Technical ReportPDF Available

Solar Potential Analysis : High Penetration Renewables in Minnesota

Authors:
  • Clean Power Research

Abstract and Figures

The Minnesota Solar Pathways (Pathways) initiative, sponsored by the U.S. Department of Energy Solar Energy Technologies Office, is a three-year project designed to explore least-risk, best-value strategies for meeting the State of Minnesota’s solar goals. As part of this aim, the Pathways Team is modeling renewable generation costs, examining ways to streamline interconnection, and evaluating technologies that can increase solar hosting capacity on the distribution grid. For more details about the MN Solar Pathways project, including published reports and a list of project partners, please visit mnsolarpathways.org. This report summarizes the modeling of future renewable generation costs as accomplished by the Solar Potential Analysis (SPA).
Content may be subject to copyright.
Solar Potential Analysis Report | November 14, 2018 | Page
Solar Potential Analysis
Report
NOVEMBER 15, 2018
PREPARED FOR
MINNESOTA DEPARTMENT OF COMMERCE
AND THE MINNESOTA SOLAR PATHWAYS PROJECT
PREPARED BY
CLEAN POWER RESEARCH
MORGAN PUTNAM
MARC PEREZ
Solar Potential Analysis Report | November 15, 2018 | Page i
Support
This material is based on work supported by the Minnesota Department of Commerce,
State Energy Office and made possible by a grant from the U.S. Department of Energy,
Office of Energy Efficiency and Renewable Energy (EERE) Solar Energy Technologies
Office, under Award Number DE-EE0007669. The U.S. Department of Energy Solar
Energy Technologies Office supports early-stage research and development to improve
the flexibility and performance of solar technologies that contribute to a reliable and
resilient U.S. electric grid. The office invests in innovative research efforts that securely
integrate more solar energy into the grid, enhance the use and storage of solar energy,
and lower solar electricity costs. Learn more at https://energy.gov/solar-office.
Stacy Miller, Project Manager
(651) 539-1886
Clean.Energy@state.mn.us
Disclaimer
This report was prepared as an account of work sponsored by an agency of the United States
Government. Neither the United States Government nor any agency thereof, nor any of their
employees, makes any warranty, express or implied, or assumes any legal liability or
responsibility for the accuracy, completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use would not infringe privately owned
rights. Reference herein to any specific commercial product, process, or service by trade
name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its
endorsement, recommendation, or favoring by the United States Government or any agency
thereof. The views and opinions of authors expressed herein do not necessarily state or
reflect those of the United States Government or any agency thereof.
Solar Potential Analysis Report | November 15, 2018 | Page ii
Acknowledgements
Core Team (alphabetic by organization)
In addition to Clean Power Research, the following partner organizations are the project leads
under Minnesota Solar Pathways: Illuminating pathways to 10% solar. These organizations
contributed to the Solar Potential Analysis development and final report as described below.
Center for Energy and Environment (CEE)CEE contributed to the technical analysis in the SPA
(through development of the load shifting analysis) and to the generation of the SPA Final
Report. In particular, CEE provided multiple cover-to-cover reviews that greatly enhanced the
clarity and quality of the report. Josh Quinnell served as CEE’s project lead and his diligent and
detailed work greatly improved the overall quality of this work.
Clean Energy Resource Teams (CERTs)The CERTs team served as a key project catalyst
(making connections, providing presentation space and IT resources, developing a website and
handling outreach efforts) and developed a number of key figures in the report. Notably, these
figures would not be nearly as clear or visually pleasing without Dan Thiede’s talented eye for
communication and design.
Great Plains Institute (GPI)GPI facilitated the Pathways’ Technical Committee meetings and
ran stakeholder engagement. In particular, Brian Ross’s years of stakeholder experience were
quite helpful for producing thoughtful discussion among the Technical Committee members.
MN Department of Commerce (Commerce)Commerce served as the MN Solar Pathways
project manager and a member of the Technical Committee. Stacy Miller served in both of
these roles for Commerce and provided extensive feedback on the SPA, including multiple
cover-to-cover reviews of the SPA Final Report. Clean Power Research and the Core Team are
greatly indebted for Stacy’s contributions to the project.
Solar Potential Analysis Report | November 15, 2018 | Page iii
MN Solar Pathways Technical Committee (alphabetic by organization):
The SPA benefited tremendously from feedback received during meetings of the MN Solar
Pathways Technical Committee. Members of the Technical Committee are listed below.
Organization
Committee Member
City of Saint Paul
Jim Giebel
City of Duluth
Erik Birkeland
Clean Grid Alliance
Beth Soholt
Energy Systems Integration Group
Mark Ahlstrom
Fresh Energy
Allen Gleckner
Hennepin County
Leah Hiniker
Innovative Power Systems
Ralph Jacobson
Lake Region Electric Cooperative
Lloyd Nelson
Metropolitan Council
Cameran Bailey
Midcontinent Independent System Operator
Brandon Heath
Minnesota Citizens Utility Board
Annie Levinson Falk
Minnesota Department of Commerce
Stacy Miller
Minnesota Power
Jennifer Peterson
Minnesota Solar Energy Industries Association
David Shaffer
National Renewable Energy Laboratory
Bethany Frew
Otter Tail Power Company
Nate Jensen
Renewable Energy Systems Americas
Matt Boys
Rochester Public Utilities
Dru Larson
Target
Holly Lahd
Xcel Energy
Patrick Dalton
Individual Contributors
The Core Team received support from a number of individuals (many of whom are listed above)
during the development of the SPA and the writing of the report. These individuals took time to
meet one-on-one, discuss modeling issues, and provide feedback on the final report that
greatly helped to clarify technical concepts.
Solar Potential Analysis Report | November 15, 2018 | Page iv
Table of Contents
Support............................................................................................................................................. i
Disclaimer ......................................................................................................................................... i
Acknowledgements ..........................................................................................................................ii
Table of Contents ............................................................................................................................ iv
List of Figures ................................................................................................................................... v
List of Tables .................................................................................................................................. vii
Executive Summary ......................................................................................................................... 1
SPA Terminology ............................................................................................................................. 6
MN Solar Pathways Overview ......................................................................................................... 8
Solar Potential Analysis (SPA) ....................................................................................................... 11
SPA Data Inputs ............................................................................................................................. 16
SPA Scenarios ................................................................................................................................ 21
SPA Results .................................................................................................................................... 27
Discussion of SPA Results .............................................................................................................. 38
Conclusion ..................................................................................................................................... 43
Appendix A: Production Requirements ........................................................................................ 44
Appendix B: Electrification and Load Shifting Models .................................................................. 49
Appendix C: Electrification and Load Shifting Results .................................................................. 67
Appendix D: Spatial Allocation of Solar ........................................................................................ 76
Appendix E: Additional SPA Datasets ........................................................................................... 79
Appendix F: Land Use .................................................................................................................... 82
Appendix G: Cost of Capital .......................................................................................................... 85
Appendix H: Scalability of the Hourly Results ............................................................................... 86
Appendix I: Cost of Natural Gas Generation Resources ............................................................... 87
Appendix J: Benefits of Additional Capacity ................................................................................. 90
Solar Potential Analysis Report | November 15, 2018 | Page v
List of Figures
Figure 1. MN Solar Pathways Core Team and Technical Committee ............................................. 9
Figure 2. Organization of the MN Solar Pathways Project ........................................................... 10
Figure 3. Timeline and Scope of MN Solar Pathways Technical Analyses .................................... 10
Figure 4. Overview of the Solar Potential Analysis ....................................................................... 12
Figure 5. Service Territories Included in Creation of Minnesota Load Data ................................. 17
Figure 6. Construction of SPA Scenarios from Choice of Production Requirements, level of
Technology Development, and Solar Distribution ........................................................................ 21
Figure 7. Dispatchability of SPA Production Requirements .......................................................... 22
Figure 8. Illustration of Solar, Wind, and Storage Production Requirements .............................. 23
Figure 9. Spatial allocation of solar for the Utility-Led and All Sectors scenarios ........................ 26
Figure 10. Influence of Additional Capacity coupled with Energy Curtailment on Generation Cost
and Resource Deployment ............................................................................................................ 31
Figure 11. Storage State of Charge (GWh) minimum state of charge plotted for each day in a
calendar year ................................................................................................................................ 32
Figure 12. Effect of Utilizing Other Generation Resources during Periods of Low Renewables
Production ..................................................................................................................................... 33
Figure 13. EV Load in 2050 with L1 and L2 chargers.................................................................... 36
Figure 14. Area Required for Solar Deployment Compared with Existing Land Use in MN ......... 41
Figure 15. Carbon Intensity of Minnesota’s Electric Sector ......................................................... 42
Figure 16. Unconstrained production profile plotted on an annual basis and a weekly basis .... 44
Figure 17. Predictable Production Profile plotted on an annual basis and a weekly basis .......... 45
Figure 18. Seasonal-Diurnal Production Profile plotted on an annual basis and a weekly basis . 46
Figure 19. Seasonal Production Profile plotted on an annual basis and a weekly basis .............. 47
Figure 20. Hourly Production Profile plotted on an annual basis and a weekly basis................. 47
Figure 21: Forecasted market penetration of controlled DHW units ........................................... 52
Figure 22: Forecasted market penetration of controlled EV units ............................................... 52
Figure 23: Forecasted market penetration of controlled residential heating units ..................... 53
Figure 24: Example scenario of working range of DHW load shifting .......................................... 55
Figure 25. Driving Behavior for Agent 20004480:2 in the NHTS .................................................. 58
Figure 26. EV Battery State of Charge Associated with Agent 20004480:2 ................................. 59
Figure 27. EV Battery Charge/Discharge Profile Associated with Agent 20004480:2 .................. 59
Figure 28. 2025 EV Load Impact for the Low Technology Development scenarios ..................... 60
Figure 29. EV Battery Charge/Discharge Profile with Load Shifting (Agent 20004480:2) ............ 60
Figure 30. EV Battery State of Charge with Load Shifting (Agent 20004480:2) ........................... 61
Figure 31. Aggregate EV Load Impact with and without Load Shifting ........................................ 61
Figure 32. Demonstration of HVAC load shifting capabilities....................................................... 65
Figure 33. Plot of the indoor temperature against time as HVAC load is shifted ........................ 66
Solar Potential Analysis Report | November 15, 2018 | Page vi
Figure 35. EV Load in 2025 with L1 and L2 chargers .................................................................... 68
Figure 36. EV Load in 2050 with L1 and L2 chargers .................................................................... 69
Figure 39. Residential Heating Load in 2050 during a Cold Winter Week.................................... 71
Figure 40. Daily Residential Heating Load in 2050 ....................................................................... 72
Figure 41. DHW load shifting in 2050 High Technology Development scenario .......................... 73
Figure 42. EV load shifting in 2050 High Technology Development scenario with L2 chargers... 74
Figure 43. EV load shifting in 2050 High Technology Development scenario with L1 chargers... 74
Figure 44. Illustration of the calculation process implied by Equation 1 ..................................... 76
Figure 45. Illustration of the creation of the Non-Deployment Zone Filter ................................. 77
Figure 46. Example Utility and Non-Utility Solar Allocations ....................................................... 78
Figure 47. Solar to Wind Optimization Curve ............................................................................... 80
Figure 48. Aggregation of hourly dispatch profiles for a week in January ................................... 81
Figure 49: Horizontal Irradiance across the state of Minnesota in kWh/m2/yr .......................... 82
Figure 50: Existing Land Use in Minnesota ................................................................................... 83
Figure 51: Comparison of Required Area for PV with Existing Land Use in Minnesota ............... 84
Figure 52. Daily Load and Solar Production in GWh ..................................................................... 90
Figure 53. Impact of Capacity Overbuilding on Required Storage Capacity ................................. 91
Solar Potential Analysis Report | November 15, 2018 | Page vii
List of Tables
Table 1. Must-Run Resources in the 2025 Timeframe ................................................................. 18
Table 2. Cost Forecasting Sources ................................................................................................ 19
Table 3. SPA Technology Costs for the 2025 and 2050 Timeframes ............................................ 23
Table 4. SPA Technology Adoption for the 2025 and 2050 Timeframes ...................................... 24
Table 5. Solar Capacity by Type of Solar for the Solar Distribution Scenarios ............................. 25
Table 6. Key SPA Results in the 2025 Timeframe ......................................................................... 27
Table 7. 2050 SPA Results without Other Generation Resources ................................................ 29
Table 8. 2050 SPA Results with 10% Other Generation Resources .............................................. 34
Table 9: Specific loads considered for electrification and load shifting and their fraction of
Minnesota’s total load (current and 2050 forecast assuming significant load electrification) .... 49
Table 10: Number of participating units and their aggregate load for the SPA scenarios ........... 53
Table 11. Temperature Dead-bands for HVAC modeling ............................................................. 64
Table 12. SPA Results for the Hourly Production Requirements with No Other Generation
Resources ...................................................................................................................................... 85
Table 13. SPA Results for the Hourly Production Requirements with 10% Other Generation
Resources ...................................................................................................................................... 85
Solar Potential Analysis Report | November 15, 2018 | Page 1
Executive Summary
The Minnesota Solar Pathways (Pathways) initiative, sponsored by the U.S. Department of
Energy Solar Energy Technologies Office, is a three-year project designed to explore least-risk,
best-value strategies for meeting the State of Minnesota’s solar goals. As part of this aim, the
Pathways Team is modeling renewable generation costs, examining ways to streamline
interconnection, and evaluating technologies that can increase solar hosting capacity on the
distribution grid. For more details about the MN Solar Pathways project, including published
reports and a list of project partners, please visit mnsolarpathways.org.
This report summarizes the modeling of future renewable generation costs as accomplished by
the Solar Potential Analysis (SPA).
What is the Solar Potential Analysis?
The SPA is a modeling tool that estimates and optimizes
the generation cost and resource capacities (e.g., solar
capacity) to serve a specified percentage of Minnesota’s
electrical load with given production requirements (e.g.,
production that is aligned with a day-ahead forecast).
One purpose of the SPA is to provide key insights into
transforming solar and wind generation into dispatchable
generation resources. The purpose of the SPA is not to
make decisions regarding specific generation resources:
the SPA is not a resource plan.
The Pathways Team used the SPA to model the
generation cost (in $/MWh) to achieve 10% of
Minnesota’s electricity from solar by 2025 and 70% of
Minnesota’s electricity from solar and wind by 2050.
The Pathways Team modeled a number of scenarios to
identify the likely range of generation costs and resource
capacities for the 10% and 70% targets of interest. These
scenarios included different: 1) future technology costs; 2) solar distributions (spatial allocation
and type of installation); and 3) production requirements.
A number of important findings from the SPA are discussed in the body of the report. Key
findings related to achieving 10% of Minnesota’s electricity from solar by 2025 and 70% of
Minnesota’s electricity from solar and wind by 2050 are presented in the Executive Summary.
Solar Potential Analysis Report | November 15, 2018 | Page 2
10% Solar by 2025
The SPA results indicate that Minnesota could achieve its goal of 10% solar at costs
comparable to the cost of natural gas generation.1
Modelled generation cost for 10% solar by 2025 ranged from $33/MWh to $66/MWh.2 The
broad range is a result of cost forecasts and production requirements that were meant to
bound the likely futures in
Minnesota. The lower-end of the
generation cost range is comparable
to the variable cost of natural gas
generation and the upper-end of the
range is comparable to the levelized
cost of new natural gas generation
(as presented in Appendix I: Cost of
Natural Gas Generation Resources).
The ranges of solar and storage capacity reflect the different solar production requirements
modeled. Scenarios with minimal production requirements (e.g., solar could produce as long as
there was load to serve after accounting for wind and must-run resources) required 5 GW of
solar capacity and no storage capacity. Scenarios with modest production requirements (e.g.,
solar was expected to match the forecasted day-ahead production) required 6 GW of solar
capacity and up to 2 GWh of storage.
70% Solar and Wind by 2050
The SPA results reveal that the expected cost declines of solar, wind, and storage will enable
Minnesota to achieve 70% solar and wind by 2050 with generation costs comparable to
natural gas generation costs.
This is a particularly notable result given the conservative production requirements used when
modeling the 2050 results: Minnesota-sited solar, wind, and storage were asked to match
Minnesota’s hourly load profile. An exception to this requirement was made during brief
periods of low-solar and low-wind production, during which time Other Generation resources
were used to support generation requirements.
1 The Pathways Team uses natural gas as an accepted benchmark of cost comparison in common use, in part
because natural gas is the leading traditional resource being developed in the U.S. The comparison does not imply
that the SPA includes a full analysis of renewables vs. gas in terms of performance or adequacy.
2 Costs are in current (nominal) dollars.
Solar Potential Analysis Report | November 15, 2018 | Page 3
SPA Key Findings
1. Solar and wind can serve 70% of Minnesota’s electrical load in 2050.
o Solar and wind can serve 70% of Minnesota load at generation costs that are
comparable to the levelized generation cost of new natural gas generation.
2. Additional Capacity coupled with energy curtailment is considerably less expensive
than, and a viable alternative to, long-term or seasonal storage in a high renewables
future.
o Declining costs of solar and wind generation (<$20/MWh) will enable solar and
wind to be economically curtailed during periods of high production and low
load.
o The ability to curtail surplus renewable production removes the need to use
energy storage to seasonally shift renewable energy production to serve load.
3. Flexible Other Generation resources used in limited amounts support a high
renewables future.
o The strategic use of Other Generation resources during brief periods of low-solar
and low-wind production significantly reduced the storage, solar, and wind
capacities used to serve Minnesota’s hourly load. As a result, the generation cost
for 70% solar and wind was reduced by nearly half.
4. Storage is an important part of a high renewables future; it expands the dispatch
capabilities of wind and solar assets.
o Sufficient quantities of storage smooth out the intra-hour variability of solar and
wind.
5. Shifting of key flexible loads may further decrease generation costs.
o Load shifting of new electric vehicle and residential domestic hot water loads
demonstrated a potential 10-20% decrease in generation costs. While these
figures are promising, results demonstrate further study is necessary.
Solar Potential Analysis Report | November 15, 2018 | Page 4
Scope of the SPA
The electrical grid can be described in terms of its Services, Markets, and Regulations:
1) Services The technical services that allow the grid to operate in a stable manner.
Services include: energy, capacity, balancing, frequency, and voltage stability.
2) Markets The markets that address the procurement and compensation of services.
3) RegulationsRegulations that define how resources can or cannot participate in
markets, including operator protocols.
Markets and Regulations are implementation and policy instruments for the delivery of energy.
The SPA is a technical analysis focused on Services. Within Services, the SPA addresses energy
and capacity at the hourly level. The SPA did not address balancing, frequency, or voltage
stability. Minnesota decision-makers need to consider Services, Markets, and Regulations when
evaluating possible energy futures. As such, the SPA results are one part of a larger picture.
Constraints of the SPA
The SPA did not include integration with the Midcontinent Independent System Operator
(MISO) market. Not modeling integration with MISO allowed the SPA to estimate generation
cost without speculating about the flexibility the MISO market could provide. This meant that
the model did not consider opportunities to export excess solar and wind production to MISO,
or to import excess solar and wind production from MISO into Minnesota. There were
members of the Minnesota Solar Pathways Technical Committee who felt that the lack of
integration with MISO led to a conservative analysis with higher generation costs and resource
capacities. This belief was borne out by reduced generation costs for SPA scenarios that
included Other Generation resources.3
The SPA did not include transmission or distribution costs. Studies evaluating the impact of
increasing solar and wind on transmission costs have been done and are on-going. Studies
evaluating the impact of solar and wind on distribution costs in Minnesota are needed. With
regard to transmission costs, two key studies worth noting are Minnesota Renewable Energy
Integration and Transmission Study (MRITS)4 and MISO’s Renewable Integration Impact
Assessment (on-going). The MRITS study assumed integration with MISO and found thatthe
addition of wind and solar (variable renewable) generation to supply 40% of Minnesota’s annual
electric retail sales can be reliably accommodated by the electric power system” and that
3 Allowing Other Generation resources to ramp-up during periods of low solar and wind production is similar to
using the surrounding MISO region as a generation resource.
4 Minnesota Renewable Energy Integration and Transmission Study. http://mn.gov/commerce-stat/pdfs/mrits-
report-2014.pdf Oct 31, 2014.
Solar Potential Analysis Report | November 15, 2018 | Page 5
“further analysis would be needed to ensure system reliability at 50% of Minnesota’s annual
electric retail sales from variable renewables.”
The SPA did not include a resource adequacy model. Existing resource adequacy models in
common use today do not holistically examine solar, wind, and storage resources. The strategic
combination of solar, wind, and storage resources to provide generation capacity is currently
not considered. Improved resource adequacy models are needed to fairly evaluate the resource
adequacy of combinations of solar, wind and storage resources.
The SPA did not address rate structures. SPA generation costs are derived from installation and
operating costs and do not consider market payments and/or customer rate structures. The
next model being developed under the Pathways initiative, the Solar Deployment Strategy, will
evaluate the effect of rate structures and their impacts on value propositions of different solar
deployment scenarios.
Use of SPA Results
At the highest-level the SPA results indicate that solar, wind, and storage resources can
reasonably and cost-effectively serve a majority of Minnesota’s load. As such, the SPA results
can change the way solar, wind, and storage are evaluated as part of Minnesota’s energy future
a process that many energy stakeholders in Minnesota have already identified as a need.
The SPA results provide important insights into the solar, wind, and storage capacities required
to achieve a future where solar and wind served 70% of Minnesota’s annual load. Under a 70%
scenario, the SPA results suggest that Minnesota could expect to have tens of GW of solar and
wind with just tens of GWh of storage capacity; not 1000’s of GWh of storage capacity. In this
way, the SPA results can be used along with other studies as decision-makers and stakeholders
anticipate and plan for expanded solar, wind, and storage capacities.
The SPA results also highlight the future need for discussion about solar and wind
compensation policies that account for Additional Capacity coupled with energy curtailment. To
this end, the SPA results can be used to explore this important issue before it is pressing. Note
that the Pathways project is agnostic on the numerous potential solutions, but rather raises this
as a point of discussion. Early discourse will prove valuable since any new compensation
policies will ultimately need to move through a regulatory process.
Finally, the SPA results provide some insight into the effect of different solar distributions on
Minnesota’s possible energy futures. However, a deeper analysis is needed. The second phase
of the Minnesota Solar Pathways project seeks to undertake such an analysis through the
development of the Solar Deployment Strategies modeling tool, which will consider the value
propositions of different solar deployment scenarios in addition to costs.
Solar Potential Analysis Report | November 15, 2018 | Page 6
SPA Terminology
The Minnesota Solar Pathways project is focused on bringing together a diverse set of
stakeholders (cities, corporations, non-profits, consumer representatives, solar installers, and
electric utilities). And while that diversity of opinion is one of the project’s greatest strengths, it
also creates a challenge when we use different words to mean the same thing or the same
words to mean different things as has happened on occasion during the Minnesota Solar
Pathways project.
Throughout this report we have sought to use clear and consistent language. Additionally, we
define key terms or concepts in call-out boxes. A few of these are below. They are provided not
only as an example, but as key terms which have already been used in the Executive Summary.
Dispatchable generation generally refers to the ability of a generation resource to flexibly
respond to match the load shape in real time. While dispatch is a technical term with specific
meaning for different groups, in practice all generation exhibits both some degree of
dispatchability and limits on that dispatchability. For example subject to some limitations,
utility-scale solar and wind generation resources are already dispatching into energy markets
around the country.
The aim with the use of the term dispatch is to highlight the increased flexibility and
expanded dispatch capabilities of the solar and wind resources when paired with the
strategies implemented in the SPA.
energy needs but still beneficial for improving the economic dispatchability of the solar/wind
portion of the Hourly Production Requirements during brief periods of low-solar and wind
resources.
If the SPA had included integration with MISO, solar and wind resources outside of
Minnesota would have also been considered ‘other generation’ resources
Generation Cost includes the installation and operational costs of solar, wind, and storage
resources plus the operational costs of Other Generation resources.
Solar Potential Analysis Report | November 15, 2018 | Page 7
What is Additional Capacity?
The Solar Potential Analysis (SPA) uses the concept of Additional Capacity to describe the
capacity needed to cost effectively maximize the dispatchability of solar and wind energy. The
“addition” is measured from the amount of capacity needed to meet annual solar production
targets (10% by 2025 or 70% solar/wind by 2050). Additional solar and wind is capacity
designed to ensure sufficient generation when solar or wind resources are low - such as cloudy
days or near sunrise or sunset or calm days.
Peaks and Valleys. All generating capacity that serves a peaking or balancing function, such as
a “peaking plant,” may be operated intermittently, even only a few times over a year. Peaking
plants are “additional” generation capacity that serves a specific peaking function. Even though
these facilities are idle for long periods of time, we do not consider them as “overbuilt,” nor do
we say that we are “curtailing” the output of the power plant when the plants are not running.
Peaking plant capacity has a critical grid function that enables the most cost effective
deployment of other resources.
Additional solar or wind capacity similarly serves a critical function on the grid; producing
energy at the needed time, as does a peaking plant. But instead of filling the need for more
production when demand peaks, Additional Capacity fills the need for more production when
solar and wind resources are low, effectively filling a valley.
Energy storage could be used in place of Additional Capacity, but the Additional Capacity is a
cheaper solution than building more energy storage to perform the same peak/valley grid
function.
Curtailment. Additional Capacity means additional energy and most likely curtailment. During
periods of high solar and wind production, energy production greater than load will exist.
Energy not utilized will be curtailed. However, it will be economical to curtail this excess energy
due to the low cost of solar and wind production.
Solar Potential Analysis Report | November 15, 2018 | Page 8
MN Solar Pathways Overview
Minnesota is a longstanding, nationally recognized leader in energy efficiency and wind
development. In recent years, Minnesota has established leadership in solar deployment as
well, including hosting the most community solar capacity in the country5 and a 1.5% solar
energy standard.6 The State also adopted a goal of meeting 10 percent of the state’s electricity
needs with solar by 2030.7
The Minnesota Solar Pathways (Pathways) initiative, sponsored by the U.S. Department of
Energy Solar Energy Technologies Office8, is a three-year project designed to explore least-risk,
best-value strategies for meeting the State of Minnesota’s solar goals. As part of this aim, the
Pathways Team is modelling renewable generation costs, examining ways to streamline
interconnection, and evaluating technologies that can increase solar hosting capacity on the
distribution grid.
The Pathways Team is comprised of a Core Team and a Technical Committee. The Core Team
consists of MN Department of Commerce (Commerce), Center for Energy and Environment
(CEE), Clean Energy Resource Teams (CERTS), Clean Power Research (CPR), and the Great Plains
Institute (GPI). The Technical Committee is the foundation for the project’s stakeholder
collaboration process and is comprised of the 22 organizations. These organizations include
cities, corporations, non-profits, consumer representatives, solar installers, and utilities. See
Figure 1 for a list of organizations involved.
Responsibilities of the Core Team and the Technical Committee
To accomplish the Pathways goals, each of the five organizations making up the Core Team
takes a lead role in various aspects of the project.
Commerce is the project manager and fiscal agent responsible for reporting to the U.S.
Department of Energy.
The Great Plains Institute is the lead facilitator for the Technical Committee and other
stakeholder work.
Clean Power Research is responsible for the development of two models and leads all
technical work with input from the Technical Committee.
5 Morehouse, Catherine. Minnesota community solar hits 400 MW. Utility Dive.
https://www.utilitydive.com/news/minnesota-community-solar-hits-400-mw/531305/. Aug 2018.
6 Minn. Stat. § 216B.691.
7 ibid.
8 The U.S. Department of Energy Solar Energy Technologies Office supports early-stage research and development
to improve the flexibility and performance of solar technologies that contribute to a reliable and resilient U.S.
electric grid. Learn more at energy.gov/solar-office.
Solar Potential Analysis Report | November 15, 2018 | Page 9
Center for Energy and Environment is the lead on quality control and supports Clean
Power Research with data needs.
Clean Energy Resource Teams is the lead partner responsible for communications
including dissemination of project results and outreach.
Technical Committee members received and agreed to numerous conditions for participation,
including meeting bi-monthly throughout the project to inform technical decisions that form
the basis of the modeling. Members work collaboratively to make recommendations regarding
inputs and variables to strengthen project results. The Technical Committee was instrumental
in defining the scenarios and informing the analysis described in this report.
Figure 1. MN Solar Pathways Core Team and Technical Committee.
Although the Technical Committee often reached agreement on key project inputs and
recommendations, consensus was not a primary goal as modelling allowed for multiple
scenarios to be run and compared.
The process for taking input and developing the SPA model was iterative as the Core Team
completed work with input from the Technical Committee and reported back. See Figure 2 for
the various roles and structure for completing technical work under Pathways.
Solar Potential Analysis Report | November 15, 2018 | Page 10
Figure 2. Organization of the MN Solar Pathways Project.
Technical Analyses and Reports
In addition to this report, the MN Solar Pathways includes three other technical analyses that
address: interconnection; hosting capacity; and solar deployment strategies. The timeline and
scope of these technical analyses are described in Figure 3. Assessing Opportunities and
Challenges for Streamlining Interconnection Processes9 was published in December 2017, and
the Enhanced Hosting Capacity10 report was published in October 2018. For Pathways reports
and more information on these studies please visit: mnsolarpathways.org.
Figure 3. Timeline
and Scope of MN
Solar Pathways
Technical Analyses.
9 Electric Power Research Institute. Assessing Opportunities and Challenges for Streamlining Interconnection
Processes. http://mnsolarpathways.org December 2017.
10 Smarter Grid Solutions. Enhanced Hosting Capacity Analysis. http://mnsolarpathways.org. October 2018.
Solar Potential Analysis Report | November 15, 2018 | Page 11
Solar Potential Analysis (SPA)
The SPA is a modeling tool that estimates the generation costs and resource capacities (e.g.,
solar and storage capacity) to serve a specified percentage of Minnesota’s electrical load with
given production requirements (e.g., production that matches the day-ahead forecast).
Goal of the SPA
The goal of the SPA is to ask and answer questions related to the deployment of increasing
amounts of solar (and wind) energy. Such questions include:
What range of generation costs might we expect to serve a percentage of Minnesota’s
load with solar? With solar and wind?
What resource capacities (solar, wind, and storage) would be required under various
deployment scenarios?
Can the strategic combination of solar, wind and energy storage provide load-following
generation?
How does electrification of transportation and heating impact generation costs?
Can load shifting reduce generation costs?
By answering these questions the SPA provides key insights for transforming solar and wind
generation into dispatchable generation resources that can ultimately be relied on to serve load
for nearly every hour of the year.
How the Pathways Team Used the SPA
The Pathways Team used the SPA to model the generation cost (in
$/MWh) to achieve 10% of Minnesota’s electricity from solar by
2025 and 70% of Minnesota’s electricity from solar and wind by
2050.
The Pathways Team modeled a number of different scenarios using
the SPA to identify a range of generation costs and resource
capacities for the 10% and 70% targets of interest.
In particular, the Pathways Team used the SPA to evaluate scenarios
with different: 1) future technology costs; 2) solar distributions; 3)
production requirements; and 4) levels of electrification and load
shifting (not present in all scenarios). Notably, future technology
costs and production requirement inputs significantly affected
modeled generation costs and resource capacities.
Solar Potential Analysis Report | November 15, 2018 | Page 12
How the SPA Operates
SPA Analysis. The SPA is an optimization tool that contains two key components: 1) an hourly
energy balance containing resource dispatch algorithms and 2) an economic engine to calculate
the system-wide generation cost of a given amount of solar, wind, and storage capacity.
The SPA operates by first finding sets of resource capacities that will satisfy the specified
production requirements on an hourly basis. For example, the SPA may find that a specified
production requirement can be satisfied by 1 GW of solar, 1 GW of wind, and 3 GWh of storage
and also by 2 GW of solar, 2 GW of wind, and 1 GWh of storage.
Having found sets of resource capacities that satisfy the specified production requirements, the
SPA’s economic engine then evaluates the generation cost of the different sets of resource
capacities.
As a final step, the SPA then searches through the sets of resource capacities to find the set of
resource capacities with the lowest generation cost.
The detailed operation of the SPA involves iteratively solving the hourly energy balance to
produce 100 to 1000’s of sets of resource capacities that satisfy the specified production
requirements.
Figure 4. Overview of the Solar Potential Analysis.
Solar Potential Analysis Report | November 15, 2018 | Page 13
SPA Inputs. The SPA uses three different types of inputs: hourly data; cost forecasts; and
electrification and load shifting data (not included in all analyses and not shown in the graphic).
Hourly data and cost forecasts are described in detail in SPA Data Inputs. Electrification and
load shifting data are described in detail in Appendix B: Electrification and Load Shifting Models.
SPA Scenarios. As previously noted, the Pathways Team evaluated a number of scenarios with
the SPA. These scenarios were comprised of different assumptions about 1) future technology
costs, 2) solar distributions, 3) production requirements, and 4) levels of electrification and load
shifting (not present in all scenarios). Further description of the different SPA scenarios is
provided in SPA Scenarios.
SPA Outputs. The key SPA outputs were generation cost (on a levelized cost of energy basis)
and resource capacity (solar, storage, and wind). Another key SPA metric was the amount of
energy curtailment. Additional SPA outputs, including optimization curves, hourly dispatch
profiles, and ramp rate distributions, are discussed in Appendix E: Additional SPA Datasets.
Electrification and Load shifting in the SPA
The SPA was capable of evaluating the impact on generation cost and resource capacities from
the potential electrification of energy end-uses not currently power by electricity. The SPA was
additionally capable of evaluating the potential for load shifting to reduce generation cost and
resource capacities.
The Pathways Team chose to evaluate
the potential electrification and load
shifting of three such energy end-uses:
Electric Vehicles (EVs), Domestic Hot
Water (DHW) and HVAC (Heating,
Ventilation and Air Conditioning).
Each of these technologies added a
unique electric load profile to the
existing Minnesota load, as detailed in Appendix B: Electrification and Load Shifting Models.
Furthermore, each of these technologies had its own load shifting capabilities as defined by
end-user needs, its technical specifications, and (in the case of HVAC) the weather.
an entity outlaying capital to build generation resources would require to break-even if
Solar Potential Analysis Report | November 15, 2018 | Page 14
SPA Scope
The SPA scope is focused on estimating generation costs of increased levels of solar and is
intended to complement prior studies. The SPA assumptions were developed with stakeholder
input from the Technical Committee and refined by the Core Team. There are a few notable
aspects of the SPA scope as discussed below:
The SPA is designed to provide insights, not decisions. The SPA is not a resource plan.
o The SPA does not perform an economic dispatch analysis to determine the
specific assets that should be built or the economic returns that specific assets
could expect. Rather, the SPA finds sets of resources that meet the specified
production needs and then calculates the generation cost of the entire set of
resources, including solar and storage capacities in 2025 and solar, wind, and
storage capacities in 2050.
The SPA only considered solar and wind generation within Minnesota11 and did not
include integration with the MISO market.
o As noted earlier, excluding integration with MISO allowed the SPA to estimate
generation cost without speculating about the flexibility the MISO market could
provide. This meant that the model did not consider opportunities to export
excess solar and wind production to MISO, or to import excess solar and wind
production from MISO into Minnesota. Some members of the Minnesota Solar
Pathways Technical Committee felt that the lack of integration with MISO led to
a conservative analysis with higher generation costs and resource capacities. This
belief was borne out by reduced generation costs for the SPA scenarios that
included Other Generation resources.
The SPA only considers generation costs. It does not consider transmission and
distribution costs.
o The SPA does not perform a power-flow analysis to examine whether the
existing transmission and distribution infrastructure is sufficient to handle
increasing penetrations of solar and wind.
o It is important for Minnesota energy stakeholders to evaluate generation,
transmission, and distribution costs when evaluating possible energy futures As
such, the SPA results are one part of a larger picture. Studies evaluating the
impact of increasing solar and wind on transmission and distribution costs have
been done and are on-going. With regard to transmission, two studies worth
11 Existing wind generation in North Dakota and South Dakota was considered if it was owned or contracted by
Minnesota utilities.
Solar Potential Analysis Report | November 15, 2018 | Page 15
noting are Minnesota Renewable Energy Integration and Transmission Study12
and MISO’s Renewable Integration Impact Assessment13.
The SPA calculates generation costs based on installation and operational costs, not
market compensation mechanisms (e.g., rate structures).
o The installation and operation costs of solar, wind, and storage assets are
considered.
o The operational costs of Other Generation resources are considered (including
fuel). The installation costs of Other Generation resources are not considered. It
is assumed that Other Generation resources are already built.
o Installation and operation costs of the load shifting resources (EVs, hot water
heaters, smart thermostats) are not considered in the load shifting analysis, as
discussed later.
The SPA accounted for Minnesota’s must-run resources (discussed in the next section).
12 GE Energy Consulting. Minnesota Renewable Energy Integration and Transmission Study.
http://mn.gov/commerce-stat/pdfs/mrits-report-2014.pdf Oct 2014.
13 An on-going study to assess possible renewable energy-driven impacts on the reliability of the electric system.
Solar Potential Analysis Report | November 15, 2018 | Page 16
SPA Data Inputs
As noted above, the SPA contained two main types of data inputs: hourly data and cost
forecasts. Hourly data consisted of solar production data, wind production, and Minnesota load
data. Cost forecasts were used for solar, storage, and wind resources. Hourly data and cost
forecasts are described in this section.
Electrification forecasts and load shifting capabilities for domestic hot water (DHW), electric
vehicles (EVs), and heating, ventilation and air conditioning (HVAC) were also data inputs into
the SPA. These data inputs are described in Appendix B: Electrification and Load Shifting
Models.
SPA Hourly Data
Solar Production Data
Hourly solar irradiance and production data were obtained
from SolarAnywhere. SolarAnywhere is a commercial solar
irradiance and weather data set produced by Clean Power
Research. SolarAnywhere irradiance estimates are derived
from satellite data and are available from 1998 through
present.
SolarAnywhere additionally includes proprietary solar
production simulation capabilities that are derived from
Sandias PV Form model (the same base model for PV Watts).
Historical solar production (2014-2016) was simulated for
every tile of a 10-km grid across the State of Minnesota.
Solar systems were assumed to be south-facing with a 30-
degree tilt.
Wind Production Data
Hourly wind production data from 2014-2016 was obtained
from MISO for existing wind assets owned by Great River Energy
(GRE), Minnesota Power MP), Otter Tail Power (OTP), Southern
Minnesota Municipal Power Authority (SMMPA), and Xcel
Energy (Xcel).
The hourly wind production data was then combined with wind plant commissioning data from
the above utilities to create a wind production profile that could be scaled to a desired wind
Solar Potential Analysis Report | November 15, 2018 | Page 17
capacity. For example, the wind production data was scaled to serve 25% of Minnesotas load
for the SPAs 2025 analyses, as previously noted.
Minnesota Load Data
Hourly load data from 2014-2016 was obtained from MISO for five balancing areas that serve
the overwhelming majority of Minnesota’s load: GRE, MP, Xcel, OTP, and SMMPA. The
geographic coverage of these five balancing areas is represented below and serves 86% of
electricity customers in the state14, including the state’s largest urban centers.
The aggregation of the hourly load from these balancing areas is used to represent the hourly
load for the State of Minnesota.
Figure 5. Service Territories Included in Creation of Minnesota Load Data.
Use of Historical Data
The SPA used historical production and load data. The SPA did not use typical year data.
The use of historical data enables the time-correlation of production data and load data. Time-
correlation is critical for accurately capturing the impact of solar and wind production
variability.
14 Minnesota Department of Commerce 2014 data.
Solar Potential Analysis Report | November 15, 2018 | Page 18
Must-Run Resources (2025 Only)
An indirect component of the hourly datasets described above was the identification of must-
run resources. Must run resources are resources that must-run independent of the hourly load
and/or solar and wind production. For example, nuclear generation resources are assumed to
run at their rated production capacity.
The proper accounting of must-run resources by the SPA is important during periods of high
solar and wind production and low load. During these periods, solar and wind production may
exceed the load available to be served after accounting for must-run resources. As such, the
SPA either stored or curtailed excess solar and wind production when these conditions
occurred.
The Pathways Technical Committee determined that must-run resources were applicable in the
2025 timeframe, but that no must-run resources were applicable in the 2050 timeframe. The
agreed upon must-run resources in the 2025 timeframe are provided in Table 1. Notably, as a
conservative assumption, wind was included as a must-run resource in the 2025 timeframe to
account for wind generation assets that have already been built or are in development. The
minimum generation for wind was assumed to be its generation for the hour of interest.
Table 1. Must-Run Resources in the 2025 Timeframe.
Utility
Plant
Plant Type
Capacity
(MW)
Min Gen. Cap.
(MW)
Multiple Multiple Wind 6109 n/a
MP Boswell 3 & 4 Coal 940 423
OTP
Big Stone
Coal
257
115
Xcel
Monticello
Nuclear
671
671
Xcel
Prairie Island
Nuclear
1100
1100
Xcel/
SMMUA
Sherco 3 Coal 876 394
Xcel New CC15 Gas 785 353
Total
(excluding wind)
4629 3056
15 Xcel Energy notes that its new combined cycle plan is considered a must-run resource but will likely operate in a
more flexible manner.
Solar Potential Analysis Report | November 15, 2018 | Page 19
As evident in Table 1, the different resources had different minimum generation capacities.
Coal and gas resources were assumed to have minimum generation capacities equivalent to
45% of their rated production capacity, while nuclear resources were assumed to have
minimum-generation capacities equivalent to their rated production capacity.
Not listed in Table 1 is Coal Creek Station, a 1,146 MW coal generation plant owned by GRE. Per
its 2017 IRP, GRE is “beginning to more flexibly operate the station to match lower market
prices and to provide energy when intermittent resources are not available.” Additionally, GRE
has announced an accelerated depreciation schedule for Coal Creek Station that would allow it
to retire as early as 2028. Given Coal Creek Station’s increased flexibility of operation and its
potential retirement date near the 2025 timeframe it was not included as a must-run resource.
Cost Forecasts
The solar, wind, and storage cost forecasts used by the SPA are provided in the Forecasted
Technology Costs sub-section of the SPA Scenarios section. This section discusses the
development of the SPA cost forecasts.
The SPA cost forecasts were developed based on a collection of cost forecasts provided by
NREL. There were six key sources utilized for solar cost forecasts, two key sources for wind cost
forecasts, and four key sources for storage cost forecasts. The sources of the cost forecasts
provided by NREL are in Table 2.
Table 2. Cost Forecasting Sources.
Solar Potential Analysis Report | November 15, 2018 | Page 20
Cost forecasts were generally not available for community solar. As such, community solar costs
were assumed to fall halfway between commercial solar and utility-scale solar costs given that
community solar systems are generally between commercial (50kW-500kW) and utility-scale
solar (5-20 MW) systems in size and that system cost declines with system size for large PV
systems.16
The storage cost forecasts were largely for Li-ion batteries, though some storage cost forecasts
included costs for flow batteries. Most forecasts did not distinguish between the size/type of
storage systems deployed (e.g., residential, commercial, or utility-scale). As such, the
technology cost projections focused on the cost forecasts for Li-ion batteries and did not
differentiate cost by battery size/type.
As a final note, the wind cost forecast from NREL’s 2017 Annual Technology Baseline was for a
wind resource with a weighted average wind speed of 7.5 m/s (TRG 5), which was viewed as
appropriate for wind projects in Minnesota.
16 NREL System Cost Benchmark Q1 2016 (http://www.nrel.gov/docs/fy16osti/66532.pdf)
Solar Potential Analysis Report | November 15, 2018 | Page 21
SPA Scenarios
The Pathways Team used the SPA to evaluate scenarios with different: 1) Production
Requirements; 2) future technology costs and adoption levels (collectively, Technology
Development); and 3) Solar Distributions. Notably, future technology costs and Production
Requirements significantly affected modeled generation costs and resource capacities.
Figure 6. Construction of SPA Scenarios from Choice of Production Requirements, level of Technology
Development, and Solar Distribution.
Having produced SPA results from the initial set of scenarios, two sensitivity studies were
conducted on two of the SPA scenarios. The first sensitivity study examined the change in the
SPA results if Minnesota load was served with Other Generation resources during periods of low
solar and low wind production. The second sensitivity study examined the effect of the cost of
capital on the SPA results.
Note, the SPA scenarios studied were developed by the MN Solar Pathways Core Team in
coordination with the Technical Committee. The Core Team presented to the Technical
Committee the different components that would comprise the SPA scenarios studied (namely
Production Requirements, level of Technology Development, and Solar Distribution) and then
worked with the Technical Committee to develop the initial set of SPA scenarios to be
modelled.
Production Requirements
The choice of production requirements strongly influences the SPA’s calculated generation cost
and resource capacities. As such, the production requirements selected by the Pathways Team
were an important part of the process for bounding the generation cost and resource capacities
to achieve the 10% and 70% targets of interest in the 2025 and 2050 timeframes, respectively.
Solar Potential Analysis Report | November 15, 2018 | Page 22
The Pathways Team selected production requirements that represented a range of production
dispatchability from a set of production requirement with almost no dispatchability
(‘Unconstrainedproduction requirements) to a set of production requirements that was nearly
fully dispatchable (the Hourlyproduction requirements). The full list of production
requirements studied and their relative dispatchability are shown in Figure 7. These production
requirements are also discussed in detail in Appendix A: Production Requirements.
Figure 7. Dispatchability of SPA Production Requirements.
As an illustrative example, the ‘Hourly’ production requirements are shown in Figure 8. For the
Hourly production requirements, solar, wind, and storage are tasked with guaranteeing delivery
of a constant percent (e.g., 70%) of Minnesota’s load for each hour of the year.
Dispatchable generation generally refers to the ability of a generation resource to flexibly
respond to match the load shape in real time. While dispatch is a technical term with specific
meaning for different groups, in practice all generation exhibits both some degree of
dispatchability and limits on that dispatchability. For example subject to some limitations,
utility-scale solar and wind generation resources are already dispatching into energy markets
around the country.
The aim with the use of the term dispatch is to highlight the increased flexibility and
expanded dispatch capabilities of the solar and wind resources when paired with the
strategies implemented in the SPA.
Solar Potential Analysis Report | November 15, 2018 | Page 23
Figure 8. Illustration of Solar, Wind, and Storage Production Requirements (‘Hourly’ Production
Requirements are shown).
Technology Development
Clean Power Research developed high and low Technology Development scenarios for both the
2025 and 2050 timeframes. Each of these Technology Development scenarios contained
forecasted technology costs and forecasted technology adoption components.
Forecasted Technology Costs
Technology cost forecasts were developed as described in the Cost Forecasts sub-section of the
SPA Data Inputs section. The forecasted technology costs are shown in Table 3 for the 2025 and
2050 timeframes.
Table 3. SPA Technology Costs for the 2025 and 2050 Timeframes.
Solar Potential Analysis Report | November 15, 2018 | Page 24
Forecasted Technology Adoption
Technology adoption forecasts for domestic hot water, domestic heating, ventilation and air
conditioning, and electric vehicles were developed by the Center for Energy and Environment
(CEE), as described in Appendix B: Electrification and Load Shifting Models. CEE based its
adoption projections on current penetrations, published growth projections from a number of
sources, and an aggressive electrification campaign for the high adoption estimates.
Forecasted technology adoption is provided in Table 4 for the 2025 and 2050 timeframes. The
adoption forecasts assume significant electrification of residential heating and transportation
loads over time.
Table 4. SPA Technology Adoption for the 2025 and 2050 Timeframes.
Solar Distribution
The solar distribution scenarios consist of two components. The first component is the type of
the solar, and the second is the spatial allocation of the solar. The spatial allocation of solar
depends partly on the type of solar. For example, utility solar has siting limitations that differ
from rooftop solar and community solar.
Do the Cost Forecasts include either the Investment Tax Credit (ITC) or the Production Tax
Credit (PTC)?
The cost forecasts do not include an ITC or PTC they are unsubsidized costs.
Additionally, the SPA does not include an ITC or PTC.
Are the Cost Forecasts in Real or Nominal dollars?
The cost forecasts are in nominal (aka current) dollars.
Solar Potential Analysis Report | November 15, 2018 | Page 25
Type of Solar
Clean Power Research developed two solar scenarios using results from its Technical
Committee survey and Technical Committee feedback on draft scenarios. In the first scenario,
the Utility-Led scenario, most of the solar capacity
comes from utility solar. In the second scenario, the
All Sectors scenario, each type of solar (residential,
commercial, community, and utility) makes a
meaningful contribution to the overall solar
capacity. The percentage of solar capacity in each
solar sector is provided in Table 5 for the Utility-Led
and All Sectors scenarios.
Table 5. Solar Capacity by Type of Solar for the Solar
Distribution Scenarios.
Spatial Allocation of Solar
The Core Team surveyed the Technical Committee for key factors related to the deployment of
solar capacity within Minnesota. Through the survey and discussions at the Technical
Committee meeting, the following key factors were identified:
Population density
Proximity to transmission
Annual irradiance
Exclusion of wetlands, forest, and open-water
Population density was identified as a key factor since residential and commercial solar capacity
is located on the customer’s property. Proximity to transmission was identified as a key factor
for utility-scale solar since it is connected to the transmission grid and thus should be located
close to existing transmission capacity if possible. The average annual irradiance was included
because of its direct effect on the economic value of the solar asset. Lastly, wetlands, forest,
and open-water were excluded given the sensitive nature of these land types.
Once these key factors were determined, Clean Power Research developed a spatial allocation
algorithm for the Utility-Led and All-Sectors scenarios, as detailed in Appendix D: Spatial
Allocation of Solar.
Using the algorithm the solar spatial allocations shown in Figure 9 were developed. For the
Utility-Led spatial allocation, more of the solar capacity is in southwestern Minnesota. For the
All-Sectors spatial allocation, much of the solar capacity is in the Minneapolis Saint Paul
metropolitan area.
Solar Potential Analysis Report | November 15, 2018 | Page 26
Figure 9. Spatial allocation of solar for the Utility-Led and All Sectors scenarios.
Solar Potential Analysis Report | November 15, 2018 | Page 27
SPA Results
The SPA work produced an extensive set of results. The final results included nearly 100
scenarios, each of which has associated hourly datasets and key summary criteria.
In this section, we discuss:
2025 SPA Results
2050 SPA Results
Electrification and Load Shifting
10% Solar by 2025
Table 6 presents SPA results for the 2025 timeframe. Two levels of Technology Development
(High and Low); three sets of Production Requirements (Unconstrained, Predictable, and
Seasonal); and two Solar Distributions (Utility-Led and All Sectors) were studied. Key results
include: generation cost ($/MWh), solar capacity (GW), and storage capacity (GWh). As a
reminder, the reported generation cost is a levelized cost of energy that is calculated based on
the set of resources (solar, wind, and storage) required to guarantee delivery of the specified
production requirements.
Table 6. Key SPA Results in the 2025 Timeframe.
Solar Potential Analysis Report | November 15, 2018 | Page 28
The SPA results for the Seasonal Production Requirements are presented in light-grey to
deemphasize their relevance in the 2025 timeframe. This was a decision made by the Core
Team after realizing that the Seasonal SPA results were heavily influenced by periods of low-
solar production. (This dependence was also found for the 2050 SPA results as discussed in the
next section). While this is appropriate at high-penetrations of solar and wind generation, it is
not relevant at a 10% solar penetration where existing generation resources could be utilized
during a multi-day period of low solar production.
Examining Table 6 we note the following:
As expected, generation cost and resource capacities increased slightly between
scenarios with Unconstrained Production Requirements (with no storage requirements)
and scenarios with Predictable Production Requirements (increased dispatchability with
solar plus storage sufficient to meet the day ahead hourly forecast and any shortfall in
actual production).
6 GW of solar capacity and 2 GWh of storage capacity would enable solar to meet its
day-ahead forecasted production.
Generation costs are 15%-20% higher when comparing All Sectors scenarios with their
Utility-Led counterparts.
Overall, the 2025 SPA results indicate that Minnesota could achieve its statutory but non-
binding goal of 10% solar by 2030 at a cost that is comparable with natural gas generation costs
(see Appendix I: Cost of Natural Gas Generation Resources). Note, this is not to say that the sets
of solar and storage resources above can be dispatched in a manner equivalent to a natural gas
generation resource or vice-versa.
production from solar resources: the solar production cannot exceed the Minnesota load
Requirements but impose an additional constraint to account for day-ahead production
resource but requires that solar and storage smooth out the daily variability of solar to
Solar Potential Analysis Report | November 15, 2018 | Page 29
70% Solar and Wind by 2050
In this section, we present 2050 SPA results with and without Other Generation resources and
discuss four of the five key findings from the 2050 SPA results. The fifth key finding is discussed
in Electrification and Load Shifting.
2050 SPA Results without Other Generation Resources
Table 7 presents SPA results for the 2050 timeframe. Two levels of Technology Development
(High and Low); two sets of Production Requirements (Seasonal and Hourly); and two Solar
Distributions (Utility-Led and All Sectors) were studied.
Table 7. 2050 SPA Results without Other Generation Resources.
Solar Potential Analysis Report | November 15, 2018 | Page 30
Generation costs for the High Technology Development scenarios are comparable to or slightly
higher than natural gas generation costs under a range of fuel and capital costs (see Appendix I:
Cost of Natural Gas Generation Resources). Generation costs for the Low Technology
Development scenario are roughly twice natural gas generation costs.
As expected, an increase in generation cost (5-10%) was observed between scenarios with
Seasonal Production Requirements (seasonally dispatchable) and scenarios with Hourly
Production Requirements (dispatchable). The Pathways Team elected to focus on the Hourly
Production Requirements as being of most interest given the ultimate need to serve
Minnesota’s hourly load.
Similar to the 2025 SPA results, the All Sectors scenarios exhibited higher generation costs than
their Utility-Led counterparts. However, the optimization across solar, storage and wind in the
2050 results meant that the SPA could utilize more wind and less solar in the All Sectors
scenario than in the Utility-Led scenarios. This resulted in generation costs that were only
slightly higher (~5%) for the All Sectors scenarios in 2050.
The most notable results in Table 7 are the storage capacity and energy curtailment. The SPA
found the optimal storage capacity to be roughly 200 GWh (about 20 hours of Minnesota’s
average load). This is a significant amount of storage capacity but at the same time one-
hundred fold less than existing studies of 100% renewable energy suggest17 (i.e., storage
capacity equivalent to weeks of load). The SPA also found that energy curtailment is key to
minimizing generation costs. As discussed below, this is because energy curtailment
significantly reduces the need for storage capacity.
Key Finding #1: Additional Capacity coupled with energy curtailment is considerably less
expensive than, and a viable alternative to, long-term or seasonal storage in a high
renewables future.
A key finding of the 2050 SPA results is that Additional Capacity coupled with energy
curtailment is an important strategy in a high renewables future. Figure 10 illustrates the effect
of energy curtailment on the 2050 SPA generation cost.
17 Roberts, David. “Is 100% renewable energy realistic? Here’s what we know.” Vox. Feb. 7th 2018.
https://www.vox.com/energy-and-environment/2017/4/7/15159034/100-renewable-energy-studies
Solar Potential Analysis Report | November 15, 2018 | Page 31
As the SPA adds solar and wind capacity, their respective generation costs increase. However,
as additional solar and wind capacity are added, storage capacity (and storage capacity costs)
significantly decrease. The net effect is that the total generation cost (the SPA generation cost)
initially decreases as additional solar and wind capacity are added, levels as the optimal set of
capacities is reached, and finally rises as the cost increases associated with additional solar and
wind capacity outpace the cost reductions associated with storage capacity. For the Hourly
Production Requirements the SPA generation cost-minimum occurs around 50% energy
curtailment, due in part to the stringent requirements of the Hourly Production Requirements.
The results shown are for Hourly Production Requirements with High Technology Development
and a Utility-led Solar Distribution. However, similar results were found for Hourly Production
Requirements using Low Technology Development scenarios and All Sector Solar Distribution
scenarios.
Figure 10. Influence of Additional Capacity coupled with Energy Curtailment on Generation Cost and
Resource Deployment.
Additional Capacity coupled with energy curtailment is a strategy that runs counter to the goal
of sizing renewable capacity in a manner that avoids ‘wasting’ it. However, designing for 100%
use of renewable energy would require a storage capacity sufficient to shift a large quantity of
energy over a seasonal time-period (as shown in Appendix J: Benefits of Additional Capacity).
Additional Capacity is solar and wind capacity over and above that needed to meet annual
energy needs but still beneficial for improving the economic dispatchability of the solar/wind
Solar Potential Analysis Report | November 15, 2018 | Page 32
By contrast, energy curtailment alleviates the need to seasonally shift renewable energy, thus
enabling much smaller storage capacities.
Key Finding #2: Flexible Other Generation resources used in limited amounts support a high
renewables future.
Another key finding from the 2050 SPA results is the ability of Other Generation to support a
high-renewables future. The strategic use of Other Generation resources during brief periods
of low-solar and low-wind production significantly reduced the storage, solar, and wind
capacities used to serve Minnesota’s hourly load. As a result, the generation cost for 70% solar
and wind was reduced by nearly half. This is best understood by examining how a few brief
periods of low solar and low wind production are responsible for over half the storage capacity
specified in the SPA results without Other Generation resources.
Figure 11 illustrates the storage state of charge for the Hourly Production Requirements, High
Technology Development, Utility-Led Solar Distribution scenario. For most of the year, the 195
GWh of storage capacity maintains state of charge levels above 74% of the maximum energy
capacity (145 GWh). However, during a few periods of low wind and solar production the
storage is discharged more fully. The total storage capacity and costs are determined from the
largest drawdown, a total discharge in early January.
Figure 11. Storage State of Charge (GWh) minimum state of charge plotted for each day in a calendar
year. Dashed line in gold denotes 74% or 145 GWh state of charge.
portion of the Hourly Production Requirements during brief periods of low-solar and wind
resources.
If the SPA had included integration with MISO, solar and wind resources outside of
Solar Potential Analysis Report | November 15, 2018 | Page 33
Considering these results, it is apparent that Other Generation resources could significantly
reduce the SPA’s storage capacity (and generation cost) by providing generation during
periods of low solar and wind production.
The ability to utilize Other Generation resources during periods of low solar and low wind
production was added into the SPA. Other Generation resources could include existing non-
renewable resources in Minnesota or imports from MISO. For the purpose of economic
modeling, Other Generation resources were assumed to be existing natural gas resources (fuel
and O&M costs were included, capital costs were not included).
Figure 12 plots the SPA’s calculated generation cost against the fraction of total energy
generation provided by Other Generation resources.18 Initially, generation costs fall markedly as
Other Generation resources are allowed to serve up to 10% of the SPA’s annual load target
(e.g., 10% of 70%). However, generation cost declines less significantly beyond 10% of the
annual load target. Focusing on the range where Other Generation resources serve 5-10% of
the annual load target, the Pathways Core Team notes that generation costs are reduced either
35% (High Technology Development) or 55% (Low Technology Development) from the cost to
meet the Hourly Production Requirements using solely solar and wind resources in Minnesota.
This significant reduction in costs is a general indication of the value of the flexibility from Other
Generation resources or the MISO market.
Figure 12. Effect of Utilizing Other Generation Resources during Periods of Low Renewables Production.
18 Results shown are for Hourly Production Requirements and Utility-led Solar Distribution.
Solar Potential Analysis Report | November 15, 2018 | Page 34
2050 SPA Results with 10% Other Generation
The SPA results that include Other Generation resources (Table 8) provide significant value by
reducing the SPA’s 2050 generation cost range from $56-$132/MWh with no Other Generation
resources to a much tighter and more actionable range of $36-59/MWh with Other Generation
resources serving 10% of the annual target load.
Table 8. 2050 SPA Results with 10% Other Generation Resources.
Note, when Other Generation resources are used during periods of low solar and low wind
production, the Hourly Production Requirements need to serve a greater percentage of the
hourly Minnesota load. When Other Generation resources serve 10% of the annual target load,
Minnesota solar and wind would serve 78% of the hourly Minnesota load (except during
periods of low solar and low wind production when they would serve less of the load) in order
to serve 70% of Minnesota’s annual load.
Key Finding #3: Storage is an important part of a high renewables future; it expands the
dispatch capabilities of wind and solar assets
The storage capacity found by the SPA (16 to 50 GWh equal to roughly one to five hours of
Minnesota’s average hourly load) is sufficient to address the intra-hour variability of solar and
wind, as well as of Minnesota’s load. It is further sufficient to shift solar and wind production to
meet load on a daily basis. Other Generation resources can be utilized when production shifting
beyond a daily basis is required.
Solar Potential Analysis Report | November 15, 2018 | Page 35
Key Finding #4: Solar and wind can serve 70% of Minnesota’s electrical load
Overall the 2050 SPA results indicate that the expected cost declines of solar, wind, and storage
will enable Minnesota to achieve 70% solar and wind by 2050 with generation costs
comparable to natural gas generation costs ($39/MWh to $64/MWh, see Appendix I: Cost of
Natural Gas Generation Resources).
This is noteworthy given the stringent production requirements used when modeling the 2050
results: solar, wind, and storage were modeled with a requirement to serve a constant fraction
of Minnesota’s hourly load profile (for every hour of the year). An exception was made during
periods of low-solar and low-wind production, during which time Other Generation resources
were used to support generation requirements, reducing costs to meet the Hourly production
profile by approximately half.
Electrification and Load Shifting
In this section we discuss how electrification and load shifting can impact the SPA results.
Electrification is the conversion of fleets of appliances and transportation powered by burning
fuel like natural gas and gasoline to electricity. Examples considered in the SPA included
transportation, space heating and domestic hot water (DHW). Load shifting refers to the ability
to move energy consumption to a different hour, or in limited cases, a different day. The
greater the time shift, the more difficult it can be for the end user to accommodate.
Prior to discussing SPA results, we summarize how residential electrification could impact
Minnesota’s hourly load profile by 2050 assuming nearly complete electrification of residential
load. We also summarize the degree to which electrified loads can be shifted. For greater
discussion of these findings please see Appendix C: Electrification and Load Shifting Results.
Electrification:
DHW would add ~1 GW of load (on average) from 2.8 million electric water heaters
o DHW load peaks in the morning and is lowest over-night
EVs would add roughly ~4 GW of load (on average) from 4.9 million electric vehicles
Solar Potential Analysis Report | November 15, 2018 | Page 36
o EV load peaks in the early evening and the sharpness of the peak is heavily
dependent on charger type (e.g., L1 chargers vs. L2 chargers)see Figure 13
HVAC would add ~3 GW of load (on average) during the coldest weeks of the year
o Timing of HVAC load peaks are weather dependent and vary on a daily basis
Load shifting:
DHW could shift ~10 GWh of load throughout a day
EVs could shift ~30 GWh of load throughout a day and a few GWh over a 2-3 day
period
HVAC could shift a few GWh of load over 1-3 hours
Figure 13. EV Load in 2050 with L1 (right) and L2 (left) chargers. (The dashed green line is the EV load
with High Technology Development, the solid blue line is the load with Low Technology Development).
SPA Results with Electrification and Load Shifting
The electrification results indicate that electrification of DHW, residential heating (using air
source heat pumps), and transportation has a minimal effect on the generation costs found by
the SPA. Generation costs of supplying the Predictable Production Requirements in 2025
decrease by about 5% with electrification. Generation costs of supplying the Hourly Production
Requirements in 2050 increase by about 5% with electrification. Resource capacities increase to
cover the additional load in both timeframes.
Key Finding #5: Shifting of key flexible loads may further decrease generation costs.
Load shifting residential DHW and EV loads in 2025 provides 10% generation cost reductions
from electrified generation costs for the Predictable Production Requirements. Load shifting
Solar Potential Analysis Report | November 15, 2018 | Page 37
residential DHW and EV loads in 2050 provides either 20% generation cost reductions from
electrified generation costs (High Technology Development) and 3% generation cost reductions
(Low Technology Development) for the Hourly Production Requirements.
This result represents an optimistic scenario for the possible benefits from electrification paired
with load shifting for three reasons now discussed. First, EV chargers were assumed to be
available whenever a vehicle was parked. Second, the SPA did not attribute a cost to load-shift
resources. Third, the SPA modified the existing load profile with electrification to the point
where the load profile with load shifting would likely impinge on transmission and distribution
capacity limits. Coordinated load control will help reduce potential impacts.
Solar Potential Analysis Report | November 15, 2018 | Page 38
Discussion of SPA Results
SPA Scope:
After examining the SPA results, it is worth considering the effect of two of the constraints
associated with the SPA scope on the SPA results:
The SPA does not include integration with MISO
The SPA is based on installed and operational costs (it does not consider rate structures)
In discussing the effects of the constraints, there were members of the MN Solar Pathways
Technical Committee who felt that the lack of integration with MISO led to a conservative
analysis with higher generation costs and resource capacities. This belief was borne out by
reduced generation costs for SPA scenarios that included small amounts of Other Generation
resources.19 Conversely, there were members of the Technical Committee who noted that the
use of installed and operational costs (as compared to costs derived from market payments
and/or current rate structures) would underestimate total costs, especially for the All-Sectors
scenarios. The MN Solar Pathways Core Team acknowledges these perspectives. The net effect
will be subject to market and regulatory decisions and is unknowable at this time. The Core
Team additionally notes that the next model being developed under the Pathways initiative, the
Solar Deployment Strategy, will evaluate the effect of rate structures and their impacts on
markets along with value propositions of different solar deployment scenarios.
The Value Proposition of Additional Capacity coupled with Energy Curtailment
At present there is a significant industry focus on the use of generation flexibility (i.e., fast-
ramping resources) and load shifting to accommodate the production variability of solar and
wind resources. By comparison, relatively little attention has been paid to the use of Additional
Capacity (solar and wind) combined with energy curtailment.
The SPA results suggest that Additional Capacity offers a low-cost way to overcome the
production variability of solar and wind resources. Specifically, Additional Capacity is uniquely
positioned to address seasonal and multi-day periods of production variability that:
cannot be provided by load shifting due to the multi-day time-scales required
cannot be provided by natural gas resources without carbon emissions
19 Allowing other in-state generation resources to ramp-up in periods of low resource is similar to using the
surrounding MISO region as a generation resource.
Solar Potential Analysis Report | November 15, 2018 | Page 39
cannot be provided by storage alone without significant storage costs20
A final note on Additional Capacity: while the Pathways Team was quick to recognize its value
proposition, the Pathways Team understands that existing solar and wind compensation
mechanisms would need to be adjusted to accommodate the associated energy curtailment.
Cost of Capital
The cost of capital is an important input into the SPA. The results presented above assumed a
cost of capital of 5%. SPA results with cost of capital between 4% and 6% are presented in
Appendix G: Cost of Capital for a subset of SPA scenarios (Hourly Production Requirements,
Utility-Led Solar Distribution scenarios in the 2050 Timeframe). The SPA results presented in the
appendix indicate that cost of capital has a moderate effect on generation costs, but a much
smaller effect on the type (solar, wind, storage) and capacity (GW or GWh) of resources that
are deployed.
Historical Load Data
The SPA utilized historical load data from three years: 2014, 2015, and 2016. The original goal
of the study was to model these years as a three-year period. Due to time and resource
constraints it was only possible to model each year individually. The results presented in this
report are for 2016. SPA results for 2014 and 2015 exhibited similar trends and costs. Clean
Power Research notes that multi-year modeling is important as the optimized resource set will
vary slightly from one year to the next. As such, the generation cost and resource capacities
would be slightly higher for a resource set that meets the needs of a multi-year period.
Utility-Led vs All Sectors
Generation costs for the Utility-Led scenarios were 5-20% lower than generation costs for the
All Sectors scenario. The lower cost of the Utility-Led scenarios is expected given the difference
in the modeled resource costs between the Utility-Led and All Sectors scenarios (Table 3).
Overall these results indicate that Minnesota has flexibility in the distribution of solar
resources.
Resource capacities were similar between Utility-Led and All Sectors. Some All Sectors scenarios
had greater wind capacities (relative to solar capacities). This makes sense when considering
that solar costs are higher relative to wind costs in the All Sectors scenarios.
20 The forecast for storage cost reductions assumed in the SPA is aggressive. Unless there are unexpected cost
declines in Lithium Ion or other storage technologies, the Pathways Team asserts the value of ‘additional’ solar and
wind.
Solar Potential Analysis Report | November 15, 2018 | Page 40
The Pathways Team acknowledges that there may be reasons to invest in distributed solar (as
with the All Sectors scenarios) notwithstanding the higher generation cost relative to utility-
scale solar (Utility-Led scenarios). The SPA specifically examines generation costs and is not a
tool for evaluating the value of one type of deployment over another. The Solar Deployment
Strategy model being developed in the latter half of the Pathways project is intended to be
useful in comparing costs and benefits of different deployment strategies.
Limitations to Existing Generation Capacity Credit Methods
During discussions of the SPA results with the Technical Committee, the question was raised
whether the resource capacities found by the SPA would satisfy generation resource adequacy
requirements. The Pathways Team did not evaluate this question in detail. However, a quick
look at existing capacity credit methods reveals an obvious challenge for the SPA results.
Existing generation capacity credit methods look at resources individually, as opposed to
holistically. For example, solar and wind would receive fixed capacity credits (currently 50% and
15%, respective class averages, in MISO),21 while storage would not receive a capacity credit
under current MISO resource adequacy methodologies as it is not a generation resource.
Ongoing discussions at Federal Energy Regulatory Commission (FERC) center around how to
value storage in markets and include related conversations about eligibility and rules in capacity
markets and for resource adequacy. FERC Order 841 requires independent system operators to
implement a method of assigning capacity credit to storage. 22 To fairly evaluate the generation
resource adequacy of a set of solar, wind and storage resources, a holistic generation capacity
credit method is needed.
Area Required for Solar Deployment
During discussions with the Technical Committee, the question arose as to whether there was
sufficient roof-space within the Minneapolis Saint Paul metropolitan area to accommodate the
allocated solar capacity in the All Sectors scenario. Great Plains Institute has previously studied
the rooftop solar potential through the Local Government Project for Energy Planning project,
based on Minnesota’s LiDAR-based (1-meter resolution) solar resource map. GPI found that
cities typically had enough economic rooftop solar resource to generate an equivalent of 40-
60% of the total electric consumption within the city, which exceeds the distributed solar
capacity modeled for the All Sectors and Utility-Led scenarios in both timeframes (without
21 After a year of operation, a unit’s capacity credit is based on historical coincidence with peak.
22 Order 841 Electric Storage Participation in Markets Operated by Regional Transmission Organizations and
Independent System Operators. Federal Energy Regulatory Commission.https://www.ferc.gov/whats-new/comm-
meet/2018/021518/E-1.pdf. Feb. 15, 2018.
Solar Potential Analysis Report | November 15, 2018 | Page 41
considering any accessory use ground-mount capacity). GPI’s measured analysis aligns with
NREL findings that modeled the rooftop solar resource across the nation. NREL estimated that
the national technical potential (gross resource) of rooftop solar was approximately 39% of
national electric usage.23
Clean Power Research separately calculated the area required for the extreme case of serving
all of Minnesota’s electrical load with solar energy (82,000 acres). See Appendix F: Land Use.
While 82,000 acres might sound large, it is actually quite small when compared with existing
land use in the State of Minnesota. As can be seen in Figure 14 below, 82,000 acres is
comparable to the area of Barren Land24 in Minnesota and over 10-fold less than the area of
Developed Land in Minnesota, based on data from the National Land Cover Database.
Figure 14. Area Required for Solar Deployment Compared with Existing Land Use in Minnesota.
84% Reduction of Minnesota’s Carbon Intensity
The 2050 SPA results represent an 84% reduction in the carbon intensity of Minnesota’s electric
sector from 2005 levels if the Other Generation resources are solely comprised of natural gas
(blend of CT and CCGT). The reduction in carbon intensity would be even greater if the Other
Generation resources included nuclear resources and/or solar and wind resources from outside
of Minnesota.
23 NREL; Rooftop Solar Photovoltaic Technical Potential in the United States. (2016)
24 Barren Land (Rock/Sand/Clay) areas of bedrock, …, gravel pits, and other accumulations of earthen material.
Solar Potential Analysis Report | November 15, 2018 | Page 42
Figure 15. Carbon Intensity of Minnesota’s Electric Sector.
Minnesota's Carbon Intensity
Generation Resource
2005
Generation
(%)
CO2 Emissions
(Metric Tons
/MWh)
2050
Generation
(%)
CO2 Emissions
(Metric Tons
/MWh)
Nuclear
24%
0.00
Coal
62%
0.98
Natural Gas Blend
5%
0.44
Wind, Hydro, Biomass
7%
0.00
Petroleum
2%
0.79
Solar and Wind
70%
0.00
Hydro, Wood, Biomass
5%
0.00
Gas - CT
8%
0.53
Gas - CCGT
17%
0.35
Total (%)
100%
n/a
100%
n/a
Weighted Average CO2
Emissions (Metric
Tons/MWh)
n/a 0.64 n/a 0.10
Reduction from 2005 Carbon Intensity
84%
Solar Potential Analysis Report | November 15, 2018 | Page 43
Conclusion
The SPA afforded a unique opportunity for a broad set of energy stakeholders in the State of
Minnesota to meet regularly over an 18-month period to develop a shared understanding of
the opportunities and challenges associated with increasing levels of solar and wind generation.
At the highest-level the SPA results indicate that solar, wind, and storage resources can
reasonably serve a majority of Minnesota’s load in terms of generation costs. As such, the SPA
results can be used to justify the thorough evaluation of solar, wind, and storage as part of
Minnesota’s energy future a process that many energy stakeholders in Minnesota have
already begun.
The SPA results provide important insights into the solar, wind, and storage capacities that
would achieve a future where solar and wind serve 70% of Minnesota’s annual load. Under a
70% scenario, the SPA results suggest that Minnesota could expect to have tens of GW of solar
and wind with just tens of GWh of storage capacity; not 1,000’s of GWh of storage capacity. In
this way, the SPA results can be useful in combination with other studies, as decision-makers
and stakeholders anticipate and plan for expanded solar, wind, and storage capacity.
The SPA results also highlight the value of Additional Capacity (solar and wind) and the future
need for discussion about solar and wind compensation policies that account for Additional
Capacity (and its associated energy curtailment). To this end, the SPA results can be used to
begin examining this important issue before it is pressing. Note that the Pathways project is
agnostic on the numerous potential solutions, but rather raises this as a point of discussion.
Early discourse will prove valuable since any new compensation policies will ultimately need to
move through a regulatory process.
Finally, the SPA results provide some insight into the effect of different solar distributions on
Minnesota’s possible energy futures. However, a deeper analysis of the effects of different
solar distributions is needed. The second phase of the Minnesota Solar Pathways project seeks
to undertake such an analysis through the development of the Solar Deployment Strategies
modeling tool, which will consider the value propositions of different solar deployment
scenarios in addition to costs.
Solar Potential Analysis Report | November 15, 2018 | Page 44
Appendix A: Production Requirements
The SPA relies on user specified Production Requirements or “production profiles.”
A production profile is the production that will be delivered (by solar, wind, and storage) for
each hour of the year and can vary greatly in terms of dispatchability. In this report, the SPA
included production profiles that varied from minimal constraints for achieving 10 percent solar
(low dispatchability) to serving 70 percent of Minnesota’s load every hour of the year (high
dispatchability).
Each production profile has 8760 values (24 hrs/day * 365 days/year) and may also be referred
to as an ‘8760’ profile by utility engineers. Each of the five production profiles specified by the
Technical Committee and analyzed in this report is defined below.
2025 Production Profiles
Unconstrained Production Profile (2025 Timeframe Only)
The Unconstrained production profile imposes only a single constraint on the production from
solar resources: the solar production cannot exceed the Minnesota load after subtracting out
the must-run resources. As can be seen in Figure 16 below, the Unconstrained production
profile has significant variability relative to Minnesota’s hourly load on an annual basis (left)
and a weekly basis (right).
Figure 16. Unconstrained production profile plotted on an annual basis (left) and a weekly basis (right).
Solar Potential Analysis Report | November 15, 2018 | Page 45
Predictable Production Profile (2025 Timeframe Only)
The Predictable production profile is like the Unconstrained production profile but imposes an
additional constraint to account for production forecasting error. The requirement of the
Predictable production profile is that there must be adequate solar and storage capacity to
make up the difference between the solar generation forecasted on the previous day for the
current day (a day-ahead forecast) and the actual generation that occurred. As can be seen in
Figure 17, the Predictable production profile does not differ from the Unconstrained production
profile on the annual basis (left). The Predictable production profile does differ from the
Unconstrained production profile on the hourly basis (right) but this difference is hard to
observe as plottedexamine July 4th to observe a difference in the production profiles. Despite
their similarities, the difference between the Predictable and Unconstrained production profiles
at the hourly level does produce meaningful generation cost and resource requirement
differences in the SPA results.
Figure 17. Predictable Production Profile plotted on an annual basis (left) and a weekly basis (right).
Seasonal-Diurnal Production Profile (2025 Timeframe Only)
The Seasonal-Diurnal production profile represents a profile that allows for the seasonal
variability of the solar resource but requires that solar and storage smooth out the daily
variability of solar to match the expected seasonal production of the resource. This is illustrated
in Figure 18, where the hourly and daily variability of the solar resource is eliminated with the
Seasonal-Diurnal profile (right), while the seasonal variability is still present (left).
Solar Potential Analysis Report | November 15, 2018 | Page 46
Figure 18. Seasonal-Diurnal Production Profile plotted on an annual basis (right) and a weekly basis
(left).
2050 Production Profiles
Seasonal Production Profile (2050 Timeframe Only)
Like the Seasonal-Diurnal production profile, the Seasonal production profile smooths out the
hourly and weekly production variability of the solar (and now wind) resources while allowing
for seasonal variability. Unlike the Seasonal-Diurnal profile, the Seasonal production profile
blends in a portion of Minnesota’s hourly load profile. This ensures that the Seasonal
production profile produces some energy every hour of the year as can be seen in Figure 19
(right). Blending Minnesota’s hourly load into the Seasonal-Diurnal profile also limits the
magnitude of the ramp rates to which the non-solar and non-wind resources would be exposed
(not shown). Additionally, it ensures that the production profile partially tracks the daily (but
not seasonal) variability of the load, producing more energy on days with greater load and less
energy on days with lower load this effect can be seen most easily when looking at the annual
basis (left).
One of the key rational behind the design of the seasonal production profile is that it allows the
solar and wind resources to produce more during periods of the year when their production is
naturally the greatest (e.g., the summer months for solar and certain months of the year for
wind). It was expected (and to a lesser degree found) that this would reduce the generation
costs and resource requirements found by the SPA in comparison with a production profile that
more closely tracked Minnesota’s hourly load across the year.
Solar Potential Analysis Report | November 15, 2018 | Page 47
Figure 19. Seasonal Production Profile plotted on an annual basis (left) and a weekly basis (right).
Note, because the seasonal production profiles of solar and wind are different, the seasonal
production profile depended on the exact capacity of solar and wind found in the optimized
SPA solution.
Hourly Production Profile (2050 Timeframe Only)
The Hourly production profile, shown in Figure 20, is simply 70% of Minnesota’s hourly load for
each hour of the year. The Hourly profile does not consider any of the production variability
associated with solar and wind assets. Rather, it forces solar and wind to exactly match the
shape of the Minnesota load profile, independent of the solar and wind production.
Figure 20. Hourly Production Profile plotted on an annual basis (left) and a weekly basis (right).
Solar Potential Analysis Report | November 15, 2018 | Page 48
The Seasonal production profile was intended to be an easier production profile than the
Hourly production profile.25 However, the Seasonal production profile did not reduce
generation cost and/or resource requirements to the degree expected. The reason for this
became clear to the Core Team after realizing the importance of brief periods of low solar and
low wind production on the generation cost and resource requirements found by the SPA. The
Seasonal production profile did result in different resource allocations (more solar and less
wind) than the Hourly production profile.
The Hourly production profile was chosen in part because it would ensure that synchronous
generation26 resources to serve at least 30% of the load each hour of the year. 27 Allowing
synchronous generation resources to serve a fraction of the hourly loud addresses one of the
concerns that grid operators have with increasing levels of renewable penetration. The Hourly
production profile would not be the most cost-effective way to serve 70% of Minnesota’s
annual load since it would limit the production of renewables to no greater than 70% of
Minnesota’s hourly load.
25 Lower generation cost and resource requirements.
26 Generation resources that are coupled to the grid through the magnetic field from a spinning rotor. Synchronous
generation acts as a stabilizing force on the grid.
27 The percentage decreases to 22% when Other Generation resources serve 10% of the annual load, except during
periods of low solar and low wind production when Other Generation resources would turn on and increase the
fraction of synchronous generation above 22%.
Solar Potential Analysis Report | November 15, 2018 | Page 49
Appendix B: Electrification and Load Shifting Models
SPA scenarios include assumptions about strategic electrification and the availability of these
loads as a load shifting resource. Initial models were developed outside the SPA by Center for
Energy and Environment to characterize the loads and their flexibility. Some of this work was
incorporated into the SPA with modification by Clean Power Research.
Load Resources Considered
Several loads were considered for electrification and load shifting. These loads were
predominately the larger well-characterized residential and commercial electricity loads that
make up about half of Minnesota load. These are listed by sector and size in Table 9.
Table 9: Specific loads considered for electrification and load shifting and their fraction of Minnesota’s
total load (current and 2050 forecast assuming significant load electrification).
% Total Load
Residential Load
Current
Estimate
2050 “High”
Scenario
Commercial Load
Current
Estimate
2050 “High”
Scenario
HVAC
3.6%
8.9%
Lighting
5.9%
3.4%
DHW
1.6%
5.6%
Ventilation
5.2%
3.0%
Refrigeration
0.5%
0.3%
Refrigeration
4.7%
2.7%
Electric Vehicles
0.0%
25.7%
Cooling
3.8%
2.2%
Computer
3.1%
1.8%
Of these loads, three end-use loads were considered: residential domestic hot water (DHW),
heating & cooling systems (HVAC), and electric vehicles (EV). These loads were selected
because they are currently or have the potential to be among the largest end-use loads on the
electrical grid and they are all inherently shift-able.
Ultimately, the SPA was able to incorporate the electrification effect of DHW, EV, and
residential heating loads and the load shifting effect of DHW and EV loads (but not residential
heating).
Definition of Load shifting
In this report, load shifting is defined as moving energy demand to another point in time, either
before or after that time where the load would otherwise occur. Load shifting does not rely on
behind the meter storage; it requires either the natural ‘inertia’ unique to certain loads (e.g.
HVAC or DHW) or true flexibility over time (e.g. EV charging) to meet the same demand only at
a different point in time. As applied in the SPA model, these load shifting resources have no
procurement or operational costs. In practice, there are costs associated with these resources,
but existing programs using them are generally deemed cost effective. The perspective taken
Solar Potential Analysis Report | November 15, 2018 | Page 50
here is that future demand response resources will confer multiple benefits to utilities and grid
operators. Distributing costs appropriately across these benefits is beyond the scope of this
study. Thus, load shifting results here may represent a technical potential that may be lowered
by additional economic constraints.
General Methodology Used by Electrification and Load Shifting Models
Each SPA Technology Development Scenario included assumptions about the relative pace of
technological development. For the electrification and load shifting models, it was generally
assumed that a high rate of Technology Development would yield an accelerated rate of
electrification and load shifting compared to a low rate of Technology Development.
Each estimate of the amount of load shifting resource is comprised of a fleet of end-use devices
that can be turned on and off within a specified range, determined by the device characteristics
and load requirements. There are three major parts to each electrification and load shifting
estimate:
1. Market Penetration For each load shifting resource, the market penetration is an
estimate for the number of units that are available for control at the SPA scenario
baseline year. Market penetration estimates are comprised of current penetrations and
published growth projections from various agencies and sources. This portion of the
estimate has the largest uncertainty of all three components due to the inherent
uncertainty in forecasting changes 8 to 33 years in the future. These projections
ultimately depend on unknown policy, market, and consumer forces.
2. Opportunity The load shifting opportunity generally describes how much capacity
(energy units, kWh) and how much load (power units, kW) are available from the fleet
of devices. This figure is constructed from the bottom up. This means that it considers
the power and capacity metrics associated with the average end-use loads that make up
the fleet of shift-able load. This portion of the estimate has a relatively low uncertainty
because it is based on published data for actual units. Uncertainty is limited to how
these figures may change between now and SPA baseline years.
3. Constraints This estimate sets the limits or the range of control for these technologies.
It is a bottom up estimate based on the requirements of each load. For example,
domestic hot water systems need to provide sufficient hot water, electric vehicles need
sufficient capacity to meet driving requirements, and heating and cooling systems must
keep occupied spaces conditioned. Additionally, there are maximum power and energy
limits of the fleet. This estimate has a relatively low uncertainty as these requirements
are derived from typical aggregated use patterns. The fleet of shift-able loads is
constrained such that it must meet the same capacity requirements as the
uncontrolled load.
Solar Potential Analysis Report | November 15, 2018 | Page 51
Market Penetration Summary
Figures 21-23 present the forecasted market penetration of DHW, EV, and residential heating
units with load control from 2018 through 2050 for the low and High Technology Development
scenarios. These projections were made by first combining existing equipment penetration
rates 28 with Minnesota population and household growth forecasts29. DHW and HVAC growth
projections are developed from this baseline coupled with the electrification of natural gas and
propane based systems and adjusted based on MISO’s Independent Load Forecast Update30.
Growth forecasts for electric vehicles were developed by blending recent national EV
forecasts31. Load shifting participation rates are based on those found from existing programs32
The 2050 High Technology Development scenario assumes full electrification of domestic hot
water, light-duty transportation, and 50% electrification of the single-family residential heating
load to bound upper potential of this resource.
28 Residential Energy Consumption Survey 2009. (2009). Energy Information Administration.
https://www.eia.gov/consumption/residential/. Accessed May 2017.
Residential Energy Consumption Survey 2015. (2015). Energy Information Administration.
https://www.eia.gov/consumption/residential/ Accessed May 2017.
Commercial Energy Consumption Survey 2012. (2012). Energy Information Administration.
https://www.eia.gov/consumption/commercial/ Accessed May 2017.
29 Minnesota State Demographic Center Database, https://mn.gov/admin/demography/data-by-topic/population-
data/our-projections/, Accessed May 2017.
Minnesota State Profile and Energy Estimates, (2016), Energy Information Administration,
https://www.eia.gov/state/seds/data.php?incfile=/state/seds/sep_use/res/use_res_MN.html&sid=MN,
Accessed May 2017.
30 Gotham, D.J. et al. (2016). MISO Independent Load Forecast Update. State Utility Forecasting Group. The Energy
center at Discovery Park, Purdue University. West Lafayette Indiana.
31 Annual Energy Outlook 2017. (2017). Light-Duty Vehicle Stock by Technology Type: Case: Reference Case. U.S.
Energy Information Administration, https://www.eia.gov/outlooks/aeo/data/browser/#/?id=49-
AEO2017&sourcekey=0. Accessed May 2017.
Electric Vehicle Outlook 2017 (2017). Bloomberg New Energy Finance, July 2017. https://about.bnef.com/electric-
vehicle-outlook/ Accessed July 2017.
Shepard, S. and Abuelsamid. S. (2016). Electric Vehicle Geographic Forecasts: Battery and Plug-In Hybrid Electric
Vehicle Sales and Populations in North America, https://www.navigantresearch.com/wp-
assets/brochures/DB-EVGEO-16-Executive-Summary.pdf. Accessed July 2017.
32 Ottertail Power, Private Conversation, May 2017.
Hledik, R. et al. (2016) The Hidden Battery: Opportunities in Electric Water Heating. Technical Report.
Kaluza, S. et al. (2016) BMW i ChargeForward: PG&E’s Electric Vehicle Smart Charging Pilot. BMW Group, Pacific
Gas and Electric Company, 2016. http://www.pgecurrents.com/wp-content/uploads/2017/06/PGE-BMW-
iChargeForward-Final-Report.pdf Accessed July 2017.
Solar Potential Analysis Report | November 15, 2018 | Page 52
Figure 21: Forecasted market penetration of controlled DHW units.
Figure 22: Forecasted market penetration of controlled EV units.
Solar Potential Analysis Report | November 15, 2018 | Page 53
Figure 23: Forecasted market penetration of controlled residential heating units.
The number of participating units and their aggregate load are provided in Table 10 for the 2025
and 2050 timeframes, Low and High Technology Development scenarios. Each of these loads
represents new electrification beyond load growth estimates, excluding the 2025 DHW scenario
which represents the current installed base of controllable DHW units.
Table 10: Number of participating units and their aggregate load for the SPA scenarios.
2025
Low
High
Load
Units
Load (GWh)
Units
Load (GWh)
Res. HVAC
289,000
500
433,000
1,450
Res. DHW
140,000
600
330,000
800
EV
32,250
200
110,000
600
2050
Low
High
Load
Units
Load (GWh)
Units
Load (GWh)
Res. HVAC
789,000
3,500
1,579,000
15,500
Res. DHW
600,000
2,200
2,807,000
9,800
EV
349,000 2,100
4,930,000 29,600
Solar Potential Analysis Report | November 15, 2018 | Page 54
Domestic Hot Water Analysis
Summary:
Electric resistance water heaters (ERWH) and heat pump water heaters (HPWH) convert
electrical energy into heated potable water for domestic hot water (DHW) consumption. These
systems are typically coupled with storage tanks to meet regular hot water consumption
requirements for 4 to 24 hours without drawing electrical power. Storage times and capacity
can be increased with the use of mixing values and higher storage temperatures. When coupled
with remote controllers, these units can be turned on and off at coordinated times to increase
or decrease the load. This flexibility enables load shifting of the controlled DHW loads to
accommodate periods of low or high solar or wind generation.
This section outlines the approach taken for estimating the load shifting potential of DHW units
in the SPA model.
1. Market Penetration The market penetrations of controllable DHW units is divided into
two groups. The first market considers the existing electrical DHW units. These loads
are already part of the SPA load model and load shifting control is enabled on some
portion (market penetration). The second group considers new electrical DHW units, for
example those units that may switch over from liquefied petroleum gas (LPG) or natural
gas fuels under strategic electrification scenarios. These two markets must be
accounted for separately because they represent existing and new electrical loads,
respectively.
2. Opportunity ERWH systems have standardized around uniform capacities (50-80 Gal)
and loads (<4.5kW), especially those units that have been used in existing demand
response programs. Furthermore, their utility for load shifting is substantially increased
by including mixing valves, such that the tanks can be charged to higher capacity (135
160 F), yet only discharge 120 F water. The load shifting opportunity reflected by DHW
is comprised of different unit types and charging temperatures depending on the
Technology Development scenario.
3. Constraints Much research has been performed on DHW consumption behavior to
produce estimates of average DHW load. The assumption used here is that these
existing profiles represent, on average for the load shifting DHW units, the minimum
amount of energy that need be available to meet hot water requirements. In other
words, the controlled DHW units must have the same capacity available on an hourly
basis as the uncontrolled units. Additionally, the controlled units have a maximum
capacity (when they are full of hot water at set point) and a maximum load (all turned
on at the same time).
The technical details of how these units are controlled are beyond the SPA model. Nonetheless,
there currently exist smart water heater controllers and demand response controllers that can
Solar Potential Analysis Report | November 15, 2018 | Page 55
provide similar functionality at relatively low cost. The incremental cost of compatible units is
assumed insignificant in the model.
Key Assumptions:
An average DHW profile is sufficient to describe the minimum required energy
necessary to meet residential hot water requirements for a large fleet of units.
Strategic electrification initiatives will add new electric load by converting natural gas
and LPG DHW units to electric heat pump DHW units.
Example Scenario
The 2025 Low Technology Development scenario provides a conservative estimate of the
number of units available for load shifting. It assumes current rates of participation in existing
DHW demand response programs in Minnesota under modest load growth, yielding about
140,000 units. The exact distribution of unit types is not known, so they are conservatively
assumed to be 50 Gal ERWH. Coupled with earlier assumptions and data sources, these values
sufficiently outline the extent of operation to which the fleet of units can be turned on or off to
accommodate solar energy while meeting consumer requirements. An example of this
operational space and load shifting flexibility is given in Figure 24.
Figure 24: Example scenario of working range of DHW load shifting.
High Level Calculation Steps
Maximum Flexible Load
   ×    ()=    ()
The load for these units can range between 0 and the maximum flexible load.
0
200
400
600
800
1000
1200
1400
1600
1800
2000
0
100
200
300
400
500
600
700
800
900
1000
0 5 10 15 20
DHW Capacity (MWh)
DHW Power (MW)
Hour of day
Non-adjust MW
Max Rate MW
Controlled Load MW
Min Capacity MWh
Max Capacity MWh
Solar Potential Analysis Report | November 15, 2018 | Page 56
Minimum Storage Capacity
    (/ℎ)×   ×
=   ()
The minimum storage capacity is based on the assumption that hot water needs are met, on
average, if the controlled DHW units have the capacity necessary to match the uncontrolled
(regular) DHW load.
Maximum Storage Capacity
    ()×  
=   ()
The maximum storage capacity is the fleet-wide sum of the energy each tank can hold.
Solar Potential Analysis Report | November 15, 2018 | Page 57
Electric Vehicles
Summary
Electric vehicles (EVs) currently comprise a negligible amount of electrical load in Minnesota,
but they are anticipated to grow at exponential rates through the timeframe of the SPA model.
Given that the magnitude of transportation energy consumption is of the same order as
building energy, it is assumed EV charging loads will be significant in the timeframe of this
study. Most vehicles are used to commute between work and home and the energy
requirements of these commutes are typically only a fraction of an EV’s energy storage
capacity. Therefore, EV chargers can be turned on and off at coordinated times to increase or
decrease electric load. This flexibility enables load shifting to accommodate solar energy
production goals. In the SPA, only the time and the rate at which the EV’s charge is considered
flexible. In other words, EVs are not considered to be true Vehicle-to-Grid (V2G) resources and
therefore cannot directly replace the need for grid-connected storage on a 1:1 basis.
This section outlines the approach taken for estimating the load shifting potential of electric
vehicles in the SPA model.
1. Market Penetration The market penetrations of controllable EVs were based on their
projected growth between now and the SPA baseline years of 2025 and 2050. There are
multiple published growth rates that were used to develop high and low estimates for
the adoption of electric vehicles in the state of Minnesota.
2. Technological Performance Currently EVs come with a wide variety of battery sizes
and potential charging rates. It was assumed that these capacity and charging rates will
increase toward values currently represented at the high end of the EV market. In order
to not over-bias the results towards the high-end, vehicle efficiency (miles per kilowatt
hour) is set to the weighted average of EVs available in the market today.
3. BehaviorAn agent-based model was created that randomly samples behavioral
profiles from the 2009 National Household Travel Survey (NHTS)33. Roughly 3500 agents
were selected, each with their own unique travel behavior between home, work and
“other”. The constraining requirement is that each agent must guarantee meeting their
travel requirements but is flexible in terms of when, at what magnitude and for how
long they charge their vehicles.
Key Assumptions
Charging is available at home, work and ‘other’
33 National Household Travel Survey. (2009) Trip Chaining Dataset, U.S. Department of Transportation, Federal
Highway Administration. https://nhts.ornl.gov/download.shtml. Accessed December 2017
Solar Potential Analysis Report | November 15, 2018 | Page 58
Charging is not available when driving between any of the aforementioned locations
Two unique sets of scenarios were run: (1) using 100% saturation of Level 1 chargers
(charge rate up to 1.9 kW at 120 V) and (2) 100% saturation of Level 2 chargers (charge
rate up to 19.2 kW at 240 V), respectively
Agent based model with 3500 agents selected at random from the 2009 NHTS
Before shifting, agents are assumed to charge their batteries until full as soon as they
stop driving
Battery size for each agent is assumed to be 60 kWh
EV performance for each agent is assumed to be 3.27 miles/kWh (the weighted average
rate based on current EV market share)
Average annual mileage of 14,600 miles (the mean extracted from the NHTS)
100% of vehicles are assumed to participate in a utility-controlled EV charging load
shifting program
Charging time and magnitude is shifted based on the magnitude of surplus renewable
generation
Baseline Charging Behavior for an Individual Agent and on the Aggregate
Figure 25 illustrates driving behavior for Agent 20004480:2 in the NHTS (a typical agent). The
agent leaves home in the morning, heads to work, drives to lunch, and then drives home in the
evening. This agent is available to charge when they are not driving (the blue sections).
Figure 25. Driving Behavior for Agent 20004480:2 in the NHTS.
Solar Potential Analysis Report | November 15, 2018 | Page 59
The nominal state of charge of the agent’s battery is shown in Figure 26 and the agent’s charge-
discharge profile is shown in Figure 27 (blue: charging, red: discharging). One can see that this
agent is able to quickly recharge their battery each time they stop driving.
Figure 26. EV Battery State of Charge Associated with Agent 20004480:2.
Figure 27. EV Battery Charge/Discharge Profile Associated with Agent 20004480:2.
The aggregate EV load impact in Minnesota is determined by repeating this analysis for each of
the 3500 agents selected from the NHTS, aggregating the results, and scaling the aggregate
results to account for the forecasted EV adoption.
For the 2025 timeframe, 32,250 EVs are expected in Minnesota under the Low Technology
Development scenario. The load impact from these EVs is shown in Figure 28. In the Low
Technology Development scenario EVs are expected to add 50 MW of load in the late afternoon
and early evening.
Solar Potential Analysis Report | November 15, 2018 | Page 60
Figure 28. 2025 EV Load Impact for the Low Technology Development scenarios.
Load Shifting
Agent 20004480:2 is now asked to charge their EV in a manner that is proportional to the
amount of renewable production surplus, as shown in Figure 29 for a day with excess solar
production. The blue area of the subplot (indicating when and how much this agent is charging
their EV) shifts to hours with solar production. Note, it does not exactly match the shape of the
solar surplus (not shown) due to the fact that the agent briefly uses their vehicle in the middle
of the day. A plot of the battery state of charge (Figure 30) reveals that in order to consume
excess solar production during the middle of the day, the EV battery must start and end the day
in a partially discharged state (for this agent on this particular day).
Figure 29. EV Battery Charge/Discharge Profile with Load Shifting (Agent 20004480:2).
Solar Potential Analysis Report | November 15, 2018 | Page 61
Figure 30. EV Battery State of Charge with Load Shifting (Agent 20004480:2).
Figure 31 shows the aggregate load impact with and without load shifting for a sequence of
days in the summer for a 100% solar scenario (a 100% solar scenario is used to provide visual
simplicity). Each agent attempts to shift charging to hours with excess solar production. This
reduces the amount of solar energy that is curtailed and also reduces the load that solar needs
to serve in the late afternoon and early evening. At the same time, this increases the peak load
associated with EV charging from 50 MW to roughly 80 MW (for the low EV adoption scenario
in the 2025 timeframe).
Figure 31. Aggregate EV Load Impact with and without Load Shifting.
Solar Potential Analysis Report | November 15, 2018 | Page 62
Heating, Ventilation and Air Conditioning Summary
Heating, Ventilation, and Air Conditioning (HVAC) systems are the largest energy loads in
Minnesota buildings. Today heating loads are mainly served using natural gas, but
improvements to cold climate heat pump technology and decreasing grid emissions factors
increase the potential for the widespread adoption of high efficiency heating under strategic
electrification initiatives. Heat pumps consume electricity in order to transfer heat from inside
to outside (air conditioning) or from outside to inside (heat pumping). Modern heat pumps
have the ability to run at variable speed such that they can be ramped up or down to meet the
load. Combined with occupant flexibility, these loads can be modulated by changing the set
point or the dead band temperature range. In this way, these units can be used to pre-
condition spaces or allow space temperatures to float beyond the set point. When coupled with
remote controllers, this control increases or decreases the current energy load and shifts it to
another time. This flexibility enables load shifting of the controlled HVAC loads to
accommodate solar energy production goals.
Heat pump-based residential and commercial cooling systems for large retail, medium office,
and K-12 schools were initially tested for load shifting capabilities but neglected in the final
model.
This section outlines the approach taken for estimating the load shifting potential of residential
& commercial HVAC units in the SPA model and the basis for not including them in the final
results.
1. Market Penetration The market penetrations of controllable HVAC units is divided into
two groups. For commercial systems, the existing air conditioning systems are
considered for the following DOE reference building types: medium office, school, and
retail. These loads are already part of the SPA load model and load shifting control is
enabled on some portion (market penetration). The second group considers new
residential electrical HVAC units or heat pumps, for example those units that may
switch over from natural gas heating under strategic electrification scenarios. The
difference between the two groups is seasonal; existing air conditioning systems can be
leveraged in cooling season and future heat pump systems can be leveraged year-round.
2. Opportunity Heat pump systems come in a variety of sizes and more importantly
residential single-family construction has a large variety of heating and cooling loads.
Nonetheless, there exists an average home that when multiplied by the total number of
single-family dwellings yields a representative net HVAC load for the purposes of the
SPA model. Residential utility energy consumption, average house size, and typical
thermal capacity of stick- frame construction are sufficient to characterize this average
home. Assuming heat pump systems are properly sized, this representative home
approximately requires a 4 Ton heating and 2.5 Ton cooling system.
Solar Potential Analysis Report | November 15, 2018 | Page 63
Commercial buildings have more variation in thermal properties and cooling system
types compared to residential buildings. To simulate commercial loads that may become
available for load shifting, four types of DOE reference buildings are used. The reference
buildings are run through an EnergyPlus simulation using SPA weather data from Clean
Power Research. From these simulations, thermal properties and cooling system sizes
are extracted to construct the load shifting model. These per-building data are then
aggregated across the state by scaling the loads to the average building size for the
reference building type and estimated number of each building type.
3. Constraints HVAC loads are constrained based on the need to properly condition a
space. The assumption used here is that occupants will tolerate changes in set points
and dead bands depending on occupied state: 1) home, 2) away or 3) night. The
minimum amount of energy that need be available to meet space conditioning
requirements changes based on occupancy and weather. Additionally, the controlled
units have a maximum capacity (furthest beyond the set point) and a maximum load
(all HVAC systems activated simultaneously). These dead band temperature limits and
schedules vary for each of the building types (residential single family, medium office,
school, and retail).
HVAC loads are likely to require some real time weather forecasting to properly plan HVAC
consumption with respect to renewable energy production. The technical details of how these
units are controlled are beyond the SPA model. Nonetheless, there currently exist smart
thermostats and demand response programs that can provide similar functionality at relatively
low cost. The incremental cost of compatible units is assumed insignificant in the model.
Key Residential Assumptions:
An average residential home with respect to size, heating and cooling loads, and
thermal capacitance and the incidence of participating homes is sufficient to estimate
the thermal inertia that can be leveraged for load shifting.
Strategic electrification initiatives will add new electric load by converting natural gas
and LPG space heating units to heat pump DHW units.
Occupancy schedules are assumed from average commuting data (residential) or DOE
reference building schedules (commercial). The dead bands are given in Table 11.
Solar Potential Analysis Report | November 15, 2018 | Page 64
Table 11. Temperature Dead-bands for HVAC modeling.
Heat Residential Medium
Office
Retail School
Occupied
±2
n/a
n/a
n/a
Unoccupied
±4
n/a
n/a
n/a
Night
+4 -8
n/a
n/a
n/a
Cool
Residential
Medium
Office
Retail School
Occupied
+2 -4
+2 -7
±2
+1 -3
Unoccupied
+6 -6
+5 -7
±4
+5 -6
Night
+6 -6
+5 -7
+10 -6
+10 -6
Example Scenario
The 2025 Low Technology Development scenario provides a conservative estimate of the
number of existing air conditioning systems that will be available for load shifting. It does not
assume any kind of strategic electrification program. It assumes 20% of eligible single-family
homes will participate, yielding about 290,000 units. Unlike other the other shift-able loads,
HVAC shifting strongly depends on the weather. More specifically, the size of the heating or
cooling load compared to the size of the equipment.
High Level Calculation Steps
The calculation treats the fleet HVAC systems and buildings as a thermal battery with limits that
change depending on occupancy. The first step is to determine the heating (or cooling load) on
the building. This is followed by a calculation of the thermal capacity of the battery, which
changes depending on the season, time of year, and occupancy as it floats with respect to
outdoor conditions. The minimum power for the fleet of HVAC systems is that necessary to
maintain the bottom of the dead band range in the given time step (i.e. low temperature in
heating or high temperature in cooling). The maximum power for the fleet of HVAC systems is
that necessary to maintain the top of the dead band range in the given time step (i.e. high
temperature in heating or low temperature in cooling). The SPA model can manipulate the total
power of the system within this range and shape the demand curve in response to solar and
wind energy production. The effects of the current power draw are carried forward into the
next time step with potentially new HVAC loads, new thermal battery capacities, and new
generation constraints.
The following figure shows an example of residential load shifting during a two-week period in
July defined by TMY3 weather for Minneapolis, MN. In this period there are heating and cooling
loads. The yellow area shows how the capacity of the thermal battery (MWh) changes in
response to the type of load, occupancy, and day and night periods. The black line is the
aggregated power of the controlled homes at the grid level. In this plot, the grid is randomly
Solar Potential Analysis Report | November 15, 2018 | Page 65
requesting zero power, maximum power, or the nominal power required to maintain the
default set point, in order to demonstrate possible variations.
Figure 32. Demonstration of HVAC load shifting capabilities. The black line represents shifted load.
(Unshifted load is not shown). The yellow region shows the capacity (MWh) of the thermal battery
change as thermostat set points, dead bands temperatures, and outside air temperatures vary over
time. The black line is the power (MW) input into the thermal battery as a function of its capacity,
heating and cooling requirements, and grid demands.
A better understanding of the model is available by looking at the average value of indoor
temperature for the thermal battery during this period as seen in Figure 33 (next page). The
available thermal battery capacity is represented by t