Metrics for Quantifying Flexibility

in Power System Planning

TECHNICAL PAPER SERIES

Power System Transformation

9802731
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Executive Summary
In power systems, flexibility is used generically to refer to the ability to This white paper summarizes the development of flexibility assessment
adapt to changing conditions while providing electricity safely, reliably, methods which augment traditional resource adequacy planning pro-
affordably, and in an environmentally responsible manner. This paper cesses. Methods range from post-processing techniques that use the results
focuses on operational flexibility, which is the ability to ramp and cycle of other planning approaches and then calculate various flexibility
resources to maintain a balance of supply and demand on timescales adequacy metrics, to modification of existing resource adequacy methods
of hours and minutes through reliably operating a system at least cost. themselves to account for variable generation and other drivers. Key
Increased flexibility in power systems is needed as a result of growth in conclusions include the following:
variable generation and the impacts of this growth on power markets,
• The need to consider system flexibility arises for systems facing grow-
projected future uncertainty in fuel prices (primarily natural gas) for
ing penetrations of wind and solar generation, or where the nature of
power generation, changing patterns of demand, and environmental
the demand has changed sufficiently to cause growing forecast errors.
regulation. Potential consequences of insufficient flexibility include
reduced reliability, increased wear and tear on power system equipment, • EPRI’s role is to lead, support development of, and transfer to the
higher electricity production costs, and difficulty integrating renewable industry flexibility metrics and tools that can help utilities and
generation. system operators plan future power systems to operate reliably, safely,
environmentally responsibly and at least cost.
Recent regulatory actions are considering power system flexibility. In
states like California, or countries like Ireland, where renewables are • While flexibility metrics and tools have been developed, strong col-
playing a greater role in the energy mix, regulators are examining the laboration between industry (electricity providers and ISOs/RTOs)
risks of system inflexibility as part of market design and planning ap- and research partners is required to demonstrate, refine, and transfer
plications. As regulators, utilities, and independent system operators/ the knowledge garnered to date.
regional transmission organizations (ISO/RTOs) respond, they require
methods to quantitatively assess the operational flexibility of their power
system in order to evaluate the optimal means of ensuring sufficient flex-
ibility in planning processes.

Table of Contents
Executive Summary......................................................... 2 Flexibility Assessment...................................................... 8
Background.................................................................... 3 EPRI InFLEXion Tool..................................................... 8
Definition of Flexibility................................................ 3 Post-Processing Analytical Flexibility Metric Calculation... 9
Drivers of Increased Need for Operational Flexibility...... 3 EPRI Tool Flexibility Metrics.......................................... 10
Potential Consequences of Insufficient Flexibility............. 5 Deterministic Production Costing Simulation Engines....... 11
Flexibility is Not Necessarily An Issue In Every System.... 5 Monte Carlo-Based Resource Adequacy Calculation....... 12
The Need to Measure Flexibility................................... 5 Monte Carlo Production Costing Simulation................... 13
Regulatory Trends: Flexibility In the Planning Process........ 6 High-Level Screening Tools.......................................... 13
Planning Processes..................................................... 6 Transmission Deliverability Assessment.......................... 14
Storage Mandates...................................................... 6 Actions for the Industry and for R&D................................ 14
Traditional Long-Term System Planning............................. 7 Acronyms....................................................................... 15
Overview.................................................................. 7 References...................................................................... 15
Planning Challenges: Flexibility.................................... 7

9802731
Metrics for Quantifying Flexibility in Power System Planning 2 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Background
Definition of Flexibility (NERC) Integration of Variable Generation Task Force released a
In power systems, flexibility is the ability to adapt to changing special report entitled: “Accommodating High Levels of Variable
conditions while providing electricity safely, reliably, affordably, and Generation.” The report highlights the fact that as the penetration
in an environmentally responsible manner. In this paper, the defini- of variable generation reaches relatively high levels, the character-
tion of operational flexibility is refined to describe a power system’s istics and operation of the bulk power system will be significantly
ability to ramp and cycle resources to maintain a balance of active altered. The primary driver of this change is the increase in overall
power supply and demand through reliably operating a system at system variability [1]. For more information on these drivers, refer
least cost. These changes can be both upward and downward ramps to the forthcoming EPRI white paper on “Electric Power System
over a wide variety of time scales, ranging from minutes to hours. Flexibility: Challenges and Opportunities [2].”

Note that there are other aspects of power system flexibility related Figure 1 shows an example of how the variability of wind and
to time scales that are not addressed in this paper. For example, over solar generation requires significant system flexibility in Germany
time scales longer than the time required to start the slowest starting over the course of three days in August 2013. It also suggests the
unit (e.g., a nuclear unit), longer term flexibility is required to ad- complexity of planning for flexibility needs – at times during this
dresses seasonal variation in hydro, wind or solar generation. At the period wind and solar generation results in significant increases in
other end of the time spectrum, also not addressed in this paper, are flexibility needs, while at other times the opposite is true. Installed
generation and transmission contingency events on the order of tens German photovoltaics (PV) and wind generation reached in excess
of seconds to fractions of seconds which require frequency response, of 65 GW in 2013. The flexibility requirements over the course of
frequency regulation and other sorts of flexibility. the three day (Saturday to Monday) period can be broken down as
follows:
Drivers of Increased Need for Operational
1. No morning ramp materializing on Saturday morning: PV
Flexibility
negates the demand rise and later causes a midday valley in
The following factors are driving the requirement for operational
net load (net load is demand minus variable generation). This
flexibility in power system supply and demand:
requires forecasting to avoid excess capacity online on Saturday
• Growth in variable generation (e.g., wind and solar photovol- morning and down ramping to accommodate the midday valley.
taic), which by its variable and uncertain nature presents chal-
2. Saturday afternoon valley for 2 hours: Keep sufficient units
lenges for reliable power system operation, drives the need for
online to meet the approaching evening ramp up in net load.
flexibility.
This poses a challenge for minimum down time of de-committed
• Environmental regulations impose constraints on conventional units and requires conventional generation to have the ability to
generation operations and generator retirements that can also turn down to lower output.
impact provision of system flexibility.
3. Saturday evening peak: A 13-GW ramp up for a 2-3 hour peak
• Variable generation significantly impacts ISO/RTO power net load and then a ramp down to a night valley. This poses
markets and utility system operation through the decommitment challenging upward and downward ramping, as well as mini-
of flexible generation, and these effects in turn, further drive the mum online time of committed units.
need for flexibility from quick start resources.
4. Sunday morning flat-line net load: Night time and morning
• Consumer adoption of energy efficient appliances, electric net load remains relatively constant as solar generation matches
vehicles, zero net energy buildings, distributed resources, and the demand shape on this day.
other changes might lead to a more variable load, sharper peaks,
5. Sunday afternoon valley: As demand drops off, more non-vari-
or may provide extra flexibility.
able generation has to be run at part load, minimum generation,
In April 2009, the North American Electric Reliability Corporation or de-committed. The instantaneous penetration of variable

9802731
Metrics for Quantifying Flexibility in Power System Planning 3 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

generation (VG) reaches 50%. The lack of synchronous genera- 8. Monday peak net load reduction: At 9 am, VG contributes to
tion may be challenging for the stability of the system and for peak demand reduction for the remainder of the daylight hours,
minimum offline times of de-committed generation. reducing the effective ramping requirement on conventional
generation.
6. Sunday evening ramp: After prolonged, depressed net load,
units must be committed and ramped to meet an 18 GW up- In this single example, the system required fast upward and down-
wards ramp over 7.5 hours. This is challenging for start up times ward ramp rates, short minimum online and offline times, low
and upward ramping limits. minimum generation levels, generator cycling, and quick start up
times from its conventional generation fleet. In addition, benefits
7. Monday night time valley and morning rise: After the steep
can be gained from forecasting capabilities, flexibility-oriented oper-
ramp for the evening peak, net load falls off with reduced
ating practices and the ability to trade with neighboring systems.
demand, before ramping up again at 3 am to meet the 20 GW
demand pick up on Monday morning. VG does not signifi-
cantly affect the timing or the rate of increasing net load in this
instance.

70,000 German Demand and Net Load

65,000 August 17-19, 2013
60,000 Source data: transparency.eex.com and entsoe.eu 8
55,000
50,000
MW

45,000
40,000 1 3 6 7
35,000
30,000 4 Demand
25,000 2 5 Net Load
Saturday Sunday Monday
20,000
0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00
60 Instantaneous VG Penetration %
Penetration (%)

50
40
30
20
10
0
0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00
Hour
Figure 1 – Wind and PV Variability Increases the Need for System Flexibility. (Source: Constructed from transparency.eex.com and entsoe.eu)

9802731
Metrics for Quantifying Flexibility in Power System Planning 4 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Potential Consequences of Insufficient Flexibility generation, acquire power from neighboring utilities or increase
Failure to incorporate sufficient flexibility into the power system can incentives to electricity customers to be an active participant in
cause some combination of the following potential consequences: the market. Insufficient flexibility can give rise to sharp price
spikes and elevated ancillary service prices. These costs are typi-
• VG curtailment and difficulty meeting renewable energy tar-
cally passed on to ratepayers and may ultimately result in higher
gets: Without sufficient flexibility, utilities may have difficulty
rates.
accommodating the required levels of renewable power genera-
tion, without compromising some combination of reliability, In an indication of the immediacy of the need for flexibility, some
increased wear and tear on equipment, production costs, and power systems are experiencing system conditions that are rare in
environmental regulations. One solution would be to curtail VG the history of their operation. These system conditions illustrate
frequently, however this reduces the ability of the system to meet the increasingly end-to-end nature of flexibility challenges that the
policy targets. industry faces (from a closed nuclear plant to increased customer
consumption of gas) and the interrelated nature of the electricity
and gas supply and delivery systems [3]. Internationally, the Bulgar-
Flexibility is Not Necessarily an Issue ian system operator ESO is one example of a system operator who
in Every System is currently experiencing flexibility issues to such an extent that VG
curtailment is required for flexibility reasons.
The need for flexibility depends on the particular region’s
generation mix, forecasting accuracy, scheduling and market
design, demand response resource and regulatory environ- WHILE YOU’RE HERE
ment. Some systems have low penetration of variable gen-
• How will the net load profile evolve in your region
eration due to less aggressive renewable portfolio standards
in 5/10/20 years time?
and relative lack of wind and solar availability year round.
• Has your company assessed the potential risks
As a result, flexibility may not yet be a significant issue (but
associated with insufficient operational flexibility?
may become an issue in the future). Another factor is the
• Will your role in the industry change as a result of
flexibility of non-renewable energy resources in the service
requirements for system flexibility?
territory. Systems with a high capacity of natural gas plants,
for example, may already incorporate significant operating
flexibility. The Need to Measure Flexibility
Traditionally, public utilities’ reliability focus has been on resource
• Reduced reliability: Lack of flexibility may result in the inability adequacy and security. NERC and utilities have established metrics
to meet net load from time to time. This may result in a greater that utilities employ to evaluate these aspects of reliability for their
dependence on emergency procedures to maintain service. systems. Today, flexibility has emerged as a significant consideration
in some systems, due to the drivers summarized above. As a result,
• Increased wear and tear on power system equipment: To
there is a corresponding need to quantify future requirements for
meet fluctuating load, power system equipment that is forced
flexibility and assess a system’s ability to meet those requirements, as
into cycling or peaking duty (and not designed for this mode
well as to maintain traditional resource adequacy to meet peak de-
of operation) may wear at a more rapid rate than anticipated in
mand. The purpose of this paper is to summarize various techniques
its planned operating mode. This may result in faster consump-
to quantify flexibility in the form of metrics, to summarize EPRI
tion of equipment life, or create the need for extensive repairs,
research in this area, and to identify future research needs.
upgrades, or replacements earlier than anticipated, increasing
O&M costs.

• Higher electricity production costs: To reliably meet demand

and avoid potential equipment damage, utilities may need to
purchase additional reserves, change the dispatch of conventional

9802731
Metrics for Quantifying Flexibility in Power System Planning 5 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Regulatory Trends: Flexibility in the Planning Process

Planning Processes The first effort is the Flexible Resource Adequacy Criteria and
A significant challenge for utilities is assessing how much opera- Must-Offer Obligation (FRACMOO), wherein the California Inde-
tional flexibility already exists and how much will be needed – and pendent System Operator (CAISO) is working with the California
when. Resource planning and procurement processes typically are Public Utility Commission (CPUC) and local regulatory authorities
not focused on delivering flexibility in real time operations. A June to ensure flexible capacity resources are available to reliably operate
2012 report from the Western Governors’ Association suggests new the grid while fulfilling state energy mandates. The work includes
metrics and methods are needed to assess flexibility of resource port- developing the tariff changes necessary for the ISO to accommodate
folios and resource capabilities needed in the future [4]. the resource adequacy flexible capacity requirements adopted by
regulators. This includes establishing availability, must offer obliga-
tions, and default provisions for entities that fail to procure their
flexible capacity allocations [5].

The other effort, California’s Long Term Procurement Planning

(LTPP), reviews and approves plans for the utilities to purchase
energy, establishes policies and utility cost recovery for energy
purchases, ensures that the utilities maintain sufficient capacity,
with the right characteristics, to meet the forecasted demand profile
within a long-term energy planning process. Starting in 2010, this
biannual planning process included a significant focus on flexibility
needs [6].

Similar efforts are also underway in Oregon, where the state PUC is
requiring that utilities start examining how much variable genera-
Recently, regulators have begun considering, or have already
tion they expect to integrate in the coming years, and what re-
stipulated, that utilities consider flexibility as part of their integrated
sources they have to balance it. All three utilities in Oregon’s PUC
resource planning (IRP) or other planning processes. In California,
service territory – Portland General Electric, Pacific Power, and
two distinct efforts are currently underway.
Idaho Power – were planning to begin the inquiry alongside typical
planning for power generation in their IRP process [7].

Storage Mandates
Energy storage is one potentially important strategy to man- Other states, including New York and Minnesota, are explor-
age and balance the inherent variability of renewable electric- ing similarly targeted initiatives. For example, Consolidated
ity generation. In September 2010 California Assembly Bill Edison Company of New York has initiated a Distribution
AB2514 was signed into law. AB2514 required the California Load Relief Program, governed by the New York State Public
Public Utility Commission (CPUC) to open a new proceed- Service Commission, which can include use of storage to
ing to “establish procurement targets, if any, for each load temporarily reduce electricity usage to help maintain reliabil-
serving entity to procure viable and cost-effective energy ity [9].
storage systems and, by October 1, 2013, to adopt an energy
As storage is mandated to meet flexibility needs in Califor-
storage procurement target, to be achieved by each load-serv-
nia and potentially in other states, utilities will need a way
ing entity by December 31, 2015, with a second procurement
to measure the flexibility of their system and the flexibility
target to be achieved by December 31, 2020 [8].”
contribution of storage.

9802731
Metrics for Quantifying Flexibility in Power System Planning 6 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Traditional Long-Term System Planning

Overview
The red-shaded steps in Figure 2 show, at a
WHILE YOU’RE HERE
conceptual level, the steps in traditional long-
term power system planning processes. Blue- • What parts of this planning framework concern your company?
shaded steps represent the four levels of flexibility • What are the business risks which are associated with flexibility for
assessment which may need to be added into an your company?
integrated planning framework. These levels are
discussed subsequently in the paper.

The traditional planning process generally begins

Long Range Forecast
with long-range forecasting, which includes load
forecasting and now renewable resource forecast-
ing. These forecasts are fed into a series of plan- Capacity Level 1
ning models that evaluate the adequacy of supply Adequacy Flexibility
to meet system demand and of the transmission Assessment Requirements
to deliver that supply to loads. If the existing
resource and network plans are found to be
inadequate, expansion options are evaluated. The Resource No
exact implementation of the analyses presented Expansion Is Adequate?
in Figure 2 may alter from system to system. Options/Screening
For example, different organizations may have Yes
responsibility for certain parts, or the order of the
processes may be altered. This process calculates
Transmission Level 2 Flexible
various reliability indices to assess whether the
Adequacy Resource
combination of resources, transmission, and load Assessment Characteristics
will meet planning reserve margins and other reli-
ability requirements.

Transmission
Planning Challenges: Flexibility No Yes Production Cost
Expansion Is Adequate?
Traditional planning approaches typically do not Simulation
Options
fully model the variable and uncertain nature
Yes
related to certain renewables. These approaches
thus do not usually consider the generators’
flexibility characteristics such as start up times or Level 4 Flexibility Level 3 System
ramp rate levels. Traditional planning processes
Delivery Flexibility
Assessment Analysis
also do not provide indices that indicate the flex-
ibility requirements or overall flexibility of the
power system.
Cost Assessment & Planning Decisions

Figure 2 – Conceptual Example of Power System Planning Process (RED) with New Flexibility
Processes

9802731
Metrics for Quantifying Flexibility in Power System Planning 7 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Flexibility Assessment
In light of the need for enhanced power system operational flexibil- issues occur. Here, we focus on a particular set of tools and metrics
ity, a number of different entities – utilities, ISOs/RTOs, regulators, developed by EPRI research to analyze flexibility issues.
academia, research bodies, consultants, and software firms – have
begun to develop tools and techniques that consider and quantita-
tively measure system operational flexibility in a planning context. WHILE YOU’RE HERE
Figure 2 shows four different points at which flexibility assessment • Do you need to consider flexibility as part of your
could be integrated within a conceptual long-term planning process. planning process?
These four points correspond to different flexibility assessment • Are your planning processes ready to adapt to
stages or levels based on where in the planning process they occur. consider flexibility needs?
It may be the case that not all levels of assessment would be neces- • Have you identified the tools, capabilities and
sary each time a planning process is conducted if the results from knowledge required to prepare for this transition?
an earlier step indicate to the user’s satisfaction that no flexibility

EPRI InFLEXion Tool

InFLEXion is a multi-level flexibility assessment tool that can InFLEXion tool enables decision makers to understand flex-
be used to support and validate power system planning (see ibility needs and future power system performance. Future
the sample screen capture below). Starting from high-level developments will include consideration of the impact of
flexibility requirements screening in Level 1 and building transmission networks on system flexibility. For more infor-
up to calculation of system flexibility adequacy metrics (i.e., mation on this tool, contact EPRI regarding the Strategic and
PFD, EUR, IRRE and well-being metrics defined below), the Flexible Planning Research Project in Grid Planning Program
40.

9802731
Metrics for Quantifying Flexibility in Power System Planning 8 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Post-Processing Analytical Flexibility Metric resource capabilities. The goal of this level is to help the user to
Calculation understand the potential capabilities of a resource fleet in a given
EPRI has developed an analytical tool which can be used to assess system. The resource fleet characterization, coupled with user
flexibility related issues in each of the four levels outlined in Figure knowledge, helps to inform decisions for further planning actions,
2 and includes a calculation of system flexibility metrics. The tool such as resource adequacy, production cost model setup and re-
operates as a post-processing algorithm in that it has the capability source requirements.
to analyze the flexibility of a system based on inputs and outputs
Level 3 includes analyses of the flexibility available from the resourc-
of the other planning processes listed above. One example is the
es once they have been dispatched, and the flexibility required by
use of the simulated generator dispatches from the production cost
the system at that time. Environmental limits are enforced through
modeling stage to calculate system flexibility metrics in Level 3. The
constraints in the scheduling process. Dedicated flexibility adequacy
outcome from each level of analysis can be used at future points
metrics are calculated at this point. The goal of this level is to assess
in the planning process, so as to assist users in ensuring that both
the flexibility that the resource fleet is technically capable of offering
capacity adequacy and flexibility are considered when determining
at each period, and measuring the adequacy of that capability rela-
future resource requirements.
tive to the flexibility required in the system. This is achieved with
the implementation of the flexibility metrics described below.

Level 4 builds on Level 3 analysis to include the constraints of the

transmission network when delivering flexibility. The deliverability
of the available flexible resources can be determined as part of the
planning process. The goal of this level is to determine the extent
to which the transmission network impacts the ability of the system
operator to balance the system using the available resources. The
insight gained from these analyses can then be used as input into
transmission expansion planning activities and to guide the place-
ment for new resources.

The first three levels of analysis are implemented in the InFLEXion:

The analysis methods provided in the tool are post-processing meth- Flexibility Assessment Tool, with transmission system impacts to be
ods that depend on the outcome of previous steps, such as resource implemented in level 4 during 2014.
expansion or production cost simulation. Development is also
EPRI and member utilities are in the vanguard in terms of develop-
being undertaken to include flexibility concerns within some exist-
ing and implementing flexibility assessment on these four levels,
ing steps, such as the integration of flexibility concerns in capacity
and are in extensive collaboration with other organizations. These
adequacy calculations, rather than as a separate step.
include the International Energy Agency (IEA), Lawrence Berkeley
Level 1 involves analysis of the flexibility requirements, driven by National Laboratory (LBNL), National Renewable Energy Labo-
net load ramping due to demand, wind, and solar. The purpose is to ratory (NREL), CPUC, and Northwest Power and Conservation
give the user an understanding of how ramping requirements may Council.
change. The analysis incorporates changes to frequency of occur-
rence, magnitude, and timing of net load ramps over a variety of
time horizons. This information can be used in decision making for WHILE YOU’RE HERE
either planning simulations or for resource selection strategies. • If flexibility will become an issue for your system,
do you have a plan to adopt new analyses?
Level 2 includes analyses of the flexible characteristics of resources.
• Flexibility assessment is data intensive; will the right
Analyses include the operating range of resources, distributions
data be available when you need it?
of start-up times and ramp rates, and a qualitative comparison of

9802731
Metrics for Quantifying Flexibility in Power System Planning 9 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Period of Flexibility Deficit (Periods)

1,400
1,200
1,000
800
600
400
200
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Time Horizon (Hours)
Figure 3 – Number of Periods in Flexibility Deficit for Various Time Horizons [10]

EPRI Tool Flexibility Metrics Figure 3 shows the number of periods in flexibility deficit (PFD)
The EPRI tool calculates four flexibility adequacy metrics as part for various time horizons in hours. This figure was constructed
of the functionality in Levels 3 and 4 : the periods of flexibility using the EPRI InFLEXion tool for a test system under the assump-
deficit (PFD), expected unserved ramping (EUR), insufficient ramp tion that the 98th percentile of the net load ramp distribution
resource expectation (IRRE), and a flexibility well-being assessment. occurred at each dispatch period (i.e., not the absolute worst case,
but nearly the worst case starting from each dispatch position). The
Metric: Period of Flexibility Deficit (PFD) figure shows that for short time horizons (e.g., 1 hour in length),
The PFD is a measure of the number of periods when the available there are a large number of periods when a flexibility deficit would
flexible resources were less than the required flexibility for a given occur. For time horizons of approximately 8-10 hours, there are a
time horizon and direction. The PFD is calculated for a range of moderate number of periods with flexibility deficits due to the large
user selected time horizons in the chosen direction based on the magnitude of net load ramps in this horizon and lack of units that
flexibility available from simulated production time series. As well as can start in less than 9 hours. For other time horizon lengths, the
identifying the number of periods of flexibility shortages, the PFD number of flexibility deficits is relatively low. This shows that over
helps to identify the time horizons associated with the shortages. the course of a year, the primary concern in terms of number of
This is important because the solution to each flexibility issue may flexibility deficits is in short duration ramps [9]. Results for other
require different actions. Deficits over shorter time periods can be systems could be very different than for this test system.
mitigated by improved short-term forecasting capabilities, dynamic
and probabilistic reserve procurement strategies, reserve sharing Metric: Expected Unser ved Ramping (EUR)
between connected systems, and new fast-start, high-response In addition to the frequency of flexibility deficits, the magnitude of
resources. Over longer time periods, such as 3 to 8 hours, better any shortage is important. The expected unserved ramping is the
long-term forecasts aid in the scheduling of large, slow starting gen- total magnitude of the deficit of net flexibility. The EUR is similar
eration and intertie flows. In both cases operational practices can be but not identical to the expected unserved energy metric in capac-
adopted first, but if there is still a flexibility deficit, new resources ity adequacy planning. Smaller magnitude shortfalls of flexibility
may be required. may be met by operational changes or might be determined to be

9802731
Metrics for Quantifying Flexibility in Power System Planning 10 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

immaterial in the context of the operation of the system, depending Metric: Flexibility Well-Being Assessment
on the risk level. Higher EUR values may require new operational A well-being analysis combines the information in the PFD and
modes for existing resources or large-scale demand side, generation, EUR metrics to determine whether a system is in a user-defined
or transmission resource additions to supplement the system’s cur- safe, warning, or dangerous state. This could be extended to con-
rent capability. sider other metrics. The definitions of safe, warning and danger
zones have not been established to date. This requires further testing
Metric: Insufficient Ramp Resource Expectation (IRRE) and verification on systems that currently operate reliably. This
The Insufficient Ramp Resource Expectation (IRRE) is a second combination of the two metrics allows users to plot the frequency
measure of the frequency of flexibility shortfalls over a variety of and magnitude of flexibility deficits using an intuitive method of
time horizons. The key difference between the PFD and IRRE categorizing the values.
is that the IRRE uses a probabilistic approach to determine the
Figure 4 shows these zones and data points for a test case. For one
likelihood of meeting each net load ramp drawn from a distribu-
selected time horizon in the study, the PDF and EUR coordinates
tion at each period in time. In lieu of a direct comparison between
are outside the safe and warning zones. This plot enables determina-
the available flexibility and the net load ramps, a distribution of
tion of combinations of frequency and cumulative magnitude of
available flexibility is created. Using this, the probability of meet-
flexibility deficiencies that merit attention [10].
ing each ramp, as well as the expected value over a time series, can
be calculated. While the PFD and IRRE give similar results, the
IRRE can be a more conservative approach to flexibility assessment. Deterministic Production Costing Simulation Tools
Hence, out of a time series that is 8760 hourly periods in duration, Traditional production costing models can play a role in defining
the IRRE values would be expected to be consistently higher than power system flexibility metrics. The dispatch and commitment
the PFD values because of the probabilistic treatment of ramp- outcomes of production cost simulation are required for post
ing requirements. processing analysis or can be the basis for Monte Carlo analysis
methods. The CAISO has conducted extensive simulations using a

Danger Warning Safe DATA POINT

Period of Flexibility Deficit (Periods)

16%

12%
Potential Problem Time Horizon

8%

4%

0
0 2.6% 5.2% 7.8% 10.4% 13.0%
Expected Unserved Ramping (Percent)

Figure 4 – This Flexibility Well Being Analysis Plots the Number of Periods of Flexibility Deficit against the Sum of the Magnitudes of the Deficits over a Year
[10].

9802731
Metrics for Quantifying Flexibility in Power System Planning 11 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

production simulation tool to examine flexibility sufficiency of the conventional resource adequacy assessments (and hence their output
system, as part of the CPUC Long Term Procurement Process. This metrics) to also incorporate the impact of variable generation.
process means that each utility in California will need to consider
Resource adequacy assessments typically calculate the metrics such
flexibility needs; as of yet, the final methodology to do so is not fully
as loss of load expectation (LOLE), which results in a temporal ex-
defined. A collaborative review of models for determining flexibility
pectation of a system’s inability to meet system load. By adapting
has been published by a group led by Pacific Gas and Electric in
the LOLE methodology, a similar expectation can be calculated for
2014 [11].
a system’s inability to provide the required flexibility. This method
carries out traditional generation adequacy reliability analysis using
Monte Carlo-Based Resource Adequacy Calculation
Monte Carlo simulations of a number of variables: outages, wind
As an alternative to the post-processing method described above
output levels, etc.
to explicitly calculate flexibility metrics, flexibility requirements
can be incorporated into a traditional resource adequacy plan- EPRI recently collaborated with Astrape, Inc. to enhance a hybrid
ning approach using adequacy metrics that consider the flexibility resource adequacy and production cost tool to include consider-
component. These assessments are typically a Monte Carlo-based ation of forecast uncertainties for load, solar, and wind resources
(probabilistic) production simulation approach, in contrast to the on day-ahead, multi-hour ahead, and intra-hour bases. EPRI also
analytical or deterministic approach to produce the production time performed a case study using this enhanced version of the tool for a
series used in the post-processing method described above. system based on the state of California to understand how both the
system cost and reliability impacts of renewable resource integration
Traditional resource adequacy assessments evaluate the capability of
may change in future.
an electric system to meet firm load obligations under a wide range
of generation outage scenarios and peak load conditions. However, Key results include the quantification of how wind, solar, and
these assessments do not capture the impact of the uncertainty of load uncertainty affect LOLE and the Expected Unserved Energy
variable generation or the constraints on moving from one period (EUE). The study also demonstrated the ability to more accurately
to the next. The approach described in this section is to enhance characterize the Effective Load Carrying Capability (ELCC) for

0.300 With Uncertainty

Without Uncertainty
0.250
LOLE (Days/Yr)

0.200

0.150

0.100
Additional capacity required
0.050 to negate uncertainty
0
0 440 880 1320
Additional Capacity Requirement (MW)
Figure 5 – Incorporation of Wind and Solar Uncertainty into the Calculation Increases the LOLE Metric [10]

9802731
Metrics for Quantifying Flexibility in Power System Planning 12 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

wind and solar – it was shown to be lower when uncertainty is energy, expected over-generation, and ramping deficiencies. This
considered, by between 2.5% and 5%. The analysis also calculated utilizes a probabilistic unit commitment and economic dispatch
the year-to-year variation in wind and solar ELCC, provided the considering forecast uncertainty in wind, solar and load.
ability to quantify the trade-off between additional generation being
installed to reduce LOLE/EUE and the costs of doing so, and the High-Level Screening Tools
ability to quantify how different operational measures of increased The IEA Grid Integration of Variable Renewables (GIVAR) study
reserve can improve resource adequacy but additionally increase highlighted the issue of flexibility in its assessments of a variety
costs [10]. of categories of power systems. Using some commonly available
information about a given power system, the GIVAR study aimed
Figure 5 shows that when the effect of wind and solar uncertainty
to identify the main aspects of that system which, if changed, would
is included in the calculation of the LOLE metric, the result is a
decrease in the level of reliability calculated; this lower level is the
more realistic number. This shows that short term VG uncertainty
can impact adequacy calculations and hence should be incorporated
into analyses. For this particular system, the graph shows that for an
LOLE of 0.16 days per year, approximately 880 MW of additional
conventional capacity (here, combustion turbines) is required to
negate the effect of VG uncertainty; this does not mean that VG
has no capacity credit, but rather that it is slightly lower than what
it would be if uncertainty was not included in the calculation. It is
also not decreasing overall system reliability, but rather the contri-
bution to reliability is smaller than would otherwise be the case if it
was assumed to be perfectly predictable. The amount of additional
capacity required is technology and system dependent because of
difference in the impact of generation and system flexibility charac-
increase the limit of variable generation which can be integrated.
teristics (e.g., 100 MW of combustion turbines is more flexible than
The target user, in this case, is the high-level policy planning groups
100 MW of coal) [10].
with the aim of making the concept of flexibility accessible at an
early stage in planning. This approach, while it had some insight-
Monte Carlo Production Costing Simulation
ful qualitative outputs for the user, was limited in the quantitative
Southern California Edison (SCE) has developed a probabilistic
results it could achieve because of the limited data available.
modeling approach to estimate the need for additional capacity and
flexibility in the CAISO due to the variability in customer load, This high-level assessment of the flexibility of a power system uses
wind and solar generation, and fleet availability through sampling generic assumptions on the availability and capability of generat-
from a distribution of potential days. SCE’s approach was developed ing units to determine a system’s capability in best and worst case
in connection with the 2012 CPUC LTPP proceeding Track 2 to scenarios. The resulting feasible range of operation is an indication
determine whether the California system is sufficiently flexible and of whether a system is likely to face operational challenges [13].
to determine the effect of a variety of remedial actions. SCE used
a “stratified sampling” technique to generate a set of representative Transmission Deliverability Assessments
load, solar and wind days which are then modeled deterministically A transmission deliverability assessment is a critical piece of the pro-
using the PLEXOS simulation tool. Based on the simulated avail- cess (see the level 4 box in Figure 2). Most of the flexibility assess-
ability and behavior of generating units, the flexibility sufficiency of ments conducted so far have been limited to assessing the flexibility
the system is then determined. of generation assets only. Analysis tools and methods are currently
in development to assess the extent to which a system’s flexibility is
Other consultants and institutes have built on production cost sim-
impacted by its transmission network. EPRI is working with ECCO
ulations to calculate flexibility metrics, including expected unserved

9802731
Metrics for Quantifying Flexibility in Power System Planning 13 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

International to develop a first of its kind case study on transmis-

sion deliverability of flexibility for a large power system. To do this,
EPRI is adapting existing tools to consider the need for flexibility in
transmission reliability assessment. These methods will contribute
to more accurate flexibility assessments using the proposed metrics.

Actions for the Industry and for R&D

As power systems continue to experience a period of transition, the
planning assumptions and rules of thumb that have served industry
well over the last century are being revisited. In some cases, existing
practices will continue to fulfill system needs. However, as variabil- refinement of flexibility tools and metrics across all the relevant ar-
ity and uncertainty become more prominent features of generation eas. This will continue to be accomplished via new research projects,
and demand, flexibility assessment may be explicitly required to case study implementation and transfer of the methods and tools
ensure the reliability and cost effectiveness of a power system in the to the industry. EPRI encourages electricity providers, ISO/RTOs,
future. and other interested stakeholders to participate with EPRI in this
process.
As with the development of most technologies, with use over time,
the variety of the proposed flexibility metrics will be refined to best EPRI will continue to play a key coordination role on the issue
meet the needs of effective planning processes. System planners in of flexibility, strengthening our coordination with national labs,
areas where significant generation variability/uncertainty is expected academia, industry groups, and with regulatory and policy bod-
or where inflexible generation is being added to a system, will ies. Only through collaboration can these metrics and approaches
benefit from implementation and testing of the existing flexibility be refined and implemented to the benefit of both industry and
assessment methods and will provide feedback that will lead to consumer. By combining the experience within the industry with
needed refinements. collaborative R&D insights, the process of refining flexibility met-
rics can be expedited.
EPRI’s role is to support the industry (generators, utilities, ISO/
RTOs, researchers and others) in the continued development and
WHILE YOU’RE HERE
• What are the key flexibility questions your company
need answered?
• Which type of tools and metrics will help you
answer those questions?
• Are there skills or insights your team requires to
successfully assess flexibility needs?
• How will you consider the fact that both
transmission and generation resource can provide
flexibility?

9802731
Metrics for Quantifying Flexibility in Power System Planning 14 July 2014
 Insert Photo

Metrics for Quantifying Flexibility in Power System Planning

Acronyms
CPUC California Public Utility Commission LBNL Lawrence Berkeley National Laboratory
ELCC Equivalent load carrying capability LOLE Loss of load expectation
EPRI Electric Power Research Institute LTPP Long Term Procurement Planning
EUE Expected unserved energy NERC North American Electric Reliability Corporation
EUR Expected unserved ramping NREL National Renewable Energy Laboratory
FRACMOO Flexible Resource Adequacy Criteria and PFD Periods of flexibility deficit
Must-Offer Obligation PUC Public utility commission
GIVAR Grid Integration of Variable Renewables PV Photovoltaic
IEA International Energy Agency RTO Regional Transmission Organization
IRP Integrated resource planning SCE Southern California Edison
IRRE Insufficient ramp resource expectation VG Variable generation
ISO Independent System Operator

References
1. North American Electric Reliability Corporation, Flexibility http://sustainablebusinessoregon.com/articles/2012/03/oregon-
Requirements and Metric for Variable Generation: Implications for utilities-asked-to-plan-for.html?page=all.
System Planning Studies, August 2010, http://www.nerc.com/
8. California Energy Storage Alliance, Energy Storage in Cali-
files/IVGTF_Task_1_4_Final.pdf.
fornia’s Grid of the Future, September 2013, www.iepa.
2. EPRI, Electric Power System Flexibility: Challenges and Opportu- com/2013AnnualMeeting/IEPA_2013_Don_Liddell.ppt.
nities, to be published, 2014.
9. Con Edison, Distribution Load Relief Program Manual, Summer
3. Michael R. Blood, Associated Press, Low on Natural Gas, 2010, http://www.coned.com/energyefficiency/DLRP_Manual.
California Told to Power Down, press release, February 7, 2014, pdf.
http://abcnews.go.com/US/wireStory/natural-gas-shortage-
10. EPRI, Power System Flexibility Metrics Framework, Software
affects-calif-power-supply-22399432.
Tool and Case Study for Considering Power System Flexibility in
4. Western Governors’ Association, Meeting Renewable Energy Planning, 3002000331, http://www.epri.com/abstracts/Pages/
Targets in the West at Least Cost: The Integration Challenge, June ProductAbstract.aspx?ProductId=000000003002000331.
2012, http://www.westgov.org/reports/doc_download/1602-
11. Pacific Gas and Electric, Collaborative Review of Collaborative
meeting-renewable-energy-targets-in-the-west-at-least-cost-the-
Review of Planning Models, April 2014.
integration-challege.
12. Energy+Environmental Economics, “Renewable Energy Flex-
5. California ISO, Flexible Resource Adequacy Criteria and Must-
ibility (REFLEX) Model, http://ethree.com/public_projects/
Offer Obligation, February 2014, https://www.caiso.com/Docu-
reflex.php.
ments/DraftFinalProposal-FlexibleResourceAdequacyCriteria-
MustOfferObligation.pdf. 13. Mueller, Simon, Evaluation of Power System Flexibility Ad-
equacy: The Flexibility Assessment Tool (FAST2), 12th Interna-
6. California Public Utilities Commission, Long-Term Procure-
tional Workshop on Large-Scale Integration of Wind Power
ment Plan History, http://www.cpuc.ca.gov/PUC/energy/Pro-
into Power Systems as well as on Transmission Networks for
curement/LTPP/index_2012.htm.
Offshore Wind Power Plants, November, 2013.
7. Sustainable Business Oregon, Oregon utilities asked to plan for
flexible power resources to back up renewables, March 21, 2012,

9802731
Metrics for Quantifying Flexibility in Power System Planning 15 July 2014
 The Electric Power Research Institute, Inc. (EPRI, www.epri.com) conducts
research and development relating to the generation, delivery and use of
EPRI RESOURCES
electricity for the benefit of the public. An independent, nonprofit organi-
Aidan Tuohy, Technical Leader/Project Manager zation, EPRI brings together its scientists and engineers as well as experts
872.829.3395, [email protected] from academia and industry to help address challenges in electricity,
Eamonn Lannoye, Engineer II including reliability, efficiency, affordability, health, safety and the envi-
704.595.2794, [email protected] ronment. EPRI also provides technology, policy and economic analyses to
drive long-range research and development planning, and supports
Grid Operations, Grid Planning and Bulk System research in emerging technologies. EPRI’s members represent approxi-
Integration of Renewables mately 90 percent of the electricity generated and delivered in the United
States, and international participation extends to more than 30 countries.
EPRI’s principal offices and laboratories are located in Palo Alto, Calif.;
Charlotte, N.C.; Knoxville, Tenn.; and Lenox, Mass.

Together . . . Shaping the Future of Electricity

3002004243 July 2014

Electric Power Research Institute

3420 Hillview Avenue, Palo Alto, California 94304-1338 • PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774  • 650.855.2121 • [email protected] • www.epri.com
© 2014 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research Institute, EPRI, and TOGETHER . . . SHAPING THE FUTURE OF ELECTRICITY are
9802731
registered service marks of the Electric Power Research Institute, Inc.