cdi AVERT
AVoided Emissions and geneRation Tool
www.epa.qov/avert
epa.gov/avert
AVoided Emissions and geneRation Tool
(AVERT)
User Manual
Version 2.3
May 2019
U.S. Environmental Protection Agency
Office of Air and Radiation
Climate Protection Partnerships Division
&EPA
-------
Acknowledgments
AVERT was developed by Synapse Energy Economics, Inc., under contract to EPA's
Climate Protection Partnerships Division and under the direction of Robyn DeYoung of
EPA. EPA thanks the staff at Synapse who developed AVERT and the user manual,
particularly Jeremy Fisher, Ph.D., Patrick Knight, Ariel Horowitz, Ph.D., Avi Allison, Jamie
Hall, Elizabeth A. Stanton, Ph.D., and Bruce Biewald. Eastern Research Group, Inc. (ERG)
provided production and logistical support and developed the web-based version of
AVERT, with thanks to Chris Lamie, Courtney Myers, Bryan Neva, Emma Borjigin-Wang,
and several ERG colleagues. ERG and Synapse provided these services under EPA
contracts #EP-BPA-12-H-0025 and #EP-BPA-12-H0036. EPA also thanks Abt Associates
staff Kait Siegel and Jonathan Dorn, Ph.D., and Anna Belova, Ph.D., currently with
Cognistx, who provided support for developing the PM2 5 emission rates under contracts
#EP-W-11-003 and #EP-W-17-009.
-------
kI/VeI
AVERT User Manual
Version 2.3
Contents
1. INTRODUCTION 1
The Challenge of Estimating Displaced Emissions 2
Basic Method: eGRID Non-baseload Method 3
Basic Method: Capacity Factor Approach 4
Intermediate Method: Historical Hourly Method 4
Sophisticated Method: Energy Modeling 4
Using AVERT 5
Example Use A: Air Quality Planner Quantification of EE/RE Emissions Impacts for
SIP Compliance 6
Example Use B: Stakeholder Review of Multiple EE Options for Emissions Impacts 6
Cautionary Note 6
Benefits of Using AVERT 7
Key Abbreviations 9
2. THE AVERT ANALYSIS STRUCTURE 10
Limitations and Caveats 12
3. AVERT MAIN MODULE: AN OVERVIEW 15
AVERT Regions 15
EE/RE Program Characteristics 17
Displaced Emissions Output 20
4. AVERT MAIN MODULE: STEP-BY-STEP INSTRUCTIONS 23
AVERT Welcome Page 23
Step 1: Load Regional Data File 25
Step 2: Set Energy Efficiency and Renewable Energy Data 27
Manual User Input 29
Reduce Generation by a Percent in Some or All Hours 30
Reduce Generation by Annual GWh 31
Reduce Each Hour by Constant MW 32
Renewable Energy Proxy 32
Step 3: Run Displacement 32
Step 4: Display Outputs 35
Summary Tables 35
Charts and Figures 39
COBRA Text File 44
SMOKE Text File 45
&EPA
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Advanced Outputs 45
APPENDIX A: INSTALLATION INSTRUCTIONS 46
Main Module 46
System Requirements 46
Installation 46
Launching AVERT's Main Module 47
Technical Assistance 47
Statistical Module 47
System Requirements 47
Installation and Launching 47
Technical Assistance 47
Future Year Scenario Template 48
System Requirements 48
Installation 48
Launching and Working with the Future Year Scenario Template 48
Technical Assistance 48
APPENDIX B: DATA 49
APPENDIX C: PROXY RENEWABLE ENERGY HOURLY PROFILES 53
APPENDIX D: OVERVIEW OF AVERT'S STATISTICAL MODULE 55
Parsing Generation Demand into Fossil-Fuel Load Bins 55
Collecting Statistical Information 58
Frequency of Operation by Fossil-Fuel Load Bin 58
Generation Level by Fossil-Fuel Load Bin 58
Heat Input and Emissions by Generation Level 61
Extrapolation to Higher and Lower Fossil-Fuel Loads 61
Extrapolating the Probability of Operation 62
Extrapolating the Generation Level 63
Statistical Analysis 65
Statistical Output 69
APPENDIX E: AVERT'S STATISTICAL MODULE: STEP-BY-STEP INSTRUCTIONS 70
Step 1: Determine Windows Operating Environment 70
Step 2: Download the Statistical Module Executable 70
Step 3: Download CAMD Database 71
Step 4: Install MATLAB Compiler Runtime (MCR) 71
Step 5: If Desired, Complete a Future Year Scenario Template 71
Step 6: Launch the AVERT Executable 71
Step 7: Choose a Base Year for Analysis 73
FPA
State and Local Energy * »
and Environment Program
-------
AVERT User Manual
Version 2.3
Step 8: Choose a Base- or Future Year Scenario 73
Step 9: Choose Region(s) of Interest 74
APPENDIX F: AVERT'S FUTURE YEAR SCENARIO TEMPLATE 75
Retirement of Existing EGUs 75
Addition of Proxy EGUs 75
Pollution-Control Retrofits 76
Running Future Year Scenarios in AVERT 77
APPENDIX G: AVERT REGIONS AND INSTRUCTIONS FOR STATES THAT CROSS
REGIONAL BOUNDARIES 78
APPENDIX H: FREQUENTLY ASKED QUESTIONS 81
Web-Based AVERT 81
Renewable Energy and Energy Efficiency 81
AVERT Outputs 86
AVERT Statistical Module 87
Future Year Scenarios 90
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
1. Introduction
The U.S. Environmental Protection Agency (EPA) recognizes that many states are adopting,
implementing, and expanding cost-effective energy efficiency (EE) and renewable energy (RE)
policies and programs. States are investing in EE/RE policies and programs to achieve benefits
including lowered customer costs, improved electric supply reliability, and diversified energy supply
portfolios. EE/RE can also reduce pollution of criteria air pollutants and greenhouse gases,
especially on high-electricity-demand days that typically coincide with poor air quality.
EPA's Roadmap for Incorporating Energy Efficiency/Renewable Energy Policies and Programs in
State and Tribal Implementation Plans describes basic, intermediate, and sophisticated methods
for quantifying the emissions impacts of EE/RE programs.1 Basic methods entail a simple
calculation: multiplying the amount of generation or electricity consumption displaced by the EE/RE
policy or program by the "non-baseload" emissions rate indicated for a specific pollutant in a region
(e.g., eGRID subregion or electricity market area). Intermediate methods, like the AVERT tool
described in this manual, are transparent, credible, free, and accessible tools. Sophisticated
methods cannot be implemented without a detailed understanding of the electricity grid and electric
generator dispatch dynamics, and/or energy modeling expertise. EPA is committed to helping state
air quality planners calculate the emissions benefits of EE/RE policies and program so that these
emissions reductions can be incorporated in Clean Air Act plans to meet National Ambient Air
Quality Standards (NAAQS) and other clean air goals.
AVERT works by estimating the "displaced generation" from EE/RE programs—that is, the
generation at fossil fuel power plants that will not take place because EE or RE is meeting
consumers' energy needs. The quantification of EE/RE programs' "displaced emissions"—or
emissions that would have been created by the generation that has been displaced—is unlike
direct measurement of emissions at an electric generator's smokestack from stack controls.
Emissions reductions from stack controls can be measured directly; emissions reductions due to
demand reductions are indirect, or derived from model scenarios.
Electricity from numerous electric generators is dispatched onto a "grid" that immediately responds
to changes in demand for power from residential, commercial, and industrial customers throughout
a broad geographic area. Reducing consumption through EE/RE programs makes some fossil fuel
generation unnecessary. Electric generating units (EGUs) that are not required to generate
electricity are not "dispatched" as often, so some of their emissions are avoided.
Specific EE/RE programs and technologies have hourly load2 profiles (hour-by-hour schedules of
expected reductions in electricity demand or increases in electricity production for a year).
Understanding the hour-by-hour relationship between specific EE/RE programs and the dispatch of
fossil fuel EGUs (that is, which power plants are called on to generate electricity in a given hour) is
essential to the estimation of the magnitude and location of emissions reductions from EE/RE.
EPA has developed a credible, free, user-friendly, and accessible tool to estimate emissions
impacts of EE/RE policies and programs so that air quality planners can incorporate those impacts
1 See Appendix I at https://www.epa.aov/enerav-efficiencv-and-renewable-enerav-sips-and-tips/enerav-
efficiencvrenewable-enerav-roadmap.
2 "Load" is the term used throughout this manual to describe regional demand for electricity.
&EPA
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
into their NAAQS SIPs.3 The AVoided Emissions and geneRation Tool (AVERT) quantifies the
displaced emissions of sulfur dioxide (SO2), nitrogen oxides (NOx), carbon dioxide (CO2), and
particulate matter with diameter of 2.5 microns or less (PM2 5) associated with EE/RE policies and
programs within the contiguous United States (Alaska, Hawaii, and U.S. Territories are not
modeled). AVERT captures the actual historical behavior of EGUs' operation on an hourly basis to
predict how EGUs will operate with additional EE/RE implemented.
AVERT users can analyze how different types of EE programs, as well as wind, solar, and other
renewable energy technologies, affect the magnitude and location—at the county, state, and
regional level—of emissions. AVERT is a flexible modeling framework with a simple user interface
designed specifically to meet the needs of state air quality planners and other interested
stakeholders.
The Challenge of Estimating Displaced Emissions
Estimating the location of displaced generation and its associated emissions impacts presents
several challenges:
• The balance of electricity supply and demand varies hour by hour and by season.
• Multiple EGUs are dispatched to supply demand for electricity over a broad region.
• Different EE/RE programs and technologies save or generate energy at varying times
throughout the day and seasonally.
Within each region across the country, system operators decide when, how, and in what order to
dispatch generation from each power plant in response to a) customer demand for electricity in
each moment and b) the variable cost of production at each plant.4 Electricity from the power plants
that are least expensive to operate is dispatched first, and the most expensive plants are
dispatched last. That is, given a cohort of EGUs, the lowest variable-cost units are brought online
first; as the load increases through peak (high-demand) hours, increasingly expensive units are
brought online. (Ideally, given no other constraints—e.g., transmission, voltage support, ramp
rates, maintenance outages—EGUs will dispatch into an electric system in a regular economic
order based on the cost of fuel, the units' heat rates, and other variable costs of production.) In this
"economic dispatch" decision-making process, EE/RE resources generally have low variable costs
or are considered "must-take" resources, the operation of which is determined by sun, wind, the
flow of a river, or efficiency program designs.5 EE/RE resources displace higher variable-cost,
emission-producing fossil-fuel generation. While electricity planners typically think about a single
"marginal" resource (that is, the highest-cost unit that is required to meet customer demand at any
See Appendix I of the Roadmap for Incorporating Energy Efficiency/Renewable Energy Policies and Programs
fhttps://www.epa.qov/enerqv-efficiencv-and-renewable-enerqv-sips-and-tips/enerqv-efficiencvrenewable-enerqy-
roadmap) for details on how this approach can be used in the different NAAQS SIP pathways.
"Variable costs" are the costs realized in the hour-to-hour operation of a generator that vary with the amount of
energy that the unit produces. They typically include the cost of fuel, maintenance costs that scale with output,
and the cost of emissions. Power plants also have "fixed costs"—such as staffing and regular maintenance—that
must be met regardless of whether or not their units are generating power.
"Must-take" resources are so named because they cannot generally be centrally dispatched (i.e., a controller
cannot determine when they provide power), and as such they must be taken when they provide power. These
resources can be curtailed under unusual circumstances, such as during periods of excess supply. These periods
are not the norm; for the most part, controllers operate dispatchable resources around the must-take resources.
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
particular time), there are often several EGUs with similar variable costs that are simultaneously on
or near the margin. EE load reductions and low-variable-cost RE generation can change the level
of generation dispatched at multiple marginal EGUs at the same time across state boundaries.
Different EE/RE resources displace generation in different hours or seasons. EE load impact
profiles describe the hourly changes in customer demand resulting from a program or measure, or
the combined impact of a set of programs or measures. For example, load impact profiles can
represent a portfolio of programs used to meet a policy target, such as the Energy Efficiency
Resource Standards adopted by 27 states.6 Generation profiles for RE technologies, such as wind
or solar photovoltaic (PV), also vary by hour and season.
Determining which cohorts of EGUs are most likely to be displaced during particular hours or under
certain conditions is a complex endeavor. It is not possible to definitively predict how an EE/RE
resource will affect any given power plant. There are, however, several ways to estimate which
EGUs would be displaced when and by how much based on EE/RE resources' load impact
profiles, and EGUs' historical operational behavior or projected information on cost and other
factors affecting dispatch in each regional electricity market.
Methods for estimating projected displaced emissions range from basic to sophisticated (see
Figure 1). The first three methods (two basic methods and the AVERT intermediate-complexity
method described in this manual) use historical operations and profiles to estimate likely displaced
emissions. The more sophisticated fourth approach predicts future electricity market conditions and
emissions displacement. Each method attempts to identify the group of EGUs that would be
displaced as a result of
EE/RE programs or
measures that vary in terms
of their sizes, geographic
areas, and timing during the
day and the year.
Basic Method: eGRID
Non-baseload Method
This basic method calculates
the average non-baseload
emissions reductions of clean
energy policies and programs
in one of EPA's "eGRID" subregions.7 Annual electricity generation or sales displaced by an EE/RE
program or measure are multiplied by the "non-baseload" emissions rate for each pollutant in each
eGRID subregion.8 The non-baseload emissions rate for an eGRID subregion is the current
average emissions rate for the EGUs most likely to be displaced by EE/RE.
Figure 1. Emissions quantification methods.
c
eGRID
Non-baseload
Method
Assumptions are Simpler
Historical
Hourly
Method
Methods are More Sophisticated
U.S. EPA Energy and Environment Guide to Action, Chapter 4.1 (2015), available at
https://www.epa.aov/statelocalenerav/enerav-and-environment-auide-action.
eGRID data can be found at https://www.epa.gov/enerav/earid.
Grid loss factors approximates the line losses that occur between the electric generating facilities and the
buildings that purchase the electricity. They should be included in this calculation.
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Basic Method: Capacity Factor Approach
The capacity factor approach estimates emissions reductions for an EGU based on its current
capacity factor (i.e., its production of electricity in the most recent year as a percentage of the
maximum energy that it can produce).9 The capacity factor is used as a proxy for the likelihood that
any given EGU will be displaced by EE/RE. Infrequently dispatched EGUs with low capacity factors
are more likely to be displaced than EGUs with higher capacity factors.
Intermediate Method: Historical Hourly Method
The AVERT method described here uses historical hourly emissions rates based on recent EPA
data on EGUs' hourly generation and emissions reported through EPA's Acid Rain Program.10 This
method couples historical hourly generation and emissions with the hourly load reduction profiles of
EE/RE resources to determine hourly marginal emissions rates and hourly emissions reductions.
AVERT can be used to predict EE/RE-related emissions reductions in a current or near-future
year—though it is based on historical behavior rather than predicted economic behavior and,
therefore, does not use projections of future fuel or electricity market prices.
Sophisticated Method: Energy Modeling
The most sophisticated method, energy modeling, is the use of highly complex simulation models
that predict individual EGU dispatch, commitment, and emissions based on economic dispatch.11
Energy models that simulate unit-by-unit dispatch and attempt to replicate decisions made by
controllers and operators are called "production-cost" models, and will often include operational
and transmission constraints. Operating economic dispatch models require the modification and
validation of extensive input datasets, significant expertise to operate proprietary models, and
ultimately a fairly high cost to evaluate individual EE/RE scenarios.12
9 See, for example, the eCALC model: Texas A&M Energy Systems Laboratory. 2004. Texas Emissions and
Energy Calculator (eCALC). Documented at http://oaktrust.librarv.tamu.edU/handle/1969.1/2079.
10 https://ampd.epa.gov/ampd/.
11 Fisher, J., C. James. N. Hughes, D. White, R. Wilson, and B. Biewald. 2011. Emissions Reductions from
Renewable Energy and Energy Efficiency in California Air Quality Management Districts. Available at http://uc-
ciee.org/downloads/CAEmissionsReductions.pdf.
12 A variety of utility-standard models that are available to estimate the impact of new EE/RE resources on dispatch.
Generally, the models best suited for this purpose in near-term years are termed "production-cost" models,
including such systems as Market Analytics - Zonal Analysis (http://new.abb.com/enterprise-software/enerov-
portfolio-management/market-analvsis/zonal-analvsis), PROMOD IV (http://new.abb.com/enterprise-
software/enerav-portfolio-manaaement/market-analvsis/promod-iv). and PLEXOS
(http://www.energvexemplar.com). Other models are designed to optimize resource build-out (i.e., new capacity
additions), and may be appropriate for examining long-term displacement of specific new resources. These
"capacity expansion" models include such platforms as EGEAS (https://www.epri.com) and System Optimizer and
Strategist (https://new.abb.com/enterprise-software/energv-portfolio-management/commercial-energy-
operations/capacitv-expansion). These models are all proprietary and require either licensure or specific project
contracts to operate for most users. Large-scale, integrated assessment models such as ICF's Integrated
Planning Model (IPM; https://www.epa.gov/airmarkets/clean-air-markets-power-sector-modeling), the National
Renewable Energy Laboratory's Regional Energy Deployment System (ReEDS;
https://www.nrel.gov/analvsis/reeds/) and the U.S. Department of Energy's National Energy Modeling System
(NEMS; https://www.eia.gov/forecasts/aeo/) are appropriate for testing the implications of large-scale policies and
initiatives over longer periods, but use highly simplified representations of electricity dispatch and generally
aggregate units for computational efficiency.
4 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Using AVERT
AVERT is a free tool that allows users with minimal electricity-system expertise to easily evaluate
county-level emissions displaced by EE/RE programs. AVERT is primarily designed to estimate the
impact of new EE/RE programs on emissions from large (greater than 25-megawatt [MW]),
stationary fossil-fired EGUs. It uses public data, is accessible and auditable, and can be used for
quantifying emissions impacts in NAAQS SIPs.13
To estimate displaced emissions using AVERT, users will need to know the type of EE/RE program
or measure to be analyzed, or the program's load impact profile. An annual EE/RE load impact
profile can be presented in 8,760 hourly intervals, or more coarsely in a few intervals (for example,
peak, off-peak, and shoulder periods). For EE policies and programs, users will need the expected
annual load reduction and an understanding of the temporal profile (e.g., would the EE program
save energy throughout the year or primarily during peak periods). For RE programs, users will
need to know the capacity of the solar or wind resource they are analyzing. These annual profiles
are used to identify more precisely what specific generation resources are displaced by EE/RE
programs or measures.14 In the absence of specific data on the load impact, planners need to use
their judgement to approximate the timing of impacts.
Using these inputs, AVERT automatically estimates emissions displacements in a region. The user
then can view various outputs, maps, charts, and tables useful for many different types of analysis.
Users can choose outputs that show regional-, state-, and county-level emissions impacts, with the
option of highlighting high-electric-demand days. Expert air quality modelers assessing the benefits
of PM2 5, NOx, and SO2 emissions impacts can use the SMOKE (Sparse Matrix Operator Kernel
Emissions) output function to produce hourly, EGU-specific air-dispersion-model-ready data.
AVERT users assessing the public health impacts of the criteria pollutant reductions can use the
COBRA (CO-Benefits Risk Assessment) ouput function to produce model-ready county-level
emissions impact data.
AVERT is best suited to analyze the emissions impacts of state-wide or multi-state EE/RE policies
and programs. Since AVERT modeling is conducted in one of 10 large regions that represent
electricity markets and does not account for transmission constraints within each region, this tool is
not recommended for estimating the emissions displaced by small local programs. Smaller
programs can use AVERT-generated emission factors to estimate impacts within an AVERT
region; these factors are available at www.epa.gov/avert. (See Appendix H for more details on
determining the upper and lower bounds for EE/RE program sizes to be modeled in AVERT.) In
addition, the tool is equipped to predict displacement in near-term future years by estimating each
unit's generation, heat input, and emissions in the event that other EGUs are retired, newly brought
online, or retrofitted with pollution controls.
AVERT provides useful information to both expert energy or air quality planners, and interested
stakeholders. In 2018, EPA released a web-based version of the AVERT Main Module on EPA's
website (www.epa.gov/avert). It provides a streamlined interface with much of the same
functionality as the downloadable Excel-based Main Module, without the need to use Excel
13 See EPA's Roadmap for Incorporating Energy Efficiency/Renewable Energy Policies and Programs in SIPs for
details on other regulatory requirements.
14 U.S. EPA. 2010. Assessing the Multiple Benefits of Clean Energy: A Resource for States. Chapter 3, page 64.
5 ipsiFPA
State and Local Energy m m
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
software or upload separate Regional Data Files. The web edition relies on the most recent year of
input data. Refer to Appendix H for a comparison between the Excel- and web-based Main
Modules' functionality and available display outputs.
Example Use A: Air Quality Planner Quantification of EE/RE Emissions Impacts for
SIP Compliance
Air quality planners can use AVERT to quantify the expected emissions of a new wind farm, solar
initiative, or energy efficiency program for the purposes of Regional Haze Rule or NAAQS State
Implementation Plan (SIP/TIP) compliance. For example, planners can evaluate a program that
could displace NOx during the summer ozone season, efforts to bolster a state wind energy
program, or proposed additional incentives for an EE program that targets peak energy usage
(e.g., high-efficiency air conditioner replacement).
Using AVERT, planners can input estimates for the amount of energy a wind farm of a particular
size and output typical of the region can produce or the amount of energy that could be avoided
from an EE program. Among other outputs, AVERT can estimate annual SO2 and PM2 5 emissions
as well as ozone-season NOx emissions reductions or a pounds (Ibs)-per-day 10-day average of
NOx emissions in counties selected, allowing a comparison of the effectiveness of these programs
against other SIP measures. Advanced AVERT users are able to incorporate expected retirements
and changes in emissions rates expected in future years, and establish new baseline conditions.
Because AVERT has the capability to output SMOKE-formatted emissions estimates for each EGU
in the region in each hour of the year, planners can also assess the air quality improvements using
an air dispersion model. Following EPA guidance, this information could be incorporated into a
SIP.15
Example Use B: Stakeholder Review of Multiple EE Options for Emissions Impacts
A second use of AVERT is for stakeholders' development and testing of multiple types of energy
efficiency programs in a state or group of states within an AVERT region to compare potential
reductions in PM2 5, NOx, SO2, or CO2 emissions. Stakeholders can use AVERT to quickly test
different types of energy efficiency load profiles and estimate the resulting displaced emissions.
Users would modify input parameters to simulate baseload, peakload, or total EE portfolio
reductions, or hour-by-hour load reduction profiles. This type of analysis allows stakeholders to
review estimated emissions benefits from a wide variety of programs, to help consider adopting
and/or implementing programs with the greatest improvements to air quality.
Cautionary Note
Please note that AVERT should only be used to assess EE/RE impacts—not to assess changes to
an EGU fleet. For example, AVERT is not equipped to examine the changes in emissions that
result from retirements, changes to heat rates, or specific fuel changes. AVERT uses data based
on historical dispatch patterns and cannot credibly estimate emissions reductions resulting from
changes to the overall pattern of dispatch.
15 For more information on EPA's guidance, refer to EPA's Roadmap for Incorporating Energy Efficiency/Renewable
Energy Policies and Programs in SIPs/TIPs at https://www.epa.qov/enerqv-efficiencv-and-renewable-enerqy-sips-
and-tips/enerav-efficiencvrenewable-enerav-roadmap.
6 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Benefits of Using AVERT
AVERT combines historical hourly generation data with EE/RE load impact profiles, making it
possible for users to:
• Compare the emissions benefits of different types of EE/RE programs or technologies.
• Incorporate EE/RE policies and programs into air quality models; public health impact
tools, such as EPA's Co-Benefits Risk Assessment (COBRA) Screening Model; and SIPs
to demonstrate Clean Air Act compliance.
• Estimate emissions impacts during peak energy demand periods in a historical year or
near-term future year.
For example, wind and solar power have
different hourly and seasonal operational
profiles. AVERT can compare the emissions
impacts between these two RE technologies at
different times of the year. Similarly, various
EE programs have different hourly load
profiles. AVERT can also help users analyze
different EE programs or portfolios of
programs that offer different energy savings
and emissions displacement throughout the
year. Using this information, air quality
planners could, for example, assess which EE
programs provide the greatest air quality
improvement on high ozone days. For smaller
programs, users can use AVERT-generated
emission factors to get a general estimate
within an AVERT region. (See Appendix H for
more details on determining the upper and
lower bounds for EE/RE program sizes to be modeled in AVERT.)
AVERT is driven entirely by historical, publicly available data reported to EPA. It uses statistically
driven "behavior simulation" to estimate near-term future emissions displacement based on the
recorded historical behavior of existing EGUs in the recent past. Using this dataset alone, the
model derives unit generation behaviors (i.e., how these EGUs respond to load requirements),
EGUs that have a must-run designation,16 and forced and maintenance outages. In addition,
AVERT accurately represents the recent historical relationship between unit generation and
emissions, with characteristics such as a decreasing heat rate (i.e., increasing efficiency) at higher
levels of output, higher emissions from EGUs that are just warming up, and seasonally changing
emissions for EGUs with seasonal environmental controls. The derivation of unit behavior and its
application to AVERT is described in detail in Appendix D of this user guide.
Emission Factors from AVERT
EPA has used AVERT to produce marginal
emission factors for each AVERT region
and a weighted average for the nation for
each year from 2007 to 2018. These
emission factors are available at
https://www.epa.qov/statelocalenerqy/avoid
ed-emission-factors-qenerated-avert and
can be used for quick estimates of avoided
emissions, especially for very small EE/RE
programs. They were calculated by
assuming a 0.5% displacement of the
existing regional demand and are divided
into four categories: wind, utility PV,
portfolio EE, and uniform EE.
16 A must-run designation indicates that a unit is required to operate for reliability reasons, and often operate at
minimum levels to maintain the ability to meet higher load in later hours.
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
AVERT has many advantages but requires several simplifying assumptions. Unlike traditional
electricity system simulation dispatch or production cost models, AVERT does not use operating
costs to estimate how and when an EGU dispatches to meet load requirements. As a result, there
are important electric system dynamics that AVERT cannot capture: temporal characteristics (i.e.,
EGU minimum maintenance downtime and ramp-rates), changing economic conditions (i.e., rising
or falling fuel or emissions prices), and explicit relationships between EGUs (i.e., units that
substitute for one another). AVERT should not be used to assess these types of changes in the
electric system or overall dispatch. These limitations are discussed in more depth in the
"Limitations and Caveats" section on page 12 of this manual.
AVERT operates in a basic computer environment and leads users through the process step-by-
step. Detailed instructions for the Excel version of the Main Module can be found in Section 4 of
this guide or EPA's AVERT online tutorial.17 In 2018, EPA launched a simplified web-based version
of the Main Module. Refer to Appendix H for a comparison between the web and Excel versions.
17 AVERT's online tutorial provides video demonstrations and information about how to run AVERT. Visit
https://www.epa.aov/statelocalenerav/avert-tutorial-homepaae.
&EPA
mm
State and Local Energy
and Environment Program
8
-------
AVERT User Manual
Version 2.3
Key Abbreviations
• AEO: Annual Energy Outlook
• AMPD: Air Markets Program Data from CAMD
• AVERT: AVoided Emissions and geneRation Tool
• CAMD: EPA Clean Air Markets Division
• CO2: carbon dioxide
• COBRA: CO-Benefits Risk Assessment Health Impacts Screening and Mapping Tool
• DOE: U.S. Department of Energy
• EE: energy efficiency
• EE/RE: energy efficiency and renewable energy
• EGU: electric generating unit
• EPA: U.S. Environmental Protection Agency
• EWITS: Eastern Wind Integration Transmission Study
• GW: gig a watt
• GWh: gigawatt-hour
• ISO: Independent System Operator
• lb: pound
• MMBtu: million metric British thermal units
• MW: megawatt
• MWh: megawatt-hour
• NOx: nitrogen oxides
• PM2.5: particulate matter with a diameter of 2.5 microns or less
• PV: photovoltaic
• RE: renewable energy
• RTO: Regional Transmission Organization
• SMOKE: Sparse Matrix Operator Kernel Emissions Model
• SO2: sulfur dioxide
• WWSIS: Western Wind and Solar Integration Study
State and Local Energy
and Environment Program
9
-------
AVERT User Manual
Version 2.3
2. The AVERT Analysis Structure
AVERT has three components:
• An Excel- and web-based Main Module allows users to estimate the displaced emissions
likely to result from new EE/RE programs, policies, or projects. The Excel-based version
requires users to select Regional Data Files (RDFs) generated by the Statistical Module to
analyze EE/RE scenarios in reference to either a historical base year or a future year. (See
Sections 3 and 4 for a detailed description of the Main Module.) The web-based version of
the Main Module provides a streamlined interface with much of the same functionality as
the downloadable Excel tool, without the need for Excel software or separate RDFs. The
web version relies on the most recent year of input data and generates a subset of display
outputs of state and county level emission impacts. Refer to Appendix H for a full
comparasion of the Excel- and web-based Main Modules. Except where noted otherwise,
this user manual describes features available in the Excel version.
• The MATLAB®-based Statistical Module performs statistical analysis on historical
generation, heat input, and emissions data collected in the EPA Clean Air Markets
Division's (CAMD's) Air Markets Program Data (AMPD)18 to produce the statistical data
files used by AVERT's Main Module. The Statistical Module is available to users as a
stand-alone executable. (See Appendices D and E for a detailed description of the
Statistical Module.)
• The Excel-based Future Year Scenario Template allows users to modify base-year
AMPD information with specified retirements and additions of power plants, as well as
changes in emissions rates due to pollution controls. This modified data can be input into
AVERT's Statistical Module to produce scenario-specific statistical data files, which are
then fed into the Main Module. (See Appendix F for a detailed description of the Future
Year Scenario Template.)
AVERT analyzes how hourly changes in demand in a user-selected historical base year change
the output of fossil EGU.19 Using detailed hourly data from AMPD, AVERT probabilistically
estimates the operation and output of each EGU in a region based on a region's hourly demand for
fossil-fired generation. This statistical information is used to predict EGUs' likely operation in
response to load impacts from EE/RE resources. Figure 2, below, shows the flow of data from its
source, through various processing tools, to its end-use in the Main Module.
In general, hourly "prepackaged" data from AMPD are input into AVERT's Statistical Module.
Hourly generation, heat input, and emissions of PM2 5, SO2, NOx, and CO2 from each EGU
reporting to AMPD (a requirement for fossil-fuel EGUs 25 MW and greater) are read from monthly
or quarterly files.
18 https://ampd.epa.gov/ampd/.
19 AVERT's "base year" is modeled from recent, detailed hourly generation and emissions data collected by EPA for
each U.S. fossil-fired generator with capacity greater than 25 MW. Base-year data is usually made available in the
second quarter of the following year.
10
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Figure 2. Schematic of AVERT.
Excel workbook
Text files
Future Year Scenario
Template
User interface for
retirements, additions,
and retrofits
Raw Hourly
Generation and
Emissions Data from
Air Markets Program
(AMP) Dataset
Inputs AMP
data, performs
statistical
analysis, outputs
new Regional
Data Files
AVERT
Statistical
Module
User interface
for creating
EE/RE load
curves,
performing
displaced
emissions
analyses, and
creating output
charts and tables
AVERT Main
Module
Contains annual
hourly load data
and unit-level
statistics on
generation and
emissions data
Regional Data
Files
MATLAB code
Text files
Excel workbook or
web-based version
AVERT's Statistical Module can analyze either raw data for a base year or modified data created in
the AVERT's Future Year Scenario Template.20 AVERT's Future Year Scenario Template allows
users to create a scenario for a year in the near-future21 by modifying data representing a historical
year. Users designate existing fossil-fuel EGUs that will no longer be in operation or wili have
different emissions rates as a result of pollution-control retrofits, and add new fossil-fuel EGUs
based on the characteristics of proxy existing EGUs. (See Appendix F for a more detailed
description of AVERT's Future Year Scenario Template.)
After receiving either base- or future year scenario data, AVERT's Statistical Module performs a
statistical analysis of how each EGU responds to variations in regional fossil load, and simulates
the average generation, heat input, and emissions of each EGU across a range of possible load
levels, from zero to the maximum coincident fossil capacity.22 These EGU and load-level-specific
averages are stored in the AVERT Regional Data Files, which are the input files used into AVERT's
Main Module. (See Appendices D and E for a more detailed description of AVERT's Statistical
Module.)
AVERT's Main Module is accessible as an online tool or a downloadable Excel workbook. It allows
users, regardless of their level of electricity modeling expertise, to quickly estimate the displaced
emissions likely to result from EE/RE programs in a chosen year.23 AVERT's Main Module provides
a simple interface that guides users through inputting an EE/RE load impact profile depicting
20 2007 through 2018 data are provided with this version of AVERT. It is expected that data for future years will
continue to be provided as additional data from CAMD are released.
21 It is recommended that future year scenarios be designed for no more than five years forward from the base data
year to account for changing emissions control technologies, changing fuel prices, and retirements and additions
into the system. Caution should be exercised in reviewing future year scenarios to ensure reasonable results in
light of known or expected system changes.
22 Maximum coincident fossil capacity is equivalent to the sum of each and every fossil generator producing its
maximum output in a single hour.
23 The web-based version of AVERT's Main Module only has the most recent data year available and limited display
outputs. Refer to Appendix H for a complete comparison.
State and Local Energy
and Environment Program
11
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
electricity demand in every hour of a year. The user is then prompted to launch automatic
calculations that result in final output tables and charts for one of the 10 AVERT regions and, if
desired, smaller areas within the region. In addition, AVERT's Excel-based Main Module also
outputs SMOKE-formatted data for advanced air modeling applications. Both the Excel and web
versions output COBRA-formatted data for public health modeling applications. (See Sections 3
and 4 for a more detailed description of AVERT's Main Module and instructions for its use.)
Limitations and Caveats
There are several key limitations to the use of AVERT displaced emissions results.
• Snapshot analysis: AVERT provides a representation of the dynamics of electricity
dispatch (i.e., which EGUs are put into operation in which hours) in a historical base year.
However, it does not model changes in dispatch due to transmission resources, fuel prices,
emissions allowances, demand for electricity, or the variable running cost of individual
EGUs.24 The use of AVERT to estimate forward-looking dispatch decisions is made more
difficult when there are changes to the electrical grid (e.g., new transmission resources,
EGU retirements, pollution control retrofits, or new EGUs), commodity prices (such as fuel
or emissions allowances), or operational restrictions (e.g., "reliability must run"
designations, curtailment due to new emissions caps). AVERT characterizes EGU
retirements, pollution control retrofits, and new EGUs in its Future Year Scenario Template,
but the scenarios created are only as good as the user's predictions of future conditions.
• No explicit ramping or cycling: AVERT does not model changes in ramping (periods
when EGUs are changing to a new generation level) and cycling (fluctuating generation
levels) behavior resulting from EE/RE programs, retirements, environmental controls, or
new EGUs.25 AVERT does not capture the changes in the frequency of ramping and
cycling of fossil-fuel EGUs that can result from variability in wind- and sun-powered
generation. In addition, it does not capture the ability of slow- or fast-cycling plants to
respond to hour-to-hour changes in demand.
• Average outcome: AVERT generates an average outcome for each EGU, rather than a
specific and precise trajectory. The default Regional Data Files produce generation and
emissions levels that are averaged across 5,000 hypothetical scenarios of a recent past
year. These levels are the statistically expected outcome, and should not be mistaken for
an assertion of what did happen in a past year or what will happen in a future year.
• Limited resolution for generation: AVERT estimates regional displaced emissions. To
do so it predicts the most likely generation levels for individual EGUs given a particular
regional fossil-fuel load level and the most likely emissions rates for individual EGUs given
a particular generation level. Results at the individual EGU level (and for counties
containing small numbers of EGUs) have very limited "resolution"; the accuracy of the
results is limited at small spatial and temporal scales.
24 For example, new emissions controls may entail additional variable costs incurred by specific units. These
additional costs could affect dispatch, but are not captured by AVERT.
25 Models that do not capture ramping or cycling dynamics are generally referred to as having non-chronological
dispatch—i.e., there is no explicit sense of time or timing built into the model.
12
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
• Limited resolution for small EE/RE programs: Due to the limited resolution of
generation, when focusing on smaller-scale EE/RE programs, AVERT may return a higher
level of "noise" in the displaced emissions results—that is, a greater divergence between
desired reductions in generation and modeled reductions in generation. Small changes
may be swamped by random effects, such as historical non-economic forced outages and
weather events, artifacts in the data, or even random perturbations in the Monte Carlo
analysis. Users are encouraged to use emission factors pre-generated from AVERT for
small-scale projects.26 There is no hard limit on the smallest project that can or should be
reviewed in AVERT, but results for EE/RE programs under several hundred MW in
capacity should be reviewed carefully.27 Appendix H discusses reasonable minimum levels
of demand reduction for the purposes of obtaining useful results from AVERT.
• Limited ability to capture dispatch implications of very large EE/RE programs:
AVERT is designed to review the impact of marginal changes in load requirements. Very
large-scale energy efficiency and renewable energy programs may fundamentally change
the way in which dispatchers operate a system. In particular, there is little precedent in the
United States for understanding how high penetrations of variable renewable resources
(such as wind and solar) impact other EGUs in a system. In some cases, very high
penetrations of these resources may result in patterns that are not often observed in the
historical dataset, such as the curtailment of slower-cycling coal plants, or an increase in
the dispatch of fast-cycling peaking plants to smooth irregularities.28 Appendix H discusses
reasonable maximum levels of demand reduction for the purposes of obtaining useful
results from AVERT.
• Precision of results: AVERT reports results rounded to the nearest 10 units (i.e.,
megawatt-hours [MWh], lbs of PM2 5, SO2, and NOx, or tons29 ofCC>2). In general terms,
users should consider the number of significant figures in their specified MW load
reduction, and limit their use of AVERT results to that number of significant figures.
• Non-communicating regions: AVERT models one region at a time, assuming that each
region generates sufficient electricity to meet its own requirements; imports and exports of
electricity between regions are assumed to stay constant with changing load
requirements). Similarly, displaced emissions are restricted to the confines of the AVERT
region selected for the analysis. The basis of this assumption is that analyses on smaller-
sized regions would risk missing important interdependencies between EGUs across
26 Pre-generated emission factors for each year from 2007 to 2018 are available at
https://www.epa.qov/statelocalenerqv/avoided-emission-factors-qenerated-avert.
27 The smallest size that AVERT can resolve appropriately will vary by region and program, depending on how the
program is distributed across time and the number of units that reduce output in response. AVERT allows users to
review the impact of noise on expected outputs via a post-run diagnostic (discussed under "Signal-to-noise
diagnostic" on page 40). In addition, AVERT rounds reductions from EE/RE programs to the closest 10 units
(MWh, lbs of PM2.5/SO2/NOX, or tons of CO2); very small EE/RE programs will effectively report little or no specific
reductions. Ultimately, the user must judge if the results return adequate information or appear reliable. For more
information, see Appendix H.
28 See Brown, P. 2012. U.S. Renewable Electricity: How Does Wind Generation Impact Competitive Power Markets?
Congressional Research Service. R42818. Available at https://www.fas.org/sqp/crs/misc/R42818.pdf. See also
National Renewable Energy Laboratory. 2016. Eastern Renewable Generation Integration Study. Available at
https://www.nrel. qov/qrid/erqis. html.
29 In AVERT, all references to tons are short tons (2,000 lbs), not metric tons.
13
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
larger, well-integrated regions. Using yet larger regions than those in AVERT, however,
would spread the influence of load reductions too widely, making it difficult to ascribe load
reductions to particular EGUs.
• Unconstrained transmission: AVERT looks at the dynamics of each region as a whole
regardless of transmission constraints.30 The model represents how the regional electricity
system actually operated in the base year given the existing transmission infrastructure,
but is completely insensitive to the physical location within a region of new EE/RE
programs, as well as to the location within a region of retirements, environmental retrofits,
and new EGUs modeled in the Future Year Scenario Template. In contrast, actual
electricity dispatch decisions may be quite sensitive to the specific locations of resources,
including (but not limited to) whether renewable resources are located close to consumers
(at "load center") or in sparsely populated areas.
• Limited capture of individual EGU dynamics: Fossil-fuel EGUs, especially those using
steam cycles, tend to operate at higher efficiencies and with lower emissions rates while in
steady-state operation at or near their maximum output (although NOx emissions in
particular may be exacerbated by high-output operations). The AVERT approach does
account for emissions rates appropriate to different levels of generation (which may be
associated with periods of fast-ramping or cycling by fossil-fuel EGUs), but does not
account for inefficiencies that may be associated with rapid cycling.
30 Transmission is the infrastructure to transport electricity from generators to load centers (i.e., from the source of
generation to electricity consumers). It can be "constrained" when the thermal (or other) limits prevent as much
energy as is needed from moving across wires. When transmission is "bound" under these circumstances,
dispatch begins to reflect local requirements, rather than regional requirements. In other words, generators may
dispatch in a non-economic fashion when transmission is constrained.
State and Local Energy
and Environment Program
14
-------
AVERT User Manual
Version 2.3
3. AVERT Main Module: An Overview
This section bypasses technical details and how-to instructions to give a broad, simplified overview
of key selections made by AVERT users (i.e., the region and other characteristics of the EE/RE
program to be analyzed) and the displaced emissions results available to users. See Section 4 for
detailed, step-by-step instructions and Appendix A for detailed installation instructions. Appendix D
describes, in detail, how AVERT performs its calculations. Appendix H provides a comparison
between the web- and Excel-based versions of AVERT's Main Module.
AVERT's Excel-based Main Module estimates the emissions displaced by EE/RE resources for
every individual fossil-fuel EGU in a region and aggregates these displacements to the county
level.31 The Main Module uses two key pieces of information:
• Expected emissions at every load level—the likely generation level and emissions of all
but the smallest fossil-fuel EGU in a region in a base- or future year scenario (as modeled
in AVERT's Statistical Module and input automatically into the Main Module).
• A change in load level for every hour of the year—a user-created load impact profile
depicting EE/RE-driven changes in the regional fossil-fuel load for every hour of the year.
The Main Module estimates how much each fossil-fuel EGU changes its generation output and
emissions in response to an EE/RE program as compared to the base- or future year scenario
without the program. The difference between emissions resulting from the base- or future year load
curve and the emissions resulting from the same year adjusted to include the load impact profile of
an EE/RE program is the "displaced emissions." The Main Module presents results in summary
tables for quick comparison, and in graphs and maps for rapid visual assessment.
Section 4 presents detailed, step-by-step instructions on the process of identifying a region for
analysis; obtaining and importing data; designing an EE/RE load impact profile; launching the
automatic displaced emissions calculations; and accessing tables, graphs, and maps summarizing
the output. Once an EE/RE load impact profile has been designed, typical processing takes 1 to 10
minutes depending on the size of the region of interest and the processing speed of the computer.
AVERT Regions
Because customers' electricity demand is met jointly by generation resources throughout a region,
emissions displacements from EE/RE programs take place region-wide. All AVERT analysis,
therefore, is conducted at a regional level. For users, designating one of the 10 AVERT regions of
analysis is "Step 1" in using AVERT's Main Module. A map of these regions is shown in Figure 3.
31 Excludes EGUs smaller than 25 MW that do not report to AMPD.
mm
State and Local Energy
and Environment Program
15
&EPA
-------
AVERT User Manual
Version 2.3
Figure 3. Map of AVERT's regions.
Northwest
Jjmi) /
m Midwes
(WMW)
I Rocky
/Mountains ,
L ,RM» Midwest
reat Lakes I Mid
Atlantic (EMW)|
California
(CA)
Twenty-six states and the District of Columbia each fall within a single region, while 22 states are
split between two or more regions each. Table 1 describes each region in detail.
It is generally recommended that air quality managers for states that are split between more than
one AVERT region evaluate the displaced emissions from regions in which 5 percent or more of
their generation is located. Appendix G includes further discussion of the regions and more
complete instructions for users analyzing EE/RE impacts in split states.
Table 1. AVERT regions, abbreviations, and states.
AVERT Region
Abbreviation
Full States
Partial States
Northeast
NE
Connecticut, Massachusetts,
Maine, New Hampshire, New
York, Rhode Island, and
Vermont
New Jersey
Great Lakes / Mid-
Atlantic
EMW
District of Columbia,
Delaware, indiana, and
Maryland
liiinois, Kentucky, Michigan, New Jersey,
Ohio, Pennsylvania, Virginia, Wisconsin, and
West Virginia
Southeast
SE
Alabama, Florida, Georgia,
North Carolina, South
Carolina, and Tennessee
Arkansas, Kentucky, Louisiana, Missouri,
Mississippi, Oklahoma, Texas, Virginia, and
West Virginia
Lower Midwest
SC
Kansas
Arkansas, Louisiana, Missouri, New Mexico,
Oklahoma, and Texas
Upper Midwest
WMW
Iowa, Minnesota, North
Dakota, and Nebraska
Illinois, Missouri, Mississippi, Montana,
South Dakota, and Wisconsin
Rocky Mountains
RM
Colorado
Wyoming
State and Local Energy
and Environment Program
16
&EPA
-------
J^Vert
missions and q«neRaiion Tuol^^
avviM.Hiid.yWduul ™
AVoid«d Emissions.
AVERT User Manual
Version 2.3
AVERT Region
Abbreviation
Full States
Partial States
Texas
TX
Oklahoma and Texas
Southwest
AZNM
Arizona
New Mexico, Nevada, and Texas
Northwest
NW
Idaho, Oregon, and
Washington
Montana, Nevada, Utah, and Wyoming
California
CA
California and Utah
EE/RE Program Characteristics
AVERT users should understand the characteristics of the EE/RE program that they want to
analyze in terms that can be input into the model; this is Step 2 in the AVERT Main Module.
AVERT analyzes the difference in hourly load (regional electricity demand) between a baseline
scenario and an EE/RE scenario. Starting with a baseline schedule of load for every hour of a
reference year (the "load profile"),32 the Main Module guides the user to create or input a load
impact profile to represent their EE/RE program—i.e., the amount of load that will be reduced by
the EE/RE program on an hourly basis. For details and instructions, refer to Step 2 in Section 4.
Designing future year scenarios is an advanced function of AVERT. Users can currently use 2007
through 2018 data as their baseline year.33 Every AVERT user, however, must create or modify a
load impact profile to represent a specific EE/RE program or group of programs. Advanced users
can manually enter or cut and paste a load impact profile (i.e., an expected reduction for every hour
of the year), but many users will use the Main Module's tools to create a load impact profile
representing a set of EE/RE policies or programs.
Users are strongly encouraged to create and adopt their own load impact profiles, representing
EE/RE programs specific to their interest or area of concern. The following load impact profiles are
built into the Main Module:
• Reduction of fossil-fuel generation by a chosen percent in some or all hours. This
option is recommended to represent a mix of energy efficiency programs that target some
or all hours of the year, but preferentially target higher hours with greater demand.
• Reduction of annual fossil-fuel generation by total gigawatt-hours (GWh) or by a
constant MW each hour. This option is recommended to represent a rough approximation
of baseload-only reductions where the total number of GWh reduced over the course of a
year is known and is expected to be equally distributed over all hours of the year.
• Renewable energy proxy. With this option, users can model wind, utility solar, and rooftop
solar resources that are broadly representative of the selected region.
32 Technically, within AVERT, the load profile represents aggregate fossil generation for a region, and not end use
consumption.
33 Future years will be released as data is made available through CAMD.
ffHK! 17 **FPA
State and Local Energy m m
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
• Combination of EE/RE programs. Users can also mix and match from the above options,
as well as combine the pre-set load reduction options with manually entered reductions
Note that this is not a comprehensive set of load reduction options. Rather, these are simple,
illustrative examples.
AVERT combines all of the user's inputs to generate a single EE/RE profile with 8,760 hourly
values.34 This profile feeds into the calculations in the next step. Since the release of version 1.5 of
AVERT's Main Module, AVERT adjusts the EE/RE profile to account for avoided transmission and
distribution line losses associated with certain EE/RE resources: specifically, EE and distributed PV
systems.35 Hourly load reductions associated with each of these resources are adjusted upward by
the following formula:
Adjusted load reduction = User's input / (1 - x),
where x is the regional average line loss percentage. Starting with Main Module version 2.3 in
spring 2019, AVERT uses line loss factors from the Annual Energy Outlook (AEO) published by the
U.S. Department of Energy's Energy Information Administration (EIA).36 AVERT uses the historical
line loss factors that correspond to the year and region of the user's analysis, as shown in Table 2.
Table 2. Transmission and distribution line loss factors used in AVERT.
Eastern
Western
Data Year
Texas
Interconnect
Interconnect
2007
6.89%
7.48%
8.09%
2008
5.09%
7.67%
8.60%
2009
5.03%
7.70%
8.65%
2010
6.53%
6.92%
6.86%
2011
4.62%
7.42%
10.94%
2012
4.62%
7.74%
9.52%
2013
5.63%
7.80%
9.18%
2014
5.41%
7.43%
9.55%
2015
5.89%
7.32%
8.83%
2016
6.53%
7.00%
9.17%
2017
5.60%
7.00%
8.13%
2018
4.83%
6.74%
8.54%
The Eastern Interconnect corresponds to the following AVERT regions: Great Lakes/Mid-Atlantic,
Lower Midwest, Northeast, Southeast, and Upper Midwest. The Western Interconnect corresponds
to the following AVERT regions: California, Northwest, Rocky Mountains, and Southwest. The
Texas region in Table 2 corresponds to the AVERT Texas region.
This adjustment has the effect of increasing displaced fossil load and displaced emissions
associated with EE programs and with distributed RE generation. This adjustment provides more
accurate results. Without this adjustment, AVERT would underestimate emissions. For example,
34 Or 8,784 in the case of leap years.
35 AVERT treats wind and utility-scale PV as centralized resources that still require transmission and distribution to
end-users; thus, while they displace fossil generation, a simple assumption is that they do not avoid any line loss.
36 AEO data can be downloaded from https://www.eia.gov/outlooks/aeo/. Each AEO provides historical data for the
preceding year. The transmission and distribution line loss factors are calculated as ((Net Generation to the Grid +
Net Imports - Total Electricity Sales)/Total Electricity Sales).
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
AVERT without adjustments would assume that 100 MWh of EE or 100 MWh of onsite (distributed)
PV generation in the Eastern Interconnect in 2018 avoids 100 MWh of fossil generation, whereas it
actually avoids approximately 107 MWh of fossil generation after accounting for the additional
power that would have been generated and lost during transmission in order to deliver 100 MWh to
the end-user.
Once a load impact profile has been designed, the user is prompted to launch the automatic
displaced emissions calculations in Step 3 of the AVERT Main Module.
State and Local Energy
and Environment Program
19
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Displaced Emissions Output
Step 4 of AVERT's Main Module reports the difference between the baseline and EE/RE scenario
through the following outputs:
• Summary tables:
o Annual displaced generation and emissions
o Displacement data for top 10 fossil-fuel generation days
o Annual displacement data by county
o Monthly displacement data by county
o Daily NOx displacement data by county
• Charts and figures:
o Displaced generation and emissions map
o Hourly displacements by week
o Monthly displacements by selected geography (region, state, or county)
o Signal-to-noise diagnostic
• COBRA text file generation (for public health impact modeling)
• SMOKE text file generation (for air quality modeling)
For assessing the air quality implications of EE/RE, the location of air emissions reductions
resulting from load reduction programs can be critical. The example shown in Figure 4 represents a
2,000 MWwind program in the Upper Midwest AVERT region compared with 2012 base-year data.
The map displays annual displaced SO2 emissions from specific EGUs as blue circles; larger
circles indicate greater displaced emissions. Where multiple EGUs overlap on the map (i.e.,
multiple units at one plant or several plants close together), the circles appear darker.
Figure 4. Map of annual SO2 emissions reductions from an example wind program in the Upper
Midwest region.
Annual Change in S02 (lbs)
0 1,780,000 lbs
magnitude of a unit's change in generation / emissions.
Circles are semi-transparent; darker areas occur in
regions with overlapping units, Negative changes are
indicated with blue circles; positive changes are indicated
The diameter of each circle indicates the magnitude of an EGU's change in emissions. Circles are semi-
transparent; darker areas occur in regions with overlapping EGUs. Emissions reductions are indicated with
blue circles; increases in emissions are indicated with black-bordered white circles.
20
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Increases in emissions are shown as black-bordered white circles. Increasing emissions may occur
because higher load is programmed into AVERT (e.g., for testing a higher baseline if reviewing the
impact of existing renewable portfolio standards (RPS), or if shifting load from peak to trough
hours), or due to aberrations in the statistical dataset.37
Many users will be interested in displaced emissions results for smaller areas within a region.
Monthly output can be displayed by region, state, or county. Figure 5 displays monthly SO2
emissions displaced in a single county in Illinois, continuing the same example shown in Figure 4.
Figure 5. Monthly SO2 emissions reductions for Madison County, Illinois, from an example wind
program in the Upper Midwest region.
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
I 11111111111
-30,000 ® ™ |
-100,000
Some users may wish to view hourly displaced generation or emissions to understand the behavior
of the system at a finer scale or during specific hours of the year. Using the same 2,000 MW of
wind reduction program, Figure 6 displays the hour-by-hour fossil generation displaced in the week
of August 1-7 in the Upper Midwest AVERT region from the wind project. Individual EGUs are
color-coded along a gradient from dark blue (baseload EGUs) to light blue (peaking EGUs).38 The
EGUs' individual generation reductions are shown in stacked bar plots and the net total contribution
is shown with a yellow line. The yellow line represents the hourly energy displaced by the wind
project. Note that peaking (gas) EGUs are primarily displaced during daytime hours, while
baseload (usually coal) EGUs are displaced during off-peak hours.
37 Some units show increasing generation even as regional load decreases. This usually occurs with baseload units
during trough hours. It is primarily explained by maintenance outages that occur during periods of low demand,
but not necessarily at the lowest demand hour of the year. Statistically, these units appear likely to generate
slightly more at the lowest-load hours than at medium-low-load hours. Therefore, reducing demand from a
medium-low load to a very low load could appear to increase the output of these units.
38 The color coding is illustrative only. It is not portrayed on an absolute scale, and is relative to all units in the region
(i.e., if a region comprises 200 units, they will all be parsed along the gradient from light blue to dark blue. If a
region comprises 1,000 units, they will similarly be parsed along the same gradient). Units are sorted based on
annual capacity factor. Units with uniform generation across high and low load levels are closer to "high-capacity-
factor" behavior, while units with high output at high load levels and none at low load levels are closer to "low-
capacity-factor" behavior.
21
State and Local Energy
and Environment Program
&EPA
-------
J&/&T
missions and qenaRaiion Tool^^
.UllkW.Ulld.uWtl.Ul 9
AVoidod Emissions.
AVERT User Manual
Version 2.3
Figure 6. Hourly generation reductions in the week of August 1 from an example wind program in the
Upper Midwest region.
Change in Generation (MW) in Week of 8/1
i j4l_j
5 -500
E - .000
Negative numbers indicate displaced
generation and emissions.
Total Change in
Generation (MW)
Total fossil-fuel load,
pre-EERE
n
High capacity factor units
Low capacity factor units
Figure 7 displays SO2 emissions displaced in the same week from the same wind project. Note
that almost all of the EGUs portrayed in this figure are dark blue in color; SO2 reductions are
primarily captured from reductions at high-capacity-factor coal EGUs, which are displaced during
off-peak hours.39
Figure 7. Hourly emissions reductions in the week of August 1 from an example wind program in the
Upper Midwest region.
Change in S02 (lbs) in Week of 8/1
A. ..-toilt A-lMUl mlMl. rlL -Ui jL J
•i.J
£> -2,000
n
O -3,000
-. ;
¦6,' A:
7.000
-9,000
Negative numbers indicate displaced
generation and emissions.
Total Change in S02
(lbs)
Total fossil-fuel load,
"pre-EERE
n
High capacity factor units
Low capacity factor units
39 High-capacity-factor baseload units operate during most hours of the year, and are (by definition) the bulk of the
units (or only units) online during offpeak hours, and therefore are the units that will be displaced in offpeak hours.
This analysis indicates that baseload units are displaced during offpeak hours, but not during the daytime. During
the day, gas plants (with no, or negligible, SO2 emissions) are displaced, and therefore do not appear in this
graphic.
22
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
4. AVERT Main Module: Step-by-Step instructions
This section provides step-by-step instructions for using AVERT's Excel-based Main Module to
estimate the emissions displaced by EE/RE resources.
To begin, download two files and save them to your computer:
• The Main Module workbook: "AVERT Main Module.xlsx." Download the workbook at
https://www.epa.aov/statelocalenerav/download-avert.
• The Regional Data File for the region under analysis.
o Default Regional Data Files developed for use by EPA are labeled
"AVERT RDF [DataYear] BaseEPA ([Region]).xlsx"; they can be obtained at
https://www.epa.aov/statelocalenerav/download-avert.
o Regional analyses developed by advanced users using AVERT's Statistical
Module will be saved, by default, in a folder of the Statistical Module titled "AVERT
Output." These files use the following naming convention:
"AVERT RDF [DataYear] [RunName] ([Region]) [RunDateTime].xlsx."
For more detailed installation instructions and model specifications, see Appendix A of this manual.
AVERT Welcome Page
When launched in Excel, the Main Module opens to its "Welcome" page as shown in Figure 8.
Welcome to AVERT's Main Module
AVERT is an EPA tool that quantifies the emission impacts of energy efficiency and
renewable energy policies and programs within the continental United States. Please
refer to the AVERT user manual for details on step-by-step instructions, appropriate
uses and assumptions built into the tool.
NOTE
Please ensure macros are enabled on your computer.
AVERT requires Excel 2Q07 or higher in Windows and Excel 2Q11 or higher on Mac.
AVERT
vvEPA
H Synapse
energy economics, Inc.
AVERT v.2.3
This version accounts for transmission and distribution line loss calculations for EE and
residential solar projects and can estimate PM2 s emissions impacts.
Developed by Synapse Energy Economics, Inc., May 2019
Use the blue entry to
Editor:
Date edited:
Edition name:
Edition description:
describe each scenario and keep track of multiple versions of AVERT
Click here to
begin
Click here to restore
default Excel
functionality
Figure 8. AVERT Main Module "Welcome" page.
State and Local Energy
and Environment Program
23
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
The Main Module is primarily driven by macros to conserve memory and processing time.40 Before
making any selections or beginning calculations, macros must be enabled on your computer. This
must be done first, before any other steps; if macros are not enabled before you load a Regional
Data File (Step 1), you will need to close the workbook, re-open it, and then enable macros to
continue.
To enable macros in Excel 2007 for Windows:
(Sl
1. Click the Microsoft Office Button J , then click "Excel Options."
2. Click "Trust Center," click "Trust Center Settings," then click "Macro Settings."
3. Click on "Enable all macros."
To enable macros in Excel 2010 or later for Windows:
1. Click the "File" menu (Office Backstage), then click "Options" in the left sidebar.
2. Click "Trust Center" in the left sidebar, then click the "Trust Center Settings" button in the
main window.
3. Click "Macro Settings" in the left sidebar.
4. Choose the "Enable all macros" option and hit "OK."
To enable macros in Excel 2011 or later for Mac, select "Enable macros" in the dialog box that
appears when opening the file.
Next, we recommend that you personalize the Welcome page with details about the user, the date
of use, and the EE/RE program for which displacements are to be estimated. This version
specification is very useful in keeping track of multiple versions of AVERT saved to the same drive.
Please note that the version of AVERT available from EPA does not contain Regional Data File,
and is thus fairly small in size (<6 MB). When a Regional Data File is loaded into AVERT and
displacement is calculated, the program grows substantially to 60 to 100 MB.
A final note on the Welcome page: The Main Module has been designed with the goal of providing
a clear user interface. While it is an Excel file, the tabs that drive the calculations for AVERT are
hidden to enhance the usability and appearance of the tool, making it more similar to a web-based
or executable program. For users who prefer full Excel functionality while using AVERT, there is a
button at the lower right-hand side of the Welcome page that reads "Click here to restore default
functionality." You can complete the steps required to estimate displaced emissions regardless of
whether or not full Excel functionality is visible.
Click on the button labeled "Click here to begin" to move on to AVERT's first step.
40
Due to the large number of calculations—typically several hundred EGUs times 8,760 hours times five output
variables, anywhere from 4 to 40 million complex calculations—storage in a dynamic Excel environment would be
analytically burdensome and space-intensive. Therefore, most calculations are performed once in a Visual Basic
environment, and are not stored in memory.
24
State and Local Energy
and Environment Program
&EPA
-------
ERT
laRaiion Tool^^
epa.gov/avart
AVERT User Manual
Version 2.3
Step 1: Load Regional Data File
Next, choose the region of analysis in Step 1 (as shown in Figure 9) and load the corresponding
Regional Data File. There are 10 AVERT regions to choose from, described in detail in Section 3.
(For details on how the regions were developed, refer to Appendix B.)
Figure 9. Image of AVERT Main Module "Step 1: Import Regional Data File" page.
Step I: Import Regional Data File
Select region
Select a region for analysis by using the
dropdown or by clicking the map.
If you haven't yet downloaded a Regior
0
Enter filepath
Double-dick below to enter the focation
of the Regional Data File.
Load data
Click here to load the
Regional Data File
I. Regional Data
File
Next->
4- Back
The choice of a region determines both which EGUs are included in the analysis and which default
renewable resources are available for modeling the effects of EE/RE programs.
The 10 AVERT regions are:
• California
• Great Lakes/Mid-Atlantic
• Lower Midwest
• Northeast
• Northwest
Rocky Mountains
Southeast
Southwest
Texas
Upper Midwest
If you have not yet downloaded a Regional Data File, you can either select a region of interest from
the dropdown menu or by clicking on the region in the map, then click the hyperlink in the bottom
left of the page.
Then double-click on the filepath box and navigate to the folder where the Regional Data File has
been saved. Once these steps have been completed, click on the green button labeled "Click here
to load the Regional Data File." This step may take several minutes to complete, depending on the
size of the region and the speed of the computer used. When the Regional Data File has been
loaded, a pop-up box appears containing information about the loaded AVERT region and
indicating "Import complete."
dHH 25 i»FPA
State and Local Energy Km m
State and Local Energy
and Environment Program
-------
BsdVERT
AVOidod Emission* and qcnaRation Tool^^
www.epa.gov/avart
AVERT User Manual
Version 2.3
The pop-up box confirms the region and data year loaded by the user and indicates the number of
reporting units in the analysis. The pop-up box also specifies the states that are fully represented
and partially represented in a loaded AVERT region, as shown in Figure 10.
Figure 10. Regional Data File import pop-up box example for 2011 Upper Midwest AVERT region
AVERT
Import complete.
You have loaded the 2011 Upper Midwest (WMW) Regional Data File. This region
contains 348 fossil units.
Generation from the following states is fully represented in this AVERT region:
- Iowa
- Minnesota
- North Dakota
- Nebraska
- South Dakota
Generation from the following states is only partially represented in this AVERT
region:
- Illinois [61%)
- Missouri (45%)
- Mississippi (1%)
- Montana [2%)
- Wisconsin (55%)
Appendix G of the User Manual describes a rule of thumb that users analyzing
partially represented states should consider for assessing the impact of EE/RE over
multiple AVERT regions. The Upper Midwest (WMW) region may include
generation from units in states with a representation too small to be considered
significant for this analysis.
Click the red "Next" button to continue.
OK
Regional Data Files produced prior to summer 2017 do not contain PM2 5 emission data, and they
include generation data in "gross" rather than "net" (corrected for parasitic losses) terms. If you load
a Regional Data File produced in 2017 or earlier, another pop-up box will alert you to these
considerations and suggest that you download a newer Regional Data File from EPA's website
(Figure 11).
State and Local Energy
and Environment Program
26
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
www.epa.gov/avert
AVERT User Manual
Version 2.3
Figure 11. Regional Data File import pop-up example for data files produced prior to summer 2017.
AVERT X
Note that this regional data file does not include PM2.5 data and quantifies
emission impacts based on gross generation, To obtain inputs with PM2.5 data
and net generation, click on the hyperlink under the AVERT map,
OK
Some states are divided by AVERT regions. For example, parts of Illinois are considered in the
Upper Midwest Region, while other parts are in the Great Lakes / Mid-Atlantic Region.41 Appendix
G of this document describes the process that one should use to determine the impact of EE/RE
programs in states that are partially represented in any one AVERT region. Users analyzing these
states should assign the impact of EE or RE programs proportionally based on the fraction of the
state's generation within each relevant AVERT region. These fractions are shown in Table 3 in
Appendix G and, as a reminder and for clarity, in the pop-up box.
After you load a Regional Data File, the blue AVERT header bar will indicate the region and data
year in the top left corner (e.g., "Upper Midwest, 2011"). The blue footer bar will also indicate the
name of the AVERT run that has been loaded in the Regional Data File (e.g.,
"EPA_NetGen_PM25").42 By default, the Regional Data Files provided by EPA are called
"EPABase." When you run the Statistical Module, you will be able to give runs alternative names.
All four of AVERT's steps have a navigation panel on the right-hand side. You can click on any step
or click "Next" or "Back" to move through the steps of the model. To move on to the next step, click
"Next."
Step 2: Set Energy Efficiency and Renewable Energy Data
In this step, you will create a load impact profile (schedule of changes to electricity demand for
every hour of a year) depicting the load reductions expected from an EE/RE program as shown in
Figure 12. In this example, the user has input a 2,000 MW wind program, the displacement profile
of which is being displayed in the chart in the bottom right hand corner
41 This division in Illinois is primarily a function eastern part of the state falling under the PJM RTO, while the rest of
the state falls into MISO.
42 Prior to summer of 2017, EPA used "EPABase" model runs to produce Regional Data Files. To recognize
methodological changes, including the addition of PM2.5 data and adjustment from gross to net generation values,
EPA's default model runs are now named "EPA NetGen PM25."
27 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
J&/&T
missions and qenaRaiion Tool^^
.UllkW.Ulld.uWtl.Ul 9
AVoidod Emissions.
AVERT User Manual
Version 2.3
Figure 12. Image of AVERT Main Module "Step 2: Set EE/RE Impacts" page
Upper Midwest, 2011
Step 2: Set Energy Efficiency and Renewable Energy Impacts
DIRECTIONS: Enter the EERE load for one or a group of EERE policies and programs.^^^^^^^^_^^^^^^^^_
To include the impacts of hourly data manually, click the green button on the right.
Each entry is additive and will create a portfolio of EE/RE impacts.
For further instructions consult Section 4 of the AVERT user manual.
Enter EE impacts based on the % reduction of regional fossil load
Reduce generation by a percent in some or all hours
Apply reduction to top X% hours:
0%
% of top hours
Reduction % in top X% of hours:
0.0%
% reduction
And/or enter EE impacts distributed evenly throughout the year
Reduce generation by annual GWh:
o
GWh
OR
Reduce each hour by constant MW:
0.0
MW
And/or enter annual capacity of RE resources
Wind Capacity:
2000
MW
Utility Solar PV Capacity:
0
MW
Rooftop Solar PV Capacity:
0
MW
Selected EERE Profile Portfolio:
c -£ cL is" c _ =? £- S
•2U.
*
-500
-1,000
-1,500
-2,000
The currently entered reduction profile equals 7,222 GWh,
or 2.7% of regional fossil load.
Welcome
1. Reqional Data
File
3. Run
Displacement
Next ->
Back
2. Set EERE Profile
4. Display Outputs
EPA NetGen PM25
If you enter an EE/RE program that exceeds 15% of regional fossil load in any given hour, you will
be shown an alert highlighting the hours of exceedance (Figure 13), but you can still proceed with
the calculations.
Figure 13. Image of AVERT Main Module "Step 2: Set EE/RE Impacts" page with program size resulting
in a more than 15% reduction in fossil load in some hours.
Upper Midwest, 2011
Step 2: Set Energy Efficiency and Renewable Energy Impacts
DIRECTIONS: Enter the EERE load for one or a group of EERE policies and programs.
To include the impacts of hourly data manually, click the green button on the right.
Each entry is additive and will create a portfolio of EE/RE impacts.
For further instructions consult Section 4 of the AVERT user manual.
Enter EE impacts based on the % reduction of regional fossil load
Caution1! EERE profile exceeds 15% of fossil load in one or more
hours (see below).
Selected EERE Profile Portfolio:
Reduce generation by a percent in some or all hours
Apply reduction to top X% hours:
0%
% of top hours
Reduction % in top X% of hours:
0.0%
% reduction
And/or enter EE impacts distributed evenly throughout the year
Reduce generation by annual GWh:
o
GWh
OR
Reduce each hour by constant MW:
0.0
MW
And/or enter annual capacity of RE resources
Wind Capacity:
6000
MW
Utility Solar PV Capacity:
0
MW
Rooftop Solar PV Capacity:
0
MW
5
The currently entered reduction profile equals 21,667 GWh,
or 8.1% of regional fossil load.
1. Regional Data
File
2.
3. Run
Displacement
Next ->
Back
2. Set EERE Profile
4. Display Outputs
EPA NetGen PM25
Since the release of AVERT S Main Module version 1.5, AVERT adjusts the EE/RE profile to
account for avoided transmission and distribution line losses associated with certain EE/RE
resources: specifically, EE and distributed PV systems.43 This adjustment has the effect of
increasing displaced fossil load and displaced emissions, which provides more accurate results.
Without this adjustment, AVERT would underestimate emissions.
43 AVERT treats wind and utility-scale PV as centralized resources that still require transmission and distribution to
end-users; thus, while they displace fossil generation, a simplified assumption is that they do not avoid any line
loss.
State and Local Energy
and Environment Program
28
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
The Step 2 page allows you to estimate a load reduction from basic characteristics:
• Enter hourly data manually (see green button in right-hand corner)
• Reduce fossil-fuel generation by a percent in some or all hours
• Reduce fossil-fuel generation by total GWh (flat)
• Reduce each hour by a constant MW each hour
• Renewable energy proxy
• Combination of EE/RE programs including combining pre-set options with manual entry
Choose the option(s) that works best for your EE/RE resource and fill in the necessary data. Each
option is described in more detail below. If you choose more than one option (including manual
entry), the selected options will be combined into a portfolio of programs. For any program, or
combination or programs, the Step 2 page returns a total annual energy reduction (in GWh)
achieved by the EE/RE program(s). Total hourly reductions can be found in the manual input page.
Note that it is not recommended for EE/RE profiles to exceed 15 percent of fossil load in any given
hour. If a user-entered EE/RE profile exceeds these recommended limits, a caution message will
appear. The graph in the "Set EE/RE" section will also indicate the affected hours. Exceeding the
15 percent threshold does not prohibit the user from proceeding with the calculations.
Manual User Input
If the hourly load reductions expected from a particular EE/RE policy, program, or measure are
known, a manual stream of load reduction values can be entered for every hour of the year
(consisting of 8,760 hourly values for a non-leap year or 8,784 hourly values for a leap year). For
example, you might use this approach to test the measured or modeled impacts of a particular
known wind farm or EE program. To enter data manually or cut and paste data from another
source, click on the green button that reads "Enter hourly data manually." Data are entered as a
single column of values from midnight on January 1 through 11 p.m. on December 31. On this
page, as with other AVERT inputs, positive values represent displacements. Users who wish to
model scenarios with increases in load (for example, a retroactive "what-if scenario to see what
would have happened if a particular EE/RE policy or program had not been implemented) can
enter negative displacement values to represent increased loads.
This page also includes two columns that indicate whether a user has exceeded the recommended
and/or calculable ranges of hourly load changes. Alerts will appear in these two columns if the data
entered by the user a) produces a cumulative load change that exceeds 15 percent beyond the
upper and lower limits of each hour's original fossil load, or b) produces a new load that is outside
the range of AVERT's ability to calculate displacements.44
The last column on the manual input page shows the total aggregate hourly reduction from the
programs input or selected by the user.
44
AVERT can estimate generation and emissions displacements within the range of actual observed load for a
certain year. In addition, it uses extrapolated data to estimate the displacements that could occur down to a load
level of 0 MW and up to a maximum load level associated with all plants within a single region running at full
capacity. It is unable to estimate displacements outside of this maximum and minimum range.
29
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Figure 14. Image of AVERT Main Module Manual EERE Data Entry page.
Manual EERE Data Entry
Step 2
When complete, click here to return to
Enter Energy Efficiency and Renewable Energy Data
Delete all manual data |
Date M "
Hour
Day of Wee -
Regional Fossil Load (MW) *
Manual EE RE Profile (MW *
Total Change (MW)
Larger than 15%?
Outside of Range?
1/1/2017
1
Sunday
11,878
0
1/1/2017
2
Sunday
11,839
0
1/1/2017
3
Sunday
11,777
0
1/1/2017
4
Sunday
11,466
0
1/1/2017
5
Sunday
11,421
0
1/1/2017
6
Sunday
11,742
0
1/1/2017
7
Sunday
12,469
0
1/1/2017
8
Sunday
12,662
0
1/1/2017
9
Sunday
13,375
0
1/1/2017
10
Sunday
13,297
0
1/1/2017
11
Sunday
12,881
0
1/1/2017
12
Sunday
12,796
0
1/1/2017
13
Sunday
12,647
0
1/1/2017
14
Sunday
12,576
0
1/1/2017
15
Sunday
12,340
0
1/1/2017
16
Sunday
12,493
0
1/1/2017
17
Sunday
13,518
0
1/1/2017
18
Sunday
14,657
0
1/1/2017
19
Sunday
15,477
0
1/1/2017
20
Sunday
15,367
0
1/1/2017
21
Sunday
15,177
0
1/1/2017
22
Sunday
14,520
0
1/1/2017
23
Sunday
13,719
0
1/1/2017
24
Sunday
13,608
0
1/2/2017
1
Monday
13,352
0
1/2/2017
2
Monday
12,979
0
1/2/2017
3
Monday
12,842
0
1/2/2017
4
Monday
12,747
0
1/2/2017
5
Monday
12,825
0
1/2/2017
6
Monday
13,501
0
1/2/2017
7
Monday
14,299
0
1/2/2017
8
Monday
14,596
0
1/2/2017
9
Monday
14,907
0
1/2/2017
10
Monday
15,406
0
Reduce Generation by a Percent in Some or All Hours
To estimate the impacts of a broad-based energy efficiency or demand response program targeting
high-cost peak fossil-consumption hours, enter two values: a) the fraction of hours to which load
reduction is applied (with reductions applied to the highest fossil-fuel generation hours first) and b)
the percent by which those hours should be reduced. For a broad-based efficiency program, the
fraction of hours that experience reductions is very high; for a demand response program, it is very
low.
Note that, when the percentage options are used, reductions are a share of fossil-fuel generation
and not total system load (i.e., consumption). Therefore, using a 2 percent reduction per hour
effectively reduces load by 2 percent of fossil-fired generation.
To simulate a broad-based efficiency program, enter 100 percent as the "% of top hours" and an
estimated load reduction fraction in the cell labeled "% reduction." A graph of the selected EE/RE
profile—which combines both manual entries and user selections made on the Step 2 page—is
shown on this page.
To simulate a peak-reduction targeting program such as demand response, enter a fraction of high-
demand hours that the program is expected to affect, and the load reduction (as a fraction of
peaking load) that would be targeted in those hours. This type of scenario is recommended for
programs that emphasize reductions in peak hours. Generally, reductions exceeding 20 percent of
fossil-fired generation are not recommended.45
45 AVERT is designed to review marginal changes in a system, not large-scale changes. At some point, significant
changes in load will result in non-marginal changes, such as the decommitment of units.
30 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Figure 15 shows a baseline load impact profile (in black) and two example load reductions for three
days in August. The blue, dotted line represents a 3 percent reduction in fossil-fired generation
requirements in every hour of the year. The red, dashed line represents an 8 percent reduction in
the 5 percent of hours with the highest demand for fossil-fired generation.
Figure 15. Examples of two load reduction programs as a percentage of some or all hours.
TJ
c
rt
E
cu
Q
~rt
c
O
"w>
cu
OL
70,000
60,000
50,000
40,000
30,000
20,000
10,000
0 i—
Aug I,
12 AM
¦8% reduction in top 5th percentile hours
3% reduction in all hours
¦Original fossil-fuel load
Aug I,
12 PM
Aug 2,
12 AM
Aug 2,
12 PM
Aug 3,
12 AM
Aug 3,
12 PM
It is important to note that reducing load by a percent in some or all hours measures a reduction
relative to hourly fossil-fuel load, and not system demand (i.e., consumption). It is important to
ensure that the total reduction (in annual GWh) or peak reduction (in MW) comports with
expectations of a reasonable fractional reduction. For example, if fossil-fuel load only accounts for
50 percent of regional generation, a user attempting to find the emissions impact of a regional load
reduction of 3 percent should double the size of the reduction specified in AVERT's Main Module
(i.e., to 6 percent) to scale from total fossil-fuel load to regional load. Similarly, to estimate the
effect of a portfolio reduction (i.e., a percentage change) where you have an expectation of the total
annual reduction (in MWh or GWh), find the correct percentage reduction such that the total
amount of energy reduced is equal to the expected quantity.46
Reduce Generation by Annual GWh
You may have an estimate of the total amount of energy that is targeted or required to be reduced
by a program in a given year, but lack information about the distribution of those reductions over
46
For a program expected to accomplish an annual target reduction through a portfolio efficiency program, the user
should find the percent reduction that will accomplish this target. For example, using Texas data from 2012, a
portfolio EE program expected to reach 10,000 GWh of annual reduction (Ereq) would require a fractional reduction
of 3.79 percent in all hours of the year (Freq). To find this percentage value (Freq), the user can choose an arbitrary
estimated fraction (e.g., 2 percent, Fest), and review the text below the graphic on the Step 2 page indicating the
GWh reduction (Eest)- Divide the required annual reduction (Ereq) by the achieved estimated reduction (Eest), then
multiply this value by the estimated fraction (Fest)- The resulting value is the percentage that should be entered into
the "% reduction" box to achieve the desired energy reduction. Check to ensure that the new total energy
reduction value (in the text below the graphic) is consistent with the desired results.
31
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
the course of the year. "Reduce generation by annual GWh" simply distributes those savings
evenly over all hours of the year. The user inputs a total number of GWh expected to be saved in a
single year. This may be a highly erroneous assumption if savings are targeted from residential or
commercial customers, for whom energy efficiency measures tend to target peak use reductions.
However, an industrial or refrigeration efficiency program may be well represented by a constant
reduction across most hours of the year. Use this option with close attention to the types of
programs assumed in your analysis.
Reduce Each Hour by Constant MW
This option is identical in effect to the annual load reduction by total GWh above. The user selects
a constant reduction for every hour of the year in MW.
Renewable Energy Proxy
This option allows you to use previously designed load impact profiles for region-specific wind,
utility-scale solar PV, and rooftop-scale solar PV. Select the annual capacity (maximum potential
electricity generation) for each type of resource, measured in MW. The model applies these values
to "hourly capacity factors" that vary by resource type and region. Hourly capacity factors are the
probability that a resource is generating at its full capacity in a given hour of the year. For example,
solar panels might have a 90 percent or higher capacity factor on an August afternoon, but a 0
percent capacity factor at midnight any day of the year. Note that these are proxy load impact
profiles for a year based on data from 2011-2013 hourly wind speed datasets gathered from a
variety of model locations across each AVERT region. These data do not represent actual or
proposed projects. The data and methodology used to develop these capacity factors are
described in Appendix C.
After the EE/RE load impact profile has been designed, click "Next" to move on to the next step.
Step 3: Run Displacement
Step 3 launches the automatic calculation of hourly displaced generation, heat input, and
emissions for each EGU in the region as shown in Figure 18.
Click on the box labeled "Click here to calculate displaced generation and emissions." Because of
the large amount of data being processed, this calculation may take several minutes to complete
depending on the size of the region and the processing speed of the computer. A status bar in the
lower left corner indicates the share of processing completed.
Note that two separate alert boxes may pop up when this button is selected. The first alert informs
the user that for at least one of the hours under consideration, the load decrease or increase is
more than 15 percent of the original hourly load. Users may choose to click "OK" and modify their
load changes to be within the recommended range, or click "Cancel" to proceed with the
displacement calculation.
32 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Figure 16. Pop-up alert for an EERE profile that exceeds a 15 percent change in load.
AVERT X
You have entered an EERE profile that exceeds the recommended range
of displacement. Press OK to adjust the magnitude of displacement, or
Cancel to continue with the displacement calculation.
OK Cancel
The second alert informs the user that in at least one of the hours where load has been adjusted,
the displacement will be outside the calculable range for AVERT. In these situations, the user must
return to Step 2 and reduce his or her load adjustments to avoid producing this alert.47
Figure 17. Pop-up alert for an EERE profile that exceeds AVERT's calculable range.
AVERT
X
You have entered an EERE profile that exceeds the calculable range for
hourly load impacts. Change your hourly load impactto eliminate any
errors in the Outside of Range column.
OK | Cancel
A pop-up box that reads "Calculation complete" will appear once the calculations are complete.
Click "Next" to move on to the final step.
47 To remedy this error, users may find it useful to review the right-most column on the "Manual User Input" page,
which produces an alert for specific hours where the load change has exceeded AVERT's calculable range.
&EPA
mm
State and Local Energy
and Environment Program
33
-------
J/VeI
AVoidod Em.s6.ont and qenaRaiion Tc
AVERT User Manual
Version 2.3
Figure 18. Image of AVERT Main Module "Step3i Run Displacement" page.
Upper Midwest, 2011
Step 3: Run Displacement
Click below to calculate displaced generation and emissions.
NOTE
Please be patient.
This calculation may take up to ten minutes to run on older machines.
During this time your screen may go blank or a "not responding" error
may occur - please disregard and allow the calculation to continue.
AVERT
Welcome
1. Reqional Data
File
2. Set EERE Profile
Click here to calculate displaced generation and emissions
3. Run
Displacement
52
4. Display Outputs
EPA NetGen PM25
Next -»
<- Back
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Step 4: Display Outputs
Step 3 of AVERT's Main Module generates raw data in the form of hourly estimated reductions in
generation, heat input, and emissions of PM25, SO2, NOx, and CO2 for each EGU in the region.
These data are aggregated in charts and tables, which are then accessed through the "Step 4:
Display Outputs" page, as shown in Figure 19.
Figure 19. Image of AVERT Main Module "Step 4: Display Outputs" page.
Step 4: Display Outputs
Summary tables
Annual regional
displacement data
Displacement data for
top ten peak days
Annual displacement
data by county
Monthly displacement
data by county
Daily NOx displacement data by county
Charts and figures
Displaced generation
and emissions map
Hourly displacements
by week
Monthly displacements by
selected geography
Signal-to-noise
diagnostic
COBRA text file generation
Enter a filepath, then
click the button to save a
COBRA text file.
NOTE
Please be patient.
This calculation may take
up to twenty minutes to
run on older machines.
Generate COBRA
text files
SMOKE text file generation
Enter a filepath, then
click the button to save
SMOKE text files.
NOTE
Please be patient.
This calculation may take
up to twenty minutes to
run on older machines.
Generate SMOKE
text files
1. Regional Data
File
2. Set EERE
Profile
3. Run
Displacement
4. Display
Outputs
Back
E P A_N etG e n_P M25_Laram i e
To display outputs, click on each of the green boxes under the headings: summary tables, charts
and figures, COBRA text file generation, and SMOKE text file generation. The sample tables and
figures shown below use the same example of a 2,000 MW wind installation in the Upper Midwest
AVERT region.
Summary Tables
Annual displacement summary
Figure 20 shows a high-level summary of the results of the analysis. This table displays the total
annual generation and emissions from the region's fossil generation fleet as reported in the base
year ("Original") and as calculated by AVERT's Main Module after the EE/RE reduction ("Post-
EERE"). The difference between these two scenarios is the total annual expected reduction from
the user-specified EE/RE program. The chart also calculates total annual average fossil-fuel
emissions rates for PM2 5, SO2, NO* , and CO2 before and after the EE/RE program.
All numerical results are shown rounded to the nearest 10 unit.48 Dashes ("-") indicate that AVERT
reported a value greater than zero, but lower than the level of reportable significance. True zeros
are reported as zero (0) values. Increasing an EE/RE program size will increase the amount of
! Data are reported by AMPD in integer units of MWh (generation), ibs (NOx and SO2), tons (CO2), and MMBtu
(heat input). Output data in AVERT are rounded to the closest 10 MWh, Ibs PM2.5, NO* and SO2, tons CO2, and
MMBtu fuel input.
35
State and Local Energy
and Environment Program
&EPA
-------
BsdVERT
AVOidod Emission* and qcnaRation Tool^^
www.epa.gov/avart
AVERT User Manual
Version 2.3
output above the reportable significance level (i.e., reduce the number of dashes in the output
datasets).
Figure 20. image of annual displacement summary table for an example wind program in the Upper
Midwest region.
Upper Midwest, 2011 AVERT
Output: Annual Regional Displacements
Click here to return to Step 4: Display Outputs
Original
Post-EERE
Impacts
Generation (MWh)
267,436,050
260,240,200
-7.195.850
Total Emissions from Fossil Generation Fleet
S02 (lbs)
1.301,793,140
1,268,694.580
-33.098,570
NO* (lbs)
540,761,980
527,072,950
-13,689,030
CO2 (tons)
300,935.610
293,707,130
-7.228.480
PM2.5 (lbs)
43,319,100
42,280,340
-1,038,760
Fossil Generation Fleet Emission Rates
S02 (Ibs/MWh)
4.868
4.875
NOx (Ibs/MWh)
2.022
2.025
C02 (tons/MWh)
1.125
1 129
PM2.5 (Ibs/MWh)
0.162
0.162
Negative numbers indicate displaced generation and emissions.
All results are rounded to the nearest ten. A dash ("—') indicates a result greater than z era,, but lower than
the level of reportable significance.
Annual displacement data by county
Figure 21 shows a summary of the displaced generation and emissions for each of the counties
from each of the states in the region. The Upper Midwest region, for example, contains EGUs in
Illinois, Iowa, Michigan, Minnesota, Missouri, Montana, Nebraska, North Dakota, South Dakota,
and Wisconsin. A line for each county containing an EGU is displayed.
For each county, the following annual output statistics are given:
• Peak Net Generation Post-EE/RE (MW): The peak (maximum) hourly generation
produced by an EGU in the base- or future year scenario after EE/RE programs have been
applied.49
• Annual Net Generation Post-EE/RE (MWh): The total annual generation of an EGU in
the base- or future year scenario after EE/RE programs have been applied.
49 Note that generation is counted on a "net" basis. Generation at the level of the boiler, prior to parasitic use by the
plant or generator, is corrected to "net" generation exported to the grid using technology-specific loss factors.
Parasitic use may include use for fans, pumps, and heating and cooling, and emissions control equipment.
AVERT uses different parasitic loss factors for natural gas-fired combined cycle units, combustion turbines, and
coal-fired steam units with and without controls for sulfur emissions.
36
State and Local Energy
and Environment Program
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
• Capacity Factor (Calculated Post-EE/RE): The capacity factor of an EGU in the base- or
future year scenario after EE/RE programs have been applied. The capacity factor is the
annual generation divided by the product of 8,760 and the peak generation level.
• Annual Change in Generation (MWh): The EGU's estimated change in generation from
baseline conditions to post-EE/RE conditions over a full year (i.e., the annual displaced
generation of this unit).
• Annual Change in Heat lnput/PM2.5/S02/N0x/C02 (MMBtu, lbs, or tons): The EGU's
estimated change in heat input or emissions from baseline conditions to post-EE/RE
conditions over a full year (i.e., the annual displaced heat input or emissions of this unit).
• Ozone Season Change in SO2/NOX/PM2.5 (lbs): The EGU's estimated change in
emissions from baseline conditions to post-EE/RE during the ozone season (May to
September, inclusive).
• Ozone Season, 10 Peak Days Change in SO2/NOX/PM2.5 (lbs): The EGU's estimated
change in emissions from baseline conditions to post-EE/RE during the 10 highest fossil
generation days in the ozone season (May to September, inclusive).
• Daily NOx (lbs) Displacement Data by County: Users can select from one to 10 days of
a given year to display the pounds of NOx for each day. The tool automatically calculates
the daily total for dates of interest and the average across those dates in every county
within one AVERT region.
All results (except for peak generation) are shown rounded to the nearest 10. Dashes ("—")
indicate results greater than zero, but lower than the level of reportable significance.
Figure 21. Annual displacement data by county for an example wind program in the Upper Midwest
region.
Upper Midwest, 2011
Output: Annual Displacement Data by County
^^C]icJUTereJojjeturnJo^te£_4j_Dis£la^Out£uts^_^J
State
H
County
H
Peak Gross
Generation, Post-
EERE (MW) \7\
Annual Gross
Generation, Post- Annual Displaced
EERE (MWh) [> Generation (MWh »
Annual Displaced
S02 (lbs) 0
Annual Displaced
NOx (lbs) 0
Annual Displaced
C02 (tons) p]
IA
Allamakee
222
1,292,900
-60,810
-461,920
20,200
-70,380
IA
Appanoose
6
2,080
-310
-390
-3,600
-400
IA
Audubon
53
10,560
-1,510
-
-2,510
-850
IA
Black Hawk
128
52,230
-7,440
-25,070
-47,710
-9,840
IA
Cerro Gordo
484
413,000
-60,350
-260
-9,500
-26,900
IA
Clay
56
23,300
-1,660
-13,330
-7,390
-1,460
IA
Clinton
200
890,070
-34,220
-258,580
-36,830
-37,950
IA
Des Moines
193
1,157,200
-30,720
-223,460
-35,030
-38,500
IA
Dubuque
57
109,800
-11,110
-117,260
-109,760
-17,230
IA
Louisa
713
4,178,900
-13,870
-68,670
-88,050
-14,690
IA
Marshall
46
14,570
-1,590
-1,370
-10,390
-1,600
IA
Muscatine
252
1,082,980
-69,580
-844,970
-248,620
-91,790
IA
Polk
487
194,130
-27,290
-2,820
-42,330
-16,650
IA
Pottawattamie
1,567
11,559,290
-198,780
-421,570
-345,050
-202,000
IA
Scott
105
770,450
-17,570
-131,640
-30,370
-18,010
IA
Story
88
326,080
-5,440
-23,470
-24,430
-7,110
IA
Union
29
11,980
-1,810
-
-10,700
-1,720
IA
Wapello
637
3,229,680
-144,690
-1,095,560
-216,570
-197,500
IA
Woodbury
1,468
9,421,680
-310,380
-1,790,420
-1,025,230
-329,520
Displacement data for top 10 load days
Figure 22 shows a summary of the 10 days in the region featuring the highest level of fossil fuel
load. Separate columns show the total fossil generation in each day, the expected displaced
generation, the simulated displaced generation, and displaced emissions. All results are shown
37 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
rounded to the nearest 10. Dashes ("—") indicate results greater than zero, but lower than the level
of reportable significance.
Figure 22. Displacement data for top 10 load days for an example wind program in the Upper Midwest.
Output: Displacement Data for Top Ten Peak Days
I Click here , return ,o Step 4: Display Outputs I
Total Fossil Generation
Expected Displaced
Displaced Generation
Displaced NO,
Displaced SOj
Displaced CO2
Displaced PM2 5
Day Rank
Date
(MWh)
Generation (MWh)
(MWh)
(lbs)
(lbs)
(Tons)
(lbs)
1
Jul 18
1.048,930
-13,490
-13.860
-39,530
-17,890
-8,730
-1.770
2
Jul 19
1,039,940
-17,190
-17,270
-61,370
-25,000
-13,150
-2,440
3
Jul 21
1,024,750
-15,770
-15,820
-58,630
-23,210
-13,000
-2,350
4
Jul 20
1,018,680
-16,810
-16,810
-57,920
-22,360
-13,410
-2.590
5
Jul 22
1,003,160
-12,930
-13.030
-44,900
-17,970
-10,620
-1,960
6
Aug 02
993,440
-12,690
-12,700
-46,380
-18,640
-10,370
-1,860
7
Aug 01
988,190
-12,610
-12.650
-46,790
-19,670
-10,530
-1,880
8
Aug 03
983,760
-12,070
-12,110
-34,960
-14,660
-9,550
-1,830
4
Jul 28
979,210
-10,310
-10,290
-36,970
-15,170
-8,330
-1,440
10
Jul 29
975,520
-8,010
-8,130
-25,300
-10,340
-6.580
-1,210
Negative numbers indicate displaced generation and emissions.
All results are rounded to the nearest ten. A dash indicates a result greater than zero, but lower than the level of reportable significance.
Monthly displacement data by county
Figure 23 shows a summary of the displaced generation and emissions for each of the counties
from each of the states contained within the region, broken out by month and with an annual total.
All results (except for peak generation) are shown rounded to the nearest 10. Dashes indicate
results greater than zero, but lower than the level of reportable significance.
Figure 23. Monthly displacement data by county for an example wind program in the Upper Midwest
region.
Upper Midwest, 2011 AVERT |
Output: Monthly Displacement Data by County
Click here to return to Step 4: Display Outputs | Negative numbers indicate displaced generation and emissions. All results are rounded to the nearest ten.
A dash (*¦—') indicates a result greater than zero, but lower than the level of reportable significance.
Displaced
'Ft
R
R
Generation
Pr
Displaced S02
Displaced NOx
Displaced C02
(Tons) ~
Displaced PM:
State
County
Month
(MWh)
(lbs)
(lbs) ~
(lbs)
IA
Allamakee
3
-6,990
-53,440
4,080
-8,120
-2,350
IA
Allamakee
4
-7,640
-57,470
6,140
-8,660
-2,490
IA
Allamakee
5
-6,030
-44,440
-1,270
-6,840
-1,970
IA
Allamakee
6
-3,290
-24,500
-80
-3,780
-1,090
IA
Allamakee
7
-1,830
-13,710
-530
-2,130
-610
IA
Allamakee
8
-2,050
-15,390
-430
-2,380
-690
IA
Allamakee
9
-4,970
-37,160
-2,130
-5,740
-1,650
IA
Allamakee
10
-6,880
-52,320
3,830
-7,890
-2,280
IA
Allamakee
11
-6,400
-48,680
3,010
-7,410
-2,130
IA
Allamakee
12
-5,230
-39,850
2,760
-6,070
-1,750
IA
Allamakee
Annual
-60,810
-461,920
20,200
-70,380
-20,280
IA
Appanoose
1
-20
-10
-80
-10
-
IA
Appanoose
2
-20
-10
-40
-10
-
IA
Appanoose
3
-40
-30
-310
-30
-10
IA
Appanoose
4
-10
-10
-110
-10
-
IA
Appanoose
5
-10
-20
-170
-20
-
IA
Appanoose
6
-10
-30
-230
-30
-10
IA
Appanoose
7
-70
-130
-1,230
-130
-30
IA
Appanoose
8
-20
-50
-430
-50
-10
IA
Appanoose
9
-30
-40
-380
-40
-10
IA
Appanoose
10
-30
-40
-340
-40
-10
IA
Appanoose
-20
-20
-180
-20
" "I""'
IA
Appanoose
12
-10
-10
-100
-10
-
IA
Appanoose
Annual
-310
-390
-3,600
-400
-80
IA
Audubon
1
-80
-
-50
-30
-
IA
Audubon
2
-10
-
10
-
-
Daily NOx displacement data by county
Figure 24 shows a summary of displaced NOx emissions in the counties of each state contained
within the region, broken out by day and with a daily average. Users enter up to 10 days for the
State and Local Energy
and Environment Program
38
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
analysis year in (MM/DD) format. All results are shown rounded to the nearest 10. Dashes
indicate results greater than zero, but lower than the level of reportable significance.
Figure 24. Daily NOx displacement data by county in Upper Midwest region.
[Upper Midwest, 2011 AVERT|
Output: Daily NOx Displacements (lbs)
Negative numbers indicate displaced generation and emissions. All results are rounded to the nearest ten. A dash indicates a result greater than zero, but lower than the level of reportable significance.
Enter up to ten dates in the header column. This page will calculate the NOx displacement associated with each day in each county, as well as the average for each county. Use the filters to select individual states or counties.
Enter dates of int
-rest (MM/DD)
St*e
County
8-Aug
7 Jul
6-Jun
5-May
| 8-Sep |
8Jul
6-Sep
5-Jun
19-Aug
20-Aug |
Average
IA
Allamakee
46
-5
-4
-86
-77
-9
-58
50
-6
-123
-27
IA
Appanoose
-29
-5
-13
-7
-1
-16
12
-5
4
8
-6
IA
Audubon
-31
-22
-25
-3
-2
-16
-3
-5
-23
-2
-13
IA
Black Hawk
-311
-227
-303
-52
-54
-214
-99
-81
-243
-81
-166
IA
Cerro Gordo
-7
-4
-7
-17
-11
1
-11
-22
-2
-20
-10
IA
Clay
-44
-57
-41
-81
-40
8
-11
25
5
-29
IA
Clinton
13
-38
-68
-73
75
-100
-41
-95
-158
-51
IA
Des Moines
6
80
23
115
1
30
-18
-35
-9
6
20
IA
Dubuque
-121
-90
-102
-249
-163
-106
-204
-367
-130
-115
-165
IA
Louisa
-111
-418
-640
160
-587
23
-249
-222
-560
-570
-317
IA
Marshall
-81
-63
-67
-50
5
-49
-23
21
-57
-20
-38
IA
Muscatine
-298
-162
-253
-487
-307
-259
-233
-441
-216
423
-308
IA
Polk
-288
-213
-299
-141
8
-203
2
-26
-209
-54
-142
IA
Pottawattamie
-161
-51
-282
-861
-470
-339
-20
-800
-207
-591
-378
IA
Scott
-51
-60
-38
-288
52
-28
-65
-53
-70
-66
-67
IA
Story
-146
-121
-117
-140
-24
-93
-64
-6
-143
40
-89
IA
Union
-91
-53
-84
2
-9
-57
-7
-5
-50
1
-35
IA
Wapello
-197
-289
-431
441
-512
-315
-389
-741
-307
-124
-374
IA
Woodbury
-564
-752
-1,115
-2,248
-2,683
-548
-2,359
-1,775
-720
-639
-1,340
IL
Clay
-22
-14
-16
0
-3
-11
-1
4
-17
-2
-10
IL
Crawford
-198
-151
-174
-509
-131
-113
-181
-276
-189
-211
-213
IL
Fayette
-9
-5
-6
0
0
-4
0
0
-6
0
-3
IL
Ford
-9
4
-6
4
-6
-6
-1
4
4
-3
-3
Charts and Figures
Displaced generation and emissions map
This dynamic map (shown in Figure 25) shows where emissions have been displaced within the
selected region. You can choose from the following options in a dropdown menu:
• Annual Change in Generation (MWh)
• Annual Change in Heat Input (MMBtu)
• Annual Change in SO2 (lbs)
• Annual Change in NOx (lbs)
• Annual Change in PM2 5 (lbs)
• Annual Change in CO2 (tons)
• Ozone Season Change in SO2 (lbs)
• Ozone Season Change in NOx (lbs)
• Ozone Season Change in PM2 5 (lbs)
• Ozone Season, 10 Peak Days Change in SO2 (lbs)
• Ozone Season, 10 Peak Days Change in NOx (lbs)
• Ozone Season, 10 Peak Days Change in PM2 5 (lbs)
Click on "Refresh map with selected variable" after making a selection. The map displays the
annual, seasonal, or peak displaced emissions at specific sources in the region. The size of the
circles indicates the relative displacement of each resource. Circles are semi-transparent. If
multiple sources are near the same location, the circle is darker. Occasional negative values (i.e.,
an emissions increase) are shown with black outlines and white interiors; these are often the result
of the timing of maintenance outages in the base-year data (see Appendix G for details).
I11H
State and Local Energy
and Environment Program
39
&EPA
-------
AVERT User Manual
Version 2.3
Figure 25. Displacement generation and emissions map for an example wind program in the Upper
Midwest region.
Upper Midwest, 2011
Output: Displaced Generation and Emissions Map
Select variable to display: |Annual Change in CQ2 (tons) ]y I I Refresh map""
Note: The diameter of each circle indicates the
magnitude of a unit's change in generation /
emissions. Circles are semi-transparent; darker
areas occur in regions with overlapping units.
Negative changes are indicated with blue
circles; positive changes are indicated with
black-bordered white circles.
Annual Change in C02 (tons)
Upper Midwest
45 ° -
40° -
# 9,500 MWh
The diameter of each circle indicates the magnitude of an EGU's change in emissions. Circles are semi-
transparent; darker areas occur in regions with overlapping EGUs. Emissions reductions are indicated with
blue circles; increases in emissions are indicated with black-bordered white circles.
Displacement data by month
Monthly output can be viewed over the entire region, or a specific state or county within the region
(see examples in Figure 26 and Figure 27). First select "Region," "State," or "County" in the top
dropdown menu. If selecting a state, choose the state in the next dropdown menu. If selecting a
county, choose both the state and the county in the next two dropdown menus.
State and Local Energy
and Environment Program
40
&EPA
-------
J/VeI
AVoidod Em.s6.ont and qenaRaiion Tc
AVERT User Manual
Version 2.3
Figure 26. Displacement data by month chart for an example wind program in the Upper Midwest
region.
Upper Midwest, 2011
Output: Monthly Displacements by Selected Geography
|^_^_CHclUTereJoj2turnitoi^teg^iDisgla^iOutgiJts__^_|
Select level of aggregation
Select state:
State
a
c
o
o
0
-200,000
-400,000
-600,000
-800,000
0
-100,000
-200,000
-300,000
0
-50,000
-100,000
-150,000
0
-10,000
-20,000
-30,000
-40,000
Monthly Emission Changes, Upper Midwest (IA)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
111 n i
, ¦.
¦ i
Mil
J
Jan Feb Mai Apr May Jun Jul Aug Sep Oct Nov Dec
¦ ¦
Jan Feb Mar Apr Ma/ Jun Jul Aug Sep Oct New Dec
i
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Now Dec
i
~
~
i
Figure 27. Displacement data by month table for an example wind program in the Upper Midwest
region.
Monthly Emission Changes, Upper Midwest (IA)
Gen (MWh)
SOj (lbs)
NO, (lbs)
CO; (tons)
PMz 5 (lbs)
Jan
-90,940
-514,390
-208,440
-96,620
-23,080
Feb
-73,610
-450,810
-176,470
-82,480
-21,930
Mar
-79,370
-477,900
-195,080
-88,990
-24,280
Apr
-114,200
-650,670
-252,620
-126,120
-31,000
May
-100,980
-560,910
-221,810
-110,090
-26,860
Jun
-77,470
-420,200
-175,320
-84,010
-20,460
Jul
-53,240
-208,720
-116,120
-51,790
-10.910
Aug
-57,860
-227,730
-114,160
-56,390
-11,250
Sep
-76,440
-411,480
-170,550
-82,770
-20,420
Oct
-98,070
-547,520
-227,010
-108,790
-27,470
Nov
-97,290
-557,430
-226,030
-108.030
-27,790
Dec
-79,650
-453,030
-190,250
-88,030
-22,600
Total
-999,000
-5,451,000
-2,274,000
-1,084,000
-268.000
Negative numbers indicate displaced
generation and emissions. AH results are
rounded to the nearest ten. A dash
indicates a result greater than zero, but lower
than the level of reportable significance.
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Hourly displacements by week
Figure 28 is a dynamic representation of hourly displacement from each EGU in a region. Individual
plants are stacked as gradated bar plots, from high-capacity-factor "baseload" EGUs in dark blue to
low capacity factor peaking EGUs in light blue.50 The total contribution of all EGUs sums to the
yellow line. As noted above, some EGUs can show a net increase in emissions as regional load is
reduced, often due to the timing of maintenance outages in the base-year data.
The second chart in Figure 28 shows the same week-long load impact profile as above, but
presents the displaced load in reference to the total fossil-fuel load. The purpose of this chart is to
illustrate the degree of change represented by the EE/RE program relative to the baseline. The
solid line represents the total fossil-fuel load by hour in the baseline; the dashed line represents the
fossil-fuel load after the user-specified EE/RE reduction has been modeled. At the bottom of the
graphic, the gray area graph represents the hourly EE/RE reduction as specified by the user.
Select which variable to review (changes in generation, heat input, or PM2 5, SO2, NOx, or CO2
emissions), and the first day of the analysis week. Once these variables are chosen, click on
"Refresh displacement chart and load chart" to view the results.
50
Gradations are on a relative scale. Within a region, the unit with the highest capacity factor sets the darkest
gradation end (baseload), while the unit with the lowest capacity factor sets the lightest gradation end (peaking).
Units are sequentially partitioned into color blocks. Given that most regions include several hundred units, this
gradation will likely be similar in most regions with true baseload units at the darkest end, and true peaking units at
the lightest end.
42
State and Local Energy
and Environment Program
&EPA
-------
J^EF
AVERT User Manual
Version 2.3
AVERT
AVoidod Emission* and qenaRaiion Tool^^
www.epa.aov/a'/ts!
Figure 28. Hourly displacement data for an example wind program in the Upper Midwest region.
Output: Hourly Displacements by Week
C
^lic^ere^o^tum^o^te^^U^la^Out^uts^
Select variable to display:
First day to display (MM-DD):
Change in Generation [MW]
August 1
J
Refresh displacement chart and load chart
Change in Generation (MW) in Week of 8/1
Mjl
sf
£r
8
I
O
Fossil-fuel load, pre- and post-EERE, in Week of 8/1
60,000
50,000
40,000
30,000
20,000
10,000
D
T
3
t
I
s
nsjmb&rs/h'Jfcate
dtipfecedg&riSf&ljhn -and
tp/mssfons:
Total Change in
Generation (MW)
Total fossil-fuel
load.
High capacity factor units
Low capacity factor units
_ Total fossil-fuel
load.
_ Total fossil-fuel
load, post-EERE
]] Load displaced
Signal-to-noise diagnostic
The signal-to-noise diagnostic shown in Figure 29 has a different structure from the time-series
images shown previously. This chart is a scatterplot of every hour of the year (8,760 or, in a leap
year, 8,784 points), showing calculated total generation reduction in each hour (y-axis) against
user-input EE/RE load reduction in each hour (x-axis). Ideally, AVERT perfectly matches unit
generation reductions to the amount of EE/RE load reduction requested by the user. This graphic
shows where that assumption hoids and where it does not hold, and to what extent. If the
generation reduction is well-matched to the EE/RE load reduction, the graphic will show a straight,
line with little scatter. If the reductions are not well matched, the line will have significant scatter.
Overall, the quality of fit (i.e., how well the generation reduction captures the EE/RE load reduction)
can be judged from the R2 metric shown in the chart title.51 Highly scattered data points should be
viewed with less weight than well-constrained data points.
51 R2 values indicate the quality of fit of a line, or how well dependent variables describe independent variables (i.e.,
how well the y-axis value describes the x-axis). Random points have an R2 value of zero (0), while perfectly
dHQ 43 i»FPA
State and Local Energy mK m
State and Local Energy
and Environment Program
-------
LAVPPT AVERT User Manual
BAYfcKI Version 2.3
Figure 29. Signal-to-noise diagnostic for an example wind program in the Upper Midwest region.
Upper Midwest, 2011
Output: Signal-to-noise diagnostic
I Click here to return to Step 4: Display Outputs
Reduction in Total Unit Generation Relative to EERE Load
Reduction (MW) for All Hours, R2=0.98
Expected EERE Load Reduction (MW)
-2,000 -1,800 -1,600 -1,400 -1,200 -1,000 -800 -600 -400 -200 0
-200
-400
-600 i"
1
c
-800 %
a
B
=
-1,000 5
c
c
-1,200 5
0
3
¦p
-1,400 a
a»
c
-1,600 1
SJ
C£
-1,800
-2,000
/
A
v
/
r
r
The above chart is a scatterpiot of every hajr >ofthe year; it contains eft/ler &.. TSOorS.. 734data points.. one for each hour. Cfrartedpoints sho n ' tfie total generation reduction
fy-am?) i
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
https://www.epa.gov/statelocalenerqv/co-benefits-risk-assessment-cobra-health-impacts-
screeninq-and-mappinq-tool and EPA's COBRA user manual.
SMOKE Text File
AVERT allows you to create files for use in SMOKE: enter a filepath in the bottom center of the
"Step 4: Display Outputs" sheet, then press the "Generate SMOKE text files" button. Twenty-four
text files are then generated in this filepath, two for each month. One of these files is for the
selected region and year, pre-EE/RE displacement, while the second set of files details information
post-displacement. For more detailed instructions on how to interpret and use SMOKE outputs
from AVERT, see the tutorial and step-by-step instruction slides available on EPA's website at
https://www.epa.qov/statelocalenerqv/avert-tutorial-usinq-avert-outputs-smoke-overview.
Advanced Outputs
Though AVERT does track the estimated output and emissions from specific EGUs, we
recommend only using these outputs for SMOKE processing and/or quantitative validation
purposes.
To access annual output on a unit-by-unit basis, restore default Excel functionality via the button on
the Welcome page. Data are available for each EGU in the region of interest in the "Summary"
worksheet. EGUs are identified by their ORISPL number,52 unit number, and unit name, fuel type,
state, county, and geographic location (latitude/longitude). Summary data are provided for each
EGU in a similar format to the county data, as described previously.
In addition, detailed data are available in the worksheets labeled "Gen" (generation), "Heatlnput,"
"SO2," "NOx," "CO2," and "PM25." These worksheets record displaced emissions and generation for
each EGU in the region for each hour of the modeled year. Hours are arrayed vertically; EGUs are
arrayed horizontally.
In contrast to the results shown in the summary tables, charts, and figures, which have been
rounded to the nearest 10, results shown in the advanced outputs have not been rounded.
Unrounded results should only be used after due consideration of their significance.
52
The ORIS or ORISPL number is a code used by DOE and EPA to identify specific generating plants, where a
"plant" is a site that may include multiple EGUs. Each ORIS number is unique and (usually) persistent. "Unit
numbers" are assigned to generators and boilers by DOE and EPA, respectively, and are subject to change or
modification by accounting agency or reporting entity.
45
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Appendix A: Installation Instructions
AVERT is divided into three components: an Excel-based platform for user-specified analysis of
displaced emissions (called the Main Module), a MATLAB®-based statistical analysis program
(called the Statistical Module), and a second Excel-based spreadsheet for creating user-specified
future year scenarios (called the Future Year Scenario Template). This section provides
installation instructions for each AVERT module. More detailed information on AVERT components
is provided in Section 2 of this manual, "Understanding AVERT."
Main Module
AVERT's Main Module estimates the displaced emissions likely to result from EE/RE programs in
reference to a base-year or future year scenario.
System Requirements
The Main Module requires Excel 2007 or newer to run in Windows. The Main Module can also be
used in Excel 2011 or newer for Mac; it has been verified to work up to Excel for Mac 2016 v16.23.
Macros must be enabled. You do not need to install the Statistical Module and the Future Year
Scenario Template to use the Main Module to estimate displaced emissions for EE/RE programs
modeled in reference to a historical base year; however, you will need all three AVERT modules to
model displaced emissions with reference to user-created future years.
The Main Module has no special requirements for hard drive space or RAM on the computer
running it, but it will run faster on computers with more RAM and higher-speed processors. Excel
files generated in the Main Module can exceed 100 MB in size, depending on the number of EGUs
in the region of analysis. Analyzing data for large regions may take over 10 minutes on some
computers.
Installation
To use the Main Module, download two files and save both to the same folder on a local computer
or drive:
• The Main Module workbook: "AVERT Main Module.xlsx". Download the workbook at
https://www.epa.aov/statelocalenerav/download-avert.
• The Regional Data File for the region under analysis.
o Default Regional Data Files developed for use by EPA are labeled
"AVERT RDF [DataYear] EPA_NetGen_PM25 ([Region]).xlsx"; they can be
obtained at https://www.epa.aov/statelocalenerav/download-avert.
o Regional analyses developed by advanced users using AVERT's Statistical
Module will be saved, by default, in a folder of the Statistical Module titled "AVERT
Output." These files use the following naming convention:
"AVERT RDF [DataYear] [RunName] ([Region]) [RunDateTime].xlsx."
In the Regional Data File:
• "Region" refers to one of 10 regions defined for the purposes of this tool. AVERT's regions
are described in Section 3 of this manual, under "AVERT Regions" (page 15). The
46 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
"BaselineYear" tag indicates the base year (the year upon which the analysis is based).
Generally, for contemporary or forward-looking analyses, this should be the most recent
full year of data available from CAMD's Air Markets Program, although data years 2007
through 2018 are currently available for input.
• "RunDateTime" indicates when the data file was generated by the Statistical Module.
Launching A VERT's Main Module
To launch the model, open the Main Module workbook in Excel and follow the step-by-step
instructions in Section 4 of this manual.
Technical Assistance
For more information, please contact EPA's State and Local Energy and Environment Program at
avert@epa.gov.
Statistical Module
AVERT's MATLAB®-based Statistical Module performs statistical analysis on AMPD to generate
output files used to model displaced emissions in the Main Module. Running the Statistical Module
is not required to operate the Main Module; it is anticipated that most AVERT users will not run it.
Users creating specific future year scenarios, however, will need to run the Statistical Module.
For more information on AVERT's Statistical Module, see Appendix D.
System Requirements
The Statistical Module requires a machine capable of running Windows XP or higher.
It is recommended that computers operating the Statistical Module have at least 2 GB of memory
available. Processing time for individual regions depends on the number of EGUs in the analysis
and the number of processors available for use by the MATLAB® platform. In development of
AVERT, it was found that for full-scale runs, larger regions could take over two hours to analyze
with four processers dedicated to the operation.
The Statistical Module can perform analysis either with a pre-loaded base-year dataset from 2007
through 2018, or with a revised electricity generation fleet created in the Future Year Scenario
Template, used in conjunction with a pre-loaded base-year dataset.
Installation and Launching
To use the Statistical Module, follow the instructions in Appendix E.
This output file can be used directly in the Main Module to analyze displaced emissions from
EE/RE programs.
Technical Assistance
For more information, please contact EPA's State and Local Energy and Environment Program at
avert@epa.qov.
47 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Future Year Scenario Template
AVERT's Future Year Scenario Template allows the user to modify the list of EGUs analyzed by
the Statistical Module. EGUs can be added or retired, or have their emissions rates modified.
Newly added EGUs are copied from existing EGUs (proxy units), but can be scaled to a desired
capacity, and given a location (county or latitude/longitude) in a different location.
System Requirements
The Future Year Scenario Template requires Excel 2007 or newer to run. It has been designed for
Windows and has not been tested on a Mac, as the companion Statistical Module requires
Windows. You do not need to install the Statistical Module to design scenarios within the Future
Year Scenario Template, but you will need it to analyze those scenarios and estimate their future
emissions in the Main Module.
The Future Year Scenario Template has no special requirements for hard drive space or RAM, but
it will run faster on computers with more RAM and higher-speed processors. Scenarios saved by
the Future Year Scenario Template are likely to be between 14 and 25 MB in size, depending on
the number of EGUs being added in a new scenario.
Installation
The Future Year Scenario Template is packaged with the Statistical Module executable package.
Instructions on obtaining this package can be found in Appendix E.
On downloading and unpacking the package, you will be presented with a folder entitled "AVERT
Future Year Scenarios." This folder contains a number of example templates illustrating
retirements, additions, and retrofits.
Launching and Working with the Future Year Scenario Template
To launch the Future Year Scenario Template, open the workbook in Excel and follow the step-by-
step instructions in Appendix F.
Technical Assistance
For more information, please contact EPA's State and Local Energy and Environment Program at
avert@epa.gov.
I11H
State and Local Energy
and Environment Program
48
&EPA
-------
AVERT User Manual
Version 2.3
Appendix B: Data
AVERT uses Air Markets Program Data (AMPD) from the EPA Clean Air Markets Division
(CAMD).53 This version of AVERT includes the data necessary to use 2007 through 2018 as the
base year of analysis.54
For the purposes of tracking and verifying emissions, and monitoring emissions trading programs,
AMPD collects extensive operational data from nearly all operating fossil-fuel EGUs with
generating capacities greater than 25 MW in the lower 48 states (i.e., excluding Alaska and
Hawaii).55 Data collected in AMPD include reported gross generation (in megawatt hours per hour,
or MWh/h),56 steam output (in tons, from combined heat and power facilities), heat input (in million
metric British thermal units, or MMBtu), and emissions of sulfur dioxide (SO2), oxides of nitrogen
(NOx), and carbon dioxide (CO2). Each quarter, CAMD consolidates information from the previous
quarter (i.e., there is typically a three-month delay in releasing data) and produces text-based
datasets for each of these factors for each fossil-fuel EGU in each state.57
A MATLAB®-based preprocessing engine converts these hourly text files into compact data arrays
and a reference EGU records file.58 The preprocessing engine calls an Excel-based spreadsheet
populated with ancillary information about each EGU, with most information gathered from AMPD
"facility information" records. The spreadsheet is populated with ancillary lookup information about
each EGU that has reported to CAMD from 2007 to 2018, which allows the model to be backward-
compatible with prior year data.
Gross generation is converted to net generation within the preprocessing engine using unit-specific
parasitic loss factors. These factors were calculated based on a comparison of by-plant gross
generation59 and by-plant net generation60 using 2015 data.61 Different loss factors are used for
coal-fired steam units with and without sulfur controls (8.3% and 6.9%, respectively); natural gas-
fired combined cycle units (3.3%) and combustion turbines (2.2%); and natural gas- or oil-fired
steam units (7.7%). For example, a sulfur-controlled coal steam unit with an annual gross
generation of 100 GWh is assumed to export a total of 91.7 GWh to the grid, while a natural gas-
fired conbined cycle unit with the same gross generation is assumed to export 96.7 GWh.
53 https://ampd.epa.gov/ampd/.
54 Future years will be released as data is made available from CAMD.
55 For the purposes of AMPD collections, "units" are typically individual boilers, but sometimes represent either a
single emissions source (i.e., smokestack) from several attached boilers or the consolidated output of a single
generator with multiple boilers. AMPD unit designations are often, but not always, the same as U.S. Department of
Energy unit designations.
56 Gross generation is measured at the level of the boiler, prior to parasitic use by the plant or generator. Parasitic
use may include use for fans, pumps, heating and cooling, and emissions control equipment. Therefore,
generation seen by the grid may differ from the values in this database by 0 to 10 percent, depending on the unit.
Emissions, however, are "at stack" and represent total emissions released to the atmosphere.
57 AMPD collects data from most fossil-fired electrical generating stations over 25 MW in the lower 48 contiguous
states (i.e., excludes Alaska, Hawaii, and territories). This data set generally does not include data from biomass
generation or most small diesel backup generators.
58 https://www.mathworks.com/products/matlab.html.
59 As reported to AMPD.
60 As reported to the U.S. Department of Energy's Energy Information Administration on Form 923
(https://www.eia.gov/electricitv/data/eia923/).
61 Empirical parasitic loss factors were found to be comparable to those published in the literature.
— 49 mFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
BsdVERT
AVOidod Emission* and qcnaRation Tool^^
www.epa.gov/avart
AVERT User Manual
Version 2.3
The six data arrays store two-dimensional matrices of net generation, steam output,62 heat input,
PM2.5, SO2, NOx, and CO2 organized by EGU and by hour of the year. Figure 30 shows an
example two-dimensional data array for base-year hourly generation (8,760 or 8,784 hours across
the horizontal axis) for each of the 4,734 fossil-fuel EGUs (down the vertical axis) for which AMPD
collected emissions data in 2011. Black areas represent hours during which particular EGUs are
not in operation (or are operating at very low levels, i.e., less than 10 MW). Figure 30 also includes
detail from the data array that focuses in on 10 EGUs and hours 3,000 through 4,000 in the base
year.
Figure 30. 2011 gross generation output (in MW) for each EGU in each hour of the year.
1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000
Hour of Year
Little Gypsy 1
Little Gypsy 2
Little Gypsy 3
Ninemile Point 1
Ninemile Point 2
Ninemile Point 3
Ninemile Point 4
Ninemile Point 5
Sterlington 7AB
Sterlington 7C
3,000
100
200
300
400 500 600
EGU Level of Output (MW)
700
800
900 1,000
or greater
3,100
3,200
3,300
3,400 3,500 3,600
Hour of Year
3,700
3,800
3,900
4,000
Colors represent the level of EGU generation.
The reference EGU records file (a structural array in MATLAB®) holds the name, ORISPL number
(a value assigned to each plant site by the Department of Energy [DOE]), EPA unit ID, a lookup
table (LUT) pointer to the two-dimensional matrices, locational information, and fuel information for
each EGU. Table 3 shows an example record for the Handley Generation Station, Unit 5. The
62
AVERT does not use steam output.
HSIffl 50 aFPA
State and Local Energy Km m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
"Unique ID" shown in Table 3 is a unique identifier created within AVERT, consisting of the ORISPL
number concatenated with the EPA unit identification number as a string.63 In addition, each record
stores a lookup value (not shown in Table 3), which codes for the location of the plant in the two-
dimensional data files.64
Table 3. Example record in the reference EGU records file.
Name
Handley Generating Station
UnitID
5
NERCSub
ERCT
NERCSub Ix
26
State
TX
State Ix
44
LUTValue
4022
ORISPL
3491
Lat
32.7278
Lon
-97.2186
County
Tarrant
FuelPrimary
Pipeline Natural Gas
FuelSecondary
Diesel Oil
PrimeFuelType
Gas
UniquelD
3491|5
CSIRegion
TX
CSIRegionIX
9
EGU records also include a plant-specific emission rate for PM2 5, in pounds of particulate matter
per MMBTU of heat input, as well as including expected CO2 emissions data for units that do not
report CO2 to AMPD on an hourly basis.
PM2 5 emissions rates were calculated using plant-specific point source emissions data from the
2014 National Emissions Inventory,65 combined with heat input data as reported in AMPD. For prior
data years, plants without PM2 5 are assigned the average rate of similar plants (i.e., the same
prime mover and fuel type) within the same AVERT region. PM2 5 emission rates are used within
the AVERT statistical module to calculate unit-specific PM2 5 emissions for each unit within each
load bin.
Expected hourly CO2 emissions data were calculated within the AVERT statistical module only for
units that do not report CO2. Expected CO2 emissions for these units were calculated as the
product of an assumed fuel-specific CO2 content factor and the unit's heat input for each hour.
CO2 factors in tons of CO2 emitted per MMBTU of fuel consumed were calculated using the
"unspecified coal", "natural gas", and "distillate fuel oil" carbon content values codified for EPA's
Greenhouse Gas Reporting Program (GHGRP) in 40 CFR Part 98, Subpart C. Units with a fuel
63 The pipe character ("|") is used to separate the ORISPL and Unit ID for legibility.
64 Due to coding limitations in Excel and MATLAB, a small number of units have modified UnitlDs relative to the
UnitID that appears in AMPD. For example, units with a UnitID of "1-1" in AMPD may instead use a UnitID of "N1"
in AVERT.
65 https://www.epa.aov/air-emissions-inventories/national-emissions-inventorv-nei
51 **FPA
State and Local Energy m m
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
type other than coal, oil, or gas were assumed to have the same carbon emissions factor as oil-
fired units.
AVERT analysis is conducted by region, with the continental United States divided into 10
reasonably autonomous electricity-market trading and dispatch areas. These AVERT regions are
generally aggregations of the eGRID subregions used by EPA, and are similar, but not identical, to
North American Electric Reliability Corporation regions.66 For a map of AVERT regions, see Figure
3 (page 16).
Several of the AVERT regions represent electricity market areas or balancing authorities. The
Northeast region represents a combination of the New England and New York Independent System
Operators (ISOs), the Great Lakes/Mid-Atlantic region represents the PJM Regional Transmission
Organization (RTO), the Upper Midwest region encompasses most of the Midwest ISO; the Texas
region represents ERCOT, and California represents the California ISO region. The Southeast
region includes the SERC and Florida reliability regions. Both the Northwest and Southwest
regions encompass several large interconnected utilities and major transmission systems that have
numerous interdependencies. The Rocky Mountain and Lower Midwest regions operate largely
autonomously relative to the other AVERT regions shown here.
Analysis based on smaller regions, such as eGRID regions, risks missing important
interdependencies between the EGUs in a larger region (e.g., the impact of New Jersey load
reductions on Ohio EGUs). Using still larger regions, such as the Eastern Interconnect, spreads the
influence of load reductions too widely, making it difficult to ascribe load reductions at a particular
location to a reasonable cohort of EGUs.
66
AVERT regions separate most of the major ISO and RTOs, such as Texas (ERCOT), the California ISO, and the
Southwest Power Pool (SPP, or Lower Midwest here). The Midwest ISO is generally aggregated into the Upper
Midwest region, with the exception of Michigan, which is interconnected with the PJM RTO (called here the Great
Lakes/Mid-Atlantic region). The Southwest region is primary composed of several large multi-state utilities
(Duke/Progress, Southern Company, TVA, Entergy, and Florida Power and Light). While the western region of the
United States is highly interconnected as well, most of California is independently dispatched, and few resources
in the Rocky Mountain region are used for export purposes to either the Northwest or Southwest. In addition, there
are few connections between the Northwest and Southwest.
52
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
Appendix C: Proxy Renewable Energy Hourly Profiles
AVERT's Main Module provides example proxy hourly capacity factors for generic solar and wind
projects. These capacity factors are meant to provide quickly accessible options to review example
renewable project portfolios in each of the regions discussed here. The user is encouraged to
develop site-, state-, or region-specific renewable energy load profiles. Where such information is
not available or for the purposes of exploration, the proxy capacity factors in the Main Module
provide a reasonable basis for expected wind and solar hourly profiles.
Annual hourly capacity factors for rooftop photovoltaic (PV) and utility PV were obtained from the
National Renewable Energy Laboratory's PVWatts v.1 tool.67 Each hourly capacity factor
assembled for each AVERT region is based on the average PV capacity factor for four to 10 cities
in the region. The number and location of the sampled cities were chosen to provide a
representative distribution of the AVERT region's insolation (energy from sunlight) at the largest
load centers.
Wind capacity factors were developed from annual 6-hour datasets of modeled wind speeds at 80-
meter turbine (hub) heights obtained from the Global Model Database developed by AWS
Truepower for 2011 through 2013. Depending on the size of the region, between five and 15
locations were used to provide a representative distribution of hypothetical wind turbine
installations. Once hourly wind speed data for each site were created by interpolating each of the
6-hour intervals, 2011-2013 hourly wind speed datasets were averaged and were then applied to a
power density curve for a Vestas V112 3 MW Wind Turbine.68 These hourly data were divided by
total regional wind nameplate capacity to produce hourly capacity factors. Hourly capacity factor
datasets from all the sites within a region were then averaged to produce a regional hourly dataset
for capacity factors. The flow of data is shown in Figure 31.
Users are encouraged to develop site-specific capacity factor profiles for renewable energy options
whenever the data are available. It is important to note that AVERT is not a tool for formal
greenhouse gas accounting or establishing who may take claim credit for emission reductions of
RE programs or projects. It is recommended that companies follow the protocols from the World
Resources Institute's GHG Protocol and the Federal Trade Commission's Guides for the Use of
Environmental Marketing Claims for the purposes of greenhouse gas and carbon footprint
accounting.69
67 National Renewable Energy Laboratory, n.d. PVWatts: A performance calculator for grid-connected PV systems.
Accessed December 14, 2012. Available at https://pvwatts.nrel.gov/.
68 Vestas Wind Systems. 2013. 3 MW Platform. Available at https://www.nhsec.nh. aov/proiects/2013-
02/documents/131212appendix 15.pdf.
69 Federal Trade Commission. 2012. Guides for the use of environmental marketing claims.
https://www.ftc.gov/sites/default/files/documents/federal register notices/guides-use-environmental-marketing-
claims-areen-auides/greenauidesfrn.pdf.
53
State and Local Energy
and Environment Program
&EPA
-------
al^EF
AVERT User Manual
Version 2.3
feAVERT
I EmiaMont and qeneRaiion Tool^^
www.epa.gov/avert
Figure 31. Flowchart for converting wind speed data into hourly regional capacity factors.
6-hour windspeed dataset (m/s)
WS(t) = {s0,s6,s18,-,sn}
Hourly Interpolation
WS(t) = {s1,s2,s3,-,s5}
Hourly Average
_ S71.2011 + ^n,2012 + sn
Vestas wind power density curve
/Max MW
§
T
MW6 ~
Z3
MW5 -
o
MW4 -
-------
AVERT User Manual
Version 2.3
Appendix D: Overview of AVERT's Statistical Module
For each region, the MATLAB®-based Statistical Module provides the model's core statistical
analysis. (For installation instructions, see Appendix A.)
Data analysis within the Statistical Module is conducted in five steps, described briefly in the sub-
sections below:
1. Parsing the base year into "bins" of hours with similar levels of total regional fossil-fuel
load.
2. Collecting statistical information (probability distributions for generation, heat input, and
emissions) on how each fossil-fuel EGU has responded to regional load requirements in
each hour of the base year.
3. Extrapolating this statistical information to extend to potential lower and higher fossil-fuel
loads not experienced in the base year.
4. Estimating the ranges of generation, heat input, and emissions likely to be experienced by
each EGU for each fossil-fuel load bin (or approximate regional load).
5. Preparing outputs for export to AVERT's Main Module.
Appendix E provides step-by-step instructions to using the Statistical Module.
The Statistical Module is also equipped to estimate how fossil-fuel EGUs respond to regional load
requirements given changes in the fleet of EGUs available in future years. The module inputs
information from AVERT's Future Year Scenario Template to identify existing EGUs to be retired,
the expected impact of pollution-control retrofits on existing EGUs' emissions rates, and new EGUs
coming on line. AVERT then re-estimates all statistical information based on each region's
projected fleet of fossil-fuel EGUs for a particular future year scenario. (See Appendix F for a more
detailed description of the Future Year Scenario Template.)
Parsing Generation Demand into Fossil-Fuel Load Bins
In its first step, the Statistical Module sums up all fossil-fuel generation in each hour under analysis
to arrive at a total regional fossil-fuel load by hour (see Figure 32, which includes a detail of hours
3,000 to 4,000).70
70 Hour 3000 = May 5th. Hour 4000 = June 15th.
mm
State and Local Energy
and Environment Program
55
&EPA
-------
J&/&T
missions and qenaRaiion Tool^^
.UllkW.Ulld.uWtl.Ul 9
AVoidod Emissions.
AVERT User Manual
Version 2.3
Figure 32. 2011 hourly sum of fossil-fuel generation in the Texas region.
so.ooo
50.000
2 40,000
r 30.000
20.000
10,000
2,000
3,000
7,000
4,000 5,000
Hour of the Year
50,000
50,000
40.000
£ 30,000
20,000
10,000
3400 3500 3600
Hour of Year
These hourly sums of fossil-fuel generation are sorted from lowest to highest generation level and
grouped into 41 "fossil-fuel load bins" for the purpose of collecting statistics for each EGU at each
approximate load level (see Figure 33).71 Thirty-seven of the bins contain 224 or 225 hours; the
second lowest and second highest bins contain 204 or 205 hours; and the bins for the lowest and
highest fossil-fuel loads contain just 20 hours each to best represent these extreme load levels.72
Bin thresholds (the fossil-fuel load levels dividing the bins) and bin medians vary by region.73
Figure 33 illustrates how the bins are formed relative to total system fossil-fuel load. The figure
shows a typical "load duration curve" in dark red for fossil generation in Texas in 2011. This curve
represents all fossil-fuel load levels of the year (8,760 data points) in declining order. The horizontal
axis represents the fraction of time that fossil generation is at or above a certain level (e.g., fossil
generation only exceeds the 20 percent marker in 20 percent, or 1,752, hours). The vertical axis
shows the fossil-fuel load for each point on the curve.
71 "Load" always refers to regional, system-wide demand, and never to individual unit generation. The fossil-fuel load
bins group together hours that have similar generation levels ignoring their chronological order.
72 The ranges of the historical fossil-fuel load bins are determined is as follows: A region's 8,760 one-hour loads are
sorted from low to high, and then divided into 39 bins each containing 224 or 225 hours, depending on rounding.
For each bin (excluding the highest and lowest of the 39 bins, described below), the maximum threshold is the
MW load of the highest-load hour in the bin, and the minimum threshold is the MW load of the highest-load hour
the next lower bin. Bin "widths" are the high bin threshold in MW minus the low bin threshold. Bin medians are the
load (in MW) of the median hour of the bin. The highest and lowest of the 39 bins are each divided into two parts,
such that there are 41 fossil-fuel load bins from historical data in every region. The lowest of the 39 original bins is
split into the 20 hours with the lowest loads and the remaining 204 or 205 hours; the highest bin is split into the 20
hours with the highest loads and the remainder.
73 AVERT output includes additional fossil-fuel load bins designed to capture regional load levels that did not occur
in the base year (see the "Extrapolation to Higher and Lower Fossil-Fuel Loads" subsection below).
56
State and Local Energy
and Environment Program
&EPA
-------
kI/VeI
AVERT User Manual
Version 2.3
55,000
50,000
=- 45,000
l/l
O
-5 40,000
= 35,000
5
T3
ro
2 30,000
"qj
3
¦= 25,000
20,000
15,000
Fossil Fuel Load Duration Curve
Figure 33. 2011 fossil fuel load duration curve for Texas region indicating load bins.
60,000
10,000
0%
30% 40% 50% 60% 70%
Percentage of Year At or Above Load Level
100%
The horizontal axis has 42 light blue lines on it, representing the outside thresholds of the 41 fossil-
fuel load bins. Thirty-eight of these lines are evenly spaced from zero percent to 100 percent,74
capturing 224 or 225 hours each. In other words, each of these bins represents slightly over 2.5
percent of the hours in the year (again, grouped according to total fossil load in that hour rather
than by chronology). At the extreme ends, there are two additional light blue lines very near to the
zero and 100 percent markers. These additional lines fall 20 hours from the extremes; therefore
there are two bins at the extremes with 20 hours each, and two bins just prior to the extremes with
204 or 205 hours each.75
Wherever a percentage threshold crosses the fossil-fuel load duration curve, it creates a horizontal
line, representing a fossil-fuel load bin threshold. These are the horizontal grey lines shown on the
chart above, closely spaced in the lower middle of the graph and spreading out toward the highest
and lowest loads. This is because the majority of hours experience total fossil load that is neither
extremely high nor extremely low. In other words, there are more hours represented in the middle
of the fossil load range (for example, the regime from ~20,000 to ~32,000 MW in Figure 33) than at
very high or very low fossil loads. In order to capture an approximately equal number of hours in
each bin, each bin in the middle of the fossil load range captures a narrower range of MW. This is
shown in Figure 33 based on the spacing between the grey horizontal lines, where the points along
the load duration curve that fall between two horizontal threshold lines are the points in the
74 Each line represents a demarcation of 2.56 percent.
75 20 hours is represented by 0.23 and 99.77 percent on this axis.
57 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
corresponding fossil-fuel load bin, and the points in each bin (apart from the end binds, as
discussed above) represent roughly 2.5 percent of hours in the year.
Collecting Statistical Information
Next AVERT gathers statistics about how each EGU responds to the generation requirements of
each fossil-fuel load bin. Three types of probability distributions are constructed: frequency of
operation by fossil-fuel load bin, generation level by fossil-fuel load bin, and heat input and
emissions by generation level.
Frequency of Operation by Fossil-Fuel Load Bin
In this first set of probability distributions, AVERT calculates the share of hours within each fossil-
fuel load bin for which a particular unit is turned on (i.e., has generation greater than zero). Figure
34 shows the frequency of operation for three EGUs in the Texas region in 2011.
Figure 34. 2011 frequency of operation by fossil-fuel load bin for three indicative EGUs in the Texas
region.
W A Parish WAP6
701 MW
Victoria Power Station 9
286 MW
T H Wharton
73 MW
70%
60%
50%
40%
30%
20%
10%
^-oooooooooo
.V '
•
• •• • • #
•
•
••
•
•
*
•
•
•
#
•
%
—, , , ,
10,000 20,000 30,000 40,000 50,000
Fossil-Fuel Load Bins (MWh)
20,000 30,000 40,000 50,000
Fossil-Fuel Load Bins (MWh)
20,000 30,000 40,000 50,000 60,000
Fossil-Fuel Load Bins (MWh)
In the figure above, the 701-MW coal-fired EGU shown on the left operates in nearly every hour of
the year, with its probability of operation dropping below 90 percent only at the lowest fossil-fuel
load requirements. This pattern is typical of a baseload EGU that operates continually with the
exception of maintenance outages scheduled to occur at low load requirement levels. The middle
EGU, a 286-MW gas-fired station, operates only rarely at low load requirements, but its frequency
of operation increases steadily with regional demand. At fossil-fuel load levels above 40,000 MW,
this EGU operates in nearly every hour. This pattern is typical of an intermediate-load EGU such as
a combined-cycle EGU. The 73-MW gas turbine on the right is a peak-load EGU, operating only at
the highest load requirements.
Generation Level by Fossil-Fuel Load Bin
The second set of probability distributions calculated by AVERT describes generation output for
each EGU in operation in each fossil-fuel load bin.76 AVERT divides each EGU's generation into 19
evenly-spaced "unit generation bins."77 Figure 35 depicts the intersection of these two types of
76 For each fossil-fuel load bin, AVERT filters out the units which did not generate, and reviews only the operational
units.
77 The thresholds between unit generation bins are unit-specific.
B
58
State and Local Energy
and Environment Program
&EPA
-------
kI/VeI
AVERT User Manual
Version 2.3
bins. Smaller fossil-fuel load bins (where the vertical lines are closer together) indicate a higher
concentration of hours at those load levels.
Figure 35. Schematic of unit generation bins and fossil-fuel load bins.
Unit Generation Bins (MW)
(n = 19, width = maximum unit
capacity (MW) / 19)
II II II II II
Low Extrapolation
(n = varies by
region, width =
median historical
bin width)
Historical
(n = 41, width =
regional historical
fossil-fuel load
(MW), divided
into bins by hours)
High Extrapolation
(n = varies by
region, width =
median historical
bin width)
Fossil-fuel Load Bins (MW)
For each of the 41 fossil-fuel load bins, AVERT determines the number of hours in which the unit
generated at an amount within each of the 19 unit generation bins. In this way, the model creates a
discrete probability distribution of generation for each fossil-fuel load bin during all hours in which
the EGU is in operation.
Figure 36 shows the probability distributions of generation at two EGUs in the Texas region. The
axis to the bottom right of each plot represents the region's fossil-fuel load bins. The axis to the
bottom left represents unit generation bins, from zero to the EGU's maximum generation in the
base year. The vertical axis is the probability that the EGU is operating at the given unit generation
level in each fossil-fuel load bin.
I11H
State and Local Energy
and Environment Program
59
&EPA
-------
mdVEi
AVERT User Manual
Version 2.3
Figure 36. 2011 generation level by fossil-fuel load bin and unit
generation bin for two indicative EGUs in the Texas region.
W A Parish WAP6
A
J 0.4
10,000
,000
Lake Hubbard 2
B 533 MW
? 0.6 s
O 0.4 s
d: 500
10,000
State and Local Energy
and Environment Program
60
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
Heat Input and Emissions by Generation Level
The final set of probability distributions relate EGU heat input and PM2 5, SO2, NOx, and CO2
emissions to unit generation. For heat input and emissions of PM25, SO2, NOx, and CO2, statistics
for the ozone season and non-ozone seasons are gathered and stored.78 AVERT creates eight
discrete probability distributions—ozone season SO2, NOx, and CO2 emissions and heat input,
and non-ozone-season SO2, NOx, and CO2 emissions and heat input—for each EGU at each of
the 19 unit generation bins. In addition, PM2 5 data is calculated by multiplying the heat input data
by unit-specific PM2 5 emission factors, as explained above. Probability distributions are not a
function of regional fossil-fuel load. Figure 37 displays a single EGU's emissions of SO2 and NOx
relative to its generation level.
Figure 37. 2011 ozone-season emissions of SO2 (right graph) and NOx (left graph) by generation level
at an indicative EGU in the Texas region.
Oklaunion Power Station 1
MW
600
5 2 400.
C ^ 200.
'^>4 2000
700
200 reoe
-------
kI/VeI
AVERT User Manual
Version 2.3
level for each fossil-fuel load bin. A flexible number of fossil-fuel load bins are constructed below
the regional minimum load (with a lowest bound of zero), and above the regional maximum, such
that the new maximum bin threshold is the coincident maximum generation of all of the fossil-fuel
EGU on the system—that is, the level of load that could be reached if every fossil-fuel EGU were
operating at its maximum output.79 Bin thresholds and medians vary by region. Theoretically, the
regional extrapolated maximum can be reached by the simultaneous use of every EGU in the
system, but in practice load curves that reach this maximum are unlikely.
Extrapolating the Probability of Operation
The minimum amount of potential generation is zero, a level that would require zero generation
from all EGUs in the region. If an EGU is already at a zero probability of operation at a low load
requirement, the zero value is maintained into the lower potential fossil-fuel load bins. Any
probability of operation above zero is extrapolated linearly down to zero from the probability at the
lowest recorded load level.
The potential maximum generation is the combined simultaneous maximum output of all EGUs in a
region; to reach that maximum point, therefore, all EGUs in the region need to be operating at their
full capacity. EGUs that have a 100 percent probability of operation at the base-year's highest
fossil-fuel load bin maintain that probability of output. Any probability of operation lower than 100
percent is extrapolated linearly up to the potential maximum from the probability at the highest
recorded load level.
Figure
Texas
38. 2011 base year and extrapolated probabilities of operation for three indicative EGUs in the
region.
Tenaska Kiamichi CTGDB1
349 MW
Bayou Cogeneration CG804
81 MW
70%
Permian Basin CT3
89 MW
10,000 20,000 30,000 40,000 50,000 60,000
Fossil-Fuel Load Bins (MWh)
10,000 20,000 30,000 40,000 50,000 60,000
Fossil-Fuel Load Bins (MWh)
10,000 20,000 30,000 40,000 50,000 60,000 70,000
Fossil-Fuel Load Bins (MWh)
Black points represent the probability of operation during base-year load periods. Gray points are the
probability of operation at potential low and high loads beyond the base-year range.
Figure 38 displays extrapolated values for the probability of operation for three EGUs in the Texas
region. In this figure, black points represent the probability of operation at the base-year fossil-fuel
load, and gray points represent the probability of operation at potential high and low fossil-fuel
loads beyond the base-year range. Extrapolation is simple in the figure to the left (Tenaska), as the
EGU does not operate at all during the lowest loads and operates continually during the highest
79
The number of fossil-fuel load bins outside of the base-year range is determined as follows: For each region, the
median of the fossil-fuel load threshold times four sets the MW size of the extrapolated bins. Bins of this size are
extended below the base-year minimum until zero is exceeded and above the base-year maximum until the
coincident maximum peak load is exceeded. The lowest and highest bins are truncated to begin at zero and end
at the coincident maximum, respectively.
62
State and Local Energy
and Environment Program
&EPA
-------
AVERT User Manual
Version 2.3
load periods. The middle figure (Bayou) requires downward extrapolation to a zero probability of
generation at a zero load, and the figure to the right (Permian Basin) requires upward extrapolation
to meet the highest load requirements.
Extrapolating the Generation Level
EGUs not only run more often at higher load requirements, but also need to generate higher levels
of output to meet the requirements of the higher fossil-fuel load bins. Within the base-year range,
EGU generation is described as a series of discrete probability distributions for each fossil-fuel load
bin.
The extrapolated space encompasses fossil-fuel load bins that extend to the highest possible
coincident peak generation (i.e., all EGUs in a region operating at maximum capacity
simultaneously) and down to zero. New fossil-fired load bins are created at intervals equal to four
times the size of the median load bin.
To extrapolate to potential higher fossil-fuel load bins, AVERT assumes that each unit will have
100% probability of generating at its highest output if fossil load is at its theoretical maximum, and
0% of generating at any other level of output given maximum fossil load. For each level of
generation (i.e., within each unit generation bin), AVERT determines the slope of the line
connecting the unit's probability of generating at that level at the highest historical fossil load bin to
either 100% probability (for the highest unit generation bin) or 0% probability (for all other
generation bins). The unit's probability of generating at that level is then extrapolated accordingly.
Once all levels of generation have been extrapolated, AVERT normalizes the height of the
extrapolated load bins such that the total value of all points in each load bin sums to one. A similar
process is repeated for lower extrapolated fossil-fuel load bins, except that AVERT assumes that
the unit's output will be zero if total fossil load is zero.
The example in Figure 39 (below) shows an extrapolation of EGU generation to potential lower and
higher fossil-fuel load bins. The graphs show base-year unit generation bins (on the left-hand
horizontal axis) for any fossil-fuel load bin (on the right-hand horizontal axis). The height of the
surface represents the probability of operating at a given generation level in a particular fossil-fuel
load bin. Below about 15,000 MW (the lowest fossil-fuel load bin median in 2011) and above about
56,000 MW (the highest fossil-fuel load bin median in 2011), the surface is the figure showing data
for the historical year (A) is blank, indicating that no hours fell into those bin combinations in the
base year.
To extrapolate to higher and lower fossil-fuel load levels, a line is extended linearly towards the
corner constraints described above—100% probability of generating at full output given maximum
fossil load, and no unit output at a zero fossil load. The bottom graph (B) shows the results of this
extrapolation. From 60,000 MW to the peak load, this method returns generation exclusively at this
EGU's maximum, 701 MW. When the probability of operation is zero, the generation output is
automatically set to zero as well.
I11H
State and Local Energy
and Environment Program
63
&EPA
-------
mdVEi
AVERT User Manual
Version 2.3
Figure 39. 2011 base year (A) and extrapolated (B) probabilities of
generation levels for an indicative EGU in the Texas region.
W A Parish WAP6
653 MW
CD
Load
10,000
20,000
30,000
40,000
50,000
60,000
F0ss»^Bin
SSliC!
State and Local Energy
and Environment Program
64
&EPA
-------
AVERT User Manual
Version 2.3
Statistical Analysis
The fourth step in the Statistical Module is to estimate the predicted range and expected (average)
generation, heat input, and PM25, SO2, NOx, and CO2 emissions for each EGU at each of the
fossil-fuel load requirement bins, from zero MW up to the coincident maximum generation of all of
fossil-fuel EGUs in a region.
AVERT's Monte Carlo analysis (contained within the Statistical Module) uses discrete probability
distributions to estimate key variables' range and expected value for each EGU in each fossil-fuel
load bin. For each EGU and fossil-fuel load bin, AVERT draws three random numbers between
zero and one:
• The first number drawn is compared to the EGU's probability of operation at the selected
fossil-fuel load bin. If the number drawn is greater than the probability of operation, the
EGU is turned off and draws two and three are not conducted. If the number drawn is less
than the probability of operation, the EGU is turned on. Figure 40 illustrates this first draw.
Starting in fossil-fuel load bin number 15 (representing a particular system-wide fossil load
level), the simulator randomly draws a value of 0.25. This value is slightly lower than the
probability of operation in bin 15 (approximately 0.40), and this EGU is "turned on."
Figure 40. EGU frequency of operation and example of random draw selection.
• The second number drawn is compared to the discrete cumulative distribution function of
EGU generation for each fossil-fuel load bin. The model selects the EGU's unit generation
bin as the next highest EGU output greater than the cumulative probability value indicated
by the number drawn. Figure 41 illustrates this step: the random draw is 0.33, which results
in the selection of unit generation bin 12.
Random Draw#1
= 0.25
OOOOOOOOO
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41
Fossil-fuel Load Bin
I11H
State and Local Energy
and Environment Program
65
&EPA
-------
AVERT User Manual
Version 2.3
Figure 41. EGU generation histogram, cumulative probability distribution, and example
of random draw selection.
1.00
0.90
0.80
0.70
2,000
1,800
1,600
12
3
1,400
0
I
14—
1,200
O
> 1,000
c
a>
800
3
a-
£
600
LL
400
200
0
/
I
. /
/
v\
III
.
—i
Random
Draw #2
33
J
/
L
w
-j-l-f
p
1
1
1
1
1
1
0
1
C
=3
0.60 S"
0.50 gj 5
0.40 05 °
0.30 i
0.20
0.10
0.00
3
o
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Unit Generation Bin
The final number drawn is compared to the seven discrete cumulative distributions for heat
input and emissions in the unit generation bin identified in the previous draw. In Figure 42,
the third random draw is 0.80 and unit emissions bin 15 is selected.
Figure 42. EGU SO2 emissions histogram, cumulative probability distribution, and example
of random draw selection.
1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
10
(A
n
9
CM
0
8
CO
T3
7
a>
T3
i_
6
O
O
a>
5
£
0
4
>
0
3
c
a>
3
2
a-
£
1
LL
Random Draw #3
=0.80
1
V
/
/
/
/
f
/
r
~
/
—, , . jJ
s
llllllllllllll
0
>< V)
£=9
"E
~ o
CD O
1 ®
E a:
o
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Unit Emissions Bins (S02, Ozone Season)
mim
State and Local Energy
and Environment Program
66
&EPA
-------
J^VeRT
AVOidod Emission* and qcnaRation Tool^^
www.epa.gov/avart
AVERT User Manual
Version 2.3
Generation, heat input, and emissions output from each EGU at each fossil-fuel load bin is
recorded for 5,000 Monte Carlo runs.80 Each of these runs repeats the process described above of
drawing and applying three random numbers in sequence. Examples of the projected generation
and NOx emissions at a hypothetical fossil-fuel load of 30,000 MW for the 270 fossil-fuel EGUs in
the Texas region are shown in Figure 43 and Figure 44.
Figure 43. Generation (MW) for 1,000 Monte Carlo runs at 270 EGUs in the Texas region at a fossil-fuel
load of 30,000 MW (2011).
Monte Carlo Run
Figure 44. NOx ozone season emissions (lbs) for 1,000 Monte Carlo runs at 270 EGUs in the Texas
region at a fossil-fuel load of 30,000 MW (2011).
Monte Carlo Run
AVERT takes the average (expected value) generation, heat input, and PM25, SO2, NOx, and CO2
emissions for each EGU within a region across 5,000 Monte Carlo runs and records these values
in a new structural array.
Figure 45 depicts ranges of percent differences between fossil-fuel load bin medians and the
regional sum of generation in that same bin. For each of the 10 regions, three bar charts are
shown. The middle bar represents the minimum, mean, and maximum percent differences between
80 The base dataset provided by EPA uses 5,000 Monte Carlo runs. The default for users of the Statistical Module is
1,000 runs.
HSIffl 67 AFPA
State and Local Energy Km m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
measured or expected fossil-fuel generation81 (e.g., fossil-fuel load bin medians) and the sum of
generation for the bin as predicted by AVERT. The left bar in each region represents the same set
of statistics for the fossil-fuel load bins interpolated between 0 MW and the actual system
generation minimum, while the right bar represents the statistics for the load bins interpolated
between the actual system generation maximum and the potential system generation maximum.
Values close to zero indicate that AVERT has closely matched the amount of generation in one or
more fossil-fuel load bins, while values greater than zero indicate that AVERT has under-predicted
the amount of generation in one or more fossil-fuel load bins. Some bias for both high and low
interpolation bins is expected as EGUs experience both forced outages and imperfect responses to
regional demand; therefore, this statistical approach cannot capture a situation in which each EGU
in the system could reasonably be expected to generate at full capacity simultaneously. At the
lowest load levels, the generation output of many baseload EGUs becomes binary (i.e., the EGU
cannot ramp down below a minimum threshold, and simply turns off). AVERT therefore predicts a
large number of simultaneous curtailments, and may under-predict generation at low load levels.
Emissions and generation at these extrapolated load bins should therefore be viewed cautiously.
Underneath the name of each region is a set of numbers indicating the number of bins associated
with the low interpolation, the actual data, and the high interpolation, in that order.
Figure 45. 2011 percent difference between measured or expected EGU generation (fossil-fuel load bin
median values) and mean predicted AVERT generation by region. Mean across bins, and minimum
and maximum percent difference in any bin.
20.00%
15.00%
oj 10.00%
<_>
O)
-15.00%
c
_o
ro
c c
.2 .2
ro
-C
00
if
5
o
.c
DO
if
5
o
f
5
o
-C
00
if
5
o
.c
DO
if
5
o
.c
00
X
5
o
.c
00
if
5
o
-C
00
if
5
o
.c
00
X
California
Great
.ower
Northeast
Northwest
tocky
Southeast
Southwest
Texas
Upper
[9, 41, 7] Lakes/Mid-
Midwest
[2, 41, 9]
[2, 41, 3]
Mountains
[6, 41,
3]
[5, 41, 4]
[4, 41, B]
Midwest
Atlantic
[7, 41, 5]
[2, 41,11]
[5, 41, 5]
[6, 41, 8
Bracketed numbers are the number of high-interpolation, actual data, and low-interpolation fossil fuel bins.
Values above zero indicate that AVERT model under-predicts total sum generation.
81
"Expected" values for interpolated space.
68 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Statistical Output
The final step is to generate an Excel output file to store the expected values of generation and
emissions of each EGU at every level of base-year and potential load for each region, and a time
series of the total base-year fossil-fuel load in each hour of the year; this file becomes an input file
to the Main Module. Eleven sections of the output file each are composed of a matrix of the names
and identifiers of each EGU and the expected value at each fossil-fuel load bin, with one section
devoted to each of the following:
• Generation (MW)
• Heat input (MMBtu, ozone season)
• Heat input (MMBtu, non-ozone season)
• SO2 emissions (lbs, ozone season)
• SO2 emissions (lbs, non-ozone season)
• NOx emissions (lbs, ozone season)
• NOx emissions (lbs, non-ozone season)
• PM2 5 emissions (lbs, ozone season)
• PM2 5 emissions (lbs, non-ozone season)
• CO2 emissions (tons, ozone season)82
• CO2 emissions (tons, non-ozone season)
82 CO2 emissions are divided into ozone and non-ozone seasons to maintain algorithmic consistency with PM2.5, SO2
and NO* emissions. In AVERT results, displaced CO2 emissions are presented in terms of annual totals.
&EPA
mm
State and Local Energy
and Environment Program
69
-------
AVERT User Manual
Version 2.3
Appendix E: AVERT's Statistical Module: Step-by-Step
Instructions
This section provides step-by-step instructions for using AVERT's Statistical Module to prepare
inputs for AVERT's Main Module.
Step 1: Determine Windows Operating Environment
The Statistical Module is designed to work in a 64-bit operating system environment, so you will
first need to determine if your Windows system operates in a 32-bit or 64-bit environment.
Generally, this information is displayed among the "Properties" of "My Computer" in Windows XP,
or "Computer" in Windows Vista, Windows 7, or Windows 8. Instructions for determining your
Windows environment can be found at https://support.microsoft.com/en-us/help/15056/windows-7-
32-64-bit-fag.
Step 2: Download the Statistical Module Executable
Download the following two files and save them to your computer:
1) AVERT's Statistical Module executable package: "AVERT StatMod [Year]
64bit_package." Download this MATLAB executable at
https://www.epa.qov/statelocalenerqy/download-avert.
2) The MATLAB Compiler Runtime (MCR). Download the Windows 64-bit version of the
MCR for R2012b from the Mathworks website at
https://www.mathworks.com/products/compiler/matlab-runtime.html.
Do not download a newer version of the MCR even if one is available from the
Mathworks website. It is important to download the exact version listed here (R2012b, also
.known as version 8.0). This is necessary because the AVERT executable is packaged in a
form that can only be compiled by an MCR of the same vintage.
Once the AVERT package is downloaded, we
recommend creating a folder on your computer titled
"AVERT Statistical Module." Place the AVERT
executable package in the folder and run the file. The
package will decompress to three files and three
subfolders. These folders must stay in the same folder
as the Statistical Module for the program to operate
successfully.
The folders are:
• AVERT Future Year Scenarios: This folder contains the future year scenario template.
Other versions of the future year scenario template must be saved in this folder in order for
the AVERT Statistical Module to see them.
• AVERT Output: This folder will hold Regional Data Files generated by the Statistical
Module.
AVERT Future Year Scenarios
AVERT Output
CAMD Input Files
[¦ 1 AVERT StatMod 2012 vl G4bit_package
* AVERT_RegionNames
AVERT_StatM o d_2012_^j64 b it
_ readme
I11H
State and Local Energy
and Environment Program
70
&EPA
-------
AVERT User Manual
Version 2.3
• CAMD Input Files: This folder contains MATLAB-formatted flat data files with hourly
generation and emissions from each fossil EGU in the United States. The most recent year
of data is packaged by default with the Statistical Module. Other years of data can be
obtained from EPA, and must be put in this folder to be accessed by the Statistical Module.
The three files are:
• AVERT StatMod [Year] 64bit_package: the executable that will run the Statistical Module
once the MATLAB compiler is installed.
• AVERT_RegionNames: A library file required to run the Statistical Module. Do not remove
this file.
• readme.txt: basic instructions on the folders
and instructions for obtaining the MATLAB
compiler.
Step 3: Download CAMD Database
The Statistical Module package contains, by default,
the most recent data year of data. If another year is
desired, additional AMPD compatible with AVERT
are available at
https://www.epa.gov/statelocalenerqv/download-
avert. For most purposes, users will want to obtain
the most recent data year. Download the file and
save it in the subfolder "CAMD Input Files."
Step 4: Install MATLAB Compiler Runtime (MCR)
The Statistical Module executable requires additional, free MATLAB software in order to function.
For additional instruction on how to verify that the MCR is installed properly on your computer,
consult the readme.txt file. As noted on the previous page, it is critical to download and install the
correct version of the MCR. Using a different version will give you an error message when you try
to run the executable file.
Step 5: If Desired, Complete a Future Year Scenario Template
The Statistical Module can create either a base-year or a future year scenario (i.e., a scenario in
which some EGUs are retired, new EGUs are brought online, and other EGUs change emissions
rates). The process of creating a Future Year Scenario Template is described in Appendix F. A
template for the Future Year Scenario Template is available in the folder "AVERT Future Year
Scenarios."
Step 6: Launch the AVERT Executable
Click on the AVERT executable to launch the Statistical Module. A window labeled "Input for
AVERT Model" will open. In this window, select:
0 Input for AVERT Model i 1=1 1
Avoided Emissions and Generation Tool (AVERT) Statistical Module
Synapse Energy Economics, March 2013
Enter number of Monte Carlo runs:
tiro
Enter number of generation-only Monte Carlo runs:
500
Minimum annual generation to participate (MWh):
1000
Write output file?
Y
Please name this run.
OK Cancel
I11H
State and Local Energy
and Environment Program
71
&EPA
-------
AVERT User Manual
Version 2.3
• The number of Monte Carlo runs (default value is 1,000).
• The number of generation-only Monte Carlo runs (default value is 500).
• The minimum annual generation for an EGU to be considered in the model (default value is
1,000 MWh).
• Whether or not an output file should be written. Choose "Y" to create an input file for
AVERT's Excel-based Main Module; choose "N" to skip writing an output file (typically used
for test runs only).
• Designate a run name. This name will be part of the output file name.
Tip: the number of Monte Carlo runs directly influences how long a run takes to execute. To ensure
that inputs and outputs are correctly read, perform a test run with a small number of Monte Carlo
runs and generation-only Monte Carlo runs (10 each). For final runs where the output will be used
in the Main Module, use at least 1,000 Monte Carlo runs and 500 generation-only Monte Carlo
runs. The base dataset supplied by EPA includes 5,000 Monte Carlo runs.
State and Local Energy
and Environment Program
72
-------
AVERT User Manual
Version 2.3
Step 7: Choose a Base Year for Analysis
A window labeled "Choose CAMD Dataset" will open. In this
window, choose the data file for your desired base year. A
second window showing a progress bar will also be visible.
Q MENU
23
Choose CAMD Dataset
AVERT_CAM DArray_20 0 S. mat
AVE RT_CAM DArray_200 S. mat
AVE RT_CAM DArray_2010. mat
AVE RT_CAM DArray_2011 .mat
AVE RT_CAM DArray_2012. mat
Step 8: Choose a Base- or Future Year Scenario
A window labeled "Choose Future Year
Scenario" will open. In this window, choose a
base- or future year scenario for this analysis. A
second window showing a progress bar will also
be visible.
Note that each historical baseline year has a
unique Future Year Scenario Template. Use
only the one associated with the historical
baseline year of interest. In other words, having
chosen a 2012 base year and using a future
year scenario, ensure that the base year of the
template is also 2012. In the example here,
three scenarios have been created for the 2012
base year, titled "10PctRetire," "MidwestCTs,"
and "S02RateRed."
H MENU
Choose Future Year Scenario
AVERT Future Year Scenario 2012 v1.10 - 10PctRetire.xlsx
AVERT Future Year Scenario 2012 v1.10 - MidwestCTs.xIsx
AVERT Future Year Scenario 2012 v1.10 - S02RateRed.xtsx
AVERT Future Year Scenario Template 2008 v1.10.xlsx
AVERT Future Year Scenario Template 2009 v1.10.xlsx
AVERT Future Year Scenario Template 2010 v1.10.xlsx
AVERT Future Year Scenario iemplate2011 v1.10.xlsx
AVERT Future Year Scenario Template 2012 v1.10.xlsx
AVERT Future Year Scenario Template 2013 v1.10.xlsx
Present year analysis (no modifications)
State and Local Energy
and Environment Program
73
&EPA
-------
BsdVERT
AVOidod Emission* and qcnaRation Tool^^
www.epa.gov/avart
AVERT User Manual
Version 2.3
Step 9: Choose Region(s) of Interest
A window labeled "Choose one or more regions:" will open.
Choose the region for analysis and click "OK," or choose
"Select All" to launch an analysis of each of the 10 AVERT
regions in turn. Two additional windows showing progress
bars will be visible.
The selection of a region launches the full Monte Carlo
analysis, which can take up to several hours to complete
depended on computer processing speed, region selected,
and number of Monte Carlo runs selected.
A progress bar will inform the user of how far the program
has progressed through various analysis stages, with the
Monte Carlo runs usually taking the longest time. After all
Monte Carlo runs are complete, an output file will be written
if selected at the start of the program.
Each region is complete when the status bar reads "Finished
with [Region]." If more than one region is chosen, the
program automatically proceeds to the next region.
f I
1 ---- P
H
i=i
Choose one or more regions:
Southwest
California
Great Lakes I Mid-Atlantic
Northeast
Northwest
Rocky Mountains
Lower Midwest
Southeast
Texas
Upper Midwest
Select all
OK
Cancel
lEBffl 74 AFPA
State and Local Energy Km m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Appendix F: AVERT's Future Year Scenario Template
AVERT is equipped to estimate displaced emissions in scenarios projecting a future year by
making adjustments to the regional fleet of fossil-fuel EGUs before calculating the probability
distributions and expected values discussed in Appendix D.
The user implements these changes before running AVERT's Statistical Module using a
spreadsheet called "AVERT Future Year Scenario Template [Year] v.3.0," where year is the data
year associated with the file. The most recent historical baseline year template is stored in a
subfolder of the Statistical Module called "AVERT Future Year Scenarios." To access and use the
Future Year Scenario Template, download the Statistical Module, following instructions in Appendix
E. To access other historical baseline year datasets and future year scenario templates, visit
https://www.epa.aov/statelocalenerav/download-avert.
For each user-defined scenario, the user is strongly
recommended to save this file (in the same
subfolder) using the following naming convention:
"AVERT Future Year Scenario [Year] v.3.0
[Scenario X]" where "Scenario X" is a user-defined
name and "Year" is the historical baseline year.
The Future Year Scenario Template workbook
stores ancillary information about each EGU in the
system, and also includes worksheets that the user
modifies directly and then inputs into the Statistical
Module. This section describes AVERT's process for projecting three types of user-specified
adjustments to the fossil-fuel generation fleet:
• Retiring existing EGUs
• Adding additional "proxy" EGUs
• Changing emissions rates for existing EGUs to represent pollution-control retrofits
Users can make adjustments for multiple regions simultaneously; AVERT will correctly associate
each EGU with its appropriate region.
Please note that each historical baseline year has a unique Future Year Scenario Template.
Use the template associated with the historical baseline year of interest.
AVERT Future Year Scenarios
AVERT Output
CAMD Input Files
|i 1 AVERT StatMod 2012 vl 64bit_package
* AVERT_RegionNames
AVERT_StatM o d_2012_vl_64 b it
_| readme
Retirement of Existing EGUs
EGUs can be retired from the analysis using the "Retires_Modifications" worksheet. The user finds
the EGU of interest and selects "yes" in the dropdown menu under "Retire?"
Addition of Proxy EGUs
New EGUs can be added to the fossil-fuel fleet in the "Additions" worksheet. This process is more
complex than that for retirements and retrofits, and requires some knowledge of the types of EGUs
expected to be added into the system. To add a new EGU, the user finds a "proxy" EGU for which
statistics are already recorded in AVERT and modifies this proxy to meet the requirements of the
ffHK! 75 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
user-defined scenario. In most cases and in most regions, the diversity of EGUs is sufficient to
provide a proxy for most traditional fossil-fuel generating resources. If completely new types of
resources are to be added (i.e., advanced combined cycle, integrated gasification combined cycle,
or fossil-fuel backup for wind plants), the proxies available for selection may be insufficient.
In the "Additions" worksheet, the user copies existing rows to create one row for each new EGU
required (see Figure 46).
Figure 46: 2011 Screenshot of example EGUs in the "Additions" worksheet.
Ether select a county from the
dropdown, or enter manually
0
Region
Fuel
Type
Unit
Type
Unit
ORSPL
UNIT ID
Description
(Note that "0 MW" units did not run in 2011.)
Capacity
(MW)
State
County
Lat-
County
Lon -
County
1
TX
Gas
CC
Bayou Cogeneration Plant
CG802
10298
CG802
This is a 84 MW unit. It is located in Harris
County, TX. In 2011, it ran for 516 GWh at a
TX
Bastrop
30.126
-97.296
capacity factor of 70%.
2
TX
Gas
CC
Bayou Cogeneration Plant
CG802
10298
CG802
This is a 84 MW unit It is located in Harris
County, TX. In 2011, it ran for 516 GWh at a
capacity factor of 70%.
TX
Bastrop
30.126
-97.296
3
TX
Gas
CC
Bayou Cogeneration Plant
CG802
10298
CG802
This is a 84 MW unit It is located in Harris
County, TX. In 2011, it ran for 516 GWh at a
TX
Bastrop
30.126
-97.296
capacity factor of 70%.
Bayou Cogeneration Plant
CG802
Bayou Cogeneration Plant
Additions
This is a 84 MW unit It is located in Harris
County, TX. In 2011, it ran for 516 GWh at a
capacity factor of 70%.
This is a 84 MW unit. It is located in Harris
County, TX. In 2011, it ran for 516 GWh at a
30.126 -97.296
30.126 -97.296
Dropdowns EPA_AMP RefTables eGRID PLNT09 CapacityGen
For each proxy EGU, the user selects from dropdown menus in the relevant cells:
• The region in which the EGU is to be placed
• The fuel type (gas, oil, coal, or other)
• The unit type (combined cycle, combustion turbine, steam, or other)
The user then chooses an appropriate proxy from a list of available EGUs that meet these selected
criteria, and the worksheet automatically fills in the ORISPL code and Unit ID of the proxy and
creates a brief description including the base-year capacity, output, and capacity factor of the EGU.
Next, the user adapts the proxy to more closely meet requirements by selecting:
• The desired capacity of the new EGU (i.e., it does not have to be the same size as the
historical EGU)
• The state and county in which the EGU will be located
The worksheet automatically fills in the latitude and longitude center of that county as the new
EGU's default location. If a more precise location is known, the user can override this
latitude/longitude selection by manually entering the correct coordinates. The only purpose of this
location selection is to map EGUs in AVERT's Main Module. The latitude and longitude serve no
function in model calculations.
Pollution-Control Retrofits
Expected changes in emissions rates due to pollution-control retrofits are also made in the
"Retires_Modifications" worksheet. The user finds the EGU of interest, selects "Yes" in the
76 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
dropdown menu under "Revise Emissions Rates?" and inputs new rates in Ibs/MWh for SO2 and
NOx, in tons/MWh for CO2, and in tons/MMBTU for PM25 in columns I, J, K, or L, respectively. To
leave the rate for a particular pollutant unchanged, the user leaves the relevant cell blank. New
rates entered by the user must be greater than zero. These adjusted emissions rates will be
employed in AVERT as single point estimates of the mean rate; no probability distribution for
adjusted emissions is developed for retrofit EGUs.
Running Future Year Scenarios in AVERT
When running AVERT's Statistical Module, the user is presented a menu of future year scenario
files saved in the "FutureYearScenario" subfolder. The user can choose one of these files or select
"Present Year Analysis (no modifications)." If the user selects a "Present Year Analysis," the model
does not read or use any changes to the dataset, including retirements, additions, or changed
emissions rates. If the user selects a particular future year scenario, the retirements, additions, and
emissions modifications from that scenario's workbook are read into the Statistical Module. Once a
region has been selected for analysis, the Statistical Module reports the individual EGUs that have
been removed from or added to the region.
Note that each historical baseline year has a unique Future Year Scenario Template. Use the
template associated with the historical baseline year of interest.
Future year scenarios require an additional level of calculations before the five steps described in
Appendix D can be carried out. For each fossil-fuel load bin, the average generation level of each
EGU (retired and active, and including new units) must be determined in a separate Monte Carlo
analysis. The results of this analysis change the system fossil-fuel load perceived by all of the
remaining EGUs to generate the correct output. Because the generation of each EGU is
independently derived, each unit's generation is not affected by the generation levels of other
EGUs.
If the net change to generating capacity of retiring old and adding new EGUs results in an increase
in total generation, the region will incorrectly generate an amount greater than the required fossil
load. AVERT determines how much to back down the "perceived" fossil-fuel load in each bin to
output the appropriate amount of generation for that bin.83 For net reductions in generation, the
algorithm is simply reversed: perceived system fossil-fuel load is increased, allowing each EGU to
generate more than it otherwise would and make up the gap left by retired EGUs.
83
In this case, "perceived" load is the fossil-fuel load bin for which the model assigns generation and emissions.
ffHKI 77 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Appendix G: AVERT Regions and Instructions for States
that Cross Regional Boundaries
AVERT regions are aggregates of EPA's eGRID subregions and based on regional boundaries
used by the North American Electric Reliability Corporation (NERC). Each region generally
represents an electrically autonomous area. Each EGU is assigned to exactly one AVERT region.
These assignments may change over time as a result of changes in the dynamics of supply and
demand. The current version of AVERT assigns EGUs to regions based on the assignments in the
2010 edition of eGRID.
Electrical boundaries do not necessarily represent political boundaries, and as such only 26 states
and the District of Columbia are encompassed entirely within one AVERT region. The remaining 22
states in AVERT (Hawaii and Alaska are excluded) are split across AVERT boundaries. With the
exception of Missouri and Oklahoma, which are split across three AVERT regions, and Texas,
which is split across four, the remaining states are split across two AVERT regions. Refer to Table
1 (on page 16) for the AVERT regions and the states they contain, either in whole or in part.
This section provides instructions for states that are split across more than one AVERT region.
AVERT results represent the impacts of the programs only on generators that are contained within
that AVERT region. AVERT regions are defined not by state geography but by the generators that
fall within their borders. To capture the impacts of a state-wide EE/RE program across two or more
AVERT regions, the EE/RE program must be parsed between the two (or more) AVERT regions
straddled by the state. It is often not possible within a state to readily ascribe different load centers
or EE/RE programs to one region or another. Instead, as a rule of thumb, the impacts of EE/RE
programs are assigned, pro rata, to AVERT regions based on the proportional generation provided
by EGUs in each AVERT region.
Figure 47 shows a schematic of such an example. In this case, State A crosses AVERT regions 1
and 2, and thus is only partially represented in each. However, the vast majority of State A's
generation is located in and serves AVERT region 2. With some exceptions, as an approximation,
the effects of the EE/RE program should be split into the two AVERT regions ratably, such that 90
percent of the program (represented by 90 percent of the generation) is run within AVERT region 2,
and 10 percent of the program is run within AVERT region 1.
Figure 47. Schematic of recommendation to states that cross AVERT regions.
AVERT Region 1
State A
AVERT Region 2
The exception to this rule is if the user has explicit knowledge of the location of new EE/RE
programs and can readily identify the region into which they will fall using the map in Figure 3.
Table 4 indicates, by state, the approximate fraction of fossil generation found in each AVERT
region. This table was constructed by reviewing how much fossil electricity was generated from
2010 to 2013 (inclusive) in each AVERT region.
ffHK! 78 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
EPA recommends that states that are split between more than one region execute the proportional
split for any AVERT regions that contain more than 5 percent of the state's generation. By this
logic, 16 states require multiple AVERT region runs. For example, New Mexico and Oklahoma
have generation in multiple AVERT regions, but only one of the AVERT regions in which they are
located has more than 5 percent of the state's generation; therefore, the other regions can be
excluded under most circumstances.
For example, an air quality planner in Arkansas reviewing the displaced emissions benefit of 2,000
MW of wind would run AVERT twice—once with 1,780 MW in the Southeast region, and once with
220 MW in the Lower Midwest region—and then aggregate the results of these runs.
A New Mexico air quality planner, conversely, would run AVERT only once, with all of the EE/RE
attributed to the Southwest region, which is in the Western Interconnect. While New Mexico does
have territory in the Lower Midwest region, which is in the Eastern Interconnect (SPP RTO), it has
relatively little generation in that region. However, wind developments in New Mexico are near the
eastern boundary of the state, which is located in the Lower Midwest region. If a planner wanted to
review the impact of wind in eastern New Mexico, it would be more appropriate to run AVERT for
the Lower Midwest region than for the Southwest region, recognizing that little of the displacement
will occur within New Mexico.
Texas covers four AVERT regions, but the fraction of Texas generation that occurs within the
Southwest region is small (1 percent), and therefore EPA recommends that all EE/RE be parsed
into the other three regions: Texas, Southeast, and Lower Midwest. To determine the fraction of
EE/RE attributable to each region, users should normalize the remaining fractions such that they
add to 100 percent; in Texas's case, this means dividing each region's fractional coverage by 99
percent. EE/RE in the Texas AVERT region should be apportioned as:
• Texas: 82 percent (or 81.6 percent divided by 0.99, rounded).
• Southeast: 6 percent (or 6.0 percent divided by 0.99, rounded).
• Lower Midwest: 12 percent (or 11.7 percent divided by 0.99, rounded).
— 79 mFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
J^Vef
AVERT User Manual
Version 2.3
I AVERT
I Fir yi:i»~. :"I:i ¦ [•¦IH-R.ll inr TiidI^^
Table 4. State apportionment in AVERT regions, based on fossil generation from 2010 to 2013.
State
(# regions)
Northeast
Great Lakes / Mid-
Atlantic
Southeast
Lower Midwest
Upper Midwest
Rocky Mountains
Texas
Southwest
Northwest
California
Alabama
100.0%
Arkansas (2)
88.7%
11.3%
Arizona
100.0%
California
0.3%
99.7%
Colorado
100.0%
Connecticut
100.0%
District of Columbia
100.0%
Delaware
100.0%
Florida
100.0%
Georgia
100.0%
Iowa
100.0%
Idaho
100.0%
Illinois (2)
38.8%
61.2%
Indiana
100.0%
Kansas
100.0%
Kentucky(2)
9.4%
90.6%
Louisiana (2)
76.1%
23.9%
Massachusetts
100.0%
Maryland
100.0%
Maine
100.0%
Michigan
99.6%
0.4%
Minnesota
100.0%
Missouri (3)
21.0%
33.8%
45.2%
Mississippi (1)
98.9%
1.1%
Montana (1)
2.3%
97.7%
North Carolina
100.0%
North Dakota
100.0%
Nebraska
100.0%
New Hampshire
100.0%
New Jersey (2)
23.4%
76.6%
New Mexico (1)
2.9%
97.1%
Nevada (2)
72.0%
28.0%
New York
100.0%
Ohio
99.7%
0.3%
Oklahoma (1)
4.1%
92.8%
3.1%
Oregon
100.0%
Pennsylvania
100.0%
Rhode Island
100.0%
South Carolina
100.0%
South Dakota
99.7%
0.3%
Tennessee
100.0%
Texas(3)
6.0%
11.7%
81.6%
0.7%
Utah (2)
65.1%
34.9%
Virginia (2)
5.1%
94.9%
Vermont
100.0%
Washington
100.0%
Wisconsin (2)
45.2%
54.8%
West Virginia (2)
87.7%
12.3%
Wyoming (2)
38.3%
61.7%
I11H
State and Local Energy
and Environment Program
4>EPA
-------
AVERT User Manual
Version 2.3
Appendix H: Frequently Asked Questions
Web-Based AVERT
What are the differences between the web-based and Excel-based versions of the
AVERT Main Module?
The downloadable Excel-based Main Module and the web-based version (both available at
www.epa.gov/avert') use the same underlying methods, calculation algorithms, and regional data
inputs, and they reflect the same assumptions. The web version offers much of the same
functionality as the Excel version, but with a streamlined interface and a somewhat more limited
range of input and output options. Specifically:
• The web edition relies on the most recent year of input data, whereas the Excel version
can incorporate data for any base year from 2007 forward.
• The web version does not allow the user to manually input a custom load profile with 8,760
hourly values, like the Excel version does.
• The web version uses built-in default RDFs, so it does not support future year scenario
planning.
• The web version provides a more streamlined range of outputs than the Excel version,
consisting of two data tables, three graphs, downloadable CSV data, and a COBRA CSV
file. However, the graphs in the web version have some dynamic capabilities that allow the
user to customize the geographic area displayed and save a variety of formats (e.g., jpeg,
pdf) to display in presentations or reports.
The web-based version provides some accessibility advantages because it can run in any major
web browser, without the need for Excel software or the need to download, save, and upload
separate RDFs. Ultimately, some users will find that the web edition meets their needs, while
others who wish to use different data years, custom load profiles, future RDFs, or additional
outputs should use the Excel version.
Renewable Energy and Energy Efficiency
Are users restricted to the EE/RE profiles created within AVERT's Main Module?
No. The Main Module maintains simple wind and solar profiles for various regions of the United
States for the convenience of users, but does not restrict users to these profiles. Users are
encouraged to create EE/RE profiles that reflect their regions and assumptions. Such profiles can
be copied into the manual entry page of the Main Module.
Can I review renewable energy options other than wind and solar generation?
Yes. New non-intermittent, must-take renewable generation, such as hydroelectric generation or
geothermal generation, can be approximated using either the manual hourly data entry or the
preset EE impact sections. For example, if you want to model the impact of a new hydroelectric
generator, you could click on "Enter hourly data manually" in AVERT's Step 2 and enter the
expected hourly generation curve. If you assume the non-intermittent resource functions as a
purely baseload resource, you could use the "annual GWh" setting as a proxy.
81 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Is there a way for baseload renewables to be included?
You can model non-emitting, must-take baseload renewables like geothermal or hydroelectricity in
AVERT using the "annual GWh" setting in Step 2.
How do you handle biomass, waste combustion generators, or combined heat and
power (CHP) generators in AVERT?
If biomass, waste combustion, or CHP generators are emitting and have capacities greater than 25
MW, they are included in the EPA's Air Markets Program Dataset (AMPD).
AVERT is not currently equipped to estimate the emissions of emitting generators that do not report
to AMPD. However, if you know the expected generation and emissions from a new biomass,
waste, or CHP generator, you could review the estimated displaced emissions and generation from
the inclusion of that generator using AVERT (assuming an hourly load impact shape is known for
the new EGU) and then add in that generator's emissions post-hoc. To do so, follow the steps
below:
1. Determine estimated load impact profile for CHP generator and associated stack
emissions.
2. Input load impact profile for CHP generator into AVERT under "manual EERE data entry"
in Step 2.
3. Run Main Module to determine emissions offset due to new CHP generator.
4. Subtract CHP stack emisssions from emissions offsets to determine total emissions
impact.
Net emissions reduction from CHP generator = AVERT displaced emissions + CHP stack
emissions.
There is no current option to review emissionsdisplaced from new biomass, waste, or CHP
generators if they do not already report to AMPD.
Are there any plans to incorporate electricity production from biogas facilities into the
Tool?
If the facility is an emitting generator and has a capacity greater than 25 MW, it is currently included
within the AMP EGU dataset. Otherwise, there are no current plans to incorporate electricity
production from these types of facilities. Often, these facilities may generate electricity according to
their on-site needs and fuel supply and may not be affected by regional changes in load or
dispatch.
Can storage technology be modeled in AVERT?
Yes. Energy storage, such as batteries, pumped hydroelectric generation, and compressed air
storage, is used to capture excess energy at low-cost (often low-demand) periods, and release that
energy at high-cost (often high-demand) periods. AVERT draws statistics from, and dispatches
against, a base year time-series of demand. You can create an energy storage assumption (how
much energy is drawn off a renewable resource, when it is released, and what the associated
losses are) and create a manual EE/RE profile that reflects this energy storage assumption. As
with all other EE/RE assumptions in AVERT, users are encouraged to create a time-series of
EE/RE generation that fits their region and assumptions.
82 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
kI/VeI
AVERT User Manual
Version 2.3
The U.S. Department of Energy has an emissions calculator that uses AVERT as its engine. This
web-based tool is called Grid Impact Emissions Quantification (GRIDPIQ). It specifcally models
storage and other distributed energy resources scenarios; however, the outputs are only regional.
For more information, visit https://qridpiq.pnnl.qov/app/#!/.
How does AVERT account for the dispatch of new RE into the existing system?
RE generation sources typically have very low variable operating costs; in other words, they are
very inexpensive to operate once they are constructed. Typically, low-operating-cost resources are
dispatched first, and increasingly expensive resources are dispatched thereafter. RE sources are
assumed to dispatch first (a common assumption across many economic dispatch models), and
thus can be modeled as an equivalent reduction in demand. AVERT simply compares the
generation and emissions of all fossil resources before the new RE resource (i.e., at the equivalent
of full demand in each hour) and after the new RE resource (i.e., at the equivalent of a reduced
demand in each hour). The difference in generation and emissions between the before and after
scenarios represents the emissions displaced by RE.
How does AVERT account for the dispatch of new EE into the existing system?
EE usually results in a reduction in demand (in some cases and for some types of programs, it may
result in a shifting of demand to off-peak hours). AVERT simply compares the generation and
emissions of all fossil resources before the new EE resource (i.e., at the equivalent of full demand
in each hour) and after the new EE resource (i.e., at the equivalent of a reduced demand in each
hour). The difference in generation and emissions between the before and after scenarios are the
emissions that are displaced by EE.
Is there a bound on the smallest EE/RE program that is appropriate to model?
No. In the current version of AVERT, users can review the output chart titled "Hourly Contribution to
Reduction in Generation" for an indication of how closely their expected reduction is captured in
hour-to-hour unit reductions. At very small scales of production, this graphical interface will indicate
a rougher hour-to-hour displacement impact profile—i.e., the amount of generation displaced will
look less like the amount of EE/RE implemented. For a more comprehensive check, the user
should view the "Signal-to-noise diagnostic," found on the "Display Outputs" page of AVERT's Main
Module. As described in this manual (p. 44), this scatter plot shows the generation reduction
calculated by AVERT (on the y-axis) against the EE/RE load reduction implemented by the user.
More reasonable results (from a program size perspective) will appear closer to 1:1 lines. Smaller
load reduction have more noise (i.e., scatter) in this plot, while larger load reductions have a
straighter line relationship. The R2 value in the title of the chart indicates how much of the
generation reduction can be explained by the EE/RE load reduction. For examples, an R2 value of
0.9 indicates that AVERT has captured 90 percent of the generation reduction required by the user,
while a value of 0.7 indicates that AVERT has only correctly captured 70 percent of the EE/RE
required by the user (i.e., noise accounts for 30 percent of the observed variability).
Figure 48 below shows two different load impact profile with very different R2 values from the same
region, and designed similarly. The graph on the left is a 1.5 percent load reduction during the peak
20 percent of hours. The reduction is sufficiently sized such that the generation reduction is able to
match the requirement very closely—over 99 percent of the reduction in generation is a direct
result of the input EE/RE program. The graph on the right is a 0.25 percent load reduction during all
hours of the year. The reduction is insufficiently sized in this case and results in a wide range of
ffHKI 83 **FPA
State and Local Energy m m
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
uncertain results. Only 88 percent of the generation reduction can be attributed to the EE/RE
program—the rest is noise.
Figure 48. Examples of two different load reductions with different R2 values in the signal-to-noise
diagnostic. Left: 1.5 percent load reduction in peak 20 percent of hours. Right: 0.25 percent load
reduction in all hours.
Reduction in Total Unit Generation Relative to EERE Load Reduction
(MW) for All Hours, R2= 1.00
Reduction in Total Unit Generation Relative to EERE Load Reduction
(MW) for All Hours, RJ=0.88
•
.
J
0
0^ *
'.
Note that all numerical results are shown rounded to the nearest 10 unit.84 Dashes indicate
that AVERT reported a value greater than zero, but lower than the level of reportable significance.
In some cases, no reasonable sized EE/RE program will result in reportable reductions. For
example, the review of monthly displacement output for a single small county in a low load month
may often result in low significance results. However, the user can use discretion to determine if an
EE/RE program has resulted in an acceptable level of significance based on the signal-to-noise
diagnostic and the degree to which critical results are below the level of acceptable significance
(i.e. are obscured by dashes in numerical results).
Is there a bound on the largest size EE/RE program that is appropriate to model?
There is not a formal bound on the largest size EE/RE program that should be modeled in AVERT.
Realistically, at very low loads (i.e. below those historically experienced in the base data year),
AVERT will generally under-predict the expected generation reductions achieved by EE/RE
programs, and thus under-predict emissions changes. See Figure 45 for details on biases at low
extrapolated load bins. It is recommended that programs generally not exceed 15 percent of fossil
generation in any given hour. Users should note that AVERT is designed to review marginal
operational changes in load, rather than large-scale changes that may change fundamental
dynamics.
In addition, users may encounter situations in which their selected change to load produces a new
hourly load that is outside the range calculable by AVERT. For load displacements, this may occur
in situations where load is reduced by 50 percent or more relative to the hour with the lowest load
(i.e., well outside the recommended threshold of 15 percent). In situations where load is increased,
users may encounter this issue at load increases from 3 to 14 percent. In these situations, users
84 Data are reported by AMPD in integer units of MWh (generation), lbs (NO* and SO2), tons (CO2) and MMBtu (heat
input). Output data in AVERT are rounded to the closest 10 MWh, lbs NO* and SO2, tons CO2, and MMBtu fuel
input.
84 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
should refine their load changes such that an error is not produced on the "Manual User Input"
page.
For reference, Table 5 provides each AVERT region's total annual load, in GWh. Knowing the total
annual load in each region may be helpful when using AVERT's "reduce generation by X% in all
hours" option to model the impacts of a broad-based EE program. Figure 49 identifies the
distribution of hourly loads in each of the 10 AVERT regions, using data from 2018. Table 5 also
shows the maximum and minimum possible load levels able to be modeled in AVERT. These
values provide helpful context if one knows the absolute size of a project, policy, or program in
units such as MW and one wants a sense of the percentage of regional load that it represents.
Note that fossil loads in Table 5 and Figure 49 have not been adjusted to reflect the T&D losses
inherent to each region. Demand-side measures (such as energy efficiency or distributed solar)
avoid not only electricity demand, but also the electricity associated with T&D losses. As a result,
demand-side programs increase the avoided fossil load by an additional 5-11 percent, depending
on the region and the year being analyzed.
Table 5. Total Regional Fossil Loads in AVERT Regions, 2018
AVERT Region
Total Annual
Fossil Load (GWh)
Maximum Possible
Hourly Load (MW)
Minimum Possible
Hourly Load (MW)
California
76,643
26,311
1,690
Great Lakes/Mid-Atlantic
551,481
119,048
31,826
Lower Midwest
149,096
36,986
6,178
Northeast
102,486
32,547
3,040
Northwest
105,052
19,517
3,413
Rocky Mountains
49,765
9,725
2,962
Southeast
790,693
156,567
46,854
Southwest
113,305
22,869
6,258
Texas
269,072
60,417
10,239
Upper Midwest
257,932
49,417
12,815
I11H
State and Local Energy
and Environment Program
85
&EPA
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
www.epa.gov/avert
AVERT User Manual
Version 2.3
Figure 49. Characteristics of Regional Hourly Fossil Loads, 2018
3
0
= 5
1 =
o
(U
oo
c
ns
cc
175,000
150,000
125,000
100,000
75,000
50,000
25,000
0
b 0 a
a
^ ^ ,
s / / / / y
B
* .#
r jr ^
^ /
Maximum
Average
Minimum
AVERT Region
Are EE/RE reductions applied over the whole region? Is there a way to apply them at
the state, county, or municipal level only?
Because AVERT does not model transmission constraints within a region, EE/RE reductions are
assumed to have region-wide impacts. A limitation of AVERT is that it is insensitive to the physical
location within a region of new EE/RE programs, despite the fact that real-world dispatch decisions
may be quite sensitive to specific locations of new EE/RE resources as well as EGUs. AVERT
assumes that EE/RE programs are spread across the modeled region and cannot currently identify
the differential impacts of local versus regional EE/RE programs. Such differentiation requires the
use of a production cost model.
Please see the user manual, Chapter 2, Section "Limitations and Caveats," for more information.
Detail on displacements at the state and county level are available on the output sheet "Annual
Displacement Data by County."
AVERT Outputs
Does AVERT account for losses?
Yes, AVERT accounts for two types of losses, starting with Version 1.6 of the Main Module. First,
gross generation as collected in the AMPD database is corrected to account for parasitic
consumption of energy onsite at fossil-fired EGUs. AVERT applies a parasitic loss factor to each
EGU based on unit and fuel characteristics and subsequently calculates emissions based on each
unit's "net" output of energy exported to the grid. Second, reductions in fossil load due to energy
efficiency and distributed photovoltaic are corrected to account for avoided grid (transmission and
distribution) losses, using region-specific, year-specific grid loss factors. Wind and utility-scale PV
86 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
profiles are not corrected for these losses, as it is assumed they are located at a similar distance
from load centers as fossil-fired EGUs.
How can users assess the accuracy of the results returned by AVERT's Main Module?
The current version of the Main Module is not equipped to return information on the accuracy or
uncertainty of the model results.
While the Monte Carlo analysis run by AVERT creates information useful for some forms of
uncertainty analysis, using this information to assess the accuracy of the results returned by the
Main Module would require simultaneously performing a Monte Carlo analysis on the baseline
scenario and on the EE/RE modified scenario, and returning uncertainty metrics associated with
the difference between these two scenarios. The current version of AVERT does not contain this
information. EPA is exploring a future version of AVERT that could perform explicit uncertainty
analyses and allow users to assess the accuracy of results returned by the model.
What is the accuracy of the map chart in AVERT's Excel-based Main Module?
The map is a visual cue only, and should not be used as a precise rendering of the location or
influence of displaced generation or emissions from any EGU or cohort of EGUs. Maps could be
used for visual presentations to show the general location of emissions.
Why do some EGUs show positive increases in generation with decreases in system
load?
Some EGUs show positive increases in generation with decreases in load because the EGU
statistics indicate either a slight increase in the probability of operation at very low loads, or an
increase in generation at very low loads. Spot checks indicate that most of the EGUs that show
generation increases with decreases in system load are due to baseload EGUs that show a lower
probability of generation at mid-range loads then at either very high or very low loads. In other
words, these EGUs counterintuitively increase the probability of operation as system load levels
become very low. Further inquiry into these EGUs suggests that they have prolonged maintenance
outages during spring or autumn—i.e., during periods of generally low load, but possibly not the
lowest in the year. Therefore, the EGU will register as non-operational through a wide swath of
medium-low loads, but may operate during the very lowest loads of the year. Therefore, the
statistics capture this behavior and increase expected generation by a small margin when system
load is reduced from very low load periods. This pattern is almost always observed in trough
periods.
AVERT Statistical Module
Why is the model driven by system fossil generation instead of by demand or total
load?
AVERT is a statistically based model that tracks and reproduces EGU behaviors. EGUs are
forecast to operate in the near future much as they operate today. In electrical system dispatch,
determing how an EGU operates is largely driven by two factors—total demand on the system and
the cost of operation for any given EGU. In economic dispatch, more expensive EGUs are
dispatched at higher levels of demand.
87 **FPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
In AVERT, the degree to which an EGU will be dispatched in the future is assumed to be the same
as the degree to which it has been dispatched at a historical level of demand .This assumption
incorporates the relative operational cost of different EGUs. In other words, EGUs that dispatched
at high demand were likely more expensive to operate and thus are likely to be on the margin at
the same level of demand.
However, if AVERT were to observe the behavior of EGUs against total system demand, it would
miss an important factor: the large number of non-fossil EGUs that are dispatched at very low
operating costs—such as solar, wind, hydroelectric, and nuclear operations.These non-fossil EGU
have the same effect as lowering system demand, or specifically, system demand that needs to be
met by fossil generation. Therefore, AVERT dispatches against demand for fossil resources, rather
than total system demand. New EE/RE programs are assumed to reduce the demand for fossil
resources. Figure 50 below illustrates the difference between total system demand and demand for
fossil resources.
Figure 50. Diagram schematic of system demand over two days, divided into fossil and non-fossil
components illustrating system and fossil demand.
a
E
Total System
Coal
Demand
Oil
Nuclear
Hydro
Total Fossil
Demand
Does the sum of all unit generation in any given Monte Carlo run add up to the size of
the load bin (i.e., the expected fossil generation)?
Not necessarily. The generation from each EGU is calculated independently in each Monte Carlo
run, meaning that there is no constraint that forces the output of all EGUs to equal the exact size of
the load bin. The total sum of all unit generation from any given Monte Carlo run may be slightly
larger or smaller than the load bin. However, over large numbers of Monte Carlo runs, the average
output of each EGU will sum quite closely to the expected fossil generation, or the size of the load
bin.
Is it possible to displace baseload EGUs in AVERT?
Yes. AVERT treats baseload EGUs, or units that run during most hours, including during baseload
hours of the year, the same as all other EGUs. There is no distinguishing characteristic that either
promotes or prevents an EGU from running during any given hour, except for how it has operated
in the past. If an EGU has experienced little downtime in the past and operates continuously even a
low levels of load, the model will replicate this behavior going forward. This type of EGU is unlikely
to be displaced by EE/RE in AVERT. However, an EGU that ramps from high output in the daytime
to low output on offpeak hours may show a displacement if system demand declines due to RE or
EE.
88 ipsiFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
AVERT User Manual
Version 2.3
Can AVERT capture or replicate ramping behavior?
No. AVERT performs a separate calculation for each hour of the year and does not evaluate the
rate at which an EGU increases or decreases generation. Capturing this behavior requires a
chronological dispatch model.
Can AVERT capture or replicate spinning reserves behavior?
Yes. AVERT captures historical generation patterns. EGUs that maintain a spinning reserve (i.e.,
maintain a minimum level of generation in most operating hours) will reflect this pattern in the
statistics gathered by AVERT. The Lake Hubbard EGU shown in Figure 36 is a classic example of
an EGU that appears to maintain a spinning reserve. It maintains an output of about 150 MW per
hour for most hours of the year. However, when system demand climbs above 45,000 MW, it
quickly climbs towards an output of 500 MW.
Can AVERT capture transmission constraints or changes in transmission?
Generally, no. AVERT operates on the simplifying assumptions that there are no transmission
constraints between load centers and EGUs within a region and that regions are independent of
each other. Therefore, AVERT is insensitive to the location where new EE/RE resources are
placed within a region, and thus does not capture transmission constraints. However, the behavior
of some EGUs may be influenced by historical transmission constraints, and this behavior is
captured by AVERT. For example, in "load pockets," or areas of constrained inbound transmission,
reliability EGUs may run at lower regional load levels than would otherwise be dictated by
economic dispatch. Because AVERT is not an economic model, it simply replicates the behavior of
these EGUs, which may capture some elements of current transmission constraints.
Due to the same simplifying assumptions that prevent AVERT from operating as a transmission-
constrained dispatch model, AVERT cannot capture future changes in transmission, which typically
change which future EGUs can compete to provide the lowest-cost energy in a particular area.
Are recent fuel prices reflected in AVERT?
Yes. To the extent that fuel prices have influenced dispatch during the base data year you choose,
AVERT will reflect those dispatch decisions. AVERT cannot, however, change dispatch based on
future economic or regulatory conditions, such as expected fuel prices, emissions prices, or
specific emissions limits. AVERT should not be used for this type of analysis, as such changes
require an economic dispatch model.
Are emissions control technologies reflected in AVERT?
Yes. To the extent that emissions controls were in operation at the time that data was collected in
the base data year, emissions will reflect operational (and operating) control technologies. To the
extent that a user requires a review of dispatch with different emissions rates, they can override
observed emissions rates using the "Future Year Scenario Template" as described in Appendix F.
Modeling emissions prices, specific emissions limits, or fuel switching requires an economic
dispatch model.
Are predicted changes in fuel or emissions prices reflected in AVERT?
No. AVERT should not be used for this type of analysis; capturing this behavior requires an
economic dispatch model.
— 89 mFPA
State and Local Energy m m
State and Local Energy
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
What other tools are available to me to estimate displaced emissions aside from
AVERT?
You can model generation and emissions displaced by new EE/RE programs in a production-cost
dispatch model. These programs simulate real dispatch decisions based on explicit costs and
operational constraints, optimizing generator use to minimize costs. Some of these models are
sensitive to EGU ramp-rates, transmission constraints, and outage schedules.
For the most part, these models are highly detailed, proprietary, and require specialized labor and
licensure to operate and use, as well as some degree of proprietary knowledge for fuel costs and
operational constraints.
AVERT provides an alternative, publicly available tool to estimate displaced emissions in near-term
years. Users who wish to conduct analyses more than 5 years from the baseline must use
AVERT's statistical module and future year scenario template. This type of analysis requires
access to future-year hourly, unit-specific generation and emissions data (e.g., from an electric-
sector dispatch model designed to forecast future generation) that can be entered in place of
AVERT's historical data.
Future Year Scenarios
Why is there a different future year template for each historical baseline year?
AVERT is sensitive to the composition of the electric fossil fuel fleet. Every year, the composition of
the fleet changes slightly as new EGUs are added or retired. To accommodate this changing fleet,
AVERT creates a new future year scenario template for each historical baseline year. Using a
mismatched pair in AVERT's Statistical Module (e.g., a historical baseline year of 2009 but a future
year template of 2012) risks accidentally using proxy "new" EGUs that did not exist in 2009, and
thus will not be incorporated into a 2009 analysis.
Why are some generators excluded from AVERT's Future Year Scenario Template?
AVERT considers EGUs that report to EPA's Air Program Markets Dataset (AMPD) only. This may
exclude generators with less than 25 MW of capacity or generators that did not operate in a
particular year.
Why are some generators included in the Future Year Scenario Template but do not
show up in the Regional Data File?
AVERT's Statistical Module allows users to exclude small, low-generation units from consideration
in the displaced emissions analysis. Small peakers have statistics that may be non-representative
of expected generation patterns (i.e., they cannot be readily extrapolated or interpreted outside of
specific events). By default, AVERT excludes units that have generated less than 1,000 MWh per
year. For a 25-MW unit (the smallest reporting unit), this would be the equivalent of 40 hours of
generation over the year, or less than one-half of one percent of all possible operational hours.
In the future scenario demo, does the total for avoided emissions include the impact of
the retirements, or just the EE/RE program impacts adjusted based on retirements?
Results from AVERT runs using the Future Year Scenario Template do not include displaced
emissions from user-specified retired units. These units are assumed to be retired in both the
"before" and "after" cases.
IBM 90 &EPA
State and Local Energy m m
and Environment Program
-------
JtVert
AVishKmI Eir ¦¦¦• I>r -. 1-kI q'-n.-R.itinr T>>i>l^^
.uu..unpjyo#/d»ul ™
AVERT User Manual
Version 2.3
The purpose of the retirements category is to exclude from consideration any units that are likely to
be non-operational in the future year, regardless of the EE/RE program selected.
Does EPA provide projections for use in AVERT's Future Year Scenarios?
At this time, EPA is not providing EPA projections for use in AVERT. Future scenarios are meant to
be developed by users. You can use AVERT's future scenario template to make known changes in
the regional data set. Users who wish to conduct analyses more than 5 years from the baseline
must use AVERT's statistical module and future scenario template. This type of analysis requires
access to future-year hourly, unit-specific generation and emissions data (e.g., from an electric-
sector dispatch model designed to forecast future generation) that can be entered in place of
AVERT's historical data.
Along these lines, EPA has partnered with the Eastern Regional Technical Advisory Committee
(ERTAC), a group of state environmental agency senior staff and multi-jurisdicational organizations
(e.g., LADCO, MARAMA, WESTSTAR, SESARM, NESCAUM), to provide AVERT-compatible
RDFs for ERTAC-specified custom future years.
EPA will issue periodic updates to the historical data files available for download, but will not
release stand-alone future scenarios. At this time, EPA anticipates releasing new Regional Data
Files in the second quarter of each year.
— 91 mFPA
State and Local Energy m m
State and Local Energy
and Environment Program
------- |