plan, which we believe is
more beneficial now than ever – given the pace of technological innovation and cost
reductions we have observed and expect to continue into the future. Therefore, we
believe that our Preferred Plan will keep our rates as low as practicable, given future
market and other uncertainties as we have described in this Plan.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

C.

Resource Plan – Chapter 4
Page 89 of 139

Environmental Effects

Xcel Energy is leading the nation with an ambitious goal of serving our customers
with a completely carbon-free resource mix by 2050 – and on our way to reaching this
goal, reducing our carbon emissions by 80 percent from 2005 levels by 2030. Our
Preferred Plan is a critical step in planning to meet these ambitious carbon reduction
goals, while also keeping in mind the socioeconomic impacts of retiring large baseload
generation units. We have proposed to close all of our remaining coal units by the
end of 2030, which will significantly reduce the amount of carbon attributable to our
system. At the same time, we are also planning to vastly expand the amount of
renewable energy capacity on our system, and continue to operate Monticello through
2040 – both of which will provide substantial amounts of clean energy to serve our
customers. All of these actions work to ensure we achieve our environmental goals
and carbon-free vision.11
D.

Socioeconomic Impacts

We recognize that plant closures can have a significant impact on our plant employees
and the communities that have hosted our plants for many years – and we bore this in
mind when developing our Preferred Plan. By announcing our plans to retire Sherco
3 and King far in advance of proposed retirement dates, employees will have time to
build additional skills and transition to other parts of the Company, if desired. Like
we have in the past, we will work with these employees to support their transitions.
We will also continue working with the host communities and other stakeholders
around those communities. We are currently working, in conjunction with CEE and
Minnesota Power, to understand the socioeconomic effects of the plants and their
closure on host communities. This commitment is evidenced in the work we have
been doing in the Sherco area, after closure of our Sherco Units 1 and 2 was approved
in our last Resource Plan. Among others things, we plan to develop our own gas CC
plant on the Sherco site. We have been working to draw new development to the
area, and with existing and prospective large industrial customers to locate facilities in
the area, which will help to offset tax and other impacts from the closure of the coal
units. These efforts will generate socioeconomic benefits through facility
construction and ongoing operations. Our clean energy efforts also generate
socioeconomic benefits, both in preserving key nuclear jobs through the Monticello
Note that resources we include our Preferred Plan meets and exceeds Minnesota greenhouse gas reduction
goals under 216H.02, the renewable energy standard under 216B.1691 and the solar energy standard under
216B.1691, 2f.

11

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 90 of 139

life extension, and spurring a large amount of new renewable construction over the
planning period.
E.

Flexibility to Respond to Change

Our Preferred Plan positions the Company well in the current planning landscape –
meeting near-term needs and creating flexibility for the future. As we have described,
planning constructs, policies, and technology costs are all creating uncertainty, which
lead us to prioritize strategic flexibility in our plans to preserve the most value for our
customers. For example, even though our modeling results show that extending the
operating life of Prairie Island would be cost-effective under today’s market
conditions, we also know those conditions can change rapidly. Thus, we have
deferred a decision regarding Prairie Island extension until the next Resource Plan.
We also have said that we are open to meeting our firm and dispatchable capacity
needs in the out-years of our Plan with options other than natural gas units, to the
extent technologies sufficiently develop and are economically-favorable at that time.
This flexibility enhances our ability to respond to changes in our planning landscape
that would affect our operations during the planning period and preserves some agility
for us to respond and adapt to these factors
F.

Limiting Risks

Much like the flexibility to respond to change, the strategic flexibility inherent in our
Preferred Plan limits the risk of adverse effects on the Company and our customers
from factors beyond our control. For example, the Reliability Requirement we
developed and incorporated into our modeling ensures that we have planned for
adequate firm and dispatchable energy to meet our customers’ needs until current
planning constructs adapt.
Our Preferred Plan represents the best option to meet customers’ needs in light of the
planning landscape for the planning period, presents the best path forward for the
Company, our customers, and the energy future of our Upper Midwest system, and is
thus in the public interest.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 4
Page 91 of 139

VI. CONCLUSION
The Preferred Plan we propose in this 2020-2034 Upper Midwest Resource Plan
reflects extensive collaboration with stakeholders as well as independent expert
analysis. Our Preferred Plan proposes to eliminate coal, add even more renewables,
and continue our industry-leading EE and DR programs, all while preserving
reliability and affordability for our customers. It also meets the varied interests of our
five-state Upper Midwest region. And by planning ahead and charting an orderly,
gradual transition of our generation fleet, we believe we can achieve all of these goals
while managing the impacts to our host communities and employees, preserving the
reliability and stability of our system, and maintaining affordability for our customers.
For these reasons, and those discussed throughout this filing, we believe our Preferred
Plan is in the public interest and merits Commission approval.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 92 of 139

CHAPTER 5
ECONOMIC MODELING FRAMEWORK
We have used the Strategist Resource Planning model to perform our economic
analyses since 2000. We use Strategist as our primary resource planning software to
estimate the costs of various resource expansion plan options, evaluate specific
capacity alternatives, and measure the potential risks of new environmental legislation
and other policy scenarios. Strategist results are a decision support tool to guide
development and selection of a Preferred Plan and test the robustness of the Plan
under a variety of assumptions and sensitivities.
To ultimately identify and refine our Preferred Plan, we created 15 scenarios that
examined different combinations and timing of baseload unit retirements, and the
resulting size, type, and timing of new resources we would need to add in order to
continue meeting customers’ needs and achieve our 2030 carbon reduction goals. We
refer to these scenarios as “baseload study scenarios.”
As we developed the various baseload study scenarios, we also conducted DR and EE
Bundle analyses. This is the first Resource Plan in which DR and EE Bundles are
considered supply-side resources, and as such, we had to undertake iterative analysis
alongside our scenario analysis to analyze these options appropriately. Finally, after
this analysis was completed, we used the outcomes and sensitivity tests to select and
refine a Preferred Plan.
We discuss our assumptions, scenarios, sensitivities and how these inputs guided
selection of our Preferred Plan in more detail below.
I.

ASSUMPTIONS

There are several assumptions included in our baseline data inputs that are common
across all scenarios studied. These factors may, in some cases, be varied within
sensitivities, but are largely kept constant across the default study of each scenario.
Important starting assumptions in our analysis include:
Load Forecast. The Company employs standard probabilistic analyses to determine our
load and energy demand forecasts. Our resource planning process takes the 50
percent probability level forecasts for both peak demand and energy requirements as
an input, we provide a detailed description of our load forecasting methodology as
Appendix F1.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 93 of 139

In the past, these forecasts have included adjustments to account for the effects of
EE as a load modifier. In order to accommodate modeling EE as a supply-side
resource in this Resource Planning process, we have not included any going forward
EE impacts in the load forecast for the 2020-2034 period. These energy and demand
savings are now included in the three EE Bundles that we evaluate as supply-side
resources.
We also incorporated an effective planning reserve margin of 2.98 percent, per MISO
requirements. As discussed in Chapter 3, MISO instituted an 8.4 percent planning
reserve margin requirement in the 2018-2019 planning year, and our system has a 95
percent MISO system coincident factor. Thus, our effective reserve margin is
calculated in the following manner:
Figure 5-1: Effective Reserve Margin Used in Strategist Modeling
  95 

 

 2.98 

 

 

    1
 

8.4 

 

 

 

 1 

Existing Fleet. We develop forecasts to model our existing fleet’s cost and
performance assumptions (such as variable O&M, heat rate, forced outage rate,
maintenance requirements, etc.) based on historical data, with adjustments for known
future changes where applicable. Additional operational and performance
assumptions include:
 Retirement of Sherco Units 1 and 2 in 2026 and 2023, respectively, as approved
in our last Resource Plan;
 Remaining coal units are dispatched economically beginning in 2028, reflecting
our expectations that MISO transitions to a multi-day commitment approach
that more efficiently commits resources in accordance with load serving needs
over a longer time horizon;
 Retirement of all other facilities at their current expected end of life if within
the resource planning period, unless we have specifically included costs of life
extension (e.g. for nuclear units in scenarios that include life extension);
 Continuation of our existing PPAs until their contractual termination dates, and
 Continued operation of the Company’s owned hydroelectric resources based
on historical performance.
Additional cost –related assumptions include:

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 94 of 139

 Costs are escalated based on corporate estimates of expected inflation rates,
 Costs associated with early retirement of the existing baseload coal units (King
and Sherco 3), as well as costs for early retirement or re-licensing the nuclear
plants were developed for use in the Baseload Study modeling.
Renewable Energy. In addition to the 1,850 MW of wind we are in the process of
adding to the NSP System since our last Resource Plan, we have assumed:
 Currently approved and/or operating renewable facilities (including both those
facilities we plan to own and those we plan to contract) are assumed to be
replaced at their end of life or contract expiration with the equivalent amount
of similar energy from generic wind and solar resources (i.e. wind would
replace wind, solar would replace solar). We refer to this as “no going back;”
 Accreditation of wind resources based on the 2018-2019 Planning Year 15.6
percent MISO ELCC, accreditation of solar resources at the default 50 percent
ELCC. For modeling purposes we assume these values remain the same
throughout the modeling period;
 No extension of the federal production tax credit (PTC) or investment tax
credit (ITC)1 past the expiration dates as per current law.
Markets. We run scenarios in Strategist both with “markets on” (i.e. where we can buy
and sell energy in the MISO wholesale market) and “markets off” (i.e. where we
cannot sell to the market, but purchases are still modeled). We use the “markets on”
view as a default assumption because this is more reflective of our realistic operations.
Sensitivities with markets off help us test the effects of this assumption on the various
scenarios.
Wholesale electricity price forecasts. Our electric power market prices are developed from
fundamentally-based forecasts from external analysts Wood Mackenzie, CERA and
PIRA. The forecasts we receive from these third party analysts provide monthly
average on- and off-peak market pricing at the Minn Hub. We then use that market
data to create an hourly shape for each month, based on the amount of thermal
generation dispatched on our system. The methodology results in lower hourly
locational marginal prices (LMPs) during times when a significant amount of
renewable energy is on the system and higher hourly LMPs when amounts of available
renewable energy are lower. Shaping the hourly prices in this manner provides a more
conservative view of potential benefits we may realize from selling excess generation
to the market.
1

The ITC reverts to 10% in 2022 and beyond, per current law.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 95 of 139

Purchase and sales limits. In our Strategist model, we include a limit as to the amount of
energy that we are able to either purchase from or sell into the MISO market. This
limit was developed using results from the 2018 MISO Transmission Expansion Plan
(MTEP) model results and evaluating maximum levels of market interaction achieved
in that modeling. For 2020-2023 we assume a market interaction limit of 1,800 MW
which grows to 2,300 MW after 2023, based on the anticipated in service of the
Cardinal to Hickory Creek transmission line which is expected to increase
transmission outlet in our region.
Emissions rates and costs. Emission rates for existing and planned resources are
consistent with historical and expected performance. We assume the following costs2
and apply them to emitting resources as relevant:
 Achievement of an 80 percent reduction in CO2 serving retail customers, as
measured from 2005 levels, by 2030. The overall carbon emissions are allowed
to increase slightly from these levels at the retirement dates of the nuclear fleet,
which vary by scenario;
 $ 25.00 per ton CO2 as a regulatory cost, starting in 2025 and escalating at
inflation, with the high CO2 externality value used prior to 2025. The societal
value of CO2 as an externality and other combinations of externality and
regulatory costs were included as sensitivity cases;
 The Minnesota Commission’s high externality values for other specified
emissions.
Generic Resources. Strategist uses generically-defined resources to meet future demand
when our already existing and approved resources are not sufficient in a given year.
Generic resources are modeled as incremental units of a certain installed capacity size,
but these sizes are chosen based on the amount of UCAP, or the MISO accredited
capacity value the units would yield. For example, although the generic unit size for
solar is rather large (500 MW installed capacity), the resource adequacy or MISO
capacity credit value we would expect to receive for a plant of that size is half that
(250 MW), which is more comparable to a generic thermal or storage plant we may
assess. Similarly, wind UCAP values are discounted to 15.6 percent of ICAP.

Note: As further discussed below, these costs are not used in evaluating the cost of our Preferred Plan for
North Dakota. See our discussion regarding the North Dakota Plan in Chapter 4: Preferred Plan.

2

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 96 of 139

Generic units ICAP values included in modeling are as follows:3
 331 MW gas-fired combustion turbine peaking unit (CT),
 206 MW gas-fired combustion turbine peaking unit (CT),
 856 MW gas-fired combined cycle intermediate unit (CC),
 331 MW energy storage project, with costs and performance comparable to
lithium-ion battery technology,
 750 MW wind project
 500 MW grid-scale single-axis tracking solar project
 100 MW distributed solar project
Appendix F2: Strategist Modeling Assumptions & Inputs, provides more detail on
Strategist assumptions. Please see Appendix F6: Resource Options, for additional
discussion on supply-side resource options included in the analysis.
Customer Programs. Incremental customer programs for DR and EE were included as
potential resources in the Strategist model. The derivation of these three DR and
three EE “Bundles” are described in Appendix F6.
It is important to note that these Bundles represent generic DSM additions and
therefore may not perfectly align with the size and timing of actual DR or EE
additions to the system in the future. These Bundles were developed immediately
after receiving third party studies for incorporation into modeling, without the benefit
of time to develop detailed implementation plans to achieve the levels of DSM in each
Bundle. Therefore, for incremental DR resource additions in particular, while the size
and timing of the first Bundle generally achieves the ordered 400 MW by 2023, the
actual implementation plans which detail the specific size, type, and timing of
incremental additions will likely differ. Procurement plans are illustrated in Appendix
G1: Demand Side Management.
DER is modeled with base and high adoptions assumptions, using similar levels as
provided in our 2018 IDP filing. We discuss our DER forecasts in Appendix F1:
Load and Distributed Energy Resource Forecasting.

The cost and performance data for these units are based on consultant’s estimates, publicly available thirdparty data, and internal company data. Availability dates are selected based on our estimates of the lead time
needed for regulatory approvals, financing, permitting and construction.

3

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

II.

Resource Plan – Chapter 5
Page 97 of 139

SCENARIOS

As noted above, we created 15 scenarios to examine combinations and timing of
baseload unit retirements, and the resulting size, type, and timing of new resources we
would need to add in order to continue meeting customers’ needs and achieve our
2030 carbon reduction goals. We describe key parameters of these scenarios below.
A.

Reference Case Scenario

We describe the development of our Reference Case in Chapter 3: Minimum System
Needs. The Reference Case (Scenario 1) is an extension of our 2015 Resource Plan,
in that all of the baseload units retire at their currently scheduled retirement dates, and
serves as our starting point. The approved 1,850 MW wind portfolio that is in
progress is included, along with generic wind and solar units added to the plan to
ensure that we do not fall below the current level of renewables we have on our
system (i.e. a “no going back” portfolio). Additional renewable units are evaluated
and optimized in the modeling and added where economic. In the original phases of
the modeling, the DR and EE Bundles were evaluated as optimized economic
alternatives, as were distributed solar, storage, and thermal CT and CC resources. The
Sherco CC unit and owned MEC CC unit were included in the Expansion Plan. Firm
peaking resources were included in the plan as needed to maintain the Reliability
Requirement criteria.
To determine the optimal strategy regarding the future of the baseload fleet, we
developed additional scenarios with varying combinations of baseload resource
retirement dates. The resulting system needs were then met with a Strategist modeloptimized portfolio of new resources. Internal finance, energy supply, and nuclear
subject matte experts worked to develop a robust set of assumptions and potential
retirement dates for the baseload units. These input assumptions include: ongoing
capital expenditures, O&M expenses and decommissioning and/or life extension
costs. We also incorporated the planning level estimates from the MISO Y2 studies
performed as part of our Baseload Study that informed our Preferred Plan. See
Appendix J1 for more details regarding this study. The scenarios we evaluated can be
generally grouped into families, as described below.
B.

Early Coal Family

This family of scenarios is designed to evaluate the economics (i.e. revenue
requirement impacts) of retiring King and/or Sherco 3 early. We did not study life
extension for coal facilities. For the early coal retirement scenarios, the early
retirement date for King was assumed to be the end of 2028, and for Sherco 3 the
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 98 of 139

early date was the end of 2030. We chose these retirement dates because they
generally allow an orderly and staged transition, with a major coal retirement every
two to three years. In the all early coal scenarios, for example, the retirement schedule
is Sherco 2 in 2023, Sherco 1 in 2026, King in 2028 and Sherco 3 in 2030.
 Scenario 2 (Early King) – King is retired at the end of 2028. Sherco 3 and the
nuclear units are unchanged.
 Scenario 3 (Early Sherco 3) – Sherco 3 is retired at the end of 2030. King and
the nuclear units are unchanged.
 Scenario 4 (Early All Coal) – King is retired at the end of 2028, Sherco 3 is
retired at the end of 2030, and the nuclear units are unchanged.
C.

Early Nuclear Family

This family of scenarios is designed to test the economics of retiring Monticello
and/or Prairie Island early, either alone or together, and with the combination of early
coal retirements. For the early nuclear retirement scenarios, the early retirement date
for Monticello was assumed to be the end of 2026 and for Prairie Island 1 and 2 it
was the end of 2024 and 2025, respectively. We chose these retirement dates as we
felt they best balanced the need for adequate lead time to enable an early major
nuclear retirement with the desire to evaluate retirement scenarios that occur well
ahead of the existing retirement dates in the 2030s.
 Scenario 5 (Early Monticello) – Monticello is retired at the end of 2026. Coal
and Prairie Island is unchanged.
 Scenario 6 (Early Prairie Island) – Prairie Island is fully retired by the end of
2025. Coal and Monticello is unchanged.
 Scenario 7 (Early All Nuclear) – Prairie Island and Monticello are both retired
early per the years above, the coal units are unchanged.
 Scenario 8 (Early All Baseload) – All baseload units, including coal and
nuclear, are retired early per the years indicated above.
D.

Extend Nuclear Family

This family of scenarios is designed to test the economics of re-licensing Monticello
and/or Prairie Island and extending operational life by 10 years over the current
retirement dates. For the extend nuclear scenarios, the revised date for Monticello
was assumed to be the end of 2040 and for Prairie Island 1 and 2 was the end of 2043
and 2044, respectively.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 99 of 139

 Scenario 9 (Early Coal, Extend Monticello) – All coal was retired at the early
dates and Monticello is extended for 10 years. Prairie Island is unchanged.
 Scenario 10 (Early King, Extend Monticello) – King was retired at the early
date and Monticello is extended for 10 years. Sherco 3 and Prairie Island are
unchanged.
 Scenario 11(Early Coal, Extend Prairie Island) – All coal was retired at the
early dates and Prairie Island is extended for 10 years. Monticello is
unchanged.
 Scenario 12 (Early Coal, Extend All Nuclear) – All coal was retired at the
early dates and both Monticello and Prairie Island are extended for 10 years.
 Scenario 13 (Extend Monticello) –Monticello is extended for 10 years. King,
Sherco 3 and Prairie Island are unchanged.
 Scenario 14 (Extend Prairie Island) – Prairie Island is extended for 10 years.
King, Sherco 3 and Monticello are unchanged.
 Scenario 15 (Extend All Nuclear) –Both Monticello and Prairie Island are
extended for 10 years. King and Sherco 3 are unchanged.
III.

FUTURES SCENARIOS AND SENSITIVITIES

To determine how changes in our assumptions impact the costs or characteristics of
different plans, we have historically evaluated how the plan responds to changes in
individual input assumptions. This testing helps us assess the “robustness” of each
scenario in the face of future uncertainty, meaning that we want to test how resilient
the scenario is to changes in one or more key assumptions. Generally, if a given plan
is extremely sensitive to changes in assumptions, it would not represent a prudent
course of action for the Company to pursue, because it would subject our customers
to excessive risk. While we believe there is value in evaluating the individual
sensitivities, and have provided a comprehensive analysis of those sensitivities in
Appendix F3, we took a slightly different approach to stress test our results in this
particular Resource Plan.
A.

Futures Scenarios

Consistent with the MISO MTEP Process, we adopted a scenario-based planning
approach to our sensitivity analysis that we have incorporated for the first time in this
Resource Plan. Since many of the input assumption variables in our modeling are
correlated, we believe there is more value in looking at a combination of variable

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 100 of 139

sensitivities as opposed to “one-off” sensitivity runs. Evaluating one sensitivity at a
time may isolate the impacts of the variable in question, but may not necessarily
reflect a realistic future scenario.
We developed four Futures Scenarios, using the 2018 MTEP Futures as guideposts.
The first two Futures Scenarios (Base PVSC and Base PVRR) represent our base
assumptions, with and without carbon costs, as we have consistently provided this
view as part of previous Resource Plans and are required to provide the PVRR view
for our North Dakota stakeholders. The High Electrification and High Distributed
Solar cases represent our new approach, in which we adjusted multiple sensitivities in
each Futures Scenario to assess the combined effect of these changes. While there are
certainly many assumptions we could have adjusted, we focused on the four most
important variables which include fuel price forecasts, load forecasts (or variables
impacting the load forecast like distributed solar), carbon and externality costs, and
new resource capital costs. The assumptions made for each Futures Scenario can be
seen in Table 5-1 below:
Table 5-1: 2019 Resource Plan Futures Scenarios
Futures Scenario

Description

No Carbon (PVRR)
High Electrification &
Low Tech Costs
(PVSC)

Base Case with Carbon
Costs, Similar to MISO
MTEP Continued Fleet
Change (CFC) Scenario
No Carbon Costs
Similar to MISO MTEP
Accelerated Fleet
Change (AFC) Scenario

High Distributed Solar
Deployment, Low
Tech Costs (PVSC)

Similar to MISO MTEP
Limited Fleet Change
(LFC) Scenario

Base Scenario (PVSC)

Load Forecast

Carbon &
Externality
Costs

New
Resource
Capital
Costs

Base

Base 50/50

High/High

Base

Base

Base 50/50
High
Electrification
Forecast
High DG Solar
Forecast &
Higher EE
Levels

None

Base

High/High

Low

High/High

Low

Gas, Power,
Coal Prices

High

Low

Note: bolded and underlined parameters indicate assumptions that have been modified from the Base Scenario

For the High Electrification Scenario, we examine a case in which higher load levels
are expected to stimulate higher fuel demand and consequently higher overall fuel
prices. To construct this Scenario, we used a high electrification forecast provided by
E3, informed by their Minnesota PATHWAYS study provided as Appendix P3 to this
Resource Plan, to assess the impacts of high load, high fuel price, and a low
technology cost environment.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 101 of 139

Conversely, for the High Distributed Solar Deployment Scenario, lower load levels
driven by higher levels of offsetting distributed solar could reasonably be expected to
drive down fuel demand and result in lower overall fuel prices. To construct this
Scenario, we used an internally developed high customer adoption based distributed
solar forecast to assess the impacts of low load, low fuel price and low technology
cost environment. We also forced in all three EE Bundles to further reduce the load
forecast and evaluate a future that truly stresses our baseload decision options.
In both the High Electrification and the High Distributed Solar Scenarios, we
assumed low new resource capital costs. We believe this is an appropriate assumption
to test, because trends have indicated that the market has previously underestimated
realized cost reductions in renewables and other new technologies, and we feel this
could continue to occur going forward. Likewise, in both these Futures Scenarios, we
included carbon and externality costs, consistent with resource planning principles in
Minnesota.
Figure 5-2: Peak Demand, Net of EE Impacts, by Futures Scenario (MW)
14,000

Peak Demand (MW)
* includes 2.5 percent EE impacts

13,000
12,000
11,000
10,000
9,000
8,000
7,000

Base PVSC/PVRR
High Distributed Solar PVSC

July 1, 2019

High Electrification/PVSC

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 102 of 139

$/mmBtu

Figure 5-3: Fuel Price Assumptions, by Futures Scenario
$11
$10
$9
$8
$7
$6
$5
$4
$3
$2

High Distributed Solar Gas
Base PVSC/PVRR Gas
High Electrification Gas

July 1, 2019

High Distributed Solar Coal
Base PVSC/PVRR Coal
High Electrification Coal

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 103 of 139

$70

$12

$60

$10

$50

$8

$40
$6
$30
$4

$20
$10

$2

$0

$0

Battery - Levelized Cost ($-kW-mo)

Wind/Solar - Levelized Costs ($/MWh)

Figure 5-4: New Resource Cost Assumptions, by Futures Scenario
($/MWh; $/kW-mo)

Base PVSC/PVRR Wind
Base PVSC/PVRR Solar
High Electrification & Distributed Solar - Wind Cost
High Electrification & Distributed Solar - Solar Cost
Base PVSC/PVRR Battery
High Electrification & Distributed Solar - Battery Cost

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 104 of 139

Figure 5-5: Carbon Emissions Cost Assumptions
($/Short Ton)
$60

$/Short Ton

$50
$40
$30
$20
$10
$0

PVSC, High Electrification and High Distributed Solar - CO2 Costs
PVRR - CO2 Costs

It is important to note that these Futures Scenarios were designed to test the
performance of our baseload retirement decisions under plausible future states. These
Futures Scenarios are not, however, intended to test which future is overall least cost
for our system. We do not have full control over the level of distributed solar or
electrification growth on our system, and have no control over variables such as fuel
prices and new resource capital costs. As demonstrated in the next section, the
Futures Scenario analysis shows that our Preferred Plan baseload decisions to retire all
coal by 2030 and extend Monticello are likely to yield customer benefits relative to the
Reference Case, even in a future where multiple key assumptions change
simultaneously.
B.

Traditional Sensitivities

While our primary focus has shifted to Futures Scenarios as opposed to the traditional
single sensitivities, we still believe the individual sensitivities provide insights on
potential plan performance. Therefore, consistent with previous Resource Plans, we
tested the following individual sensitivities. Detailed results for these sensitivities can
be found in Appendix F3.
 Load. The low load sensitivity includes high customer adoption-based DER
growth and higher EE savings (i.e. it includes all three EE Bundles), which
reduces load. The high load sensitivity includes high electrification load.
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 105 of 139

 Fuel Price/Market Costs. High and low price sensitivities were performed by
adjusting the growth rate up and down, respectively, by 50 percent from the
base forecast starting in year 2022.
 CO2 Values. To examine the effect of CO2 pricing, we tested high and low cost
sensitivities. We also performed a sensitivity evaluating no CO2 cost. The
PVSC Base Case CO2 values are based on the high externality cost values for
CO2 as determined by the Minnesota Commission through 2024.4 The PVSC
Base Case values starting in 2025 are based on the “high” end of the range of
regulated costs.5 Below is the list of carbon sensitivities.
o Low Externality
o Low Externality, Low Regulatory
o Mid Externality, Mid Regulatory
o High Externality
o PVRR, or No Externality or Regulatory
 Externalities. Criteria pollutants values are derived from the high and low values
for each of the three geographic locations in the Minnesota Commission
Order,6 with existing plants assigned the appropriate area and generic units
assigned to “rural.” The midpoint externality costs are the average of the low
and high values. The high, low and midpoint externality costs are used in
conjunction with the CO2 sensitivities described above.
 Resource Costs. For wind, solar and battery energy storage we use NREL’s
Annual Technology Baseline (ATB) 2018 report to provide high and low
technology cost sensitivity inputs. For wind and solar, we use the costs
projected by the ATB directly. For batteries, we take a slightly different
approach. Low and high battery costs are based the percent difference in the
NREL ATB base, low and high battery cost forecasts, with this percent
difference applied to the Company’s base battery cost forecast. We did not
adjust capital costs for thermal resources such as the generic CC or CTs, so all
scenarios include our base cost assumptions for those resources.
 Markets Sales Off. As previously discussed, we assume that markets are “on” for
each scenario. The “markets off” sensitivity represents a view in which we
4 Minnesota Commission Order Updating Environmental Cost Values in Docket No. E999/CI-14-643 issued
January 3, 2018.) at 31.
5 Minnesota Commission Order Establishing 2018 and 2019 Estimate of Future Carbon Dioxide Regulation
Costs in Docket Nos. E999/CI-07-1199 and E999/DI-17-53 issued June 11, 2018) at 12.
6 Minnesota Commission Order Updating Environmental Cost Values in Docket No. E999/CI-14-643 issued
January 3, 2018.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 106 of 139

cannot access the market to sell energy outside our system.
IV.

STRATEGIST ANALYSIS AND RESULTS

After identifying the scenarios and sensitivities for analysis, we used Strategist to
identify the expansion plans for each of the 15 primary scenarios, and their resulting
cost and emissions impacts. We faced a number of challenges as we tested the full
capabilities of the model and attempted to be responsive to stakeholder requests to
include a robust set of supply-side and demand-side resource options for
consideration. As described below, we undertook an extensive process which
attempted to balance the inclusion of a comprehensive set of resource options with
the model’s limitations, to arrive at a plan that demonstrates a high level of diligence
and investigation. This analysis was performed iteratively in several rounds, as we
refined the process and results, with each round informing modeling parameters for
subsequent rounds. Achieving an appropriate balance between resource options and
modeling runtime efficiency required the following adjustments, which are described
in more detail in the sections below:
 Removal of our generic distributed solar resource options, with the
understanding that if future DG solar growth exceeds our embedded
assumptions, it can simply displace utility scale solar additions identified in the
Preferred Plan.
 Shortening the modeling period to end in 2025, from the original 2057 end
date.
 Testing EE and DR selection via optimizations, evaluating different
combinations of the three EE and three DR Bundles for cost effectiveness, and
then locking the optimal mix of Bundles in final optimizations.
 Manually inserting CT additions as a proxy placeholder for the firm,
dispatchable resource needs driven by the Reliability Requirement. In reality,
because these additions happen in the post-2030 timeframe, we expect the need
will be met by a combination of firm dispatchable resource options. These may
include battery storage, pumped hydro, DR, natural gas, and/or others.
A.

Initial Full Optimization

In the first initial round of modeling, all technology alternatives (wind, solar,
distributed solar, storage, DR, EE, CC, CT) were made available to the model and we
developed a fully optimized expansion plan for each scenario through the end of the
available years in Strategist (2057). We found that this stretched the capabilities of the
Strategist tool. Due to the large number of alternatives, these runs took a significant
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 107 of 139

amount of time to complete; on average, five days of processing time per scenario.
Further, the results were significantly truncated.7 In these initial plans, no DR, EE, or
distributed solar alternatives were selected for any of the scenarios. Although this set
of initial runs was not reliable enough for drawing final conclusions, due to the
truncation issues, we identified several refinements for the next round of modeling.
First, when comparing utility-scale solar and distributed solar, the model did not select
any distributed solar. We reviewed the modeling data for these two alternatives and it
was clear that, because both utility-scale and distributed solar have identical capacity
accreditation8 values and similar capacity factors, the model would only ever select the
lower cost utility-scale solar. For modeling purposes, therefore, we removed
distributed solar from the optimization in order to improve model runtime and reduce
the number of truncated results. We note that this does not imply distributed solar is
not a resource we anticipate will be added to our system – only that from a modeling
perspective, distributed solar will not appear as cost-effective relative to utility-scale
solar in the modeling process, and retaining both types of resources in the model for
future runs would reduce the quality and runtime of the modeling process. As we
have explained, any growth in distributed solar we experience on our system beyond
what is in our embedded forecasts will simply serve to displace some of the utilityscale additions identified in our Preferred Plan.
Truncation challenges also informed the duration through which we modeled our
plans. While our initial runs experienced truncation issues fairly early (beginning in
the late 2020’s), the further out we attempted to optimize portfolios, the more
truncation occurred and the slower the simulation became. Based on this
observation, we determined it was prudent to shorten the modeling period, using
2045 as the end year rather than 2057. We believed that this, in combination with
adding 10 years of “end effects” in the modeling, would inform plans through the
planning period (2020-2034) that were more robust and valid than the longer
simulation would provide. Additionally, the availability, cost and performance
assumptions for technologies become increasingly subjective far into the future, and
we would not adequately account for new technologies that may develop within that
timeframe.9 Thus the results of modeling in that extended period would not be
particularly robust and would likely misstate the resource mix and cost required to
Truncation occurs when the Strategist model has more viable plans for a given year than the internal
memory is able to store. The total collection of plans is sorted by accumulated cost up to that year, and the
highest cost plans are discarded and not analyzed further. The Company’s model is set to a maximum of
2,500 saved plans per year.

7

8

I.e. Effective Load Carrying Capability, or ELCC.

We note that an independent analysis from consultant E3 highlighted similar concerns and also conducts
expansion plan modeling to 2045.

9

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 108 of 139

meet the Company’s longer term vision of 100 percent carbon-free energy by 2050.
Further, including the DR and EE Bundles in full model optimizations proved to be a
significant challenge for the Strategist model runtime efficiency. Given that the initial
full optimizations resulted in no DSM being selected, we decided to pursue an
alternative modeling path. For the next round of modeling, the first DR and the first
EE Bundle were forced into the plans to test if “seeding” the model with these
Bundles would lead to the second or third Bundles being chosen within the economic
optimization.
B.

Revised Targeted Optimization

The model revisions discussed above resulted in a somewhat improved modeling
process, shortening runtimes and reducing truncation. However, the second round of
modeling still took over two days per simulation, and displayed significant truncation,
such that we determined additional refinements were needed. This process did,
however, help us derive more information from our model runs that informed the
final stages of the modeling process.
First, in almost all 15 scenarios, once the first DR and EE Bundles were forced in, the
second EE Bundle was selected economically. No additional DR was selected. This
result indicated that there was indeed a modeling bias (most likely due to truncation)
that prevented selection of DR and EE, as defined by the Bundles, in the fully
optimized results. We concluded that some other method of “manual” testing would
be necessary to determine these resources’ true cost-effectiveness.
Additionally, some of the scenario outcomes in this revised modeling process relied
almost entirely on non-dispatchable or use-limited resources (wind, solar, storage) for
the full capacity expansion plan. At this point, resource planning consulted with
operations and engineering, and worked together to develop and implement a
modeling element that would ensure the portfolio resulting from each scenario
retained sufficient firm dispatchable generation to reliably serve customer capacity and
energy requirements. We describe the Reliability Requirement in Appendix J2.
To incorporate the Reliability Requirement into the modeling, we added firm
dispatchable load supporting resources, represented currently with CT resources as a
proxy, to the expansion plan in specific years10 to ensure the portfolio maintained the
minimum level of firm dispatchable, load supporting resources as defined by the
Reliability Requirement. Given the manual addition of firm dispatchable resources,
10

Applicable years vary by scenario.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 109 of 139

we also reduced the number of new resource alternatives available in certain years to
further improve model run times. We believe this did not sacrifice or reduce the
model’s ability to find optimal solutions for the expansion plan. As an example, the
large CC unit option was removed as an alternative in years where the incremental
capacity need relative to the previous year was small. This targeted “pruning” of
alternatives yielded faster run times and less truncation.
Accounting for all the aforementioned factors, we repeated scenario modeling while
including: (1) the Reliability Requirement, (2) the targeted resource “pruning” and (3)
one DR and one EE Bundle forced in, while still allowing the incremental DSM
Bundles to be selected in the optimization. The second EE Bundle was almost always
selected, while no additional DR was selected. Additionally, early coal retirement and
nuclear extension scenarios emerged as potential preferred options, as they showed
favorable PVSC and PVRR, when compared to other scenarios.
C.

Energy Efficiency and Demand Response Analysis

In the next phase of modeling, we worked to refine the DR and EE analysis to
identify the most cost effective Bundle combinations. After reviewing initial
modeling results, we were confident that two Bundles of EE would likely be selected
across all scenarios but wanted to conduct an additional round of tests to confirm.
Given we observed strong PVSC and PVRR performance of early coal retirement and
extended nuclear scenarios in previous rounds of modeling, we initially conducted DR
and EE testing using Scenario 9 (Early King and Sherco 3 retirement with Monticello
Extension) and Scenario 10 (Early King retirement with Monticello Extension) as a
test. To adequately analyze the Bundles, we developed PVSC and PVRR matrices by
selecting a scenario and performing optimizations that included each permutation of
the three DR and three EE Bundles. This manual process eliminated the potential for
DR and EE truncation, thus allowing us to conduct a robust analysis of each option.
We show results for Scenario 9 as an example below, and note that in both cases the 0
DR/2 EE combination returns the lowest PVSC and PVRR results.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 110 of 139

Table 5-2: Scenario 9 (Preferred Plan) DR and EE Cost Effectiveness Analyses
($2019 millions)
0 DR
0 EE $48,486
1 EE $45,390
2 EE $45,173
3 EE $45,847

PVSC
1 DR
$48,203
$45,670
$45,512
$46,166

0 DR
0 EE $40,029
1 EE $37,657
2 EE $37,476
3 EE $38,374

PVRR
1 DR
$40,216
$37,910
$37,784
$38,589

2 DR
$48,502
$45,947
$45,726
$46,389

2 DR
$40,478
$38,182
$37,925
$38,802

3 DR
$48,745
$46,152
$45,910
$46,596

PVSC Deltas (as compared to 0 DR/2 EE)
0 DR
1 DR
2 DR
3 DR
0 EE $3,313 $3,030 $3,329 $3,572
1 EE
$217
$497
$774
$979
2 EE
$339
$553
$737
3 EE
$674
$993
$1,217 $1,423

3 DR
$40,653
$38,344
$38,143
$39,009

PVRR Deltas (as compared to 0 DR/2 EE)
0 DR
1 DR
2 DR
3 DR
0 EE $2,554 $2,741 $3,003 $3,177
1 EE
$181
$435
$706
$869
2 EE
$308
$450
$668
3 EE
$899
$1,113 $1,327 $1,533

Based on this result, subsequent model runs for the baseload analysis locked in 0 DR
Bundles and two EE Bundles, and removed consideration of the remaining DR and
EE Bundles from the optimization process. As discussed further in Section V and
elsewhere in this Resource Plan however, we ultimately included the first Bundle of
DR as part of the Expansion Plan.
D.

Final Scenario Analysis

The last round of baseload scenario modeling incorporated the results of the previous
rounds into defining and executing a final analysis, which we used to draw
conclusions on the relative economics and operational performance of the 15
baseload scenarios. For the final model runs, two EE Bundles were manually added
to the plans, and the remaining Bundles were removed from the optimization, per our
previous findings that they would not be selected. The Reliability Requirement was
included as a constraint, and the number and timing of alternatives were reduced as
previously described, in order to improve model run performance without sacrificing
the ability to effectively optimize remaining resource options. We then created
expansion plans for all 15 scenarios, using PVSC assumptions, and completed the full
set of sensitivities.11

11

Minn. Stat. § 216B.2423

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

E.

Resource Plan – Chapter 5
Page 111 of 139

Modeling Results and Conclusions

Completing baseload scenario runs, as described above, allows us to examine Scenario
outcomes side-by-side, to evaluate their benefits and drawbacks. Among other
factors, we examine each Scenario’s resource expansion profile and carbon emissions
outcomes, present value costs, and several indicators of risk.
1.

Capacity Additions and Emissions Reductions

The cumulative expansion plan additions through the planning period for the 15
scenarios are shown below in Figure 5-6.
Figure 5-6: Expansion Plans by Scenario
(GW, Cumulative Nameplate Capacity Resource Additions by Fuel Type)

Base (2034)

2.0

2- Early King

2.0

2.0

4.0

0.8

1.7

10.6

3- Early Sherco 3

2.0

2.0

4.0

0.8

1.7

10.6

4- Early Coal

2.0

5- Early Monti

2.0

6- Early PI

2.0

3.5

3.5

0.8

1.7

7- Early Nuclear

2.0

3.5

3.5

0.8

1.6

8- Early Baseload

2.0

9- Early Coal; Extend
Monti

2.0

1.2

4.0

0.8

10- Early King; Extend
Monti

2.0

1.2

4.0

0.8

11- Early Coal; Extend PI

2.0

1.2

4.0

0.8

12- Early Coal; Extend
Nuclear

2.0

1.2

4.0

0.8

1.0

13- Extend Monti

2.0

1.2

4.0

0.8

0.2 8.3

14- Extend PI

2.0

1.2

4.0

0.8 0.6 8.7

15- Extend Nuclear

2.0

1.2

2.7

4.0

1.2

0.8

4.5
2.7

0.8
0.8

4.0

3.5

10.7
1.0

0.8

0.8 - 7.6

10.6

2.1

4.0

2.7

1.0

1.7
1.0

10.6
11.6
11.4

2.7

12.3

9.8
EE

9.1

1.7

9.8
9.1

Wind
Solar
CC
Firm
Peaking

As the 80 percent carbon emissions reduction target was included as a modeling
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 112 of 139

parameter, all the scenarios achieve this goal and remain under the emissions
threshold from 2030 throughout the planning period. There is minimal variability
between Scenarios on this measure, other than the timeframe in which they first reach
80 percent reduction levels.
2.

Present Value Costs

In general, plans that favored early coal retirements and nuclear extensions were the
lowest cost plans, both in terms of PVSC and PVRR. The results for the 15 scenarios
from the final modeling runs are shown below in Figures 5-7 and 5-8. The figures
show the net present value (NPV) delta of modeled costs compared to Scenario 1 (the
Reference Scenario), with negative values representing customer savings relative to the
Reference Scenario and positive values representing increased costs.12
Figure 5-7: Scenario PVSC Deltas from Reference Case
($2019 millions)
12- Early Coal; Extend all Nuclear
15- Extend Nuclear
11- Early Coal; Extend PI
14- Extend PI
10- Early King; Extend Monti
9- Early Coal; Extend Monti
13- Extend Monti
4- Early Coal
2- Early King
3- Early Sherco 3
5- Early Monti
8- Early Baseload
6- Early PI
7- Early All Nuclear
($1,200)

12

($700)

($200)

$300

$800

Note that these PVRR and PVSC deltas shown depict NPV for 2020-2045.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 113 of 139

Figure 5-8: Scenario PVRR Deltas from Reference Case
($2019 millions)
14- Extend PI
15- Extend All Nuclear
11- Early Coal; Extend PI
12- Early Coal; Extend all Nuclear
10- Early King; Extend Monti
13- Extend Monti
9- Early Coal; Extend Monti
2- Early King
4- Early Coal
3- Early Sherco 3
5- Early Monti
8- Early Baseload
6- Early PI
7- Early All Nuclear
($900)

($600)

($300)

$

$300

$600

$900

In addition to evaluating the total NPV of each scenario, we also examine the timing
of the relative costs and benefits. Examining NPV over a time series helps us make
relative comparisons of changes in costs or benefits at key “transition points” of the
plans (such as retirement or extension dates, significant renewable additions, etc.).
The cumulative total PVRR costs or savings for each of the scenarios are shown
below in Figure 5-9. Each line on the chart illustrates the running total of the annual
PVRR costs or benefits by year for a specific scenario, compared to the Reference
Case. The end point of the lines in 2045 corresponds to the final PVRR deltas shown
in Figure 5-8 above.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 114 of 139

Figure 5-9: Cumulative PVRR Cost or Savings Deltas by Scenario,
Compared to the Reference Plan
($2019 million)

$1,000
$800

PVRR ($2019 millions)

$600
$400
$200
$0
($200)
($400)
($600)
($800)
($1,000)

2- Early King

3- Early Sherco 3

4- Early Coal

5- Early Monti

6- Early PI

7- Early Nuclear

8- Early Baseload

9- Early Coal; Extend Monti

10- Early King; Extend Monti

11- Early Coal; Extend PI

12- Early Coal; Extend Nuclear

13- Extend Monti

14- Extend PI

15- Extend Nuclear

In addition to PVSC and PVRR, the Company completed sensitivity analyses for all
15 scenarios. The results of these analyses are shown in Appendix F3.
3.

Additional Risk Metrics for Baseload Scenario Evaluation

While present value costs are one factor we use to inform our Preferred Plan, we also
consider other factors and elements of risk exposure for each potential Scenario.
Consistent with the Resource Plan objectives presented in Chapter 4, we evaluated
each scenario with regard to achieving our cost, environmental, risk and reliability
goals. It is essential that our Preferred Plan meets our carbon reduction goals while
also ensuring that the system remains reliable and the plan remains affordable for
customers. Likewise, it is important we evaluate various risks, because forecasting
into the future is inherently uncertain and we want to ensure our selected Plan
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 115 of 139

remains robust, even if some key factors change. We describe the objective risk
measures we used to evaluate the scenarios in Table 5-3 below.
Table 5-3: Scenario Modeling Portfolio Risk Metrics
Objective

Metric

Definition
Traditional NPV measure of total 2020-2045 PVRR or PVSC
Base PVRR and
costs to determine least cost plan. Plans showing cost savings
Base PVSC
are preferred.
Cost
Measure of worst case potential cost outcomes across the four
Worst Case
Futures Scenarios so provides insight into plan cost risk. Plans
Futures Scenario
still showing cost savings in worst case Futures Scenario are
Cost
preferred.
Measures the absolute value of average annual total market
interaction (purchases plus sales plus dump energy) associated
Energy Risk
with each plan, to assess market energy risk exposure. Plans
with lower market energy exposure are preferred.
Measures average annual net capacity position associated with
Risk
each plan, to assess market capacity risk exposure. Typically,
plans with lower net capacity positions are considered
Capacity Risk
favorable, and our rankings reflect that. However, we also take
into account certain factors that are specific to this Resource
Plan, which affect the weight we place on this metric. These are
discussed further below in Section V.
Carbon
All plans achieve acceptable levels of carbon reduction, as a
Environmental
Emissions
result of including an 80 percent carbon reduction (relative to
Reduction
2005 levels) constraint in modeling.
Firm,
All plans achieve acceptable levels of reliability, measured by
Dispatchable
the amount of firm, dispatchable resources available, as a result
Reliability
Resource to
of including the Reliability Requirement in modeling.
Peak Load Ratio

As noted, all scenarios meet the environmental and reliability objectives, given these
targets are included as constraints in our modeling. Thus, our Scenario evaluation
focuses on the cost and risk objective metrics noted above.
V.

PREFERRED PLAN SELECTION AND ASSESSMENT

As described previously in this Chapter and in Chapter 4, we evaluated the PVRR and
PVSC results of our 15 baseload scenarios, and how effectively each potential plan
would meet our planning objectives, to determine which Scenario should form the
basis of the Preferred Plan. Based on these outcomes, we selected baseload Scenario
9. Our plan charts the path toward achieving ambitious carbon reduction goals,
reflects substantial stakeholder input and consensus, and ensures reliability and
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 116 of 139

affordability for our customers. The baseload aspects of this plan include an early
King retirement in 2028, Sherco 3 early retirement in 2030 and extension of our
Monticello nuclear facility to 2040. We also took into account additional
considerations regarding DR when finalizing the Preferred Plan. We discuss more
detail regarding how we selected and evaluated our Preferred Plan below.
A.

Baseload Study Analysis Results

From a modeling perspective, the PVSC and PVRR results are primary indicators of
the various scenarios’ economic favorability. Figures 5-8 and 5-9 shown above
indicate that the nuclear extension scenarios paired with early coal retirements yielded
the most attractive customer value relative to the Reference Case.
We note that while Baseload Scenario 9 was not the least cost of our 15 scenarios,
several lesser cost scenarios included an extension of Prairie Island’s operating license.
However, as discussed previously, Prairie Island’s license does not expire until the
2033-3034 timeframe just outside the planning period, and we believe there is risk
avoidance value in deferring a decision on Prairie Island extension until a future
Resource Planning process. As a result, we eliminated from consideration cases that
include a Prairie Island extension, as shown below.
Figure 5-10: Scenario PVSC Deltas from Reference Case,
PI Extension Cases Eliminated
($2019 millions)
12- Early Coal; Extend all Nuclear
15- Extend Nuclear
11- Early Coal; Extend PI
14- Extend PI
10- Early King; Extend Monti

Note: all scenarios that
included Prairie Island
extension were removed
from consideration, as we
believe it prudent to defer
Prairie Island extension
consideration to future
resource planning cycles.

9- Early Coal; Extend Monti
13- Extend Monti
4- Early Coal
2- Early King
3- Early Sherco 3
5- Early Monti
8- Early Baseload
6- Early PI
7- Early All Nuclear
($1,200)

July 1, 2019

($700)

($200)

$300

$800

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 117 of 139

After screening out the baseload scenarios that include Prairie Island extension, we
evaluated the remaining scenarios using the cost and risk metrics discussed previously,
including savings or costs achieved in a “worst-case” Futures Scenario, and average
energy and capacity exposure. Of the remaining baseload scenarios available for
selection, Scenarios 9 (Early Coal; Extend Monti), and 10 (Early King; Extend Monti)
achieve the most favorable risk profile overall.
Further, Minn. Stat. § 216B.2422, subd. 2(c) requires that we “include the least cost
plan for meeting 50 and 75 percent of all energy needs from both new and
refurbished generating facilities through a combination of conservation and renewable
energy resources.” The Preferred Plan (Scenario 9 - Early Coal; Extend Monti)
satisfies the statute’s first requirement (50 percent of energy needs from conservation
or renewables) because it is economically optimized and meets approximately 64
percent of energy needs with renewables and conservation. Our baseload scenario
analysis satisfies this statute’s second requirement (75 percent of energy needs from
conservation or renewables), as Scenario 4 (Early Coal) yields the least cost plan for
meeting at least 75 percent of all energy needs from both new and refurbished
generating facilities through a combination of conservation and renewable energy
resources. Because this scenario does not include a nuclear extension, it enables
greater levels of renewable additions than the Preferred Plan that meet or exceed the
75 percent threshold.
B.

Early Retirement of Sherco 3 by 2030

Excluding the Prairie Island extension scenarios, Scenarios 9 and 10 become the
optimal least cost options. Both Scenarios 9 and 10 assume an early King retirement
and Monticello extension; however, Scenario 9 includes an early Sherco 3 retirement
while Scenario 10 does not. Both Scenarios are beneficial on a PVSC and PVRR
basis, and are the most resilient of the remaining scenarios to the potential worst case
evaluated, continuing to maintain customer benefits relative to the Reference case.
Given the proximity of overall customer savings and risk considerations between
Scenario 9 and 10, we ultimately considered which case would fit best with our
strategic objectives and our understanding of stakeholder interests. We selected
Scenario 9 as our Preferred Plan, which includes the retirement of all remaining coal
units. Scenario 9 provides the best fit for our carbon goals and helps mitigate the
potential for regulatory or legislative action around carbon costs or carbon reduction
levels. Further, general market trends toward increasing levels and decreasing costs of
renewables, low natural gas prices, the need for more flexible resources, and other
factors are expected to make it more and more difficult for coal resources to operate
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 118 of 139

in an efficient and economic manner beyond 2030. Finally, our interactions with
customers, stakeholders, and shareholders alike have shown increasing interest in
achievement of carbon reductions and other environmental solutions. From a
financial risk perspective, we believe it is beneficial for the Company to reduce carbon
risk exposure, and we view transitioning our generation fleet away from coal assets is
one of the best ways to achieve that goal.
C.

Demand Response Adjustment to Scenario 9

As noted previously, the model optimization exercise did not select any of the DR
Bundles provided as options. However, the Order approving our last Resource Plan
included direction to add 400 MW of incremental DR resources. Therefore, the final
step in developing our Preferred Plan was to include DR Bundle 1 as part of Scenario
9. This addition increases our net long capacity position, where after 2025, our
position remains long by a range of 500-1,000 MW through the remainder of the
planning period. As mentioned previously, we typically view a long capacity position
as less favorable; however, we believe this is an acceptable path forward given
alignment with our risk mitigation planning objective, discussed further below.
D.

Futures Scenarios Results

As previously discussed, a final step in our analysis process evaluated the performance
of the Preferred Plan under the Futures Scenarios. Table 5-4 below provides a
summary of the Futures Scenario results. Under all of these Futures Scenarios, the
Preferred Plan provides savings relative to the Reference Case,13 which suggests that
the Plan is robust under a range of potential future conditions.
Table 5-4: Preferred Plan NPV Savings under Different Futures Scenarios
($2019 millions)

Delta

Base PVSC

Base PVRR

High
Electrification
Scenario
PVSC

(461)

(203)

(81)

High Distributed
Solar Scenario PVSC

(51)

As demonstrated in the baseload scenario analysis and in Table 5-4, the Preferred Plan
yields customer value of $203 million in the base PVRR and $461 million in the base
PVSC scenarios. Early coal retirements paired with the Monticello extension yield
Note: Each NPV result compares the Preferred Plan with Future Scenarios assumptions applied to the
Reference Case with those same assumptions applied.

13

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 119 of 139

benefits to customers particularly when carbon costs are included. In both the High
Electrification and the High Distributed Solar Futures Scenarios, customer value is
marginally reduced from the Base PVRR and PVSC scenarios with $81 million of
savings in the High Electrification Scenario and $51 million of savings in the High
Distributed Solar Scenario.
As shown in Figure 5-11 below, the Preferred Plan consistently results in customer
savings relative to the Reference Case in all Future Scenarios through 2030. At that
time, however, the Base PVSC and PVRR scenarios diverge from the High
Electrification and High Distributed Solar Scenarios mainly in the 2030 timeframe. A
number of factors impact the annual deltas in these Scenarios and drive the
divergence. Assumed low new resource capital costs in the Electrification and High
Distributed Solar Scenarios likely functions as the biggest driver in upward cost
pressure on the Preferred Plan in the 2030s, as in those Scenarios Monticello can be
replaced with cheaper renewables. Even under these conditions,, however, the results
demonstrate that over the entire planning period and across multiple Futures
Scenarios, the Preferred Plan provides overall customer savings relative to the
Reference Case. This demonstrates that the Plan is robust and beneficial to
customers, yielding savings under a host of potential future conditions.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 120 of 139

Figure 5-11: Preferred Plan Annual Costs or Savings
Compared to the Reference Case, by Scenario
($2019 millions)

$150
$100
$50
$
($50)
($100)
($150)

2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045

Savings/Cost in $2019 millions

$200

Base PVSC
High Electrification PVSC

Base PVRR
High Distributed Solar PVRR

The expansion plans for the Preferred Plan under all of the Futures Scenarios analyses
are provided below. In the High Electrification Scenario, higher load growth drives
incremental solar and firm dispatchable resource additions above what is included in
the Base PVSC/PVRR expansion plans. Specifically, the High Electrification
Scenario yields an additional 1,000 MW of solar and 748 MW of firm dispatchable
additions. In the High Distributed Solar Scenario, utility-scale solar is displaced by
incremental distributed solar, as well as additional EE resources, per the inclusion of
the third EE Bundle. Total large solar additions are decreased from 4,000 MW in the
base scenarios to 2,500 MW total in the High Distributed Solar Scenario.
Table 5-5: Preferred Plan Base Expansion Plan
(MW)
Grid-scale Solar
Battery
CC
Firm Dispatchable
DR
EE
Wind
Distributed Solar

July 1, 2019

2020
0
0
0
0
270
115
0
154

2021
0
0
0
0
20
130
0
22

2022
0
0
0
0
21
116
0
22

2023
0
0
0
0
10
133
126
21

2024
0
0
0
0
17
143
45
21

2025
500
0
0
0
41
145
70
21

2026
500
0
0
0
12
154
66
21

2027
1000
0
835
0
14
157
72
20

2028
500
0
0
0
15
155
10
20

2029
500
0
0
0
17
140
107
20

2030
500
0
0
0
19
138
16
20

2031
0
0
0
606
20
136
56
20

2032
500
0
0
0
21
129
31
20

2033
0
0
0
374
22
126
523
20

2034
0
0
0
748
23
126
81
19

Total
4,000
0
835
1,728
542
2,041
1,202
442

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 121 of 139

Table 5-6: High Electrification Scenario Expansion Plan
(MW)
2020
0
0
0
0
270
115
0
154

Grid-scale Solar
Battery
CC
Firm Dispatchable
DR
EE
Wind
Distributed Solar

2021
2022
0
0
0
0
0
0
0
0
20
21
130
116
0
0
22
22

2023
0
0
0
0
10
133
126
21

2024
0
0
0
0
17
143
45
21

2025
500
0
0
0
41
145
70
21

2026
500
0
0
0
12
154
66
21

2027
1000
0
835
0
14
157
72
20

2028
500
0
0
0
15
155
10
20

2029
500
0
0
374
17
140
107
20

2030
500
0
0
0
19
138
16
20

2031
500
0
0
606
20
136
56
20

2032
500
0
0
374
21
129
31
20

2033
0
0
0
374
22
126
523
20

2034 Total
500 5,000
0
0
0
835
748
2,476
23
542
126
2,041
81
1,202
19
442

Table 5-7: High Distributed Solar Scenario Expansion Plan
(MW)
Grid-scale Solar
Battery
CC
Firm Dispatchable
DR
EE
Wind
Distributed Solar

2020
0
0
0
0
270
151
0
154

2021
0
0
0
0
20
171
0
166

2022
0
0
0
0
21
152
0
9

2023
0
0
0
0
10
174
126
11

2024
0
0
0
0
17
186
45
9

2025
500
0
0
0
41
189
70
14

2026
0
0
0
0
12
202
66
93

2027
500
0
835
0
14
207
72
19

2028
500
0
0
0
15
203
10
127

2029
0
0
0
0
17
185
107
27

2030
0
0
0
0
19
183
16
62

2031
500
0
0
606
20
181
56
64

2032
500
0
0
0
21
171
31
67

2033
0
0
0
374
22
167
523
70

2034
0
0
0
748
23
166
81
72

Total
2,500
0
835
1,728
542
2,687
1,202
964

For simplicity, the table below shows cumulative expansion plan additions by resource
type. It is important to note that while DR, EE and battery storage are reflected as
separate categories, and no incremental additions for these resources are shown, they
would be considered to fill any firm dispatchable needs identified in the expansion
plans. Similarly, evolving economics and value could also shift the mix of wind and
solar additions.
Table 5-8: Cumulative 2020-2034 Additions by Resource Type and Scenario
(MW)

Large Scale Solar
Battery
CC
Firm Dispatchable
DR
EE
Wind
Distributed Solar

Base Preferred
Case
4,000
0
835
1,728
542
2,041
1,202
442

High
Electrification
5,000
0
835
2,476
542
2,041
1,202
442

High Distributed
Solar
2,500
0
835
1,728
542
2,687
1,202
964

 

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

E.

Resource Plan – Chapter 5
Page 122 of 139

Preferred Plan Benefits

We believe our analysis supports selection of Scenario 9, including the early retirement
of all of our coal resources by 2030 and extension of Monticello nuclear facility to
2040, as our Preferred Plan. While all of our scenarios meet the carbon goal and
Reliability Requirement we established, we believe cost and risk considerations elevate
Scenario 9 above the rest as an appropriate path forward.
1.

Cost

As demonstrated in our modeling analysis, the Preferred Plan achieves customer
value, not only under the our Base PVSC ($461 million) and PVRR ($204 million)
analysis but also under more challenging future conditions as evidenced in our High
Electrification ($81 million) and High Distributed Solar ($51 million) Futures Scenario
analysis. In addition, the Preferred Plan yields customer value under all of the
individual sensitivities run, even in the Futures Scenario that results in the worst case
customer savings outcome (High Distributed Solar Scenario). Lastly, from a customer
rate impact perspective, the Preferred Plan results in annual rate increases of just over
one (1) percent, which is below the rate of inflation.14 Altogether, we believe the
Preferred Plan delivers tangible customer savings while taking industry-leading steps
towards a carbon free future.
2.

Risk

In addition to beneficial cost outcomes, the Preferred Plan addresses major risks by
maintaining portfolio diversity, retaining optionality and effectively managing market
exposure. The Plan incorporates significant capacity additions to replace retiring
resources, consisting of a diverse portfolio of DSM, nuclear extension, solar, wind,
and firm dispatchable resource additions. Ensuring we do not become too dependent
on a single fuel source mitigates risk. In addition, deferring a decision on a potential
Prairie Island license extension affords us additional flexibility to reevaluate in future
resource planning cycles, as technology costs and other key assumptions can change
quickly.
We also evaluate factors such as energy market exposure and portfolio length. All of
our baseload scenarios show high levels of market interaction, driven in part by
significant renewable additions; but our selected Plan minimizes them relative to other
scenarios and attempts to carefully balance and pace renewable additions with other
resources. Further, we typically try to achieve a closer supply-demand balance than
14

As noted in Chapter 4: Preferred Plan and discussed further in Chapter 6: Customer Rate and Cost Impacts

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 5
Page 123 of 139

any of our baseload scenarios offer, although the 500-1000 MW of length in any
scenario is relatively minimal compared to our overall system. We believe our
Preferred Plan’s portfolio length is warranted at this time, however, and creates an
effective hedge for our customers against two key risk factors:


Capital Investment Wind Down At Retiring Plants. The retirement of all 2,400 MW
of our coal assets, in addition to a few other units by 2030 exposes our
customers to some risk as we wind down operations and reduce capital spend
at these plants. In the event of an early outage, excess capacity will give us the
option to flex the retirement dates as needed if we find that a capital
investment is not in our customers’ best interests at that time.



Renewable and Use-Limited Resource Capacity Accreditation. Solar capacity
accreditation is assumed at 50 percent credit in all years of our modeling. We
expect this to change as MISO changes its approach to forward capacity
accreditation – recognizing that as solar penetration increases in the footprint,
the accredited value of solar will decline. The same also applies for use-limited
resources like DR. We discuss this emerging MISO recognition in the
Baseload Study provided as Appendix J1, in conjunction with the Reliability
Requirement in Appendix J2, and in discussion of our Supporting
Infrastructure – Transmission & Distribution, provided as Appendix I.

VI.

CONCLUSION

Considering the above we believe our modeling and analysis fully supports selection
of the Preferred Plan, and strikes a strong balance in meeting our planning objectives,
in service of our customers’ needs. The plan sets us on a path to deliver tangible
savings to our customers, while transitioning our system to meet both our 2030
carbon reduction objectives and longer term carbon-free goals.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 124 of 139

CHAPTER 6
CUSTOMER RATE AND COST IMPACTS
Overall, our Preferred Plan results in an estimated average annual increase in revenue
requirements less than the Reference Case and just over one (1) percent overall. In
other words, we can achieve significant CO2 emissions reductions, with cost impacts
that are roughly half of the expected national average increase in electricity prices.
Both the Reference Case and the Preferred Plan are designed to meet the Company’s
goal of reducing carbon emissions 80 percent by 2030, compared to 2005 levels. We
did not do the full rate impact calculations discussed in this Chapter on a resource
portfolio that does not meet our 80 percent carbon reduction by 2030 goal, we did
run a “no 80 by 30” portfolio in our Strategist model and and confirmed that the
impacts are in line with our Preferred Plan. In other words, our carbon goals do not
materially increase costs for our customers.
To show the cost impact of our proposal over the long-term, we provide a
Compound Average Growth Rate (CAGR) comparison of our Preferred Plan
compared to the national average nominal cost CAGR for the NSP System in Figure
6-1 and Minnesota in Figure 6-2.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 125 of 139

Figure 6-1: Preferred Plan Average Nominal Cost Comparison
(NSP System)
18
17
Nominal $ cents/kWh

16
15

Average Nominal Rate of Increase (2020-2040)
EIA Forecasted National Average
NSP System Base Case
NSP System Preferred Plan

2.4%
1.2%
1.1%

14
13
12
11
10
9
8
2020

2022

2024

NSP System Base Case

2026

2028

2030

2032

NSP System Preferred Plan

2034

2036

2038

2040

EIA Forecasted National Average

* Notes: National energy cost forecast from Energy Information Administration (EIA) Annual Energy Outlook 2019, Table Energy
Supply, Disposition, Prices and Emissions – Reference Case. End use prices, all sector average.1 The Preferred Plan and Reference
Plan lines include the costs of Solar Rewards*Community.

See https://www.eia.gov/outlooks/aeo/data/browser/#/?id=8-AEO2019&region=00&cases=ref2019&start=2017&end=2050&f=A&linechart=~ref2019-d111618a.70-8AEO2019&ctype=linechart&sid=ref2015-d021915a.70-8-AEO2015~ref2019-d111618a.70-8AEO2019&sourcekey=0 The EIA’s Annual Energy Outlook was published in January 2019. The report is
available at https://www.eia.gov/outlooks/aeo/pdf/aeo2019.pdf.
1

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 126 of 139

Figure 6-2: Preferred Plan Average Nominal Cost Comparison
(State of Minnesota)
18
17
Nominal $ cents/kWh

16
15

Average Nominal Rate of Increase (2020-2040)
EIA Forecasted National Average
State of MN Base Case
State of MN Preferred Plan

2.4%
1.1%
1.1%

14
13
12
11
10
9
8
2020

2022

2024

State of MN Base Case

2026

2028

2030

2032

State of MN Preferred Plan

2034

2036

2038

2040

EIA Forecasted National Average

We derived this long-term projection using a combination of a shorter-range financial
forecast and the Strategist model.
We note that a detailed analysis of rate impacts in a resource planning process with
long-time horizons is difficult to produce due to changes in our rates and resource
needs that will occur over time. Because of the simplifying assumptions made in both
the calculation methodology and the input variables, these estimated impacts may not
correspond with actual rates that the Commission sets for various rate classes in the
future. That said, aside from updated inputs to the class cost allocation factors used
in this analysis, the methods are the same as those we used in our last Resource Plan.
In this Section, we explain how we approximated a baseline level of revenue
requirements associated with our Reference Case and measured the incremental cost
impacts of our Preferred Plan at the NSP system, State of Minnesota, and individual
State of Minnesota customer class levels. This is generally consistent with the
methodology we used in our last Resource Plan.
I.

REFERENCE CASE REVENUE REQUIREMENTS FORECAST
METHODOLOGY

To calculate the long-term rate impacts of the Preferred Plan as compared to the
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 127 of 139

Reference Case, we first developed a forecast of total rates under Reference Case
assumptions. This forecast leveraged our detailed five-year financial forecast and a
specific approach to identify costs through the end of the planning period (2034)
using the CAGR of generation and fuel costs from the Strategist model.2 Next, we
analyzed the annual cost differences by category (i.e. fuel, purchased power, capital
expenditures, operating and maintenance costs, taxes, depreciation, etc.) from the
Strategist model results for the Reference Case and the Preferred Plan to determine
the aggregate system cost impacts and jurisdictional and rate class breakouts.
To determine the overall impact to Minnesota customers and individual customer
classes in Minnesota, we converted the differential in annual expenses and capital
spend of the Preferred Plan compared to the Reference Case into a differential
revenue requirement forecast. We then jurisdictionalized the differential revenue
requirements and applied class allocation principles to calculate impacts on individual
Minnesota customer classes. We provide various rate impact analyses and discuss the
methodologies below.
II.

ESTIMATED RATE IMPACTS AND METHODOLOGY

The primary Strategist model captures only the generation-related portion of the
business, or around 50 percent of the total revenue requirements. Developing a total
rate forecast beyond 2023 when detailed Company financial models are not available
is dependant on making assumptions for capital expenditures and O&M costs for all
areas of the business, including generation (both new and existing), transmission,
distribution and corporate support services. Many of these assumptions are
speculative, and the resulting total rate forecast would be similarly speculative.
A.

Methodology

To calculate the rate impacts of the Preferred Plan, we started with the 2018 budget
forecast of total revenue requirements for the 2019-2023 period.3 To estimate
customer impacts for the immediate five-year period, we estimated revenue
requirements similar to a Jurisdictional Cost of Service (JCOSS) for each year, and
then performed an estimated Class Cost of Service (CCOSS) analysis – both of which
required us to make a number of assumptions.

The Strategist model for the Reference Case is the same model used for the Strategist aspect of our
Baseload Study, Scenario 1 Reference Case, with the exception that the first Demand Response (DR) bundle
was added to the plan, as was also added to the Preferred Plan.
3 Developed in July 2018 and updated in November 2018.
2

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 128 of 139

To determine the JCOSS, we had to make a number of assumptions, including the
following:
 Full recovery of the Company’s internal five-year forecasts of capital, O&M,
and sales,4
 Return on Equity (ROE) of 9.20 percent,5
 A forecast of debt and equity ratios and debt rates appropriate for the five-year
modeling term,
 Estimated historical regulatory adjustments made in rate cases.
To calculate longer-term rate impacts of the Preferred Plan, we used a combination of
the Company’s 5-year financial forecast and the Strategist model to project total
system revenue requirements for extended periods. For the period beyond 2023, we
escalated the capital and O&M costs in the last year of the 5-yr model by the CAGR
of the Reference Case as modeled by Strategist.6 This approach avoids speculation on
areas of the business not related to resource planning and modeling, while still using
the detailed generation-related information from the Strategist model to create a
“business as usual” long term rate projection. Finally, we calculated the annual
difference between the Preferred Plan and the Reference Case to estimate the total
rate impact of our Preferred Plan.
B.

Estimated Overall Rate Impacts

Figure 6-3 below illustrates the State of Minnesota estimated rate impacts of the
Preferred Plan compared to the Reference Case over the long-term.

Data as of November 2018.
The Company acknowledges the recent decision in the TCR docket requiring a 9.06% calculation to be used
in future filings and will implement that practice once the order is received.
6 The Reference Case is Resource Plan Scenario 1; see Appendix F2 for additional details.
4
5

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 129 of 139

6.00%
4.00%
2.00%
0.00%
-2.00%

2040

2039

2038

2037

2036

2035

2034

2033

2032

2031

2030

2029

2028

2027

2026

2025

2024

2023

2022

-6.00%

2021

-4.00%
2020

Percent diffrence from Reference
Case

Figure 6-3: Annual Percent Change in Revenue Requirements (2020-2040)
Preferred Plan above Reference Case – State of Minnesota

The modeling includes accelerated depreciation costs associated with the early
retirements of Sherco 3 and King. However, consistent with the Commission’s
actions in the approval of the early shutdown of the Benson biomass plant (Docket
No. E002/M-17-530), a regulatory asset is another tool that could be used to
accompany these early retirements. The use of a regulatory asset for the remaining
costs of these plants, including a cost of capital return on those assets, would be an
appropriate alternative to accelerating the depreciation because it would keep the
Company whole over the remainder of the plants’ remaining lives. This would also
serve to smooth the projected rate impacts over the planning period.
C.

Key Drivers

The major inflection points in the delta of revenue requirements (and rates) is driven
entirely by the differences in the set of resources that comprise each the Preferred
Plan and Reference Case; these points coincide with key differences in baseload plant
retirement dates between the two cases and the timing of replacement resources. The
reduction in revenue requirements associated with the early coal unit retirements helps
to offset a portion of the ongoing nuclear revenue requirements in the Preferred Plan
in the early 2030s, as discussed in more detail below:
 Extension of Monticello. In 2028, costs associated with the 10-year license
extension begin to ramp up in the Preferred Plan, and capital revenue
requirements and O&M costs continue through 2040. In contrast, the
Reference Case does not have ongoing capital and O&M costs for Monticello
beyond 2030 as it is retired in that case; this results in an approximate $295
million difference between the two cases in fixed costs for Monticello,
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 130 of 139

beginning in 2031.
 Retirement of Coal Units. The Reference Case contains ongoing capital and O&M
costs for King and Sherco Unit 3, whereas in the Preferred Plan the costs for
King terminate in 2029 and Sherco Unit 3 in 2031 due to early retirement. This
results in savings of approximately $45 million in fixed costs in 2029, increasing
to $110 million in 2031.
 Load Supporting Resources. The Preferred Plan has some load-supporting,
dispatchable resources added in the early 2030’s associated with the Reliability
Requirements Proxy discussed in the Baseload Study in this Resource Plan.
With the early retirement of King and Sherco Unit 3, the Preferred Plan has a
load supporting, dispatchable resource deficiency of approximately 400 MW in
that time frame. The Preferred Plan extension of Monticello helps to offset
some of this capacity deficiency. The net cost of the load supporting,
dispatchable resources in those years ranges from approximately $35 million to
$70 million.
The rate increase seen in 2031-2033 reverses in 2034 and the Preferred Plan remains
an annual savings producer thereafter. The cost savings from the Preferred Plan are
due to the extension of Monticello, which maintains the NSP system 80 percent
carbon reduction after Prairie Island retires, without the need to add significant
renewables. In the Reference Case, the model adds 2,250 MW of wind in 2034-35 to
maintain the 80 percent carbon reduction level, which adds significant costs.
D.

These Estimates are not Directly Comparable to Rate Impact Analysis
in a Rate Case

We caution that this information should not be interpreted as directly comparable to
the customer rate impact information we would provide as part of a rate case filing
for reasons including the following:
 The internal forecast for 2019-2023 is not prepared at the level of detail
necessary to support a rate case,
 While the forecast includes typical regulatory adjustments, we have not
attempted to remove one-time effects or other one-time adjustments that are
not specifically known at this time,
 We have made no assumptions of a rate case filing schedule over this period;
the forecast provided assumes full recovery of annual deficiencies, suggesting a
full rate case annually, and
 All factors of the Cost of Capital, including debt rates, return on equity, and
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 131 of 139

debt-equity ratios, are subject to change over the period.
III.

ESTIMATED RATE IMPACTS BY CLASS PER YEAR

After determining the incremental revenue requirement impacts from the Preferred
Plan and Reference Case for the Minnesota jurisdiction, we determined class revenue
requirement impacts. We provide the estimated impacts below, then discuss the
methodology and calculations that we used. The incremental revenue requirement
impact of the Preferred Plan versus the Reference Case is shown in column 3 of Table
6-1 below. Column 4 of the below Table also shows the incremental impact of the
Preferred Plan as a percent of the total State of Minnesota revenue requirement.
We calculated rate impacts in $ per kWh by dividing each class’s revenue requirement
in each year by the forecasted sales in each year.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 132 of 139

Table 6-1: Estimated Incremental Impact of Preferred Plan
State of Minnesota – All Customers
1

2

Year

State of MN Total
Revenue Req ($000)

2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040

$3,241,019
$3,309,662
$3,407,431
$3,531,080
$3,567,006
$3,614,422
$3,662,468
$3,711,153
$3,760,484
$3,810,472
$3,861,124
$3,912,450
$3,964,457
$4,017,157
$4,070,556
$4,124,665
$4,179,494
$4,235,052
$4,291,348
$4,348,392
$4,406,195
$4,464,766

3
Incremental Impact of
Preferred Resource Plan
($000)

4
Incremental
Impact
(%)

-$20,307
-$18,905
-$18,646
-$18,163
-$17,880
-$21,259
-$17,597
-$16,389
-$8,389
-$26,054
-$22,932
$78,174
$41,948
$47,089
$4,568
-$36,862
-$58,945
-$61,023
-$68,746
-$2,809
-$66,728

-0.61%
-0.55%
-0.53%
-0.51%
-0.49%
-0.58%
-0.47%
-0.44%
-0.22%
-0.67%
-0.59%
1.97%
1.04%
1.16%
0.11%
-0.88%
-1.39%
-1.42%
-1.58%
-0.06%
-1.49%

We visually portray this information in Figure 6-4 below.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 133 of 139

Figure 6-4: Incremental Rate Impact of Preferred Plan
State of Minnesota – All Customers
$0.0030
$0.0020

$ per kWh

$0.0010
$0.0000
-$0.0010
-$0.0020
-$0.0030

A.

Methodology and Calculations

We determine class revenue requirement impacts by allocating incremental costs to
rate classes for each year in the planning period (2020-2034). After costs are
allocated, we then calculate revenue requirement impacts for each customer class.
We apply ratemaking treatments to expense items that are impacted by the Resource
Plan, as follows:
 Fuel Costs
 Purchased Energy
 Production O&M Expenses
 Property Taxes
 Deferred Income Taxes
 Tax Depreciation and Removal Expense,
 Decommissioning Accruals
 Plant In Service and Associated Depreciation, Construction Work in Progress
(CWIP), and Accumulated Deferred Income Taxes
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 134 of 139

 Bulk Transmission Costs
We discuss our treatment of these expense items for purposes of this rate impact
analysis below.
B.

Fuel Costs and Purchased Energy

Fuel and purchased energy costs are allocated to classes using the E8760 energy
allocator approved in our most recent Minnesota rate case, as provided below:7
Table 6-2: E8760 Energy Allocator
MN

Residential

100.00%

29.27%

Commercial NonDemand
3.04%

C&I Demand

Lighting

67.24%

0.44%

The E8760 allocator is calculated by taking the forecast hourly load for each of the
8,760 hours of the test year for each customer class, then weighting the hourly load by
the forecasted hourly marginal energy cost in each respective hour.
C.

Production Expense, Property Taxes, Deferred Income Taxes, Tax
Depreciation and Removal Expense and Decommissioning Accrual

These expense items are split into energy-related and capacity/demand-related
components using the Company’s plant stratification analysis approved in our most
recent Minnesota rate case.8 We provide the approved plant stratification analysis that
we applied to production O&M expenses for each plant type below:

See Docket No. E002/GR-15-826, In the Matter of the Application of Northern States Power Company for Authority
to Increase Rates for Electric Service in the State of Minnesota, FINDINGS OF FACT, CONCLUSIONS, AND ORDER,
(June 12, 2017).
8 Id.
7

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 135 of 139

Table 6-3: Stratification Analysis by Plant Type
Plant Type
Combustion Turbine
Fossil
Nuclear
Combined Cycle
Wind
Solar

Replacement
Value $/kW
$825
$2,089
$4,286
$1,079
$15,847
$8,182

Capacity Ratio
$825 / $825
$825 / $2,089
$825 / $4,286
$825 / $1,079
$825 / $15,847
$825 / $8,182

Capacity/Demand
Percentage
100.0%
39.5%
19.3%
76.5%
5.2%
10.1%

Energy
Percentage
0.0%
60.5%
80.7%
23.5%
94.8%
89.9%

The plant stratification approach begins by comparing the replacement cost of each
type of generation plant (fossil, combined cycle, nuclear, etc.) to the replacement cost
of a CT. CT are 100 percent capacity/demand-related since they are the generation
source with the lowest capital cost and the highest operating cost. For each
generation type, the percent of total generation costs that exceeds the cost of a CT
peaking plant are classified as being energy-related. These costs are in excess of the
capacity/demand-related portion, and as such, were not incurred to obtain capacity,
but rather to obtain lower cost energy.
After production O&M costs originating from each type of generation plant are split
into capacity-related and energy-related components based on the percentages shown
in Table 6-3 above, those costs that have been classified as being energy-related are
allocated to class using the E8760 energy allocator provided in part 1 above.
The capital costs that have been classified as being capacity- or demand-related are
allocated to customer class using the D10S capacity allocator approved by the
Commission in our most recent rate case.9 The D10S allocator is simply each class’s
load that is coincident with the NSP system peak load. We provide the approved
D10S class allocator percentages below:
Table 6-4: D10S Capacity Allocator

9

MN

Residential

100.00%

36.14%

Commercial
Non-Demand
3.28%

C&I Demand

Lighting

60.59%

0.00%

Id.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

D.

Resource Plan – Chapter 6
Page 136 of 139

Generation Rate Base Costs Including Plant in Service, Depreciation,
CWIP and Accumulated Deferred Income Taxes

Rate base related costs from each type of generation plant are also split into energyrelated and capacity/demand-related components using the Company’s plant
stratification analysis approved in our most recent Minnesota rate case.10 As was true
with the expense items listed in part 2 above, rate base costs classified as being energyrelated are allocated to class using the E8760 energy allocator. Likewise, the capital
costs that have been classified as being capacity or demand-related are allocated to
customer class using the D10S capacity allocator.
E.

Bulk Transmission Costs

As ordered by the Minnesota Commission, all rate base and expense items related to
bulk transmission are classified as being capacity or demand-related and are allocated
to customer class using the Commission-approved D10S capacity allocator.11
IV.

DETERMINING CLASS RATE IMPACTS

In order to show the estimated impacts of the Preferred Plan on customer rates and
bills, we provide a breakdown by customer class for the 2020-2040 period, and in
more detail for the immediate five-year 2019-2023 period at the Minnesota customer
class levels.
Figure 6-5 below shows the estimated incremental impacts of our Preferred Plan over
the long-term by customer class.

10
11

Id.
Id.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 137 of 139

Figure 6-5: Incremental Rate Impact of Preferred Plan
by Customer Class – State of Minnesota
$0.0040
$0.0030
$0.0020

$ per kWh

$0.0010
$0.0000
-$0.0010
-$0.0020
-$0.0030
-$0.0040

Residential

Commercial Non Demand

C&I Demand Billed

Table 6-5 below provides a more detailed view of near-term estimated rate impacts
for Minnesota customer classes.

July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 138 of 139

Table 6-5: Preferred Plan Estimated Rate Impacts by Class per Year
Comp'd
Rate Class Impacts \1

2019

2020

2021

2022

2023

2024

Incr/Yr

Residential (avg rate, ¢/kWh)
Cumul Increase (¢/kWh)
Cumulative Increase (%)

14.488¢

14.367¢
-0.121
-0.84%

14.506¢
0.018
0.12%

14.847¢
0.359
2.48%

15.377¢
0.889
6.14%

15.526¢
1.037
7.16%

N/A
N/A
1.39%

$ Impact/Month, @

($0.79)

$0.11

$2.33

$5.78

$6.74

N/A

N/A

13.218¢
-0.052
-0.39%

13.167¢
0.293
2.22%

13.511¢
0.727
5.50%

13.946¢
1.380
10.44%

14.599¢
1.636
12.38%

14.855¢
N/A
2.36%

$2.93

$7.27

$13.80

$16.36

N/A

N/A

9.370¢

9.300¢
-0.070
-0.75%

9.707¢
0.336
3.59%

10.040¢
0.669
7.14%

10.471¢
1.100
11.74%

10.570¢
1.199
12.80%

N/A
N/A
2.44%

($26.30)

$126.15

$250.98

$412.56

$449.71

N/A

N/A

25.290¢

25.027¢
-0.262
-1.04%

24.668¢
-0.622
-2.46%

24.917¢
-0.372
-1.47%

25.624¢
0.334
1.32%

26.079¢
0.790
3.12%

N/A
N/A
0.62%

($0.37)

($0.22)

$0.20

$0.47

N/A

N/A

650

Sm Non-Dmd (avg rate, ¢/kWh)
Cumul Increase (¢/kWh)
Cumulative Increase (%)

13.218¢

$ Impact/Month, @

($0.52)

1,000

Demand (avg rate, ¢/kWh)
Cumul Increase (¢/kWh)
Cumulative Increase (%)
$ Impact/Month, @

37,500

Street Ltg (avg rate, ¢/kWh)
Cumul Increase (¢/kWh)
Cumulative Increase (%)
$ Impact/Month, @

60

($0.16)

Using the methodologies described above, the incremental costs in the last year of the
period (2024) for the Preferred Plan would be expected to increase the average
Residential rate by about 1.39 percent on a compounded annual basis through 2024.
That is equivalent to a total increase of $6.74 per month above the current rate level.
The impact to the average Large Demand Billed rate would be an increase of about
2.44 percent on a compounded annual basis through 2023, which is equivalent to an
increase of 1.199 cents per kWh above the 2019 level.
V.

FACTORS IMPACTING NEAR-AND LONG-TERM RATE
ESTIMATES

We note that the following factors could have an impact on the estimated rate impacts
in the planning period:
July 1, 2019

2020-2034 Upper Midwest Resource Plan

 Xcel Energy

Resource Plan – Chapter 6
Page 139 of 139

Depreciation Expense for Coal Closures. The modeling and estimated rate impacts reflect
accelerated depreciation associated with the early retirement of the Allen S. King and
Sherco Unit 3 plants. This is consistent with the Company’s current method of
recovery for Sherco 1 and 2. As noted previously, however, and consistent with the
Commission’s actions in the approval of the early shutdown of the Benson biomass
plant, a regulatory asset is another tool that could be used to accompany these early
retirements. An alternative regulatory treatment such as this would impact this
analysis.
Generation Ownership: Owned and Purchased Power Agreement resources will have
different cost patterns, which will impact this analysis to the extent a resource addition
differs in terms of ownership from what was modeled.
Taxes. This analysis is based on present tax conditions. Any tax changes will impact
the modeling underlying this analysis and thus the rate impact results.
Nuclear Decommissioning Trust. There are several items regarding the NDT that may
have a material impact on costs included in this resource plan.
Pending or Future Regulatory Decisions. Rate case and resource acquisition outcomes have
the potential to impact rates and system needs.
Large Customer Changes. The loss or addition of a large business customer has the
potential to impact both rates and system needs.

July 1, 2019

2020-2034 Upper Midwest Resource Plan