Redefining Resource Adequacy
for Modern Power Systems
A Report of the Redefining
Resource Adequacy Task Force
2021
ES
EnErgy SyStEmS
IntEgratIon group
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group ii
ES
ENERGY SYSTEMS
INTEGRATION GROUP
About ESIG
The Energy Systems Integration Group is a nonprofit organization
that marshals the expertise of the electricity industry’s technical
community to support grid transformation and energy systems
integration and operation, particularly with respect to clean
energy. More information is available at https://www.esig.energy.
ESIG’s publications
This report, as well as other reports and briefs by ESIG, is available
online at https://www.esig.energy/reports-briefs.
Get in touch
To learn more about the recommendations in this report,
please send an email to resourceadequacy@esig.energy.
To learn more about the Energy Systems Integration Group,
send an email to info@esig.energy.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group iii
Authors
Prepared by Derek Stenclik, Telos Energy
The core members of the Redefining Resource Adequacy Task Force are:
Aaron Bloom, Energy Systems Integration Group
Wesley Cole, National Renewable Energy Laboratory
Armando Figueroa Acevedo, Black & Veatch
Gord Stephen, National Renewable Energy Laboratory and University of Washington
Aidan Tuohy, Electric Power Research Institute
Acknowledgments
The task force collaborated closely with a project team made up of industry experts.
The task force would like to acknowledge the valuable input and support of the following
individuals regarding the concepts discussed in this report.
Chris Dent, University of Edinburgh
Rob Gramlich, Grid Strategies
Elaine Hart, Moment Energy Insights
Karin Matchett, consulting writer/editor
Michael Milligan, independent consultant
Mark O’Malley, Energy Systems Integration Group
Nick Schlag, Energy and Environmental Economics
Suggested citation
Redefining Resource Adequacy Task Force. 2021.
Redefining Resource Adequacy
for Modern Power Systems.
Reston, VA: Energy Systems Integration Group.
https://www.esig.energy/reports-briefs.
Design: David Gerratt/NonprofitDesign.com
© 2021 Energy Systems Integration Group
Redefining Resource Adequacy
for Modern Power Systems
Report of the Redefining Resource Adequacy Task Force
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group iv
Contents
1 Evolving Reliability Needs for a Decarbonized Grid
1 Elements of Resource Adequacy Under Rising Levels of Renewables
1 Wake-Up Calls from California and Texas
3 Reassessing the Resource Adequacy Methods
4 Traditional Resource Adequacy Analysis Problems and Their Causes
5
Why
Reliability Events Occur Is Changing
7
When
Reliability Events Occur Is Changing
8
How
Reliability Events Occur Is Changing: It’s All About the Weather
9 The Need for a Modified Approach
10 Principle 1: Quantifying size, frequency, duration, and timing of capacity
shortfalls is critical to finding the right resource solutions
13 Principle 2: Chronological operations must be modeled across many
weather years
18 Principle 3: There is no such thing as perfect capacity
21 Principle 4: Load participation fundamentally changes the resource
adequacy construct
22 Principle 5: Neighboring grids and transmission should be modeled
as capacity resources
25 Principle 6: Reliability criteria should be transparent and economic
29 Looking Forward
30 References
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 1
Evolving Reliability Needs
for a Decarbonized Grid
FIGURE 1
The Elements of Grid Reliability
Source: Energy Systems Integration Group.
Resilience
Cyber and other
human-caused
attacks
Storms and
other extreme
weather
Operational
reliability
Lack of
flexibility
Operating
reserve
deficiencies
Distribution
reliability
Natural
event
Equipment
failures
Transmission
stability
Low short-
circuit strength
Frequency
and voltage
deviations
Affordability
Reliability Sustainability
Resource
adequacy
Loss of
interties
Generator
failures
Load
uncertainty
Weather
variability
A
s grids around the world continue to decarbonize
and integrate renewable energy, it is critical that
power system planners, policymakers, and regula-
tors continue to balance three pillars of power system
planning: aordability, sustainability, and reliability
(Figure1). While some stakeholders may have dierent
priorities across the three pillars, each one is critical to
ensuring a smooth clean energy transition.
Ask any grid operator their top priority and the answer
is simple: reliability. Our society has come to expect, and
require, uninterrupted power—even on the hottest days
and coldest nights and through the longest storms. ese
expectations remain as the grid transitions to high
variable renewable energy; reliability is paramount. With
increased variability and uncertainty, how can we ensure
there are enough resources to serve electricity customers
whenever and wherever they need power?
Elements of Resource Adequacy Under
Rising Levels of Renewables
One dimension of grid reliability, that taking the longest
view, is resource adequacy: having enough resources in
the bulk power system available to the system operator
to meet future load, while accounting for uncertainty in
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 2
The increased role of wind, solar, storage,
and load flexibility requires the industry to
rethink reliability planning and resource
adequacy methods.
FIGURE 2
Generation Additions and Retirements from 2014 through 2020,
Plus Planned Retirements
Source: Energy Systems Integration Group; data from APPA (2021).
Change in installed capacity (GW)
Additions Retirements Announced retirements Net change
80
60
40
20
0
–20
–40
–60
–80
–100
–120
Natural gas Coal Nuclear Other fossil Wind Solar
both generation and load. Some uncertainties are becom-
ing more important, such as correlated generator outages
and changes in the weather. By evaluating these uncer-
tainties statistically, grid planners project their resource
needs to reach an acceptably low level of risk of capacity
shortages. Risk metrics can then be used to determine
how much investment our power grids require, how much
new generation should be built, what type of generation
should be built, and which generation can retire.
e power system has always been heavily inuenced by
the weather. Extreme temperatures determine the timing
of peak demand, winter cold snaps can limit natural gas
supply, gas turbine reliability and output are aected by
ambient conditions, and hydro output varies seasonally
and annually. However, as the grid increasingly relies on
variable renewable energy such as wind and solar, the
attention to reliability and weather conditions is
increasingly important.
e industry has more than two decades of experience
integrating variable resources while maintaining—and
even improving—grid reliability. However, a notable
trend is occurring. While early wind and solar capacity
constituted incremental expansions of the grid’s installed
capacity, the industry is now seeing a large swath of fossil
generator retirements, including coal, nuclear, and legacy
gas assets (Figure2). As a result, portfolios of wind,
solar, storage, and load exibility are increasingly used
as replacements to conventional fossil capacity.
ese new resources are being utilized not only for
energy, but also for the grid services required to maintain
grid reliability. e increased role of wind, solar, storage,
and load exibility requires the industry to rethink reli-
ability planning and resource adequacy methods and
to reconsider analytical approaches. Computational
approaches developed in the 20th century are limiting
our collective ability to evaluate reliability and risks
for modern power systems. e conuence of changes
requires new data, methods, and metrics to better
characterize evolving reliability risks.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 3
Wake-Up Calls from California and Texas
Two recent events underscore the importance of
modernizing our thinking on resource adequacy in an
era of changing generation mixes and changing weather.
e rst occurred in California in August 2020, when a
heat storm resulted in two separate days of involuntary
rolling outages across California’s power system. e sec-
ond occurred in Texas in February 2021, when extreme
winter weather resulted in very high electricity demand
while also causing natural gas fuel supply shortages, low
wind output, and widespread equipment failures across
all generation types. Both of these reliability failures
showed how susceptible the grid can be to inadequate
supply as well as the economic, political, and social fall-
out that can occur when grid reliability is jeopardized.
Steve Berberich, the chief executive ocer of the
California Independent System Operator (CAISO) dur-
ing the August 2020 events, summarized the changing
needs of resource adequacy analysis, stating:
ere doesn’t have to be a tradeo between reliability
and decarbonization. What caused the [August black-
outs] was a lack of putting all the pieces together.
You have to rethink these old ways of doing things,
and I think that’s what didn’t happen. . . . e resource
adequacy program in California is now not matched
up with the realities of working through a renewable-
based system, and, in a nutshell, needs to be redesigned
(Hering and Staneld, 2020).
Reassessing the Resource
Adequacy Methods
e objective of this report is to move that redesign
forward by providing an overview of key drivers chang-
ing the way resource adequacy needs to be evaluated,
identifying shortcomings of conventional approaches,
and outlining rst principles that practitioners should
consider as they adapt their approaches.
is report focuses specically on the resource adequacy
analysis and methods that measure system reliability and
risk, and it intentionally stops short of translating that
analysis into procurement decisions. Ultimately, system
planners, regulators, and policymakers need to ensure
there are enough resources to serve load. Resource
adequacy analysis provides the tools to determine
whether there are enough resources and, if not, what
type of resource is needed to meet reliability needs.
While this report is comprehensive in its treatment
of resource adequacy methods, it intentionally does not
address capacity accreditation, which determines how
to assess reliability contributions of specic resources,
or capacity procurement and market mechanisms,
which require further analysis.
Resource adequacy methods have not changed con-
siderably in the past few decades, despite rapid changes
of the resource mix. e central message for practitioners,
regulators, and policymakers is, what got us here won’t
get us there.
Resource adequacy methods have
not changed considerably in the past
few decades, despite rapid changes of
the resource mix. The central message
for practitioners, regulators, and
policymakers is, what got us here
won’t get us there.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 4
Traditional Resource Adequacy
Analysis Problems and Their Causes
FIGURE 3
Two Driving Factors That Require New Approaches to Resource Adequacy
Source: Energy Systems Integration Group.
Chronological grid operations Correlated events
Variable renewable
energy
Energy-limited
resources
Hybrid resources
Load flexibility and
demand response
Weather impacts
Combined outages
Climate trends
Modular technology
reduces correlation
A
t its core, the challenge with resource adequacy
analysis is that, while the methods and metrics
used by the industry today originated in the last
several decades of the 20th century, they have only been
improved incrementally while the resource mix transi-
tioned appreciably. For example, early tools evaluated
only single peak load periods and did not assess risk of
shortfalls across an entire year. ese tools often assumed
static loads and did not consider energy limitations of
most resources. Traditional resource adequacy analysis
also made a simplifying assumption that reliability events
were uncorrelated and that mechanical failures of gener-
ating equipment occurred at random, thus assuming that
the probability of multiple failures occurring simul-
taneously was low.
However, for a grid with high levels of renewables,
energy-limited resources, and load exibility, reliability
is strongly aected by chronological operations and
weather-inuenced correlated events (Figure3). ese
are two driving factors requiring the industry to mod-
ernize frameworks for resource adequacy analysis.
Chronological grid operations: Traditional resource
adequacy analysis often evaluates only individual peak
load hours and does not consider the full year of
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 5
FIGURE 4
Example of Capacity Outage Probability Table
Source: Calabrese (1947).
The table shows the probability of multiple units being on outage simultaneously. Looking down a single column, the probability
of multiple units on outage simultaneously drops precipitously. As the total fleet size increases (moving from left to right along a
row), the probability of a large percentage of the overall fleet on outage (e.g., six units out of 18) is a one-in-a-million event.
operation. is has two problems: it presupposes that
the highest risk period occurs during peak load, and
it fails to account for the sequential operating charac-
teristics of resources. For example, the usefulness of
battery storage as a resource depends on the weather
(and resulting generation) in the preceding days and
expectations for needs in subsequent hours. Likewise,
the use of demand response as a resource depends
on how long the system has already been asking
customers to provide demand response.
Correlated events: While historical resource
adequacy analysis focused on probabilities of discrete
independent mechanical or electrical failures (modeled
with randomly occurring forced outages), weather-
inuenced correlated events should now be recog-
nized as a driving factor of reliability.
Why
Reliability Events Occur Is Changing
Historically, determining whether there were enough
resources available to meet load was a straightforward
analysis, with the foundation rooted in probabilistic
assessment. With the power system made up of many
large, centralized fossil fuel generators for which fuel
availability was rarely a concern, the availability of a
generator was largely based on discrete maintenance
and mechanical failures (forced outages). Each generator
could be characterized with a maintenance outage rate
(%) and a forced outage rate (%), which were used to
determine the likelihood the unit would be unavailable
to serve load. Because these were mechanical failures and
largely uncorrelated (with one another, the weather, or
other factors), probabilistic assessments could quantify
the likelihood that many generators would be on outage
at the same time, thus increasing the risk of a shortfall
and failure to meet load.
An example of this analysis, referred to as the convolution
method, can be found in a capacity outage probability
table from a seminal work on resource adequacy from
the mid-20th century (Figure4). is table shows the
probability, in millionths, that the indicated number
of units would be out simultaneously for eets having a
given number of units when the outage rate is 2 percent.
It shows that as the number of units goes up, the like-
lihood of a large portion of the eet being on outage
decreases quickly—so for an interconnected power
system, the probability of capacity shortfall events
diminishes noticeably as system size increases.
A probabilistic approach may have been appropriate
or the historical power system, where reliability risk
stemmed largely from mechanical failures of large gen-
erating units that could mean many hundreds, or even
thousands, of megawatts (MW) lost due to a single
While historical resource adequacy
analysis focused on probabilities of dis-
crete independent mechanical or electrical
failures, weather-influenced correlated
events should now be recognized as
a driving factor of reliability.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 6
FIGURE 5
Correlated Outages for Natural Gas Generators by Cause During the ERCOT
February 2021 Event
Source: Electric Reliability Council of Texas (2020c).
Gigawatts (GW)
30
25
20
15
10
5
0
■
Weather
related
■
Equipment
issues
■
Fuel
limitations
■
Miscellaneous
■
Existing
outages
Sunday (2/14) Monday (2/15) Tuesday (2/16) Wednesday (2/17) Thursday (2/18) Friday (2/19)
Note: Extreme cold temperatures began on Monday morning.
failure. Coal and nuclear generation were the primary
fuel sources and had weeks’ worth of fuel storage on site,
so fuel availability was not a concern and output was
not variable.
However, while randomly occurring forced outages are
still important to consider, it is increasingly important to
consider correlated generator failures and outages, due
to either the underlying weather or other root causes.
First, a large shift from coal to gas capacity has increased
risks associated with fuel supply. e electric power sector
is now tightly coupled with the natural gas delivery system,
which delivers fuel on demand, with little or no storage
located at the power plant. As a result, correlated outages
due to fuel supply failures are now a key reliability risk,
especially during the winter months when multiple
power plants may experience interrupted fuel supplies
simultaneously. ese same time periods see signicant
increases in load and mechanical failures. is conuence
of factors is leading some system operators, like the New
York Independent System Operator (NYISO), to require
dual-fuel capability for natural gas generators and others,
like the Electric Reliability Council of Texas (ERCOT),
to discuss potential winterization requirements.
Second, the gas turbine technology in wide use today is
more dependent on ambient temperature than are steam
turbine technologies. is is especially true at high
summer temperatures. Both extreme high and extreme
low temperatures derate the maximum output of the
machines, correlating their availability to the underlying
temperature. Mechanical failures are also more likely
during extreme cold events for most technologies and
fuel types in common use today. e correlation in these
types of outages was clearly evident in the February 2021
event in Texas, as shown in Figure5 (ERCOT, 2021c).
ird, for a grid with higher levels of wind, solar, storage,
and load exibility, the actual events that are correlated
have very dierent characteristics. Unlike a fossil fuel–
powered generator, which can lose hundreds or even
thousands of MWs of capacity to a single failure, the loss
of capacity from the disconnection or failure of small,
modular resources is much smaller and more geographi-
cally dispersed. Wind, solar, and storage plants are made
up of many independent inverter-controlled resources.
While any individual wind turbine may fail, the probabil-
ity of an entire plant failing is much lower. is modular-
ity shifts the analysis from discrete generator forced
outages to evaluations of the likelihood of correlated
events and common mode failures.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 7
FIGURE 6
Shifting Periods of Risk in MISO with Increasing Levels of Solar Photovoltaics
Net load diurnal profile (GW)
Source: Midcontinent Independent System Operator (2021).
90
80
70
60
50
40
30
20
10
0
0.06%
0.05%
0.04%
0.03%
0.02%
0.01%
0.00%
Loss of load probability
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
■
Base
■
10%
■
30%
■
50%
■
100%
Hour (Eastern Standard Time)
When
Reliability Events Occur
Is Changing
e changing resource mix is also aecting when reli-
ability events are more likely to occur. In traditional
resource adequacy analysis, periods of higher probability
of a shortfall were almost always associated with peak
loads. Because generator outages were assumed to be
random and variable resources constituted a small part
of the resource mix, generator availability was assumed
to be relatively uniform across the year. As a result, peak
risk occurred during periods of higher demand. Across
most of North America, this usually occurred during hot
summer afternoons or cold winter mornings or evenings.
However, time periods with a risk of shortfall are shift-
ing. e periods of risk we’re used to keeping our eye
on may no longer be the most challenging. In the case
of solar, the diurnal pattern causes a drop in solar pro-
duction at the end of the day correlated among all solar
plants in the area, and extended cloud cover can reduce
output as storms pass through a region. For wind gen-
erators, wind speeds can be correlated as dierent atmo-
spheric conditions or storm fronts pass through a region.
As a result, in a system with high levels of wind and solar
resources, there are both predictable lulls in production
as well as other, weather-inuenced times during which
production across the eet is well below average.
ese dynamics were evident in the involuntary rolling
outages in California in August 2020, which occurred
late in the evening after the sun had set and solar
resources dropped o, several hours after peak load
occurred in the middle of the day. e shifting periods
of shortfall risk are illustrated in Figure6 (MISO, 2021).
As levels of solar generation increase, the periods of
risk shift from 3 p.m. to 6 p.m. due to changes in
resource availability.
Winter cold snaps are also increasingly challenging, as
seen in the Texas event in February 2021, which occurred
in a historically summer-peaking system that has a high
winter reserve margin. While load is higher than normal
during periods of extreme cold, for most summer-peaking
systems these winter loads still tend to be lower than the
annual peak. However, the challenge also manifests itself
on the supply side with increased probabilities of equip-
ment failures, wind turbine icing, and natural gas supply
that is stressed by heating demand. us, in these winter
periods, shortfalls do not have a single root cause, but
are rather a correlation of multiple challenges.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 8
Periods of risk can also be common during o-peak
periods if extended cloud cover across a region reduces
solar availability and weather patterns reduce wind
speeds. ese declines in solar and wind production can
align with extreme cold and heat, and therefore higher
loads. Finally, even periods once characterized as low
risk, like the spring and fall seasons, may have increased
risk. Given that large fossil generators are typically taken
oine for maintenance during these periods, if an outlier
weather event occurs, the probability of a shortfall can
increase signicantly even though loads are considerably
lower than they are in peak periods.
Taken together, the shifting periods of risk mean that
planners can no longer bypass analysis and evaluate only
peak load periods. A broader evaluation across all hours
of the year is necessary to accurately capture shifting
periods of risk of shortfall. Given the energy limitations
of storage and demand response and the operational
characteristics of other resources like start-up times
and ramp rate limitations, the all-hour approach must
be combined with a chronological assessment of grid
operations across an entire year. From a modeling per-
spective, the disciplines of production cost modeling and
resource adequacy modeling are increasingly blurred.
How
Reliability Events Occur Is Changing:
It’s All About the Weather
e power system has always been heavily inuenced
by the weather—extreme temperatures determine the
timing of peak demand, winter cold snaps can limit
natural gas supply, gas turbine reliability and output are
aected by ambient conditions, and hydro output varies
seasonally and annually. However, as already discussed, as
the grid increasingly relies on variable renewable energy,
like wind and solar, the attention to reliability and
weather conditions is increasingly important.
Traditional resource adequacy analysis typically evalu-
ated weather as a driver of system load. Weather changes
could move peak demand periods and created uncertainty
in planners’ load forecasts. ere was some recognition
that weather could lead to correlated outages of the
fossil eet, but rarely was this trend evaluated explicitly.
Instead, the outage rate assigned to generators was based
only on forced outages for unexpected mechanical
failures and planned maintenance.
e increased dependence on weather that accompanies
the shift to more wind and solar on the system causes
multiple issues. e rst and most obvious is that weather
variability aects the availability of these generation
resources. Hour-to-hour changes in weather and elec-
tricity generation mean that a system’s probability of
a capacity shortfall is constantly changing. Given that
serving load in a high-renewables power system also
involves the use of energy-limited resources such as
storage and demand response, a chronological perspective
on system modeling and simulation is required, rather
than the static analysis used in traditional analysis.
In addition, while weather is constantly changing, so is
climate—the weather conditions prevailing in an area in
general or over a long period. If a changing climate leads
to changes in weather, temperature, and extreme events,
it changes the overall resource adequacy risk prole.
Traditional resource adequacy analysis relied solely on
historical weather data; however, the use of historical
data to characterize load and renewable resources may
not be appropriate for gauging future risks aected
by climate change.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 9
The Need for a Modified Approach
T
o overcome the limitations in traditional resource
adequacy analysis, a fresh look is required. While
decades of resource adequacy analysis can be used
as a reference point for reliability planning moving for-
ward, future methods will need to evolve, and a set of
rst principles can be a useful guide. While each region
and system require a unique process, guiding principles
can help ensure a consistent approach in terms of the
objectives, structure, and process of resource adequacy
planning. Consistency in approaches to resource adequacy
can better allow for sharing of insights and best practices,
interregional resource coordination, and a smoother
regulatory process for resource procurement.
e rst principles listed below are based on a few
simple questions: if the approaches to resource adequacy
analysis started from scratch, without a backdrop of
100 years of power system planning and conventional
approaches, how would resource adequacy be evaluated
for modern power systems? How should risk and reli-
ability be evaluated in a power system with large shares
of wind, solar, storage, and load exibility? How can
methods be developed in a technology-neutral manner,
to ensure the methods evolve with a changing resource
mix and new technologies? Responses to these questions
point to six principles of resource adequacy for modern
power systems.
e objective of these principles is to clearly articulate
evolving resource adequacy concepts to system planners,
regulators, and policymakers in order to encourage a
consistent approach to complex challenges. e princi-
ples are not meant to be overly prescriptive; instead, they
are designed to provide a guiding framework that can
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 10
be used by system planners around the world, regardless
of the unique system attributes. ese principles are
designed to help system planners do three things: rst,
to better understand and quantify the reliability shortfalls
that a modern power system is more likely to experience;
second, to identify ways that such shortfalls can be miti-
gated and responded to; and third, to understand what
the resource adequacy analysis means for resource
procurement.
Principles 1 and 2 address the new needs in our under-
standing of capacity shortfalls. Principles 3, 4, and 5 focus
on new understandings of capacity and resource types
in a modern power system. Principle 6 calls for the inclu-
sion of economic considerations in reliability analyses.
PrinciPle 1: Quantifying size, frequency,
duration, and timing of capacity shortfalls
is critical to finding the right resource
solutions.
As the power system’s resource mix changes, resource
adequacy metrics need to be transformed as well. e
conventional resource adequacy metric, loss of load
expectation (LOLE), quanties the expected number
of days when capacity is insucient to meet load. A
common reliability criterion is one day of outage in
10 years, often simplied to 0.1 days per year LOLE.
But LOLE is an opaque metric when used in isolation.
It only provides a measure of the average number of
shortfalls over a study period and does not characterize
the magnitude or duration of specic outage events. It
also does a poor job of dierentiating shortfalls, which,
depending on their length and duration, can have un-
equal impact on consumers and can require dierent
mitigation options. For example, since LOLE only
quanties frequency, a shortfall of 1 percent of load for
10 hours is measured the same way as a shortfall of 10
percent of load for 10 hours. In addition, there is very
little consistency in this metric’s application, as dierent
planners in dierent regions interpret the criterion
dierently, and each region has dierent institutional
and regulatory requirements that determine what
probability of unserved energy is acceptable.
Similar metrics also provide information on the
probability of a shortfall event but limited information
regarding what the shortfalls look like. Loss of load
hours (LOLH) counts the average expected number of
hours of shortfall, loss of load events (LOLEv) is similar
to LOLE but allows for multiple “events” to occur in a
single day or a single event to span multiple days, and
expected unserved energy (EUE) calculates the average
amount of energy unserved.
Looking Beyond LOLE
e reliance on the LOLE metric was adequate in
traditional resource adequacy analysis because shortfalls
tended to share similar characteristics, largely occurring
during peak load events and caused by randomly occur-
ring forced outages of the conventional fossil eet. In
addition, the resource solutions implemented when the
LOLE measure was exceeded were one size ts all. e
combustion turbine was the de facto resource used to
meet reliability needs, as it was the lowest capital cost
way to get more “steel in the ground,” and operating
costs (based on fuel eciency) were not a concern
because the units were rarely utilized. However, the
resource options available to system planners today
are numerous. Energy storage, demand response, and
load exibility provide competitive alternatives to
the combustion turbine approach for many types
of shortfall events.
Energy storage, demand response, and load
flexibility provide competitive alternatives
to the combustion turbine approach for
many types of shortfall events.
In addition, the reliability events are now more varied;
therefore, understanding the size, frequency, duration,
and timing of potential shortfalls is essential to nding
the right resource solutions. LOLE is an inadequate
metric in a world of more varied shortfall events because
it provides limited information on shortfall events’ size
and duration. is makes it dicult to know the true
impact of potential shortfalls and nearly impossible to
determine the types of resources necessary to reduce
the number of shortfalls.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 11
FIGURE 7
Building Blocks of Resource Adequacy Metrics
Source: Energy Systems Integration Group.
Example 1— Same LOLEv and LOLH, but very different events Example 2— Same LOLH and EUE, but very different events
MW
A
hrs
LOLEv = 1
LOLH = 4
EUE = 12
Max MW = 5 MW
Max MWh = 12 MWh
Duration = 4 hr
MW
B
hrs
LOLEv = 1
LOLH = 4
EUE = 4
Max MW = 1 MW
Max MWh = 4 MWh
Duration = 4 hr
MW
C
hrs
LOLEv = 3
LOLH = 3
EUE = 6
Max MW = 4 MW
Max MWh = 4 MWh
Duration = 1 hr
MW
D
hrs
LOLEv = 1
LOLH = 3
EUE = 6
Max MW = 2 MW
Max MWh = 6 MWh
Duration = 3 hr
Each block represents a one-hour duration of capacity shortfall, and the height of the stacks of blocks depicts the MW of unserved
energy for each hour. A: a single, continuous four-hour shortfall with 12 MWh of unserved energy; B: a single, continuous four-hour
shortfall with 4 MWh of unserved energy; C: three discrete one-hour shortfall events with 6 MWh of unserved energy; D: a single,
continuous three-hour shortfall with 6 MWh of unserved energy.
Differentiating Capacity Shortfalls
Systems with the same LOLE and LOLH can have very
dierent risk proles, types of shortfalls, and mitigation
options. Figure7 illustrates four dierent capacity short-
fall events. On the x-axis of each chart is time and on
the y-axis is the MW of a shortfall event. Each block
represents a one-hour duration of capacity shortfall, and
the height of the stacks of blocks measures the amount
of unserved energy. ese building blocks show how
dierent shortfall events can be and thus how easily
traditional metrics can fail to capture them.
e two charts on the left (Figure 7A and B) show
how simple expected value metrics can fail to distinguish
between very disparate events. ese charts show a single
continuous capacity shortfall event of equal duration
(four hours). Both of these events would count toward
the aggregate loss of load events (LOLEv) metric as one
event, since they occur within the same day, and both
would count toward LOLH with four hours. From an
LOLEv and LOLH perspective, then, the events are
indistinguishable. However, the top event is three times
larger in terms of unserved energy and ve times larger
in terms of the maximum unserved energy at a single
point in time. e events have very dierent impacts on
customers and may require dierent mitigation strategies
on the part of system operators.
While EUE is better at dierentiating individual events,
this metric too can have challenges. e charts on the
right (Figure 7C and D) show consistent unserved energy
and loss of load hours, but the top plot shows three
distinct events (LOLEv of 3), whereas the bottom plot
shows a single event. In this case, the corresponding
EUE and LOLH metrics are identical, but the LOLEv
metrics are three times larger in the top example. Separate
events could be mitigated by energy storage that can
re-charge between events, but may be further challenged
by demand response programs that may be limited by
the number of allowable calls.
Without the use of multiple metrics, as well as additional
information on the size (both in MW and megawatt-
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 12
quanties the expected aggregate size (amount of energy)
and duration of shortfall events as opposed to only quan-
tifying the probability or frequency of one occurring.
However, EUE still provides only a single average metric
that cannot distinguish between the individual events.
In addition, resource adequacy analysis should pay atten-
tion not just to the expected values, but to potential tail
events. While high-impact, low-probability events are
very rare—and system planners intentionally do not plan
to mitigate all potential risk—these events’ impact on a
high-renewables grid is important to assess given their
potentially devastating impact on customers.
Unfortunately, traditional resource adequacy metrics’
simplication of hundreds or thousands of power system
simulations into a single average oers little insight into
the distribution of potential resource adequacy shortfalls
that the system could experience. A system that has rare
but very large events could appear to have the same level
of reliability as a system with more frequent, smaller
events, causing current metrics to fail to account for
the much greater impact on consumers—and society in
general—of the large events. Future resource adequacy
analysis should move beyond expected values and provide
information on the distribution of individual events.
e chart in Figure8 (p. 13) quanties the number of
shortfall events (each represented as a dot) for a single
system simulated across three dierent resource mixes.
hours (MWh)), frequency, and duration of individual
events, determining appropriate mitigation actions is
dicult. For example, the event on the top left would
require at least three times more energy storage and
demand response than the event on the bottom left.
For the events on the right, a battery resource of 4 MWh
could avoid all of the unserved energy in the top right
event (provided it could recharge between events), but
would be insucient to avoid the bottom right event
(where we would need 6 MWh of storage). is infor-
mation would be impossible to ascertain by LOLE,
LOLH, and EUE metrics alone. Resource adequacy
metrics that can quantify size, frequency, duration,
and timing of shortfall events are critical to nding
the right resource solutions.
Achieving Deeper Insights into
Resource Adequacy Metrics
One of the biggest limitations of LOLE, LOLH,
and EUE metrics is that they provide only an average
measure of system risk across many hundreds or thousands
of samples. ey do not provide information on the
full distribution of shortfalls. New methods in resource
adequacy analysis should expand to provide additional
insights into not only the average (expected value) re-
source adequacy events, but also the characteristics of
the individual events themselves. System planners require
this type and quantity of information to ensure that they
can right-size mitigations to meet the system’s specic
reliability needs.
New methods in resource adequacy
analysis should expand to provide addi-
tional insights into not only the average
(expected value) resource adequacy
events, but also the characteristics
of the individual events themselves.
It is too early to tell whether entirely new metrics need
to be developed, but what is certain is that planners need
to extract more information and details from existing
ones. Increased use of EUE is a good rst step, as it
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 13
FIGURE 8
Scatter Plot of Size, Frequency, and Duration of Shortfall Events
with Energy-limited Reliance on Energy Limited Resources
Size (MW)
Source: Energy Systems Integration Group.
900
800
700
600
500
400
300
200
100
0
Frequency
Scenario 1 Scenario 2 Scenario 3
Duration (hours) Duration (hours) Duration (hours)
7000
6000
5000
4000
3000
2000
1000
0
Magnitude
(MWh)
Each resource mix has very dierent underlying
resources but the same LOLE of 1-day-in-10-years (i.e.,
the same number of dots). However, despite having the
same LOLE, the systems have very dierent risk proles.
An improperly planned high-renewables grid may
experience much larger shortfall events than those we
are used to planning for due to sustained periods of low
renewable production. is could cause longer and larger
disruptions—even if the probability of these events
occurring is lower than historical norms. Improved use
of resource adequacy metrics can avoid this challenge.
Improved utilization of existing metrics and visualiza-
tions must move beyond average values. ey must pro-
vide information on the distribution of events as well as
provide emphasis on individual, rather than aggregate,
event characteristics. Relying on multiple metrics and
visualizations of the size, frequency, duration, and timing
of shortfall events will allow planners to select mitigations
and resources that are appropriately sized to t system
needs and avoid over-procurement of resources.
PrinciPle 2: Chronological operations
must be modeled across many weather
years.
Historically, traditional resource adequacy analysis
evaluated only periods of peak demand for reliability risk.
is was in part due to the more limited computational
capabilities of the time as well as to a resource mix that
did not uctuate much seasonally or hourly, making the
uctuations of load the main variable. In addition, there
was limited energy storage on the system with which to
smooth out demand. e small amount that was installed
was pumped hydro, often with 12 or more hours of
energy storage, and energy limitations were less of a
concern. As a result, systems included few short-duration
and energy-limited resources that would not be able to
provide extended support during reliability events. ere-
fore, if generation on the system was adequate during
the period of highest load, it would be adequate during
the rest of the year as well.
2 4 6 8 2 4 6 8 2 4 6 8
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 14
Today, the increased reliance on variable renewable
energy and energy-limited resources is changing the
resource adequacy construct. Periods of risk are no longer
conned to peak load conditions, but are shifting to
other time periods due to abnormal weather events, the
daily setting of the sun, and the fossil eet undergoing
increased maintenance during fall and spring. e in-
creased levels of variable renewable energy mean that
resource analysis requires specic attention to hourly,
seasonal, and inter-annual resource variability. e se-
quence of the variability is key, given that energy-limited
resources such as batteries or demand response require
either a preceding period or subsequent period of high
production to be useful for grid reliability. is will
require increased reliance on weather and power fore-
casting and integrated storage scheduling that considers
forecast uncertainty to ensure that storage can be
available when needed.
As a result, the conventional approach of designing a
system solely to meet peak load conditions—via a static
planning reserve margin—is no longer appropriate.
A simple planning reserve margin that is used to procure
a certain amount of capacity above and beyond peak load
does not ensure that the system will be reliable during
other times of the year given changes in the resource mix.
Importance of Chronological Evaluation
of All Hours
e California rolling blackouts in 2020 are a good
example. California’s resource adequacy construct and
planning reserve margin are based on the peak gross
load, which occurs in the middle of the day during sum-
mer months. However, periods of peak risk in California
now occur in the evening hours as solar resources decline
and loads remain relatively high. is is clearly illustrated
in Figure9, which shows the gross and net load for
CAISO for the August days when rolling blackouts
occurred (CAISO, 2021). e conventional assumption
that peak risk is aligned with peak load is no longer
true, requiring a chronological evaluation of all hours
of the year so that the times of risk of shortfall can be
accurately identied.
FIGURE 9
Gross and Net Load During the 2020 California Reliability Event
Gigawatts (GW)
Source: Energy Systems Integration Group; data from California Independent System Operator (2021).
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
August 14, 2020
50
45
40
35
30
25
20
15
10
5
0
August 15, 2020
■
Stage 3 emergency Demand Net demand
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 15
FIGURE 10
Example of Chronological Resource Adequacy Simulations with a Shortfall Event
Source: Hawai’i Natural Energy Institute (2020).
■
Fossil
■
Solar
■
Wind
■
Storage
■
Unserved energy
One Week of Grid Operations
Generation (MW)
is chronological assessment is required to ensure that
the energy storage and demand response will be available
for enough hours to get the system through periods
of scarce supply. Energy-limited resources may reduce
reliability risk in some periods (when the storage is dis-
charging or when load is reduced), but only if they in-
crease risk in other periods (when the storage is charging
or when load is shifted to earlier or later times). Hour-
to-hour operations and scheduling ensure that energy
storage and demand response will be around long enough,
and can fully recharge, to support the system through
reliability challenges. Chronological assessment is essen-
tial to highlight resource adequacy needs and necessary
procurement of long-duration storage resources.
Modeling sequential grid operations is critical to cap-
ture the whole picture: the variability of wind and solar
resources along with the energy limitations of storage
and load exibility. Chronological stochastic analysis is
thus increasingly important, simulating a full hour-to-
hour dispatch of the system’s resources for an entire year
of operation across many dierent weather patterns, load
proles, and random outage draws. An example is shown
in Figure10, which illustrates a week of chronological
commitment and dispatch of a power system, and a
shortfall that occurs when there is insucient storage
available to extend through the late evening hours.
Despite load being signicantly lower in the late evening
hours, the probability of a shortfall is higher (HNEI,
2020).
Need for Many Years of Weather Data
In addition to modeling chronological grid operations,
resource adequacy analysis for modern power systems
requires the incorporation of many years of weather data.
Many years of synchronized hourly weather and load
data are necessary to understand correlations and inter-
annual variability between wind and solar generation,
outages, and load. e same weather conditions can
aect wind and solar output, whose probabilities are
driven by irregular and complex weather patterns, and
load and thermal unit derates—requiring that the
Chronological operations and scheduling
ensure that energy storage and demand
response will be around long enough, and
can fully recharge, to support the system
through reliability challenges.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 16
FIGURE 11
ENSTO-E Example of Monte Carlo Simulation Principles
N
Climate
Ye a rs
N x M
Monte-Carlo
Simulations
Source: European Network of Transmission System Operators for Electricity (2020a).
weather data be consistent across these inputs. e
California event in August 2020 stemmed, at least in
part, from a widespread heat wave that seemed highly
improbable based on historical patterns but may be more
likely now and into the future due to climate change.
More changes to resource adequacy analysis and model-
ing are needed to address both potential conditions
and resource availability during these conditions.
An example of this process is shown in Figure11, which
depicts the European Network of Transmission System
Operators for Electricity’s (ENTSO-E) regional grid
planning methodology. In this approach, random unit
outages are sampled across many years of synchronized
weather data and across many years of annual variations
in wind, hydro, solar, and load (ENTSO-E, 2020a). Each
weather year is simulated against the same number of
stochastic generator outage proles to create a matrix
of weather years and outage draws. e total number of
samples is the product of the two, and average resource
adequacy statistics are calculated across them. is pro-
cess allows system planners to identify whether certain
weather conditions lead to increased probabilities of
shortfall events.
e methodology also helps ensure resource adequacy
across an entire range of potential operations, as opposed
to just the peak load periods or average weather conditions.
Using stochastic production cost methods—combining
both chronology and varying weather across a full 8,760-
hour analysis—is necessary to help identify times and
situations of peak risk. Given that low-probability events
drive resource adequacy challenges, a long historical
record of weather data is necessary to identify the prob-
ability of potential extremes. With higher renewable
energy and storage capacity on the grid, these periods are
likely to be made up of more combinations, across more
variables, than planners were accustomed to in the past.
Data Limitations in Weather Modeling
Analysts and policymakers should be cognizant, however,
of data limitations. is methodology is data-intensive
and requires a convergence of power systems and meteo-
rological expertise. System planners often have access to
long historical records of solar and hydro resources, but
may be limited on wind data. In addition, historical data
may be available for system load, but underlying changes
to consumer behavior, load growth, and distributed energy
resources may limit the usefulness of legacy load data
from several years in the past. Where long historical
records of correlated wind, solar, hydro, and load are not
available, planners will need to either use a limited data
sample or develop methods that can bootstrap a larger
dataset based on correlation of a smaller, but complete,
dataset to a longer dataset such as temperature.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 17
In addition, past observations may no longer be good
predictors of future conditions with a changing climate.
Part of reliability planning is ensuring that the system
can maintain reliability during potential—and credible—
weather events. e California heat wave saw some of
the highest average temperatures in the past 35 years,
spread across most of the western United States. Simi-
larly, during the 2021 winter events, Texas saw tempera-
tures well below the near-term historical record, sus-
tained for many days longer than a similarly cold event
in 2011. Just because our recent weather data do not
include a weather event doesn’t mean system planners
do not have to prepare for one in the future.
To assume that historical trends continue into the future
can also be problematic due to climate change, for two
reasons. A changing climate will likely cause weather
conditions to diverge from their historical norms and
may shift load and renewable generation away from
expectations. And climate change may increase the
frequency of extreme weather events, which can increase
the probability of resource adequacy shortfalls. e Euro-
pean members of ENTSO-E, for example, have identied
climate change as a key contributor to resource adequacy
risk and are planning to incorporate a climate change
trend as a baseline assumption in their resource
adequacy process:
e impact of climate change on adequacy assess-
ments can be signicant, considering that an impor-
tant element of the adequacy models is the underlying
climate-dependent data used as input. ENTSO-E
is working with climate and data experts to prepare a
database that will reliably reect the impact of climate
change on climate variables and, thus, on adequacy
simulation results. . . . Our eorts will continue during
the upcoming three years, targeting to reliably incor-
porate in our models the impact of climate change
by the end of 2023 (ENTSO-E, 2020b).
Given the uncertainty in the weather, limited data across
a long historical record, and potential climatic changes,
system planners should identify and evaluate potential
drivers of resource adequacy risk, even if they have not
occurred or stressed the system in the past. While it
will be impossible to assign probabilities to these events,
and thus use them to quantify conventional resource
adequacy metrics, these drivers can be used to under-
stand potential periods of risk for further investigation
and contingency planning.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 18
More planning should be focused on identifying poten-
tial situations where the traditional data-driven statistical
modeling has limitations, and on testing system reliability.
Future resource adequacy analysis should evaluate poten-
tial situations that may not have occurred in the past but
could reasonably occur in the future. Identication of
these high-impact, low-probability events can then be
evaluated in isolation to determine whether and how
they should be mitigated.
PrinciPle 3: There is no such thing
as perfect capacity.
As Principle 1 suggests, some capacity shortfalls may
consist of frequent but short-duration events, while
others may be infrequent but long-duration events.
Mitigation strategies will need to be specied accord-
ingly, because dierent resources bring dierent capa-
bilities. Battery energy storage may be well suited to
solve frequent, short-duration shortages, while demand
response may be better suited for infrequent, but challeng-
ing, events. Additional resources like long-duration stor-
age, hydro, and thermal generation may be required for
long-duration capacity shortages spanning days or weeks.
However, gas plants are not always available on demand,
as they experience planned as well as weather-related
outages. e false dichotomy between the perfect resource
and resources with only partial “rm capacity” is due
to be replaced by analysis applying the eective load-
carrying capability (ELCC) metric to all resource types.
ELCC measures the amount of load that can be added
to a system given the addition of a resource, while main-
taining the same level of reliability as the system prior
to the resource addition.
Weather-Dependence of Thermal Generators
e bias toward centering resource adequacy around
“rm capacity” and treating a gas turbine as a perfect
capacity resource (having an ELCC of 100 percent)
causes several problems. First, it assumes that combus-
tion turbines and similar fossil technology are available
on demand, and rarely assigns an ELCC to these tech-
nologies in a similar manner as wind, solar, storage,
and demand response technologies. In some cases, the
fossil technology is discounted, but only based on the
equivalent forced outage rate on demand (EFORd). For
example, a gas turbine with a 5 percent forced outage
rate would receive 95 percent capacity credit toward
the planning reserve margin.
However, as discussed above, there are times when cor-
related outages occur on the gas eet, which increases
reliability risk substantially. All generation sources are
weather-dependent to some degree. e light blue
segments of the bar chart in Figure12 (p. 19) provide
the average forced outage rate of resources throughout
the year, whereas the dark blue bar segments show the
increase in forced outage rates during extreme cold con-
ditions. ermal generators, including nuclear, require a
water supply which can be threatened by extended drought
conditions, and extreme temperatures can force reduced
operations. Gas turbines have ambient derates due to
high temperatures, forced outage rates that are consider-
ably higher during extreme cold conditions, and a fuel
supply that can be jeopardized by competition with gas
heating demand. Coal piles can freeze solid. Availability
considerations due to weather, supply, and intra-resource
correlations should be applied to all resource types. If
ELCC is used for capacity accreditation, the methodology
should be applied to all resource types, not just variable
renewable energy and energy-limited resources.
Different resources bring different capa-
bilities. Battery energy storage may be well
suited to solve frequent, short-duration
shortages, while demand response may
be better suited for less frequent events.
Unfortunately, traditional resource adequacy analysis is
designed around a one-size-ts-all approach to resource
adequacy additions. Conventional system planning has
often treated a natural gas combustion turbine as peaking
“rm capacity” and, therefore, a near-perfect capacity
resource that could be added to improve reliability. If a
system was determined to be short of capacity, combus-
tion turbines were often used as the default resource to
bring the system to the reliability criteria. is is because
these represented a low-installed-cost resource and could
eectively put more “steel in the ground” for reliability.
Under this construct, resources like wind, solar, and
storage are given partial “rm capacity” credit.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 19
Weather-driven
outages,
3–4% seasonal
average
Expected
winter outage
rate; typically
used in resource
adequacy
analysis
FIGURE 12
Total Unplanned Outages During
Recent Cold Weather Events
Percentage of capacity out of service
Source: Energy Systems Integration Group.
PJ M
Jan. 2014
50
45
40
35
30
25
20
15
10
5
0
Includes forced outages plus
derates for all technology types
MISO
Jan. 2019
ERCOT
Feb. 2021
technologies can be used to ensure that enough resources
are available when needed, despite the inherent uncer-
tainty of individual resources.
Recognition of Resources’ Limitations
and Strengths
ELCC is a useful metric to evaluate reliability contribu-
tions and correlated output within and between resource
types. Combinations of resources and the interactions
between them are important to understand, though
currently dicult to quantify. Numerous studies suggest
that the ELCC of a resource type is highly dependent on
the underlying resource mix and the load prole—both
of which change continuously. Figure13 shows how the
ELCC of solar and storage are both higher when evalu-
ated in combination than when evaluated separately.
ese interactive eects may be either antagonistic,
where each increment of solar provides successively lower
capacity value, or synergistic, where a portfolio of solar
and storage likely provides more value than the sum
of its parts (Schlag et al., 2020). In addition, the way
in which a system is operated can have an impact on
ELCC, especially for energy-limited resources. It is
possible, for example, to operate a storage system to
maximize resource adequacy, which could dier at times
from operation that maximizes revenues. e ability to
accurately forecast system conditions can also change
ELCC accreditation.
FIGURE 13
The ELCC of Solar Alone, Storage Alone, and the Two Resources in Combination
Source: Energy and Environmental Economics (E3) / Schlag et al. (2020).
Load (GW)
45
40
35
30
25
20
15
10
5
0
0 5 10 15 20 0 5 10 15 20 0 5 10 15 20
Hour of day
Hour of day
Hour of day
Solar only 4-hour storage only Solar + 4-hr
storage portfolio
Second, the conventional “perfect capacity” approach
assumes that a resource is needed during all hours and
must be dispatchable at any moment to be eective for
reliability. In reality, a balanced portfolio of resources,
including wind, solar, storage, load exibility, and fossil
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 20
FIGURE 14
Resource Adequacy Capacity Credit vs. Actual Generation During the Texas 2021 Event
Source: Eamonn Lannoye, Electric Power Research Institute; data from Electric Reliability Council of Texas (2021a, 2021b).
ERCOT electricity generation versus seasonal expected availability (February 15-18, 2021).
* ERCOT’s Seasonal Assessment of Resource Adequacy (SARA) attributes an expected available capacity to each generation type, considering seasonal factors.
†
During peak demand (18-23h); over all hours: 259%
Natural
Gas
Coal
Nuclear
Wind
Solar
Future resource adequacy analysis should explicitly
recognize that all resources have limitations based on
weather-dependence, potential for outages, exibility
constraints, and common points of failure. is was
abundantly evident in the extreme cold events in Texas
in February 2021. In this case, all of the resources on the
system were similarly strained, and no resource was able
to contribute as planned for reliability, demonstrating
that there is no such thing as perfect capacity. As shown
by the Electric Power Research Institute’s analysis in
Figure14, natural gas, coal, nuclear, and wind resources
all provided energy well below their expected levels
when compared against the ERCOT System Assessment
of Resource Adequacy report.
Resource adequacy analysis for modern power systems
should also recognize that each resource brings dierent
capabilities that may work best in specic situations.
For example, frequent but short-duration events could
be best addressed by battery storage or load exibility
(via time-of-use rates), whereas infrequent, short events
could be best addressed by load-shed demand response
programs. Frequent and long events may require long-
duration resources like fossil generation, long-duration
storage, or hydro resources, and infrequent, long events
may be best handled through coordination with
neighboring grids or emergency procedures.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 21
PrinciPle 4: Load participation
fundamentally changes the resource
adequacy construct.
In traditional resource adequacy analysis, load was treat-
ed as a static—that is, uncontrollable—input into the
modeling and simulations. Load was increasing year
over year, and the purpose of resource adequacy analysis
was to determine whether there was enough generation
capacity to serve a xed load. e question for planners
was simply, do we have enough generation to meet
our load requirement? While there was uncertainty
in the load forecast, a higher load forecast just meant
additional supply-side resources were needed.
But the historical notion that a specic amount of
generation capacity is required to meet a static load is
no longer relevant. Load exibility is increasing quickly.
e decreasing costs of distributed sensors and controls,
increased proliferation of distributed energy resource
aggregation, and increased visibility into behind-the-
meter load consumption have all made loads more
exible, price responsive, and intelligent.
Using Load Flexibility for Resource Adequacy
e proliferation of energy storage, demand response,
electric vehicles, and dynamic rate design bring with
them new options for load exibility and should be
evaluated in a similar context as generation resources,
including uncertainty and availability. However, future
load exibility is based on customers’ economic decisions.
Real-time markets, with a high degree of participation
from price-responsive demand, may place more attention
on economic considerations rather than reliability needs,
as customers can determine and dierentiate which
loads matter most. It is therefore important to ensure
that reliability benets of exible loads are not lost, and
that various forms of load exibility—and their asso-
ciated reliability benets—are included in resource
planning assessments.
Two examples in Hawaii include procurement of
capacity grid services from virtual power plants and
price-responsive loads. A recent utility procurement
is leveraging a virtual power plant of 6,000 residential
photovoltaic+battery systems (25 MW, 80 MWh) to
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 22
provide both “load build” and “load reduce” grid ser-
vices; this allows the utility to call on behind-the-meter
batteries to charge or discharge when grid conditions
require. e utility also recently announced another
60 MW procurement target for virtual power plants
and price-responsive loads to provide resource adequacy
benets for an upcoming coal retirement (Pickerel,
2021).
System Planners’ Data Needs
for Load Resources
As load becomes more exible, the options to balance
both the supply and demand sides of the resource ad-
equacy equation become much more dynamic. However,
the industry does not have the same institutional knowl-
edge base and experience with demand-side resources as
it does with supply-side generation. If system planners
and operators are going to rely on load exibility to the
same extent as supply-side resources, more information
is needed on potential unavailability of load exibility,
uncertainty in participation, and scheduling constraints
that may aect load resources’ utilization.
In other words, the modeling of demand-side resources
will require the same level of inputs used to model
generation. is includes an equivalent to planned and
forced outage rates, seasonal derates, energy and duration
limitations, constraints on the number of starts per year/
month/day, variable operating costs, and other character-
istics typically used to simulate a generation resource.
Fortunately, as demand-side load exibility continues
to proliferate, experience and data are also growing. ese
resources constitute a exible, modular, and dynamic
resource for solving resource adequacy challenges with-
out installing more generation that would be used
sparingly, if ever, for reliability needs.
PrinciPle 5: Neighboring grids
and transmission should be modeled
as capacity resources.
Resource adequacy modeling can be complex and is
often computationally challenging; a large power system
must typically be simulated across hundreds or thousands
of Monte Carlo samples. is challenge is further ampli-
ed by the increasing need to model full chronology
across an entire year of operations (Principle 2). To make
this problem tractable, simplications are required. Often
that means only limited representation of neighboring
power systems and the transmission network in general.
However, resource sharing can be a signicant, low-cost
alternative to procuring new resources. Imports from
neighboring regions are likely to become more valuable
for resource adequacy due to the increased diversity of
Resource adequacy and power system
planning should consider load flexibility as
a supply-side resource capable of reducing
system risk of shortfalls.
Another potential mechanism for increased load par-
ticipation for resource adequacy benets is through
energy-only markets with price scarcity driving customer
behavior. For a full energy-only market to work, value
of lost load–based scarcity pricing is needed, along with
a market structure that ensures that market participants
have both the incentive and ability to procure power in
advance or can fully handle any risk of paying scarcity-
based prices if they wind up with a short position.
Scarcity pricing, even without an underlying resource
adequacy construct, creates two incentives for resource
adequacy. First, it provides a clear incentive to reduce
loads or switch to behind-the-meter generation during
scarcity events. Second, it provides a clear incentive for
load-serving entities to enter into bilateral contracts
for capacity as a hedge against price volatility.
Additional mechanisms exist to increase load exibility.
is can be done dynamically, via real-time pricing
and direct distributed energy resource aggregation and
control, or more passively via time-of-use rates, critical
peak pricing, energy eciency programs, and education.
Regardless of the method, resource adequacy and power
system planning should consider load exibility as a
supply-side resource capable of reducing system risk
of shortfalls.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 23
chronological wind, solar, and load patterns over a much
larger area. A typical wind plant output tends to have
little correlation with other wind plants a few hundred
miles away. Solar output varies with cloud cover and time
zones. Load diversity is greater across large areas. While
extreme weather can happen anywhere, it does not
happen everywhere at once.
Neighboring systems are often simplied in resource
adequacy analysis because the system planners’ and regu-
lators’ perspective has traditionally been that the power
system should be self-reliant and able to serve load
without requiring imports from neighbors during critical
time periods. is mindset is not unreasonable; ultimate-
ly the utility or system operator is responsible for reliably
meeting its customers needs, regardless of what happens
in neighboring regions.
However, the preference for self-reliance leads to a
potentially large and expensive overbuild of capacity. If
every region is carrying its own margin, which is only
used sparingly for reliability, the cost of surplus resources
is amplied across the interconnected power system. e
value of sharing between adjacent regions was a major
driver for the formation of independent system operators
and regional transmission organizations, which allowed
many smaller, vertically integrated utilities to pool their
resources and reduce coincident peak load (achievable
when the peak regional load is lower than the sum of
individual localities’ peaks due to load diversity).
Ultimately, it is up to regulators, policymakers, and
system planners to determine the level of reliance on
neighbors that is acceptable, given the local conditions
and resource mix. ere is no right or wrong answer.
An Economic Opportunity Too Large to Ignore
ere is a very large economic opportunity in increasing
regional coordination, sharing of resources, and relying
on imports to meet reliability needs. Major benets
include:
• Staggeredpeaks.Load diversity increases with
large geographies, varied weather patterns, multiple
time zones, and demographic dierences. e larger
the system, the less likely peak loads are to occur
simultaneously.
Key to unlocking this economic oppor-
tunity is transmission, to enable flows
between regions and create interregional
resource diversity.
• Moreconsistentrenewablegeneration.As the
planning footprint increases in size, the wind and
solar variability diminishes. While the skies can
be cloudy and winds can be calm anywhere, it will
not likely be cloudy and calm everywhere.
• Lesschanceofsimultaneousoutages. A larger
portfolio of resources means lower probability of
simultaneous outages across a large portion of the
resource mix, as each region has access to a larger
number of generating units and higher installed
capacity.
• Lesschanceofoutagescausedbyfuelshortages.
ere is a lower probability of outages due to fuel
shortages because a larger region likely has multiple
fuel delivery paths, such as natural gas pipelines.
Key to unlocking this economic opportunity is trans-
mission, to enable ows between regions and create
interregional resource diversity. Transmission assets
should therefore be clearly identied as having
resource adequacy benets.
is principle is clearly illustrated by market data from
the February 2021 extreme cold weather event when
ERCOT and MISO experienced capacity shortfalls due
to cold temperatures (Figure15, p. 24). As the middle
of the country struggled to meet load, much of the east
coast was experiencing normal temperatures and had
surplus capacity, indicated by signicantly lower energy
prices and a clear gradient across the MISO-PJM seam
( JCM, 2021). Additional transmission capacity between
regions could have mitigated some of the resource
adequacy failure.
e same benets can be had locally, available to zones
within a single balancing authority. For example, NYISO
and PJM both have nested capacity zones. In the New
York example, the lower Hudson Valley (Zones G-J),
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 24
FIGURE 15
Real-Time Energy Prices During a MISO Resource Adequacy Shortfall Event
Source: Joint and Common Market (2021).
New York City (Zone J), and Long Island (Zone K)
have local capacity requirements because there is limited
transmission capability into the zones, which also have
the highest loads. As a result, from a resource adequacy
perspective these zones have higher probability of a
shortfall and thus require additional local capacity. If
increased transmission capability was constructed, these
zones could share resources with neighboring zones, thus
decreasing the resource adequacy risk and lowering the
amount of local generating capacity needed for reliability.
Modeling and Policy Needs for
Transmission Coordination
While the pooling of resources improves reliability,
it raises questions about how to appropriately share
resources during times of resource adequacy risk. is
introduces a policy and regulatory challenge about how
to balance reliance on neighbors and self-suciency.
Politicians and system operators are beholden to their
own constituents and customers for reliability—not
those in other regions. However, constituents and
customers also rely on them for aordable reliability.
e desire for self-reliance must be informed by the
aordability oered by the option of using neighbors’
resources rather than investing in redundant resources.
Integrating neighboring areas into resource adequacy
analyses requires some advances in the modeling and
policy/regulatory arenas. First, conventional resource
adequacy analysis tends to do a poor job of modeling
neighboring systems due to the preference of self-
reliance and to computational limitations. So, some
modeling simplications may be necessary to make
the problem size tractable, and this will need to be done
with care and deliberation. But ultimately, more precise
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 25
modeling of neighboring systems can lend condence
that, statistically speaking, some resources in neighboring
systems will likely be available during times of highest
reliability risk.
To address these issues, resource adequacy analysis for
modern power systems should include two things. First,
transmission assets should be evaluated as a capacity
resource if they allow additional ow to enter into a
capacity- and transmission-constrained region. is pro-
vides an alternative resource, beyond just local generation
and load exibility, to meet resource adequacy require-
ments. Second, resource adequacy analysis should pro-
vide a detailed representation of the neighboring systems
so that the same probabilistic assessments can be made
in neighboring regions in order to provide more delity
in the availability of imports. is will be computation-
ally challenging but necessary, given the degree to which
the transition to higher levels of variable renewable
energy is enabled by taking advantage of geographic
diversity.
is type of transmission coordination is not just a
matter of increased interties between regions, but just as
importantly requires market mechanisms that allow for
transparent sharing of resources across balancing areas,
regional transmission organizations, and other jurisdic-
tions. ere are a wide variety of market mechanisms
available, ranging from voluntary capacity agreements
and bilateral contracts to formal capacity markets.
Regardless of the form, establishing clear market rules
for capacity sharing is a critical regulatory and policy
need for the coming years.
PrinciPle 6: Reliability criteria should
be transparent and economic.
Conventional resource adequacy analysis largely ignores
economic principles and excludes the nancial impacts
on consumers of a system’s reliability choices. In this
context, reliability is binary: a system’s resource mix is
considered either reliable or not when compared to a
reliability criterion. e cost of achieving this reliability
has not typically been taken into account.
Today, there are many more pieces to the reliability
puzzle. Mitigations now include fossil resources, solar,
wind, various storage technologies, hybrid resources,
various congurations of demand response and load
exibility, and transmission. Dierent combinations
of resources oer distinct reliability proles, with
reliability trade-os and dierent costs.
Lack of Transparency Around
an Arbitrary Criterion
As discussed above, a common reliability criterion used
by many system planners is 1-day-in-10-years LOLE.
However, this criterion was developed in the middle
of the 20th century, with limited rationale as to its
selection and limited evaluation of the costs and benets
of achieving this denition of reliability. e arbitrary
nature of the 1-day-in-10-year LOLE criterion is con-
cerning, despite its use as the de facto reliability standard
across a wide range of dierent systems having hetero-
geneous resource mixes, consumer needs, regulatory
structures, and markets.
Simply put, the 1-day-in-10-year LOLE criterion is an
arbitrary line in the sand. System planners and regulators
set the criteria and determine a portfolio to be reliable
or not, regardless of the costs incurred to ratepayers.
Decisionmakers are left without knowledge of the costs
necessary to achieve the target reliability, and they rarely
consider the costs and benets of measures taken to
increase reliability.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 26
FIGURE 16
Comparing the Cost with the Value of Adding Resources for Reliability
Implied value of lost load per MWh
Source: Regulatory Assistance Project / Hogan and Littell (2020).
0 1 2 3 4 5 6 7 8 9
Loss of load expectation (hours per year)
$500,000
$450,000
$400,000
$350,000
$300,000
$250,000
$200,000
$150,000
$100,000
0
One event
in 10 years
(representative)
One day
in 10 years
VoLL = $25,000/MWh
e implications of this lack of awareness are great.
Resource adequacy analysis sets the foundation for
resource procurement and investment decisions by ver-
tically integrated utilities, and it sets the quantity needs
for competitive capacity markets. Although the nancial
impact of meeting the reliability criteria is large, the
current lack of transparency around the costs of dierent
approaches to reliability makes it impossible to perform
a rigorous cost-benet analysis.
Nonlinear Relationship Between
Reliability and Costs
The modern power system is much more dynamic than
systems of the past. For example, as consumers become
increasingly aware of their energy consumption, costs,
and alternative objectives like environmental impact, load
exibility becomes an important resource. As a result,
new resource adequacy analysis should be designed to
increase cost transparency so that regulators, policymakers,
and consumers understand the relative costs of dierent
levels of and approaches to reliability and can make
informed investment decisions.
Figure16 shows an example of the relationship between
cost and reliability. On the x-axis is LOLH per year for a
given system. To make the system more reliable (moving
from right to left), additional gas turbine capacity is add-
ed, reducing LOLH but leading to an increase in costs.
On the y-axis is the implied (implicit) value of lost load,
or the incremental change in cost relative to the change
in reliability (Hogan and Littell, 2020).
e chart shows a highly non-linear relationship between
reliability and cost, and illustrates that a 1-day-in-10-year
reliability criterion could be much more expensive to
consumers than higher levels of reliability achieved by
other means. While value of lost load (VoLL) is a highly
debated metric—and varies considerably based on cus-
tomer type—transparency in the costs of the reliability
criterion is critical. Although it may be impossible to
identify an economically ecient reliability level because
it is hard to speculate how much reliability is worth to a
diverse group of customers, there needs to be a clear un-
derstanding among policymakers, regulators, and system
planners of what incremental reliability costs consumers.
Such transparency could reveal that, in some cases, incre-
mental reliability is relatively aordable and worth the
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 27
investment, while in other cases it is extremely expensive
but purchased anyway because it’s hidden beneath an
arbitrary reliability requirement. Transparency in the cost
versus reliability relationship will allow wiser decisions
around reliability improvements going forward.
Understanding Resource Adequacy’s
Share of Overall Reliability
Factors other than resource adequacy also play a role in
power system reliability, of course. For example, failures
can be due to distribution outages, transmission outages,
network instability, and cyber attacks. Setting reliability
requirements for resource adequacy must be balanced
with allocating resources toward other forms of reliability.
Ultimately, the consumer does not dierentiate between
reasons for lost power. System planners need to make
sure that the benets assumed from a resource adequacy
requirement and capacity procurements are actually those
needed to ensure grid reliability, as opposed to invest-
ments in transmission and distribution infrastructure,
grid hardening, and cyber defense.
FIGURE 17
Lost Load Energy in Australia by Reliability Type from 2007 to 2016
Supply interruptions (GWh)
Source: Australian Energy Market Commission Reliability Panel (2020).
07/08 08/09 09/10 10/11 11/12 12/13 13/14 14/15 15/16 16/17 17/18 18/19
Financial year
300
250
200
150
100
50
0
■
Distribution
■
Transmission
■
Security
■
Reliability
Without translating the reliability requirements to
economic costs as a common comparison, it is impossible
to know whether reliability dollars are being allocated
eciently. Without this economic consideration, system
planners risk over-procuring capacity without signicantly
increasing system reliability. Given limited resources,
system planners need to allocate investment appro-
priately across other facets of reliability.
Figure 17, showing the sources of supply interruptions
in the Australian National Energy Market from 2007
through 2019, indicates that only a tiny fraction (0.3
Without translating the reliability
requirements to economic costs as a
common comparison, it is impossible
to know whether reliability dollars are
being allocated efficiently.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 28
percent of all lost load) was due to capacity shortfalls
(AEMCRP, 2020). e vast majority of lost load and
customer outages were due to failures and outages on the
distribution system, and a single security-related event
(South Australia blackout) in 2016. is situation is not
uncommon, indicating that the industry may be focusing
too much on reliability based on resource adequacy and
too little on distribution-system reliability. Transparently
showing the economic costs of incremental resource
adequacy improvements is critical to understanding
the dierent sources of reliability for each system.
e costs of achieving 100 percent resource adequacy
on a high-renewables grid would be innite, and sense-
less for most consumers when the same money could be
spent on other reliability mitigations. A single resource
adequacy criterion centered solely on the number of
MW, absent economic considerations, is therefore
unjustied. Grid planners and regulators should have
a clear understanding of the costs associated with achiev-
ing dierent reliability targets in dierent ways, to ensure
that the value provided to the customer is worth the cost
of a given investment—that the resource adequacy for
which customers are being asked to pay is actually the
type of reliability needed on the grid.
Grid planners and regulators should
have a clear understanding of the costs
associated with achieving different
reliability targets in different ways.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 29
Looking Forward
I
f the rolling blackout events in California and Texas in
2020 and 2021 teach us anything, it is that the industry
cannot continue to approach resource adequacy as we
have in the past. ese were not failures of the evolving
resource mix, but rather failures of planning. Existing
methods that have served the industry well historically
are not adequate as the resource mix changes to one of
variable renewable energy, energy storage, and exible
loads, and as power systems experience increased cor-
relation of generator outages due to weather. Today, the
industry, and ultimately consumers, are paying the price
of limited planning and analytical shortcuts that do
not capture chronological operations and weather-
inuenced correlated events.
Many of the metrics currently used, such as the traditional
planning reserve margin, are not adequate for modern
power system planning. e industry needs new options.
For now, we must rely on more in-depth analysis of real
systems, or general rules that can be applied. As electric-
ity system stakeholders, we need to roll up our sleeves
and do the hard analytical work. What we learn will help
us develop heuristics and new rules to make resource
adequacy analysis easier and less costly to conduct and
simpler to understand.
While considerable work is needed to fully dene what
robust resource adequacy looks like, some basic rst steps
can lead to improved resource adequacy analysis now.
ese steps include:
• Considering how the rst principles of resource
adequacy should be applied to the specic system
being examined
• Makingtheresourceadequacyanalysispublicand
easily accessible, so that the community of stakeholders
can benet from seeing a diverse set of case studies
from regions around the world with dierent resource
mixes, load proles, and characteristics of system risk
• Collectingasmuchchronologicalandcorrelated
hourly historical weather and load data as possible,
and then considering whether the available historical
data are suciently representative of possible future
events, including consequences of climate variability
and change
• Reportingabroadersetofresourceadequacymetrics
than simply an average LOLE, including hourly EUE
and additional information on the distribution of out-
ages. Metrics should also be used to develop detailed
statistics on the shortfall events themselves in order
to better characterize the size, frequency, duration, and
timing of events so that mitigation measures can be
properly sized.
Consistency in resource adequacy analysis and reporting
will provide the necessary data and better insight on what
shortfall events look like across many systems. Such
consistency will help the industry better understand
how resource adequacy risk shifts with changes in the
underlying resource mix of increased variable renewable
energy, energy storage, and load exibility of modern
power systems.
Consistency in resource adequacy analysis
and reporting will provide the necessary
data and better insight on what shortfall
events look like across many systems.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 30
References
AEMCRP (Australian Energy Market Commission Reliability Panel). 2020. Annual Market
Performance Review, 2020. Sydney, Australia. https://www.aemc.gov.au/data/annual-market-
performance-review-2020/sources-of-supply-interruptions-in-the-nem-2008-09-to-2018-19.
APPA (American Public Power Association). 2021. America’s Electricity Generation Capacity, 2021
Update. Washington, DC. https://www.publicpower.org/system/les/documents/Americas-
Electricity-Generation-Capacity_2021-update.pdf.
Calabrese, G. 1947. “Generating Reserve Capacity Determined by the Probability Method.”
AIEE Transactions 66:1439-1450.
CAISO (California Independent System Operator). 2021. Final Root Cause Analysis: Mid-August 2020
Extreme Heat Wave. Folsom, CA. http://www.caiso.com/Documents/Final-Root-Cause-Analysis-
Mid-August-2020-Extreme-Heat-Wave.pdf.
ENTSO-E (European Network of Transmission System Operators for Electricity). 2020a. Mid-term
Adequacy Forecast 2020: Appendix 2; Methodology. Brussels, Belgium. https://eepublicdownloads.entsoe.
eu/clean-documents/sdc-documents/MAF/2020/MAF_2020_Appendix_2_Methodology.pdf.
ENTSO-E. 2020b. Resource Adequacy Implementation Road Map. Brussels, Belgium. https://
eepublicdownloads.entsoe.eu/clean-documents/cep/implementation/BZ/201210_ERAA_
implementation_Road_Map.pdf.
ERCOT (Electric Reliability Council of Texas). 2021a. Daily Balancing Authority Operations Report.
Austin, TX. http://mis.ercot.com/misapp/GetReports.do?reportTypeId=13458&reportTitle=Dai
ly%20Balancing%20Authority%20Operations%20Report&showHTMLView=&mimicKey.
ERCOT. 2021b. Final Seasonal Assessment of Resource Adequacy for the ERCOT Region (SARA) Winter
2020/2021. Austin, TX. http://www.ercot.com/content/wcm/lists/197378/SARA-
FinalWinter2020-2021.pdf.
ERCOT. 2021c. Update to April 6, 2021, Preliminary Report on Causes of Generator Outages and Derates
During the February 2021 Extreme Cold Weather Event. Austin, TX. http://www.ercot.com/content/
wcm/lists/226521/ERCOT_Winter_Storm_Generator_Outages_By_Cause_Updated_
Report_4.27.21.pdf.
Hering, G., and J. Staneld. 2020. “‘You have to rethink these old ways’: Parting Advice from
CAISO’s Retiring CEO.” S&P Global Market Intelligence, September 25. https://www.spglobal.com/
marketintelligence/en/news-insights/latest-news-headlines/you-have-to-rethink-these-old-ways-
parting-advice-from-caiso-s-retiring-ceo-60481529.
Redefining ResouRce AdequAcy foR ModeRn PoweR systeMs EnErgy SyStEmS IntEgratIon group 31
HNEI (Hawai’i Natural Energy Institute). 2020. Report to the 2021 Legislature, Annual Report from
the Hawai’i Natural Energy Institute. Honolulu, HI. https://www.hnei.hawaii.edu/wp-content/
uploads/2021-HNEI-Annual-Report.pdf.
Hogan, M., and D. Littell. 2020. Get What You Need: Reclaiming Consumer-Centric Resource Adequacy.
Montpelier, VT: Regulatory Assistance Project. https://www.raponline.org/knowledge-center/get-
what-you-need-reclaiming-consumer-centric-resource-adequacy.
JCM ( Joint and Common Market). 2021. JCM Contour Map. Carmel, IN. https://www.miso-pjm.
com/markets/contour-map. Accessed February 15, 2021.
MISO (Midcontinent Independent System Operator). 2021. MISO’s Renewable Integration Impact
Assessment, Summary Report. Carmel, IN. https://cdn.misoenergy.org/RIIA%20Summary%20
Report520051.pdf.
Pickerel, K. 2021. “Swell Energy to Establish 6,000-battery Virtual Power Plant for Hawaiian
Electric.” Solar Power World, January 19. https://www.solarpowerworldonline.com/2021/01/swell-
energy-to-establish-6000-battery-virtual-power-plant-for-hawaiian-electric.
Schlag, N., Z. Ming, A. Olson, L. Alagappan, B. Carron, K. Steinberger, and H. Jiang. 2020. Capacity
and Reliability Planning in the Era of Decarbonization: Practical Application of Eective Load Carrying
Capability in Resource Adequacy. San Francisco, CA: Energy and Environmental Economics. https://
www.ethree.com/wp-content/uploads/2020/08/E3-Practical-Application-of-ELCC.pdf.
PHOTOS
Cover: © Creative Commons/Simeon W
p. 3: © Alamy/Micael Malmberg
p. 8: © Creative Commons/NASA-NOAA-NPP-VIIRS
p. 9: © Creative Commons/Ian Dick
p. 12: © Creative Commons/Jim McDougall
p. 17: © iStockphoto/Stegarau
p. 21: © iStockphoto/conceptualmotion
p. 25: © Creative Commons/Eva Cristescu
p. 28: © Wonderlane
The report is available at https://www.esig.
energy/reports-briefs.
To learn more about the recommendations
in this report, please send an email to
resourceadequacy@esig.energy.
The Energy Systems Integration Group is
a nonprofit organization that marshals the
expertise of the electricity industry’s tech-
nical community to support grid transfor-
mation and energy systems integration and
operation, particularly with respect to clean
energy. More information is available at
https://www.esig.energy/reports-briefs
or info@esig.energy.
Redefining Resource Adequacy
for Modern Power Systems
A Report of the Redefining Resource Adequacy Task Force
of the Energy Systems Integration Group
ES
ENERGY SYSTEMS
INTEGRATION GROUP
