Volume 19 • Number 2 • March/April 2021

magazine

Volume 19 • Number 2 • March/April 2021

www.ieee.org/power

©SHUTTERSTOCK.COM/VLADGRIN
©SHUTTERSTOCK.COM/PETR VACLAVEK

46
on the
cover
features
18 Operational Security 56 A Vision to Enhance Transmission
By Julia Matevosyan, Shun Hsien Huang, Security
Pengwei Du, Nitika Mago, and Rochie Guiyab By Evangelos Vrettos, Marc Hohmann,

contents
and Marek Zima
28 Energy Insecurity Due to Gas Supply
Availability 69 A Novel Approach to Transmission
By Guillermo Bautista Alderete Bottleneck Management in Japan
By Noritaka Ohashi, Kazuki Terashima,
37 The Impact of Renewables on and Makoto Tanaka
Operational Security
By Ivan M. Dudurych 79 The Fragile Grid
By Pierluigi Mancarella and Farhad Billimoria
46 Energy Security Through Demand-Side
Flexibility
By Jacob Østergaard, Charalampos Ziras,
Henrik W. Bindner, Jalal Kazempour,
Mattia Marinelli, Peter Markussen,
Signe Horn Rosted, and Jørgen S. Christensen

columns & departments

4 Editors’ Voice 91 Society News
10 Leader’s Corner 93 Book Reviews
16 Guest Editorial 96 Calendar
89 Awards 100 In My View

IEEE POWER & ENERGY SOCIETY (PES)

The IEEE Power & Energy Society is an organization of IEEE members whose principal interest is the advancement of the
science and practice of electric power generation, transmission, distribution, and utilization. All members of the IEEE are
eligible for membership in the Society. Mission Statement: To be the leading provider of scientific and engineering information on
electric power and energy for the betterment of society, and the preferred professional development source for our members.

Digital Object Identifier 10.1109/MPE.2020.3043609

march/april 2021 ieee power & energy magazine 1


Editor-in-Chief
Steve Widergren (Pacific Northwest National
Laboratory, USA), pem-eic@ieee.org
Immediate Past Editor-in-Chief
Michael Henderson (Power System Analysis and
Training, USA)
Associate Editors
John Paserba (Mitsubishi Electric Power Products,
Inc., USA), History
Antonio Conejo (The Ohio State University, USA),
Ning Lu (NC State University, USA), Barry
Mather (National Renewable Energy Labora-
tory, USA), Issues
Jianhui Wang (Southern Methodist University,
USA), Submissions
Assistant Editor
Susan O’Bryan (USA)
IEEE POWER & ENERGY SOCIETY (PES)
Governing Board
F. Lambert, President
J. Bian, President-Elect
M. Armstrong, Vice President, Chapters
V. Vittal, Vice President, Technical Activities
B. Enayati, Vice President, Education
B. Pal, Vice President, Publications
W. Bishop, Vice President, Meetings
J. Romero Aguero, Vice President, Membership
& Image
S. Bahramirad, Vice President, New Initiatives/
Outreach
N. Hadjsaid, Treasurer
J. Sykes, Secretary
S. Rahman, Past-President
D. Diaz, Region Rep., U.S. & Canada
J. Milanovic, Region Rep., Europe, Middle East,
& Africa
S. Leon, Region Rep., Latin America
D. Sharafi, Region Rep., Asia & Pacific
G. Gonzales, Member-at-Large
R. Kappagantu, Member-at-Large
A. Sannino, Member-at-Large
J. Yu, Member-at-Large
IEEE Division VII Director
M. Sanders
IEEE Division VII Director-Elect
C. Cañizares
PES Executive Director
Open
Standing Committee Chairs
Open, Constitution & Bylaws
M.T. Correia de Barros, Fellows
N. Hadjsaid, Finance
A. Apostolov, History
J. Bian, Long-Range Planning
S. Rahman, Nominations & Appointments
Region Representatives
D. Diaz, M. Gosalia, B. Gwyn, J. Khan,
I. Kuzle, S. Leon, J. Milanovic, M. Papic,
D. Sharafi, C. Wong, X. Wu
Chapter Committee Chairs
C. Diamond, Awards & Resources
Z. Wang, Chapters Website
N. Nair, Distinguished Lecturer Program
Z. Wang, Electronic Communications
T. Ribeiro de Alencar, Student Chapters Coordinator
Membership & Image Committee Chairs
A. Vaccaro, Awards & Recognition
S. Fattah, Humanitarian Activities
J. Benedict, Marketing
M. Seif, Membership Development
L.F. Gaitán, Social Media
J. Cardoso, Web Site Development
R. Li, Women in Power
S. Akar, Young Professionals
Digital Object Identifier 10.1109/MPE.2020.3043610
2 ieee power & energy magazine
Spanish Editorial Board
magazine Enrique Tejera (Panama Canal Authority, Panama),
Editor-in-Chief
Editors: M. Baquedano (GRA - University of
Florida Power Lab, Honduras), J. Cerda (University
of Michoacan, Mexico), G. Gonzalez (National
Editorial Board Secretariat of Energy, Panama)
L. Barroso (Comillas Pontifical, Brazil), J. Feltes
(Siemens Power Technologies International, USA), Advertising
N. Hatziargyriou (National Technical University of Erik Henson, Naylor Association Solutions
Athens, Greece), T. Hong (University of North Car- +1 352 333 3443, fax: +1 352 331 3525
olina at Charlotte, USA), B. Johnson (University ehenson@naylor.com
of Idaho, USA), B. Kroposki (National Renewable
Energy Laboratory, USA), D. Lew (Energy Sys-
tems Integration Group, USA), M. Miller (National IEEE Periodicals/Magazines Department
Grid, USA), D. Novosel (Quanta Technology, 445 Hoes Lane, Piscataway, NJ 08854 USA
USA), M. O’Malley (University College Dublin, Geraldine Krolin-Taylor, Senior Managing Editor
Ireland), N. Ochoa (University of Melbourne, Janet Dudar, Senior Art Director
Australia), A. Orths (Energinet, Denmark), Gail A. Schnitzer, Associate Art Director
C.E. Root (Velco, USA), H. Rudnick (Systep Theresa L. Smith, Production Coordinator
Ingenieria y Disenos Limitada, Chile), G.B. Sheblé Felicia Spagnoli, Advertising Production Manager
(EPPMT LLC, USA), J.C. Smith (ESIG, USA), Peter M. Tuohy, Production Director
E. Uzunovic (Worcester Polytechnic Institute, Kevin Lisankie, Editorial Services Director
USA), S.S. Venkata (Venkata Consulting Solutions Dawn M. Melley, Senior Director, IEEE Publishing
LLC, USA) Operations
IEEE Trans. on Energy Conversion, A. Tessarolo
Technical Council IEEE Trans. on Power Delivery, F. DeLeon
V. Vittal, Chair, H. Chen, Vice Chair IEEE Trans. on Power Systems, N. Hatziargyriou
D. Watkins, Secretary, F. Rahmatian, Past-Chair IEEE Trans. on Smart Grid, C. Canizares
IEEE Trans. on Sustainable Energy, B. Chowdhury
Technical Committee Chairs IEEE Open Access Journal of Power and Energy, F. Li
K. Schneider, Analytical Methods for IEEE Power & Energy Magazine, S. Widergren
Power Systems eNewsletter, S. Fattah
J. Yagielski, Electric Machinery Editor-in-Chief at Large, W. Xu
J. Yale, Energy Development & Power Generation Marketing, E. Batzelis
B. Chalamala, Energy Storage & Stationary Battery PES Representative to IEEE Press, K. Bhattacharya
H. Geene, Insulated Conductors Website, R. Rana
D. Harmon, Nuclear Power Engineering
K. Fodero, Power System Communications Meetings
& Cybersecurity Committee Chairs
L. Lima, Power System Dynamic Performance General Meeting Steering, W. Cassel
E. Hanique, Power System Instrumentation Joint Technical Committee Meeting Steering,
& Measurements S. Ward
F. Li, Power System Operation Planning Technically Cosponsored Conferences Steering,
& Economics A. Borghetti
M. Yalla, Power System Relaying & Control Transmission & Distribution Conference &
R. Melton, Smart Building, Load & Exposition Committee for North America
Customer Systems Steering, C. Segneri
P. Fitzgerald, Substations Innovative Smart Grid Technologies North America,
S. Hensley, Surge Protective Devices J. Romero Aguero
K. Flowers, Switchgear Innovative Smart Grid Technology Conference–
B. Forsyth, Transformers Europe, D. Van Hertem
S. Santoso, Transmission & Distribution Innovative Smart Grid Technology Conference–
Asia, D. Sharafi
Technical Council Coordinating Committees PowerAfrica Steering Committee, H. Louie
H. Sun, Energy Internet Website, Open
D. Houseman, Intelligent Grid & Emerging
Technology Education
D. Alexander, Marine Systems Committee Chairs
A. Leon, Renewable Systems Integration Power & Energy Education, A. Srivastava
PES Scholarship Plus, J. Hoffman, C. Root
Technical Council Standing Committees PES University, L. Hennebury
F. Rahamatian, Awards Website, Open
Open, Industry Education
D. Watkins, Organization & Procedures New Initiatives and Outreach
T. Burse, Standards Coordination Committee Chairs
H. Chen, Technical Sessions Executive Advisory Council, G. van Welie
Open, Webmaster IEEE Smart Cities, D. Novosel
IEEE Smart Grid, P. Wung
Publications IEEE Smart Village, J. Nelson
Publications Board Chair, B. Pal Industry Tech Support Leadership, D. Novosel
Editors-in-Chief Media Engagement, D. Kushner
IEEE Electrification Magazine, L. Fan Website, K. Anastasopoulos
IEEE Power Engineering Letters, M. Fotuhi-
IEEE prohibits discrimination, harassment, and bullying. For more infor-
Firuzabad mation, visit http://www.ieee.org/web/aboutus/whatis/policies/p9-26.html.
IEEE Power & Energy Magazine
IEEE Power & Energy Magazine (ISSN 1540-7977) (IPEMCF) is published bimonthly by the Institute of Electrical and
Electronics Engineers, Inc. Headquarters: 3 Park Avenue, 17th Floor, New York, NY 10016-5997 USA. Responsibility for  the
contents rests upon the authors and not upon the IEEE, the Society, or its members. IEEE Operations Center (for orders, sub-
scriptions, address changes): 445 Hoes Lane, Piscataway, NJ 08854 USA. Telephone: +1 732 981 0060, +1 800 678 4333.
Individual copies: IEEE members US$20.00 (first copy only), nonmembers US$89 per copy. Subscription Rates: Society members
included with membership dues. Subscription rates available upon request. Copyright and reprint permissions: Abstracting is permit-
ted with credit to the source. Libraries are permitted to photocopy beyond the limits of U.S. Copyright law for the private use of
patrons 1) those post-1977 articles that carry a code at the bottom of the first page, provided the per-copy fee indicated in the code
is paid through the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA; 2) pre-1978 articles without fee.
For other copying, reprint, or republication permission, write Copyrights and Permissions Department, IEEE Operations Center, 445
Hoes Lane, Piscataway, NJ 08854 USA. Copyright © 2021 by the Institute of Electrical and Electronics Engineers, Inc. All rights
reserved. Periodicals postage paid at New York, NY, and at additional mailing offices. Postmaster: Send address changes to IEEE
Power & Energy Magazine, IEEE Operations Center, 445 Hoes Lane, Piscataway, NJ 08854 USA. Canadian GST #125634188
Printed in U.S.A.
Promoting Sustainable Forestry
march/april 2021
SFI-01681
 OPTIMISE AND EVALUATE THE DISPATCH
OF YOUR SYSTEM WITH POWERFACTORY!
UNIT COMMITMENT AND DISPATCH OPTIMISATION
 Time sweep optimisation
 Consideration of contingencies (n-k)
 Storage schedule optimisation
 Full consideration of network constraints, exchange flows and generation unit limitations
such as ramp rates, up- and downtimes
 Support of contingency constraint optimisation and consideration of planned outages

OPTIMAL POWER FLOW

 Linearised DC optimisation (including contingency constraint/optimisation)
 AC load flow optimisation with interior point method
 Active power and reactive power and voltage profile optimisation
 Various controls such as: Generator active and reactive power, transformer, quad booster
and shunt taps, static var systems

PowerFactory provides flexible results analysis and visualisation with dedicated reports,
colouring modes and diagrams.

For more information visit POWER SYSTEM SOLUTIONS

www.digsilent.de/powerfactory
MADE IN GERMANY
Serving more than 160 countries.
 editors’ voice
A
AS THE WORLD FACED THE
g r i p s of t h e C OV I D -19 p a n -
demic, the mere idea of potential
lockdowns sent packages of toi-
let paper flying off the shelves.
That has been followed by the
loss of jobs and income that stag-
gered economies globally. Fortu-
nately, power and energy system
operation has remained largely
stable. But imagine the further
hoarding and pandemonium that
would ensue should the electrified
world’s societies confront short-
ages in access to the energy on
which they have become so dependent.
Think of the economic collapse that
would transform the orderly operation
of businesses and governments. Anxi-
ety at the thought of the reactions to
such situations does more than lead to
lack of sleep. Many of us in this busi-
ness have jobs that directly consider
the short- and long-term availability of
uninterrupted energy access. Most of
the rest of us work to see that energy is
delivered safely at an affordable price.
These two aspects form the foundation
for exploring energy security.
A host of global, regional, and local
issues inspires socioeconomical trends
and policies that drive changes in the en-
ergy supply chain. Our energy systems
are evolving in response to these trends
and policies as spurred on by seemingly
faster leaps in technology-based capabili-
ties. The responses to climate change, and
the factors that may encourage it, have
given rise to advancement in renewable
resources being produced at cost-effective
Digital Object Identifier 10.1109/MPE.2020.3043566
Date of current version: 19 February 2021
4 ieee power & energy magazine
Barry Mather and Steve Widergren
energy security
ensuring affordability & availability
©SHUTTERSTOCK.COM/KATYR
rates. This leads to an associated reduc-
tion in more traditional fossil supply sys-
tems characterized by an abundance of
spinning mass to which we have applied
decades of engineering advancement in
design and paradigms for system op-
eration. Innovation in batteries, sensing,
communications, and information tech-
nology gives rise to economic drivers for
cooperative operating arrangements that
harness greater efficiencies and offer fu-
ture promises of robustness and resilience
in response to system operation pertur-
bations. But this new type of system is
not so familiar. Now, as in the past, our
industry is responding to change to better
understand the complex interplay of these
emerging shifts so that we may attain
some level of proficiency in managing
these evolving power and energy systems.
May our success in this endeavor afford
us a moment to rest as we summon new
strength for the next set of energy security
challenges that are sure to follow.
In This Issue
The articles in this issue explore the op-
erational security of the world’s power
systems in the face of change and as-
sociated system challenges as seen
by the higher penetration levels of re-
newable energy generation, increased
adoption of new grid technologies,
and market developments for grid
services to address an expanding suite
of operational issues. Guest Editors
Antonio Conejo and Federico Milano
have assembled an international
contingent of experts in various spe-
cialized areas of operational security
to provide a truly comprehensive per-
spective of the wide range of chal-
lenges and solutions evolving in this
space. In total, the seven feature articles
include perspectives from six countries
and seven separate power systems. The
feature articles discuss the following:
✔ the operational challenges being
managed in Texas, given sus-
tained growth in renewable en-
ergy generation capacity
✔ an in-depth analysis of the inter-
dependency of California’s natu-
ral gas and electric power system
to ensure the secure operation of
the electric power system as it is
modernized
✔ the challenges to and impacts on
power system short-term stabil-
ity being experienced and ad-
dressed in Ireland as higher lev-
els of renewable energy operate
within the system
✔ a view of how the Danish power
system can meet operational se-
curity and decarbonization goals
harmoniously
✔ the operational security issues real-
ized in the strongly interconnected
Swiss power system as the power
systems of both Switzerland
march/april 2021
 Ready
To
ARC FLASH EV
EVALUATION
SwiTch? Design safer power systems while ensuring compliance with NEC,
OSHA, NFPA, IEEE, NESC, and CSA standards

Receive up to 60% off Fully integrated with Short Circuit, Over-Current Coordination,
and Equipment Evaluation modules
when you switch Evaluate alternatives quickly and easily to establish an optimal design
to SKM. Improve safety margins with user-definable arcing fault tolerance
Save time by automatically generating arc flash labels and work permits
You’ll also receive Determine required PPE levels instantly
free service to convert Increase flexibility by creating custom labels in any size with
competing software user-defined logos, text, comments, and field placement.
Avoid potential fines, post productivity, and increased insurance
projects over to SKM! and litigation costs.
Compare results from multiple project scenarios in a single table

ADDITIONAL SOFTWARE FROM SKM

“The main advantage with
Power*Tools for Windows is
its user-friendliness. PTW is
easy to use and the ready-
to-use standard models
enable you to speed up
your work. Changing data is
simple and makes evaluating
DAPPER® Integrated Electrical Analysis Software. Comprehensive
Three Phase Short-Circuit Studies, Load Flow Study, Demand Load
Study, Feeder and Transformer Sizing Study, Impact Motor Starting
Study, and Load Schedules. (Unbalanced Studies also available)
CAPTOR® Time-Overcurrent Coordination. Graphical Time-Overcurrent
Coordination. Integrated with one-lines, short-circuit modules, Equipment
Evaluation, and Arc Flash. Comprehensive protective device library.
AUTO COORDINATION & EVALUATION Instantly identify
coordination and protection issues by evaluating each
equipment for adherence to the National Electric Code (NEC).
Automatically make changes to resolve violations.
CABLE AMPACITY Accurate calculation of ampacity and
temperature rise in various power cable installations. Based
on and fully compliant with IEC 60287.
A_FAULT ANSI Short-Circuit Study. Three Phase Short-Circuit Study
AUTO DESIGN Automatically sizes Cables, Transformers,
Buses, and Protective Devices to achieve proper size and
ratings that meets criteria limits set by the Equipment Evaluation
study. Utilize user-defined Design libraries. Best suited for new system
designs or to resolve undersized and/or overdutied equipment for
existing systems. (Requires Equipment Evaluation module)
I*SIM Dynamic Simulation and Transient Stability. Model, Simulate,
and Analyze Dynamic Responses to Power System Electro-Mechanical
Disturbances. Includes Generator Sizing and Stability, Rotor Angle and
Voltage Stability, Flux Level Machine Representation, Identify Fault
Clearing Times. User Defined Graphical Models for Exciter, Turbine
Governor, PSS, and other controllers.
DISTRIBUTION RELIABILITY Calculates the reliability indices of
individual load points and the overall distribution systems with either
radial or loop configuration.

different cases very easy.”
based on the ANSI/IEEE C37 Standards. Separate solutions for low,
medium and high voltage systems and for symmetrical, momentary
and interrupting calculations. (Unbalanced Studies also available)
EQUIPMENT EVALUATION Equipment Evaluation Report. Automatically
compares short-circuit ratings, withstand ratings. Applies de-rating
adjustments and user defined pass/marginal/failed criteria. Includes
error checking for input data and topology.
TMS Transient Motor Starting Simulation. Time-based motor starting
simulation with graphical output. Includes reduced voltage and
capacitor starting, graphical motor and load models.
HI_WAVE Power Quality Investigation and Filter Design. Frequency
Scan, Harmonic Current, Voltage Distortion, Harmonic Load Flow and
Interactive Filter Design. Evaluate and correct harmonic distortions.
DC SYSTEMS ANALYSIS Battery Sizing, Load Flow, & Short Circuit
Analysis. Evaluate all loading conditions for DC duty cycle loads and
AC emergency loads. Complies with IEEE std. 485, 1115, 399, 946,
and IEC std. 61660.
GROUND MAT Substation Ground Grid Design and Analysis. Optimizes
grid design using general purpose finite element algorithm for
potential analysis and graphical facilities to validate grounding
systems efficiency.
Data Exchange Seamlessly share project data between
SKM projects and other third-party software. Exchange
format includes Tab Delimited, CSV, and SKM XML. Also includes
an interface to exchange data with Autodesk Revit and Intergraph
SmartPlant Electrical.

SKM Power*Tools Training Classes

Visit skm.com to find a class near you today! 101/102 Introduction to SKM Power*Tools
Hurry, classes fill up fast! 103/104 Intermediate Topics in SKM Power*Tools
201/202 Power System Application – Coordination & Arc Flash
Training more than five engineers? 204/205 Motor Starting and Harmonics
One Pearl Street
Email training@skm.com for a 206 DC System Analysis & Arc Flash
Redondo Beach, CA 90277
personalized On-Site Training quote. 207 Grounding and Applications
1-800-500-4SKM
www.skm.com • sales@skm.com Chosen By The Top 40 eleCTriCal engineering Firms in The World
 and its neighbors evolve, includ-
ing the need for new adaptive
control mechanisms
✔ a detailed account of the nation-
al-level implementation of ad-
vanced protective relaying meth-
ods to enable higher levels of
renewable generation while co-
ordinating generation response
to contingencies in Japan
✔ a comprehensive look at the oper-
ational security of the Australian
power system, focusing on how
system security can be achieved
by balancing technology and
market solutions, which can both
serve to improve or diminish
overall system security.
In addition, the “In My View” column
discusses some of the difficult issues re-
lated to power system security, such as
how security is defined, its relation to op-
erational risk, what will change and what
will stay the same as grids evolve, and
how societies can achieve power system
security given all of its many facets.
Leader’s Corner
The Asia and Pacific area—Region 10
of the IEEE Power & Energy Society
(PES)—encompasses the most populat-
ed areas of the world as reflected in the
size and growth of its PES membership.
This region is represented on the PES
governing board by Dean Sharafi, who
joins us in this issue to provide a glimpse
of the extent of diversity in activities,
events, and initiatives covered across
this varied landscape. The influence and
impacts on our profession are sure to
increase from Region 10, as we see re-
ported in contributions to this magazine.
Book Reviews
We take a break from the “History” col-
umn this month to bring you two book
reviews. Mohammad Shahidehpour
provides a summary of and insight into
Infrastructure Asset Management with
Power System Applications by Lina
Bertling Tjernberg. Given the vast out-
lay of equipment needed to supply and
deliver energy through the electric pow-
er system, the lifecycle of its acquisition,
maintenance, and retirement is complex
and offers rich opportunities for analysis
methodologies and tools to improve as-
set management. The expanded deploy-
ment of smart technology in sensing
and control is contributing to changes
in the asset mix. Smart technology
also provides sources for more data
from which to extract information on
the status and predictions of health for
better diagnostics and efficient mainte-
nance strategies.
Our second book review concerns
the growing asset area associated
with renewable energy. Gerry Sheblé
evaluates the educational textbook
Renewable Energy Technologies and
Resources by Nader Anani. This intro-
duction to renewable energy resourc-
es, with equipment explanations and
worked examples on how to model and
analyze them, offers a valuable source
of information as practical concerns
such as power system interconnection
become commonplace.

World Leader in Developing

BIGWOOD SYSTEMS, INC. Powerful and Field
Field-Proven
www.bigwood-systems.com
+1.607.257.0915 On--line and Off
On Off--line Tools
sales@bigwood-systems.com for Power Grid Modeling, Analysis, Assessment,
NY, USA
Head Office Branch Offices Enhancement and Optimization
Ithaca, NY Taipei, TWN | Beijing, CHN
Innovation prevails!

Network Applications for the Control Center

“Bigwood Systems’ Online Voltage
Stability product is simply the best in the
field.” - CAISO Project Leader
“BSI On-line Transient Stability
Assessment is a superb product.”
- TEPCO Manager
“The BSI VAR Management System is a
best practice tool for all utilities.”
- SERC Audit Staff
• Real-time and Look-ahead Static and Dynamic Security
Assessment (patented screening, ranking, and detailed
analysis)
• Reactive Power Reserve Monitoring, Management &
Enhancement
• Actionable Control Recommendation engine to
eliminate violations and mitigate unstable contingencies
• New-Generation State Estimation for Transmission and
Distribution Networks with a low number of measurements
using SCADA & PMU data
• Global Optimal Power Flow (for over 250,000 control variables with
AC/DC constraints)
• Distribution Network Management Tools
• Renewable Energy Forecasting, Uncertainty Modeling,
Representation & Management
• Advanced Measurement-Based Power System Modeling
(Generation, Transmission & Load)
World-Wide Customer Base
Over 40 electrical utilities served world-wide, including long-standing
customers such as California ISO, PJM Interconnection, Tokyo Electric
Power Company, Tennessee Valley Authority, and UK Power Networks,
among others. BSI’s advanced technology, innovation and satisfied
customers are key differentiators.
Society News and Awards
The “Awards” column recognizes the
PES members elevated to the class of
2021 IEEE Fellows. We list all 31 new
Fellows who are members of PES or
were elevated by PES together with a
statement of each member’s contribu-
tion. IEEE Fellow is our highest class
of membership, offered to only a select
few. Hats off!
In other news, the 10th IEEE PES
Innovative Smart Grid Technologies Eu-
rope Conference was held in an online,
virtual setting on 26–28 October 2020, as
organized and hosted by the Delft Uni-
versity of Technology, The Netherlands.
European attention on climate change-
mitigating initiatives set the tone for this
conference. Thanks in part to its virtual
nature, nearly one-third of the online at-
tendees came from outside Europe.
Wrap Up
While we have acknowledged the im-
pacts of the COVID-19 pandemic in edi-
torial columns and reports of conferenc-
es in this magazine, this is the first issue
in which articles examining its impacts
on power and energy systems and the
implications for systems operations are
beginning to appear. It will not be the
last. As the rollout of vaccines to com-
bat the novel coronavirus expands, there
is light on the horizon. But the effects
of the pandemic on human life and our
economies will be concerns that we will
be confronting for months and years to
come. We look forward to bringing you
more information and perspectives as
clarity emerges regarding these impacts
and their effects on power and energy
systems operations and planning.
The successful production of this
issue of IEEE Power & Energy Maga-
zine rests on the contributions of many,
starting with Guest Editors Antonio
Conejo and Federico Milano, who
developed the theme for the issue, re-
cruited a stellar crew of authors, and
managed the creation of the technical
articles from these authors through the
editorial process. Our sincere apprecia-
tion goes out to them and the team that
they assembled. Recognition is also due
to Associate Editor Barry Mather, who
led the technical review of the articles
and contributed to this column, as well
as to Assistant Editor Susan O’Bryan,
who provided language editing support
and managed our editorial process and
its handoff to the professional staff at
IEEE Publishing. The remaining forces
for continued excellence that drive this
magazine forward include Associate
Editors John Paserba, Antonio Conejo,
Ning Lu, and Jianhui Wang, who pro-
vided counsel and contributed to up-
coming issues in the production cycle
as this one goes to press. Finally, credit
is due to past Editor-in-Chief Mike
Henderson, who continues to calmly
answer the phone when called about all
matters of concern that this new editor-
in-chief stumbles to address.
p&e
 Leaders in power
systems prevent
and survive cyber
attacks with the
RTDS® Simulator

IT’S NOT JUST WEATHER OR WILDLIFE. PREPARE FOR CYBER

INCIDENTS THAT CAN DISRUPT SERVICE AND DAMAGE EQUIPMENT.

Worldwide, the RTDS Simulator is a crucial component of cybersecurity

testbeds, offering users the safest, most controllable way to simulate attacks
on the power system.
Connect a real-time model of the grid to a communication network simulator
and quantify the impacts of a cyber attack. Anticipate the behaviour of
real protection, control, and measurement equipment through closed-loop
testing. Results are authentic. Service is uninterrupted. This provides you
with a realistic, flexible, and contained environment for the validation of
energy system security technologies.
Learn more about cybersecurity project opportunities at rtds.com/applications/cybersecurity

YOUR WORLD IN REAL TIME.

RTDS.COM
 leader’s corner
A
A FEW YEARS AGO, WHEN THE
late Patrick Ryan was in Australia dur-
ing the 2011 IEEE Power & Energy
Society (PES) Innovative Smart Grid
Technologies Asia (ISGT Asia 2011)
Conference, we had a conversation on
the final night of the event. We were
on a ferry ride along with conference
delegates, enjoying the evening lights
and sounds of Perth on the Swan River.
Ryan asked my opinion about some-
thing that was just an idea at the time
but eventually became an important
PES program, namely, the PES Schol-
arship Plus Initiative.
I remember that the conversation
turned to volunteering and how im-
portant volunteers are in achieving
any objective in an organization such
as PES. Ryan said everything starts
and ends with volunteers! His words
are still ringing in my ears. The PES
Scholarship Plus Initiative now sup-
ports many students in North America
pursuing their education in the electri-
cal power industry and provides intern-
ship opportunities for them in power
companies. I expressed my support for
the idea, thinking about how this ini-
tiative could benefit many countries in
Asia-Pacific. Now, in addition to North
America, the Scholarship Plus Initia-
tive also supports students in Italy; for
the last few years, with the same pas-
sionate approach that our PES leaders
have shown in India, this program is
in its fourth year of implementation
Digital Object Identifier 10.1109/MPE.2020.3043649
Date of current version: 19 February 2021
10 ieee power & energy magazine
Region 10
its past, present, & future
there. I hope that we can have this pro-
gram in other countries in Asia-Pacif-
ic, as there has been great interest
expressed, especially from our PES
Chapter in Indonesia.
I started with a story of gratitude
for Ryan to signify the importance of
volunteering. I now want to give read-
ers of IEEE Power & Energy Maga-
zine a quick snapshot of Region 10.
Asia-Pacific is one of IEEE’s Regions
with a strong PES presence. The region
has the largest population among all
of IEEE’s Regions and the strongest
membership of the PES. We now have
46 PES Chapters, 179 Student Chap-
ters, and 11,185 members with a steady
membership growth trend. Asia-Pa-
cific incorporates a vast geographical
area, from Pakistan to New Zealand,
with several time zones. When our
members are preparing their dinner in
Auckland, our members in Islamabad
are about to start their workday. This
wide range of time zones resembles the
diversity of culture, education, public
infrastructure, engineering, industry,
and power system perspectives in the
numerous countries of the region.
Historically, IEEE has had a sig-
nificant presence in Asia-Pacific. The
formation of Region 10 was approved
by the IEEE Board of Directors in
1966; however, the Institute of Radio
Engineers had a section that was created
in Asia-Pacific before the formation of
Region 10, namely, Japan (Tokyo) start-
ing in 1955. New Zealand and Pakistan
followed in 1968, with India forming in
1969. Initially, Region 10 covered other
Dean Sharafi
areas, including a large part of Africa,
but in 1981, those areas of Africa be-
came a part of Region 8.
It has been a great pleasure for me
to represent this wide and enduring re-
gion before the PES Governing Board
since 2017. I had been involved with
PES for many years before this ap-
pointment, and my focus in this role
has been on addressing the question of
what can be done to serve PES mem-
bers in this new capacity.
In PES, Chapters are at the core of
serving members and the wider PES
community. An active Chapter is an
enabling element in offering programs,
activities, professional networking
opportunities, special events, work-
shops, webinars, and conferences to
support local members in achieving
their personal and professional goals.
My focus has been on supporting ex-
isting PES Chapters and developing
future ones in Region 10. Because of
the strong leadership in our Chapters,
we have achieved continued growth in
PES membership. Our Chapter leaders
support local members and demon-
strate the benefits of PES membership
through high-quality technical and
professional activities.
PES is growing in Asia-Pacific with
new Chapters in the Philippines and
in several cities in China. Some new
Chapters in China are in the process
of forming. During the last two years,
PES has also formed the PES China
and PES India Councils.
PES councils facilitate the activities
of members and coordinate the work of
march/april 2021
• •
• •
• •
 PES Chapters in a country or region.
China and India qualified for such
councils as they have vast geographical
coverage with numerous Chapters and
specific technical interests. The coun-
cils manage the interactions between
PES Chapters and local industry as well
as the relationship between PES and lo-
cal engineering organizations. These
councils provide leadership to Chapters
through training workshops, tutorials,
webinars, and so on to make PES vis-
ible to the engineering sector and grow
its membership. They have also coordi-
nated technical and standards-related
activities to support our industry and
provide solutions to specific challenges
facing the industry in their respective
countries. Both the PES China and PES
India Councils have proved their value
in the two years of their operation.
The PES China Council has coor-
dinated PES Chapter activities and
cooperated with local engineering as-
sociations, such as the Chinese Society
of Electrical Engineers. The council
has established numerous committees,
such as the IEEE PES Energy Storage
and Stationary Battery Satellite Com-
mittee, the IEEE PES Technical Satel-
lite Council, and the Power and Energy
Cloud Forum.
The PES India Council organized the
PES Global Workshop and welcomed
Bangalore Electricity Supply Company
Limited and Schneider Electric as PES
corporate members. The council also
set up a team of mentors to support the
growing student membership as well as
a team of industry professionals to help
academic institutions deliver industry-
driven talks and workshops.
Recently, strong interest has been
expressed by the Association of South-
east Asian Nations in establishing a
PES council, and I am working with
Chapter leaders to determine how we
can support them in that objective.
The Asia-Pacific region has also
hosted numerous conferences and work-
shops in the last decade, including ISGT
Asia, the Asia-Pacific Power and Energy
Engineering Conference, and the inau-
gural IEEE PES Generation, Transmis-
sion, and Distribution Conference and
Exposition Asia 2019 held in Bangkok,
Thailand. This PES event, the largest
outside of North America, had more
than 11,000 participants, accommodated
274 exhibition booths from 25 countries,
and enjoyed the expertise of more than
300 authors and speakers from various
countries around the world.

Modeling of renewables and their controls in DSAToolsTM

Dynamic Security Assessment Software
Offering a simple, user-friendly interface
with extensive analysis options, DSATools TM
is a suite of software tools for power system
analysis. The software is designed for
applications in both off-line studies and
on-line dynamic security assessment.
www.dsatools.com
• Supporting latest generic models for dynamic simulations
• Block-diagram based approach to develop user-defined models and
templates using the UDM Editor application
• Ability to link to user-written C/C++ code (DLL)
• White-box (open) and black-box (closed) model development
Experience and support
• Provider of hybrid simulation platforms to perform TSA-EMT hybrid
studies with TSAT & PSCADTM or TSAT & RTDSTM
• Engineers highly experienced in developing dynamic models for
transient stability studies and small signal stability studies
• Consulting services for custom model development and generic
model validation/tuning

TSAT SSAT

Contact us:
dsainfo@powertechlabs.com

12 ieee power & energy magazine march/april 2021

Mag W h e re i n d ) Days?
ne t W
ire Ma i n the World Ca n I F n (7
de to Order n Se v e
and Shipped Withi

Craig Wire Products ~ Frequently Asked Questions

• When was Craig Wire started? November 2007
• Who founded Craig Wire? Jack Craig, Brad Archambeau, Linda Belcher, Brett Portwood
• Is Craig Wire a joint venture? No. CWP is a privately held independent LLC.
• What is the breadth of CWP’s product line? Please visit our website www.craigwire.com for a full review
of our product offering.
• What is the minimum order quantity? For 99% of our product line the minimum is 250 pounds.
We will readily take orders for 1,000 pounds or more.
• What is the delivery time? Our standard shipment time after receipt of order is 14 days.
In an emergency CWP can ship within 7 days.

Call Craig Wire

• Customer Service: 770.920.2222
• Sarah Berkowitz: 770.558.7301
• Brad Archambeau: 770.500.7199
• Jason Craig: 312.799.0894

Craig Wire Products

1055 Shadix Industrial Way • Douglasville, Georgia 30134
www.craigwire.com
 New conferences with innovative
themes have also been initiated from
the region with great potential to be-
come global events, such as Energy
Internet and Energy System Integra-
tion (EI2), which focuses on address-
ing emerging challenges and utilizing
the opportunities of grid modernization
through advanced technologies and
the Internet for the integration of new
types of energy resources. EI2 has also
been a great platform for cooperation
between PES; the Chinese Society of
Electrical Engineers; and other Chi-
nese associations, universities, and
research institutes. PES members in
Asia-Pacific are staying ahead of in-
dustry challenges and facilitating the
cooperation of industry and academia
to serve the PES community and re-
solve problems that are important to the
electric power sector.
And importantly, these days we
cannot discuss anything without men-
tioning the impact of COVID-19. Dur-
ing the coronavirus pandemic, many
conferences and events went virtual,
and we lost face-to-face contact and
networking opportunities. The silver
lining was that these events became
more widely accessible. For example,
in 2020, Chapter chair training benefit-
ted from a much larger audience com-
pared to previous meetings, at no cost
to PES.
I look forward to working with
Chapter leaders and my colleagues on
the PES Governing Board to serve the
Asia-Pacific PES community during
the next year that I am in this role. I am
sure that it will be a successful period
for PES and our members, and hope-
fully a better world with no concern for
the pandemic.
Yuval Noah Harari, the historian
and philosopher of our times, believes
it was only because of the specific abil-
ity of Homo sapiens to be grouped and
organized in large numbers that human
beings could dominate the world and
achieve what other species were inca-
pable of doing. This proves the impor-
tance of being part of an association
that pursues specific objectives. I see
no other cause more valued than what
PES has set as its objective: namely, to
undertake our activities for the better-
ment of society. Our role as PES lead-
ers and members is to demonstrate this
value proposition and serve our Society
in the best possible way, leveraging the
latest scientific, technological, and en-
gineering achievements.
p&e
Be the force behind change
Bring the promise of technology — and the knowledge and power to leverage it,
to people around the globe. Donate now to the IEEE Foundation and make a
positive impact on humanity.

· Inspire technology education

· Enable innovative solutions for social impact
· Preserve the heritage of technology
· Recognize engineering excellence

Discover how you can do a world of good today.

Learn more about the IEEE Foundation at ieeefoundation.org.
To make a donation now, go to ieeefoundation.org/donate.
 guest editorial
A
AS THE POWER INDUSTRY MOVES
from a capacity-limited to an energy-
constrained generation mix, diverse situ-
ations impact energy security. Environ-
mental considerations may prevent the
deployment of components that increase
the energy security level, e.g., nuclear
or thermal units, or they may require the
retirement of such components. Limited
transmission capacity, outdated operation-
al procedures, and/or limited storage avail-
ability may also significantly decrease the
operational security level. The interaction
of different energy systems is yet another
emerging issue. For example, natural gas
unavailability due to issues with natural
gas sources or pipelines affects the opera-
tional capability of power systems with a
significant share of renewable generation.
There exist, of course, new technolo-
gies and tools that, if properly harnessed,
may help maintain or even increase en-
ergy security. In this vein, demand flex-
ibility has great potential. The redesign
of operational procedures involving a
comprehensive use of data processing,
forecasting, and optimization also ap-
pears as a viable and effective solution.
Models that accurately represent the
physical laws (ac power flow equations)
are increasingly important to make in-
formed operational decisions. This is-
sue of IEEE Power & Energy Magazine
provides an overview of energy security
problems and remedial actions in power
systems in different regions of the world.
The Article Lineup
This issue contains seven articles and
one column. The first two articles ad-
dress energy security issues on the
Digital Object Identifier 10.1109/MPE.2020.3043650
Date of current version: 19 February 2021
16 ieee power & energy magazine
Antonio J. Conejo and Federico Milano
energy insecurity
problems & remedial actions
“electrical island” of Texas and in
California, where wind and solar pro-
duction, respectively, are the most rel-
evant concerns. The next three articles
describe energy security approaches in
Europe, including the very distinct cases
of Ireland, Denmark, and Switzerland.
The final two articles pertain to security
issues in Japan and Australia, where in-
sularity and a continental-wide service
area, respectively, call for tailored ap-
proaches. The issue concludes with the
“In My View” column, which provides
an insightful formal analysis of energy
security by Göran Andersson.
In the first article, Julia Matevosyan
and her coauthors provide a compre-
hensive description of the operational
security framework at the Electric Re-
liability Council of Texas. It includes
background, a review of issues about
high inverter-based resource penetra-
tion, market-based solutions that com-
prise an array of ancillary services,
improved forecasting, and different sit-
uational awareness tools to ensure secu-
rity from day ahead to energy delivery.
In our second article, Guillermo
Bautista Alderete introduces the California
power system to the reader. He analyzes
energy insecurity due to the coupling of
the natural gas and power systems as
a result of significant electricity production
using combined-cycle gas turbines and the
high penetration of weather-dependent
renewable sources. The article focuses on
the issues that have surfaced in southern
California as a result of the Aliso Can-
yon natural gas storage leak in 2015. The
remedial and preventive actions taken to
ensure electric energy supply security
are described. These actions are related
mostly to the coordination of the operation
of natural gas and power systems.
In the third article, Ivan Dudurych
presents the all-island Irish power sys-
tem and focuses on the impact of a high
share of wind generation on the tran-
sient, voltage, and frequency stability
of the grid. The article provides a set of
definitions and best practices adopted
by the transmission system operator
EirGrid Group to cope with the peculiar
features of the Irish grid. On one hand,
to reduce emissions, the operators allow
up to 65% of instantaneous wind gen-
eration. On the other hand, the system is
islanded and connected asynchronously
to the British system. This interconnec-
tion can supply up to 10% of the Irish
peak demand but can trip unexpectedly
and very quickly and cause fast and
large variations of frequency. The article
concludes with a discussion of the rel-
evance for monitoring of system secu-
rity metrics, such as the rate of change
of frequency, system inertia, and system
nonsynchronous penetration level.
The next article, by Jacob Østergaard
and his coauthors, provides a comprehen-
sive overview of the Danish power system
and the ambitious targets of the Danish
government regarding decarbonization.
The article explains how such targets can
be achieved through renewable generation
and load flexibility. The latter appears as
the most promising way to keep an accept-
able level of system security in a scenario
where a large quota of the generation mix
is based on weather-dependent sources. A
review of ongoing Danish research proj-
ects suggests that there are many chal-
lenges but also many possible solutions.
Key to the process toward secure decar-
bonization is the role of distribution sys-
tem operators, a multilayered landscape
of energy markets and services, and the
emerging concept of energy communities.
march/april 2021
 In the issue’s fifth article, Marek
Zima and his coauthors give a comprehen-
sive overview of the Swiss power system,
emphasizing the characteristics that make
it unique in the context of energy security.
Unlike the other systems discussed in
this issue, the transmission network of
Switzerland is strongly meshed and highly
interconnected to the grids of neighbor-
ing countries. A variety of real-world
events, spanning generation shortages to
the consequences of the lockdown caused
by the COVID-19 pandemic, illustrates
the need for flexible and adaptive mecha-
nisms that guarantee the security of the
system. The article outlines a new opera-
tional paradigm to address energy security.
In the sixth article, Makoto Tanaka
and his coauthors describe transmis-
sion bottleneck management in the two-
frequency power system of Japan. They
describe, in detail, the fundamentals of
a new methodology (the N−1 intertrip
scheme) being implemented to mini-
mize congestion issues exacerbated by
the integration of weather-dependent
BIRD DIVERTERS
FireFly® & BirdMark®
The most effective bird diverters
in independent studies.
renewable sources. Such a methodology
involves the instantaneous curtailing of
generating units in case of a fault by us-
ing an advanced protection system.
In the seventh and final article, Pierlu-
igi Mancarella and Farhad Billimoria il-
lustrate the well-known energy trilemma
of affordability, sustainability, and reli-
ability through the unique perspective of
the Australian system. The long, radial
Australian transmission system with a
high penetration of renewables appears as
an emblematic example of system brittle-
ness, where even futuristic solutions, such
as the world’s largest battery energy stor-
age system installed at Hornsdale (South
Australia), are both the solution to and
source of grid security issues. The article
concludes with an in-depth discussion on
why the economic implications of system
fragility are the most compelling challeng-
es (possibly even more than the technical
challenges) that demand a new approach.
The issue closes with an “In My
View” column by Göran Andersson.
Physical and cyberinsecurities are ana-
making life visibly safer
POWER LINE MARKERS
SpanGuard TM & SpanLite TM
Mark any line. Anywhere.
Standard & self-illuminated markers.
800-722-8078 • pr-tech.com
lyzed using stability concepts known to
power engineers. Such concepts are ex-
panded and adapted to reflect an energy-
constrained production mix. Building
on these stability concepts, the analysis
provides several relevant observations to
boost robustness and resilience in cur-
rent and future electric energy systems.
Our goal in putting together this issue of
IEEE Power & Energy Magazine is to spur
a much-needed conversation on energy
security as the power industry shifts world-
wide from a power-limited to an energy-
constrained paradigm. Such a conversation
will hopefully involve regulators, operators,
producers, consumers, researchers, and the
academic community. We hope that this
will generate innovative ideas that contrib-
ute to ensuring a secure, efficient, and en-
vironmentally friendly supply of electrical
energy. We are most grateful to Mike Hen-
derson for suggesting the creation of this
issue and nurturing its initial steps and
to Steve Widergren, Barry Mather, and
Susan O’Bryan for pertinent corrections
and insightful observations. p&e
LOW LINE MARKERS
QuikMark TM & SpanMark TM
Saving lines & saving lives.
Snaps directly onto live lines.
6874
 Operational
Security The Case of Texas

By Julia Matevosyan, Shun Hsien Huang,

Pengwei Du, Nitika Mago, and Rochie Guiyab

T
THE ELECTRIC RELIABILITY COUNCIL
of Texas (ERCOT) is the independent transmis-
sion system operator that manages the flow of
electric power to serve 90% of the state’s elec-
tric load. ERCOT’s system-wide load peaked at
74,820 MW on 12 August 2019.
The ERCOT region is one of four electric
grid interconnections in North America and
has only a limited connection of 1.2 GW of
direct current ties to its neighboring grids. It
manages power on an electric grid that includes
more than 46,500 mi of transmission lines and
more than 650 generation units. In the last 15
years, ERCOT has experienced significant
growth of renewable generation, most of which
is wind generation resources connected to the
transmission grid. Looking at the new genera-
tion interconnection requests, wind and solar
are expected to be the primary generation
resources added in the future. As of Septem-
ber 2020, ERCOT has an installed wind gen-
eration capacity of ~25 GW and a solar power
generation capacity of ~3.6 GW (Figure 1) con-
nected to the transmission grid. An additional
~13.4 GW of planned wind projects and nearly
16 GW of planned solar projects with signed
interconnection agreements are expected to
connect to the transmission grid before the
end of 2023. There is also about 650 MW of
residential and commercial rooftop solar in
ERCOT as well.

Digital Object Identifier 10.1109/MPE.2020.3043611

Date of current version: 19 February 2021

18 ieee power & energy magazine 1540-7977/21©2021IEEE march/april 2021

 ERCOT’s generation mix has dramatically changed in northernmost 26 counties, and West Texas as well as the
the past 20 years—from being heavily dominated by coal area along the Gulf Coast (Figure 3). These areas have
and gas-steam generation in the late 1990s to now having an average wind generation capacity factor of about 43%
a substantial share of renewable generation and simple and (with consideration of curtailments due to transmis-
combined cycle generation, as presented in Figure 2. Fur- sion constraints).
thermore, nearly 6,700 MW of coal and gas-steam units However, while the Gulf Coast area is generally more attrac-
were retired in the past four years. tive due to higher energy prices and favorable wind generation
patterns better correlated with ERCOT load peak (Figure 4),
Operational Challenges with High more wind generation has been built in West Texas and the Pan-
Inverter-Based Resource Penetration handle area due to land availability and a higher annual energy
To date, the highest percentage of load served by wind gen- yield. Similar considerations apply to solar projects as well. It
eration in ERCOT is 59.3% at 2:10 a.m. on 2 May 2020, should be noted that the renewable resources in the ERCOT
with a total of 19,426 MW served by wind generation. region are generally far from the load centers where most
Most wind and solar generation projects in ERCOT are conventional synchronous generation is also concentrated.
interconnected at the transmission level (i.e., 69 kV, 138 kV, This results in high power transfers over a long transmission
or 345 kV). The best wind resource potential within the path to deliver renewable power to the load centers (Figure 3).
ERCOT region is in the Panhandle, which consists of the In this environment, to ensure the reliable operation of the
ERCOT grid, adequate voltage
support must be maintained
and associated voltage stability
must be assessed.
Wind and solar resources are
connected to the grid through
inverters. These renewable in-
verter-based resources (IBRs)
have a near-zero marginal cost
and, thus, displace more expen-
sive conventional synchronous
generation in unit commitment
and dispatch. As a consequence,
the decline in system inertia and
the corresponding increase in
the rate of change of frequency
during events must also be as-
sessed. Sufficient system re-
serves and flexibility are needed
to manage variability and uncer-
tainty of renewable generation
during real-time grid operations.
To address these challenges and
efficiently and reliably adapt to
the ever-changing generation
mix, ERCOT relies on its en-
ergy market, ancillary services,
wind and solar forecasting, se-
curity assessment studies, and
tools in real-time operation.

Overview of ERCOT
Market Processes
©SHUTTERSTOCK.COM/GRAL

ERCOT is the sole energy-on-

ly market in the United States,
i.e., there is no capacity mar-
ket. Energy is co-optimized
with ancillary services in the

march/april 2021 ieee power & energy magazine 19

 60,000
57,902
57,001
Cumulative Wind MW Installed Cumulative Wind MW Planned
55,000 Cumulative Solar MW Installed Cumulative Solar MW Planned
49,989
50,000 15,435 15,833

45,000 9,576

40,000 3,633 3,633

35,992 3,633
35,000 1,529
3,633 12,961 13,461
(MW)

30,000 11,805
26,141 5,854
25,000 23,634 2,281
21,766 1,857
1,068
20,000 18,228
16,147 566
15,000
12,898
10,771 11,221 23,860
24,976 24,976 24,976 24,976
10,000 20,698 21,777
15,857 17,662
12,729
5,000 10,698 11,100

0
2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012 2012
(Year)

figure 1. The ERCOT installed and planned generation capacity as of 30 September 2020. The cumulative megawatt
(MW) planned include projects with signed interconnection agreements.

Other Other
Solar 1% Solar 1%
0% 2%
Wind Nuclear
0.3% Nuclear 5%
8% Coal
Gas 16%
CT/IC Coal
5%
29%
Wind
23%

Late 1990s 2019

Gas CC
5% Gas CT/IC Gas CC
7% 35%

Gas Steam Gas Steam

52% 11%

figure 2. The changes in the ERCOT generation mix from the late 1990s to 2019. CT: combustion turbines; IC: internal
combustion engine; CC: combined cycle.

20 ieee power & energy magazine march/april 2021

 Planned Projects by County
40–250 MW
251–450 MW
451–780 MW
781–1,400 MW
Operational Wind Units
Major Cities

Dallas

Austin

Houston

San Antonio

figure 3. An ERCOT map with operational and planned wind generation resources and major cities illustrating where
main load centers are located.

voluntary financial day-ahead

market and traded in the real- 70,000 12,000
time energy market. The ERCOT
Wind Power Production (MW)

electricity market processes have 60,000 10,000

evolved to incorporate frequent 50,000
8,000
updates on changes in system
Load (MW)

40,000
conditions. This allows effective 6,000
system adjustments to uncertainty 30,000
and power output variability of 4,000
20,000
wind and solar resources over mul-
10,000 2,000
tiple timeframes, from seven days
ahead to real-time operation. 0 0
Figure 5 illustrates ERCOT
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

energy and ancillary services mar- Hour Ending

ket processes, starting from the
Load Coastal Wind West/Panhandle Wind
day ahead and going into real time.
The day-ahead market is a vol-
untary financial market where figure 4. The differences in diurnal wind power production patterns between the
energy and ancillary services for Coastal zone and West/Panhandle zone and the system load on a typical summer day.

march/april 2021 ieee power & energy magazine 21

 Various types of reserves defined and procured as ancillary
services are deployed when needed to maintain frequency
close to the nominal frequency (60 Hz).
the upcoming day are co-optimized. After the day-ahead
market is cleared, day-ahead reliability unit commitment is
carried out to ensure that sufficient generation is commit-
ted to serving the next day’s load forecast while satisfying
reliability criteria. Day-ahead reliability unit commitment is
reevaluated every hour based on updated system conditions,
load, and renewable resource power production forecasts.
This is called hourly reliability unit commitment.
Security-constrained economic dispatch is executed every
5 min in the operating hour, and dispatch instructions (dis-
patch basepoint), effective immediately, are sent to all of
the resources. The resources adjust their output to attain the
basepoint within 5 min. Being able to adjust the system every
5 min in real-time operation significantly reduces the need
for system reserves to balance the uncertainty and variability
of renewable generation and load.
Ancillary Services
Various types of reserves defined and procured as ancillary ser-
vices are deployed when needed to maintain frequency close
to the nominal frequency (60 Hz). Annually, ERCOT reviews
and revises the methodology for determining ancillary service
requirements. In recent years, ERCOT observed declining sys-
tem inertia due to the decreasing number of online synchronous
Day 0
The hourly reliability unit commitment runs
every hour of the day.
Day- Day-Ahead
Ahead Reliability
Market Unit
for Commitment
Day 1 for Day 1
The day-ahead
reliability unit
commitment
solves for the
period from
00:00 to
23:00 for the
next day.
figure 5. The ERCOT market processes.
22 ieee power & energy magazine
generators. Declining system inertia leads to an increasing rate
of change of frequency during events and an increasing need
for faster frequency response and additional ramping to main-
tain adequate frequency performance. To address this issue,
two new ancillary services were introduced: the Fast Frequency
Response (FFR) Service and ERCOT Contingency Reserve
Service (ECRS), implemented in March 2020 and June 2024,
respectively. A summary of reserves expected to be procured
and maintained through the ERCOT ancillary services, includ-
ing two new ancillary service products, is provided next.
✔ Responsive reserve service (RRS): This is the frequen-
cy containment reserve deployed during low-frequency
events. The needed RRS amount is determined based on
the largest generation contingency (simultaneous loss of
two largest generation units, ~2,805 MW) and anticipated
system inertia conditions. RRS can be provided by genera-
tion resources through traditional governor response, also
called primary frequency response (PFR), and by load re-
sources with under-frequency relays. The recently imple-
mented FFR product is also procured as a part of the RRS.
Studies found that, during low-inertia conditions, faster
responding resources, such as load and FFR resources,
are more effective than the PFR from the generators. An
equivalency ratio was introduced, showing how many
Day 1
The hourly reliability unit commitment runs
every hour of the day.
Day- Day-Ahead
Ahead Reliability
Market Unit
for the Commitment
Next for the
Day Next Day
march/april 2021
 megawatts of PFR generation can be replaced by 1 MW
of faster responding resources, depending on prevailing
inertia conditions. Introducing this ratio allowed ERCOT
to improve efficiency and reduce the overall procurement
costs of RRS, while better aligning procured amounts
with system needs under different inertia conditions.
✔ Regulation reserve service: This is the frequency res-
toration reserve deployed every four seconds. The up-
regulation and downregulation reserves are procured
separately. The regulation reserve amount is primarily
determined based on historic 5-min net-load variability.
✔ ECRS: This is a new service to be used as a replace-
ment reserve that can be deployed within 10 min.
ECRS will be implemented in June 2024.
✔ Nonspinning reserve service: This is the existing re-
placement reserve that can be deployed within 30 min.
The amount of nonspinning reserve service is deter-
mined mainly based on historic hourly net-load forecast
uncertainty, evaluated based on wind, solar, and load
forecasts 3 h before an operating hour. The risk for net-
load ramps is also considered when determining nons-
pinning reserve service amounts.
Figure 6 provides more details about the ancillary ser-
vices listed previously and reflects recently introduced
changes to the ancillary service set.
Interconnection Requirements
for Generation Resources
ERCOT ensures operational security through a strong set
of interconnection requirements that apply to generation
resources including IBRs. These requirements include, but
are not limited to the following:
✔ Voltage support: This is used to provide leading and
lagging reactive capability with a 0.95 power factor.
The capability is determined at the maximum active
power output, and voltage support within this capabil-
ity limit has to be provided continuously, independent
of the active power output level.
✔ Frequency support: This provides the PFR with a 5%
maximum droop and a maximum frequency deadband
of ±0.017 Hz when frequency-responsive headroom
is available. Capacity reservation is not required un-
less the resource offered and was awarded RRS in the
day-ahead market. Figure 7 illustrates the frequency

Regulation No Change

FFR
• Triggered at 59.85 Hz and Gives a Full Response in 15 Cycles
• Once Deployed, Sustained for Up to 15 min; Once Recalled,
RRS Restores Within 15 min
PFR
FFR
• PFR-Capable Capacity Reserved on Generators or
Controllable Load
Load Resources on UFR
• 1,150 MW Minimum Must Be Provided by Resources Capable of PFR
PFR Load Resources on UFR
• Triggered at 59.70 Hz and Gives a Full Response in 30 Cycles
• Is Sustained Until Recalled; Once Recalled, Restores Within 3 h
• Beyond the Minimum PFR, Up to 60% of Total RRS Can Come From
Load Resources on UFR or FFR

Generation
ECRS
• Online or Offline Capacity That Can Be Converted to Energy
Within 10 min
• Dispatched by Security Constrained Economic Dispatch
10-min Ramp
Load Resources (UFR Not Required)
Load Resources • Up to 50% of ECRS Capacity Can Come From Load Resources
May or May Not Be on UFR With or Without UFR
• Once Deployed, Must Respond Within 10 min; Restoration Within 3 h

No Protocol Changes
• Proposed Methodology for Nonspin Reserve Service Quantities in
Nonspin
This Framework—Quantities Computed Using 2018 Ancillary
Services Methodology Are Reduced by ECRS Quantities

figure 6. ERCOT ancillary services. The right-hand side illustrates recent changes introduced to the ancillary services
framework. UFR: under-frequency relays.

march/april 2021 ieee power & energy magazine 23

 response of the wind generation resource to low and
high-frequency events.
✔ Voltage and frequency ride-through capability: This
is used to maintain the ability to remain connected to
the grid and for providing the necessary support during
voltage and frequency events of predefined severity.
✔ Dispatchability: This is to ensure that resources fol-
low the dispatch instructions for maintaining or ad-
justing power output in a timely and precise manner.
✔ Ramping: This is used to limit the ramp rate of each
wind or solar resource to 20% per minute of its name-
plate rating (megawatts) when responding to ERCOT
dispatch instructions.
✔ Telemetry: This provides numerous telemetry data
points to ERCOT with a temporal resolution of 2–10 s. These
data points are critical to operation studies and situ-
ational awareness tools as well as for offline analysis.
Renewable Generation Forecasting
Renewable production potential forecasting is essential to
the integration of large amounts of renewable generation.
ERCOT acquired its first wind forecast service in 2010 and
added a second in 2017 to provide a centralized prediction
for all renewable generation connected to its transmission
grid. Both of these utilize site-specific meteorological
information, telemetered by wind generation resources,
to generate wind power production forecasts for each
resource. Adding another wind forecasting service pro-
vided ERCOT’s control room with an ability to enhance
situational awareness and increased resiliency for the pro-
vision of this critical information.
With the addition of the second provider, new capabilities
of forecasting for extreme weather were procured from both
providers and made available to the control room. In addition
to the midterm wind generation forecast, which is delivered
each hour for the next 168 h, ERCOT also has an intrahour
wind forecast, which is delivered at a 5-min resolution for the
next 2 h. The averaged performance, namely the mean abso-
lute percentage error, of day-ahead and 1-h-ahead midterm
wind forecasting in 2019 was 4.6% and 2.7%, respectively
(see Figure 8).
In 2017, ERCOT started solar forecasting for solar generation
with a size greater than 10 MW, connected to the transmission
network. The averaged performance of ERCOT’s day-ahead and
1-h-ahead solar power production forecasting in 2019 was 6.55%
and 5.7% mean absolute percentage error, respectively.
Operations Analysis and Studies
To ensure reliable system operation in real time under vary-
ing system conditions, the system operator periodically car-
ries out studies and analyses, starting from the day ahead
and going into real time, and uses situational awareness tools
to monitor system conditions (Figure 9).
Day-Ahead Analysis and Studies
✔ Generic transmission limits for the day-ahead market:
In addition to well-defined tools and processes to man-
age the steady-state constraints, increasingly stability-
related constraints also need to be controlled in opera-
tions. Generic transmission constraint is a restriction
on one or more grouped transmission elements to man-
age stability-related limitations. A two-day-ahead

Wind Resource Response to Wind Resource Response to

Low Frequency 13 July 2016 High Frequency 25 August 2015
60.05 180 60.075 140

60.025 178 60.056

130
60 176 60.038
120
59.975 174 60.019

59.95 172 60 110

59.925 170 59.981

100
59.9 168 59.963
90
59.875 166 59.944

59.85 164 59.925 80

13:25 13:28 13:31 13:34 13:37 13:40 16:13 16:16 16:19 16:22 16:25 16:28

Frequency Nominal Frequency MW

figure 7. The frequency response from wind generation resources in ERCOT.

24 ieee power & energy magazine march/april 2021

 offline study is conducted to determine transfer
limits based on the projected system condition on
all existing generic transmission constraints. The
identified stability limits on generic transmission
constraints are used in the day-ahead market and in
reliability unit commitment processes and are up-
dated in real-time, if necessary, to maintain reliable
transfers in real-time operation.
✔ Next-day study: This study includes steady-state, con-
tingency, and voltage-stability analyses for the upcom-
ing operating day to ensure that ERCOT has sufficient
generation to reliably serve projected peak load. It does
so while taking into account day-ahead market and re-
liability unit commitment results, applicable system
ment run. This study, therefore, has an updated outlook
of the available generation compared to the next-day
study. The gap study uses the same process as the next-
day study and includes steady-state, contingency, and
voltage stability analyses.
✔ Real-time voltage security assessment: This is
conducted every 10 min, using real-time system con-
ditions, to determine the transfer limits of the con-
cerned regions based on steady-state voltage stabil-
ity or system strength considerations (whichever is
more binding). The determined transfer limits are
1-h-Ahead Wind Forecast Performance

Absolute Percentage Error

6
outage information, and transmission line dynamic
5

5.6%
ratings if applicable. Mitigation plans may also be de-

5.3%

5.3%
veloped for any issues that may arise. 4

Mean

4%
3

3.7%

3.6%

3.2%
Real-Time Analysis and Studies 2

2.7%
✔ Gap study: This is similar to the next-day study but car- 1
ried out closer to real time using the most updated infor- 0
2012 2013 2014 2015 2016 2017 2018 2019
mation, such as generation patterns for wind and solar, Year
updated based on the latest operating plans submitted
by each resource. The generation input for the gap study figure 8. The annual 1-h-ahead wind power production
is the result of the latest hourly reliability unit commit- forecast performance.

Operating Day-2 Next Day Operating Day

Generic Multihour Study for the Entire Day

Transmission
Constraints Day-Ahead
Day-
Ahead Reliability
Market Unit Next Study the Peak Load Hour
for Commitment Day
Day 1 for Day 1 Study
Study the Peak Load Hour
Gap
Study
Runs Every 10 min

Voltage Stability Assessment

Transient Security Assessment

Inertia Monitoring and Look-Ahead

Reserve Adequacy Monitoring and Look-Ahead

Capacity Adequacy Monitoring and Look-Ahead

Physical Responsive Capability

Runs Every 4 s

figure 9. Operational studies and situational awareness tools.

march/april 2021 ieee power & energy magazine 25

 ERCOT identified the need for a reactive power control tool in
operations to improve the control and coordination of various
reactive devices.
managed through generic transmission constraints
enforced in real-time security-constrained eco-
nomic dispatch.
✔ Real-time transient security assessment (currently
under implementation): Most dynamic stability-re-
lated generic transmission constraints are based on
offline studies that generally include conservative as-
sumptions to account for various potential real-time
system conditions. ERCOT is planning to implement
the transient security assessment in real-time opera-
tion to determine the limits based on the most recent
operating conditions.
Situational Awareness Tools
✔ Inertia monitoring tool: Since IBRs do not naturally
contribute to system inertia as the number of these re-
sources continues to increase and displace synchro-
nous generators in a power system’s generation mix,
the synchronous inertia will inevitably decline. ER-
COT identified a critical inertia level that needs to be
maintained at all times to minimize the risk of under-
frequency load shedding after a large generation trip.
In 2015, ERCOT implemented the real-time inertia
monitoring tool. It continuously calculates the current
total inertia contribution of all online synchronous
generators based on the inertia parameters of indi-
vidual units in the network model and on the online
status of the units in the energy management system.
The tool also calculates future inertia conditions for
the next 168 h on a rolling basis. This calculation is
based on the generation unit commitment plans that
every generator submits to ERCOT every hour (span-
ning the next 168 h). The tool then identifies any peri-
ods where the expected system inertia is less than the
critical level, in which case ERCOT system operators
will follow procedures to deploy nonspinning reserve
or start other synchronous generators to increase syn-
chronous inertia online.
As of now, ERCOT’s all-time-minimum inertia
level is still about 30% higher than the determined
critical inertia level. This means that ERCOT has
not yet had to deploy any of the above measures to
increase inertia.
✔ Reserve adequacy tool: Another situational aware-
ness tool developed at ERCOT is the reserve ad-
equacy tool. ERCOT publishes RRS amounts before
the start of each year for every hour of the upcoming
26 ieee power & energy magazine
year. These requirements are based on historic inertia
conditions from the past two years and vary by month
and time of day. In real time, however, the system in-
ertia may be different from what was expected based
on historic data. The reserve adequacy tool allows
operators to verify whether the available frequency-
responsive capacity, including procured reserves, is
sufficient based on actual inertia and calculated future
inertia conditions. The tool uses information from the
inertia monitoring tool and, using empirical equations
obtained from the series of dynamic studies, calcu-
lates the amount of reserves needed to protect against
the largest generation contingency. The tool then com-
pares all available frequency-responsive capacity with
the amount of reserves needed.
In this comparison, the tool also takes into account
the prevailing equivalency ratio between faster
responding resources and traditional PFR, based on
inertia conditions. If the available frequency-respon-
sive capacity, including the RRS procured in the day-
ahead market, is insufficient, ERCOT may open a
supplemental ancillary service market and procure
additional reserves as needed.
✔ Physical responsive capability monitoring tool: This
tool is a monitoring tool used in operations as an indi-
cator of overall system “health” in terms of available
frequency-responsive capability. The tool summarizes
all of the available frequency-responsive capacity
from online resources. If the available capability is
below a certain threshold and cannot be restored to
a higher level over a given time, ERCOT enters into
an energy emergency alert, and system operators fol-
low procedures and steps to recover the system back to
normal conditions with sufficient physical responsive
capability available.
✔ Capacity adequacy tool: This tool was developed
at ERCOT to assess whether the generation capac-
ity is sufficient to serve the forecasted load demand
for the next 1–24 h. It tracks and monitors online
generation capacity and reserves as well as offline
generation capacity that can potentially be started
in the future hours, up to 24 h ahead. To account
for the variability and uncertainty of intermittent
resources, the tool also allows users to evaluate the
impacts of different levels of load and wind fore-
cast uncertainties on system capacity adequacy.
Thus, operators have a better awareness of the
march/april 2021
 situation for the future grid operating conditions
and can effectively mitigate anticipated large net-
load ramp events.
Current Challenges and Work Planned
for the Nearest Future
Increasing numbers of IBRs and declining online syn-
chronous generation lead to lower system strength and
higher voltage sensitivity. Frequent adjustments of voltage
profiles and reactive power devices are expected due to
the variation of renewable generation and load patterns.
To address this challenge, ERCOT identified the need
for a reactive power control tool in operations to improve
the control and coordination of various reactive devices,
including generator voltage profile, switchable shunts,
static var compensators, and static synchronous compen-
sators, for both day-ahead and real-time operations. Cur-
rently, there is no commercial tool available to meet this
need, and ERCOT is in the process of developing and
implementing one.
ERCOT has already implemented a transient secu-
rity assessment tool in operations to assess the system’s
dynamic stability on an as-needed basis. However, due to
the increasing dynamic stability challenges in the evolv-
ing grid, ERCOT plans to implement a real-time transient
security assessment tool to assess real-time dynamic sta-
bility. Similar to the real-time voltage stability assessment,
the real-time transient stability assessment is expected to
run automatically at regular intervals, e.g., every 10 min, to
determine dynamic stability limits on defined transmission
interfaces. The identified limits will be managed through
generic transmission constraints enforced in security-con-
strained economic dispatch.
Conclusions and Recommendations
As more wind and solar generation is added to the system,
ERCOT and its market participants are introducing neces-
sary changes to the market rules, ancillary services, and
operation practices as well as developing new situational
awareness tools and real-time studies to ensure continued
system reliability. Based on ERCOT’s experience, the fol-
lowing practices are recommended to ensure operational
security with high IBRs:
✔ Granular unit commitment and economic dis-
patch must take into account the most up-to-date
load, renewable forecast, and generation unit sta-
tus information.
✔ Essential reliability support needs to be regularly
reviewed and revised to account for evolving sys-
tem characteristics. While some essential reliability
requirements may need to be implemented through
interconnection requirements or grid codes, market
solutions should also be considered where practical.
✔ Ancillary service products need to grow with the
evolving generation mix, and faster, more flexible
march/april 2021
services may be required as the share of IBRs contin-
ues to increase. The continuous refinements of ancil-
lary service methodology reflecting changing system
needs is recommended.
✔ It is recommended that possibilities to supplement
diminishing inertial response from synchronous
generation with the use of emerging technologies
are explored.
✔ Granular telemetry from all generation resources is
required to inform situational awareness tools and
studies conducted in operations as well as ex-post
analysis.
✔ Conducting several studies and analyses in day-
ahead and real-time situations allows maintaining
reliable system operation. This simultaneously al-
lows for more efficient operation under less con-
servative assumptions compared to predetermined
limits based on offline studies with worst-case sce-
nario assumptions.
✔ Real-time situational awareness tools in the control
room are essential for efficient and reliable operation.
Continuously improving existing tools and identifying
the need for new tools is paramount for adapting to the
rapidly evolving power system.
For Further Reading
“Inertia: Basic Concepts and Impacts on the ERCOT Grid,”
ERCOT, Taylor, TX, Apr. 2018. [Online]. Available: http://
www.ercot.com/content/wcm/key_documents_lists/141324/
Inertia_Basic_Concepts_Impacts_On_ERCOT_v0.pdf
“2017 Responsive Reserve Service (RRS) study for deter-
mining 2018 ancillary service RRS requirements,” ERCOT,
Taylor, TX, Sept. 2017. [Online]. Available: http://www
.ercot.com/content/wcm/key_documents_lists/108744/05
._RRS_Study_2017_Methodology_11022017.docx
“Use of generic transmission constraints in ERCOT,”
ERCOT, Taylor, TX, July 2020. [Online]. Available: http://
www.ercot.com/content/wcm/lists/197392/The_Use_of
_GTCs_in_ERCOT_July_2020.pdf
“Nodal protocols and operating guides,” ERCOT,
Taylor, TX, Nov. 2020. [Online]. Available: http://www
.ercot.com/mktrules/nprotocols/current
Biographies
Julia Matevosyan is with the Electric Reliability Council of
Texas, Taylor, Texas, 76574, USA.
Shun Hsien Huang is with the Electric Reliability Coun-
cil of Texas, Taylor, Texas, 76574, USA.
Pengwei Du is with the Electric Reliability Council of
Texas, Taylor, Texas, 76574, USA.
Nitika Mago is with the Electric Reliability Council of
Texas, Taylor, Texas, 76574, USA.
Rochie Guiyab is with the Electric Reliability Council of
Texas, Taylor, Texas, 76574, USA.
p&e
ieee power & energy magazine 27
 By Guillermo Bautista Alderete

T
THE CALIFORNIA INDEPENDENT SYSTEM OPERATOR (CAISO) IS ONE
of nine independent system operators (ISOs) in North America. Its primary
responsibility is to ensure the safe and reliable delivery of electricity on the
power grid. CAISO manages the flow of electricity for about 80% of California
and a small part of Nevada, which encompasses all of the investor-owned utility
territories and some municipal utility service areas. The ISO wholesale energy
market is comprised of day-ahead and real-time processes. The energy products
and services traded in its markets allow the organization to meet reliability
needs and the serve load at the least cost. The ISO is responsible for ensuring
fair, nondiscriminatory, and transparent access to the transmission network and
market transactions.
In 2014, CAISO, together with PacificCorp and other balancing areas,
designed and implemented the energy imbalance market (EIM). The EIM
allows participants to buy and sell power close to the time in which electricity
is consumed and gives system operators real-time visibility across neighboring
grids. Figure 1 displays the footprint of the CAISO balancing area and that
of the EIM. About 82% of the load in the Western Interconnection is either
active in or already committed to participating in EIM by spring 2022. This
real-time market improves the balancing of intrahour supply with demand at
a lower cost than could otherwise be achieved by hourly or 15-min scheduling
practices. In addition to the optimized balancing of supply and demand, EIM
manages congestion on the transmission system to maintain grid reliability and
support the integration of renewable resources. The western EIM also makes

Energy
Insecurity Due
to Gas Supply
Availability
Digital Object Identifier 10.1109/MPE.2020.3043612
Date of current version: 19 February 2021

28 ieee power & energy magazine 1540-7977/21©2021IEEE march/april 2021


excess renewable energy available to participating utilities
at a low cost rather than curtailing the production of clean
energy. This also leverages the regional and economic
diversity of the broader area and allows for more flex-
ibility in scheduling and dispatching generation. Since its
inception in 2014, the EIM has resulted in gross economic
benefits of more than US$1 billion. In the context of gas
supply availability, the EIM allows for more geographi-
cal, economical, and technological diversification in the
extended market footprint.
The Supply Mix in California
Is Rapidly Evolving
The electric grid is rapidly evolving and transitioning
toward a clean and environmentally sustainable electric sys-
tem. Grids around the world are working to meet aggressive
renewable goals. As a result of California’s SB100 legisla-
tion, passed in 2018, the state is working toward its goal
march/april 2021
Efforts to
Coordinate
Electric and
Gas Systems
of meeting 60% of its electrical supply
needs with renewable energy by 2030
and 100% of them by 2045.
With the rapid transformation to a
cleaner but more variable and energy-
limited generation fleet, CAISO is devel-
oping new policies and practices to sup-
port the unprecedented transformation of
its electric system. This transformation
is rapidly changing the landscape of the
system’s generation mix to meet demand.
Currently, the supply mix of the CAISO
system is composed of generating facili-
ties with different technologies and fuel
sources, namely nuclear, gas, hydro,
solar, wind, and, more recently, batteries.
CAISO also relies on imports from the
©SHUTTERSTOCK.COM/LIGHTSPRING
northwest that are predominantly hydro.
Clean and variable energy production is
growing in the system while more conven-
tional thermal generation is retiring as part
of environmental goals, such as meeting a
once-through-cooling compliance policy,
which results in the retirement, repower-
ing, or remediation of coastal power
plants. Figure 2 illustrates an hourly production mix on a typi-
cal 2019 summer day. Gas generation accounted for about 32%,
followed by renewables with 26%, and imports with 22%. As
renewable production steeply reduces in the evening, gas and
hydro generation, together with imports, ramp up to compensate
for it.
Supply Adequacy
and Availability
The CAISO system relies on a bilateral-based mechanism to
secure capacity, unlike some other ISOs that rely on capacity
markets. This resource-adequacy program is administered by
CAISO and developed by local regulatory authorities with the
California Public Utilities Commission (CPUC). At the onset
of this program in 2006, the generation mix of the system was
predominantly conventional generation technologies, such
as nuclear, gas, and hydro resources. Some of these resources
had a limitation on the energy they could produce due to
ieee power & energy magazine 29
 environmental regulations and air permits, but they generally This technology needs to provide an equally flexible upward
had reliable and predictable fuel sources that allowed them capacity to match the ramping capability of variable
to produce energy where and when the system needed. With resources. Figure 3 presents how the 3-h evening ramp has
these characteristics, the resource-adequacy program’s focus increased year after year. This ramp exceeded 15,000 MW
was on the overall system-capacity need, which required juris- in 2019. A large portion of this ramp is currently absorbed
dictional load-serving entities to secure the capacity to meet by gas-fired generation. When there is a limitation on gas
up to 115% of their expected peak load needs. availability, the risk of not being able to support this ramp
Over the years and with the transition to achieving envi- increases and poses a reliability concern.
ronmental goals, the generation fleet that CAISO uses to Every year, CAISO conducts its summer loads and
operate the grid has become more diverse and variable. This resources assessment to study the upcoming summer sup-
has required CAISO to refine certain aspects of its resource- ply-and-demand outlook for the CAISO balancing authority
adequacy tariff provisions to ensure it has the flexible capac- area and identify issues that could lead to potentially chal-
ity needed to operate the grid reliably. The flexible resource- lenging operating conditions. This assessment uses a variety
adequacy requirement secures capacity with the needed of inputs from external and internal sources as well as sto-
flexibility to sustain the ramping requirements arising during chastic production-simulation models and more than 2,000
the evening peak when solar production is decreasing. scenarios of forecasted load and renewable generation with
Ramping needs have been growing rapidly in the last different weather conditions.
years as more solar and variable generation come online In recent years, this assessment has included the poten-
in the CAISO system. When aggregated variable-source tial impact of operating restrictions on the Aliso Canyon gas
generations decrease production, other types of resource storage facility. As explained in greater detail in subsequent
technologies will have to ramp up to offset the reduction. sections, a leak in this gas facility had profound implications
on the supply availability of gas-
fired resources in the southern part
of the CAISO system.
Market Operator For the load, this study considers a
Puget Sound CAISO distribution of load patterns that include
Energy Powerex the range of one-in-two, one-in-five, and
EIM Entity
Seattle one-in-10-year peak-demand forecasts.
Active Participant
City Light The forecast accounts for load reduc-
Tacoma Planned EIM Entry 2021 tions from behind-the-meter solar pro-
Power Avista
Planned EIM Entry 2022 duction and energy efficiency programs.
Portland NorthWestern For generation, it considers the expected
General Bonneville Energy
Electric
generation from hydro resources based
Power
Administration Idaho on expected hydro conditions and
PacifiCorp Power
the reliance on imported resources.
The availability of imported energy is
uncertain but a function of contracted
NV PacifiCorp
BANC
Energy
resource-adequacy capacity, the load
in external areas demand, transmis-
Turlock
sion and resource outages, and other
Irrigation
District fuel supply conditions. It also consid-
California Xcel Energy-Colorado ers the variation in the contribution
ISO of renewable resources, retirements or
mothballing of existing units, and new
Arizona Public
Los Angeles Service Public Service generation reaching commercial oper-
Department of Company of ation. This effort assesses CAISO’s
Water and New Mexico
resource adequacy for capacity, ancil-
Power lary services and flexible capacity, and
Salt River
Project Tucson the potential risk for load interruption.
Electric In preparation for the summer, CAISO
Power releases its assessments to start coor-
dinating with other agencies involved
in the sector, including the Western
figure 1. The CAISO and EIM footprint. BANC: Balancing Authority of Northern Electricity Coordinating Council,
California. transmission operators, neighboring

30 ieee power & energy magazine march/april 2021

balancing areas, natural gas providers, and the California facility is a key part of the gas system in the Los Angeles
Department of Forestry and Fire Protection. Basin. The leak impacted the overall gas market, including
gas-fired generating plants in that area. Following the leak,
Gas System in California the withdrawal and injection of gas in this storage facility
The gas system in California is defined in two major areas: were largely restricted. Aliso Canyon supplies gas to 17
one in Northern California, covered by Pacific Gas and Elec- natural gas-fired power plants, accounting for more than
tric (PG&E), and one in Southern California with Southern 9,000 MW of generation capacity, while also indirectly
California Gas Company (SoCalGas). These companies sup- impacting more than 40 plants, totaling 20,000 MW in the
ply gas not only to gas-fired generators but also to residen- southern part of the system.
tial and industrial customers. Electric generators are deemed The California Energy Commission, CPUC, Los Ange-
noncore customers, and, when the gas supply is limited, cur- les Department of Water and Power, and CAISO formed
tailments apply to noncore customers first. an interagency task force in consultation with SoCalGas to
The majority of the natural gas in California is supplied assess the risks of the limited supply availability that Aliso
from locations outside of the state, including Canada, the Canyon introduced to the gas and electric markets. Among
U.S. Rocky Mountains, and the U.S. Southwest. California the various risks, this assessed the reliability risks of load
has its gas supplied through the interstate gas pipeline sys- interruption in the electric market.
tem, composed mainly of Kern River, El Paso, Gas Trans-
mission Northwest, Transwestern, Ruby, and Mojave Pipe-
lines. PG&E and SoCalGas deliver gas through the intrastate
pipeline system known as the backbone gas system, which is 40,000
further distributed through local pipelines. Large customers, 35,000
30,000
Supply (MW)
such as some electric generators, may be connected directly 25,000
to the backbone system. 20,000
In addition to some independently owned storage facili- 15,000
10,000
ties, both PG&E and SoCalGas own and manage gas storage
5,000
facilities within their service territory. Such facilities expand 0
the gas capacity to meet the demand that otherwise would
12:00 a.m.
2:00 a.m.
4:00 a.m.
6:00 a.m.
8:00 a.m.
10:00 a.m.
12:00 p.m.
2:00 p.m.
4:00 p.m.
6:00 p.m.
8:00 p.m.
10:00 p.m.
require additional pipeline infrastructure. The SoCalGas
system has four storage facilities interconnected through the
transmission pipeline system, namely Aliso Canyon, Honor
Time
Ranch, La Goleta, and Playa del Rey. PG&E owns three
main gas storage facilities: McDonald Island being the larg- Nuclear Natural Gas Large Hydro
est storage, followed by Pleasant Creek and Los Medanos. Imports Renewables Batteries Coal
Unlike electricity, natural gas moves slowly in the pipe-
lines at rates of no more than 30 mi/h. This requires the use of figure 2. A daily supply profile on a summer day of 2019.
gas from storage facilities to com-
plement the supply from the pipe-
lines to meet expected demand.
28,000
Both PG&E and SoCalGas are
operators for their gas systems. 26,000
Analogous to an electric system,
there is a need to balance the gas 24,000
supply from the pipeline system 22,000 2012
Net Load (MW)

and storage with the demand of

customers. When supply is lim- 20,000
2013
ited to meet all of the expected gas 18,000
demand, gas system operators can 2014
issue curtailments for gas deliv- 16,000 2015
2016
eries, typically starting from the 2017 3-h Ramp
14,000 of 15,639 MW on
large noncore customers, such as Net Load of 2018
5,439 MW on 1 January 2019
electric generators. 12,000 2019
5 May 2019
In October 2015, Aliso Can-
10,000
yon natural gas storage located in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Southern California experienced a
significant gas leak. This storage figure 3. The progression of net load curves for CAISO’s system.

march/april 2021 ieee power & energy magazine 31

 Commitment Costs and Energy Bids
CAISO runs both day-ahead and real-time markets using a
cost-minimization problem based on bids submitted by sup-
ply and demand. Generating costs include minimum load
costs, start-up costs, transition costs, and variable production
costs. The markets also use energy reference bids for when
market power mitigation triggers. All of these costs rely on
fuel costs. Participants can submit bids between the bid floor
and cap for the variable production component while the
bids submitted for commitment costs are capped at 125% of
the nominal value of estimated costs.
CAISO averages the natural gas next-day commodity
prices published by different providers plus the applicable
transportation costs to estimate fuel prices. The three main
gas hubs used for commodity pricing in CAISO’s market are
SoCal Citygate, PG&E Citygate, and Kern River. CAISO’s
commitment cost estimates used in the real-time market are
based on these fuel prices. Similarly, reference bids used in
the market power-mitigation process rely on these fuel prices
and include a 10% adder. The 25% headroom on commit-
ment costs and 10% adjustments on reference bids serve as
headroom to absorb any uncertainties or other actual costs
borne by generators.
For the day-ahead electric market, CAISO uses the most
recently available gas commodity prices by 8:30 a.m. Pacific
Prevailing Time before the close and run of the day-ahead
market for the next day. Any change in traded gas prices
afterward, such as those reflected in the intraday or same-
day strips in gas trading, are not reflected in CAISO’s cost
estimates. If there are strained market conditions, such as
the risk of penalties from deviations from a daily balancing
requirement, the traded gas prices during these procurement
and nomination periods are expected to increase relative to
timely trading. Under this scenario, estimated costs may
not reflect real-time conditions and could lead to inefficient
real-time commitments and dispatches as well as insuffi-
cient cost recovery for dispatched resources.
With a significant amount of gas-fired generation by gas,
gas prices will heavily influence electric prices. Figure 4
illustrates the system-wide electric price in CAISO in com-
parison to the prices at the two main hubs in CAISO’s foot-
print. Prices in Southern California are more dynamic and
directly impact electric prices. The figure displays the strong
correlation between electric prices with gas prices.
Aliso Canyon Storage
and Gas Limitations
Given the storage limitations at Aliso Canyon, CAISO initi-
ated an assessment of market mechanisms and other tools
to minimize the risks associated with managing limited
gas-fired supply in the CAISO electric system. The objec-
tive was to support reliable operation while minimizing
any adverse impact on the markets and assess how limited
gas supply affected generating facilities and what new risks
could arise under these conditions.
With a scenario of limited gas-fired supply as a result
of the need to comply with gas-balancing requirements,
CAISO identified four potential adverse market impacts in
the electric sector.
1) There could be a mismatch between actual gas de-
mand and estimated gas flow. The bid window for the
electric market closes at 10 a.m. Pacific Time; market
solutions are available by no later than 1 p.m. Pacific
Time. Scheduling coordinators managing generating
facilities will not know the electric schedule of gas-
fired plants until then. Based on the electric sched-
ules, a gas-fired plant can estimate the required gas
burn for the next day. On the gas side, the first cycle
for a gas nomination, known as the timely nomination
cycle, closes at 11 a.m. Pacific Prevailing Time, which
is before the publishing of the electric market results.
Thus, the timing between electric and gas markets
does not allow resources to make gas procurement or
pipeline nominations based on schedules determined
by the electric market. This
creates a risk of the actual gas
demand being different from
Day-Ahead Electric Price (US$/MWh)

70 280 the nominated gas flow.

Average Day-Ahead Electric Price 2) Electric dispatches from CAI-
Gas Price (US$/MMBTu)

60 240
SoCal Citygate PG&E Citygate SO could not be followed due to
50 200
gas availability or constraints.
40 160 As part of the market solutions,
30 120 the CAISO market issues com-
20 80 mitment and dispatches for re-
sources to follow. This allows
10 40
the electric market to meet de-
0 0 mand at minimum costs while
1-Jan-17
1-Feb-17
1-Mar-17
1-Apr-17
1-May-17
1-Jun-17
1-Jul-17
1-Aug-17
1-Sep-17
1-Oct-17
1-Nov-17
1-Dec-17
1-Jan-18
1-Feb-18
1-Mar-18
1-Apr-18
1-May-18
1-Jun-18
1-Jul-18
1-Aug-18
1-Sep-18
1-Oct-18
1-Nov-18
1-Dec-18
1-Jan-19
1-Feb-19

taking into account all of the

system conditions. Dispatches
for gas-fired resources are
supported with the gas supply;
figure 4. Gas and electric prices in CAISO’s system. incremental electric dispatches

32 ieee power & energy magazine march/april 2021

 CAISO worked closely with other agencies and SoCalGas to
implement market enhancements and tools to efficiently operate
the electric system while helping manage limited gas supply.
will require the corresponding incremental gas pro-
curement. If resources cannot follow instructions
because they are unable to procure gas while man-
aging gas deviations within the daily gas balancing
limits, there can be a disconnect between electric
market dispatches and what is available to meet ac-
tual demand. This disconnect may cause a balancing
problem and turn into a reliability condition.
3) There could be less-efficient commitments and dis-
patches of gas resources. The CAISO markets commit
and dispatch resources based on minimum load costs,
start-up costs, transition costs, and variable operation
costs. The markets use a daily gas price index to ac-
count for the fuel cost of gas resources. Depending
on how accurately an index price reflects the real-
time gas conditions, the electric market may produce
less-efficient commitments. In turn, these suboptimal
instructions may exacerbate the limited supply condi-
tion. For instance, using low fuel prices when there are
gas limitations can result in gas resources artificially
appearing cheaper, and, thus, they can be committed
or dispatched more frequently or to a higher level.
This leads to more gas usage, further stressing the gas
supply limitation.
4) There could be an inability of resources to recover
costs. Deviations from not following CAISO instruc-
tions are settled on real-time prices and expose resourc-
es to additional costs. Resources also settle based on
commitment costs, and if they are based on fuel prices
not reflecting the actual procurement costs, resources
may not sufficiently recover their costs associated with
meeting CAISO’s commitments and dispatches.
As a result of these assessments, CAISO worked closely
with other agencies and SoCalGas to implement market
enhancements and tools to efficiently operate the electric
system while helping manage limited gas supply. Origi-
nally, these changes were implemented temporarily, and,
since then, they have been extended annually or, in some
cases, b e c o m e p e r m a n e n t f e a t u r e s . These changes
have been gradually reassessed and enhanced over the last
years through five different phases.
The most recent SoCalGas winter assessment con-
cluded that, in the best-case scenario, there will be insuf-
ficient supplies to meet a one-in-10-year cold day demand
forecast. The SoCalGas assessment goes on to state that
“in a one-in-10-year cold day scenario, some level of non-
core curtailment may be required, either voluntary or
march/april 2021
involuntary, beginning with (electric generation) demand
following the commission-approved procedure.”
The SoCalGas assessment is consistent with CAISO’s
experience with the natural gas system in Southern Cali-
fornia since the initial Aliso Canyon restrictions. It shows
that the Southern California gas system, given the current
limitations on the gas transmission and storage system, is
insufficient to maintain natural gas reliability to electric-
generation customers during high-demand periods. These
system-level challenges are exacerbated by the require-
ment that gas companies must maintain reliable service to
core customers before serving noncore customers. CAISO
requires close coordination with SoCalGas to ensure that
electric-generation resources are dispatched appropri-
ately to maintain both gas and electric system reliability.
Different winter assessments by the CPUC and SoCalGas
have indicated that, even though critical pipelines have
come back to service, the ability to provide reliable ser-
vice to gas customers depends on the ability to withdraw
gas from Aliso Canyon and on no additional outages on
main pipelines.
In April 2020, the CPUC released the 2020 summer reli-
ability assessment for Southern California. Based on the
most recent assessment of the operation of the Aliso Canyon
facility and projected conditions, gas supply conditions have
improved following previous years. This is a result of more
gas in storage, not only in Aliso Canyon but also in the other
three storage facilities, and of the return to service of a key
pipeline after an outage in 2017. The CPUC revised its Aliso
Canyon protocol, which provides more flexibility for using
this storage facility to balance the gas system. Although
expected conditions show that the gas balance can be main-
tained by using other storage facilities, there is still a risk of
limited supply.
Given these potential risks, CAISO expects these limita-
tions to remain a constant fixture over the long term rather
than merely being short-term challenges. Accordingly, in
October 2019 CAISO sought and received the authority to
permanently implement a market enhancement related to
Aliso Canyon. These market features and tools are described
in the subsequent sections.
Gas Constraints in the Electric Market
The CAISO market enforces all required system con-
straints related to the transmission system, including opera-
tional limits for thermal and security conditions. As part
of the congestion management embedded in the market
ieee power & energy magazine 33
 clearing process, resources are required to increase or
decrease generation. This produces the right market sig-
nals through prices that incentivize resources to follow
CAISO instructions. Resources that exacerbate conges-
tion on a constraint are exposed to a negative marginal
congestion component that reduces the overall locational
marginal price of the resource.
This price is consistent with the need for the resource
to decrease production. The congestion-management pro-
tocol applies regardless of the type of generation technol-
ogy or fuel type and is based on both economical bids and
resources’ effectiveness in relieving congestion. This, in
turn, allows the reliable operation of the system while meet-
ing supply with demand and is possible because the mar-
ket mechanism has the information to enforce transmission
limits. This minimizes the risks of violating or overloading
transmission constraints.
Even when conditions change, such as when there are
transmission outages or a reduction of capacity (which
result in the need for a different operation of resources),
the market will redispatch resources to comply with the
new and reduced operating limit. When certain opera-
tional conditions are not visible to the market mechanism,
there can be a disconnect between the market instructions
and the actual operating condition of the system, which,
if not addressed, can lead to or exacerbate a reliability
condition. For this reason, system operators rely on a
variety of tools and visualizations to know actual system
conditions and adjust as needed. When a condition in the
system cannot be addressed through the market, operators
have tools to adjust for increases or decreases in genera-
tion throughout the system. It is less efficient, but it is a
mechanism of last resort.
When Aliso Canyon restrictions took effect, they effec-
tively limited the amount of gas supply available to elec-
tric generators. On one hand, if the electric system has
no visibility on these supply restrictions, there are risks
of issuing operating instructions that cannot be followed,
which could lead to an electric reliability condition. On the
other hand, if these restrictions are not followed, it could
jeopardize the gas restriction and potentially lead to a gas
reliability condition. In the absence of an explicit con-
straint in the electric market to limit electric generation to
support limited gas supply, CAISO grid operators would
need to resort to manual dispatches on potentially all gen-
erators that receive service from SoCalGas to reduce their
production and be within the gas limit without regards
to the impact on the electric operation or transmission
constraints. This imposed additional costs on the system
and was not as efficient as an explicit market mechanism.
Furthermore, CAISO seeks to utilize a market constraint
to manage the gas constraint instead of using manual
dispatches to alleviate an untenable situation for elec-
tric operators. Without an explicit constraint modeled
in the market that could drive for an optimal dispatch of
34 ieee power & energy magazine
resources, operators will need to manually manage the
constraint on a 5-min basis, requiring the adjustment of
potentially dozens of generators.
Being able to enforce a gas constraint provides CAISO
with an effective tool to accomplish the necessary redispatch
to limit the gas usage on the electric system when it needs
to avoid further stressing the already-constrained gas system
while continuing to reliably and efficiently serve the load.
In 2016, CAISO started using a maximum gas constraint
in both day-ahead and real-time electric markets. This is
a maximum gas-burn constraint that coordinates the gas
limitation with the electric dispatches for generators that
rely on such a limited gas supply. This gas constraint is
analogous to a transmission constraint. It has an upper limit
that defines the maximum gas-burn enforcement. The con-
straint has a predefined set of resources subject to the gas
limitation. This nomogram limits the dispatch of generators
in the affected areas to comply with maximum gas usage.
The constraints also limit the CAISO market dispatch of the
affected generators in the real-time market to a maximum
gas usage if there is a limitation that relates to differences
between the gas scheduled with the gas company and the
gas consumed during the operating day due to gas system
imbalance limitations.
The gas constraint definition accounts for the different
generators’ efficiencies through the heat rates. This allows
the electric system operator to attain more efficient dis-
patches by first reducing the least-efficient genera-
tors in terms of gas usage. This is information that the
market already uses in the determination of generators’
energy costs.
Analogous to a transmission constraint, the enforcement
of the gas limitation is equivalent to congestion management.
The constraint is priced in the electric market and results
in price signals to generators that are part of the restriction
definition. These price signals are consistent and incentivize
generators’ decremental dispatches.
When the market enforces a gas constraint concurrent
with the enforcement of all transmission constraints, a nat-
ural interaction between electric and gas restrictions can
occur. Resources that are part of the gas limitation can also
be effective for transmission congestion management. In
some instances, the management of transmission constraints
will require certain resources to increase production while,
at the same time, they may be part of the gas constraint
that requires limiting or decreasing production. Under such
competing generation movements, the market will effi-
ciently determine the most optimal dispatch that complies
with both constraint limitations. This also has a secondary
effect that, to comply with the gas limitation, congestion
on the electric transmission system can sometimes increase.
In some cases, this creates power flows from Northern to
Southern California, which can congest certain transmission
elements in between and result in congestion that otherwise
would not occur.
march/april 2021
 Effectively, resources inside the region of the gas con-
straint will decrease production, which is the intended
outcome to impose a gas limitation. Another effect of the
gas constraint is that it will naturally limit the system-
supply availability and cause higher energy prices in the
overall system. This has the effect of requiring generation
from the rest of the system to increase production. All
of these effects are a result of an efficient dispatch that
properly prices the additional limitations imposed in the
electric market. The maximum gas constraint produces
efficient prices for affected generators because if the
maximum gas constraint is binding, the price decreases
for these resources. This pricing logic is the same for how
a generator behind a transmission constraint is priced
lower to reflect the congestion cost associated with dis-
patching that supply.
Under certain conditions, it may not be feasible to com-
ply with the enforced limitation on a given constraint in
CAISO’s market. Generally, infeasibility occurs when there
is an insufficient supply or short-term ramping capability
to be utilized in the management of the constraint. For
such cases, CAISO provisions allow for constraints to be
relaxed up to the point needed. This infeasibility is short
term in nature and set at administrative prices defined above
any economical bid value.
With the participation of virtual bids in the day-ahead
market, the current rules utilize prices with the gas-con-
straint effect to settle only physical gas-fired generators that
are within the gas constraint. Virtual bids or congestion rev-
enue rights do not settle on those prices to eliminate the pos-
sibility of any adverse impact. For instance, a virtual trader
could take a position with a location exposed to the gas-con-
straint price and an offset position at another location unaf-
fected by the gas constraint to profit from the gas-constraint
binding between the day-ahead and real-time markets.
Use of Gas Constraints
CAISO’s markets have the functionality to enforce gas con-
straints. CAISO enforces gas con-
straints in the day-ahead market,
the real-time market, or both to
14
limit the gas burn in specific ar-
eas previously defined by CAISO Coastal East of Moreno Valley Inland
12
as deemed necessa r y. T hese Los Angeles Basin
Gas Burn (MMSCFH)
10
constraints come in the form of
potential risks of curtailments or 8
as requests for conservation to
6
noncore customers from the gas
company. CAISO enforces gas 4
constraints only after a full as-
sessment of gas and electric con- 2
ditions and in coordination with 0
the impacted gas company. This 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
ensures that limitations imposed
by the constraint in the market are figure 5. The gas usage estimated with the day-ahead solution.
march/april 2021
consistent with the limitations observed on the gas system.
The general process to apply a gas constraint in the electric
market includes the following steps:
1) In coordination with the affected gas company, CAISO esti-
mates and sets hourly limits to the gas burn by genera-
tors in the impacted region. In some cases, the limita-
tion is set as a fixed hourly limit. In others, it can be set
as a daily quota. In this latter instance, CAISO creates
an hourly profile that meets the daily limit with a shape
based on gross demand patterns, enforcing lower limits
in the middle of the day, when the load is low, and im-
posing the highest limits for peak hours, when the load
is highest. CAISO is currently expanding the options
to shape the hourly limitations based on a historical gas
burn or using net load patterns. Figure 5 illustrates the
projected gas burn in May 2020 based on CAISO’s en-
ergy schedules from the day-ahead process.
2) CAISO verifies with the impacted gas company
whether these hourly limits address the gas limitation
and do not further worsen the gas system constraint.
3) Based on the specifics of the condition, CAISO may
enforce the gas constraints in both the day-ahead and
the real-time markets as needed to ensure that the dis-
patches and commitments from the CAISO market
do not exceed the maximum gas burn in the specified
region. Since conditions can change rapidly, if CAISO
becomes aware of the need for a gas constraint after
the day-ahead market run, it may enforce the con-
straint in only the real-time market run since it can
still help manage and coordinate generation with gas
conditions in the impacted area. Depending on the
commitments derived from the day-ahead market,
when gas constraints are enforced only in real time,
the market may have reduced flexibility to fully com-
ply with the gas limitation, such as the case where all
units in the impacted region are already at their mini-
mum operating points, and temporal constraint may
prevent turning them offline.
San Diego Area
ieee power & energy magazine 35
 4) If CAISO enforced a gas constraint in the day-ahead
market and it subsequently learns from the gas com-
pany that the limitation is lifted, the company will no
longer enforce the constraint in the real-time market
but still settle the day-ahead market with the impact
of the constraint.
5) Similarly, CAISO may proactively enforce gas con-
straints if it anticipates that large imbalances between
gas schedules and the gas consumed could compro-
mise electric or gas system reliability.
6) CAISO has the flexibility to adjust the hourly limita-
tions of the gas constraints to no longer enforce these
limitations as conditions change or if it determines
that the enforcement of the constraint adversely im-
pacts the market.
Figure 6 illustrates the number of days in which the elec-
tric market enforced the gas constraint across the different
markets. Overall, the frequency of enforcing gas constraints
has been very low as expected since this only happens when
limited gas supply conditions arise.
The effectiveness of enforcing a gas constraint in the market
to limit gas use in the system greatly depends on how accurate
the limit estimates are and on how timely they are enforced.
As part of the overall electric-gas coordination improvement
with Aliso Canyon, CAISO and the impacted gas company
have close coordination to learn about gas and electric system
conditions. As part of the day-ahead process, CAISO gener-
ates and posts the expected gas usages calculated in the two-
days-in-advance market solution. This happens two days in
advance of the trading date and provides the first opportunity
to resources and gas companies to learn about the projected
gas usage from the gas-fired generators in the CAISO system.
Based on that projection, gas companies study gas conditions
to assess if such a level of gas usage can pose a challenge, and
they assess mitigating actions that can go back to ISO as maxi-
mum gas quotas are enforced in the corresponding markets.
Since the first effort to mitigate risk related to Aliso Can-
yon restrictions, CAISO has undertaken other initiatives
30
25
Number of Days
20
15
10
5 NewsUpdates/2020/Summer2020-ReliabilityAssessment_
0 Final.pdf
2016 2017 2018 2019
Year
Day Ahead Predispatch Dispatch
figure 6. The number of days with gas constraints en-
forced in the electric market.
36 ieee power & energy magazine
to ensure generators have the market mechanism to reflect
the changing gas conditions in their market bids and costs.
Some of them have been required to redesign entire portions
of market functionality in CAISO’s markets. This will also
allow resources to sufficiently recover their costs under gas-
restricted conditions.
Final Remarks
The transition toward a decarbonized grid presents unique
challenges in the management and operation of electric sys-
tems. The resource mix is quickly evolving with growing
penetrations of variable generation resources, such as the
addition of more renewable resources and the electrification
of transportation. This type of generation introduces large
ramping characteristics that conventional generation, such as
gas-fired resources, will need to help absorb. Gas resources
still represent a significant share of supply in CAISO’s sys-
tem, and efficient dispatch has direct implications on the
overall market performance and pricing. In recent years,
electric-gas coordination has become critical to ensure reli-
ability in both the gas and electric systems when gas supply
may be restricted in specific regions of the system. CAISO’s
market includes maximum gas-usage constraints in the mar-
ket clearing process concurrently with transmission conges-
tion management. This achieves efficient generation dispatch
by sending the right price signals to gas-fired resources. As
the grid continues to evolve with more and different types
of generation resources, other operational challenges may
come into place. The electric system will need to have new
and enhanced market mechanisms to properly manage and
price such conditions.
For Further Reading
“Stakeholder initiative for Aliso Canyon gas–electric
coordination,” Caiso.com. [Online]. Available: https://
stakeholdercenter.caiso.com/StakeholderInitiatives/Aliso
-Canyon-gas-electric-coordination
California Public Utilities Commission. Accessed May
2020. [Online]. Available: www.cpuc.ca.gov
“2020 summer load and resources assessment,” CAISO,
Folsom, CA, May 15, 2020. [Online]. Available: http://www
.caiso.com/Documents/2020SummerLoadsandResources
Assessment.pdf
“Summer 2020 Southern California Reliability Assess-
ment,” California Public Utilities Commission, San Fran-
cisco, Apr. 15, 2020. Available at https://www.cpuc.ca.gov/
uploadedFiles/CPUCWebsite/Content / News_Room /
Biography
Guillermo Bautista Alderete is with the California In-
dependent System Operator, Folsom, California, 95630,
USA.
p&e
march/april 2021
The Impact
of Renewables
on Operational
Security By Ivan M. Dudurych

Operating Power
Systems That Have
Extremely High
Penetrations of
Nonsynchronous
Renewable Sources

F
FOR DECADES, ELECTRICITY TRANSMIS-
sion systems were planned to operate within
the scope of active power transfers that, by their
nature, were fairly well defined and limited in num-
ber, size, and direction. However, in recent years, this
situation has been dramatically changing in front of our
eyes as a result of two main developments. The first con-
cerns new types of renewable power generation (i.e., wind,
solar, tidal wave, and so on) and their increasing share of the
generation portfolio. The second relates to the introduction AS
TO
CK

/J EM
C OM
CK.
ST O
TTER
©SHU
Digital Object Identifier 10.1109/MPE.2020.3043614
Date of current version: 19 February 2021

march/april 2021 1540-7977/21©2021IEEE ieee power & energy magazine 37

 A large interconnected power transmission system
is probably the single most complex engineering device
to be created by mankind.
and evolution of electricity markets. These factors contrib-
ute to fundamental changes in generation patterns and power
transfers in ways that were not anticipated.
This article focuses on the impacts of renewables on
operational security. Since renewable generation is inher-
ently intermittent due to the countless combinations of
loads, weather, and electricity market conditions, the variety
of generation patterns and transfers is virtually unlimited.
Therefore, base case and contingency case scenarios may
significantly vary on a minute-to-minute basis in terms of
the following:
✔ online generation portfolios
✔ power flows
✔ voltage profiles
✔ parameters of power system dynamics.
Thus, to adequately support the operational security of a
power system, offline analyses and studies should be com-
plemented by online, close-to-real-time evaluations. While
attempting to cover all possible scenarios, offline studies
produce estimated security limits that can be extremely con-
servative for one setting and insufficient for others. Trans-
mission system operators (TSOs) whose networks have a
high penetration of renewable generation face several tech-
nical challenges, including the following:
✔ Variability: The output of most types of renewable
generation changes in time frames measured from
seconds to hours.
✔ Uncertainty: Variable generation is less predictable
than conventional generation.
✔ Location: Renewable sources of energy are often lo-
cated in remote areas that have small or no local loads,
and they can even be situated offshore (e.g., wind and
wave generation). This requires long transmission
paths to deliver power to load centers.
✔ Nonsynchronicity: Renewable generation normally
uses convertor-based technologies to couple with the
grid. As a consequence, it does not inherently provide
an inertial contribution, which is crucial for synchro-
nous power system stability.
✔ Service capabilities: There is a lack of the traditional
system service capabilities that conventional genera-
tors normally have, such as
• sufficient voltage and frequency regulation
• an operational (rotating) reserve
• adequate fault ride-through (e.g., the slow active
power recovery of a wind generator following a
transmission fault clearance).
38 ieee power & energy magazine
✔ Complexity: Additional complexity is introduced into
the system configuration and operation through the
necessity of solving special problems, such as har-
monics, subsynchronous resonances, interactions with
other control systems, and the performance of special
protection systems.
Therefore, it is critical to implement additional measures
to ensure the security of a power system under all condi-
tions, especially when a high level of renewable and other
unconventional generation is integrated. In recognition of
that fact, the European Network of Transmission System
Operators for Electricity (ENTSO-e) included a require-
ment in “System Operation Guideline on Dynamic Stability
Monitoring and Assessment” (European Commission regu-
lation 2017/1485). According to Article 38, “each TSO shall
perform a dynamic stability assessment at least once a year
to identify stability limits and possible stability problems in
its transmission system. All TSOs of each synchronous area
shall coordinate the dynamic stability assessments, which
shall cover all or parts of the synchronous area.” However,
the pace of penetration of renewables and other new technol-
ogies into power systems is such that the minimum require-
ment set by the regulation is insufficient, especially when
saturation levels reach 70–80% (i.e., 70–80% of the total
generation fleet in a synchronous zone is nonconventional).
One response to these challenges can be to use online, “real-
time” dynamic security assessment technology.
The technology includes three main components. The
first is a digital model of a power system that represents the
real network at selected points in time; this includes the net-
work topology, operating conditions, and all other relevant
factors. The second is a system that, utilizing the real-time
digital model, runs millions of simulations on a digital
model of a power system (contingency analyses). The third
is an analytical tool that produces system security and stabil-
ity assessments based on the simulations. These assessments
can be made for any selected point in time—the present and
the future—and they will be used by system operators to
maintain operational security.
In the following, we will answer these questions:
✔ What is the operational security of a power transmis-
sion system?
✔ How is it affected by significant penetrations of re-
newable and nonconventional generation?
✔ How does online dynamic security assessment help
manage ever-increasing levels of renewable and non-
conventional generation in transmission systems?
march/april 2021
 Power system safety means that risks to life and property
associated with generation, transmission, and distribution
are satisfactorily mitigated.
The answers will be complemented by the experience that
TSOs have gained in the single interconnected synchronous
system of Ireland and Northern Ireland, which connects to
the United Kingdom through 500-MW high-voltage (HV)
dc in the north and 500-MW HVdc in the south. As of May
2018, the system’s generation fleet had some 9,500 MW
of conventional resources, 4,500 MW of wind resources,
a peak demand of 6,500 MW, and a minimum demand of
2,500 MW (see “First DSA Stakeholder Workshop” in the
“For Further Reading” section). The relatively small size of
the island’s synchronous system versus the increasing pen-
etration of intermittent renewables creates a “large-scale
renewable energy penetration laboratory.” Insights from
this “laboratory” will be useful for larger systems (such as
Scandinavia’s) and eventually for the European continental
synchronous system to prepare for the levels of intermittent
renewable generation penetration seen in Ireland.
Power Transmission System Operational
Security and Its Components
A large interconnected power transmission system is proba-
bly the single most complex engineering device to be created
by mankind. In many parts of the world, power transmission
systems operate for many decades, continuously supplying
electrical power to millions of customers. The goal of power
system operation is not only to reliably supply power but to
do so safely and securely. Power system safety means that
risks to life and property associated with generation, trans-
mission, and distribution are satisfactorily mitigated through
proper design, installation, maintenance, and organizational
measures. On the other hand, the security of the power sys-
tem means that the risk of customer supply interruptions is
adequately mitigated by proper planning, design, installation,
maintenance, and operation. The latter component refers to
operational security, which is the ability to retain a normal
state or return to such a state as soon as possible following a
contingency. The goal of managing operational security is to
keep a system in its normal state 99.99% of the time. Except
for the normal state, there are four others that the power sys-
tem could experience or transition through if system security
has been compromised and if exceptional external events
(e.g., extreme weather conditions) have taken place. Figure 1
represents the possible power transmission system states.
Power Transmission System States
In any modern energy management/supervisory control
and data acquisition system, power transmission network
march/april 2021
operational states are modeled by three sets of equa-
tions, one differential and two algebraic. The differen-
tial set represents the dynamic behavior of a system. The
first algebraic set consists of equality constraints, which
refer to the generation–load system balance. The second
set contains inequality constraints corresponding to the
operational limits of a system. These limits include power
transfer caps, voltage upper and lower operating restric-
tions, frequency lower and upper operating bounds, and
other controls that may be imposed to ensure operational
security. The differences between transmission states are
determined by the validity of the constraints of sets of
algebraic equations.
The Normal State
In this state, all constraints are satisfied, which means that
the balance between the generation and the load (plus losses)
is met, there is no overloaded equipment, and voltages and
frequencies are within normal operational limits. Also, a sys-
tem’s reserve margins are sufficient, and a system can with-
stand any of the credible contingencies so that it will remain
within its operational security bounds and be brought back
to normal operation following remedial actions. The set of
credible contingencies is defined by and used for the assess-
ment of power system operational security. For example, one
credible contingency concerns the trip of a large generator.
If such a trip causes a system frequency to deviate within the
security limits and if the system frequency returns to normal
operational limits after a remedial action (e.g., activating a
Normal State
Restorative
(Black Start)
Alert State
Blackout Emergency
(Red Alert)
System Not Intact
System Intact
figure 1. Power transmission system states.
ieee power & energy magazine 39
 rotating reserve), the system is considered to be in a normal
state in terms of its frequency security.
The Alert State
When a system’s security level is weakened and the prob-
ability of a disturbance increases, a credible contingency
would cause the system to reach a state where some net-
work equipment would be overloaded or voltages at key
bus bars or system frequencies would exceed the limits. In
this state, all constraints are satisfied, but there are no suf-
ficient reserve margins to make sure that parameters will
be satisfied when a severe contingency occurs. Therefore,
preventive actions can be taken to restore the system to
its normal state. The ENTSO-e Operational Security Code
defines the deficiency as being 20% or more of the fast
frequency and primary frequency reserves for 30 min or
longer and says that at least one of the credible contingen-
cies can lead to an operational limit violation even after
the effects of remedial actions. In this state, an amber alert
is sent to system participants.
The Emergency State
A system can move from an alert state to an emergency state
if a severe contingency occurs before any preventive action
is taken. In this state, some inequality constraints will be
violated, which means that certain network components
are overloaded or that voltages at key bus bars or system
frequencies are outside their limits. Nevertheless, the sys-
tem remains intact and can be restored to its normal state
or at least to its alert state if appropriate corrective actions
are taken. ENTSO-e adds that an emergency state can be
declared if at least one measure of the system defense plan is
activated and if there is a complete loss of system monitoring
and control tools and facilities (energy management system/
Operational Security
Steady State Dynamic
Short Circuit Synchronous (Rotor
Current Level Angle Stability)
Thermal Nonsynchronous
(Fault Ride-Through)
Voltage Voltage Stability
Frequency Frequency Stability
Reserves Oscillatory Stability
Ramping
Stability Limits
Static Limits
figure 2. Operational security components.
40 ieee power & energy magazine
supervisory control and data acquisition, state estimators,
and other online system analysis and monitoring tools). In
this state, a red alert is sent to system participants. A system
defense plan includes the automatic disconnection of part of
the load (load shedding) in case more than one big generator
trips, the projected operational active power reserve is not
been activated, or both.
The Blackout State
A system can move to the blackout state from the emer-
gency state if no appropriate corrective action is taken, and
it can move from the alert state to the blackout state when a
severe disturbance occurs. In the blackout state, equality and
inequality constraints are violated. Many generators may
lose synchronism, leading to cascading transmission system
component outages. A system may split into islands, and sig-
nificant parts of the transmission network may shut down.
Control actions, such as load shedding and system separa-
tion, are used for saving as much of the network as possible
from a widespread blackout. ENTSO-e defines this state as
a loss of more than 50% of the load in a TSO’s responsibil-
ity area and as the total absence of voltage in an island for
at least 3 min. A blue alert is sent to system participants,
a black-start restoration plan immediately is triggered, and
the system enters the restorative, or black-start, state. A res-
toration plan provides steps to restart the system following
a blackout. Such steps include activating a black-start gen-
erator, sequentially energizing a black-start path, gradually
picking up the load, creating frequency islands, synchroniz-
ing the islands, and restoring the load.
The Restorative (Black-Start) State
In this state, an operator performs control actions to recon-
nect facilities and restore the system load. These actions are
performed according to the black-start restoration plan to
gradually bring frequency, voltage, and other parameters
within operational security limits. A system can reach the
normal state or the alert state, depending on conditions.
Operational Security Definitions and Criteria
Operational security can be defined in terms of steady-state
and dynamic security criteria. Steady-state criteria are also
called static limits and refer to thermal, short circuit current,
voltage, and frequency bounds. Active power and ramping
reserves are included. Dynamic criteria, on the other hand,
are also known as stability limits and refer to transient sta-
bility (the rotor angle for a synchronous plant and the fault
ride-through for a nonsynchronous plant), voltage stability,
and frequency stability. We also include oscillatory stabil-
ity, which describes the ability of a power system to damp
oscillations. It deals with small-signal stability and cases of
transient and frequency stability, where disturbances could
result in unstable oscillations of active powers, voltages, and
frequencies. A proposed structure for operational security
components is shown in Figure 2.
march/april 2021
 Power system stability can be assessed by considering its
various forms, which include the following:
✔ Steady-state stability: This is the ability of a system
to arrive at a safe steady state without disconnecting
customers following the outage of one (N–1) or more
(N–k) generator, transmission line, transformer, and
load components. This includes the following:
• Thermal constraints: These involve the requirement
to maintain power flows in the elements of a system,
below their thermal ratings in all N–1 contingencies
and credible N–k contingencies; these ratings can be
variable, e.g., seasonal or so-called dynamic, where-
by they change more frequently (e.g., every hour) de-
pending on the load and weather conditions.
• Steady-state voltage stability: The is the ability of a
system to maintain secure voltage levels and voltage
gradients during all N–1 contingencies and credible
N–k contingencies.
✔ Transient stability: This concerns the ability of a sys-
tem to keep its components connected following large
disturbances, such as faults in the transmission system
that are cleared by protection, sudden trips of large
generators, the disconnection of large loads, and so
forth. Transient stability further includes
• Synchronous plant stability: This describes this
ability of a system to maintain its synchronism (ro-
tor angle stability).
• Nonsynchronous plant stability: This is the abil-
ity of a system to keep other sources of generation
connected. It is also called fault ride-through ca-
pability.
✔ Oscillatory stability: This is the ability of a system
to quickly damp out the power and frequency oscil-
lations that can occur in the power generation. Oscil-
lations can be triggered after small and large distur-
bances that cause variations in the load, generation
(especially wind, which has intrinsic variability), au-
tomatic control device settings, and so on. Oscillatory
stability includes the following:
• Small signal stability: This is the ability of a sys-
tem to maintain the synchronism of its synchronous
plants and keep its nonsynchronous plants connect-
ed following small disturbances.
• Control mode stability: This describes the ability of
a system to dampen oscillations related to turbines
and their controls.
✔ Frequency stability: This is the ability of a system to
maintain its frequency within safe limits and to
return to the permitted frequency range following
disturbances.
Each of the operational security components can be
assessed by a set of security criteria applied to analyze the
results of simulations based on a power system model and a
scenario containing particular conditions and disturbances.
If the modeled system performance parameters are within
march/april 2021
an established set of security criteria, the scenario is consid-
ered secure. If at least one security criterion is violated, the
scenario is considered unsecure, and corrective measures
will need to be taken. A version of a set of operational secu-
rity criteria is proposed as follows:
✔ Thermal criterion: Transmission lines, cables, and
transformers must be loaded at less than 100% of their
nominal thermal ratings in base case conditions and
at less than 110% in N–1 contingency conditions. For
transformers, which have a greater short-term load ca-
pacity, 130% may be acceptable.
✔ Voltage stability criterion: Voltage stability must be
maintained in base-case conditions and all studied
contingency conditions. Steady-state voltage levels
at all HV buses within the transmission system (e.g.,
110 kV and higher) must be within limits: 0.95–1.1 per
unit in base case conditions and 0.9–1.12 per unit in
N–1 contingency conditions.
✔ Synchronous plant transient stability criterion: The
maximum rotor angles of synchronous generators
must be within limits to stay in synchronism. This
corresponds to a requirement that the maximum angle
separation of any two synchronous generators in one
island must be less than 360º.
✔ Frequency stability criterion: In the event of tripping
one generator or a three-phase fault with a forced out-
age of an element (transmission line or cable), the fre-
quency deviation must be lower than 1 Hz when the
duration of the frequency deviation is longer than 0.5 s.
✔ Oscillatory stability criterion: The oscillation damp-
ing coefficient can be assessed through small-signal
analysis and as an additional outcome of the time do-
main simulation analysis. The damping factor and the
exponential decay rate of an oscillation following an
event are related, and the larger the damping coeffi-
cient is, the higher the damping factor will be. The
damping coefficient must be greater than 0.05.
✔ Additional criteria:
• System nonsynchronous penetration (SNSP) level:
This concept is valid for a synchronous system as
a whole, and it was developed in parallel with the
“All Island TSO Facilitation of Renewables Stud-
ies” (see Eirgrid 2010 in the “For Further Reading”
section). The objective of the SNSP metric is to
have an index that, at a high level, captures a range
of operational issues. It shows a percentage of the
nonsynchronous generation (including net imports
via the HVdc interconnectors, if any) in the total
generation in a synchronous power system. HVdc
interconnectors are included since, when they are
importing power, they displace generation located
in the system that would otherwise be operating.
At a constant demand, an increase in nonsynchro-
nous generation would equate to an increase in the
SNSP. The maximum level of the SNSP can be
ieee power & energy magazine 41
 Existing transmission and distribution networks
were designed using planning criteria based on a limited
number of power flow patterns.

established based on the comprehensive analysis of
the operational security of the synchronous system
and revised in line with improving the flexibility of
the system components (mainly generators’ ability
to provide inertial and fast frequency response). At
present, the maximum SNSP limit for the synchro-
nous system in Ireland and Northern Ireland is set
as 65% (see “Operational Constraints Update” in
the “For Further Reading” section). The maximum
SNSP is maintained by the curtailment of nonsyn-
chronous renewable generation.
• Rate of change of frequency (RoCoF): This is the
measure of the frequency through time following the
sudden disconnection of a generator or an intercon-
nector. The most likely time for the RoCoF to exceed
its limit occurs during the initial moments after an
event. To this end, the RoCoF is derived from the
power system swing equation and can be calculated
as the observed change in the system power mul-
tiplied by the nominal frequency and divided by
twice the postevent system inertia. At present, the
minimum and maximum RoCoF are monitored. The
minimum RoCoF could occur due to the trip of a
large generator or importing interconnector, and the
maximum RoCoF could result from the trip of an
exporting interconnector.
• System inertia: This criterion is valid for a synchro-
nous system as a whole, but it can be used for part of
a larger synchronous system that is planned to oper-
ate as a separate synchronous system during a sys-
70
60
50
40
(%)
tem separation (e.g., following a severe frequency in-
cident). Monitoring system inertia from synchronous
generation ensures that operators always maintain a
minimum number of conventional generating units
synchronized to the power system. At present, the
minimum level of inertia for the Ireland and North-
ern Ireland system is 23,000 MW (see “Operational
Constraints Update” in the “For Further Reading”
section). It can be provided by a minimum number of
conventional generating units to maintain
— enough inertia to support frequency control and
stability
— sufficient synchronous torque to maintain the sys-
tem strength and transient stability
— reactive power support to manage the voltage con-
trol and stability; for that purpose, units should
be selected in voltage-critical parts of the system.
• Active power operational reserve: This is a classi-
cal requirement for a rotating reserve that can be
complemented by new technologies (e.g., batteries).
• Available megawatt ramping: This is an assessment
of the total available ramping capability of an online
plant and an offline plant across various time hori-
zons (e.g., 3 h, 8 h, and so on).
Impact of an Extremely High
Penetration of Renewable Sources
on Operational Security
The Effects of Resource Variability
Renewable resource variability has a profound impact on
operational security. Figure 3 displays the percentage of
nonsynchronous generation (mostly wind) on the island of
Ireland. The SNSP is shown in the percentage of the total
generation for April–June 2019. It is seen that such genera-
tion is extremely volatile and changes randomly from almost
zero to 65%. This means that the scheduling of the conven-

30
tional plant will be significantly different from day to day
and hour to hour, causing the plant to cycle and even stop
20
and start much more often than during the past. More fre-
10 quent cycling and operation outside the plant’s most efficient
0 range (which is normally near the plant’s full rated output
11 April

21 April

11 May

21 May

31 May

10 June

20 June

30 June
1 April

1 May

minus the reserve provision) reduces efficiency and leads to

increased wear, and thus it can reduce reliability. It is worth
mentioning that Europe is working to improve the flexibility
figure 3. The percentage of the SNSP in Ireland’s total of existing and new gas and steam turbines and extend the
generation in April–June 2019. (Source: Eirgrid.) equipment’s lifecycle to meet the requirements of grids that

42 ieee power & energy magazine march/april 2021

have high levels of renewables. One example is the European
Union Flexturbine project (2016–2018).
Existing transmission and distribution networks were
designed using planning criteria based on a limited number
of power flow patterns. These patterns were derived from the
fundamental paradigm of unidirectional power flow from large
generators through the transmission system to the distribution
system. The latter has been considered “passive” and was
designed to convey electricity from bulk transmission system
supply points to loads. With the advent of renewable and other
distributed generation, this is no longer the case. Transmis-
sion and distribution systems now experience multidirectional
power flows, depending on the amount of instant renewable
generation in different parts of the system. The power flow
magnitude can vary, and so can its direction. When wind gen-
eration is low, power flows from the transmission system to the
distribution system, while during periods of high wind gen-
eration, energy may flow in the opposite direction, from the
distribution system to the transmission system.
It is well known that conventional power stations have been
built close to load centers. In contrast, renewable generation is
typically located far from load centers. Therefore, for example,
if there is a lot of wind generation, power flows can reverse and
reach significant values that were not originally anticipated for
the transmission circuits. Such circuits can be highly loaded,
and they can overload if there are outages in parallel circuits.
Therefore, thermal security must be monitored in real time to
avoid cascading outages due to overload conditions.
Effect of Renewables on Transient Stability
As defined in the “Operational Security Definitions and
Criteria” section, transient stability has two components:
synchronous plant rotor angle stability (as a set of synchro-
nous machines connected to the system) and nonsynchronous
plant fault ride-through performance (as a set of converter-
based energy sources connected to the system). To qualita-
tively demonstrate how rotor angle stability is affected by a
significant penetration of renewables, we will use the power
angle characteristic of a synchronous machine, as illustrated
in Figure 4.
The electromagnetic power of the machine changes sinu-
soidally with the rotor angle. In a steady state, the electro-
magnetic power of the machine is balanced by the mechan-
ical power of the turbine (operating point A in Figure 4).
During a disturbance (power unbalance), the rotor angle
changes as the rotor speed deviates from the synchronous
speed. As a measure of rotor angle stability, we can use the
concept of “synchronizing torque.” It shows how much the
rotor angle changes for the same variation in the power from
the steady-state operating point of the machine. A higher
synchronizing torque gives more stability to the synchro-
nous plant because it encourages synchronous machines to
spin together at the same frequency. As the machine gets
more loaded (the operating point moves from A to B), the
synchronizing torque diminishes. This decrease contributes
march/april 2021
to a reduction in the transient stability of the machine to
a disturbance that perturbs the rotor angle. That is, as the
machine moves to the new operating point, the forces trying
to resynchronize it to that operating point will be weaker.
Similarly, if the maximum electromagnetic power of the
machine decreases (see the orange curve in Figure 4), it has
the effect of decreasing the synchronizing power and results
in a reduction in the transient stability of the machine.
Now, consider what initially happens when only a small
portion of the synchronous generation is replaced by con-
verter-based renewables. If the remaining synchronous
generators are less loaded than normal, their synchronizing
torque increases. This improves the transient stability of the
system. However, a further increase of renewables requires
a progressive decommitting of the conventional units. The
equivalent power angle characteristic will mirror the orange
curve in Figure 4, with operating at point C. The maximum
electromagnetic power of the synchronous machine will be
reduced because it will be more difficult to “push” power
from the remaining synchronous machines to the loads, due
to the increased reactance of the now longer paths and the
poorer voltage support from the now prevailing nonsyn-
chronous generation. As discussed previously, having less
electromagnetic power will cause a reduction of the syn-
chronizing torque, and at very high renewable penetration
levels, the transient stability deteriorates rapidly because
the remaining synchronous units are forced to be highly
loaded. Note, however, that recent advances in converter-
based grid connection applications based on insulated gate
bipolar transistor technology will enable engineers to coun-
teract these phenomena.
The “All Island TSO Facilitation of Renewables Study”
(see Eirgrid 2010 in the “For Further Reading” section)
estimated the reduction of the critical clearing time (the
maximum time allowed for protection to clear a three-phase
fault at the terminals of a synchronous generator before the
Increased Loading
B Maximum Electromagnetic Power
Power
A C
Normal Loading
Rotor Angle
figure 4. A power angle characteristic with normal and
increased loading (operating points A and B) and a reduced
power angle characteristic (operating point C).
ieee power & energy magazine 43
 generator starts slipping poles) with an increasing penetra-
tion of wind generation. Based on this estimation and the
fact that clearance time is inversely related to synchronizing
power, the latter can be presented as in Figure 5. It is seen
from the figure that the rotor angle stability of the system
slightly improves until the wind penetration increases by up
to 50%. Then, it starts declining, and it quickly deteriorates
at wind penetration levels of 80% and higher. Therefore,
monitoring the remaining synchronous plants’ transient
Synchronizing Power
0 20 40 60
Nonsynchronous Plant Penetration (%)
figure 5. The synchronizing power versus nonsynchronous
plant penetration level (derived from the critical clearance
time studies in “All Island TSO Facilitation of Renewables”;
see Eirgrid 2010 in the “For Further Reading” section).
V and/or reactive power at their terminals. The best existing
A technology for this task involves connecting renewables to the
(a)
V
A′
B′
(b)
figure 6. The idealized power voltage characteristic of a
transmission corridor between two parts of an ac power
system with (a) a low penetration and (b) high penetration
of renewables.
44 ieee power & energy magazine
security (i.e., rotor angle stability) in real time will be even
more critical with higher levels of renewable penetration.
To investigate how the transient security of nonsynchronous
plants (the ability to ride through a three-phase fault) evolves
with increasing renewable generation, consider how three-
phase fault characteristics evolve. The most economical way of
connecting renewable sources to a synchronous system is via
power electronic converters. This has two main implications:
the mechanical system inertia is increasingly depleted when
renewable penetration increases, and the short circuit current
contribution from renewable sources is limited to their maxi-
mum load current. The first implication will affect a system’s
frequency security; the second causes voltage dips associated
with faults in the transmission system to propagate more widely
within a system that has reduced short circuit levels. This has a
direct impact on the fault ride-through performance of nonsyn-
chronous sources of generation. The bigger part of the system
sees a voltage dip, and a larger proportion of the nonsynchro-
nous generation can trip on undervoltage protection.
Effect of Renewables on Voltage Stability
80 100 In the early days, renewable generators did not have any volt-
age and reactive power control capability. Soon, as renewable
nonsynchronous generators displaced conventional ones, it
was recognized that conventional generators’ voltage con-
trol capabilities were being depleted. Grid codes began to be
updated to introduce at least some conventional generation-
like functionality into a nonsynchronous generation portfolio.
For example, the Irish grid code requires that all transmis-
sion-connected wind power systems be able to control voltage
B
grid via a fully rated converter. This enables modulating
the grid-side, three-phase voltage according to the required
reactive power. However, most distribution-connected
renewable sources are still fixed-power-factor machines that
do not regulate voltage. This means that voltage regulation
P
45,000
40,000
Inertia (MW)
35,000
30,000
25,000
P
20,000
0 10 20 30 40 50 60 70
Wind Generation (%)
figure 7. System inertia reduction as the renewable pen-
etration rises in the single synchronous power system of
Ireland and Northern Ireland (Source: Eirgrid).
march/april 2021
becomes increasingly difficult in a system with a high penetra-
tion of renewables. In areas with significant renewable genera-
tion, voltage collapses become more probable.
Consider the ideal power voltage characteristics of a trans-
mission corridor between two parts of a system, with the
receiving part represented by the load in the form of imped-
ance. This characteristic is described in Figure 6(a). In a
properly designed ac power system, where all generation is
being conducted by synchronous machines, the distance from
the area of normal operation (point A) to the site of instabil-
ity (point B), the “nose” of the power voltage characteristic is
large enough to accommodate possible contingencies, such as
outages in parallel circuits. Historically, the thermal conven-
tional generation portfolio was planned to be placed as close
as possible to load centers; thus, large amounts of electricity
were not designed to travel long distances. At high levels of
renewable penetration, when a significant part of the thermal
portfolio is displaced, power needs to travel farther because
renewables are normally placed away from load centers.
Consider the same transmission corridor for a system
where the conventional generation near the load center is
scheduled to be offline so that it can be replaced by distant
renewable generation. The power transfer would change
directions, and its absolute value would increase. Thus, the
power-voltage characteristic would change due to the addi-
tional impedance associated with importing electricity from
remote areas, as in Figure 6(b). Operating point A now will
be closer to instability point B because of the greater imped-
ance and the increased transfer.
Effect of Renewables on Frequency Stability
The main effect of renewable generation on frequency sta-
bility is related to the fact that renewable generation provides
little inertia to a system because it is partially or completely
50.1
49.9
49.7
49.5
Frequency (Hz)
49.3
a
49.1
48.9
b ENTSO-e. in Proc. 1st DSA Stakeholder Workshop, Brus-
48.7
48.5
c
48.3
48.1
0 10 20 30 40 50
Time (s)
figure 8. Frequency deviations due to the sudden trip of
a large generator for (a) a traditional generation portfolio,
(b) a portfolio with a large nonsynchronous generation
penetration, and (c) a portfolio with a significant nonsyn-
chronous generation penetration.
march/april 2021
mechanically decoupled. With the progressive replacement
of conventional generation with nonsynchronous renewable
generation (Figure 7), the available system inertia decreases,
meaning that if a power unbalance occurs, the frequency
deviation for the same unbalance will be bigger in a system
that has less available inertia (Figure 8). To an extent, the
issue can be alleviated by the introduction of fast frequency
response mechanisms (some of which are called synthetic
inertia), but the genuine inertia provided by a synchronous
plant can be fully replicated only by a converter-based plant
if the power imbalance on the system is directly assessed
and compensated in real time. Curve C in Figure 8 corre-
sponds to the activation of an underfrequency load shedding
scheme. This is the system’s integrity defense mechanism
designed to automatically disconnect a percentage of the
load. This is a last defense, and it is not accepted for routine
operation in modern society.
Real-Time Assessment of
Operational Security
Assessing the operational system security of a power system
becomes compulsory for TSOs that integrate large amounts
of renewables. Such assessments can be performed by using
online dynamic security assessment technology: software that
takes snapshots of power system conditions, performs secu-
rity evaluations (including the determination of stability lim-
its) in near real time and for future time frames (from hours
to days), and provides warnings about, and remedial measures
for, abnormal situations. An example of an online dynamic
security assessment system can be found in the article “Safety
in Numbers” in the “For Further Reading” section.
For Further Reading
Eirgrid. All Island TSO Facilitation of Renewables Studies.
Dublin: Eirgrid, 2010.
I. M. Dudurych, “On-line Assessment of Secure Level of
Wind on the Irish Power System,” in Proc. 2010 IEEE PES
General Meeting, Providence, RI, July 25–29, 2010, pp. 1-7.
doi: 10.1109/PES.2010.5588059.
I. M. Dudurych, A. Rogers, R. Aherne, L. Wang, F.
Howell, and X. Lin, “Safety in numbers,” IEEE Power
& Energy Mag., pp. 62–70, Mar.-Apr. 2012. doi: 10.1109/
MPE.2011.2178283.
sels, May 23, 2018. [Online]. Available: https://docstore.entsoe
.eu/Documents/Events/2018/180523_DSA_SH_Workshop
-Joint_presentation_with_conclusions.pdf
Eirgrid, “Operational constraints update version
1.80,” July 31, 2020 [Online]. Available: https://www
.eirgridgroup.com/site-files/library/EirGrid/Operational
ConstraintsUpdateVersion1_96_July_2020.pdf
Biography
Ivan M. Dudurych is with the Military Technological Col-
lege, Muscat, 111, Oman. p&e
ieee power & energy magazine 45
T
THE DANISH GOVERNMENT HAS SET VERY
ambitious binding targets regarding decarboniza-
tion. By 2030, carbon dioxide emissions must be
reduced by 70% compared to the 1990 level. This
can be achieved primarily through a predomi-
nantly renewables-based electricity system and the
electrification of energy demand.
In 2019, Denmark covered 50% of its elec-
tricity demand primarily through wind and, to
a lesser extent, solar energy. The large increase
in the share of renewable energy production
in meeting the country’s electricity demand
is illustrated in Figure 1. In 2009, only 20%
of the demand was covered by wind and solar,
whereas 10 years later this figure stands at
50%. The electrification of the transportation
sector will accelerate through the wide-scale
adoption of electric vehicles (EVs).
To reach the 70% target, it is estimated that
1.5 million cars, corresponding to 50% of all cur-
rent cars in Denmark, must be electric or hybrid-
electric by 2030. In the heating sector, 70% of
natural-gas boilers and 95% of oil-fi red boilers
used for individual heating must be replaced with
renewables-based options. As a result, it is esti-
mated that 210,000 heat pumps must be installed
by 2030. These developments will lead to the

Energy Security
Through
Demand-Side
Flexibility The Case of Denmark

Digital Object Identifier 10.1109/MPE.2020.3043615

Date of current version: 19 February 2021

46 ieee power & energy magazine 1540-7977/21©2021IEEE march/april 2021

 doubling of electricity demand from 35 to
70 TWh by that year.
As a result, the Danish power system of the
future is expected to be characterized by large
shares of intermittent renewable sources and
increased energy demand. These developments
lead to significant challenges. The gradual
phaseout of conventional power plants results
in lower rotational inertia and more volatile
frequency dynamics. Additionally, renewable
generation is less predictable compared to the
energy produced by conventional power plants,
leading to an increase in the system’s balanc-
ing needs. Since these power plants have tradi-
tionally offered the bulk of power system flex-
ibility and the required ancillary services, new
sources of flexibility are needed.
Demand-side management and the flexibility
of distributed energy resources (DERs), such as
EVs, stationary batteries, and heat pumps, can
be used to maintain the current high level of
©SHUTTERSTOCK.COM/VLADGRIN
security of supply, even in a power system that
is dominated by volatile renewable production.
Recent technological advances and cost reduc-
tions in information and communication infra-
structure, sensors, and power electronics have
made DER control more economic and reli-
able. Commercial entities, called aggregators,
manage and control large numbers of DERs or
larger flexible loads to create new streams of
revenue by participating in the energy markets. DER charac-
teristics make them suitable for providing a variety of power
system services and participating in various energy markets,
such as the day-ahead, intraday, balancing, and ancillary
services markets.
The Danish project EcoGrid 2.0 demonstrated how
demand-side flexibility can be used concurrently on differ-
ent system levels by managing 800 flexible consumers who
offered services on both distribution and transmission lev-
els. However, the distributed nature and the large number of
these resources compared to the traditional, highly central-
ized structure of power systems pose several new challenges
in power system operation.
While these new sources of flexibility can benefit the
system and increase the security of supply, their mass inte-
gration requires careful consideration of many issues. One
By Jacob Østergaard, Charalampos Ziras,
Henrik W. Bindner, Jalal Kazempour,
Mattia Marinelli, Peter Markussen,
Signe Horn Rosted, and Jørgen S. Christensen
march/april 2021
is the need for coordination between the transmission and
distribution systems. Another is the heterogeneity of the
technical characteristics of these new resources that can
result in various barriers in their integration in power sys-
tem operation. Moreover, various reports and projects have
shown that the financial benefit to DER owners from offer-
ing flexibility to the power system is not sufficient for their
engagement. This means that motives that are not purely
financial must be found to convince owners to make their
flexibility accessible and/or use it for the benefit of the power
system. To overcome this, coordinated efforts to raise con-
sumer awareness and engagement are required. The emer-
gence of energy communities, the use of automation, reward
systems, user-friendly application programming interfaces,
and mobile applications along with governmental informa-
tion campaigns can contribute to engaging asset owners in
utilizing their flexibility.
To enhance the security of supply, DERs and demand-
side management must provide power system services
within the reliability, accuracy, quality, and response time
criteria set by the system operator. Therefore, the aggrega-
tors of those resources, which act as an interface to the sys-
tem and are responsible for service delivery, must ensure that
flexible resources meet those requirements. As presented
in the following section, a considerable research effort in
Denmark has focused on investigating control strategies,
monitoring techniques, communication issues, and aggrega-
tion and optimization models. As a result of those efforts,
many technical issues related to DER controllability have
been resolved, and the capability of DERs to offer power
system services has been demonstrated. However, the lack
of standardization in terms of prequalification, monitoring,
and validation for small and diverse resources is hindering
their potential.
An often overlooked aspect is that the functionalities of
the aggregators and not the individual resources should also
be prequalified, especially under extreme scenarios, i.e.,
information technology/communication failures, particu-
larly warm or cold weather conditions, and so on. It is often
the case that the same types of flexible units participate in
power markets of different and, in many cases, neighboring
countries under different terms. This negatively affects the
business opportunities of aggregators who manage interna-
tional portfolios and decreases the efficiency of the utiliza-
tion of flexible resources on a regional or European level. To
this end, there are many efforts toward standardization in
ieee power & energy magazine 47
 the regulation of European balancing markets and toward
the harmonization of monitoring and validation of services
offered by flexible resources.
The large increase in electricity demand and the utiliza-
tion of demand-side flexibility on a system level may also
have some unintended consequences on the distribution
system. First, the expected large numbers of new loads due
to the electrification of demand will stress distribution net-
works, and costly reinforcements will be required over the
next decade. The efficient use of the flexibility and control-
lability of these new loads on a distribution level can elimi-
nate or postpone the need for these reinforcements, with an
estimated socioeconomic benefit of US$3.2 billion in Den-
mark alone by 2030. This benefit increases to US$4.8 bil-
lion if reinforcements of the transmission system are also
considered. Second, the coordinated DER response (which
is installed at low or medium voltage levels) for offering sys-
tem services may lead to the overloading of lines or trans-
formers or to unacceptable voltage levels. For example, very
low spot prices at specific hours of the day may synchronize
consumption and result in congestion on certain low-voltage
feeders. The same could happen from the activation of bal-
ancing power that is offered by DERs or by the coordinated
DER response to a large frequency deviation when such
units are used as frequency-reserve providers.
DER participation in the energy markets creates a
need for cooperation between distribution system opera-
tors (DSOs) and the transmission system operator (TSO)
to avoid jeopardizing the safe operation of the power sys-
tem. On one hand, the TSO needs the flexibility and fast
response of DERs. On the other hand, the utilization of this
flexibility may cause operational problems to DSOs due to
load synchronization.
If an appropriate TSO-DSO coordination framework to
manage these challenges during the participation of DERs in
the energy markets does not exist, then energy security may
be compromised. If congestion at the distribution network
occurs, the DSO may resort to load curtailment, disrupting
60
Share of Electricity Supply (%)
Wind
50 Solar
40
30
20
10
0 business models.
09
10
11
12
13
14
15
16
17
18
19
20
20
20
20
20
20
20
20
20
20
20
Year
figure 1. The evolution of wind and solar energy produc-
tion in Denmark over the last decade.
48 ieee power & energy magazine
the energy supply of customers and causing a significant
reduction in the customers’ quality of service. As a result
of those DSO actions, aggregators will not be able to offer
the contracted services to the TSO, putting the whole sys-
tem at risk. In other words, the absence of an operational
framework may lead to both interruption of supply at the
distribution level and insufficient service delivery at the sys-
tem level. As we will show next, a substantial part of Danish
research effort has been directed to address such DER sys-
tem integration problems.
Danish Research Efforts on Demand-Side
Flexibility in the Past Decade
Denmark is a front runner in the green transition and has
long recognized the importance that research on demand-
side flexibility plays in the integration of large shares of vol-
atile renewable production. The vital role of research activi-
ties in the Danish energy system is reflected in the country’s
leading position in smart grid-related research in the Euro-
pean Union (EU). In this section, we provide an overview of
Danish research activities regarding demand-side flexibility.
Energiforskning.dk is a database where all Danish
research, development, and demonstration funding pro-
grams within energy and climate are registered. As of April
2020, 251 projects were listed under the “Smart Grid and
Systems” category. Of those, 166 were started after 2010,
had a total budget greater than US$235,000 each, and were
funded by the four major programs: the Energy Technology
Development and Demonstration Program, Innovation Fund
Denmark, ForskEl, and Green Labs DK. In our assessment,
only nationally funded projects were considered.
A large number of those projects have dealt with various
aspects of integrating flexibility and smart grid solutions in
the electricity sector. We have used four categories to clas-
sify these projects:
✔ flexibility—component level
✔ flexibility—aggregation level
✔ system integration
✔ market design and business models.
The first category refers to the investigation of issues
related to demand-side flexibility on a component level,
i.e., the control, monitoring, and modeling of flexible
loads. The second category refers to issues related to the
aggregation and control of large numbers of demand-side
flexible resources. The third category refers to projects that
have dealt with how flexibility and smart grid solutions can
be integrated into the power system, following a system-
level approach. The last category looks into novel market
mechanisms, the redesign of the electricity markets, and
A significant number of projects did not fall into one of
these categories, i.e., high-voltage direct current systems, dis-
trict heating, gas networks, or storage technology improve-
ment, and are outside the scope of our analysis. The total bud-
get of each project has been distributed equally throughout
march/april 2021
its duration. Additionally, we have mapped the contribution investigation of the safe and efficient system integration
of each project to the four categories, depending on its focus. of DERs is greatly strengthened by demonstrations in
The total funding per category and year is presented in real environments and networks. This leads to a need for
Figure 2. While all categories experienced a substantial larger experimental setups, involving DERs and distri-
increase in funding in the first years of the decade, their bution networks, the participation of DERs in the whole-
evolution in the subsequent years has not been the same. sale markets, and the investigation of mechanisms that
System integration and markets/business models are the two ensure not only the reliable provision of services on a
categories with an almost-continuous increase, whereas the system level but also the secure operation of distribu-
funding for the components and aggregation categories has tion networks.
been declining since 2016. There is a continuous decrease in For this reason, projects dealing with system integration
the budget of projects related to flexibility components and tend to have larger budgets. A scatter plot of the projects’
aggregation, which dropped from a combined share of 60% starting year against their budget is displayed in Figure 3.
in 2010 to 30% in 2019. This difference was compensated by Only six projects exceeded a total budget of US$6 million,
an increase in the share of projects related to system integra- and all of them focused primarily on system integration. The
tion and market design.
In the early 2010s, concepts and ideas related to the uti-
lization of DER flexibility were relatively new. Most of the
30
research work was devoted to various aspects of modeling,
control, and the monitoring of flexible loads, with a focus 25
on EVs and thermal loads. The goal was to demonstrate that
demand response could be a reliable provider of flexibility. US$ (millions)
20

This was done by determining the characteristics and influ- 15

ence of the operational constraints of flexible units so that they
could be used for balancing purposes and lead to an increase 10
of energy security in the power system. Thus, it is natural that 5
earlier projects focused more on technical problems and under-
standing the capabilities of demand-side flexibility as a service 0
provider rather than system-integration issues.
10

11

12

13

14

15

16

17

18

19
20

20

20

20

20

20

20

20

20

20
Over the years, early-stage technical problems related Year
to the control of DER flexibility in power systems have
largely been solved. The relevant technology has improved, Components System Integration
Aggregation Markets
and many commercial solutions are now available, offering
advanced monitoring, control, and management capabilities,
both on individual and aggregated levels. The funded proj- figure 2. The project funding per category and year of
ects and research in the field of smart grids and flexibility nationally funded Danish projects related to demand-side
in general have shown that the main challenges associated flexibility.
with the wide-scale utilization of
DER flexibility are technical sys-
tem integration and energy market
issues as well as viable business 25
models for end users, aggregators,
EnergyLab
and system operators. 20
iPower
Nordhavn
The system integration of
EcoGrid 2.0
US$ (millions)

demand-side flexibility covers

15
many issues, mainly associated Cities
with the increasing penetration System Services
10 From Small-Scale DERs
of DERs in the power system FED
and the safe utilization of flex-
5
ibility. Central to these prob-
lems is the secure operation of
0
distribution networks because 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
it is at this level where DERs Year
are installed, and it is through
the distribution network that figure 3. A scatter plot of the project budgets and starting years of nationally funded
services will be provided. The Danish projects related to demand-side flexibility.

march/april 2021 ieee power & energy magazine 49

 three largest projects during the past decade were iPower,
EcoGrid 2.0, and EnergyLab Nordhavn. In iPower, flexibility
modeling was combined with a coherent market approach at
the system and distribution level based on well-defined, flex-
ibility-based services. This was further developed, refined,
and demonstrated in EcoGrid 2.0, where a large number of
flexible units were controlled to provide services at either the
system or distribution level, based on an optimal bidding strat-
egy for the aggregators. In EnergyLab Nordhavn, the coordi-
nated operation of electricity and district heating systems was
investigated to exploit cross-sector flexibility, and integrated
multidomain markets were assessed.
An important function of research projects is that they
can provide valuable learnings for system operators. These
could refer to operational strategies, new market mecha-
nisms, and the provision of novel system services, among
others. This knowledge can provide insight and valuable
experience regarding the direction that power markets and
system operation should or should not take. For example,
EcoGrid 2.0 revealed the real-life challenges associated with
the use of commonly agreed upon baselines in the provi-
sion of DSO services in a local flexibility market and pointed
toward the direction of explicit operational limits as a more
viable solution.
Pilot projects are also particularly important tools for sys-
tem operators because they allow collaboration and knowl-
edge sharing between system operators and market partici-
pants by testing new business models and ideas. Energinet,
the Danish TSO, brings the experience from research to
operation with the use of such pilot projects, where small-
scale participation of new technologies or changes in mar-
ket regulation are tested in real-time operation together
with market participants. The obtained experience is used
for the fast market implementation of new technologies as
well as for the identification and, subsequently, reduction
of possible regulatory barriers. In the following section, we
focus on different system integration mechanisms that have
been proposed in various research projects and are currently
being investigated in Denmark to facilitate the utilization of
demand-side flexibility.
Facilitating the Use of Demand-Side
Flexibility in the Power System
As discussed in the previous section, a substantial number
of research projects and demonstrations have proven the
capabilities of DER and demand-side management to offer
power system services. Recently, the focus has been shifting
toward the investigation of mechanisms that can unlock this
flexibility so that it can be safely and efficiently integrated
into both system operation and markets. To this end, several
mechanisms have been proposed and are currently investi-
gated in Denmark, such as
✔ time of use (ToU) tariffs
✔ dynamic tariffs
✔ a finer geographical granularity of power markets
50 ieee power & energy magazine
✔ bilateral agreements
✔ local flexibility markets.
These mechanisms should not be seen as mutually exclu-
sive. Instead, more than one mechanism can coexist because
each of them has different characteristics, addresses vary-
ing problems, and has its own limitations. In the following
section, we describe the characteristics of each mechanism,
with a focus on a Danish application perspective. We assess
the mechanisms based on four criteria:
✔ effectiveness of congestion management: how effec-
tive each mechanism is in preventing and alleviating
congestion
✔ ease of implementation: how easy each mechanism is
to implement on a technical level
✔ market compatibility: how easy it is to integrate each
mechanism into the current power market structure
✔ impact on system balancing: the impact that each
mechanism has on the function of demand-side flex-
ibility as a balancing source of renewable production
and consumption at a system level.
These mechanisms can be implemented on transmission,
distribution, and feeder levels. By feeder, we refer to the
low-voltage branch departing from a 10/0.4-kV distribution
transformer, and distribution refers to voltage levels below
50–60 kV. For this reason, their effectiveness in conges-
tion management is assessed separately for each level. The
results are presented in Table 1.
ToU tariffs are probably the oldest and simplest way,
both in terms of implementation and market compatibil-
ity, to shift consumption away from peak hours. Such tar-
iffs were only recently introduced in Denmark. In their
conception, those tariffs were used by integrated system
operators to reduce peaks and shift consumption to off-
peak periods. Nowadays, ToU tariffs are set by DSOs such
that peak consumption is reduced to minimize the risk of
the overloading of lines or transformers or of unaccept-
able voltage levels. They are the same over relatively large
geographical areas (usually the same per DSO) and change
infrequently over the year. Such rigid tariffs cannot allevi-
ate transmission-level congestion, which is usually caused
by hard-to-predict events (e.g., faults, market actions, or
large variations in renewable production from the fore-
casted values).
The same is also true on a distribution and feeder level.
Such inflexible tariffs cannot cope with the different topolo-
gies and load compositions over the distribution network or
with the high shares of renewables that lead to unpredict-
able load patterns of flexible demand, which follows highly
volatile market prices. Furthermore, ToU tariffs cannot
cope with the possible congestion that may occur because
of the activation of balancing power or ancillary services,
which can happen at any point in time. Such tariffs can also
have an impact on the use of demand-side flexibility for
system-balancing purposes because they do not adapt to the
state of the system, and their inflexible nature may lead to
march/april 2021
cases where much-needed flexibility is too expensive due to
tariff costs.
It must be mentioned that ToU tariffs are calculated by
the DSO to induce the necessary motivation for consumers to
shift consumption away from peak hours. This DSO-driven
calculation does not consider the effect of ToU tariffs on the
use of flexibility, including the aggregators’ economic per-
formance in the wholesale markets, which may deteriorate.
The appropriate balance between the financial burden on
the various uses of flexibility in the wholesale markets and
the avoidance of distribution network operational problems
via network pricing can be more easily achieved through a
coherent market-based solution.
The disadvantages of ToU tariffs can be partially alle-
viated by the use of dynamic tariffs, which are calculated
at finer time resolutions, e.g., daily, and can be much more
granular geographically. Thus, they can be more effective in
preventing congestion at the distribution level. Their imple-
mentation is significantly more complicated than simple
ToU tariffs because advanced techniques are required to
calculate their optimal values. Similar to ToU tariffs, this
mechanism cannot avoid the possible congestion at the dis-
tribution level that occurs due to the activation of balancing
or ancillary services, and it is very challenging to implement
on a feeder level.
It should be noted that the effectiveness of dynamic tariffs
on congestion management can be significantly improved by
increasing the geographical and temporal granularity, e.g.,
by changing those tariffs hourly and per feeder. However,
this comes at the expense of technical complexity and regu-
latory obstacles (e.g., fairness).
Another relevant mechanism that has been proposed
is a finer geographical granularity of power markets. This
could be done by moving from a zonal pricing system
(Denmark has two bidding zones) to a nodal pricing system.
Such a change can be effective in alleviating transmission-
level congestion. However, nodal pricing would need to be
applied on a distribution or feeder level to be effective in
alleviating congestion at those voltage levels. This change
would require a complete restructuring of power markets
and system operation, whereas the implementation of
nodal pricing on a feeder level would add immense com-
plexity. Nodal pricing is not compliant with the current
legislation and regulations that govern the operation of
DSOs in Denmark.
Bilateral agreements are “limited grid access” agree-
ments between system operators and consumers. These
agreements allow the DSO to decrease the load of a con-
sumer in the event or prospect of congestion by granting
the DSO direct controllability over the flexible asset. In
exchange, consumers receive compensation in the form of
reduced or waived connection fees. Such agreements have
been established in Denmark for industrial customers, for
instance, large electric boilers where both the connection
fees and the impact on the distribution network are substan-
tial. The new initiatives, “limited grid access” and “tempo-
rary limited grid access,” aim at introducing reduced tariffs
for customers connected at the transmission network who
agree to act as interruptible loads.
Bilateral agreements are effective in alleviating conges-
tion at both transmission and distribution levels, have rela-
tively low regulatory or implementation barriers, and are not
difficult to be integrated into a power system. However, such
a solution has scalability limitations. DSO bilateral agree-
ments would be impractical at an individual DER level due
to complexity, high costs, privacy, and user comfort issues.
For these reasons, DSOs are directed toward market-based
solutions through the EU Clean Energy Package, which
states that “DSOs shall procure flexibility services in accor-
dance with transparent, nondiscriminatory and market-
based procedures.”
Market-based solutions are sought after for congestion
alleviation at a transmission level as well, mainly due to the
high renewable power infeed, as described next. Bilateral

table 1. An assessment of the different mechanisms for unlocking flexibility

in renewables-based power systems with flexible demand.
Criterion Effectiveness on Ease of Market Impact on System  
Mechanism Congestion Management Implementation Compatibility Balancing 
ToU Tariffs Low (T, D, F) High (T, D, F) High (T, D, F) Negative
Dynamic Tariffs Low (T) Medium (T, D) Medium (T, D, F) Negative
Medium (D, F) Low (F)
Finer Geographical High (T, D, F) Medium (T) Medium (T) Neutral
Granularity Low (D, F) Low (D, F)
Bilateral Agreements High (T, D, F) Medium (T, D) High (T, D, F) Negative
Low (F)
Local Flexibility High (T, D, F) Medium (T, D, F) High (T, D, F) Neutral
Markets
Key: T: transmission; D: distribution; F: feeder.

march/april 2021 ieee power & energy magazine 51

 agreements might limit the potential of using demand-side While all of the aforementioned mechanisms can coex-
flexibility for system-balancing purposes. If system opera- ist, some are more attractive depending on the characteris-
tors can limit the grid access of flexible resources when tics of a particular power system and the regulatory frame-
flexibility is needed on a system level, the alleviation of work. Research and pilot projects can provide valuable
congestion may be achieved at the expense of flexibility uti- insights and real-life experiences and learnings, guiding
lization for system balancing. However, limited grid access all of the relevant stakeholders toward the most appropri-
also results in significant savings for both asset owners and ate and beneficial solutions.
system operators (due to reduced network upgrade costs), so
a cost-benefit analysis is necessary to determine if an unlim- Energy Security and Marketplaces for
ited grid access results in socioeconomic benefits higher Flexibility Procurement: Some Real-Life
than the required connection fees. Experience from Denmark
Local flexibility markets are envisioned to operate in par-
The mechanisms presented in the previous section intend
allel to the existing power markets, with system operators to facilitate the use of demand-side flexibility and safely
being able to procure flexibility at specified locations of the
integrate DERs and demand-side management in modern
network, both at transmission and distribution levels. Such power systems. Since the early 1990s unbundling, power
markets are more compatible with the EU Clean Energy system operation has been becoming more market based. As
Package, regarding the market-based procurement of flex- a result, the opportunities for exploiting DER flexibility and
ibility services. demand-side management are continuously expanding. This
The Lolland project tests how a local flexibility mar- leads to more revenue streams for the owners of those assets,
ket can solve congestion at a transmission level on the increased competition and operational cost reductions, more
island of Lolland, which often experiences congestion business opportunities for actors such as aggregators, and
due to very large amounts of renewable power produc- multiple benefits to the overall system operation.
tion. The project will allow the TSO to acquire balanc- Flexibility in Denmark, and particularly demand-side
ing power on an individual transformer basis, in contrast flexibility, is primarily procured and activated by the Danish
to current practice where balancing power is offered per TSO, Energinet, in a national and Nordic balancing market.
bidding zone. However, this trend is changing by establishing additional
The Danish EcoGrid 2.0 designed and demonstrated a local marketplaces for flexibility procurement in two directions:
flexibility market on the island of Bornholm. In this market, 1) downstream marketplaces, such as localized flexibil-
DSOs can acquire flexibility services by DER aggregators in a ity markets as well as local energy communities, and 2)
market-based environment. These services can increase DSO upstream marketplaces, such as regional (e.g., Nordic and
reliability, provide flexibility in case of planned maintenance
Pan-European) markets. This upcoming trend is illustrated
or faults, and resolve potential conflicts between the safe oper-
in Figure 4, including an example for each market.
ation of the distribution network and the provision of services As examples for downstream markets, two energy com-
at a system level by DERs. Since local flexibility markets canmunities in Denmark were recently demonstrated: the
operate seamlessly in parallel to the existing markets, their Energy Collective project in Bofællesskabet Svalin and
operational integration is relatively easy, whereas, by provid-
the EnergyLab Nordhavn project in the Nordhavn area of
ing a market platform for localized market procurement, the Copenhagen. Further development of energy communities
potential use of demand-side flexibility for system balancing is expected to be initiated in Denmark in the coming years.
is not negatively affected. Additionally, EcoGrid 2.0 designed and demonstrated a
local DSO-level flexibility mar-
ket, which will be described later.
As examples of upstream mar-
kets, the Nordic Balancing Model
European is the foundation for continued
Market Nordic cooperation and a market
Regional
Market for manual frequency-restoration
National e n t reserve and automatic frequency-
pm
Localized
Market D evelo
rea m restoration reserve services. The
Market Upst
Community annual socioeconomic benefits of
Market Nordic cooperation are estimated
ent
e v e lopm to be around €200 million and
D
n stream
Do w going to be modernized to ensure
a safe transition to a new way of
figure 4. The existing and upcoming marketplaces for demand-side flexibility pro- operating the Nordic power sys-
curement in Denmark. tem. It will be compliant with

52 ieee power & energy magazine march/april 2021

and able to join European-level balancing energy exchange
markets, i.e., the Manually Activated Reserves Initiative
(MARI) and the Platform for the International Coordina-
tion of Automated Frequency Restoration and Stable Sys-
tem Operation (PICASSO). It is expected that the technical
prerequisites for joining the MARI and PICASSO platforms
will be fulfilled by early 2023, allowing the Nordic manual
frequency-restoration reserve market to join the MARI dur-
ing the second half of 2023. This will provide several new
business opportunities for flexibility providers in the Nordic
area. In parallel, several Internet of Things, artificial intel-
ligence, and blockchain platforms, e.g., “Flex Platform,” are
developing in Denmark, intending to precisely model the
flexibility of each demand-side asset, unlock its flexibility,
bring it to the existing and upcoming markets, and offer var-
ious types of flexibility services on an aggregated asset level.
The island of Bornholm has been used in several of the
projects as a field testbed (Figure 5). Bornholm is a Dan-
ish island that has been a European frontrunner in the green
transition. The population of the 588-km2 island is approxi-
mately 40,000, and the electricity and district heating gener-
ation on the island is based on 100% renewable sources from
wind, sun, and biomass. The island acts as a laboratory for
the testing of new energy technologies, electricity markets,
and operational concepts.
One of these projects is the real-life demonstration of
the Across Continents Electric Vehicle Services project that
investigates the technoeconomic system benefits of large-
scale EV integration in Bornholm, augmented by real usage
figure 5. The living lab of Bornholm island, Denmark. This figure is a snapshot, showing the instantaneous level of total
wind, solar, and biomass-based productions, total demand, and import/export level at 10:21 a.m. on 7 December 2020.
(Source: PowerLabDK, https://bornholm.powerlab.dk/; used with permission.)
march/april 2021
patterns, grid data, and field testing for transcontinental
replicability. A core part of the research activity focuses
on how EVs with commercial bidirectional chargers can be
used to provide frequency support to the system. Twenty-one
vehicles used by the municipality of Bornholm, paired with
21 10-kW bidirectional chargers operated by a commercial
aggregator, are offering frequency containment reserve for
normal operation by bidding in the Danish ancillary ser-
vices market. Vehicles are driven every day for their normal
use but can still provide frequency reserves for more than
14 h/day.
Figure 6 presents a sample of the aggregated response
from the vehicles following a certain frequency profile.
The average latency between requested and provided
power is around 6 s, and the charger-vehicle pairs are
capable of following the desired set point. However, the
effect of such latencies on power system stability should
be carefully investigated and potentially update the ser-
vice specifications for ancillary services provided by
demand response. It also emphasizes the necessity of
enhancing the performance of flexible demand-side units
and their control infrastructure. In most cases, commer-
cial chargers are not designed for offering fast-frequency
control. Their capabilities may need to be adjusted and
the monitoring and control infrastructure further devel-
oped under the system operator’s requirements, espe-
cially given the progressive reduction of system inertia
that may require very small delays in the provision of
frequency reserves.
ieee power & energy magazine 53
 On the economic side, the provision of primary fre-
quency control is attractive, but network tariffs/taxes and
hardware performance (mainly losses) may negatively affect
the business case. Particular emphasis is given to battery
degradation, which is a topic of concern for vehicle owners.
Measurements on the vehicles’ health are ongoing, and early
results show limited degradation despite a rather intense
duty cycle for the provision of the service.
EcoGrid 2.0 was a large and pivotal project in the sys-
tem integration of demand-side flexibility use on both dis-
tribution and transmission levels. It proposed, implemented,
and tested a local flexibility market for the procurement of
DSO services. Such a local flexibility market can be easily
integrated into power markets and system operation, as it
Frequency (Hz)
operates in parallel to the existing markets. The demonstra-
tions involved 800 flexible residential customers with con-
trollable heating loads on the island of Bornholm.
The developed local flexibility market in the EcoGrid 2.0
project was designed such that DSOs can acquire flexibility
services from aggregators in a market-based environment.
The potential use of flexibility from the DSO side could
be to reduce consumption during periods of exceptionally
high demand. Another use of aggregator flexibility could be
avoidance of the disconnection of customers during periods
of planned maintenance of the distribution network.
A particularly attractive feature of such a market is that
potential conflicts from the use of flexibility at both DSO
and TSO levels can be resolved without the need for explicit
DSO-TSO coordination. For example, in the case of over-
loading distribution network components due to the aggre-
gator activation of flexibility at the TSO level, the DSO

50.06 can activate acquired services and alleviate the congestion.

50.04
50.02
These services have been offered by aggregators, and poten-
50 tial DSO service activations are incorporated into the aggre-
49.98 gator’s operational planning so that the consequences on the
0 1 2 3 4 5 6 7 8 9 10 delivery of other services to the TSO will be to the aggrega-
Time (min) tor’s knowledge.
(a) EcoGrid 2.0 operated the developed market with two
60 competing aggregators. These two aggregators virtually
40
participated in the Danish balancing market while, at the
Power (kW)

20
0 same time, offering DSO services through the local flexibil-
–20
–40 ity market. One of the main goals of the project was the pro-
–60
–80 vision of DSO flexibility at low aggregation levels, even at a
–100
0 1 2 3 4 5 6 7 8 9 10 low-voltage level. The results from a series of tests, typically
Time (min) involving around 100 households, are illustrated in Figure 7.
Although relying on a proven technology with limited func-
Requested Power Provided Power
tionality and flexible loads could not be metered separately
(b) (each house was equipped with one smart meter, measuring
the total net demand), DSO load-reduction services could be
figure 6. The provision of frequency containment reserve delivered with appropriate accuracy.
for normal operation from 21 EVs located on the island of
Bornholm. (Data courtesy of NUVVE.) Conclusion and Future Perspectives
Flexibility is becoming more abundant in power systems as
more DERs are being installed and technological barriers are
gradually eliminated. Numerous research projects and expe-
Delivered Load Reduction (kW)

200 riences from power market participation have demonstrated

that flexible resources can offer power system services while
150
meeting the technical requirements of system operators.
100
However, it is challenging to unlock this large potential and
use it to enhance the energy security of the power system.
50 Demand-side flexibility needs to be utilized on multiple
market levels to fully realize its potential and broaden the
0 spectrum of commercial opportunities for asset owners
0 50 100 150 200
Requested Load Reduction (kW)
and aggregators. There are currently many parallel devel-
opments that lead to an emerging multilayered landscape.
figure 7. The requested versus delivered flexibility from Energy communities are promoted by several stakehold-
households with heat pumps or direct electric heating on ers, including the EU, as a means of the democratization
the island of Bornholm. Each point represents a response of energy procurement, which empowers and engages end
from an average of 100 households. users. Denmark is very active in this field, with several

54 ieee power & energy magazine march/april 2021

projects investigating such consumer-centric decentralized
electricity market structures. At the same time, local flex-
ibility markets are advocated as an efficient market-based
approach in procuring flexibility at finer geographical gran-
ularities. Denmark is very engaged in those developments
as well with large projects, such as EcoGrid 2.0 and Lolland
Local Market.
With the current low energy prices and high equipment
costs due to the lack of standardization, the resulting small
payments per flexible resource make it challenging to estab-
lish viable business models except for perhaps a few cases,
such as the provision of frequency reserves. Standardization
is widely recognized as a key requirement for improving the
economic attractiveness of flexibility procurement. A coor-
dinated effort on a European level is underway to harmonize
TSO practices, regarding the monitoring, validation, and
standardization of communication. The recently proposed
platform, Equigy, can act as a common and standardized
interface among TSOs, aggregators, and DERs to facilitate
the prequalification, provision, and validation of ancillary
services. The Danish TSO, Energinet, represents Denmark
in this effort to utilize digitalization and blockchain tech-
nologies to lift the barriers for the mass integration of DERs
in power system operation.
DSOs are expected to play a pivotal role in this transition
to a DER-dominated power system. The substantial increase
in electricity demand and the amount of flexible demand that
follows the volatile power market prices pose a threat in the
security of the supply of end-users unless costly grid reinforce-
ments are undertaken. Demand-side flexibility can turn this
challenge into an opportunity by creating new value streams,
given that the appropriate framework is in place. The Danish
DSOs and TSO have established a market and operation col-
laboration committee to discuss such future challenges and
how to develop operational and market-based solutions.
This is unique internationally and has proven to be effi-
cient with common proposals for a future tariff structure
and the establishment of a pilot project on a local flexibility
market. The Danish Energy Authorities have also started
the Marked Model 3.0 initiative, which focuses on 1) ensur-
ing power system security via market-based solutions, 2)
enhancing flexibility via active customers and new actors,
and 3) developing the role of DSOs in the market. As part of
the initiative, the Danish market design for local flexibility
will be developed.
In this new landscape, flexibility must be utilized concur-
rently in different contexts, including the traditional electric-
ity markets, emerging local communities, local flexibility
markets, and regional/European marketplaces. As a result,
the need for research focused on system integration, market
mechanisms, and actor coordination is more relevant than
ever. There are open research questions regarding DSO-TSO
coordination and which mechanisms are more appropriate
for the efficient and secure integration of DERs. Other open
questions exist on how the aforementioned new marketplaces
march/april 2021
can fulfill their role, operate in parallel, and not jeopardize
the secure operation of the system and on how these develop-
ments can be facilitated by a future digital energy architec-
ture with open interfaces.
Future research projects at Danish and EU levels should
include DSOs and TSOs, as well as aggregators and asset
owners, in large-scale demonstrations that resemble real oper-
ations as closely as possible. Such projects can provide valu-
able insights and learnings, point out regulatory or technical
challenges that require particular consideration, and provide a
way forward to make attractive business cases for all actors.
For Further Reading
“Powering Denmark’s green transition,” The Danish Gov-
ernment’s Climate Partnerships, 2020. [Online]. Available:
https://www.danskenergi.dk/the-climate-partnership-the-
energy-and-utilities-sector
P. Pinson, L. Mitridati, C. Ordoudis, and J. Østergaard,
”Towards fully renewable energy systems: Experience and
trends in Denmark,” CSEE J. Power Energy Syst., vol. 3,
no. 1, pp. 26–35, 2017. doi: 10.17775/CSEEJPES.2017.0005.
“Danish research projects within smart grids 2010–2019,”
Zenodo, June 16, 2020. doi: 10.5281/zenodo.3866327
FinGrid, Energinet, Statnett and Svenska Kraftnät, “Lo-
cal Flexibility Projects in the Nordics,” 2020. [Online]. Avail-
able: https://energinet.dk/-/media/F3F4B2A805014164A5F
9FB8F5107176B.pdf
C. Heinrich, C. Ziras, A. L. A. Syrri, and H. W. Bindner,
”EcoGrid 2.0: A large-scale field trial of a local flexibility
market,” Appl. Energy, vol. 261, 2020, Art. no. 114399. doi:
10.1016/j.apenergy.2019.114399.
A. Zecchino, A. M. Prostejovsky, C. Ziras, and M.
Marinelli, ”Large-scale provision of frequency control via
V2G: The Bornholm Power System Case,” Electric power
System Res., vol. 170, pp. 25–34, May 2019. doi: 10.1016/j.
epsr.2018.12.027.
Biographies
Jacob Østergaard is with the Technical University of Den-
mark, Kongens Lyngby, 2800, Denmark.
Charalampos Ziras is with the Technical University of
Denmark, Kongens Lyngby, 2800, Denmark.
Henrik W. Bindner is with the Technical University of
Denmark, Kongens Lyngby, 2800, Denmark.
Jalal Kazempour is with the Technical University of
Denmark, Kongens Lyngby, 2800, Denmark.
Mattia Marinelli is with the Technical University of
Denmark, Kongens Lyngby, 2800, Denmark.
Peter Markussen is with Energinet, Fredericia, 7000,
Denmark.
Signe Horn Rosted is with Energinet, Fredericia, 7000,
Denmark.
Jørgen S. Christensen is with Danish Energy, Frederiks-
berg C, 1900, Denmark.
p&e
ieee power & energy magazine 55
 A Vision
to Enhance
Transmission
Security
The Case of Switzerland’s
Power System

I
IT IS HARD TO IMAGINE LIFE WITHOUT ELEC-
tricity. Electrical energy is constantly consumed to allow
humanity to carry out diverse activities ranging from charg-
ing mobile phones to controlling the flow of vehicles and
pedestrians with traffic lights. Due to its widespread usage
and significance, some people take the availability of
electric power for granted and expect that the lights will
always turn on when they turn on the switch. Neverthe-
less, when illuminating our rooms, we need to appreciate
that an extensive power system infrastructure is in place
to securely generate, transport, and distribute electrical
energy to our homes.
Electric power systems have been undergoing massive
changes over the last few decades. This includes the integra-
tion of high levels of small-scale energy resources in distri-
bution grids, the electrification of transportation and heat-
ing/cooling sectors, higher volatility due to uncertain power
in-feed from renewable energy generation, the international-
ization of energy markets, and increasing public opposition
to grid expansion, just to name a few. These changes make it
increasingly challenging to maintain the security of supply
in the power system.
Power system security is defined as the ability of the sys-
tem to withstand sudden disturbances or unanticipated loss

By Evangelos Vrettos, Marc Hohmann,

and Marek Zima
Digital Object Identifier 10.1109/MPE.2020.3043567
Date of current version: 19 February 2021

56 ieee power & energy magazine 1540-7977/21©2021IEEE march/april 2021


of system elements and is closely related to system adequacy,
reliability, and integrity. This article is concerned with secu-
rity issues in the operation of the high-voltage transmission
grid originating from environmental or transmission con-
straints and focuses on the case of Switzerland. Specifi-
cally, we focus on steady-state insecurity, namely violations
in transmission line flows or bus voltage limits following
a component contingency, without considering dynamic
aspects such as maintaining synchronism and system stabil-
ity following the disturbance.
First, we describe today’s operational philosophy
of Swissgrid, the Swiss transmission system operator
(TSO), which is also commonly applied by other Euro-
pean TSOs. Next, we identify potential risks of this
operational philosophy, which we exemplify through dis-
cussion of recent grid events in the Swiss power system
and link to observed trends in system operation. Finally,
we argue that a new operational paradigm is needed to
mitigate such risks in the transmission grid of the future.
This new philosophy relies on data analytics and deci-
sion support systems (DSSs) to assist the work of system
march/april 2021
©SHUTTERSTOCK.COM/STEFANO EMBER
operators, and we provide examples of recent relevant
initiatives at Swissgrid.
Security Challenges and Today’s
Transmission System Operator
Operational Philosophy
Power system security is one of the key responsibilities of
TSOs. This article elaborates on how system security is
addressed in the planning and operational procedures in
Switzerland. Despite the Swiss focus, the principles of the
operational philosophy outlined here apply to several other
European TSOs as well.
The Unique Characteristics
of the Swiss Power System
The Swiss transmission system is very strongly intercon-
nected with its neighboring countries. There are, in fact, 41
cross-border transmission lines compared to about 160 lines
that both start and end within Switzerland.
The Swiss transmission system evolved organically by
following the residential and commercial activity patterns
ieee power & energy magazine 57
 as well as the natural energy sources. Thus, the consumption
centers are located in the north, together with run-of-river
hydropower plants and nuclear stations, whereas the storage
and pumped-storage plants (in other words, the more flex-
ible power plants) are in the Alps in the south. The stron-
gest transmission path runs from the north to the south, or
rather southeast, from where it extends to Italy. Historical
flows naturally followed the structure of the transmission
lines, with the cross-border exchange joining in domi-
nantly with the imports from France and Germany and
exports toward Italy.
The Swiss power system features another characteris-
tic that is not common in most other continental European
systems: the subtransmission system and some distribution
systems do not have a radial structure but are rather strongly
meshed and connected to transmission nodes in several loca-
tions. Thus, they accommodate part of the power flows in
parallel to the transmission lines.
The installed production capacity exceeds 20 GW, with
the peak consumption being approximately 11.5 GW, so the
traditional adequacy assessment does not reveal the weak-
ness of the system, which is the energy scarcity in particu-
lar months due to the seasonality in water inflows. This is
caused by the utilization of water in the reservoirs of hydro-
power plants during winter, which only start to refill again
in the summer when snow in the Alps melts.
Switzerland has adopted a self-scheduling market design,
with the vast majority of transactions taking place “over-the-
counter” on a typically bilateral basis between the market
participants. Only a small portion of transactions are “vis-
ible” in the power exchange. There is no explicit connec-
tion between transactions and the underlying physical assets.
All technical constraints of the power plants are embedded
into the transactions or bids. Although electricity transac-
tions have to respect the net transfer capacities among coun-
tries, the transmission constraints on the level of individual
transmission lines are not explicitly included in the market
activities—the assumption of a “copper plate” is applied.
If congestion in network elements occurs as a result of the
market activities, it is the responsibility of the TSO to apply
remedial actions to mitigate them.
From an organizational point of view, the Swiss electric-
ity landscape is quite fragmented for a system with an annual
consumption of about 65 TWh. There are about 80 power
plant companies, even though they often have similar share-
holders, and more than 600 distribution system operators
(DSOs). About 100 balance responsible parties group con-
sumers and producers for scheduling and electricity transit
purposes. Switzerland is surrounded by member countries of
the European Union subjected to its legislation, which is not
directly adopted by Switzerland.
Operation and Operational Planning
The factors described in the previous section imply that the
operation of the Swiss transmission system is strongly affected
58 ieee power & energy magazine
by annual seasonality, short-term volatility, the availability of
timely and accurate information/data from other stakeholders,
and limited enforceability or intervention possibilities. These
challenges are addressed in a set of procedures spanning vari-
ous geographical ranges and time scales.
✔ Day-ahead regional (cross-border) coordinated oper-
ational planning: To capture the effect of various parts
of the interconnected system on other parts, TSOs of
continental Europe exchange their data for the coming
day and carry out a coordinated security assessment.
If they recognize a risk, which can be mitigated only
by an action in another country, they place the corre-
sponding request. First, topological measures (e.g., re-
configuring busbars in a substation) are checked, and
if the computed effect is not sufficient, a preventive
redispatch (or, more precisely expressed, countertrad-
ing) is triggered among TSOs. Alternatively, trans-
mission capacities between the countries are restricted
for the trading activities of the following day.
✔ Day-ahead operational planning in Switzerland: The
data from the stakeholders are collected first, includ-
ing expected power plant schedules, consumption
forecasts, and planned cross-border exchanges. An
N−1 security assessment in the form of contingency
screening is applied for the expected deterministic
scenario. In the case of congestions in the production
regions, congestion warnings are issued to the produc-
ers. Some topological measures may be applied. No
redispatch actions take place as further trading activi-
ties may take place in the intraday market.
✔ Intraday and real-time operation: Switzerland is not
included in the electricity market coupling of the Eu-
ropean Union member states. Thus, the Swiss system
is occasionally subjected to a sudden change of the
transmission flows with minimal previous awareness.
Also, stakeholders within Switzerland may change
their positions and actions communicated in the day-
ahead market on very short notice due to a variety
of reasons. Therefore, a rolling security assessment
forecast is carried out every 15 min for the period of
the subsequent 4 h. Operators may then decide which
countermeasures to employ and in which sequence
(preventive, corrective) to guarantee security. Again,
typically topological measures are applied first, when
available, which are mostly determined by operators
based on experience or simulation-based methodolo-
gies. If no suitable topological measure is available,
redispatch is employed within Switzerland or if no
national redispatch measure is available, cross-border
redispatch with the neighboring TSOs is applied.
The Maintenance of Transmission Assets
The maintenance of transmission assets naturally takes
place in the summer months as environmental conditions are
most favorable at that time, especially in the Alps, and the
march/april 2021
 In general, a comprehensive medium-term energy security planning
process is needed, which includes planned grid and power plant
outages, hydro storage levels, and market conditions.
electricity demand is lowest in summer, not necessitating the
availability of all power plants.
The process follows a bottom-up approach: transmission
assets are grouped according to region, with the regional
manager submitting the maintenance requests for approval.
The requests are checked by the operational planning team
in order of submission, applying rough assumptions about
the expected system conditions as the future system state is
not accurately known or forecasted. This means there is also
little information about the assumed availability of possible
corrective actions.
Power plants similarly announce their maintenance
schedules on an individual basis. Their unavailability is con-
sidered for the system security assessment, but the TSO has
only an advisory role in rescheduling, with the only excep-
tion being the power plants contracted for black start and
islanding operation capability.
Note that data exchange with DSOs takes place only to
a limited extent in the planning stage. Thus, when mainte-
nance takes place in the distribution grid, the transmission
system may be affected in real time without the detailed
awareness on the side of the TSO.
Historical Grid Events
in the Swiss System
Although today’s operational philosophy has been largely
successful in maintaining the security of the power supply,
some recent grid events have demonstrated growing poten-
tial risks due to the evolution of the changes to which the
system is subjected. This section discusses several recent
events and attempts to identify relevant aspects of the opera-
tional philosophy that could be improved to better handle
such situations in the future.
The Winter 2015–2016 Situation
A problematic supply and grid situation was identified in
December 2015 due to a combination of circumstances. Dry
weather conditions during summer and autumn led to lower
power generation from run-of-river plants. The filling lev-
els of hydro storage plants published by the Swiss Office
of Energy were below the long-term annual average. More-
over, nuclear power reactors Beznau 1 and 2 on the 220-kV
level were out of operation, totaling 720 MW of missing
power in-feed.
The missing base production from run-of-river and
nuclear plants had to be compensated by imports and produc-
tion from hydro storage plants. While the supply of Swiss
march/april 2021
consumers is mostly via the 220-kV level, up to 85% of
imported power comes in via the 380-kV level during win-
tertime. Since the transformation capacity between the
380- and 220-kV level was limited, the risk of premature
depletion of the hydro storage dams (most of them are con-
nected to the 220-kV network) followed by a lack of energy
supply increased. The problem was amplified due to the
rather high energy prices at the beginning of the winter,
which resulted in many Swiss hydro storage power plants
often exporting energy. An illustration of the problem
related to the limited production resources in the 220-kV
network is given in Figure 1.
To mitigate these problems, a temporary connection of
the transformer in Laufenburg that allowed transformation
between the 380- and 220-kV networks was carried out,
and a reserve transformer was commissioned in Tierfehd,
as depicted in Figure 2. Moreover, topological measures in
Werben and Bürs increased the import on the 220-kV level
from Austria via the Rhine valley. The outcome of these
measures was to increase the net transfer capacity to France,
Germany, and Austria during peak times. To allow for a
higher import to cover the domestic consumption in Swit-
zerland, monthly export products (cross-border capacity) to
neighboring countries were made unavailable during winter.
The measures taken by Swissgrid, together with the unusu-
ally mild, rainy winter weather and the recommissioning
of block 2 of the Beznau nuclear power plant on 23 and
24 December 2015, helped improve the grid situation from
the Christmas holidays 2015–2016 onwards.
This event demonstrated the challenges in maintaining
the overall adequacy and security of supply in an electric-
ity system with unbundled production and grid operation.
Swissgrid is responsible for maintaining network secu-
rity, but no entity in Switzerland is in charge of ensuring
the overall security of supply. In general, a comprehensive
medium-term energy security planning process is needed,
which includes planned grid and power plant outages, hydro
storage levels, and market conditions.
Grid Security Violation on 20 May 2019
Due to attractive market conditions, there was an unusu-
ally high generation of electrical power in Switzerland (up
to 12 GW) on 20 May 2019 as well as a high net export to
neighboring countries (up to 4.5 GW). In particular, there
was high planned export to Germany (up to 4 GW export,
whereas Switzerland typically imports from Germany), very
low export to Italy (1.3 GW, which is quite uncommon), and
ieee power & energy magazine 59
 import from France (600 MW). This export situation to the
north, unusual in comparison to the past, is potentially prob-
lematic for the Swiss transmission grid and might result in
congestion because production has to be transported to the
north via the 220-kV grid within Switzerland.
On top of this unusual export pattern, the concurrence
of several other factors led to a security violation. In the
distribution grid, several outages planned by a DSO were
executed, and these were not known to the TSO. In the same
region of the distribution grid, transit flows from the trans-
mission grid through the distribution grid caused the over-
loading of a distribution line and the need for an unplanned
outage of the affected line, of which the TSO was not
aware. These topological changes in the distribution grid
due to the execution of planned outages and actions to
avoid overloading the line, which were unexpected for the
TSO, required topological measures (transformer phase
shifts) in the transmission grid. These, in turn, led to load
redistribution. Furthermore, there were short-term produc-
tion increases planned and communicated only shortly
before real time. The combination of these factors led to
overloading a 220-kV network element and multiple N−1
violations on the 220-kV level of the transmission grid.
The following aspects were identified as the main causes
of this event:
✔ the uncommon export situation and high energy
production in Switzerland together with short-term
changes in power production schedules
✔ the effect of topological changes in the distribution
grid on the transmission grid as well as the effect of
topological measures in the transmission grid on the
load allocation in the lower-voltage grids.
Whereas the first point is mostly related to the expansion
planning of the Swiss power grid, the second one is related to
today’s operational philosophy. Specifically, the congestion
forecast tool was unable to predict the violations of transmis-
sion grid elements in time because the subtransmission and dis-
tribution networks are not included in the model. Therefore, the
main lesson from this grid event is the importance of having an
operational philosophy that considers the subtransmission and
distribution networks (or at least important networked parts of
them) both in outage planning and real-time operations.

Germany,
France, Austria
Nuclear Storage
Power Plants Power Plants
Import/ Import/
Leibstadt Gösgen
Export Export Imports Supply Exports

MW MW
t t GL 1 75–85% 5% 80–90%
Base Load Peak
380-kV Grid Energy
Nuclear Storage
Power Plants Power Plants
Mühleberg Beznau 1/ GL 1
Beznau 2

MW MW
t t GL 1 15–5% 95% 10–20%
Base Load Peak Energy
220-kV Grid

Run-of-River Storage GL 2
Power Plant Power Plants

MW MW MW
t t GL 3– 7
Base Load Peak Energy
Distribution Grid

Transformation Power Plants

figure 1. An illustration of the cause of the winter 2015–2016 situation with most of the load, but only limited power pro-
duction and import, in the 220-kV grid. Import was mainly via the 380-kV level, but the transformation capacity between
the 380- and 220-kV levels was limited.

60 ieee power & energy magazine march/april 2021

Minimum Production Product
in Winter 2019–2020
Following the plan for phasing out nuclear power in Swit-
zerland, the Mühleberg nuclear power plant was decommis-
sioned in November 2019. The decommissioning coincided
with limited controllability and capacity limitations for the
Bassecourt transformer and delays in the commissioning
of a new transformer in Mühleberg. As a consequence, and
assuming that the missing power was to be compensated by
additional import from the north, i.e., by increasing the net
transfer capacity to Austria, Germany, and France, some net-
work elements in Western Switzerland (especially the trans-
former in Bassecourt) were expected to be overloaded during
the low consumption period between Christmas 2019 and
New Year 2020.
To guarantee grid security in the region between Bassec-
ourt, Mühleberg, and Chamoson, Swissgrid agreed upon
a minimum level of production with selected power plants
in Western Switzerland from 20 December 2019 to 6 Janu-
ary 2020. Specifically, the selected power plants agreed
to cumulatively produce a minimum of 200 MW at off-
peak times and 400 MW at peak times. This longer-term
preventive measure was shown to be preferable and more
rm
Ge
1 St. Gallen
France
Bern
France
France
Genève
Italy
Italy
Italy
figure 2. The measures applied to relieve the stressed grid situation in winter 2015–2016. Location 1: a temporary con-
nection of the transformer in Laufenburg to allow for transformation between the 380- and 220-kV networks. Location
2: the topological measures to increase the import on the 220-kV level from Austria via the Rhine Valley. Location 3: the
commissioning of a reserve transformer in Tierfehd.
march/april 2021
economical to alternative measures, such as reducing the
net transfer capacity to Austria, Germany, and France (i.e.,
to incentivize Swiss power plants to increase production) or
resorting to many shorter-term national and/or international
redispatch calls closer to real-time operation.
This is a relevant example of real-world situations where
new measures against grid insecurity must be devised by
system operators in a short amount of time. Despite the
measure’s success, this event demonstrated the value of an
integrated approach for outage planning and the selection
of remedial actions, which can include the design of new or
customized products such as longer-term minimum power
production products. TSOs would benefit from a new opera-
tional paradigm, including such an integrated planning and
operation approach.
COVID-19 Lockdown in March–April 2020
As a measure against the spread of COVID-19, the Swiss
Federal Council ordered the closing of stores and public ven-
ues starting 16 March 2020. Neighboring European coun-
tries took similar measures around the same point in time.
As a consequence, the domestic energy consumption dur-
ing weekdays and the energy flows in and across the Swiss
any
Lines
Substations
Zürich
2
Austria
3
Chur
Lugano
Italy
ieee power & energy magazine 61
 network decreased significantly. Figure 3 illustrates the
country-wise aggregated load profile right before and after
the lockdown, where the sudden reduction in consumption is
apparent. The subsequent higher injection of reactive power
from distribution grids (because feeder cables are capacitive
at light loading conditions) and the lower inductive absorp-
tion of reactive power in transmission lines increased the risk
of local overvoltages. To mitigate this risk, additional reactive
power reserves were activated. Also, topological measures
(e.g., line switching and transformer tapping) were imple-
mented to decrease or redistribute reactive power flows.
As mentioned in the “Operation and Operational Plan-
ning” section, preventive remedial actions are typically
1
0.9
to Peak Load in March 2019
determined by the operators based on experience or com-
puted using simulation-based approaches. Due to the
advantages of optimization-based methods, such methods
are being investigated by some TSOs for determining the
best remedial actions. Optimization models are mostly
based on dc network models as they target avoiding line
overloadings. Nevertheless, the COVID-19 situation
showed the significance of an operational philosophy
allowing system operators to implement preventive mea-
sures to circumvent voltage violations in addition to typi-
cal curative measures. To efficiently and systematically
compute preventive measures against voltage violations,
optimization models that can handle ac network models
would be required together with the associated computa-
tional and algorithmic innovations.
Overall Trends in System Operation
The previous analysis of recent grid events revealed specific

aspects of system operation that would benefit from a new

0.8
operational philosophy. To evaluate the need for a new oper-
Load Relative

0.7 ational paradigm, one can also observe the overall trends in
system operation in recent years. In this section, we pres-
0.6
ent a simple analysis of the historical remedial actions in
0.5 the Swiss transmission system taken against N−1 or voltage
security violations. The remedial actions include topologi-
0.4
cal measures to optimize either power flows or voltage pro-
0.3 files as well as generation redispatch. Furthermore, we dis-
0 20 40 60 80 100 120 140 160 180 tinguish between two types of redispatch actions: national
Hour of the Week, Starting Sunday 00:00 and international redispatch. National redispatch actions
Average Consumption in March 2019 are on Swiss power plants in response to internal issues
Week Before Lockdown in the Swiss transmission grid. On the other hand, inter-
First Week of Lockdown national redispatch actions implemented by Swiss power
Second Week of Lockdown plants are part of an international coordinated redispatch
effort to resolve problems in either the Swiss or neighboring
figure 3. The impact of the lockdown measures on Swiss power grids.
electrical consumption. Figure 4 displays the annual number of topological mea-
sures from 2015 to 2019 that were implemented either to
optimize the power flows and relieve congestions or to opti-
mize the voltage profile across the grid. A linear increasing
2,000
Load Flow Optimization
trend is visible in both cases.
Number of Topological Measures

1,800
Voltage Optimization Figure 5 demonstrates the trend in redispatched energy
1,600 from 2014 to 2019. Besides the total energy values, we
Total
1,400 present the breakdown into Swiss and international
1,200 redispatches as well as generation decrease and increase
1,000 actions. Even though it is hard to identify clear patterns
800 for the various combinations, there is a general increas-
600 ing trend for the total redispatched energy. A clearer trend
400
exists in the frequency of redispatch actions, which is
quantified by the number of days when at least one redis-
200
patch action was carried out (Figure 6). Furthermore, it is
0
2015 2016 2017 2018 2019 relevant to mention that the majority of redispatch actions
Year was international (79% in terms of redispatch duration).
This observation underlines the significance of TSO–TSO
figure 4. The trend in topological measures as remedial coordination as part of today’s operational procedures, as
actions. explained previously.

62 ieee power & energy magazine march/april 2021

 Finally, there have been several recent grid situations
where avoiding N−1 violations by applying separated busbar
operations in several substations would lead to a low degree
of connectivity in the transmission grid. In the worst case
after a busbar operation, if a substation is connected to the
rest of the grid with only one transmission line, the unex-
pected tripping of this line (e.g., a lightning strike) would
cause the loss of the substation and the interruption of supply
in the underlying distribution grid.
To summarize, the number of remedial actions against
N−1 or voltage violations has been increasing over the last
years, which increases the burden on the system operators in
the control room. As mentioned earlier, remedial actions are
mostly determined based on experience and rules of thumb
derived from simulation-based approaches. We believe that
system operation would benefit from a DSS that provides the
operators with suggested remedial actions against N−1 and
analytics can be applied to evaluate the effectiveness of his-
torical remedial actions. A possible approach would start
with assessing the accuracy of the congestion forecast tool
and proceed with analyzing and classifying historical grid
snapshots. More precisely, for each grid snapshot with secu-
rity violations, the goal would be to determine the root cause
such as 1) no congestion warning was issued due to poor
forecast quality of generation and/or load; 2) congestion
was correctly predicted, but the selected remedial action
turned out to be inefficient; and 3) congestion was correctly
predicted, but there was no available remedial action to
mitigate the insecurity event. This analysis can help a TSO
300

Redispatched Energy (GWh)

250
voltage violations. More importantly, the DSS could effi-
ciently account for possible undesired side effects of actions 200
against one problem (e.g., N−1 violation) on other opera-
tional criteria (e.g., voltage profiles and degree of transmis- 150
sion network connectivity).
100
Vision for a New Operational Paradigm
The analysis of recent grid events and trends in system 50
operation demonstrate that improvements in the cur-
0
rent operational philosophy, or even a new philosophy, 2014 2015 2016 2017 2018 2019
are needed to ensure power system security in an ever- Year
changing environment. In this section, we discuss ele-
ments that shall have a central role in a new successful CH-Decrease CH-Total
operational paradigm. CH-Increase International-Total
International-Decrease Total
International-Increase
Analytics and Visualization
Secure power system operation requires taking fast and
effective actions under critical grid conditions. Typically, figure 5. The trend in redispatched energy due to reme-
many alarms are displayed on the large monitor of a TSO dial actions. CH: Swiss national.
control center in case of a grid event. In some cases, the
system operators must visually process this information
and identify the most important alarms. Analytics and
80
artificial intelligence applied to real-time monitoring data CH
can help classify the alarms into first priority (for exam- 70
International
With Redispatch Actions

ple, alarms associated with the root cause of an event) and 60 Total
second priority (for example, alarms activated as a con-
Number of Days

50
sequence of other alarms). Such methods, in combination
with modern visualization approaches that, for example, 40
display the alarms according to their assigned priorities, 30
can help the operators obtain valuable insights into the
ongoing grid conditions and increase their reaction speed. 20
Moreover, algorithms running on monitoring data could 10
help predict imminent grid events and provide recom-
0
mendations of remedial actions to the operator to further 2014 2015 2016 2017 2018 2019
reduce the reaction time. Year
Automated ex-post data analytics can help identify the
root causes of past events of grid insecurity, learn from figure 6. The trend in the number of days per year when
them, and better prepare for the future. For example, ex-post redispatch actions were implemented.

march/april 2021 ieee power & energy magazine 63

 decide where to devote more resources to improve energy Currently, Swissgrid is working on such ex-post analyt-
security, for instance, by upgrading the congestion forecast ics, and here we present some preliminary results from the
tool or by investing in a DSS for the optimization of reme- assessment of the net load forecast. As part of the intraday
dial actions. congestion forecast process, a tool generates predictions
of the aggregate net load in the Swiss system as well as
of the net load at each transmission bus. Specifically, the
Forecast Error for bus-level forecasts are obtained based on the aggregate
System-Level Active Power Injection forecast with the usage of predetermined load allocation
11
Normalized
RMSE (%)

10 factors. The predictions are updated every hour and have

9 a forecast horizon in the range of 8–24 h, depending on the
8 time of the day.
1 3 5 7 9 11 Figure 7 displays the normalized root-mean-square
Forecast Horizon (h)
error (RMSE) of the predicted net active power load for six
(a) increasing forecast horizons. The errors are computed for
Forecast Error for the period from January to April 2020. In Figure 7, (a) is
Node-Level Power Injection: 25% Quantile for the aggregate net load, whereas (b) and (c) correspond
11
Normalized
RMSE (%)

10
to bus-level predictions. Specifically, Figure 7(b) shows the
9 RMSE’s 25% quantile, and Figure 7(c) plots the median
8 value, both computed across all buses. Although the error
1 3 5 7 9 11
generally increases with the forecast horizon, the differ-
Forecast Horizon (h)
ences are rather small, and this trend is less pronounced
(b)
for bus-level predictions. In contrast to the relatively low
Forecast Error for errors for the aggregate net load, the median of the bus-level
Node-Level Active Power Injection: Median
35 RMSEs is unexpectedly high, mostly due to shifts and/or
Normalized
RMSE (%)

34 biases in the predicted time series for some of the buses.

33 This may be partly because of a significant amount of pro-
32 duction resources in and/or a meshed topology of the under-
1 3 5 7 9 11
Forecast Horizon (h) lying lower-voltage grids.
(c) Furthermore, Figure 8 illustrates the ratio of RMSE for
forecasts obtained with a 1-h lead time over those with a 7-h
figure 7. The dependence of prediction quality of net ac- lead time. The smaller the ratio, the better the relative perfor-
tive power load on the forecast horizon. mance of 1-h-ahead forecasts. While, for most transmission

Ratio Between RMSE of 1-h-Ahead

Forecast With That of 7-h-Ahead Forecast (Active Power)

1.017

47° 30′ N
0.906

Horizon 1 h: Higher Error

47° N Horizon 1 h: Lower Error
Latitude

46° 30′ N

46° N
50 km
20 mi
6° E 7° E 8° E 9° E 10° E
Longitude

figure 8. The geographical distribution and classification of transmission buses with regards to the relative accuracy of
1-h-ahead and 7-h-ahead forecasts.

64 ieee power & energy magazine march/april 2021

buses, the 1-h-ahead forecasts are better (orange circles in
the figure), there are also several buses with 7-h-ahead fore-
casts with higher accuracy than the 1-h-ahead ones. Future
investigations can focus on tracing the cause of these imper-
fections and improve the performance by either optimizing
the load allocation factors applied in the forecast model or
switching to a new forecasting approach.
By splitting the data set into the periods before and
after the lockdown in Switzerland as a measure against
the COVID-19 pandemic, we can analyze the effect of the
lockdown on the forecast quality. As seen in Figure 9, the
RMSEs are significantly lower in the prelockdown period,
especially for longer prediction horizons. This can be partly
attributed to the fact that the forecasts have a strong season-
ality component, which resulted in an overestimation of the
net load demand in the weeks right after the lockdown.
activities closer to real time within the ones in the planning
phase (indicated with black arrows in Figure 10).
Within this general vision, we specifically propose an
approach to integrate the maintenance scheduling (outage
planning) and congestion management (congestion fore-
cast, security analysis, and remedial actions) phases of the
activity chain. Today, Swissgrid’s operational planning team
receives maintenance plans from transmission asset owners
(ownership of Swissgrid), generation asset owners, distribu-
tion asset owners (only to a limited extent), and transmission
assets of neighboring TSOs. Information comes through
15
14 Before Lockdown
13 After Lockdown

Normalized RMSE (%)

DSSs 12
The second key aspect of the future operational philosophy 11
is an extensive usage of DSSs to assist the personnel involved 10
in core TSO tasks in decision making. As detailed in Figure 10, 9
the four core TSO activities to ensure system security 8
are 1) grid reinforcement and expansion, 2) maintenance 7
scheduling for transmission assets, 3) congestion manage-
6
ment, and 4) control actions in real-time system operation.
5
Today, these four activities typically follow a chronological 1 3 5 7 9 11
sequence and interact in a mostly unidirectional way from Forecast Horizon (h)
left to right, as depicted with the gray arrows in the figure.
We believe that higher effectiveness across the entire activ- figure 9. The effect of anti-COVID-19 lockdown mea-
ity chain can be achieved if the degree of bidirectional infor- sures (applied 16 March 2020) on the intraday congestion
mation flow increases, namely if there is the consideration of forecast error.

Portal
Goal: Network Security

Grid Expansion Maintenance Planning Congestion Management System Operation

10 Years Ahead, One Year Ahead, One Year Ahead, Real Time
Updated Every Updated Monthly One Week Ahead,
Four Years and Weekly One Day Ahead, Intraday

figure 10. Strengthening the links among the core TSO activities will enhance network security with a focus on mainte-
nance planning and congestion management. (This image has been designed using resources from Flaticon.com.)

march/april 2021 ieee power & energy magazine 65

 different communication channels ranging from dedicated
platforms to emails and phone calls. It then is the task of
the operational planning team to verify the feasibility of the
received maintenance schedules based on the congestion
forecast and security analysis (at various time scales: year
ahead, week ahead, and day ahead).
In case of anticipated N or N−1 violations, three types
of measures are investigated: topological actions, generation
limiting and redispatch, and maintenance rescheduling. In
the latter case, coordination with the respective asset owner
is required until an alternative solution is found, and then
security analysis needs to be repeated to check the mea-
sure’s effectiveness. This iterative process is the result of
the sequential coupling between maintenance planning and
congestion management. Notably, if anticipated congestion
can be avoided by either a real-time topological action or
by rescheduling the maintenance of a grid asset, the former
is often preferred due to the complexity in maintenance
rescheduling among several asset owners and their mainte-
nance service providers.
To improve operational efficiency and security of supply,
we envision an integrated approach to optimize the mainte-
nance schedules of a control area’s transmission grid assets
while considering possible congested network elements in
system operation as well as the maintenance needs of gener-
ation and distribution grid assets. A comprehensive mainte-
nance scheduling tool would comprise maintenance periods
as decision variables for four types of assets: 1) transmis-
sion assets of the control area, 2) DSO assets, 3) generation
assets, and 4) transmission assets of neighboring control
areas. Since the specific maintenance requirements of points
2)–4) are not known to the TSO of the given control area,
we propose that DSOs, generation companies, and neighbor-
ing TSOs submit a list of preferred maintenance windows
as well as blocking periods when no maintenance work is
possible for their assets.
As indicated in Figure 10, the submission of preferences
can be enabled by a portal that all involved parties can
access. The best maintenance window for each asset is then
determined while ensuring network security in case of con-
tingencies and satisfying the submitted preferences as much
as possible. In this way, the TSO’s operational planning team
automatically obtains a maintenance schedule that mini-
mizes the need for remedial actions against grid insecurity.
This is a pragmatic approach placed halfway between today’s
industrial practice where generation and distribution asset
owners submit fixed maintenance schedules to the TSO and
idealized academic approaches where generation and trans-
mission assets with distinct ownerships are co-optimized.
Figure 11 presents an illustration of the envisioned approach
for integrated maintenance scheduling and congestion man-
agement and contrasts it with today’s practice.
From a mathematical point of view, this is a complex
optimization problem due to the presence of integer deci-
sion variables (maintenance schedules, topological remedial
66 ieee power & energy magazine
actions, and so on) and various nonlinearities, especially if
an ac power flow model is used to capture voltage violations.
Moreover, it is a large-scale problem due to the hundreds or
thousands of buses and transmission lines and the consider-
ation of many contingency scenarios. Further, as revealed
by the analysis of some historical grid events, it is some-
times important to include models for the distribution grid
in the study, which further increases the problem size. Such
a large-scale, mixed-integer, nonlinear, and nonconvex opti-
mization problem is practically impossible to solve to full
optimality in a reasonable amount of time and with realistic
computational resources. Nevertheless, for practical applica-
tions, a good feasible solution that satisfies time and compu-
tational constraints would be sufficient.
Various methods have been proposed in the scientific lit-
erature to optimize the maintenance scheduling of power grid
generation and transmission assets (in a separated, sequential,
or joint manner). The methods to solve the resulting mixed-
integer program can be generally classified into mathemati-
cal programming, heuristic approaches, and combinations
of these. Early mathematical programming approaches were
based on dynamic programming that, however, turned out
not to be scalable due to the curse of dimensionality. More
recent approaches rely on branch and bound-based mixed-
integer program models often enhanced with decomposition
techniques (mostly Bender’s decomposition or Lagrang-
ian relaxation). Such approaches typically use a linearized
dc power flow model, thus resulting in mixed-integer linear
programs. A large variety of heuristics has been applied in
the maintenance scheduling problem, including genetic algo-
rithms, particle swarm optimization, simulated annealing,
cuckoo searches, electro search algorithms, and so on.
We believe that a hybrid approach is needed to efficiently
tackle this problem at scale, which combines elements from
mathematical optimization, heuristic optimization, artificial
intelligence, and customized approximations based on oper-
ational experience. An example of a hybrid approach would
be applying first a heuristic to provide a set of initial fea-
sible solutions and then using these to initialize branch and
bound-based algorithms executed in parallel to obtain local
optima in the neighborhood of the initial feasible solutions.
Such an approach would combine the strength of heuristics
in efficiently exploring a vast and nonconvex search space
with the guarantee of mathematical programming to return
locally optimal solutions.
Another possibility would be to use heuristics for the
discrete variables and mathematical programming for the
continuous ones, which could be done either in two steps
or iteratively. For instance, one could use heuristic opti-
mization until the discrete variables converge and then
switch to gradient-based optimization for the remaining
continuous variables. Alternatively, one could decompose
the problem into an outer loop optimizing over the inte-
ger maintenance decisions (solved with heuristics) and an
inner loop optimizing over remedial actions [basically, a
march/april 2021
 Generation
Current Approach Envisioned Approach Maintenance
Generation
Maintenance

march/april 2021
Preferences Final Schedule

Fixed Windows Integrated Maintenance Planning

and Congestion Management
Aggregation Distribution Transmission Distribution
Transmission
of Maintenance Maintenance Maintenance Optimization Maintenance
Maintenance Preferences of Maintenance Preferences
Windows
Final Windows Final
Schedule Schedule
Fixed Windows Fixed Windows
Schedule
Maintenance Adjustments Maintenance Fitness
Schedules
Decisions Evaluation

Congestion Management SCOPF for Congestion Management

Contingency Master Problem

List SCOPF With Selected Contingencies

No
Violation? Approximate Iterative Probabilistic

Yes Topological Base-Case Solution Potentially Binding

Actions Contingencies
Slave Problems
Post-Contingency Security Analysis
Remedial Actions List

Yes Contingency Network

Solved?
Filtering Compression
Redispatch
No

• Fixed Maintenance Windows Approved or Not • Maintenance Windows Optimized Considering Preferences/Constraints
• If Not Approved, Coordinate With Asset Owners • Optimal Selection of Remedial Actions Against Contingencies
and Repeat • Tradeoff Between Remedial Actions and Maintenance Rescheduling

ieee power & energy magazine

figure 11. Today’s approach and future vision for coordination between maintenance scheduling and congestion management. (This image has been designed using
resources from Flaticon.com.)

67
 security-constrained optimal power flow (SCOPF) solved
with mathematical programming].
Customized approximations aim at reducing the size
and/or complexity of the optimization problem by incor-
porating operational experience and knowledge about the
system. Ideally, an industrial-grade DSS for integrated
maintenance scheduling and congestion management
would combine multiple approaches that have shown good
performance in the literature, such as contingency filter-
ing, network compression, and adaptive linearization. The
first two are typically implemented simultaneously, where
contingency filtering selects the contingencies that are
expected to be binding at the optimal solution, and net-
work compression reduces the size of each postcontingency
network. Adaptive linearization could be employed to sim-
plify the power flow equations of selected lines in formu-
lations that use an ac network model. For instance, lines
with frequent N−1 violations or voltage violations at the
terminal buses are modeled with ac power flow equations,
whereas a dc model could be used for lines that hardly ever
experience violations.
Needless to say, the approach presented here can be
extended in various ways to account for additional issues
identified from the analysis of historical grid events. For
example, constraints on the degree of network connectivity
could be modeled in the “SCOPF for Congestion Manage-
ment” module of Figure 11 to avoid creating weak grid links
due to remedial actions. Also, customized measures, such
as the minimum production product discussed earlier in the
“Historical Grid Events in the Swiss System” section, can
be obtained as the outcome of such an optimization-based
approach by introducing the appropriate modeling variables.
Finally, this approach could provide the basis for a more com-
pressive (possibly probabilistic) medium-term energy secu-
rity planning process that integrates outage planning, conges-
tion management, and adequacy planning with consideration
of storage levels of hydro dams and market conditions.
Summary
In this article, we have addressed the challenging task of
maintaining power system security from the perspective
of the Swiss power system. We started with a review of the
distinctive characteristics of the Swiss electricity grid and
an explanation of today’s practices for maintenance schedul-
ing, operational planning, and real-time system operation.
Despite the success of these practices, we argue that enhance-
ments or even a new operational paradigm are needed in
the future to cope with the constantly changing boundary
conditions of the power grid. Based on a few recent grid
events and historical trends in power system operation, we
identified aspects of particular importance. These include
representation of subtransmission and distribution grids in
planning and operation, better coordination of maintenance
scheduling across different power system assets, preventive
measures against voltage violations, and consideration of the
68 ieee power & energy magazine
effect of topological actions to the level of network connec-
tivity. In response, we presented two key aspects of a new
operational paradigm that would help enhance system secu-
rity, namely data analytics and DSSs.
Specifically, we outlined an integrated approach for the
traditionally time-separated activities of outage planning and
congestion management. The approach enables a combined
consideration of the maintenance needs of transmission, dis-
tribution, and generation assets in the planning phase while
increasing the margin for preventive interventions against
contingencies closer to real-time operation. To realize the
envisioned approach, methodological innovations in the field
of SCOPF are needed, including hybrid approaches combin-
ing ideas from mathematical and heuristic optimization.
Acknowledgments
The authors would like to thank the following Swissgrid col-
leagues who provided valuable input for the preparation of
this article: Christophe Dunand, Stéphane Gerbex, Adam
Kai, Rosmarie Joss, Thomas Reinthaler, Bastian Schwark,
Andrin Siegenthaler, Michael Wasmer, Christian Welti, and
Franziska Zuber.
For Further Reading
“System operation guideline (SOGL),” ENTSO-E, Brussels,
Belgium, 2017. [Online]. Available: https://www.entsoe.eu/
network_codes/sys-ops/
“Analysis of the grid security violation on 20 May 2019
available,” Swissgrid, 2019. [Online]. Available: https://www
.swissgrid.ch/en/home/newsroom/newsfeed/20190923-01
.html
“Planning for winter: Winter situation 2015-16,” Swiss-
grid, 2016. [Online]. Available: https://www.swissgrid.ch/
en/home/operation/regulation/winter-planning.html
P. C. López, R. Sadikovic, H. Pinto, and F. Magnago,
“Swiss TSO experience with an AC security-constrained
optimal power flow application for real-time security man-
agement,” in Proc. IEEE Eindhoven PowerTech, 2015, pp.
1–6. doi: 10.1109/PTC.2015.7232442.
A. Froger et al., “Maintenance scheduling in the electric-
ity industry: A literature review,” Eur. J. Oper. Res., vol. 251,
no. 3, pp. 695–706, 2016. doi: 10.1016/j.ejor.2015.08.045.
F. Capitanescu, “Critical review of recent advances and
further developments needed in AC optimal power flow,”
Electric Power Syst. Res., vol. 136, pp. 57–68, July 2016.
doi: 10.1016/j.epsr.2016.02.008.
Biographies
Evangelos Vrettos is with Swissgrid Ltd., Aarau, 5001,
Switzerland.
Marc Hohmann is with Swissgrid Ltd., Aarau, 5001,
Switzerland.
Marek Zima is with Swissgrid Ltd., Aarau, 5001,
Switzerland.
p&e
march/april 2021
By Noritaka Ohashi,
Kazuki Terashima,
and Makoto Tanaka

E
ELECTRIC ENERGY SYSTEMS
throughout the world are now accommo-
dating a growing number of renewable
resources in addition to conventional gen-
eration. When responding to new intercon-
nection service requests, minimizing the
total cost of the power system, particularly
generation and transmission, is important.
Also, the maintenance of a large number
of existing power system facilities is neces-
sary, including older facilities, to take the
best advantage of their capabilities.

©SHUTTERSTOCK.COM/FANDIJKI

A Novel Approach
to Transmission
Bottleneck
Management
in Japan
Digital Object Identifier 10.1109/MPE.2020.3043569
An N−1 Intertrip Scheme

Date of current version: 19 February 2021

march/april 2021 1540-7977/21©2021IEEE ieee power & energy magazine 69

 Such a scheme would be particularly relevant in systems with—
or that expect—high renewable penetration not only in Japan but
also in many other countries/regions.
To this end, the Organization for Cross-Regional Coordina-
tion of Transmission Operators (OCCTO) in Japan has defined
the desired future evolution of the power system network as 1)
ensuring the appropriate reliability, 2) providing the smooth
use of the power system with a reasonable cost, and 3) ensur-
ing the soundness of power system facilities. OCCTO thus
commenced studies toward the Japanese version of connect
and manage, specifically N−1 intertrip and nonfirm access to
improve the efficiency of power system facilities. The phrase
“connect and manage” is also used, for example, in the United
Kingdom, where the Office of Gas and Electricity Markets
introduced a scheme in 2010 to provide earlier grid intercon-
nection to new and existing generation projects.
In this article, we focus on the N−1 intertrip scheme,
one measure in the Japanese connect and manage scheme, to
increase operational capacity by intertripping/curtailing gen-
erators instantaneously using a relay system in the event of a
single contingency. The N−1 intertrip approach can mitigate
transmission congestion and improve energy security.
In general, transmission congestion may decrease the
operational security level. This is because flexible resources
might be isolated from areas that need flexibility due to the
upper limit of operational capacity. If operational capacity is
expanded by the implementation of the N−1 intertrip scheme
and transmission bottlenecks are eliminated, the supply
(MW)
Operational
Capacity
ATC
Maximum Power
Flow (Assumed)
Probabilistic Evaluation of Power Flow
Contributes to Improving the Accuracy of the Assumed Maximum Power Flow
figure 1. An overview of the Japanese version of connect and manage. ATC: available transmission capacity.
70 ieee power & energy magazine
from flexible resources to areas that need flexibility may
increase under normal conditions, improving the normal
condition operational security level. Such a scheme would be
particularly relevant in systems with—or that expect—high
renewable penetration not only in Japan but also in many
other countries/regions.
On the other hand, the N−1 intertrip scheme requires
advanced control systems that can instantaneously locate
a failure and automatically curtail generator output to a
specific, effective level, and such instantaneous control
must be evaluated appropriately for its influence on the
reliability of the entire network. Thus, its application
requires careful design and highly reliable control sys-
tem implementation.
This article outlines the expected effects of the N−1 intertrip
scheme as well as the evaluation of the reliability required for
its application and highlights relevant points to consider when
determining the location and scale of application.
An Outline of Japanese Connect
and Manage
Overview of OCCTO
OCCTO was established in April 2015 as a government-autho-
rized organization that all electricity companies must join. The
Operational Capacity Increased
by the Application of
the N –1 Intertrip Scheme
Assumed Power Flow
Nonfirm Access
Facilitates the Use of ATC in
the System
march/april 2021
OCCTO’s goals and objectives are to maintain a stable electric-
ity supply and enhance the efficiency of supply systems as much
as possible. To achieve these objectives, OCCTO undertakes the
role of conducting studies by its advisory committee, from an
expert perspective, which is formulating regulations and codes,
developing long-term policies and cross-regional network devel-
opment plans, and studying network access application.
The Japanese Version of Connect and Manage
In the Japanese version of connect and manage, as illustrated
in Figure 1, studies are being implemented to introduce the
probabilistic evaluation of power flow and nonfirm access in
addition to the N−1 intertrip scheme. The probabilistic evalu-
ation of power flow is an effort to increase the available trans-
mission capacity (ATC) through the evaluation of a power
system based on the assumption of actual power flow, which
can be made by an evaluation of variable renewable energy
output and generation probability according to demand con-
ditions. Nonfirm access is the concept of generator connec-
tion without a network facility upgrade, in which new genera-
tors under normal conditions agree to curtail their output if
there is congestion.
Network Access Rule
In Japan, a new generator wanting to connect to a power
system is provided with a transmission capacity in the
order of their application as a general rule (i.e., a first-come,
first-served queue). Transmission and distribution utility
companies are obliged to provide network access under the
Electricity Business Act in Japan. Accordingly, the capacity
of the power system is secured in the order of application,
and when the system runs short of ATC, generation inter-
connection is granted only after the corresponding network
upgrades are completed.
Also, the present cost allocation rule requires network
upgrade expenses to be borne based on the beneficiaries-
pay principle that states that specific generators near a direct
connection attachment facility need to pay these upgrade
costs. In contrast, the network upgrade costs associated with
a bulk power system could be allocated to all network users
and not just the specific generators near the connection.
Under the access rule in Japan, the N−1 intertrip scheme
is expected to address new interconnection service requests
and minimize any increases in electricity costs.
The Basic Concept of the
N−1 Intertrip Scheme
Power System Development for
Generation Network Access
The N−1 criterion for all N−1 credible contingencies is the
basic concept for power system development of transmission
and transformation systems (7 kV or higher) with generation
network access. This concept, which is one of the typical
technical standards related to facility security and power
system stabilization, remains the same even after the appli-
cation of the N−1 intertrip scheme.
The N−1 intertrip is a mechanism to expand operational
capacity by instantaneously intertripping (or curtailing;
hereafter, intertripping includes curtailing) a generator using
a relay system in the event of a single contingency. Figure 2
illustrates the concept of the N−1 intertrip scheme.

No Contingency N–1 Contingency No Contingency N–1 Contingency

Line A Line A
Line A 100 MW/Circuit
100 MW/Circuit
100 MW/Circuit × One Circuit
Line A × Two Circuits
× One Circuit
100 MW/Circuit ATC 200 MW
× Two Circuits
100 ATC 100 MW 200
100 100 Trip Signal
(Curtailment)

Relay

100 0
100 100 100 100 100
New New Trip
Existing New Existing Existing Existing (Curtailed)
G G
G G G G G
Cannot Connect to the Grid Possible to Connect to
(Reinforcement Is Needed) the Grid
(a) (b)

figure 2. An overview of the N−1 intertrip scheme. (a) Present operation and (b) applying N−1 intertrip scheme.
G: generation plant.

march/april 2021 ieee power & energy magazine 71

 The operational capacity is the upper limit of power that ✔ a generator that has a capacity exceeding the allow-
can flow under normal conditions, and it ensures stable power able intertripping amount, which is set to ensure sup-
transmission even in the event of a single contingency. If the ply reliability
thermal capacity per line is 100 MW, the operational capacity ✔ a generator that has difficulty in restarting generation,
of a parallel two-line transmission system used to be set at with such a difficulty requiring a considerable period
100 MW considering a single contingency (short-term over- for reinterconnection and output increase to its previ-
load may be allowed in some cases). The N−1 intertrip scheme ous dispatch level
is designed to allow a 200 MW power flow (i.e., operational ✔ a must-run generator necessary to maintain system
capacity is set at 200 MW) under normal conditions on the reliability.
premise that power flow is limited to the thermal capacity per
one line (100 MW) by intertripping either an existing or new Applicable Power Systems
generator instantaneously in the event of a single contingency. Concerning the power system to which the N−1 intertrip
Figure 2 depicts an example where the new generator is sub- scheme is applied, there are concerns that are associated with
ject to intertripping. the application of the scheme to a bulk power system, including
Effective power system development can be achieved the following:
through an increase in operational capacity by the N−1 inter- ✔ A large intertripping amount is required, resulting in
trip scheme and the resulting decrease in network facility a significant reduction in supply capacity and a large
upgrades. On the other hand, a generator needs to be tripped decline in frequency.
in the event of a single contingency, requiring consideration ✔ The number of generators under control by intertrip-
to be paid to the impact on securing supply–demand balance. ping becomes extensive, resulting in a great number of
For this reason, if the impact of N−1 intertripping on reli- targets and an overly complex system.
ability is small, power system development should be imple- ✔ A significant impact is expected if intertripping fails
mented based on the application of the N−1 intertrip scheme. to operate as intended.
However, a network facility upgrade should be implemented These concerns are prominently visible in a bulk power
when intertripping has a major impact on reliability under system constituting a loop system. The concept of the appli-
generator scenarios such as cation of the N−1 intertrip scheme to power the system is
detailed in Table 1.
table 1. The concept for applying the N−1 intertrip
scheme to a power system. Intertripping Amount Allowable From
a Supply Reliability Standpoint
Transmission Power System Since a radial bulk power system may have a large inter-
(7 kV or Higher) Concept of Application
tripping amount under the N−1 intertrip scheme and its reli-
Bulk power system Loop Carefully determine for ability is affected significantly in terms of frequency and
(top two voltages) each system based on
the features of the loop reserve margin, it is necessary to set the allowable intertrip-
system1 ping amount in advance.
Radial Apply as a rule2
Power systems excluding bulk Apply as a rule2,3 table 3. An example of the allowable intertripping
power system amount for each facility subject to the N−1 intertrip
1 scheme (indicated in red).
Apply the N−1 intertrip scheme to loop systems that are
considered capable of being controlled simply. Intertripping Intertripping
2
Exclude the cases where the intertripping amount exceeds Amount Limit Amount
the allowable amount per failure of the facility subject to Necessary Limit Taking
the N−1 intertrip scheme. to Keep the Into Account
3
Exclude the cases where simple control is difficult because Frequency the Reserve
the system constitutes the loop system with a bulk power System Within Normal Margin of
system. Scale Variation Each Area
Area (MW)1 (MW)2 (MW)3
Area with large 56,530 400 2,550
table 2. The normal frequency variation range. system scale
Area Frequency Area with small 5,210 500 250
Hokkaido 50 ± 0.3 Hz system scale
1
Maximum power demand in 2018
Eastern Honsyu 50 ± 0.2 Hz 2
Normal frequency variation (low side): –0.2 Hz (–0.3 Hz
Western Honsyu 60 ± 0.2 Hz for Hokkaido and Okinawa)
3
Calculated based on the demand (2017 supply plan) of
Okinawa 60 ± 0.3 Hz each area during the light load period

72 ieee power & energy magazine march/april 2021

 In terms of frequency, the intertripping amount was set to
be within the normal frequency variation range (see Table 2) to
moderate the frequency variation immediately after failure in
consideration of the frequency of a single contingency in power
system facilities. The frequency of transmission line failures is
nearly 30 times greater than generation failures in Japan. On the
other hand, in terms of reserve margin, when a large-capacity
thermal power plant is intertripped in the event of a single con-
tingency, it takes time for reinterconnection, raising concerns
about a shortage of the reserve margin. Thus, it is also necessary
to consider the reserve margin of each utility supply area.
Accordingly, the allowable intertripping amount for each
facility subject to the N−1 intertrip scheme is, in principle, set
to be an amount within the normal frequency variation range
or an amount determined by considering the reserve margin of
each utility supply area, whichever is smaller (see Table 3). If an
intertripping amount exceeds the specified allowable amount,
the transmission and distribution utility company in each supply
area needs to carefully determine an appropriate intertripping
amount considering its impact on the power system frequency,
the failure frequency of the facility subject to the N−1 intertrip
scheme, and other factors.

Transmission Line A
200 MW/CCT × Two Lines

400 MW

Relay System

Trip Signal

250 MW
L G Generator
Offered to N –1 Settlement for
50 MW 200 MW Intertrip Scheme many generation
entities is required.
Subject to
N –1 Intertrip
Scheme

The generator becomes able to access G G G G G

the network without a network upgrade by
intertripping generators that connect to
a higher voltage transmission line. Beneficiaries (New Generators)

(a)

How to Estimate
Actual Fuel Cost Transmission
How to Collect
and Start-Up Cost and Distribution
Each Payment
Companies

Demand
Induction of
Total Losses
Payment Payment

Owner of Intertripping Generator Beneficiaries

(Existing Generator) (New Generator)
(b)

figure 3. The mechanism of the N−1 intertrip scheme to separate the target of intertrip from the payment.

march/april 2021 ieee power & energy magazine 73

 The Generators Subject to the
N−1 Intertrip Scheme
All generators newly connected to the transmission
power system (7 kV or higher) will be permitted
network access under the condition of implement-
ing the N−1 inter trip scheme, except in cases where
a network facility upgrade is necessar y, including
cases in which
✔ the generator, for example, a must-run generator nec-
essary to maintain system reliability, cannot partici-
pate in the N−1 intertrip scheme
✔ the power flow exceeds facility capacity as a result of
the connection of a new generator
✔ a network facility upgrade is economically more
advantageous than installing N−1 intertrip scheme
equipment (excluding the opportunity cost of loss
caused by the N−1 intertrip scheme).
The N−1 intertrip scheme requires the construction
of a highly reliable dedicated system because the relay
system is used instead of a network facility upgrade, and
an instantaneous response is required in the case of a
transmission contingency. In particular, if the intertrip-
ping amount, which is necessary to respond to a single
contingency in the transmission power system (7 kV or
higher), is to be provided by the relatively small-scale
generators connected to the distribution power system
(6.6 kV) (hereinafter called distribution generators), it
will be necessary to intertrip an enormous number of
distribution generators, and the N−1 intertrip system will
become more complex.
From the viewpoint of system reliability and certainty,
it is undesirable to intertrip distribution generators. But
considering that many renewable energy generators
in Japan are small scale and connected to distribution
systems, it is necessary to utilize the effect of the N−1
intertrip scheme in the connection of distribution gen-
erators and to pursue control of these assets. Thus, we
need to establish a mechanism that a generator subject
to the N−1 intertrip is not simply selected from newly
connected generators but from the most appropriate

Planned Outage Period for Maintenance

Immediately After N –1 Failure
(Including Continuation of N –1 Failure Period)
Ex-Ante Power Operational Capacity
Flow Is Large (With N –1 Intertrip Scheme)
D
Curtailment of C
Curtail All Generator Outputs
All Generator Facility Subject to N –1
B
Outputs Subject Capacity A Intertrip Scheme
to the N –1 Per
Intertrip Single Line Same as on the Left
Scheme Is
Required

Before After
Adjustment Adjustment

Ex-Ante Power Operational Capacity (With N –1 Intertrip Scheme)

Flow Is Small
Curtail on Last-In, First-Out Curtail on Pro Rata Basis
Basis Among Generator Among Generator Outputs
Outputs Subject to the N –1 Subject to the
Facility D Intertrip Scheme N –1 Intertrip Scheme
Capacity C D
B B C
Per A A B
Single Line A

Before After After

Adjustment Adjustment Adjustment

Generator Connected Within Operational Capacity (Not Subject to N –1 Intertrip Scheme)

Generator Subject to N –1 Intertrip Scheme (Prior Application)

figure 4. The operation during first-step application.

74 ieee power & energy magazine march/april 2021

generators from the viewpoint of power system impacts
and economic aspects. Moreover, generators should bear
the opportunity cost of loss associated with this measure
based on the degree of benefit they receive during nor-
mal operation. To this end, OCCTO is studying how to
calculate the opportunity cost of production losses and a
mechanism for settlement (Figure 3).
Since there is a great demand from generation entities
for early connection, the first step in the application of the
N−1 intertrip scheme started in October 2018. In the first
step, new generators seeking grid connection under the
N−1 intertrip scheme were designated as generators sub-
ject to intertrip control, i.e., the target generator of inter-
trip and opportunity loss cost payer is the same entity.

Substation
1L 2L

Transmission Line

Substation

1 2 Communication
Substation Overload Device (Sending)
Detector

3 –1 Communication Route
(Microwave, OP Wire)
Transmission Line
Relay Point
(If Needed)

3 –2
Communication Route
(Microwave, OP Wire)
Generation Station
Circuit Signal for Trip or Output Decrease
Breaker

5 4 Communication
Controlling
Device (Receiving)
G G Device

Generators

1
Overload Detector
• Detect Overload in the Event of a N –1 Failure
• Send a Trip or Output Decrease Signal According to the Amount of Overload
2
Communication Device
• Transfer Signal Generated by 1
3
Communication Route
4
Communication Device
• Receive Signal and Transfer to the Controlling Device (Owned by the Transmission and Distribution Utility Company)
5
Controlling Device
• Conduct Operation Based on the Received Signal (Owned by the Generator Owner)

* Precise Configuration Such as the Number of Communication Routes and the Way of
Communication Should Be Decided Based on the Grid Condition Subject to the N –1 Intertrip Scheme

figure 5. The system configuration of the N−1 intertrip scheme. OP: optical fiber.

march/april 2021 ieee power & energy magazine 75

 Operational Procedures for the respond after the occurrence of a failure in the short time
First-Step Application of the available. During the scheduled outage period (including a
N−1 Intertrip Scheme continuation of the N−1 failure period), curtailment on a
pro rata basis is introduced from the viewpoint of equal
Operation Method for the N−1 use (Figure 4).
Intertrip Scheme
When there are multiple generators subject to the N−1 inter- Overview of the N−1 Intertrip Scheme
trip scheme within the same power system, they are tripped System Configuration
on a last-in, first-out basis, in conformity with the power flow The basic system configuration of the N−1 intertrip scheme
condition. This is necessary for generators to immediately is depicted in Figure 5.

table 4. A demonstration case study of a possible N−1 intertrip application amount.

Available Operational Available

Capacity Capacity Capacity
Before the After the After the
Application Application Application Increase
Electrical Operational Ex-Ante of the N−1 of the N−1 of the N−1 Amount
Wire Capacity Power Flow Intertrip Intertrip Intertrip and Increase
Point Voltage Type (Single Line) Value Scheme (1) Scheme Scheme (2) Rate (2) - (1) Description
Point A 77 kV ACSR 130 130 0 200 70 +70 —
410 (100) (+54%)

Point B 275 kV TACSR 1,000 930 70 1,270 270 +200 Operational

610 (770) (770 + 500) (+20%) capacity after
the application
of the N−1
intertrip scheme
depends on its
upper control
limit
Point C 77 kV ACSR 130 100 0 200 100 +100 Available
410 (100) (+77%) capacity before
the application
of the N−1
intertrip scheme
depends on
Point F
Point D 154 kV TACSR 440 400 0 680 110 +110 Available
410 (340) (+25%) capacity
depends on
Point F
Point E 77 kV ACSR 170 100 0 260 160 +160 Available
610 (130) (+94%) capacity before
the application
of the N−1
intertrip scheme
depends on
Point F
Point F 275 kV TACSR 1,000 1,000 0 1,270 0 0 Operational
610 (770) (770 + 500) capacity after
the application
of the N−1
intertrip scheme
depends on its
upper control
limit
Total flow of Ex-ante After the After the +540 —
system H power flow intercon- intercon- (+28%)
1930 nection of nection of
the new the new
generator generator
2000 2,540
ACSR: aluminum conductors steel reinforced; TACSR: thermal-resistant aluminum conductors steel reinforced.

76 ieee power & energy magazine march/april 2021

Overload Detector
The overload detector has a specification that enables it to
minimize the number of intertripping generators by inter-
tripping an appropriate amount of power output based on
the extent of overload. The number of systems for overload
detection is determined by the degree of importance of the
target facility (e.g., a bulk power system needs two systems,
a regular use plus a spare).
Communication Device and Communication Route
The telecommunication system and number of routes of the com-
munication lines are determined by the degree of importance of
the target facility (e.g., bulk power system) and the installation
environment (e.g., mountain area, remote area, and so on).
Controlling Device
The control device receives data from the overload detector
and needs to be able to ensure reliable control. Since the con-
trolling devices belong to the generation entities, the main-
tenance (including expenses) of the devices falls under their
responsibility.
A Demonstration Case Study for the
N−1 Intertrip Scheme
Table 4 displays the result of a demonstration case study
made for the power system illustrated in Figure 6 to examine
the effect of the N−1 intertrip scheme. Operational capac-
ity is expanded by the implementation of the N−1 intertrip
scheme, and transmission bottlenecks are eliminated.
The assumptions in the case study analysis are as follows:
✔ The N−1 intertrip scheme is applied to the bulk power sys-
tem, which is radial, considering the impact on reliability.
✔ The maximum intertripping amount by the N−1 in-
tertrip scheme allowed for a single transmission line
is 500 MW.
✔ The operational capacity is doubled by the N−1 inter-
trip scheme.
✔ All generators are eligible for the N−1 intertrip scheme.
✔ The available capacity is calculated assuming that inter-
connection is made from the lower to the upper system.
The actual applicability of the N−1 intertrip scheme and
additional available capacity, which are calculated by the
transmission and distribution utility companies in each area,
can be found on each company’s website (Table 5).
Future Work
In this article, we reviewed the application of the N−1 inter-
trip scheme in Japan. The initiatives brought about through
the Japanese version of connect and manage were estab-
lished as rules after the consensus of concerned parties and
deliberation by the Cross-Regional System Development
Committee, an advisory committee of the OCCTO. Future
work includes the following.
The Application of the N−1 Intertrip
Scheme to Distribution Generators
With the start of the first-step application of the N−1 inter-
trip scheme, it becomes possible for new transmission-level

Loop Bulk Power System

500 kV 500 kV 500 kV

2,540 ← 1,930 System H
275 kV
930 →1,270 1,000 → 1,270
Point B Point F
275 kV
400 → 510 100 → 260
500
200←130 400 400 270
154 kV 77 kV
77 kV 400 →510
100 → 260
200 ←130 Point D 154 kV
Point A Point E
300 10
77 kV
100 160
70 130 Point C 100 → 200

Legend

100 : Existing Generator 100 : New Generator (After Application of N –1 Intertrip Scheme)

Numbers in Red: Power Flow After Application of the N –1 Intertrip Scheme

figure 6. A demonstration case study of the effect of N−1 intertrip scheme (i.e., a simulated system).

march/april 2021 ieee power & energy magazine 77

 table 5. An example of additional available capacity with the N−1 intertrip scheme.

Available Additional
Installed Capacity Available
Thermal Without Availability Capacity
Capacity the N−1 of the with the
Number (100% × Operational Intertrip N−1 N−1 Intertrip
Transmission Transmission Voltage of Number Capacity Constraint Scheme Intertrip Scheme
Number Name (kV) Circuits of Circuits) (MW) Factor (MW) Scheme (MW)
1 A 500 2 6,580 3,290 Thermal 2,283 Not —
capacity available
2 B 500 2 13,162 6,581 Thermal 5,737 Not —
capacity available
3 C 275 2 1,532 880 Thermal 568 Available 651
capacity
4 D 275 2 612 612 Thermal 284 Not —
capacity available
5 E 275 2 1,236 648 Thermal 334 Available 587
capacity
6 F 66 2 64 36 Thermal 0 Available 27
capacity
7 G 66 2 102 58 Thermal 0 Available 43
capacity
8 H 66 2 208 109 Thermal 0 Available 98
capacity

generators to connect to power systems without network For Further Reading

facility upgrades, and this offers a great benefit, but there still
is an issue concerning network access by distribution genera-
tors as previously described. To realize the full application of
the N−1 intertrip scheme, it is necessary to accurately calculate
the opportunity cost of production losses caused by the N−1
scheme as well as establish a cost settlement system involv-
ing many generators based on the degree of their benefit. In
response to this, OCCTO is studying a system that will be
acceptable to both the target generator of the N−1 intertrip and
the cost bearers. OCCTO plans to put it into application within
the 2022 fiscal year.
The Introduction of Nonfirm Access
The shortage of ATC in bulk power systems caused by the
increase in the amount of generation interconnection by the
N−1 intertrip scheme results in an increase in the need for
network upgrades. Under the present cost allocation system
in Japan, upgrades may lead to inefficient power system
development and an increase in electricity costs. There-
fore, OCCTO will adopt a system by which the need for a
network upgrade is appropriately determined by using an
evaluation of a cost–benefit analysis as well as comprehen-
sively taking into account the cost-effectiveness and feasi-
bility of mitigation measures. At the same time, OCCTO
is studying nonfirm access as a new generation intercon-
nection mechanism for the power system when network
upgrades are inappropriate.
“Delivering our environmental future: Annual statement,”
National Grid Electricity Transmission, Warwick, 2015.
Accessed Nov. 6, 2020. [Online]. Available: https://www
.nationalgrid.com/uk/electricity-transmission/document/
129751/download
“On the concept of first-step application of N−1 Inter-trip
scheme in the development of power system facility planning
(Network Code, Article 55) (in Japanese),” The Organization
for Cross-Regional Coordination of Transmission Operators,
Tokyo, Japan, 2018. [Online]. Available: https://www.occto
.or.jp/access/oshirase/2018/files/181001_n-1densei.pdf
“Business plan FY 2019 of OCCTO,” The Organization
for Cross-Regional Coordination of Transmission Opera-
tors, Tokyo, Japan, 2019. [Online]. Available: https://
www.occto.or.jp/en/about_occto/articles/files/Business_
Plan_FY2019.pdf
Biographies
Noritaka Ohashi is with the Organization for Cross-Re-
gional Coordination of Transmission Operators, Tokyo, 135-
0061, Japan.
Kazuki Terashima is with the Organization for Cross-
Regional Coordination of Transmission Operators, Tokyo,
135-0061, Japan.
Makoto Tanaka is with the National Graduate Institute
for Policy Studies, Tokyo, 106-8677, Japan.
p&e

78 ieee power & energy magazine march/april 2021


The Fragile Grid
The Physics and Economics of Security
Services in Low-Carbon Power Systems
W
WORLDWIDE, THERE ARE UNSTOPPABLE FORCES
toward low-carbon power systems that can support the fight
against climate change and help solve the security of supply
issues in many countries. Low-carbon grids are likely to be
characterized by substantial renewable energy sources, both
centralized and distributed, combined with intelligent and
dynamic demand-side technology and multisector electrifi-
cation (including heating, transport, and future fuels). In this
context, successfully resolving the “affordability-sustain-
ability-reliability” energy trilemma is crucial for paving the
way to low-carbon energy futures.
When looking at the “reliability” pillar of the trilemma, it
becomes apparent that these transformations are of unprece-
dented scale and pose significant threats to operational secu-
rity. Many of these are currently being experienced in the
Australian National Electricity Market (NEM), which has
By Pierluigi Mancarella and Farhad Billimoria
Digital Object Identifier 10.1109/MPE.2020.3043570
Date of current version: 19 February 2021
march/april 2021 1540-7977/21©2021IEEE
©SHUTTERSTOCK.COM/SERGIOSH
one of the highest instantaneous renewable energy penetra-
tions in the world, with 75% projected by 2025.
Due to changes in the physical characteristics of the sys-
tem and in the supply- and demand-side resources connected
to it, low-carbon grids might be more intrinsically fragile
and, therefore, both less secure and less resilient. The sys-
tem’s new physics has a profound impact on its economics
too—in the requirements for security services and in the
technical, commercial, regulatory, and policy measures to
procure such services. These aspects are thoroughly dis-
cussed in the rest of this article.
Low-Carbon Grids: Technical
Challenges and Mitigation Options
Several technical challenges are being experienced in sys-
tems that are increasingly operating with large shares of
ieee power & energy magazine 79
 renewables and inverter-based resources (IBRs). In particu-
lar, frequency control and balancing challenges are promi-
nent in systems that are weakly interconnected, connected
to other grids via high-voltage direct current links, or even
completely isolated, such as Australia’s system. Other chal-
lenges, generally associated with voltage sensitivity and sta-
bility, are more typical of systems where resources and load
centers are linked by long, relatively low-density networks,
which, again, is a widespread situation in Australia. Next
we discuss some of the major technical challenges of low-
carbon operation with real-case studies from Australia.
Frequency Control and Inertia
Frequency control is inherently more challenging in an envi-
ronment with more variable and asynchronous resources. First
of all, lower system inertia means that the system frequency
becomes much more sensitive to active power disturbances.
In particular, large disturbances associated with contingen-
cies may lead to very high initial rates of change of frequency
(ROCOF) and to undesired minimum (nadir) or maximum
(zenith) frequency values at time scales too fast for conven-
tional primary or even emergency responses to alleviate.
For example, Figure 1 illustrates how the system response’s
dynamic characteristics change with inertia levels while the
quasi-steady-state frequency level does not. The latter only
depends on the available primary frequency response (PFR),
which was assumed to be the same in the two cases. In the
case of the South Australian blackout in 2016, when South
Australia islanded from the rest of the grid and less than
40% of available generation was synchronous, the estimated
ROCOF was about –6 Hz/s. With that rate of frequency drop,
“under-frequency load-shedding” schemes were not able
to intervene. Also, with increasingly larger shares of loads
interfaced via power electronics, the frequency damping by
load will inherently reduce dramatically, again, calling for In addition to the issue of decreasing inertia, grids with
both faster and larger amounts of frequency response. Such fewer synchronous generators and more IBRs are also
50
Same Quasi-Steady-State
49.9 Lower
Frequency Value
Inertia
49.8
Higher
Frequency (Hz)
49.7 Higher Inertia
ROCOF
49.6
49.5 Lower
49.4
Frequency
Nadir and
49.3 Reduced
Nadir Time
49.2
0 5 10 15
Time (s)
figure 1. The effects of lower inertia on system frequency response after a given
contingency event (adapted from Mancarella et al., 2017).
80 ieee power & energy magazine
faster and/or larger reserves might not be readily available, or
the market or regulatory setup may be unsuitable to provide
the right incentives in the short term.
Variability, Uncertainty, Visibility, and Flexibility
Systems dominated by renewable energy sources, particu-
larly wind and solar, feature a much higher degree of opera-
tional variability and uncertainty (due to forecast errors) over
various time scales, from seconds and minutes (impacting
regulation and load-following reserves) to hours (potentially
requiring new types of flexibility and ramping reserves).
This situation is further compounded by the fact that many
distributed energy resources (DERs) embedded in distribu-
tion networks are not visible at the system level.
This means that the demand side, whose forecasting
has historically been quite accurate, becomes more dif-
ficult. Moreover, new forms of uncertainty are being wit-
nessed, including “commercial contingencies” through self-
curtailment by generators within dispatch periods in response
to low or negative prices and lack of visibility on the curtail-
ment of DERs.
All of these issues call for more active power flex-
ibility and new flexibility service providers. Conventional
resources are also starting to respond to new requirements.
For example, in Australia, coal power plants are investing in
making themselves more flexible so that they can cycle on
and off during the middle of the day (due to, for example,
low prices resulting from large amounts of solar generation).
However, there is a system risk to this as well, once again
illustrating the complexity of the energy trilemma. This is
the time when such generators might be most needed to pro-
vide inertia as well as system strength.
System Strength and Local Stability Challenges
weaker and more susceptible to
voltage disturbances. In particu-
lar, the increased voltage sensitiv-
ity to reactive (as well as active)
power d ist u rba nces a nd t he
decreasing availability of short
circuit current means that volt-
age management becomes much
more difficult, especially in long,
high-impedance networks in rural
areas where many large-scale
renewables are connected.
The term system strength has
been introduced as an umbrella
concept to indicate a general mea-
20 25
sure of power system stability,
particularly in terms of a system’s
ability to maintain voltage wave-
form and phase under generic
march/april 2021
operating conditions, including following an outage. System
strength has typically been associated with the fault current
level available at a given location and measured via metrics
such as the short circuit ratio.
However, several emerging instability issues have also been
recently witnessed in the presence of IBRs and fast control
loops, calling for new considerations and metrics besides fault
current level. For example, in strong grids, IBRs normally use
a phase-locked loop control as the voltage synchronization
mechanism that makes up for the lack of inherent synchro-
nism. However, as very weak grids with large shares of IBRs
may experience significant voltage oscillations and distortions
during and after a disturbance, existing controls may be inade-
quate to track the voltage angle and maintain synchronization.
This has led the Australian Energy Market Operator (AEMO)
to make system-strength-related security interventions, such as
the enforcement of a minimum number of synchronous gen-
erators that stay online, the curtailment of renewable output,
and the procurement of synchronous condensers.
Other solutions to reduce the negative impacts of low
system-strength conditions include the retuning of IBR con-
trollers and fast reactive power support. Several ongoing,
large-scale battery-storage projects are also being trialed
throughout Australia. Some of these include newer grid-form-
ing converters, which do not use a phase-locked-loop-based
architecture but generate the reference voltage themselves,
and virtual synchronous machines that emulate the dynamic
characteristics of conventional synchronous generators.
System Fragility and the Risk of Cascading
Also linked to decreasing system strength and inertia, many
renewables, DERs, and their protection systems may gen-
erally be more sensitive and vulnerable to both frequency
and voltage excursions and, therefore, be more prone to cas-
cading (what we can generally indicate as “fragility”). This
includes inadequate ride-through capability, insufficient
local reactive support in response to voltage transients, or
unsuitable or not fully coordinated protection settings.
Additionally, there are risks related to interlinkages
between system parameters and security conditions, such as
voltage-induced frequency dips and uncoordinated or poorly
understood dynamic response from new technologies (see
the case study next). All of these situations highlight the
greater fragility of an increasingly complex grid with large
shares of IBRs and multiple interlinked dynamics, which
calls for new operational tools and technologies to manage
system security.
The Role of New Operational Security Tools
In the face of all of the emerging technical challenges illustrated
previously, several advances have been proposed to deal with
fast changing system conditions, including new optimal power
flow and scheduling tools to securely commit and dispatch
resources in low-inertia systems as well as machine-learning-
based tools to deal with the variability and uncertainty of differ-
march/april 2021
ent natures. For example, the frequency-response security maps
developed during the Finkel Review (Figure 2), which aimed to
review the status of the NEM following the South Austra-
lian blackout of September 2016 and move forward toward a
renewables-rich future, can be used to represent the effects of
different parameters and the utilization of resources on secu-
rity requirements. Such maps depict the secure area (shaded)
in which the relevant three-frequency-response parameters
would be constrained to meet suitable security criteria (e.g.,
maximum acceptable ROCOF, minimum acceptable nadir,
and the minimum level of quasi-steady-state frequency)
for different levels of PFR and system inertia. The verti-
cal line depicts the level of inertia that corresponds to the
desired ROCOF limit while the horizontal line corresponds
to the static PFR requirement (currently implemented in the
NEM and in most jurisdictions worldwide) for the desired
quasi-steady-state frequency limit. The hyperbolic-like
curve depicts the desired limit for the frequency nadir for
under-frequency events (similar maps can be drawn for over-
frequency events and even for multiple areas in the case of
system separation).
For typical parameters that might be used to constrain the
system (e.g., a postcontingency frequency nadir not lower than
49 Hz) and typical load-damping factors (e.g., 3%/Hz) high-
inertia systems with plentiful synchronous generation are
typically constrained by static requirements that are indepen-
dent of the system inertia, i.e., the secure operating envelope
is bounded by the horizontal line. However, when transition-
ing into low-inertia systems, both the frequency nadir limit
(the hyperbolic curve) and the ROCOF limit (the vertical line)
may become binding to meet the operational security require-
ments. Therefore, the scheduled PFR should be higher than
the static requirement typically adopted in traditional sys-
tems with plentiful inertia and should also be sized (or co-
optimized), taking into account the incumbent system inertia.
This approach also provides an intuitive way to visualize (and
even formally quantify) the transition into low-inertia condi-
tions from the perspective of frequency management.
Nadir Requirement
Static Requirement
ROCOF Requirement
Secure Region
PFR
High Inertia
Low Inertia
Aggregated Inertia After Contingency
figure 2. A frequency-response security map for low-
frequency events with an illustration of the conventional
transition from “high-inertia” to “low-inertia” conditions
(adapted from Mancarella et al., 2017).
ieee power & energy magazine 81
 Frequency-response security maps have been widely used conditions for the Wholesale Electricity Market operated in
by AEMO in its 2020 Renewables Integration Study and the South West Interconnected System. These maps can also
implemented in AEMO’s control room in Western Australia be used to assess the implications of using fast frequency-
for the real-time online security assessment of low-inertia response resources to support slower generation units or
modulate the size of the largest contingency in the system to
decrease the incumbent security requirements.
In this respect, Figure 3 illustrates a 3D security map
ROCOF Requirement
considering the interaction of a variable contingency size,
0 with the amount of inertia and PFR shown. From the exam-
2,50 Nadir Requirement ple, it can be appreciated how the ROCOF and nadir require-
0 ments decrease greatly (for the nadir, very nonlinearly) with
2,00 Static Requirement
the contingency size. This suggests that ramping down the
0
1,50
PFR (MW)

output of the largest online generator could significantly help

0 meet security requirements, and this effect is much larger for
1,00
decreasing inertia conditions.
500
70
A natural implementation of the multiparametric secu-
In 0
er 60 0 rity maps discussed previously is the development of a fre-
tia 10 0 5 0 )
Af 2 30 00 W quency-response, security-constrained, optimal power-flow
te 0 4 ( M
rC 4 00
ze tool that could co-optimize energy, PFR (and fast frequency
on 0
tin 5 0
30
0 Si
ge 6 20 cy response), inertia, and the largest system contingency. The
nc 0 0 en
y( 7 0 1 t i ng largest contingency may refer to generation/demand loss as
GW 8 00 n
s) Co well as, more generally, to interconnector flows in case there
is a risk of system split. Such tools are now being considered
figure 3. A frequency-response security map to comply for implementation in both the NEM and Western Australia.
with the desired nadir, ROCOF, and quasi-steady-state There are also various progressive approaches under-
frequency limits considering variable contingency size taken by AEMO to deal with variability and uncertainty
(adapted from Puschel et al., 2020). by using data-driven artificial intelligence algorithms. For

table 1. The fragility of low-carbon grids: Challenges and potential solutions.

Risk Emergence Mitigations
Frequency • Sustained frequency excursions • Minimum inertia levels
control and (regulation) • Compulsory droop-based governor response
inertia • High ROCOF following contingency • An additional amount of PFR
• Insufficient regional inertia • The co-optimization of energy, frequency response, and
• Insufficient PFR (regional and system-level) inertia
• The risk of low inertia and insufficient PFR • The regional allocation of reserves
after separation • New sources of fast frequency response (e.g., batteries,
electrolyzers)
• The management of largest contingency and interconnector
flows (system at risk of regional separation)
Variability and • A large variation in net demand • Better forecasting
uncertainty • Insufficient short- and medium-term and • Machine learning to assess reserves (e.g., dynamic Bayesian
ramping reserves belief network tools)
• The use of more flexible resources, including energy storage
(e.g., pumped hydro)
DER visibility • Photovoltaic generation (and generally • Connection standards
DERs) trip following transmission • The setup of distributed marketplaces for DERs to interact
frequency and voltage events with wholesale markets
• Commercial security issues related to fast • Advanced transmission system operator/distribution system
price-driven response from DERs operator interface and coordination
System strength • Fault current shortage • Minimum levels of inertia and fault current (out-of-merit
• Voltage instability dispatch of conventional generators)
• Sustained voltage oscillations after fault • Synchronous condensers
• Fault ride-through issues • Static synchronous compensator and static VAR
compensators to improve voltage stability
• Improvements of internal generator controls (especially in
solar and wind farms)
• Grid-forming converters and virtual synchronous machines

82 ieee power & energy magazine march/april 2021

example, dynamic Bayesian belief networks are currently
being employed by AEMO to assess the requirements for
short- to medium-term reserves, including the forecasting
of several uncertain parameters, such as conventional gen-
eration availability, demand, renewable output, weather, and
market prices. Such a tool could potentially also be extended
to forecasting ramping and various flexibility requirements.
A summary of solutions being implemented or have been
proposed or discussed in Australia to deal with different sys-
tem security issues is presented in Table 1.
Case Study: Australian System Cascading and
Separation Event, 25 August 2018
On Saturday, 25 August 2018, lightning strikes on the
transmission towers that support the two circuits of the
330-kV Queensland–New South Wales interconnector led
both circuits to trip, leaving Queensland islanded. As the
Queensland–New South Wales interconnector was trans-
ferring 865 MW from Queensland to New South Wales,
Queensland experienced over-frequency conditions while
the remainder of the NEM experienced low frequency. The
generating units in South Australia, including the Hornsdale
battery system, increased their output to restore the system
frequency, which led to a rapid rise in active power flowing
through the Heywood interconnector. The interconnector
eventually tripped due to the activation of dynamic protec-
tion mechanisms—about 8 s after the Queensland–New
South Wales interconnector trip.
At the time of separation, Hey-
wood was transferring 430 MW
from South Australia to Victoria.
This caused a frequency rise in 51
South Australia and a further fre-
quency drop in New South Wales 50.5 QNI Trip
and Victoria, which led to the acti- 13:11:39 h
vation of multiple under-frequency
Frequency (Hz)
load-shedding schemes (around 50
1 GW in total) in these two states.
Droop response and automatic gen- 49.5
Heywood
eration control from synchronous Trip
generation then restored the fre- 49 13:11:46.901
quency in Victoria and New South 13:11:45
AUFLS-TAS
Wales, followed by the restoration (Smelter)
48.5
of loads and resynchronization of
13:11:30
the separated states within approx-
imately 2 h. The event frequency
traces for the first minutes of the
event can be seen in Figure 4.
Although the NEM as a whole Legend:
QNI – Queensland-New South Wales Interconnector
was not, strictly speaking, operat- TAS – Tasmania
ing under low-inertia conditions UFLS – Under-Frequency Load Shedding
at the time, the separation event AUFLS – Automated Under-Frequency Load Shedding
illustrates the complexity and,
to some extent, the fragility of a figure 4. The 25 August 2018 separation event frequencies in the different states
large interconnected system that (Source: AEMO).
march/april 2021
may cascade and split into different areas. In particular, the
ROCOF value in Queensland following the interconnector
trip was up to approximately 0.8 Hz/s. This relatively high
value is attributed to local low-inertia conditions in the
separated system due to 23% of the total generation being
nonsynchronous and consisting mostly of photovoltaic units.
Additionally, on one hand, while renewables and DERs
helped support system security by providing some over-fre-
quency response, uncertainty in their response underscored
the importance of both increasing DERs’ visibility and mak-
ing the system less fragile. On the other hand, given that
the interconnector’s trip was not considered a credible event,
renewables and DERs contributed (through their frequency
response) to providing system resilience to deal with a non-
credible event that might have otherwise ended up in further,
high-impact cascading behavior.
Special consideration should go to the response of the
100-MW/129-MWh Hornsdale Power Reserve plant located in
South Australia, which, at the time, was the largest lithium-
ion battery system in the world. During the separation event,
the Hornsdale plant provided an extremely fast response for
both low and high frequency when the frequency drop in South
Australia switched to a rise following the trip of the Heywood
interconnector (Figure 5). As it can be appreciated, this is an
impressive performance in terms of the subsecond speed of
activation and ramping capability to mirror the frequency
variation by droop control and, most importantly, in terms of
QNI and Heywood Trips–25 August 2018
Frequency in Queensland
Frequency in South Australia
Frequency in Victoria and
New South Wales (“Mainland”)
Frequency in Tasmania
13:11:47
UFLS-Mainland
(Smelter and Residential)
13:11:40
13:11:50
13:12:00
13:12:10
13:12:20
13:12:30
13:12:40
13:12:50
13:13:00
Time (hh:mm:ss)
ieee power & energy magazine 83
 its ability to switch the active power direction following the ing to system security. In this respect, while the challenges
frequency’s inversion. of resource adequacy and reliability often relate to critical
However, such responses might create issues in an periods over a year and are a matter of planning, keeping
interconnected system, for example, because too fast of an a system operationally secure is an everyday (and even an
increase in interconnector flows might activate protection every-second) challenge that requires careful consideration
relays and trip the interconnector. In the specific case of the within a market environment.
separation event, it is not clear whether the fast response from As previously discussed, new forms of technical risk
the battery did contribute to the trip of Heywood. However, are being introduced from intermittent and inverter-based
authors’ simulations that were analyzed to emulate the event sources of supply and demand. From a market perspective,
and study potentially similar cascaded outages suggest this scarcities are emerging in system services that were previ-
is a material possibility. Therefore, moving forward, it will ously delivered as a by-product of energy, such as inertia
be particularly important to carefully plan the location of and system strength provided by synchronous generators
fast frequency-response resources, considering interconnec- that would be online to produce energy (and make money in
tor transfer limits and coordinating their potential response energy markets) and would also provide these other services
with interconnectors’ protection settings. (and in many cases without being specifically remunerated).
Finally, increasing interactions among different power sys-
Economy, Market, and Regulatory tem parameters with the potential for cascading and corre-
Challenges and Opportunities lated system events (what we previously called “system fra-
Following the previous technical discussions and consider- gility”) are being observed.
ing all of the potential solutions to the emerging challenges The economic implications of a failure to manage system
that have been proposed, the questions now arise of what is security are serious. Scarcities of system services may force
the best way to incentivize the development of the most suit- market operators to increasingly intervene in the market.
able solution within a market environment, what the role of This may also lead to more curtailment of renewable plants,
regulation and energy policy is, and what the best approach delays to connection, and stagnating investment. This has all
would be. been witnessed in Australia.
On a starker level, inadequate service provision could
The Economics of System Security: From the lead to system outages and major economic loss, especially
Bottom Up in the context of a more fragile grid that is less resilient to
From the considerations made in the previous sections, it is high-impact, low-probability events, as illustrated previ-
clear that the transition of our electricity system to a low-car- ously in the case of the 25 August 2018 separation event. A
bon future with large shares of variable and partly uncertain key challenge for policy makers and market designers is to
renewable energy and with pervasive IBRs is profoundly develop cohesive economic and regulatory frameworks for
changing (or has already changed) the nature of risks relat- operational security that efficiently procure the services nec-
essary to secure the system to an
appropriate level of risk and that
Hornsdale Battery Plant–25 August 2018 effectively manage the integration
80 50.6 within the broader market design
for electric supply.
60 50.4
In this respect, cross-market co-
40 50.2 optimization, for energy, operating
Active Power (MW)

Frequency (Hz)

20 50 reserves, inertia, largest contin-

gency, and so on, has been shown
0 49.8 to be technically achievable. In
–20 49.6 fact, it is already being pursued in
Western Australia. However, inte-
–40 49.4
gration among other different ser-
–60
Battery (MW) Frequency (Hz) 49.2 vices is lacking. For example, sys-
–80
tem strength, which is also linked
49
to inertia, is managed as a local
13:11:30

13:11:40

13:11:50

13:11:60

13:12:10

13:12:20

13:12:30

service, in contrast with the more

regional or system level nature of
Time (hh:mm:ss) the others. Furthermore, the right
economic signals should be incor-
figure 5. The Hornsdale battery plant’s active power response versus frequency in porated to support adequate invest-
South Australia during the 25 August 2018 separation event (Source: AEMO). ment over the long term, but this is

84 ieee power & energy magazine march/april 2021

 A bottom-up approach can provide nuanced insights into the
incentive compatibility of alternative mechanism designs to foster
the participation of different, innovative technologies.
not obvious when viewed across multiple markets and tech-
nologies that may compete to provide different services.
An array of market and regulatory mechanisms is avail-
able to the policy maker, including mandatory licenses,
regulated monopoly provision, and organized markets. The
choice of mechanism, or a combination thereof, is a critical
policy decision, along with the delegation of responsibility to
the operating parties and the frameworks that guide provid-
ers and users’ incentives.
One method to designing policy for system security is
through a top-down approach that begins with the economic
design that then flows down to system parameters and control
frameworks. However, this approach risks applying an eco-
nomic framework that is inconsistent with the fast-changing,
underlying physics of the system. It also risks being bound,
unnecessarily and in some cases undesirably, to traditional
and established market constructs, which might be less
applicable to the new system dynamics being observed.
By contrast, a bottom-up approach begins with the phys-
ics of power networks. This approach can provide a more
granular appreciation of the economic characterization of
new security services that may be required during the low-
carbon transition. It can provide nuanced insights into the
incentive compatibility of alternative mechanism designs to
foster the participation of different, innovative technologies.
The Physical Characteristics
Informing the Economics
System security is a multifaceted concept and concerns
many different aspects of power system physics. It moves
well beyond the static or dynamic balancing of active and
reactive power, as could be appreciated from the previous
technical discussions. Increasingly relevant are the interac-
tions among various physical parameters of generators, net-
works (and associated equipment), and load. Next, we high-
light critical features of aspects of system security that can
inform their economic characterization.
✔ The multiplicity of provision: Unlike the classical de-
lineation between the supply, transmission, and con-
sumption of the electricity commodity, system secu-
rity has a multiplicity of providers, which can include
generators, load, storage, and network equipment.
For example, inertia can be provided by synchro-
nous generation and load, synchronous condensers,
and IBRs in a virtual or emulated form. This kind of
service is closer to a fast frequency response than to
synchronous inertia per se. This means that the pro-
march/april 2021
vision of services can also cross over between differ-
ent industrial structures, such as between competitive
electricity supply and regulated network monopolies.
Hence, the designation of economic responsibility
for aspects of security takes on added weight and should
account for the efficiency of competitive provision, rel-
ative to the efficacy of regulated monopoly provision.
✔ Complex multiuse goods: Some physical parameters ap-
ply to the system as a whole, others are highly locational,
and, finally, others have both system and locational impli-
cations. Take the example of inertia, which is essentially
the kinetic energy stored in rotational turbines. While
much of the previous analysis has focused upon the im-
pact of inertia on slowing the system-wide frequency
degradation, inertia is also important for local angular
stability and system strength. Furthermore, droop-based
PFR is relevant for frequency control but may also affect
oscillatory (small-signal) stability constraints. In eco-
nomic terms, these are complex multiuse goods in ad-
dition to being technically complex services. Thus, any
design of system security frameworks must reflect the
full contribution of the good-to-system requirements.
✔ The rivalry and excludability of consumption: Sys-
tem services are often categorized as public goods.
Public goods are generally characterized by two fea-
tures: 1) they are nonexcludable—users of the good
cannot be easily excluded from the enjoyment of the
good and 2) they are nonrival—where the consump-
tion of the good by one user does not affect the abil-
ity of others to enjoy the good. While this may apply
to some security services, such as PFR services, a
closer examination of the underlying physics suggests
that not all system services neatly fit into the classi-
fication of public goods (see Figure 6). For instance,
some system services may be nonexcludable but dis-
play rivalry, especially under congestion, such as
fault current levels and voltage response. Also, other
services may be nonrival but excludable. An example
of this is tertiary reserves (sometimes called replace-
ment reserves and linked to ramping and flexibility
reserves), where it may be possible to curtail users
that do not wish to pay for the product.
✔ The inseparability of goods: The provision of many
essential system services by synchronous resources is
inseparable. For example, the provision of inertia and
the fault current level support by a synchronous gener-
ator are inseparable. This means that the way in which
ieee power & energy magazine 85
 one good is procured affects the other. This introduces approaches to commitment (and therefore time-ahead
new questions for markets and regulation. Are some markets) and dispatch. Examples of binary services
products to be procured separately or together, i.e., are are inertia, fault current levels, and system strength.
they individual products or super products? Should par- The binary nature of the service has significant impli-
ticipants be compensated for the provision of one prod- cations for price formation.
uct when scheduled for the other? This underscores the ✔ Product definition: Another key aspect to consider
complexity of situations where system services have is the definition of the product or service itself. In
different characterizations but are inseparable. For ex- fact, for a market product or good to be clearly trad-
ample, in South Australia, synchronous generators have able, it must be well defined and measurable. This is
been directed by AEMO to commit and generate at relatively straightforward for a few services (e.g., ac-
minimum levels to contribute to local system strength; tive power services) but far from trivial in emerging
however, they also provide system inertia at the region- cases. For example, as we mentioned earlier, “sys-
al level. This obviously may have an impact on other tem strength” is usually used as an umbrella concept
technologies that, in the future, might compete to pro- that refers to general stability considerations and has
vide this service. Furthermore, as inertia is also linked been historically associated with a local fault cur-
to frequency-response requirements, these services will rent level. However, synchronous inertia also plays a
also be indirectly affected by the requirements for sys- certain role, and it is not yet fully clear how emerg-
tem strength and the relevant provision. ing forms of instability are physically linked to these
✔ The binary nature of provision: The provision of cer- parameters and others as well. Therefore, what is the
tain system security services is based on binary re- market product that we can trade when we talk about
sponse. In other words, they depend on whether the system strength?
unit is synchronized with the system. The unit pro-
vides its full response if it is synchronized and zero Market Development Versus
if it is not. These responses do not scale up or down Regulatory Approaches
with the level of generation, which is important for There are a variety of options available to policy makers to
manage power system security,
ranging from market mechanisms
Excludable such as auctions and spot markets,
nonmarket mechanisms such as
mandatory license conditions and
Club Goods Private Goods
regulated network procurement,
to operational mechanisms such
as constraints and curtailment.
Table 2 illustrates some relevant
exa mples f rom Australia and
around the globe.
Excludability Through The following considerations
Curtailment Regimes are relevant in deciding upon the
Dynamic/Tertiary optimal approach for service pro-
Reserves curement:
Nonrival Rival ✔ Product standardization and
Fault Levels
uniformity: Market-based ap-
PFR Dynamic Rival With
Voltage
proaches rely upon the speci-
Secondary Increasing
fication of a standardized
Frequency Static Voltage Congestion
product that can be traded in
Response
Subclasses of a liquid and transparent man-
Inertial
Users May ner and under a clear formu-
Impact the Value lation. Commoditization re-
Response
of the Product lies upon a clear and accepted
Common Pool
Public Goods Resources definition of the service,
which for many emerging
Nonexcludable system security phenomena
is not yet available, as previ-
figure 6. The economic properties of security services (adapted from Billimoria ously mentioned for the case
et al., 2020). of system strength.

86 ieee power & energy magazine march/april 2021

 ✔ The quantification of the service requirement: This
is critical for market-based approaches but also relevant
for regulatory approaches. While this may be carried
out for some system requirements, such as relating to
the co-optimization of inertia and frequency response,
it is challenging for other emerging security phenom-
ena, again, such as system strength in weak grids. In the
latter case, simple standardized metrics fail to capture
the technical issue in its entirety while more accurate,
dynamic simulations are computationally intensive and
highly sensitive to changes to assumptions on operating
conditions, power flows, and power system parameters.
Thus, there is a tradeoff between the accuracy of the
condition and quantification for dispatch, connection,
and planning purposes. Furthermore, even if modeling
is possible, there is a social question of how much to
procure, e.g., from minimum procurement to a full pro-
curement that supports 100% renewables.
✔ Price formation: Pricing and cost allocation need to be
granular enough to reflect the technical scarcity but also
understandable enough for contracts to be created from
it. As outlined previously, certain system services, such
as inertia and system strength, are dependent upon unit
commitment rather than generation dispatch. Because
of its binary nature, unit commitment creates noncon-
vexities in a centralized scheduling optimization. This
makes shadow prices and, therefore, locational margin-
al pricing less meaningful. Some designs are looking
into adopting linear relaxations of the integer products
to obtain meaningful pricing of inertia. In other markets,
the challenges of revenue inadequacy flowing from cen-
tralized unit commitment have relied upon “uplift and
clawback” side-payment frameworks and upon alterna-
tive price formations for energy, such as convex-hull
approximations, to minimize those uplift payments. It
also means that for thermal generator factors such as
the minimum stable generation level, start-up and shut-
down costs and no-load cost become increasingly im-
portant. This is an important open research topic in the
context of new security services.
✔ Market depth and liquidity: In the consideration of
market frameworks, a key question emerges as to
whether sufficient depth and liquidity can be created
in system service markets to allow for vibrant com-
petition. Especially as it relates to local system pa-
rameters, such as voltage, reactive power, and system
strength, the potential for the exercise of market power
in relation to pricing needs to be carefully considered.
This is even more important in the context of services
that can affect other markets, too, as discussed previ-
ously. The perception of market power may also dis-
suade competition and result in a barrier to new entry.
✔ DERs: DERs have an important role to play in the
provision of system services through the provision of
bottom-up grid-flexibility services. While wholesale
frameworks would ideally be technology and provider

table 2. Examples of different possible mechanisms to procure security services

(adapted from Billimoria et al., 2020).
Mechanism Description Examples
Mandatory Mandatory conditions imposed on • Virtual inertia provision (Quebec, Ontario)
licenses connection licenses to provide or • Primary frequency control (NEM, National Grid United Kingdom)
contract for system services • Mandatory system reserves (Spain)
• “Do no harm” generator technical requirements (NEM)
• Obligatory reactive power service (National Grid United Kingdom)
Regulated Responsibility delegated to regulated • Minimum system strength and inertia levels (NEM)
procurement network monopoly to provide system • “Delivering a Secure, Sustainable Electricity System” (DS3) System
services, procured either through self- Services Regulated (Eirgrid, Ireland)
build or contracting
Central agency Responsibility delegated to a central • System integrity protection schemes (NEM)
delegation agency (such as an independent • Network support and control ancillary services (NEM)
system operator) to procure system • System stability, voltage, and network pathfinders (National Grid
services via bilateral contracts or United Kingdom)
organized auctions • Enhanced frequency response (National Grid United Kingdom)
• DS3 Service Tender (Eirgrid, Ireland)
• Enhanced Reactive Power Service (National Grid United Kingdom)
• Megavolt amp of reactive power services tender (Belgium)
Spot markets Short-term operational spot • Fast regulation markets (PJM, Midcontinent ISO)
markets for system services, often • Ramping products (California ISO, Midcontinent ISO)
co-optimized with system risk limits • Primary frequency reserve (Wholesale Electricity Market—Western
Austral, proposed)
Market Direct operator intervention or • Residual unit commitments (United States)
constraints and imposition of constraints on system • Market intervention/directions (NEM)
interventions resources to maintain a secure
operating envelope
ISO: independent system operator.

march/april 2021 ieee power & energy magazine 87

 agnostic, the provision of services by DERs necessi-
tates a consideration of the impact upon distribution
network operations. Coordination with the relevant
distribution network operator or distribution system
operator and, as it is being discussed in Australia, dis-
tribution market operator, is therefore critical to al-
lowing local competition and mitigating local adverse
effects from the provision into wholesale markets.
✔ Co-optimization of multiple services: Any market or
regulatory design will ultimately have to relate the provi-
sion of system security services to each other and with
energy dispatch. System services are intended to pro-
vide a secure operating envelope within which energy
dispatch can occur. As such, there are varying degrees
and types of interlinkages between energy provision and
system service provision. Three forms of co-optimiza-
tion underpin constraint formulations. The first is “no
co-optimization with other system conditions,” which
is usually specified under minimum service level pro-
vision (for example, in the NEM, the minimum system
strength unit configurations in South Australia). The sec-
ond is “co-optimization with other ancillary services and
energy limits,” for instance, the development of security
envelopes through the co-optimization of inertia against
PFR and the maximum contingency sizing previously
discussed. Finally, there is the “comprehensive co-opti-
mization with other ancillary services, energy limits, en-
ergy flows, and individual or groups of resources.” These
constraint formulations are often highly locational and
resource specific, for example, in the NEM, the system
strength constraints relating to renewable generation in
Victoria and Queensland. The degree of co-optimization
can also reflect of the level of system security risk. This
may include minimum provision across localized regions
such as the co-optimization and scheduling of energy, in-
ertia, PFR, and contingency levels on regional levels, giv-
en the risks of cascading scenarios or system separation.
Concluding Remarks
The traditional physical features that have characterized elec-
tricity grids over the past 100 years and more are changing
in different ways in the transition toward low-carbon grids. In
this context, power systems around the world, and particularly
in Australia, are witnessing increasing challenges in providing
security services associated with large shares of renewables
and DERs, most of them connected via power electronic inter-
faces. However, these challenges and the resulting increasing
system fragility can be and are already being very effectively
dealt with from a technical perspective, utilizing innovative
operational solutions and new technologies with an outlook to
integrate 75% renewables in Australia by 2025.
However, fundamental challenges remain as to how new
security services should be procured and provided in a mar-
ket context as well as the role of regulation and energy policy.
While several options exist, the integrated techno-economic
88 ieee power & energy magazine
challenges that market designers and regulators face are sig-
nificant and are, as of June 2020, the object of ongoing dis-
cussions in Australia.
Given the deep and complex interactions of technical and
economic aspects, as outlined in this article, we believe that
a from-physics-to-economics approach to the low-carbon
energy market and regulatory design may be more effective
than in the past, when economic designs paid less attention to
physical considerations. This will require a greater presence
and engagement of power system engineers and researchers
at the relevant energy policy and regulation discussion tables.
For Further Reading
P. Mancarella et al., “Power system security assessment of the
future national electricity market,” Melbourne Energy Inst.,
Australia, Tech. Appendix, June 2017. [Online]. Available:
https://www.energy.gov.au/sites/default/files/independent
-review-future-nem-power-system-security-assessment.pdf
“Renewables integration study,” Australian Energy Mar-
ket Operator, Melbourne, Australia, Mar. 2020. [Online].
Available: https://aemo.com.au/en/energy-systems/major
-publications/renewable-integration-study-ris
“Queensland and South Australia system separation
on 25 August 2018,” Australian Energy Market Operator,
Melbourne, Australia, Final Rep., Jan. 2019. [Online]. Avail-
able: https://www.aemo.com.au/-/media/Files/Electricity/
NEM/Market_Notices_and_Events/Power_System_Incident
_Reports/2018/Qld—SA-Separation-25-August-2018-Incident
-Report.pdf
S. Puschel, M. G. Dozein, S. Low, and P. Mancarella,
“Separation event-constrained optimal power flow to en-
hance resilience in low-inertia power systems,” Electric
Power Syst. Res., vol. 189, Art. no. 106,678, Dec. 2020. doi:
10.1016/j.epsr.2020.106678.
F. Fahiman, S. Disano, S.M. Erfani, P. Mancarella, and
C. Leckie, “Data-driven dynamic probabilistic reserve siz-
ing based on dynamic Bayesian belief networks,” IEEE
Trans. Power Syst., vol. 34, no. 3, pp. 2281–2291, May 2019.
doi: 10.1109/TPWRS.2018.2884711.
F. Billimoria, P. Mancarella, and R. Poudineh, “Market
design for system security in low-carbon electricity grids:
From the physics to the economics,” Oxford Inst. for Energy
Studies, U.K., EL41, 2020. [Online]. Available: https://www
.oxfordenergy.org/wpcms/wp-content/uploads/2020/06/
Market-design-for-system-security-in-low-carbon-electricity
-grids.pdf
Biographies
Pierluigi Mancarella is with the University of Melbourne,
Melbourne, 3010, Australia, and the University of Manches-
ter, Manchester, M13 9PL, U.K.
Farhad Billimoria is with the Australian Energy Market
Operator, Melbourne, 3010, Australia, and the University
of Oxford, Oxford, OX2, U.K.
p&e
march/april 2021
awards
T
THE IEEE FELLOW AWARD IS A
special recognition for IEEE Members
with extraordinary accomplishments
in IEEE technical fields. The total
number of recipients each year cannot
exceed 0.1% of the total higher-grade
membership, which ensures the excep-
tional status of becoming a Fellow.
The Class of 2021 Fellows
Congratulations to the following outstand-
ing IEEE Power & Energy Society (PES)
members and to those who were evaluated
by the PES for their achievements.
Enrique Acha
For contributions to power electronics
modeling and applications in electrical
power systems.
Alireza Bakhshai
For contributions to the development of
synchronization techniques for power
electronics converters.
Yaow-Ming Chen
For contributions to grid-connected
power converters for renewable energy
applications.
Vasilis Fthenakis
For contributions to photovoltaics tech-
nology.
Digital Object Identifier 10.1109/MPE.2020.3043651
Date of current version: 19 February 2021
march/april 2021
IEEE Fellows
congratulations to the class of 2021
Wenzhong Gao
For contributions to grid integration
of wind power and electric vehicle
technology.
Hua Geng
For contributions to control of renew-
able energy power converters.
Oriol Gomis Bellmunt
For contributions to grid integration of
renewable energy sources.
Yongpei Guan
For contributions to robust and sto-
chastic optimization for power system
operations.
Dragan Jovcic
For contributions to improvements in
multiterminal high-voltage dc (HVdc)
transmission grids and development of
HVdc transformers.
Ning Lu
For contributions to load modeling and
control methods for providing demand-
side grid services.
Giovanni Mazzanti
For contributions to HVdc cable sys-
tems.
Sanjib Panda
For contributions to iterative learning
control of motor drives.
Leila Parsa
For contributions to control of mul-
tiphase permanent-magnet electrical
drives.
Jacquelien Scherpen
For contributions to nonlinear model
reduction and passivity-based control.
Kalyan Sen
For the development and application of
power flow control technology.
Ramteen Sioshansi
For contributions to energy storage in
electric power systems.
Poul Sorensen
For contributions to wind power con-
verter control and grid integration.
Suresh Srivastava
For contributions to power system se-
curity and stability.
Shinzo Tamai
For contributions to control for motor
drives and three-level converters.
Mi-Ching Tsai
For leadership in magnetic materials
for the electric motor industry.
Mohammad Uddin
For contributions to control techniques
for ac motor drives.
ieee power & energy magazine 89
 We want
Ganesh Venayagamoorthy Wenchuan Wu to hear
from you!
For contributions to the application of For contributions to energy manage-
artificial intelligence to power systems. ment, operations, and control.

Jianhui Wang Zheng Xu

For contributions to unit commitment For contributions to control and mod-
and economic dispatch with renewable eling of modular multilevel converter-
generation. based HVDC transmission systems.

Jin Wang Hamidreza Zareipour

For development of high-density power For contributions to the modeling of
converters and their use in electric cars. energy storage for system operation IMAGE LICENSED BY GRAPHIC STOCK

and planning.
Zhongdong Wang
For contributions to insulating liquids Jizhong Zhu Do you like what you’re reading?
and frequency response analysis meth- For application of optimization meth- Your feedback is important.
ods for power transformers. ods for real-time economic power
system operation. Let us know—
Fushuan Wen send the editor-in-chief an e-mail!
For contributions to fault diagnosis in
power grids. p&e

EVERYTHING YOU NEED IN ONE PLACE

Visit the IEEE PES Resource Center: resourcecenter.ieee-pes.org

The IEEE PES Resource Center is the most extensive library in the world devoted exclusively to the power and
energy industry. You don’t have to be a PES member to access the Resource Center’s vast array of content available
to assist in your research, presentations, or academic programs, and benefit your professional development!

PES Member Access: Most materials in the PES Resource

Center are free to society members - an important member
benefit.
Non-PES Member Access: Discounted prices, for IEEE
members, are listed for each product. Non-members also have
access to all the products on the IEEE PES Resource Center.

• Download PES Magazines

• Watch Video Tutorials & Archived Webinars
• Access Conference Videos & Presentations
• Update Your License with CEU/PDH credits
• Download Complete Technical Reports

Digital Object Identifier 10.1109/MPE.2021.3055439

90 ieee power & energy magazine march/april 2021

society news
Peter Palensky

ISGT Europe 2020

conference focuses on smart grid technology

D
DELFT UNIVERSITY OF TECH- discussed how innovative smart grid Ad van Wijk (adviser to Hydrogen
nology, The Netherlands, played host technologies can contribute to clean- Europe and the European Commission)
to the 10th edition of the IEEE Power energy objectives. Keynote presenta- opened the conference with a proposal
& Energy Society (PES) Innovative tions addressed different views of these that electricity and hydrogen will be car-
Smart Grid Technologies (ISGT) Eu- futuristic energy systems. bon-free symbiotic energy carriers with a
rope Conference, a flagship event
cosponsored by the PES. This
year, the symposium was orga-
nized virtually due to COVID-19.

Policy initiatives
set by the
European
Commission
aim to make
Europe climate
neutral in 2050.
One of the virtual sessions at ISGT-Europe 2020.
Participants logged in from 47
countries on five-plus continents.
More than 30% of the attendees
had an affiliation outside Europe.
2020 wa s t he yea r of t he
“European Green Deal.” Policy
initiatives set by the European
Commission aim to make Europe
climate neutral in 2050. There-
fore, the conference’s theme was
“Smart Grids: Key Enablers of a
Green Power System.” Attendees

Digital Object Identifier 10.1109/MPE.2020.3043652

Date of current version: 19 February 2021

march/april 2021 ieee power & energy magazine 91

 Leo Dijkstra (IBM) gave the presentation “The Cognitive Grid Enerprise,” one of the three keynotes.

need for smart grids to integrate them. Leo Dijkstra

(IBM) led off the second day with a keynote about
the cognitive grid enterprise, which makes smarter
grid operation possible by augmenting humans with
artificial intelligence and machine learning algo-
rithms. On the last day, Alexander Apostolov (Omi-
cron) spoke about how the smart grid is a digital net-
work in which the role of standards is crucial.
Apart from the keynotes, there were eight
s Meets FAA Specifications!
Color – Size – Shape! panel sessions, including two industry-led discus-
– International Orange sions. Topics varied from energy prosumers and de-
s Tested and approved by mand response to real-time simulation and digital
major power companies!
s Thousands still in service twins and from the flexibility of the transmission grid
after 40 years to energy markets in distribution grids. “I congratu-
s Universal attaching! Fits any
wire .1" to 1"! late the committee on a superbly well-organized
s Installs in 5 minutes! virtual conference, especially during these tough
s Withstands hail! times, when so much flexibility is required when plan-
s No maintenance! Does not
slip, oscillate, chafe, cause ning events,” said Kati Sidwall of RTDS Technolo-
electrolysis or harmonic gies. “The technical program was very strong, and
vibration.
s Ships in halves nested. the logistics were done very well.” All 2020 conference
9, 12, 20, 24, 30, and 36" balls keynote addresses and panel discussions are available
& special sizes available through the IEEE Resource Center.
Call now 573-796-3812
Save the Date
ext. 2001 The next PES ISGT Europe Conference, tentative-
Fax 573-796-3770 ly set for 18–21 October 2021, will be organized by
www.tanawiremarker.com
TANA WIRE MARKERS Aalto University, Finland.
p&e
P.O. Box 370, California, MO 65018

92 ieee power & energy magazine march/april 2021

634276_Tana.indd 1 23/03/13 10:38 AM
book reviews
T
THIS ISSUE’S “BOOK REVIEWS”
column offers two reviews. The first
discusses Infrastructure Asset Man-
agement With Power System Applica-
tions, written by Lina Bertling Tjern-
berg. The reviewer writes, “This book
is a solid contribution to the state-of-
the-art asset management application
to electric power systems.” The second
reviews Renewable Energy Technolo-
gies and Resources, written by Nader
Anani. The reviewer writes, “This
book is a welcome addition to the lit-
erature on renewable energy modeling,
computation, and integration.”
Infrastructure Asset
Management With Power
System Applications
By Lina Bertling Tjernberg
This book provides detailed highlights of
infrastructure asset management issues.
It is written by Dr. Lina Bertling Tjern-
berg, a professor in power grid technol-
ogy at Kungliga Tekniska Högskolan, the
Royal Institute of Technology in Swe-
den. Bertling Tjernberg points out
that infrastructure asset management
can be expressed as a combination of
managerial, financial, economic, and
engineering concepts applied to physi-
cal assets. Infrastructure asset manage-
ment is a wide field that covers different
expertise areas. It includes the whole
lifecycle of an asset: design, construc-
tion, commission, operation, mainte-
Digital Object Identifier 10.1109/MPE.2020.3043664
Date of current version: 19 February 2021
march/april 2021
fundamental references
infrastructure asset and renewable energy
nance, modification, decommission-
ing, and disposal. There are budget
constraints, which result in the need
for a strategy for asset management.
The electric power system is a mas-
sive infrastructure to manage the supply
of electricity from generation to con-
sumption. The electric power system
is undergoing a major reconstruction,
which is captured in the concept of a
smart grid. The main drivers are higher
reliability, climate change, and customer
participation in sustainable energy. This
book provides readers with knowledge
of the terminology and theory used. It
provides different analysis tools and ex-
amples to use for electric power system
applications.
The book focuses on energy infra-
structure asset management. It shows
systematic techniques, such as reliabil-
ity-centered maintenance (RCM) and
reliability-centered asset maintenance
(RCAM). These are used to achieve
infrastr ucture asset management
strategies. The RCM method focuses
on maintenance plans’ reliability as-
pects. Maintenance and reliability
are large costs not only as mainte-
nance tasks but also as production and
breakdowns costs. Breakdown conse-
quences affect the environment and/or
personal safety. The RCAM method
combines RCM with maintenance-op-
timization techniques. Merging these
approaches provides an instrument
for maintenance strategy assessment
and comparisons.
The author eloquently presents that
new technology developments provide
enhanced solutions for asset manage-
ment. One new general trend is control
and operation with phasor measure-
ment units located in transmission
grids or with smart meters placed at
consumer sites to enable the integra-
tion of distributed resources with dis-
tributed energy-use control. Another
trend is the development of diagnostic
measurement techniques for assessing
insulation conditions, lifetime predic-
tions of assets, and sensor network con-
dition monitoring. These trends have
the common goal to access and handle
large volumes of data (i.e., big data).
The book introduces reliability and
maintenance, including equipment
ieee power & energy magazine 93
 failure, maintenance planning, and
prevention. Chapter 3 presents is-
sues of risk assessment, lifecycle cost,
and the reliability of computer tools.
Chapter 4 elaborates upon the RCAM
method, implementation, systematic
asset management, and data analysis
requirements. Chapter 5 introduces
The author
eloquently
presents that
new technology
developments
provide
enhanced
solutions
for asset
management.
applications to electric power system
reliability assessment, load point in-
dices, deterministic versus probabi-
listic criteria, operational risk assess-
ment, and maintenance specifications
and performance. Chapter 6 presents
RCAM applications to power distribu-
tion systems, including underground
cables, cable insulations, the rehabili-
tation of water-treed cables, and rural
overhead line system aging factors, and
their effects.
Chapter 7 elaborates upon wind
power system survey failures, tur-
bine and gearbox reliability, condi-
tion monitoring for preventive main-
tenance, and RCAM implementation.
Chapter 8 explains the reliability and
maintenance of components including
hazard-rate estimation using infrared
thermography, condition-based main-
tenance management, artificial neural
network condition monitoring, and su-
pervisory control and data acquisition
alarms and warnings analysis. Chapter
9 discusses RCAM case maintenance
94 ieee power & energy magazine
optimization, including multiobjective
optimization for distribution systems,
event-driven Monte Carlo simulations,
heuristic optimization approaches, and
opportunistic optimization methods.
Chapter 10 explains the RCAM case for
hydropower systems with calculation
details at Vattenfall, risk management
and data analyses, and a comparison of
three RCM methods for case study.
This book’s vision is to encourage
collaboration and knowledge exchang-
es among participants from industry
and academia, regulators, and deci-
sion makers. The author states that ex-
changes would lead to better decisions
and higher quality asset management
results. The target audience includes
readers with varying expertise in as-
set management. The book is meant for
practitioners and academics within dif-
ferent infrastructure asset management
systems with applications to electric
power systems. The background knowl-
edge required for the reader is math-
ematical analysis and fundamentals in
electrical engineering. The author has
provided additional explanations in
the appendix and listed more refer-
ences to a large literature review for
novice readers.
This book is a solid contribution to
the state-of-the-art asset management
application to electric power systems.
I congratulate Prof. Lina Bertling
Tjernberg for amassing a much-need-
ed set of topics to educate the next
generation in this era of uncertainties
a nd evolvi ng elect r ic power sys-
tem innovations.
—Mohammad Shahidehpour
Renewable Energy
Technologies and
Resources
By Nader Anani
This exciting new resource by Nader
Anani presents comprehensive cover-
age of renewable energy technologies
and resources. Anani is an associate
professor in electrical engineering and
Digital Object Identifier 10.1109/MPE.2020.3043676
Date of current version: 19 February 2021
the head of the Electronics and Electri-
cal Engineering Division at the Univer-
sity of Chichester, United Kingdom. He
has authored dozens of journal articles
and conference papers and is a Senior
Member of IEEE. He received his
Ph.D. degree from Manchester Metro-
politan University, United Kingdom.
Renewable Energy Technologies
and Resources focuses on renewable
energy generation and describes the sci-
entific principles and physical systems
used to harness these resources. The en-
vironmental and economic impacts of
using these methods are also introduced
with worked examples and exercises.
Photovoltaics and the modeling of
these systems are presented along with
the environmental and social issues of
utilizing specified biomass as an en-
ergy source. Readers will also learn
how to effectively calculate the cost
and payback time for a given renew-
able energy plant by understanding the
factors affecting the cost of generating
electricity from a renewable energy
system. A compa r ison with tradi-
tional energy generation is included.
Based on the author’s experience in the
development and delivery of renewable
energy models, this book provides
concise and practical solutions that
will appeal to both student and pro-
fessional practitioners.
march/april 2021
 The organization and material in this
text reflect a thoroughly planned ap-
proach as demonstrated by 1) the depth
of covered topics, 2) the introductory
material included, and 3) the learning
outcomes of each chapter. As with most
textbooks, a distinctive feature is that
most of the topics are illustrated with
detailed numerical examples.
The text is organized with introduc-
1) to illustrate analytical expressions with This book is suitable for an advanced
numerical values through ranges within undergraduate course. It would be best
plots and 2) to provide examples of nu- for students who have completed the
merical values for design over a range of first course on power system analysis.
values. The interconnection with practi-
—Gerry Sheblé
cal power systems is shown with the
p&e
power electronics to connect each device.

EasyPower
tory background material first and then
with equipment explanations and mod-

Power System Analysis

els. The chapters include

This book
provides concise
and practical
solutions that
will appeal to
both student
CURRENT IN AMPERES X 100 AT 480 VOLTS
.5 .6 .8 1 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 100 2 3 4 5 6 7 8 9 1000 2 3 4 5 6 7 8 9 10000
1000 1000
800 TX-2 800
FLA
600 600
500 500

0.15
400 400

300 TX-2 300

1 / 1.288 MVA
200 13.8 - 0.48 kV 200

and professional
6%

18.71
100 BL-1 100
80
TX-2
GE AKR-50 80

60 1000 / 1288 kVA 1600/1760 60

50 6% 50

3
.1
40 SWG-4

24
40

30 BL-3 30
GE AKR-30H
20 800/400 20

1.90
practitioners.
BL-1
10 GE MVT
MV T-Plus
MVT-Plus 10

17.58
8 Sensor = 1600 8
TIME IN SECONDS

TIME IN SECONDS
Plug = 1600
6 6
C ur Set = 1.1 (1760A)
Cur

8
5 5

.4
LT
T Band = 1 M-1

19
4 4
STPU = 2
STPU .5 (4400A)
2.5 BL-5
3 STD
ST elay
ela
Delayy = IInt
nt 3
C-H HFD
ST Delay
ST y I²t
I²t = Out
2 O verride = 50000A
Override 225/150 2

1 1
.8 .8

.6 BL-3 .6
.5 GE MVTMVT-9
MV T-9 .5
.4 Sensor = 800 .4
Plug = 800 BL-1
.3
3 Cur
C ur Set = 0.5
0.5
5 (400A) .3
18627A
LT
L T Band = 1 C-6
.2 IInst
nst = 4 (3200A) .2

✔ Photovoltaics: Background Ma-

1 - 400 kcmilil CU
CU

TX-2
.1 1000 / 1288 kVA .1
.08 INRUSH .08
BL-5
.06 .06
Cutler Hammer
er Series
mer Serie C
.05 .05
HFD

terial
.04 BL-3 .04
Frame = 225A (150AT
(150AT)
(150A T
Trip = 150 29249A
.03 .03

.02 .02
BL-5
20574A
Arc Fault Arc Fault Upstream Upstream Electrode Electrode Est Arc Flash Working Incident
.01 .01 Trip Device Trip Device Equip Type Gap Boundary Distance Energy

✔ Photovoltaics: Characteristics
.5 .6 .8 1 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 100 2 3 4 5 6 7 8 9 1000 2 3 4 Bus
5 6 7 8 9 10000 Name Bus kV Configuration
Name Function (mm) (inches) (inches) (cal/cm2)
R R CURRENT IN AMPERES X 100 AT 480 VOLTS
BUS-3 13.8 R-6 51/50 Open Air VOA 152 32.1 + 26 1.7

BUS-7 13.8 R-7 51/50 Open Air VOA 152 30.5 + 26 1.5
M-1 0.48 BL-3 Other HCB 32 31.6 + 18 3.8

and Circuit Modeling of the

MAIN SWG 3.8 R-18 51/50 Switchgear VCB + HCB 152 189.3 + 18 59.3

REFINER 2.4 R-7 51/50 Switchgear VCB + HCB 104 261.9 + 18 122.8
SWG-4 0.48 R-6 51/50 Switchgear VCB + HCB 32 213.2 + 18 135.1

MCC-1 0.48 BL-2 MCC VCB 25 53.6 + 18 6.9

PV Cell
MCC-2 0.48 BL-3 MCC VCB 25 53.1 + 18 6.8

PNL-1 0.208 FS-2 Panel VCB + HBB 25 48.2 + 18 7.1

PNL-2 0.208 PNL-1 Panel VCB 25 57.8 + 18 7.7

✔ Photovoltaics: PV Arrays Opera- and Shock

Arc Flash priate PPE Requir
Risk Hazar
ed
d

PANEL PANEL PANEL Appro Arc Flash and Shoc

tion and Characteristics rd Boundary Incident Ener
gy k Risk Hazard
Flash Haza inches - Arc Flash s or arc rated cove Appropriate PPE
rall
4’ - 0” cal/cm at
18
arc rated
and
2pant Required
6.0 shirt 4’ - 0” Flash Hazard Bounda
Arc-rated ved
cover is remo 6.0 cal/cm2 at 18 inches ry
Hazard when - Arc Flash Inciden

✔ Photovoltaics: PV Generating
kV Shock oach
t Energy
0.48 Limited Appr oach
3’ - 6” Appr 0.48 kV Shock Hazard
Restricted when cover is remove
1’ - 0” (Fed By: 27B) ONLY3’ - 6” Limited Approa d
MCC-23A FIGURATI
ON 1’ - 0” ch
pme nt Name: EM CON Restricted Approa
ch
Equi SYST
NORMAL

Systems
VALID FOR Equipment Name:
SWG-4 (Fed By:
VALID FOR NORMA 18A)
L SYSTEM CONFIG
URATION ONLY

✔ Hydroelectricity

· Easy to Learn and Use

✔ Tidal Energy
✔ Wind Energy
✔ Bioenergy
✔ Costing a Renewable Energy
Project.
The book offers a list of references
for each chapter of the key literature. A
· Fast Results
thorough explanation of the operation
Try instantly online or download a free demo copy at:
for each device, with detailed versions of www.EasyPower.com/demo
key plots and figures, is included. Each
chapter ends with a set of problems to re-
view the depth of learning achieved. ®
The author provides examples and
Power made easy.
problems as a key part of the material
treatment. The data serves two purposes:

march/april 2021 ieee power & energy magazine 95

 calendar
T
T H E I EEE POW ER & EN ERGY
Society’s (PES’s) website (http://
www.ieee-pes.org) features a meet-
ings section, which includes calls
for papers and additional informa-
tion about each of the PES-sponsored
meetings. Please check the conference
website for the most current informa-
tion and any venue changes due to the
pandemic.
May 2021
IEEE International Conference
on Electrical Machines and Drives
(IEMDC 2021), 16–19 May, Hartford,
Connecticut, United States, contact Ali
Bazzi, alibassi@ieee.org, http://iemdc
-conference.org/
June 2021
IEEE Transportation Electrification
Conference and Expo (ITEC 2021),
23–25 June, Chicago, Illinois, United
States, contact Rebecca Krishnamurthy,
rebecca.k@rna-associates.com, https://
itec-conf.com
Digital Object Identifier 10.1109/MPE.2020.3043678
Date of current version: 19 February 2021
96 ieee power & energy magazine
PES meetings
for more information, www.ieee-pes.org
IEEE PowerTech Madrid (PowerTech
2021), 28 June–2 July, Madrid, Spain,
contact Tomas Gomez San-Roman,
tomas.gomez@comillas.edu, https://www
.powertech2021.com/
July 2021
IEEE PES General Meeting (GM
2021), 25–29 July, Washington, D.C.,
United States, contact Roseanne Jones,
roseanne.jones@ieee.org
August 2021
IEEE Electric Ship Technologies
Symposium (ESTS 2021), 4–6 August,
Arlington, Virginia, United States,
contact Julie Chalfant, chalfant@mit
.edu, https://ests21.mit.edu
September 2021
IEEE PES GT&D International
Conference and Exposition, Istan-
bul (GTD), 14–17 September, Istanbul,
Turkey, contact Omer Usta, usta@ieee
.org, https://ieee-gtd.org/
IEEE PES Innovative Smart Grid
Technologies Latin America (ISGT
LA 2021), 15–17 September, Lima,
Peru, contact Jorge Lafitte, dr.jorge
.lafitte@ieee.org
October 2021
IEEE PES Innovative Smart Grid
Technologies Europe (ISGT Europe
2021), 18–21 October, Espoo, Finland,
contact Pourakbari Kasmaei Mahdi,
mahdi.pourakbari@aalto.fi
November 2021
IEEE PES Asia-Pacific Power and
Energy Engineering Conference
(APPEEC 2021), 21–23 November,
Thiruvananthapuram, India, contact
Boby Philip, boby.philip@ieee.org
December 2021
IEEE PES Innovative Smart Grid
Technologies Asia (ISGT Asia 2021),
5–8 December, Brisbane, Australia,
contact Tapan Saha, saha@itee.uq.edu.au
For more information on additional tech-
nical committee meetings, webinars, and
events, please visit our IEEE PES calendar:
https://www.ieee-pes.org/meetings
-and-conferences/conference-calendar.
p&e
march/april 2021
in my view (continued from p. 100)

must consider both the consequences of
a disturbance as well as its probability.
Note that system resiliency was not
mentioned in the CIGRE paper and
consequently not defined. Today, sys-
tem resiliency is important when dis-
cussing system security. I will come
back to this.
different countries do not always have
the same legislative policies and regula-
tions. Third, there are different players
active in the system, i.e., power produc-
ers, system operators, regulators, con-
sumers, and so on, making the system
complex from an organizational point of
view. I will not elaborate more on these
challenges, but they are essential to keep
in mind in the future, too.
In this issue of IEEE Power & En-
ergy Magazine, several new constraints

Challenges—Old and New

The electric power system is continu-
ally faced with new challenges in addi-
They know exactly where they’re going.
tion to old ones, and so are all other in-
frastructure systems. These challenges Your crews can too!
are due to evolving requirements and
boundary conditions coming from so-
Route work orders and control processes with
ciety as well as the introduction and
integration of new technologies. New configurable workflows and notifications
technologies offer new and more ef-
fective solutions, but their introduction
often requires the redesign of prevail-
ing systems and procedures, which
could be a demanding task introducing
new failure modes in the system. Many
of the latest requirements from soci-
ety, imposed in the form of new laws
and regulations, are due to increasing
environmental concerns and aware-
ness. Furthermore, the requirements
should be taken into account without
any substantial increase in cost for the
end consumer. This is summarized in
the so-called “energy trilemma” opti-
mization scheme: energy security, sus-
tainability, and affordability, which all
need to be considered when designing
the system.
Before discussing the new chal-
lenges, let us quickly review some of the
salient characteristics of power sys-
tems that will remain in the future. First,
time constants from subseconds (tran-
sients of different kinds) to several years
(the expansion planning and manage-
ment of large hydro reservoirs) must be
addressed. Luckily, many processes can
be decoupled due to different timescales
so that the complexity can be managed.
Second, the physical distances in inter-
connected systems, like those in Eu-
rope, North America, and other places
in the world, create various problems.
Software that works varasset.com
Interconnected systems that comprise

march/april 2021 ieee power & energy magazine 97

 and challenges with relevance for sys-
tem security in different systems have
been presented together with solutions
and remedial actions. Many innova-
tive designs have been suggested and
implemented as a response to new dif-
ficulties, and they are good examples
of how power engineers perpetually
work to guarantee the security of sup-
ply. In my view, there are a couple of
main causes behind the development
described here, as I will explain.
First, with the large-scale intro-
duction of new weather-dependent
renewables [mainly wind power and
photovoltaics (PVs)], power system
planning, operation, and control face
new problems, and the ongoing stud-
ies on the topic are plentiful. With the
ever-increasing penetration of renew-
ables, new solutions are proposed, and
I do not think we have seen the com-
plete answer yet.
Considering the total electric ener-
gy production by wind power and PVs
from a larger geographical area and
over a longer period, e.g., a year, this
production will not show any larger
variations from year to year. The yearly
energy production of the large area can
be assumed to be fairly constant (al-
though a changing climate might imply
that this will not be true in the future).
However, the local variations from year
to year can be substantial, and so can
the temporal variations over the year.
There are, in principle, three ways to
cope with these problems: transmis-
sion lines, storage, and demand-side re-
sponse. Depending on the nature of the
variations (geographical or temporal)
and other system characteristics, one
or the other (and in many cases a com-
bination of these means) must be em-
ployed to guarantee the desired system
operation requirements and security.
Several articles in this issue de-
scribe new ways to master this problem
in different systems. Note that in some
systems, e.g., the continental European
system, interconnections were origi-
nally built and designed primarily to
enable the sharing of reserves and pro-
viding redundancy. Therefore, even if
interconnecting transmission lines ex-
98 ieee power & energy magazine
ist, these might not always be able to
fulfill the tasks required to integrate
renewables securely. A fourth possi-
bility would be to utilize the flexibil-
ity of other power plants, e.g., hydro-
power, if available, or flexible thermal
power plants.
A new security threat that has at-
tracted a lot of attention lately is re-
lated to cybersecurity. We have already
seen cyberattacks in power systems
and other technical systems. With the
increasing penetration of control and
monitoring systems, e.g., based on pha-
sor measurement units using Internet-
based communication channels, the ex-
posure to such attacks will grow. I am
convinced that cybersecurity will be
an integrated part of the overall system
security assessment in the future, and
this must be done in cooperation with
other engineering disciplines.
How to Achieve Security
To achieve the desired security level,
one can, in principle, use two design
approaches: make the system either
robust or resilient. The difference can
be captured in the following some-
what simplified descriptions: a robust
system is designed so that its users
should not be surprised, whereas a
resilient system is designed so that
its users are prepared to be surprised.
The article “Toward a Consensus on
the Definition and Taxonomy of Pow-
er System Resilience” offers the fol-
lowing definition:
The ability to withstand and re-
duce the magnitude and/or dura-
tion of disruptive events, which
includes the capability to an-
ticipate, absorb, adapt to, and/
or rapidly recover from such
an event.
A robust system is designed to cope
with several predetermined events
and disturbances within a specified
space. These events are presumed to be
known a priori, and their magnitudes,
in some metric, are assumed to be
bounded. The events to be considered
in the design and their possible maxi-
mum magnitudes are based on experi-
ence and system knowledge. In a robust
system, we try to contain the impacts
of the “known” unknowns in such a
way that the risk level of the system is
within desired limits, and this can be
said to be the traditional way of design-
ing power systems.
However, due to the increased sys-
tem complexity because of new tech-
nologies, new players, new types of
power plants, and so on, the ability to
foresee all possible threats is increas-
ingly difficult. This is corroborated
by the analyses of many of the major
power system blackouts during the last
decades. A system can be designed
to withstand many individual distur-
bances, but a specific combination and
temporal occurrence of these events
might bring the system to a complete
or a partial blackout. One could say
that an “unknown” unknown, i.e., the
combination of events, was the cause
of the power outage. Furthermore, de-
signing a system to withstand all low-
probability events, even if they can
have a very severe impact on the sys-
tem, is not economical. Examples in-
clude hurricanes or ice storms hitting
a power system with disastrous con-
sequences. History tells us that such
extreme weather events can be very
different, so it is not realistic to aim at
designing and building a system that is
robust against all these events. Thus,
other means to cope with high-impact,
low-probability events are needed,
and here, a resilience-based approach
would be more promising.
I believe that the future solution
would be to combine robustness and
resilience in system design and opera-
tion. Actively introducing resiliency is
fairly new in the design of power sys-
tems, at least when compared with ro-
bustness, and a lot of research is going
on in this area. I am sure that, in the
future, many new and innovative ideas
inspired by resiliency will be imple-
mented in the power system to increase
system security. As indicated in the
previous definition, an effective resto-
ration process will add to the resiliency
of the system. Restoration plans have
always been important to ensure sys-
tem security, and I am convinced that
march/april 2021
they will continue to be important in
the future.
Experience gained from blackouts
and other situations that put the power
system under severe stress should be
shared so that other system operators
can use it in developing and designing
their systems. Knowledge gained about
so-called hidden failures that occur un-
der stressed system conditions or as a
consequence of several unlikely events
should be used to improve power sys-
tem security.
Some Lessons From the
COVID-19 Pandemic
When a disruptive event with far-
reaching consequences for society, like
the COVID-19 pandemic, hits, the ex-
perience and knowledge obtained from
such an event may also apply to power
system operations. Even if a pandemic
caused by an infectious virus and a
power system follow different physical
laws, I think that there are some par-
allels. Both a pandemic and the mal-
function of the power system can have
dramatic impacts on social life, eco-
nomics, and other vital functions of our
society. I am sure that when the pan-
demic has been brought under control,
governments, authorities, companies,
and other organizations will analyze
the course of events, consequences, and
efficacy of different preventive actions
and try to draw conclusions that will be
useful in the containment of future epi-
demics and pandemics. Without claim-
ing to have made a more systematic
study of the current pandemic and its
consequences, I want to give two ex-
amples that I think also have relevance
for the electric power system.
One of the fundamental principles
of the European Union (EU) is a free
and nondiscriminatory market that
is open to actors from all its mem-
ber states. However, when the virus
started to spread, some countries
stopped the export of equipment es-
sential for health care to other mem-
ber states, despite earlier agreements.
march/april 2021
The needs of the nation’s health sys-
tem were deemed more important
than the laws and agreements of the
EU, and the overall coordination and
assessment within the EU were on
hold. This was later rectified, and
when the first phase of the pandemic
was over, free trade seemed to func-
tion more or less as intended.
Many countries in Europe are criti-
cally dependent on the exchange of
electric power. It can be in the form of
“pure” energy imports over a year or
seasonal power exchanges over nation-
al borders. What will happen if should
there be an energy shortage somewhere
in Europe? Will each country primar-
ily look to its own needs, or will the
overall needs of the interconnected
system have priority? If we look at
the experiences from the initial phase
of the COVID-19 crisis in Europe,
the former of these two may likely be
the initial outcome. Furthermore, the
trade of electric power system equip-
ment might also be hindered due to na-
tional concerns, and this will have an
impact but on a longer timescale. The
complex, international supply chains
that are dominating the manufacture of
many high-tech products are also used
by the producers of high-voltage equip-
ment and can affect the maintenance
and replacement of critical components
in the power system.
As far as I know, power systems
worldwide have worked more or less
as intended, and no severe interrup-
tions directly caused by the pandemic
have been reported. I believe this
is due to today’s highly automated
systems, including remote-control
operations. The daily operation of
power systems has been marginally
affected, and the end consumers have
not noticed anything. Of course, the
economic lockdown has reduced the
consumption of electric power, but the
power system has still withstood the
stresses incurred by the crisis. Also,
most of the longer-term work, e.g.,
planning, has continued without major
distraction, mostly because of home-
office work and reliable Internet con-
nections. This is a strength, but as John
A. Casazza, a prominent U.S. power
engineer, once said, “Look at a system’s
or a procedure’s greatest strength to see
the source of its greatest weakness.”
How essential processes and proce-
dures used in the power industry may
be exposed to various disturbances and
what impact they can have on power
system security both deserve a system-
atic and critical review. Also, processes
and procedures that might appear as
peripheral to the power systems opera-
tion can turn out to be critical during
long-lasting emergencies such as the
COVID-19 pandemic.
Concluding Remarks
This issue contains many interest-
ing examples of how the problems
are addressed in different systems in
the world. Each system is unique and
requires specific solutions to meet se-
curity requirements. Due to changes
and developments in technology, regu-
lations, and organizations, new solu-
tions and concepts must continuously
be contrived. The future power system
must be designed to cope with known
and unknown events and disturbances
through a combination of robustness
and resilience to meet the security and
other requirements of society.
For Further Reading
P. Kundur et al., “Definition and classi-
fication of power system stability IEEE/
CIGRE joint task force on stability terms
and definitions,” IEEE Trans. Power Syst.,
vol. 19, no. 3, pp. 1387–1401, Aug. 2004.
doi: 10.1109/TPWRS.2004.825981.
A. Gholami, T. Shekari, M. H.
Amirioun, F. Aminifar, M. H. Amini,
and A. Sargolzaei, “Toward a consen-
sus on the definition and taxonomy of
power system resilience,” IEEE Access,
vol. 6, pp. 32,035–32,053, June 2018.
doi: 10.1109/ACCESS.2018.2845378.
p&e
ieee power & energy magazine 99
 in my view
A
A SECURE, RELIABLE, EFFICIENT,
and sustainable energy system is a pre-
requisite for a well-functioning society
that can satisfy the needs of its citizens
affordably, and this applies to the elec-
tric power system in particular. This is
obvious to most of us, and I will not
discuss and motivate the correctness of
this statement further in this column
since numerous papers and articles dis-
cuss it in detail. Rather, I will take this
premise as a starting point for a dis-
course on some aspects of the security
of energy systems and, more specifi-
cally, the electric power system.
During recent years, many studies
and other initiatives have been directed
to multienergy systems, also referred
to as energy system integration or sec-
tor coupling, aiming at optimizing the
overall energy system. By considering
not only one energy-delivery system,
such as electricity, in the system analy-
sis and optimization but also includ-
ing other interrelated energy-delivery
systems, such as gas and heat-delivery
systems, the features of the different
energy-delivery systems can comple-
ment each other, resulting in a more ef-
ficient and reliable overall system. I am
convinced that future system design-
ers will increasingly consider this op-
tion. However, due to the limited space
available, I will focus on the electric
power system. Another motivation is
that in most studies of future multie-
nergy systems, the scenarios where
Digital Object Identifier 10.1109/MPE.2020.3043679
Date of current version: 19 February 2021
100 ieee power & energy magazine
power system security
why & how
electric power systems constitute the
“backbone” usually come out as the
most advantageous ones, due mainly
to the versatility and attractiveness of
electricity for the end consumer.
What Is System Security?
Formal definitions of system security
and reliability are worthwhile to estab-
lish as different engineering disciplines
use somewhat different definitions. A
joint CIGRE and IEEE task force pub-
lished a paper in 2004 that formulates
a number of definitions used in power
systems dynamics, but I think that they
are also useful in the more general con-
text we are dealing with here. The pa-
per states:
Reliability of a power system
refers to the probability of its
satisfactory operation over the
long run. It denotes the ability to
supply adequate electric service
on a nearly continuous basis,
with few interruptions over an
extended time period.
Security of a power system
refers to the degree of risk in its
ability to survive imminent dis-
turbances (contingencies) with-
out interruption of customer ser-
vice. It relates to robustness of
the system to imminent distur-
bances and, hence, depends on
the system operating condition
as well as the contingent prob-
ability of disturbances.
Stability of a power system
(…) refers to the continuance of
intact operation following a dis-
Göran Andersson
turbance. It depends on the oper-
ating condition and the nature of
the physical disturbance.
As a comment on the definition of
security, it is further postulated that
System security may be further
distinguished from stability in
terms of the resulting conse-
quences. For example, two sys-
tems may both be stable with
equal stability margins, but one
may be relatively more secure
because the consequences of in-
stability are less severe.
Other terms that one sees are “secu-
rity of supply” and “system security,”
where the latter applies specifically to
power system dynamics, as was the fo-
cus of the aforementioned paper. How-
ever, I think that we can use the previous
definition with some minor modifica-
tions. Thus, I propose the following
definition of power system security:
✔ Security of a power system re-
fers to the degree of risk in its
ability to withstand disturbanc-
es, on all connected equipment
and subsystems, without inter-
ruption of customer service. It
relates to the design of the sys-
tem and connected subsystems
as well as the contingent prob-
ability of disturbances.
A few words and phrases have been
modified. These changes will be elabo-
rated in this article. I want to empha-
size that the risk concept is kept, which
means that when assessing security, we
(continued on p. 97)
march/april 2021
advertisers index
The Advertisers Index contained in this issue is compiled as a service to our readers and advertisers: The
publisher is not liable for errors or omissions although every effort is made to ensure its accuracy. Be sure
to let our advertisers know you found them through IEEE Power & Energy Magazine.

Company page# URL Phone

ASPEN, Inc. CVR 4 www.aspeninc.com +1 650 347 3997

Bigwood Systems, Inc. 6 www.bigwood-systems.com +1 607 257 0915

Craig Wire Products 13 www.craigwire.com +1 770 920 2222

CYME CVR 2 www.eaton.com/cyme +1 800 361 3627

DIgSILENT GmbH 3 www.digsilent.de/powerfactory

EasyPower LLC 95 www.EasyPower.com/demo

ETAP 11 etap.com/20.5 +1 949 900 1000

P & R Technologies 17 www.pr-tech.com +1 800 722 8078

Powertech Labs, Inc. 12 www.dsatools.com

PowerWorld Corporation 8 www.powerworld.com +1 217 384 6330

PSCAD a product of MHI 7 pscad.com

RTDS Technologies, Inc. 9 www.rtds.com

SKM Systems Analysis, Inc. 5 www.skm.com +1 800 500 4SKM

Tana Wire Markers 92 www.tanawiremarker.com +1 573 796 3812

Varasset 97 varasset.com

YVI Labs 14 www.photoplusplus.com

445 Hoes Lane, Piscataway, NJ 08854

IEEE power & energy magazine representative

Erik Henson
Naylor Association Solutions
+1 352 333 3443, fax: +1 352 331 3525
ehenson@naylor.com

Digital Object Identifier 10.1109/MPE.2021.3054095