Volume 18 • Number 3 • May/June 2020

magazine

Volume 18 • Number 3 • May/June 2020

www.ieee.org/power

on the 18
cover

©ISTOCKPHOTO.COM/TRONAND
©ISTOCKPHOTO.COM/ERHUI1979

features

contents
18 The Tariffs of Tomorrow 49 China’s Solar Subsidy Policy
By Ahmad Faruqui and Cecile Bourbonnais By Houqi Dong, Bo Zeng,
Yuqing Wang, Yingxin Liu,
26 Flexible Network Pricing Encourages and Ming Zeng
Greater Sharing
By Furong Li, Chenghong Gu, and Ran Li 61 Utility Load Research
By Curt Puckett, Craig Williamson,
33 Distribution Network Rate Making Claude Godin, Will Gifford,
in Latin America Jonathan Farland, Tom Laing,
By Rodrigo Moreno, Bernardo Bezerra, and Tao Hong
Hugh Rudnick, Carlos Suazo-Martinez,
Martha Carvalho, Alejandro Navarro,
Carlos Silva, and Goran Strbac

columns &
departments
4 From the Editor
8 Leader’s Corner
14 Guest Editorial
72 History
80 Calendar
84 In My View

72
Digital Object Identifier 10.1109/MPE.2020.2971821

may/june 2020 ieee power & energy magazine 1


Editor-in-Chief
Michael I. Henderson, mih.psat@gmail.com
Associate Editors
John Paserba, History
Editorial Board
L. Barroso, A. Conejo, J. Feltes, N. Hatziargyriou,
T. Hong, B. Johnson, B. Kroposki, Debra Lew,
N. Lu, B. Mather, M. Miller, D. Novosel,
M. O’Malley, N. Ochoa, J. Paserba, C.E. Root,
magazine
IEEE Periodicals/Magazines Department
445 Hoes Lane, Piscataway, NJ 08854 USA
www.ieee.org/magazines
H. Rudnick, G.B. Sheblé, J.C. Smith, Geraldine Krolin-Taylor, Senior Managing Editor
M. Thomas, E. Uzunovic, S.S. Venkata, Janet Dudar, Senior Art Director
J. Wang, S. Widergren Gail A. Schnitzer, Associate Art Director
Theresa L. Smith, Production Coordinator
Spanish Editorial Board Felicia Spagnoli, Advertising Production Manager
Enrique Tejera, Editor-in-Chief Peter M. Tuohy, Production Director
Editors: M. Baquedano, G. Gonzalez Kevin Lisankie, Editorial Services Director
Dawn M. Melley, Staff Director, IEEE Publishing
Advertising Operations
Erik Henson, Naylor Association Solutions
IEEE prohibits discrimination, harassment, and bullying. For more infor-
+1 352 333 3443, fax: +1 352 331 3525 mation, visit http://www.ieee.org/web/aboutus/whatis/policies/p9-26.html.
ehenson@naylor.com
Promoting Sustainable Forestry

IEEE POWER & ENERGY SOCIETY (PES)

SFI-01681

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.

Governing Board J. Cardoso, Web Site Development Publications

F. Lambert, President R. Li, Women in Power Publications Board Chair, B. Pal
J. Bain, President-Elect J. Hofman, Young Professionals Editors-in-Chief
M. Armstrong, Vice President, Chapters IEEE Electrification Magazine, L. Fan
V. Vittal, Vice President, Technical Activities Technical Council IEEE Power Engineering Letters, M. Fotuhi-
E. Uzunovic, Vice President, Education V. Vittal, Chair, H. Chen, Vice Chair Firuzabad
B. Pal, Vice President, Publications D. Watkins, Secretary, F. Rahmatian, Past-Chair IEEE Trans. on Energy Conversion, A. Tessarolo
W. Bishop, Vice President, Meetings Technical Committee Chairs IEEE Trans. on Power Delivery, F. DeLeon
J.C. Montero Q, Vice President, Membership K. Schneider, Analytical Methods for IEEE Trans. on Power Systems, N. Hatziargyriou
& Image Power Systems IEEE Trans. on Smart Grid, C. Canizares
S. Bahramirad, Vice President, New Initiatives/ J. Yagielski, Electric Machinery IEEE Trans. on Sustainable Energy, B. Chowdhury
Outreach J. Yale, Energy Development & Power IEEE Open Access Journal of Power and Energy,
N. Hadjsaid, Treasurer F. Li
Generation IEEE Power & Energy Magazine, M. Henderson
J. Sykes, Secretary C. Ashton, Energy Storage & Stationary Battery
S. Rahman, Past-President eNewsletter, S. Fattah
H. Geene, Insulated Conductors Editor-in-Chief at Large, W. Xu
D. Diaz, Region Rep., U.S. & Canada D. Harmon, Nuclear Power Engineering
J. Milanovic, Region Rep., Europe, Middle East, Marketing, E. Batzelis
K. Fodero, Power System Communications PES Representative to IEEE Press, Open
& Africa & Cybersecurity
S. Leon, Region Rep., Latin America Website, R. Rana
C. Vournas, Power System Dynamic Performance
D. Sharafi, Region Rep., Asia & Pacific J. McBride, Power System Instrumentation Meetings
B. Enayati, Member-at-Large & Measurements Committee Chairs
R. Kappagantu, Member-at-Large F. Li, Power System Operation Planning General Meeting Steering, W. Cassel
H. Koch, Member-at-Large & Economics Joint Technical Committee Meeting Steering,
J. Yu, Member-at-Large R. Patterson, Power System Relaying & Control S. Ward
R. Melton, Smart Building, Load & Technically Cosponsored Conferences Steering,
IEEE Division VII Director A. Borghetti
M. Sanders Customer Systems
J. Gravelle, Substations Transmission & Distribution Conference &
Exposition Committee for North America
IEEE Division VII Director-Elect S. Hensley, Surge Protective Devices
Steering, C. Segneri
Open K. Flowers, Switchgear
B. Forsyth, Transformers Innovative Smart Grid Technologies North America,
PES Executive Director J. Romero Aguero
G. Chang, Transmission & Distribution Innovative Smart Grid Technology Conference­­–
Patrick Ryan, +1 732 465 6618,
fax +1 732 562 3881, p.ryan@ieee.org Technical Council Coordinating Committees Europe, D. Van Hertem
D. Houseman, Intelligent Grid & Emerging Innovative Smart Grid Technology Conference–
Standing Committee Chairs Technology Asia, D. Sharafi
Open, Constitution & Bylaws D. Alexander, Marine Systems PowerAfrica Steering Committee, H. Louie
M.T. Correia de Barros, Fellows A. Leon, Wind & Solar Power Website, Open
N. Hadjsaid, Finance Education
A. Apostolov, History Technical Council Standing Committees Committee Chairs
J. Bain, Long-Range Planning F. Rahamatian, Awards Power and Energy Education, S. Suryanarayanan
S. Rahman, Nominations & Appointments Open, Industry Education Scholarship Plus, R. Pieper
D. Watkins, Organization & Procedures Website, Open
Region Representatives T. Burse, Standards Coordination
D. Diaz, M. Gosalia, B. Gwyn, J. Khan, H. Chen, Technical Sessions New Initiatives and Outreach
S. Leon, J. Milanovic, M. Papic, Open, Webmaster Committee Chairs
D. Sharafi, C. Wong IEEE Smart Village, R. Larsen, Chair
R. Podmore, Vice-Chair
Chapter Committee Chairs Industry Outreach, J. Giri, Chair
C. Diamond, Awards & Resources Website, K. Anastasopoulos
Z. Wang, Chapters Website
N. Nair, Distinguished Lecturer Program IEEE Power & Energy Magazine
Z. Wang, Electronic Communications IEEE Power & Energy Magazine (ISSN 1540-7977) (IPEMCF) is published bimonthly by the Institute of Electrical and
T. Ribeiro de Alencar, Student Chapters Coordinator ­Electronics E ­ ngineers, 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-
Membership & Image Committee Chairs scriptions, address changes): 445 Hoes Lane, Piscataway, NJ 08854 USA. Telephone: +1 732 981 0060, +1 800 678 4333.
A. Vaccaro, Awards & Recognition Individual copies: IEEE members US$20.00 (first copy only), nonmembers US$87 per copy. Subscription Rates: Society members
S. Fattah, Humanitarian Activities included with membership dues. Subscription rates available upon request. Copyright and reprint permissions: Abstracting is permitted
J. Benedict, Marketing 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
M. Seif, Membership Development the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA; 2) pre-1978 articles without fee. For other copying,
L.F. Gaitán, Social Media reprint, or republication permission, write Copyrights and Permissions Department, IEEE Operations Center, 445 Hoes Lane, Piscat-
away, NJ 08854 USA. Copyright © 2020 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,
Digital Object Identifier 10.1109/MPE.2020.2971822 IEEE Operations Center, 445 Hoes Lane, Piscataway, NJ 08854 USA. Canadian GST #125634188 Printed in U.S.A.

2 ieee power & energy magazine may/june 2020

 OPTIMIZE YOUR GENERATION AND
DISPATCH COSTS WITH POWERFACTORY!
 Unit Commitment, Redispatch Calculation and Market Simulation based on a full network
 Minimisation of operating, emission & startup costs of generating units, as well as costs for
curtailment of renewables, load shedding and transformer switching
 Consideration of network constraints and generator limitations such as ramping, minimum
up/down times
 Support of contingency constraint optimisation and consideration of planned outages
 Time sweep optimisation with powerful algorithm settings for performance improvement
 Storage schedule optimisation
 Flexible results analysis and visualisation with dedicated reports, colouring modes and
diagrams
 Integrated optional interface to external solvers such as CPLEX, GUROBI or data exchange
via AMPL-NL file format

For more information visit POWER SYSTEM SOLUTIONS

www.digsilent.de/unit-commitment
MADE IN GERMANY
Serving more than 150 countries.
 from the editor
H
HOW SHOULD ELECTRIC-
ity rates be designed? Many
factors must be considered
to properly balance reliabil-
ity of service with the cost to
customers while also ensur-
ing the economic viability of
utility providers. Social and
political concerns add com-
plexity to regulatory decisions
that require consideration of
cost causation, public policies,
and grid modernization. Rate
making was difficult enough,
but now the transformational
changes in the electric power
industry further complicate matters, given
the growth of variable energy resources,
demand response, electrification, behind-
the-meter resources, microgrids, and the
introduction of newer technologies. Geo-
graphic diversity and types of demand
served present other issues that affect
electricity rates. Many factors influence
the service provider’s revenue require-
ments, including the required rate of re-
turn; depreciation schedules; operations
and maintenance costs; other accounting
factors; and financing, taxes, and sub-
sidies. The rates provide incentives for
desired behaviors, which influence reli-
ability and economy of service.
How can rate issues be addressed?
Data collection provides key information
for studying customer types and behav-
iors, system performance, and the need
for future capital investment. ­Advanced
Digital Object Identifier 10.1109/MPE.2020.2972664
Date of current version: 17 April 2020
4 ieee power & energy magazine may/june 2020
Michael Henderson
utility rate making
the path between theory & implementation
hicles. The role of markets,
operating and maintenance
costs, and capital costs further
complicates the situation.
In This Issue
This issue of IEEE Power &
Energy Magazine presents a
worldview of utility rate-mak-
ing theory and implementa-
tion. The traditional two-part
rates consist of fixed and vari-
able monthly energy charges
but require major reform to ad-
©ISTOCKPHOTO.COM/SILVIONKA
dress the acceleration of grid
transformation and make use
metering infrastructure (AMI) and of modern data collection and analysis
­other new technologies now provide techniques. The articles in this issue dis-
highly precise data, but new analytic cuss lessons learned from rate-making
techniques are necessary for data inter- experience and provide an outlook for
pretation and analysis. This must be done future rate structures.
in the backdrop of grid transformation, Guest Editors Ning Lu and Tao
the widespread addition of variable en- Hong compiled five articles that exam-
ergy resources, and greater penetrations ine the principles of rate making, the
of new technologies such as electric ve- need for rate reform, proposals for new
IEEE PES Transmission & Distribution Conference and
Exposition Postponed Until October
Following the declaration by the World Health Organization of COVID-19 as
a pandemic, local, regional, national, and international bodies have moved to
restrict travel and mass gatherings. In light of these developments and for the
protection, health, and safety of our transmission and distribution (T&D) com-
munity, the IEEE Power & Energy Society made the decision to change the T&D
Conference and Exposition to 12–15 October 2020 at McCormick Place South.
Please monitor the IEEE PES T&D website (https://www.ieeet-d.org/) for
updates on the reopening of registration and housing, exhibitor details, and
conference highlights.
 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
 rates, and trends of recent applications.
The articles consider a wide diversity
of regions and utilities, both with and
without electricity markets. The ar-
ticles in this issue cover the following:
✔ the reasons and means of replac-
ing two-part tariffs, including sev-
eral case studies in North America
✔ ways of dealing with uncertain-
ties in the future system by reward-
ing customers who can provide
flexibility in the United Kingdom
✔ distribution network rate making
in Latin America and a detailed
discussion of tariff structures in
Chile and Brazil
✔ the effects of rate making and
world markets on the develop-
ment of the photovoltaic industry
in China
✔ the latest developments in load
research, which remain critical
to rate design and provide vital
information to a wide array of
utility business departments.
History
Load research provides information on
consumption patterns of customers that
several utility departments and regula-
tors analyze. Our “History” column,
by Curt Puckett and Craig Williamson,
discusses the early days of load research
and major technical advances that have
occurred, including the wider use of AMI.
Society Updates
In our “Leader’s Corner,” Edvina Uzu-
novic, IEEE Power & Energy Society
(PES) vice president of Education, pro-
vides a summary of the training opportu-
nities available to our membership. The
PES University and Resource Center
offers webinars, tutorials, and materials
that can benefit audiences ranging from
novices seeking to learn about the elec-
tric power industry to well-experienced
professionals and the most advanced
academics. PES also encourages high
school students to develop an interest in
power engineering and the industry. The
Society offers a variety of scholarships
to college student members who plan to
work in our industry upon graduation.
Thanks
A special note of appreciation to Mel
Olken, who continues to provide guid-
ance and tutelage, and to IEEE pub-
lications staff who make this publi-
cation possible. Thanks to the major
contributors of this issue, especially
our guest editors, Ning Lu and Tao
Hong. A note of appreciation to Asso-
ciate Editor John Paserba and to Robert
C. Henderson, who provides edito-
rial assistance.
p&e

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.

6 ieee power & energy magazine may/june 2020

Leaders in HVDC & FACTS
have something in common.

They use the RTDS® Simulator

to test their systems.

How do you de-risk HVDC and FACTS system integration?

The RTDS® Simulator is the world standard for real time digital simulation,
used worldwide for the closed-loop testing of protection and control systems.
The Simulator is used by the major manufacturers to perform Factory System
Testing of their HVDC and FACTS projects. Leading utilities use it for pre-
commissioning testing, grid integration risk analysis, AC system interaction
studies, and more. Closed-loop testing with real time simulation avoids
project delays and increases safety, reliability, and performance.

That’s the advantage of 25 years of leadership in power system simulation.

YOUR WORLD IN REAL TIME.

RTDS.COM
 leader’s corner
V
VOLUNTEERING FOR THE IEEE
Power & Energy Society (PES) in any
capacity provides wonderful learning
experiences, occasions for meeting
knowledgeable colleagues, and the
opportunity to truly enrich our lives.
On a bittersweet personal note, 2020
is my final year as the PES vice presi-
dent of Education after serving the
typical five-year term, which is ap-
proved annually by the PES president.
PES Education provides a wide vari-
ety of training venues for our members.
Academics and industry professionals
collaborate and generate educational
materials in the form of webinars and
tutorials, fund and disseminate schol-
arships for PES students, invite stu-
dents to attend PES conferences and
cover their expenses, and participate
in many other activities that keep our
membership informed of the latest
technical developments. The following
sections provide information on several
committees within PES Education.
Power and Energy
Education Committee
The Power and Energy Education
Committee (PECC) promotes power
engineering and technology educa-
tion and professional development in
schools and industry by providing edu-
cational programs and advocating for
research opportunities. Several sub-
committees fall under the PEEC, but
I would like to emphasize the impor-
tance of the Student Meeting Activities
Digital Object Identifier 10.1109/MPE.2020.2972666
Date of current version: 17 April 2020
8 ieee power & energy magazine may/june 2020
a focus on learning
PES University & Resource Center
Subcommittee, which facilitates and
helps organize programs for students.
The subcommittee helps many gradu-
ate and undergraduate power engineer-
ing students attend the PES General
Meeting, the North American Power
Symposium, and the T&D Conference
and Exposition. For many students,
these events are their first exposure
to PES and the power industry, offer-
ing them the chance to promote their
technical interests, meet and collabo-
rate with other students, and even find
their first job. Students greatly benefit
from meeting lifelong PES members,
who eagerly share information with
upcoming generations of professionals
and each other.
PES Scholarships
PES now offers several scholarships to
its student members, who are encour-
aged to visit the PES website for details
and eligibility criteria. The programs
include the following:
✔ ✔ IEEE PES Scholarship Plus:
This program defrays educa-
tional expenses for qualified un-
dergraduates who are interested
in power and energy engineering
careers. These undergraduates
will, ultimately, replace the re-
tiring workforces in the United
States and Canada (throughout
IEEE Regions 1–6). PES pro-
vides seed money, and schol-
arships are then supported by
the energy and power industry.
Since the first year of the pro-
gram in 2011, 997 students have
Edvina Uzunovic
earned 1,691 scholarships, and
more than 500 PES scholarship
recipients are working full time.
Similar initiatives have been ex-
panded to Canada, India, Italy,
and Puerto Rico.
✔✔ The IEEE PES Erich Gunther
Future Power Innovator Schol-
arship: This fund was established
in memory of Erich Gunther to
recognize a student in power en-
gineering who pushes the bounds
of current thought to improve the
power and energy industry.
✔✔ IEEE PES Outstanding Student
Scholarships: These new awards
recognize PES student members
worldwide who are pursuing
master’s degrees and have cho-
sen an academic path leading to
an electric power and energy en-
gineering career. The criteria for
selecting recipients include their
academic achievements, contri-
butions to meeting community
and humanitarian needs, and
leadership in advancing student
engagement within PES. In 2020,
PES will begin administering
this prestigious scholarship and
plans to award up to 20 student
recipients with a plaque and prize
of US$10,000, which will be dis-
tributed directly to the student.
IEEE PES High School
Initiative
The first PES High School Initiative
was held at the PES T&D Exposition
in 2018 in Denver, Colorado. This
• •
• •
• •
 initiative acquaints high school stu-
dents with the power and energy in-
dustry, equipment, and job functions.
The first initiative was attended by 27
students from the Denver area, where
student teams presented projects at a
poster session addressing power and
energy issues. The program is sched-
uled to continue in Chicago during
the 2020 PES T&D Exposition, with
­Chicago-area high school students par-
ticipating. The PES High School Ini-
tiative is a great way to expose young
people to our industry and generate
further interest in pursuing educational
and career opportunities.
PES University webinars cover a
variety of engineering subjects, includ-
ing ongoing technical issues and state-
of-the-art topics such as microgrids,
1547 standards, and distributed energy
resources. Live webinars typically last
60 min as stand-alone presentations
or part of a formal series. Webinar at-
tendance is free, and archives are also
available for free download by PES
members. Educational credits can be
purchased for US$25 per PDH. PES
webinars are offered across four cat-
egories, shown in Figure 2.
Experienced power ­engineering
professionals benefit from PES web-
based tutorials (Figure 3), which of-
fer in-depth discussions on a wide
array of engineering topics. These
tutorials typically last approximately
80 min (60 min for the presentation
and 20  min for Q&A). They are of-
fered for US$50 to PES members, with
the chance to purchase educational
credits (0.1 CEU/1.0 PDH) for an ad-
ditional US$25 per PDH.
PES University always seeks to ex-
pand its offerings through participation

PES University and

Resource Center
PES actively offers continuing edu-
cation through numerous and timely
technical webinars and tutorials, which
are accessible online to PES members
through PES University. PES Univer-
sity began in 2019 and, together with
the PES Resource Center, hosts numer-
ous online programs, research papers,
and registration services for live edu-
cational events.
Consistent with their missions, PES
University (Figure 1) and Resource
Center support the development needs figure 1. PES University offers continuing education products. (Source: PES
of many individuals who currently University.)
work in the power and energy field or
endeavor to do so. Leading experts in
power and energy present educational
products that are available through
PES University and designed to meet
the needs of experienced engineers,
students, and nontechnical profes-
sionals.
Many of the PES webinars and
online tutorials provide professional
figure 2. The PES University webinars cover a variety of topics. (Source: PES
development hours (PDHs) and con- University.)
tinuing education units (CEUs) for
­m eeting Professional Engineering
license requirements. Moreover, for
the growing numbers of profession-
als working in the power and ener-
gy industry with no background in
power engineering, PES organizes
Plain Talk workshops, where tech-
nical aspects of the electrical power
system are explained in easy-to- figure 3. Web-based tutorials offer in-depth discussions about technical topics.
understand terms. (Source: PES University.)

10 ieee power & energy magazine may/june 2020

 pscad.com
Powered by Manitoba Hydro International Ltd.

figure 4. PES University is always looking for knowledgeable presenters.
(Source: PES University.)
by new volunteers with advanced ex-
pertise. If you would like to present a
webinar, tutorial, or Plain Talk, please
submit a proposal for consideration
(Figure 4); see https://www.ieee-pes
.org/professional-development/pes
-university/pes-university-presenter
-resources. Topics should be relevant
to engineers with a focus on growing
technical skills, providing for profes-
sional development, and informing
participants about current issues/hot
topics in the power industry. The ap-
plications for webinars, web-based tu-
torials, and Plain Talks are available at
https://www.ieee-pes.org/professional
-development/pes-university/pes-uni
versity-presenter-resources.
I hope you will take advantage
of the wealth of PES opportunities
that continuously enrich and grow
our community of power profession-
als. Thank you to all who have made
these last five years such a pleasure as
I served in my role as vice president of
Education. I look forward to seeing the
organization continue to grow and sup-
port the needs of our industry.
p&e


12 ieee power & energy magazine may/june 2020

 guest editorial
U
UTILITY RATE MAKING IS BECOM-
ing increasingly critical. The high pen-
etration of distributed energy resources
changes not only the landscape of utili-
ty operating and planning practices but
also the business models that govern
how utilities recover their service costs.
The objective of power generation and
delivery has changed from considering
only the delivery of reliable and cost-ef-
fective energy services to also account-
ing for sustainability and minimizing
adverse environmental impacts. Today’s
energy systems are evolving from tra-
ditional highly centralized suppliers to
multiple distributed small suppliers or
even self-suppliers. This evolution could
be further accelerated when major tech-
nology breakthroughs in energy stor-
age and photovoltaics (PVs) make these
systems less expensive, more reliable,
and more efficient. All of these electric
system changes greatly affect many tra-
ditional rate-making principles, meth-
odologies, and processes.
Many questions must be answered
to address the challenges posed by this
paradigm shift.
✔ ✔ What rate-making principles
need to be abolished, adjusted,
replaced, or added?
✔✔ How can a framework be created
to better connect the wholesale
market with the retail one and
pass price signals to customers
in a timely manner?
✔✔ Should regulators maintain the
status quo in retail rate making
Digital Object Identifier 10.1109/MPE.2020.2972665
Date of current version: 17 April 2020
14 ieee power & energy magazine may/june 2020
utility rate making
an intersection of economics and social good
(considering appropriate govern-
ment subsidies and policy chang-
es) instead of making aggressive
transitions to new rate systems,
such as establishing blockchain
platforms and real-time retail
markets?
✔✔ What can be done in the policy
paradigm to accelerate changes
in the electric power industry?
✔✔ How can a level playing field be
achieved for existing infrastruc-
tures, renewables, energy stor-
age technologies, and customer-
owned distributed generators to
provide grid services?
✔✔ How can reliability be best fac-
tored into rate making?
✔✔ If self-generation becomes a ma-
jor way of producing electricity,
how should transmission and
distribution be priced?
✔✔ What are the best rate-making
practices for regulated and non-
regulated utilities and for market-
and nonmarket-based systems?
In This Issue
The articles in this issue of IEEE
Power & Energy Magazine discuss ap-
proaches, new trends, and future plans
for understanding how utilities in dif-
ferent countries plan to answer these
questions. Five well-written articles
discuss various aspects of utility rate
making under different operation and
regulation rules worldwide, including
the perspectives of regulated and de-
regulated utilities in the United States,
China, Europe, and South America.
Ning Lu and Tao Hong
The electricity bills sent to most
American households today consist of
two parts: a fixed monthly charge and
an energy charge based on the total en-
ergy usage of the month measured in
kilowatt-hours. The first article, “The
Tariffs of Tomorrow” by Faruqui and
Bourbonnais, explains the shortcom-
ings of such two-part rates. The authors
also offer an outlook on tomorrow’s
tariffs structured in three-part rates
consisting of a fixed monthly charge,
a time-varying charge, and a demand
charge. Moreover, they present many
case studies from U.S. utilities, such
as Arizona Public Service, Oklahoma
Gas and Electric, and Commonwealth
Edison, to illustrate various ways of
implementing the proposed rates.
The United Kingdom’s energy reg-
ulator is actively reforming network
charges that facilitate a transition to a
low-carbon future. The existing pricing
methodologies reflect only the usage of
network assets by users but downplay
the unused capacity. “Flexible Network
Pricing Encourages Greater Sharing”
by Li et al., the second article in this
issue, highlights the importance of
long-term uncertainty and long-term
network investment in network pricing.
The authors propose new directions in
pricing principles that reward custom-
ers who reduce network uncertainties
and network operators who best man-
age the uncertainties.
“Distribution Network Rate Mak-
ing in Latin America” by Moreno et al.
discusses the regulatory reforms imple-
mented by various ­r egulators in the
Make costly nuisance outages disappear.

REDUCE NUISANCE OUTAGES WHILE IMPROVING CUSTOMER SATISFACTION WITH

THE VACUFUSE SELF-RESETTING INTERRUPTER.
™

70% of fuse operations on overhead distribution transformers are for nuisance outages. While
there’s no trick for keeping the lights on at the end of lateral lines, there is technology. With the
VacuFuse™ Self-Resetting Interrupter, you’ll have advanced fault-testing technology where you need
it most at the outer reaches of the grid, improving customer satisfaction and saving you needless
maintenance expenses.
Plan for advanced protection with S&C, the worldwide leader in reliability.

See the difference at sandc.com/vacufuse19 © S&C Electric Company 2019. All rights reserved.
 distribution sector in Latin American
countries, including Colombia, Peru,
Chile, Brazil, and Argentina. Alterna-
tive tariff structures currently being dis-
cussed in Chile and Brazil are introduced
as examples of the reformation efforts.
In “China’s Solar Subsidy Policy,”
Dong et al. discuss the impact of sub-
sidy policies on the nation’s PV indus-
try over the past 15 years. The country’s
rapid growth of a new industry was
achieved mainly by governmental ef-
forts. We hope that this article provides
you with a different perspective on
policy making for fostering new tech-
nologies and an alternative approach
for rate making, especially in countries
where utilities are closely regulated by
the government.
The rate-making process is non-
trivial. No matter what rates are being
established and applied to the end users,
they should be based on load research, a
comprehensive study of how customers
use electricity. The last article, “Utility
Load Research” by Puckett et al., fo-
cuses on the origin of rate making. The
authors introduce the load research life-
cycle as well as the interaction between
load research and various utility busi-
ness functions such as rate making, load
forecasting, and planning. They also of-
fer their outlook on the next decade of
load research areas and the key factors
for load research to continue to thrive.
In the “In My View” column, Satch-
well et al. discuss key drivers of retail
rate reforms, design considerations of
retail rates, and implications for the de-
ployment of distributed energy resources.
The broad implications of rate reforms for
utilities, customers, and society are dis-
cussed. The authors recommend that reg-
ulators and policy makers establish rate
structures that weigh these factors against
each other while balancing other consid-
erations and stakeholder perspectives.
The articles in this issue would not
be made possible without the contribu-
tions of the volunteers who wrote and
reviewed them. We also thank the en-
tire IEEE Power & Energy Magazine
editorial board for their continued sup-
port. We hope you enjoy reading the
articles and find this issue informative
and thought-provoking.
p&e

Dynamic Security Assessment Software
Offering a simple, user-friendly interface
with extensive analysis options, DSAToolsTM
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
Advanced modeling and analysis features include:
• Comprehensive capabilities for voltage, transient, small signal, and
frequency stability assessment
• Integration with EMS/WAMS for on-line DSA to assess security of
real-time system conditions
• Computation of secure operation regions (stability limits)
• Recommendation of remedial actions
Experience and support:
• Prominent provider of the on-line DSA technology with over 50
installations worldwide including 8 out of 9 ISOs in North America
• Consulting services for custom software and model development,
system studies, and training

VSAT TSAT SSAT

Contact us:
dsainfo@powertechlabs.com

16 ieee power & energy magazine may/june 2020

 Save the Date
Join us next August in Montreal for the
2019 IEEE Power & Energy Society General Meeting

August 2 - 6, 2020 — Montréal, Canada

The IEEE PES General Meeting Technical topics include:

is the premier global technical
conference for electric power and · Utility Grid of the Future
energy professionals and attracts · Microgrids
more than 3,500 people from · Cybersecurity
more than 70 countries. · IEEE Standard 1547 and Its
Impact on the Grid
Network with your industry peers · Energy Storage
from across the globe · Distributed Energy Resources
(DERs)
Attend one of over 150 technical · Digital Substations and IEC
presentations and panel 61850
discussions
· Grid Modernization
· Utility Best Practices
Enjoy technical tours of local
facilities · Big Data and Utility Analytics
· Asset Management
Earn continuing education credits · Smart Grid
(CEUs and PDHs) · and more…

Visit pes-gm.org for on-going updates and details.

Digital Object Identifier 10.1109/MPE.2020.2981787

 The Tariffs
of Tomorrow
By Ahmad Faruqui and Cecile Bourbonnais
T
TODAY, MOST AMERICAN RESIDENTIAL CUSTOMERS
pay for electric service via tariffs that are structured as
two-part rates consisting of a fixed monthly charge and a
volumetric energy charge expressed in U.S. dollars per kilo-
watt-hour. The fixed-charge component generally comprises
a small portion of the bill, which is dominated by the volu-
metric charge.
California’s energy crisis in 2001–2002 triggered an
important discussion about the need to better connect retail
and wholesale markets. Since then, scores of pilot programs
Digital Object Identifier 10.1109/MPE.2020.2972136
Date of current version: 17 April 2020
18 ieee power & energy magazine 1540-7977/20©2020IEEE
have been carried out with time-varying rates. The empirical
evidence shows that customers can understand and respond
to incentives provided by new tariffs that convey the cost
structure of electricity to customers. Additionally, smart
meters now track the energy use of half of all U.S. residen-
tial customers, removing a major barrier to the deployment
of modern tariffs.
The introduction of smart digital technologies, changing
consumer tastes, and new state policies promoting renew-
able energy sources have cast doubt on the sustainability
of the utility business model based on traditional two-part
tariffs. This shift comes at a time when many utilities are
making significant investments to modernize the grid and
may/june 2020
 Innovations in Rate Designs
integrate distributed energy re-
sources (DERs). These capital in-
vestments, along with most other
utility costs, do not vary with the
volume of electricity consumed.
They thus cannot accurately be
reflected through a volumetric
charge. This has resulted in large
residential customers effectively
subsidizing smaller customers,
who, as a result of low energy us-
age, pay less relative to the costs
they impose on the system. An
additional limitation of flat volu-
metric charges is that they fail to
capture the effects of temporal
and seasonal variability, which
result in peak periods that are
more expensive to utilities than
off-peak periods.
Consequently, a m isa lign-
ment between utilities’ cost and
rate structures currently exists,
particularly with respect to the
residential class. In response to
this challenge, many utilities
©ISTOCKPHOTO.COM/TRONAND
are introducing innovative, cost-
reflective tariffs better aligned
with both their own needs and
those of their customers. In short,
utilities seek to develop the tariffs
of tomorrow.
The Principles of Rate Making
The same seminal principles that have guided rate design for
decades should motivate regulatory designs of future tariffs.
James Bonbright’s principles for public utility rates, first
published in 1961, have remained in place in spite of various
technological advances and evolving industry trends. These
principles continue to be widely accepted (see the “For Fur-
ther Reading” section). As shown in Table 1, they can be
distilled to five key criteria.
Economic efficiency means that the resources are not
used to generate and deliver electricity in such a way that
their reallocation will increase the total cost to consumers
and producers, resulting in an overall increase in the total
cost to society. In other words, no resources consumed in
the delivery of electricity should be wasted through over-
investment and operating costs.
may/june 2020 ieee power & energy magazine 19
The second criterion, equity, refers to fairness among
customers and between the utility and customers. Although
rate design nearly always involves some degree of cross-sub-
sidy, a utility should aim to remove unintentional subsidies
between customer types. According to Bonbright, a natural
way to achieve equity among customers with different load
profiles and consumption values is through cost-reflective
rates. Under such rates, customers who incur high costs for
the system will pay proportionally higher amounts than low-
cost customers.
Revenue stability refers to the utility’s ability to recover
its costs through a sufficient and predictable level of reve-
nues. Bill stability, the fourth criterion, then stipulates that
while the utility must recover its costs, ideally through cost-
reflective rates, it must also protect customers from unman-
ageable fluctuations in their bills. Although new rates will
nearly always result in bill increases for some customers,
utilities can take steps to minimize seriously adverse and
unexpected impacts, for instance by gradually implementing
changes to rates.
Finally, customer satisfaction is needed for the successful
implementation of any changes to the pricing structure. If
not properly explained or rolled out, even simple rates can
cause confusion and subsequently trigger a backlash from
customers. Regulators and utility companies who anticipate
such an adverse reaction from customers will resist imple-
menting new rates.
Rethinking Present-Day Rate Making
Underpinning all five key criteria is the principle of cost
causation, which Bonbright considered the most important
standard of reasonable rates. In the case of electricity, costs
consist of three elements:
✔✔ a fixed cost for servicing the customer
✔✔ a capacity cost associated with the distribution grid,
the transmission network, and the power plants
✔ ✔ an energy cost associated with the production of
electricity.
Cost causation says that revenues should reflect these three
costs. However, given the ubiquity of two-part tariffs for
residential customers, this has historically not been the case.
Instead, as typified in Figure 1, utilities employing two-part
rates typically recover most of the costs of residential service
on a volumetric basis. To achieve this, they build nonvari-
able fixed and capacity costs into the energy charge using
assumptions about class load factors and applying them
equally to all customers in the class.
As a result, a common violation of the equity principle
occurs under two-part rates when “peaky” customers, who

Key Criterion
1) Economic efficiency
2) Equity
3) Revenue stability
4) Bill stability
5) Customer satisfaction
table 1. Bonbright’s principles of rate making.
Bonbright Principle
Static efficiency of the rate classes and rate blocks in discouraging wasteful use of service while
promoting all justified types and amounts of use
Reflection of all of the present and future private and social costs and benefits occasioned by a
service’s provision (i.e., all internalities and externalities)
Dynamic efficiency in promoting innovation and responding economically to changing demand and
supply patterns
Fairness of the specified rates in the approportionment of total costs of service among the different
ratepayers to avoid arbitrariness and capriciousness and attain equity
Avoidance of undue discrimination in rate relationships to be, if possible, compensatory
Effectiveness in yielding total revenue requirements under the fair-return standard without any socially
undesirable expansion of the rate base or socially undesirable level of product quality and safety
Revenue stability and predictability, with a minimum of unexpected changes that are seriously adverse
to utility companies
Stability and predictability of the rates themselves, with a minimum of unexpected changes that are
seriously adverse to utility customers and with a sense of historical continuity
The related, practical attributes of simplicity, certainty, convenience of payment, economy in
collection, understandability, public acceptability, and feasibility of application
Freedom from controversies as to proper interpretation

consume more in high-cost hours, are subsidized by less peaky

100 customers with a higher load factor. A customer with a low load
Fixed
Fixed factor may consume the same kilowatt-hours as a customer
90 with a high load factor and thus pay the same bill under a
80 two-part rate but impose much higher costs on the system.
The recent distributed generation (DG) expansion has
70 further compounded inequity concerns by introducing a
Variable
60
cross-subsidy from non-DG consumers to DG consumers
(often referred to as prosumers, since they both produce and
(%)

50 consume energy). Under net energy metering, available

Variable
40
in 43 states, a utility credits prosumers for any generation
they produce back into the grid. As a result, prosumers
30 who produce as much as they consume can achieve very
Demand
low energy bills, even as they continue to require grid ser-
20
vices from the utility. Given that fixed infrastructure costs
10 make up a large share of utility costs, this allows DG con-
sumers to avoid paying for a key service that they continue
0
Utility Costs Customer Bill to use. Meanwhile, non-DG consumers must then cover a
disproportionate share of capacity costs through their own
Cost Categories energy charges.
Fixed (US$/Month) Under most current rate designs, this misalignment will
– Metering only grow as DER penetration continues to develop, trans-
– Billing
forming the grid and the way in which customers interact
– Customer Service
with it. Dynamic pricing has been proposed as a means to
Variable (US$/kWh)
– Fuel address this misalignment, while also facilitating renewable
– Operations and Maintenance penetration. This also comes at a time when states are adopt-
Demand (US$/kW) ing increasingly aggressive goals for their renewable portfo-
– Transmission Capacity lio standards, with the city of Washington, D.C., and the
– Distribution Capacity states of California and Hawaii all targeting 100% renew-
– Generation Capacity
able energy within the next two to three decades. Although
Bonbright’s principles remain as relevant as ever, the world
figure 1. An illustrative misalignment between residential to which they apply is changing, and tariffs must now adapt
rates and costs. to meet that change.

20 ieee power & energy magazine may/june 2020

Proposing Tariffs of Tomorrow
Of course, tariffs have been slow to adapt in part because
of technological limitations. As Bonbright recognized, cost-
reflective rates require having the necessary metering infra-
structure in place, and, for decades, this infrastructure was
only available to large commercial and industrial customers.
For residential and smaller commercial and industrial custom-
ers, it was less expensive and easier for utilities to simply bill
metered kilowatt-hour usage, despite the resulting issue of
fixed-cost recovery. However, the widespread deployment of
smart meters is making it much more affordable to measure
demand, even for residential customers. As this change pro-
gresses, the implementation of cost-reflective rate designs for
all customers is becoming more accessible than ever before.
Given these advances, the tariffs of tomorrow are likely
to consist of three parts corresponding to the three elements
that comprise electricity costs: a fixed monthly charge, a
time-varying energy charge, and a demand charge.
The fixed charge (sometimes referred to as a customer
charge, service charge, or facilities charge) is expressed in
dollars per month. It reflects the costs of servicing the cus-
tomer, such as billing, metering, and customer service.
The time-varying energy charge, expressed in U.S. dol-
lars per kilowatt-hour, recovers energy costs, either in the
form of a simple time-of-use (TOU), critical peak pricing
(CPP), variable peak pricing (VPP), or real-time pricing
(RTP) rate. A simple TOU rate defines peak periods during
which prices are higher than in off-peak periods and is cur-
rently the most common form of time-varying rate. How-
ever, programs like CPP, VPP, and RTP are considered purer
forms of dynamic pricing in that they are based on actual
market conditions and thereby a better signal of customer
changes in the utility’s costs.
The demand charge, expressed in dollars per kilowatt,
recovers grid capacity costs, typically based on peak elec-
tricity consumption over a span of 15, 30, or 60 min. It may
be either coincident or noncoincident. A coincident demand
charge applies to a customer’s peak electricity consumption
at the time of the maximum system usage, whereas a non-
coincident demand charge measures a customer’s highest
usage during the month, regardless of the time of day.
Some critics of noncoincident demand charges argue that
they do not reflect the utility’s cost structure, since they may
not coincide with the capacity costs driving system peak and
the need for new infrastructure where beneficiaries should
pay. Many of these critics favor time-varying energy charges
without any demand charges. However, these two charges
are not at odds and can be offered simultaneously to supple-
ment one another. While time-varying energy charges can
dynamically recover energy costs and encourage load shift-
ing, noncoincident demand charges for residential customers
can recover distribution-related capacity costs and encour-
age overall greater efficiency. After all, customers expect
electricity service whenever they need it, at any time of the
day. As a result, a utility must build adequate infrastructure
may/june 2020 ieee power & energy magazine 21
to meet a customer’s peak usage, regardless of when it
occurs. However, a noncoincident demand charge effectively
serves as a proxy for the localized cost of connecting a cus-
tomer to the grid.
Another option for recovering the costs of the grid is to
increase the fixed charge, but a fixed charge may not fully
account for a customer’s size or send price signals the way
that both coincident and noncoincident demand charges can.
The implementation of a demand charge further addresses
the equity criterion by minimizing cross-subsidies from cus-
tomers with high load factors to those with low load factors,
who may have a low kilowatt-hour consumption but a high
kilowatt demand that a volumetric rate cannot capture.
In terms of the cross-subsidy from non-DG consumers
to DG consumers, one solution recognizes the unique fea-
tures of the manner in which such customers interact with
the grid, as exemplified by their load shapes, and to price
electricity accordingly. Thus, utilities in a number of states
are now asking their commissions for permission to estab-
lish a new class for prosumers and to price electricity to
them in a manner that reflects the cost of serving electricity
to them.
The creation of a separate class may discourage the adop-
tion of rooftop solar and other DERs. Cost-based, three-part
tariffs may also encourage the adoption of certain technolo-
gies like electric vehicles by enabling time-varying rates that
give vehicle owners the option of charging vehicles during
the less-expensive off-peak periods. Neither outcome is inher-
ently wrong. Rather, regulators need to understand that a more
advanced rate design requires a greater understanding of out-
comes. Rate design is not always the place to incent technolo-
gies. Other policies outside of those applying to rates can do
this, for instance, through tax credits, rebates, or renewable
energy certificates.
Obstacles to the Future
Undoubtedly, given the diversity of customer tastes and
expectations, these new tariffs will not appeal to all residen-
tial customers. Risk-averse customers might resist the vola-
tility of time-varying rates and the risk implicit in demand
charges. Utilities can offer these customers rate options that
better suit their risk profiles, while still incorporating the
cost-reflective concepts underpinning the tariffs of tomor-
row. Possible rate options are listed in Table 2.
From among these many options, the most risk-averse
customers would likely go with a guaranteed bill, which is
constant regardless of the volatility in their load profiles or
their electricity prices. On the other hand, risk-taking cus-
tomers would likely go with a RTP rate that, on average,
would likely give them a lower average price. Each of these
rate options presents a unique tradeoff between the bill sav-
ings that customers would experience and the risk that they
would be exposed to in the form of bill volatility. When they
are plotted out in the savings-risk space, they yield an “effi-
cient pricing frontier.”
 table 2. Rate-design options.
Rate Design Definition
GB Customers pay the same bill every month, regardless of usage.
Flat rate A uniform US$/kWh rate is applied to all usage.
Demand charge Customers are charged based on peak electricity consumption, typically over a span of 15, 30, or 60 min.
TOU The day is divided into time periods, which define peak and off-peak hours. Prices are higher during the
peak-period hours to reflect the higher cost of supplying energy during that period.
CPP Customers pay higher prices during critical events when system costs are highest or the power grid is severely
stressed.
IBR Customers are charged a higher rate for each incremental block of consumption.
PTR Customers are paid for load reductions on critical days, estimated relative to a forecast of what they would
have otherwise consumed (their baseline).
VPP During predefined peak periods, customers pay a rate that varies by utility to reflect the actual cost of
electricity.
DSS Customers subscribe to a kilowatt demand level based on the size of their connected load. If they exceed
their subscribed level, they must reduce their demand to restore electrical service.
TE Customers subscribe to a baseline load shape based on their typical usage patterns and then buy or sell
deviations from their baseline.
RTP Customers pay prices that vary by the hour to reflect the actual cost of electricity.
GB: guaranteed bill; IBR: inclining block rate; PTR: peak-time rebates; DSS: demand subscription service; TE: transactive energy.

The efficient pricing frontier reflects Bonbright’s crite- charged more in peak hours when the cost of energy is high.
rion of equity in that, while it is good to offer choices to cus- Figure 2 illustrates this tradeoff.
tomers, each rate must still recover the cost of serving each This tradeoff applies to all customers, including those
customer correctly to avoid creating intercustomer subsidies. who are disadvantaged or have a low income. In fact, low-
As a result, to capture the impact of customer load volatil- income customers tend to have flatter load profiles and thus
ity on wholesale prices, low-risk rates should yield lower are likely to benefit from rates that charge higher prices in
potential savings than higher-risk rates that better reflect the peak hours. On the other hand, critics of dynamic new rates
time-varying cost of energy. For instance, the average price argue that disadvantaged and low-income customers gener-
implicit in a standard tariff should be higher than the aver- ally have more difficulty responding to new rates, either due
age price implicit in a TOU rate, under which a customer is to physical or flexibility constraints.
Such apprehension about cus-
tomer reactions largely explains
why the adoption of smart rates has
not kept up with the adoption of
smart meters. As important as eco-
nomic efficiency may be, it remains
only one of Bonbright’s key crite-
ria, and concerns about customer
Reward (Bill Savings)

RTP
TE satisfaction can immediately
DSS
VPP halt any changes to the pricing
PTR CPP
TOU structure. However, as described in
Demand Charge more detail later in this article, pro-
Standard Tariff tections can be applied for vulner-
Higher FC able customers to mitigate risk and
ease the transition.
GB With DR
Another concern surrounding
three-part tariffs is that, by reduc-
GB Bill ing the volumetric portion of cus-
tomers’ bills, they may discourage
Risk (Bill Volatility)
energy efficiency. However, cost-
reflective tariffs are the best way
figure 2. Efficient pricing frontier. FC: fixed charge. to simultaneously promote equity

22 ieee power & energy magazine may/june 2020

between customers as well as economic efficiency in the
use of scarce capital and fuel resources. Most of the tension
between energy efficiency and rate design stems from the
recovery of fixed costs, which is best addressed through sep-
arate financial mechanisms such as revenue decoupling or
lost-revenue adjustment mechanisms. Such mechanisms pro-
tect utilities from fluctuations in sales, either by completely
breaking the link between revenue and sales or allowing
utilities to recover lost revenues resulting specifically from
energy-efficiency programs.
Case Studies
Despite concerns about customer response, several utili-
ties have already begun offering dynamic rate options with
promising initial results. In Maryland, for example, the three
largest utilities (Pepco, Delmarva Power, and Baltimore Gas
and Electric) have each implemented peak-time rebates as
the default tariff to all their customers. These programs have
resulted not only in peak load reductions for the utility but
also customer savings and satisfaction.
Figure 3 illustrates the results of 332 TOU, CPP, and PTR
pricing experiments. Some of the 332 experiments had very
high peak-to-off-peak ratios and could not be included in
the figure. These results further support the idea of customer
peak-load reductions in response to higher peak-to-off-peak
price ratios. Some experiments provided a pure price signal
to customers, whereas others included enabling technolo-
gies, such as smart thermostats, with the pricing signal.
Around 14% of all U.S. utilities, including roughly half
of all investor-owned utilities, now offer a residential TOU
rate. Among these TOU rates, 6% now also include a demand
charge on top of the time-varying volumetric charge. Although
adoption rates for dynamic tariffs remain low, the successes
of a few utilities show promise for customer willingness to
adopt future tariffs.
The Arizona Public Service (APS) has adopted TOU
and demand charges. Approximately 57% of APS’s resi-
dential customers are enrolled in TOU rates, and 20% of
these customers also pay a demand charge. To ease the
transition since it first implemented demand charges in
1989, APS made a significant commitment to providing
customers with information on the various rate options.
Ultimately, this resulted in a rate-comparison tool that
customers either could use on their own or with the assis-
tance of a customer service representative. Additionally,
APS began to provide customers with an annual analysis
of their usage, energy savings recommendations, and a
rate recommendation if they would benefit from switch-
ing. The company also provides rate comparisons for cus-
tomers who contact the call center. As APS began rolling
out more advanced metering to all customers, the rate-
comparison tool was modified to reflect an analysis of
actual load data. Over the years, APS has not marketed
any particular rate option to its customers. Instead, it has
provided them—through the rate-comparison tool—the
may/june 2020 ieee power & energy magazine 23
information they need to make an informed decision about
what rates are best for them.
APS is not alone in introducing demand charges for
residential customers. At least 51 demand charges are now
being offered in 22 states by 43 utilities. Demand charges
tend to be disproportionately prevalent among coopera-
tives, which are owned by the customers and thus have an
additional responsibility to guarantee equity among mem-
bers. Some cooperatives have even implemented manda-
tory demand charges for certain customers. For example,
the Salt River Project has mandatory demand charges for
customers with rooftop solar panels.
Successful dynamic pricing case studies are often supported
by smart technologies. In 2012, Oklahoma Gas and Electric
(OG&E) rolled out its SmartHours program, a time-based pro-
gram offered to residential customers on an opt-in basis. As
part of the program, the utility offers its customers the option to
install smart thermostats, which will adjust in response to price
signals according to customers’ programmed preferences. How-
ever, customers also maintain full control, so they can override
these settings and choose not to respond to OG&E’s alerts. A
fifth of the utility’s customers have signed up for SmartHours
and achieved significant savings on their electric bills.
ComEd has similarly partnered with Nest, a smart ther-
mostat provider, for its AC Cycling Program. Under this pro-
gram, ComEd offers customers a US$100 rebate to purchase
and install a Nest thermostat, which it can then control during
a maximum of 10 peak events each summer. The program
runs from June through September, and customers receive a
US$10 bill credit for each month of participation. The smart
thermostat is designed to reduce energy usage while main-
taining a comfortable temperature, but if customers start to
feel too warm, they can manually override it at any time.
Transitioning in the Digital Age
Increasingly widespread advancements in technology are
integral to empowering and informing customers about their
60
Peak Impact (%)
40
20
0
2 4 6 8 10
Peak-to-Off-Peak Price Ratio
Rate Design
With Technology Price Only
figure 3. Price responsiveness under dynamic pricing
experiments.
 power-supply choices. Many customers today are likely
to have not only smart meters but also smart digital appli-
ances, thermostats, and digital apps to track and optimize
their own usage patterns. In the age of constant connectivity
and the Internet of Things, newer generations of customers
are already well equipped to embrace this shift and adapt to
more open and responsive two-way communication. In fact,
many of these customers have already embraced modern
pricing designs in other aspects of their life. They encounter
dynamic pricing every day, from airlines and hotels to con-
certs and movies, and they understand how to interpret and
respond to surge pricing from ride sharing apps like Uber.
The success of all these designs is in, large part, predi-
cated on their transparency and simplicity and in the
manner in which they are clearly conveyed to customers.
However, there is no reason utilities cannot likewise com-
municate their prices and offer more transparent bills that
minimize unnecessary line items and unclear calculations.
Even relatively complex tariffs can be made simple to cus-
tomers, especially with the profusion of modern-day tech-
nologies. With such technologies, customers can respond to
dynamic rates and achieve savings without constant moni-
toring. For example, smart thermostats can automatically
adjust home temperatures with minimal programming, and
utilities can send messages and text alerts to prepare cus-
tomers for peak events.
Utilities can also use these same tools and technologies
to understand and fulfill their customers’ needs, for instance,
through pilot programs and experimental studies, which can
help predict customer response. To minimize adverse reac-
tions at the beginning of the transition, utilities can also
progressively ease customers from their old rates to the new
ones through initial bill protections or transition rates.
Any change in rates will invariably cause some customer
bills to go up and others to go down. The change in bills will
depend on each customer’s energy consumption, load factor,
and load shape as well as the structure of the new rate. For
example, a customer with a very high load factor would likely
see a lower bill on a TOU energy rate compared to a flat rate,
whereas a customer with a very low load factor would likely
see a higher bill with the new cost-reflective tariff, especially if
peak coincides with times of peak system energy prices. Cus-
tomers with low kilowatt demand would see lower overall bills
with the introduction of a demand charge, while customers with
low kilowatt demand and a high load factor would see a lower
bill with a demand charge offered in combination with a TOU
energy rate. As a result, while most bills would not change sub-
stantially, this may not be the case for all customers.
Customers whose bills do go up substantially under three-
part tariffs will likely complain, and if their complaints are
picked up by the media, the move toward new tariffs could
come to a standstill. Thus, implementing more efficient,
cost-based tariff reform always requires serious consider-
ation of customer satisfaction and education. However, given
all of the changes and technological advancements currently
24 ieee power & energy magazine may/june 2020
sweeping the industry, this should not be a reason to resist
progress and maintain an outdated status quo.
Instead, to balance these challenges and ease the transi-
tion, utilities should first analyze customer bill effects to bet-
ter mitigate these impacts and anticipate customer reactions.
Before even implementing the three-part tariff, utilities and
regulators can calculate expected bill changes for a represen-
tative sample of customers, assuming existing load profiles.
These results should be plotted in the form of a propeller chart
identifying which customers will see higher bills and which
will see lower bills. The utility can then analyze the sociode-
mographic and regional characteristics of those with signifi-
cantly higher bills and understand the degree to which this
impact affects all users.
Next, utilities can carry out simulations to predict and
incorporate customer responses. Models, such as the Price
Response Impact Simulation Model, which was initially
developed to quantify the impact of TOU and dynamic pric-
ing in California’s 2003–2004 statewide pricing pilot, can
perform a bill-impact analysis that allows for a certain amount
of demand response. The adverse bill impacts under the new
three-part tariffs should be lower than they were without
demand response, as customers shift their usage between
periods in response to new average and relative prices.
If the adverse bill impacts are still significant for a certain
group of customers, utilities can take several approaches to
ease their transition. These include the following:
✔✔ Gradualism: Roll out the new rates gradually for each
rate-design element to preserve bill stability, for in-
stance, by gradually ramping up the peak price of a
new TOU rate.
✔✔ Bill protection: Provide these customers with bill pro-
tection in the first year and then gradually phase it out
over three to four years
✔ ✔ Protections for vulnerable customers: Make the
three-part rate optional for vulnerable customers,
while making it mandatory for the largest customers
and the default rate design for all other customers. Al-
ternatively, offer customers who are vulnerable or will
experience adverse bill impacts financial assistance
for a defined period of time.
✔✔ Enabling technologies: Install enabling technologies,
such as smart thermostats, on customers’ premises
that allow them to more easily respond to price signals
under the new tariffs.
✔✔ Two-staged rollout: Structure the rate into two stages.
Under the first stage, charge customers the current
rate, if their usage resembles their historical usage
over a given reference period. Under the second stage,
charge customers the new tariffs for any deviations
from their historical usage.
Utilities can also prepare for the rollout by conducting
focus groups with customers to test possible education and
marketing campaigns and gauge customer understanding of
the new tariffs. Based on their findings, utilities can then
make appropriate modifications in language (and possibly
in the rate-design parameters, such as the magnitude of the
demand charge and the charges for energy by TOU as well
as the duration and temporal location of the peak period)
to make the new tariffs more understandable to customers.
To better test customer acceptance and the amount of
demand response to the new tariffs, utilities can conduct
additional pilots. Pilots should be designed on scientific prin-
ciples of experimental design that would preserve the inter-
nal and external validity of the results, allowing them to be
extrapolated to the population of customers and applied to
different prices than the ones being tested. Randomized con-
trols, randomized encouragement, and matching controls are
different ways of preserving pilot validity.
Given these pilot results, the utility must then decide on the
rollout strategy and whether it should be mandatory, default,
or opt in. As a point of reference, Fort Collins rolled out TOU
energy rates to all its residential customers on a mandatory
basis as of 1 October 2018. Also, as of 1 October 2018,
the Sacramento Municipal Utilities District began rolling
out time-of-use energy rates with a US$20-per-month service
charge on a default basis in California. In Ontario, Canada,
TOU rates have been rolled out as the default tariff for energy
supply since 2007 (which means that customers can opt out of
the rate). At this point, only 10% have opted out. The distribu-
tion tariff is a flat bill. Finally, the utility must track the deploy-
ment of the new tariffs, survey customers for feedback, set up
social media sites and monitor the conversation, and make any
necessary modifications to the rate design.
Conclusions
TOU rates were first tested in the United States in the late
1970s. Fourteen pilots showed that customers accepted such
rates and responded to the price signals in a fairly predict-
able fashion. The tests were motivated by the energy crisis of
the 1970s and the passage of the Public Utilities Policies Act
in the United States under the Carter administration, which
put the spotlight on TOU rates.
But not much happened, since smart meters were lack-
ing, and other priorities surfaced. Most notably, there was
a newfound interest in restructuring the electricity industry
and offering customers a retail choice of energy suppliers.
Tariff reform was put on the back burner.
The California energy crisis of 2001–2002 put the focus
back on tariff reform, along with advanced metering and
demand response. Hundreds of pilots were carried out in the
United States and abroad showing that customers accepted
such rates and responded in a fairly predictable pattern.
Smart-meter deployments began in earnest and now encom-
pass half of the residential population. Modern tariffs, how-
ever, touch fewer than 5% of customers.
Progress is stymied because of the fear of the unknown
and by a concern that rate reform will harm low-income
may/june 2020 ieee power & energy magazine 25
customers, customers with disabilities, and senior citizens.
Customer backlash remains a primary concern. Secondary
concerns relate to the inability of several billing systems to
handle the new tariffs.
The way forward is to conduct pilot programs where none
have been done to deploy the new tariffs, which could be
offered with temporary bill protection. Additionally, they
should be accompanied with customer-education and mar-
keting campaigns to ease the transition.
Ultimately, these three-part rates provide a favorable option
as the default tariff to maximize economic efficiency and inter-
customer equity. They can be paired with alternative designs
that would be offered only on an opt-in basis.
There are signs of change. Technological advancements
have already cleared some of the major barriers to tariff
reform. With slowing sales and growing DER penetration, the
day is not far off when the tariffs of tomorrow will become
the tariffs of today.
For Further Reading
J. Bonbright, Principles of Public Utility Rates. New York:
Columbia Univ. Press, 1961.
M. Everett, C. Fang, A. Ramirez, and J. Schneider, “A
modern rate architecture for California’s future,” Pub-
lic Utilities Fortnightly, vol. 156, no. 12, pp. 8–14, Nov.
2018.
A. Faruqui, H.-P. Chao, V. Niemeyer, J. Platt, and K.
Stahlkopf, “Analyzing California’s power crisis,” Energy J.,
vol. 22, no. 4, pp. 29–52, 2001. doi: 10.5547/ISSN0195-6574-
EJ-Vol22-No4-2.
A. Faruqui and J. R. Malko, “The residential demand for
electricity by time-of-use: A survey of twelve experiments
with peak load pricing,” Energy, vol. 8, no. 10, pp. 781–795,
1983. doi: 10.1016/0360-5442(83)90052-X.
A. Faruqui, S. Sergici, and C. Warner, “Arcturus 2.0: A
meta-analysis of time-varying rates for electricity,” Elec-
tr. J., vol. 30, no. 10, pp. 64–72, Dec. 2017. doi: 10.1016/j.
tej.2017.11.003.
S. S. George and E. Bell, “Key findings from California’s
recent statewide TOU pricing pilots,” Electr. J., vol. 31, no.
8, pp. 52–56, Oct. 2018. doi: 10.1016/j.tej.2018.09.013.
J. Lazar and W. Gonzalez, “Smart rate design for a smart
future,” Regulatory Assistance Project, Montpelier, VT,
July 2015. [Online]. Available: https://www.raponline.org/
wp-content/uploads/2016/05/rap-lazar-gonzalez-smart-rate
-design-july2015.pdf
Biographies
Ahmad Faruqui is with The Brattle Group, San Francisco,
California.
Cecile Bourbonnais is with The Brattle Group, San
Francisco, California.
p&e

 Making the
Grid Work
for Itself and
Distributed
Energy
Resources

C
CONSU M ERS I N T H E U N I T ED K I NG -
dom typically spend, on average, £33 billion/
year for electricity, of which a significant portion
(21% in 2017) is paid to distribution companies for deliv-
©ISTOCKPHOTO.COM/ERHUI1979
ering electricity to customer premises. To protect consum-
ers’ interest, Ofgem (Office of Gas and Electricity Markets),
the energy regulator in the United Kingdom, is actively
reforming network charges so that they remain usable for
By Furong Li, the purpose of facilitating the industry’s transition to low-
carbon emissions.
Chenghong Gu, Network charges are applied to generators and suppli-
and Ran Li ers for their use of the system to transport energy over a
distance. Utilities employ network charges as commercial
instruments that influence network users’ behavior, which

Flexible
promotes the most efficient use of the network and achieves
cost recovery. The fundamental principle for traditional net-
work pricing methodologies also reflects the use of network
assets attributable to network customers, such as distributed

Network Pricing
Encourages
Greater Sharing
Digital Object Identifier 10.1109/MPE.2020.2972137
Date of current version: 17 April 2020

26 ieee power & energy magazine 1540-7977/20©2020IEEE may/june 2020

and central generation stations and suppliers. The unused
network capacity is typically socialized.
Decarbonization and digitalization will fundamentally
disrupt this principle, which has been with the industry for
roughly 30 years. This is because significant demand would
be met by local generation, particularly if demand is of a
flexible nature, leading to significant spare capacity in parts
of the network. If the current pricing principle continues,
spare capacity in the system could become more wide-
spread. Lower power flows would lead to very expensive sys-
tem charges, which would further discourage network use,
resulting in a high risk of assets being stranded.
We believe that the current pricing and regulatory frame-
work does not account for the long-term uncertainty of the
relationship between energy customers and network invest-
ment. The absence of uncertainty in network charges will
pose a serious threat to asset utilization over time. The
incentives provided by network charges do not recognize the
potentially costly investment mistakes arising from future
uncertainty in asset utilization. As a result, network users
will not be rewarded if they can manage their generation
or demand uncertainties nor will they be penalized if they
escalate the uncertainties. Under such a pricing framework,
uncertainty will be allowed to propagate through the system,
potentially leading to costly network investments, which are
highly inefficient with uneven network utilization and would
be prone to additional future uncertainties of asset underuti-
lization. Alternatively, a failure to invest in the system could
compromise the reliability of service to customers and inter-
fere with the integration of flexible low-carbon technologies.
In other sectors, such as manufacturing, when the level of
electricity demand uncertainty becomes excessively large,
the risk of irreversible investment not paying off becomes so
great that it no longer makes sense to invest, but rather wait
until the uncertainty drops below a certain level. In parts
of the power distribution system, the uncertainty in future
generation and demand and their operations is so great that
major capital investment would face a real risk of assets
being underutilized or stranded in the future.
One viable solution to cope with large uncertainties
would be to offer flexible network access to mobilize inher-
ent resources at the customer level, which mitigate short-run
uncertainties and make better use of existing capacity in the
network. As more customers opt for flexible (intermittent
but less expensive) network access, the cost of irreversible
network investment to serve a few inflexible customers will
become excessively expensive, encouraging more customers
to join flexible access. Such a pricing framework would drive
an efficient future-proof network over time, which has less
network investment and relies more on network operation to
reduce congestion and uncertainties.
We highlight the need to establish the link between long-
term customer uncertainty and network investment in network
pricing and identify the opportunities for network custom-
ers and operators to minimize costly investment mistakes.
may/june 2020
We first examine the revolution in Great Britain in network
pricing and new challenges arising from low-carbon transi-
tion, particularly from long-term uncertainties in the energy
landscape, for example, increased unused network capacity
or congestion in parts of the network. We then analyze the
degree of uncertainties from customers to the network and
identify opportunities for managing uncertainties by the net-
work operator and network customers through greater capac-
ity sharing between different types of customers. Finally, we
outline new directions in pricing principles that could relate
network customers’ long-term uncertainty to network invest-
ment, thus rewarding customers who reduce uncertainties and
network operators who have better skills in managing uncer-
tainties. The new pricing principles would not only influence
the behavior of generation and customers in understanding and
managing their uncertainties but, critically, network operators,
to take the necessary risks to fully utilize network resources
and minimize the risk of costly investment. In doing so, the
new principle aims to optimize generation, networks, and
flexible assets for the benefit of energy consumers and better
facilitate the competition between local and distant generation.
History and New Progress
in Network Pricing
Prior to the deregulation of the U.K.’s power industry in the
early 1990s, transmission and distribution networks were
planned in coordination with anticipated generation and
demand, with the key aim of developing infrastructure net-
works in an efficient and timely fashion. This centralized
planning practice has since been replaced by uncoordinated,
profit-driven, decision-making processes among network
owners and users. Its limited coordination makes it extremely
challenging to plan the long-term development of electrical
supply systems. Although generators and demand are free to
choose their locations, when to connect, and how to oper-
ate, these profit-driven planning activities create huge uncer-
tainties in future network planning. Infrastructure network
investment is very expensive and requires a long lead time
to build. In the absence of certainty in future generation and
demand, uninformed decisions could easily lead to under or
overinvestment in the network. In the former case, the power
network may become congested, leading to high operating
costs or, worse, compromise reliability of service. In the lat-
ter case, the network investment may become unnecessarily
excessive when operators expect a much higher volume of
generation and/or demand to connect to the system, leading
to underutilized or stranded assets. In either situation, the end
users, i.e., consumers, have to pick up the extra costs.
Network operators do not have any power to dictate where
new generation and load should be connected or how they
should operate, except for affecting their behaviors through the
use of system charges. The use of network charges are charges
against generators and suppliers for their use of the system
to transport energy over a distance. Through network charges,
network companies are able to recover system investment,
ieee power & energy magazine 27
 operation, and maintenance costs and earn a reasonable rate
of return on the cost expenditure. Another key objective is to
provide forward-looking economic signals that incentivize the
most appropriate locations for new generators and load and the
most suitable operational strategies so that they would encour-
age efficient use of the existing network and cost-effective
development of future system improvements.
There are no uniform economic pricing methodologies
in place in the deregulated power industry. The method-
ologies vary dramatically from one country to another in
terms of sophistication and complexity. These methods
reflect the different characteristics of networks (location,
length, utilization, and connection voltage) and customers
(generation technologies, load density, and load profiles).
Since privatization, Great Britain’s electrical power sys-
tems underwent three major reforms, representing increased
sophistication and efficiencies in the move from aver-
age pricing to cost-reflective network pricing. As listed in
Table 1, Great Britain’s network charging changed over the
last 25 years from reflecting only circuit length to account-
ing for circuit utilization, types of generation technologies,
and demand profiles. Great Britain first introduced loca-
tional pricing to its transmission network in 2003 and its dis-
tribution system in 2006.
During the initial stage of privatization, the use of system
charges for both the transmission and distribution systems
followed the same principle: postage stamp, in which the
main purpose was to recover network cost. At that time, the
use of a simple network pricing methodology allowed for
the electric power industry to focus on developing wholesale
energy markets. The first major reform in network charges
came in 2003, when investment cost-related pricing (ICRP)
was implemented. ICRP considered the need for future net-
work investment based on the marginal increment of genera-
tion and demand at each location of the transmission system.
Under this pricing methodology, for the Great Britain trans-
mission system, in which power is predominantly transferred
from the north to the south, network companies rewarded
new generators located in the south that could reduce the net-
work power flows (i.e., reduce network usage). Conversely,
new generators in the north were heavily penalized for an
increased use of the network. Because distribution networks
were still operated in a passive fashion at that time, distribu-
tion system charges remained the same for all customers.
Since 2002, decarbonization policies have fundamen-
tally changed the energy landscape at the distribution level.
Distributed generators have increasingly connected to the
edge of the supply system. If the postage stamp approach
had continued in distribution network pricing, generation
companies would have had no incentive to locate their gen-
erators closer to load centers with associated reductions in
network costs. Great Britain’s distribution system expected

table 1. Great Britain’s network charges reform timeline. The orange color background denotes the
transition from a postage stamp to cost-reflective pricing.
1990–1992 First Major Reform Second Major Reform
Status Quo (1993) (2006) Third Major Reform (2013)
Transmission Postage stamp ICRP ICRP Improved ICRP
• The same price • Reflects the distance • Differentiates the
and network power must travel generation technologies,
from generation i.e., intermittent renewable
to the distribution generation and fossil
center generation and their
• The same treatment impact on long-term
between generation network development
and demand
Distribution Postage stamp for each DRM (EHV, HV, LRIC/FCP (EHV) LRIC (EHV)
voltage level–DRM and LV) • Reflects the distance CDCM (HV and LV)
(EHV, HV, and LV) power must travel
• The same price and to the customers’
voltage level premises
• Downstream • Reflects the degree of
customers assumed to utilization along the
use all the upstream distribution paths
assets • Treats generation and
demand the same–
CDCM (HV and LV)
• DRM pricing
• Time of use (TOU)
• Treats generation and
demand differently;
generation always
given a credit
DRM: distribution reinforcement model; EHV: extra high voltage; LV: low voltage; MV: medium voltage; LRIC: long-run
incremental cost pricing; FCP: forward cost pricing; CDCM: common distribution charging methodology.

28 ieee power & energy magazine may/june 2020

to receive at least a £5–6 billion investment between 2010
and 2015. Recognizing that flat-rate charging posed a major
barrier to the cost-effective integration of renewables, in
2015, Ofgem called for a major reform in the long-term
charging structure to provide greater incentives to minimize
distribution network development costs from low-carbon
resources. In 2006, long-run incremental cost (LRIC) pric-
ing, the first locational pricing methodology for the distri-
bution system, was developed (see Li and Tolley in the “For
Further Reading” section). Its impact assessment indicated
that a cost savings of £200 million could be achieved over
the next 20 years for electricity consumers if the industry
moved to locational pricing (see Li et al. in the “For Further
Reading” section). By 2007, Western Power Distribution,
the United Kingdom’s largest distribution network opera-
tor, had implemented LRIC pricing to its extra high-volt-
age (EHV) distribution system (from 132 down to 33 kV)
throughout its two distribution areas operated in Southwest
and South Wales. By 2011, all 14 distribution areas across
Great Britain had adopted locational network pricing for
their HV distribution areas.
The subsequent reform of the transmission system in
2013 demonstrated the U.K.’s continued efforts to refine
network charges that address new challenges introduced by
decarbonization and differentiating generation technologies,
including their impacts on long-term network investment.
Because transmission and distribution networks have tradi-
tionally displayed inherently different characteristics, their
reforms followed different developmental pathways prior
to 2017. In 2017, however, Ofgem set up the Future Charg-
ing Forum to perform overall coordination and integration
of network charges across the transmission and distribution
networks, given the rapid changes in Great Britain’s energy
systems introduced by decarbonization and digitalization. In
particular, system changes included the growth of distribution
operations, new energy resources (including both generation
and demand) at all levels of the distribution system, two-way
interactions, and indistinct boundaries between the trans-
mission and distribution systems. In addition, decarboniza-
tion and digitalization also created major uncertainties in the
long-term outlook of the energy system. This fundamentally
challenged traditional charging principles and methodologies,
which reflected deterministic network planning with a certain
future outlook.
An enduring network pricing principle should properly
account for long-term uncertainty in the energy landscape
and thus be able to minimize the risk of under and over net-
work investment. In particular, it should answer the follow-
ing questions:
1) What would be the impact on long-term investment
from customers with different uncertainties?
2) How might these uncertainties propagate through the
distribution and transmission system?
3) How could these uncertainties be managed or reduced
by network operators and network customers?
may/june 2020 ieee power & energy magazine 29
4) What would be the effect on long-term investment as
well as the use of system charges if uncertainty man-
agement strategies were put in place?
Opportunities to Reduce Uncertainties
The large-scale penetration of low-carbon technologies and
flexible demand will bring unprecedented uncertainties to
distribution network planning. Reducing uncertainty is key to
reducing costly mistakes in network asset investments. Char-
acterizing and quantifying these uncertainties across customer
and network levels can provide opportunities for reducing
and managing future uncertainties. Against an increasingly
uncertain future, network users should provide estimated cer-
tainty of their future network access needs, which can reduce
the risk of investment mistakes and alleviate the burden to
network operators of forecasting the state of the system.
Network operators who understand the system would rec-
ognize differences in use by various customers over a range
of periods and identify opportunities for capacity sharing of
the existing network, which would improve network utiliza-
tion and minimize new network investments. This would be
particularly beneficial to renewable generation and flexible
demand, as they do not require full network capacity all of
the time. Through a well-designed sharing system, the exist-
ing network could result in a substantial increase in network
utilization and thus serve a far greater volume of customers
than it does today. This would, in turn, lead to significantly
reduced network costs.
Uncertainties Across Customer Types
The degree of energy uncertainty differs among custom-
ers due to their various lifestyles or nature of businesses
and reflects: 1) temporal uncertainties, the time when
peak demand occurs, and 2) magnitude uncertainties, the
predictability of the demand. This will have an direct
impact on the uncertainty of network peak demand, which
will directly affect network capacity investment and thus
should be a key driver to be factored in network charges.
Uncertainties can be further studied across different cus-
tomer types in terms of their tariff regime and net demand
characteristics [such as owners of electric vehicles (EVs) or
photovoltaics (PVs)].
Figure 1 shows the temporal distributions of the daily peak
for four groups of residential consumers. R1 and R2 consum-
ers share a similar pattern, with consistent evening peaks from
4 to 10 p.m. For groups R3 and R4, the load peaks are ran-
domly distributed throughout the day, which suggests uncer-
tainties associated with PVs and EVs. The uncertainty of R3
consumers could be due to the intermittency of renewable
energy, while that of R4 consumers could be from variations
in charging locations and times.
Figure 2 shows the temporal distributions of daily peak
occurrence for the small- and medium-sized enterprise
(SME) consumers, which appear different from residential
customers. Most of the peaks consistently occur between
 8:00 a.m. and noon, working hours. SMEs are considered
consumers with low uncertainties due to their regular work-
ing hours.
The accuracy of a day-ahead, peak-load forecasting model
can be used to quantify the magnitude of uncertainty. An
autoregressive integrated moving average (ARIMA) model
presented in this article captures the autocorrelations of
the standardized demand. The predictabilities of different
customer types are evaluated through a mean absolute error
(MAE), with a lower MAE indicating a better forecasting
performance. Groups R1, R2, and SME are highly predict-
able, with low uncertainties resulting from regular lifestyles
or consistent working hours. Group R3, consumers with
PVs, is less predictable as a result of the intermittent nature
of renewable energy. Group R4 is the least predictable group
across customer types, which could be due to irregular EV
charging behaviors (see Table 2).
Uncertainties Across Network Voltage Levels
Uncertainties are not evenly distributed across the distribu-
tion network because they vary drastically from one level to
another. At a granular [i.e., low-voltage (LV)] level, demand
is usually more volatile and shows higher uncertainties, while
load variations tend to cancel out when they are more aggre-
gated at HV levels. In addition, different types of customers
demonstrate dissimilar demand patterns due to factors such
as energy usage behaviors, energy prices, and technologies. In
this section, we investigate the uncertainties across three net-
work voltage levels: LV (0.415–11 kV), medium-voltage (MV)
(11–33 kV), and HV (66–132 kV). Both temporal and magni-
tude uncertainties are discussed.
Data were collected from the homes and substations from
selective areas, including South Wales, Ireland, and London.
Figure 3 depicts the distribution of daily peak occurrences for
annual data collected for HV, MV, and LV levels. The tem-
poral distribution across all network voltage levels skewed
toward evening periods, which indicates a coincident peak
time. Nevertheless, the peak time is more scattered as the
voltage level drops. At the LV level, 37% of the peak times
are spread between 4:00 and 10:00 p.m., while 11% of peaks
occurred from 8:00 a.m. to noon, with the rest randomly dis-
tributed. At the HV level, more than half of the sample values
fall within a narrow interval between 4:00 and 8:00 p.m. The
standard deviations (in hours) of the peak time distributions
are 1.13, 1.55, and 2.48 for the HV, MV, and LV levels, respec-
tively, which are consistent with those shown in Figure 3.

0.4 0.15

0.3
0.1
Density

Density

0.2

0.05
0.1

0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
(Hour) (Hour)
(a) (b)
0.3 0.25

0.25 0.2
0.2
0.15
Density

Density

0.15
0.1
0.1

0.05 0.05

0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 0 2 4 6 8 10 12 14 16 18 20 22 24
(Hour) (Hour)
(c) (d)

figure 1. The temporal distribution of residential consumers’ peak demands. (a) R1: The residential customers without stim-
ulus (demand response) or special tariffs. They are billed an existing flat rate and currently the most representative residen-
tial consumers. (b) R2: Consumers billed using TOU tariffs reflecting different unit prices through the day. (c) R3: Consumers
with PVs. (d) R4: Consumers with EVs.

30 ieee power & energy magazine may/june 2020

 Table 3 lists the day-ahead peak load forecast accuracy
across different voltage levels using the ARIMA models. The
mean absolute percentage error (MAPE) is used as the mea-
surement of the forecast accuracy by calculating the relative
scale of the error as a percentage. A lower MAPE indicates a
better forecasting performance. Testing 100 LV substations/
customers and 249 MV substations resulted in MAPEs of
55.78 and 12.83%, respectively. As the voltage level increases
to high, the MAPE drops to 4.21%, indicating that magnitude
uncertainties can be reduced at aggregated levels.
investment, which would lead to high use of network charges.
For customers not requiring high certainty in network access,
their needs may be better served via intermittent and less-
expensive network access, which could be supplemented by
local customer resources and/or network energy storage.
0.4
0.3

Potential Pricing Schemes That

Density
0.2
Reduce Future Uncertainties
and Risk Investment
0.1
There is a huge diversity in energy usage across a wide cus-
tomer base. Uncertainties vary substantially among customer
types, both in terms of the magnitude and time of peak use. 0
Likewise, uncertainties across different voltage levels vary 0 2 4 6 8 10 12 14 16 18 20 22 24
significantly, with lower voltage levels having much higher (Hour)
uncertainties. However, traditional network pricing meth- (a)
odologies do not distinguish between load uncertainties and 0.5
future projections for different loads, which could lead to
unnecessary network investments, as is currently experienced 0.4
in parts of Great Britain’s network. As a result, the new round
0.3
of reforms in network charges that began in 2019 in the United
Density

Kingdom focuses on achieving a greater efficiency in existing

0.2
network use by reducing future uncertainties and managing
said uncertainties through network sharing. 0.1
Reducing uncertainties can be achieved by improving
the definition of network access for customers (see Table 4). 0
For customers requiring high certainty in network access, 0 2 4 6 8 10 12 14 16 18 20 22 24
their needs may still be served by traditional costly network (Hour)
(b)
0.4
0.4

0.3
0.3
Density

0.2
Density

0.2

0.1
0.1

0
0
0 2 4 6 8 10 12 14 16 18 20 22 24
0 2 4 6 8 10 12 14 16 18 20 22 24 (Hour)
(Hour) (c)

figure 2. The temporal distribution of SME consumers’ figure 3. The peak demand temporal distribution at (a) HV,
peak demand. (b) MV, and (c) LV levels.

table 2. The forecasting accuracy for different table 3. The forecasting accuracy across different
customer types. voltage levels.
Customer Type R1 R2 R3 R4 SME Voltage Level LV MV HV
MAE 0.0608 0.0602 0.1178 0.2624 0.0673 MAPE (%) 55.78 12.83 4.21

may/june 2020 ieee power & energy magazine 31

 table 4. A comparison of the different levels of certainty in network access required by customers.
Certainty in Customer
Network Access Network Access Offered Rate of Return to
Requirements by Network Operators Network Activity User System Charges Distribution Owner
High Firm Traditional asset High Low
investment
Low Intermittent Short-run flexible Low High
operation

One approach to managing uncertainty could be by offer-
ing flexible network access, recognizing that future distri-
bution network operators would have two core businesses:
traditional long-term network investment and short-term
flexible operation that mobilize inherent resources at the
customer and network levels to manage uncertainties. For
the latter, it is important to understand that resources have
finite capacity over limited time periods; therefore, network
operators could be rewarded if they are willing to take risks
and develop greater capability in understanding and miti-
gating uncertainties through short-term operational intelli-
gence, thus minimizing investments in new assets.
As more customers opt for less-expensive intermittent net-
work access, there could be less of a need for customers to
request firm capacity, and the cost of accessing firm capacity to
serve fewer customers would become excessively expensive. In
such a situation, local energy provisions may become competi-
tive when compared to their distant counterparts. This pricing
framework would drive an efficient future-proof network over
time, which would have less network investment and rely more
on network operations that reduce congestion and uncertainties.
Such networks would provide the capacity to deal with short-
term uncertainties rather than a network dominated by longer-
term plans for network asset additions; however, regulators
would need to consider social justice issues arising from users
requiring set electrical service, such as the elderly, the disabled,
and caregivers.
Summary
The degree of sophistication in network pricing has improved
dramatically in the last 30 years, moving away from pure
cost allocation to cost-reflective pricing, indicating both the
­characteristics of the network and its users to improve overall
network utilization. Decarbonization and digitalization have
fundamentally changed the energy landscape and created
major uncertainties in customer demand and network use as
well as the future outlook. These changes challenge the present
principles and charging methodologies of electric rates, which
do not adequately consider uncertainties in network customer
use. Many types of regulatory regimes can address this issue.
In this article, we quantified the diversity in network and
customer uncertainty, identified opportunities for mitigat-
ing uncertainty through sharing network use and customer
demand response and generator assets, and proposed future
directions for network pricing that offer flexible network
access to mitigate uncertainties. In doing so, the risk of
making mistakes in costly investment can be substantially
reduced, and network operators can be rewarded for their
improved ability to mitigate uncertainties.
For Further Reading
“Digest of United Kingdom Energy Statistics (DUKES)
2016: Main chapters and annexes,” Department for Business,
Energy and Industrial Strategy, U.K. Government, London,
Sept. 9, 2016. [Online]. Available: https://www.gov.uk/
government/statistics/digest-of-united-kingdom-energy
-statistics-dukes-2016-main-chapters-and-annexes
F. Li et al., “Distribution pricing: Are we ready for the
smart grid?,” IEEE Power Energy Mag., vol. 13, no. 4, pp.
76–86, July 2015. doi: 10.1109/MPE.2015.2416112.
“Structure of electricity distribution charges: Consulta-
tion on the longer term charging framework,” Ofgem, 2005.
[Online]. Available: https://www.ofgem.gov.uk/publications
-and-updates/structure-electricity-distribution-charges
-consultation-longer-term-charging-framework
F. Li and D. L. Tolley, “Long-run incremental cost pricing
based on unused capacity,” IEEE Trans. Power Syst., vol. 22,
no. 4, pp. 1683–1689, 2007. doi: 10.1109/TPWRS.2007.908469.
F. Li, D. Tolley, N. P. Padhy, and J. Wang, “Framework
for assessing the economic efficiencies of long-run network
pricing models,” IEEE Trans. Power Syst., vol. 24, no. 4, pp.
1641–1648, 2009. doi: 10.1109/TPWRS.2009.2030283.
“Delivering the electricity distribution structure of charg-
es project,” Ofgem, 2008. [Online]. Available: https://www
.ofgem.gov.uk/publications-and-updates/delivering-electricity
-distribution-structure-charges-project
“Project TransmiT,” Ofgem, 2011. [Online]. Available:
https://www.ofgem.gov.uk/electricity/transmission-networks/
charging/project-transmit
“Charging futures forum,” Ofgem, 2017. [Online]. Avail-
able: https://www.ofgem.gov.uk/publications-and-updates/
charging-futures-forum
Biographies
Furong Li is with the University of Bath, United Kingdom.
Chenghong Gu is with the University of Bath, United
Kingdom.
Ran Li is with the University of Bath, United Kingdom.

p&e

32 ieee power & energy magazine may/june 2020

Distribution
Network Rate
Making in Latin
America An Evolving Landscape

F
F O L L OW I N G THE TREND OB-
served in developed economies, vari-
ous Latin American governments are
committed to reducing greenhouse gas
emissions, particularly in the power
sector. In countries such as Chile, Peru,
Colombia, Brazil, and Mexico, various
regulatory policies have been issued
to meet renewable-generation integra-
tion targets and satisfy the increasing
demand from consumers for supply
quality. Meanwhile, the integration of
distributed generation (DG) in rural
and urban areas as well as the increas-
ing need to integrate electric vehicles
(EVs) in urban areas are driving impor-
tant reforms in the distribution sector.
Distribution networks are expected
to increase reliability and integrate
various distributed energy resources
(DERs) by using an array of newly
available technologies (including smart
meters, online monitoring and control,
©ISTOCKPHOTO.COM/LEOLINTANG

By Rodrigo Moreno, Bernardo Bezerra,

Hugh Rudnick, Carlos Suazo-Martinez, Martha Carvalho,
Alejandro Navarro, Carlos Silva, and Goran Strbac
Digital Object Identifier 10.1109/MPE.2020.2972667
Date of current version: 17 April 2020

may/june 2020 1540-7977/20©2020IEEE ieee power & energy magazine 33

 Regulators need to reshape frameworks that have
historically focused on low costs (rather than the
best value for money).
and other smart grid and IT advances), enabling utilities to
implement an active (rather than the historical passive or
“fit-and-forget”) approach to network operation, advanced
operational measures, and nonwire solutions that will effec-
tively displace costly investments. This environment poses
challenges, but there are opportunities too. Understanding
those opportunities and designing regulations accordingly is
a priority in Latin American countries, where maintaining
affordable energy is essential.
To ease the transition toward a more reliable and modern
electricity network, regulators need to reshape frameworks
that have historically focused on low costs (rather than the
best value for money) and expand electricity coverage and
access. Regulators in countries including Colombia, Peru,
Chile, Brazil, and Argentina have implemented or are under-
taking regulatory reforms in the distribution sector. The most
important topics of debate (relevant to rate making) include
enhancing reliability for consumers, connecting DERs, and
providing secure and flexible services between distribution
and transmission networks within an incentive- or perfor-
mance-based price-control regime.
Well-designed incentive- and performance-based regula-
tions use financial enticements to encourage distribution net-
work owners to provide their (multiple) services securely and
cost-effectively. Apart from encouraging network owners to
be efficient, network users (that is, producers, consumers,
and prosumers) should be motivated to improve their opera-
tional and deployment profiles. This is particularly relevant
Quality and
Reliability Decarbonization
Economic
Efficiency and
Equity and
Revenue
Fairness
Adequacy
figure 1. The tetralemma for distribution network regula-
tion in Latin America.
34 ieee power & energy magazine may/june 2020
in the context of DERs, and it poses challenges related to the
deployed meter technology and the fact that tariffs, to which
network users react, present a limited level of spatiotem-
poral granularity and cost reflectiveness. Other important
concerns include equity and fairness in tariff and reliability
levels, especially in rural and other underdeveloped areas.
Thus, regulators consider various policy concerns and objec-
tives when deciding on network regulations, which are sum-
marized in Figure 1. In the rest of this article, we focus on
distribution network price controls, especially remuneration
and tariffs. We summarize rate making in Latin America
countries and discuss the Chilean and Brazilian experiences.
Rate Making in Latin America
Overview
Determining the regulated revenues for a distribution net-
work company is a complex task with multiple conflicting
objectives as well as determining rates, charges, or tariffs
that the company is allowed to impose for billing and col-
lecting fees from consumers. The design of such processes
should consider that, in broad terms, remuneration will affect
network owners’ decisions (investments, operation, mainte-
nance, administration, and so forth), while tariffs will impact
network users (the locations they choose, their consump-
tion and production profiles, and so on). While traditional
objectives for appropriate remuneration center on revenue
adequacy [to ensure that income is adequate to cover capital
expenses (CAPEX) and operating expenses (OPEX), includ-
ing a reasonable return on capital] and productive efficiency
(for the regulated service to be provided at the most efficient
cost), the objectives for tariff design focus on allocative effi-
ciency (guaranteeing that effective charge signals are sent
so that users can make cost-effective decisions about their
consumption and production) and equity (ensuring that net-
work tariffs feature desirable distributional characteristics,
especially in terms of protecting vulnerable consumers). To
achieve this, there are numerous regulations that emphasize/
weigh conflicting objectives differently.
Latin America employs cost-of-service and incentive- or
performance-based regulations (with a preference for the
latter) to determine allowed revenues. In countries using
the cost-of-service approach, such as Ecuador and Costa
Rica, planners determine future network investments, and
regulators ensure that a company’s costs are fully recov-
ered through appropriate charges. In countries applying
­incentive- or performance-based approaches, such as Chile,
Brazil, Argentina, Colombia, and Peru, various regulations
are used to compel network owners to make appropriate
investment and operational decisions. In those countries,
network tariffs are usually set ahead of time, ex ante, decou-
pling future real costs from regulated future revenues to
originate incentives for companies to minimize costs. This
is because revenues will be maintained at the permitted level
for a given period of time, regardless of the company’s actual
costs. From time to time, however, ratchets will be applied,
where the permitted revenue and associated tariffs are reset
to 1) pass the achieved cost savings to network users and
2)  ensure, despite information asymmetries, that revenues
and costs are not unreasonably different.
The main features of price-control regulations in Latin
America are shown in Table 1, including whether the
mechanisms are cost-of-service or incentive based, the refer-
ential rates of return commonly used (note that under incen-
tive-based regulation, real rates of return are endogenous
and dependent on the performance of the companies), and
the duration of the control periods. We also show whether
the ownership of utilities is private, public, or mixed.
The region also features an array of tariff structures (that
is, rules to determine how to charge different consumers) that
vary across countries and user classes (such as residential,
commercial, industrial, public lighting, and so forth). These
tariffs present fixed (per user) and variable components [in
U.S. dollars per kilowatt-hour (kWh) and U.S. dollars per
kilowatt (kW)] as well as different spatiotemporal-granularity
levels (in time, including peak and off-peak hours, winter
and summer, and day and night, and in space, such as per
company, per municipality, and per voltage level), and can be
based on average or marginal (or incremental) network-cost
principles. Regarding tariff components in the electricity bill,
the energy cost is generally passed through to consumers on
a dollar ($)/kWh basis, while the infrastructure cost (poten-
tially including generation capacity and networks) is charged
on by $/kW. This, however, depends on the country and user
class. For residential consumers, for example, the combined
cost of generation, transmission, and distribution is packed
into a single $/kWh rate (which may change in time and loca-
tion, albeit to a very limited degree).
Another tariff-related practice concerns the different
treatment of transmission- and distribution network costs,
which in Chile are included in the $/kWh and the $/kW
rate, respectively (for an industrial consumer). Moreover,
networks, depending on whether they belong to the trans-
mission or distribution sector, can present completely differ-
ent remuneration and pricing frameworks. For transmission,
locational marginal prices (LMPs) can be observed in coun-
tries such as Chile, Peru, and Mexico, in which energy and
part of the network costs are recovered via nodal prices that
change with location and time according to real-time sys-
tem conditions. This regime, however, cannot be observed in
distribution networks in any country in Latin America (and
probably the world), although the effects of LMPs on distri-
bution networks are starting to be investigated in the aca-
demic literature, especially in the context of DERs. Table 2
presents three countries’ tariff structures for residential and
industrial consumers and their maximum temporal-resolu-
tion levels for residential and industrial loads.
The Good, the Bad, and the Ugly
Although Latin America has increasing levels of electrification,
as shown in Figure 2, the power supply’s reliability remains

table 1. Price-control regulation in Latin America.

Costa Dominican
Argentina Bolivia Brazil Chile Colombia Rica Republic
Utility Mixed Mixed Mixed Private Mixed Public Mixed
ownership
Price control Incentive Incentive Incentive Incentive Incentive Cost Incentive
Referential rate 8.04% According to 8.09% posttax 10% pretax 11.8–12.4% 4.24 9.02%
of return posttax public utility (benchmark (model firm) pretax posttax pretax
(Edesur) companies company)
included in the
Dow Jones
Control period 4 years 4 years 4 years 4 years 4 years 1 year 4 years
Ecuador El Salvador Guatemala Honduras Panama Peru
Utility Mixed Private Mixed Mixed Mixed Mixed
Ownership
Price control Cost Incentive Incentive Incentive Incentive Incentive
Referential rate No public 10% pretax 6.88% posttax 7–10% pretax 8.94% 12% pretax
of return information posttax
Control period 1 year 5 years 5 years 5 years 4 years 4 years

may/june 2020 ieee power & energy magazine 35

 poor. Even in Chile, which has graduated from the list of devel- Interruption Duration Index (SAIDI), is roughly 15 h, which
oping countries and is part of the Organization for Economic is approximately 15 times higher than in the United Kingdom.
Cooperation and Development, the average outage duration Importantly, this lack of reliability is one of the negative
per year for each customer, measured by the System Average consequences of regulation. In Chile, distribution network

105 90
80
100 70

SAIDI (h/year)
Access (%)

60
95 50
90 40
30
85 20
10
80 0
Chile
Cuba
Uruguay
Brazil
Argentina
Venezuela
Paraguay
Costa Rica
Mexico
Ecuador
Colombia
Dominican Republic
El Salvador
Panama
Peru
Guatemala
Bolivia
Nicaragua
Honduras

Cuba

Brazil

Argentina

Chile

Panama

El Salvador

Colombia

Peru

Costa Rica

Dominican Republic
(a) (b)

figure 2. The (a) electrification and (b) reliability levels in Latin America.

table 2. The end consumer’s bills and their breakdown in three Latin American countries (in US$). All are low voltage.
Santiago, Chile Quito,* Ecuador Rio de Janeiro, Brazil
Sector Unit Residential Industrial Residential Industrial Residential Industrial
Generation US$/month — — — — — —
US$/kWh 0.1 0.08 0.03 0.03 0.06 0.06
US$/kW/month — 6.99 — — — —
Transmission US$/month — — — — — —
US$/kWh 0.02 0.02 0.01 0.01 0.03 0.03
US$/kW/month — — — — — —
Distribution US$/month 0.85 0.99 1.41 — — —
(including
metering US$/kWh 0.02 — 0.05 0.03 0.03 0.03
and billing)
US$/kW/month — 9.2 — 4.18 — —
Others US$/month — — — — — —
US$/kWh 0.001 0.001 0.01 0.01 0.03 0.03
US$/kW/month — — — — — —
Total US$/month 0.85 0.99 1.41 1.41 — —
US$/kWh 0.14 0.1 0.1 0.08 0.16 0.16
US$/kW/month — 16.2 — 4.18 — —
Time kWh rate kWh and No time kWh rate No time No time
resolution with up to kW rates variations with up to variations variations
(maximum) three levels with three levels
in one day no time in one day
variations
*In Ecuador, there is a government subsidy of roughly US$0.05/kWh that reduces the overall bill to the values shown here.

36 ieee power & energy magazine may/june 2020

 Policy makers are pushing regulators to include
more sophisticated objectives in remuneration
and tariff regimes.

price caps have been limited to relatively small values, low- ✔✔ innovation and network modernization, evolving to-
ering costs significantly in comparison with other devel- ward the smart grid
oped countries. However, as illustrated in Figure 3, lower ✔✔ decarbonization and renewables.
costs have an impact on service quality because of tradeoffs These objectives were initially de-emphasized by regulators to
between network performance and expenses. In general, prioritize electrification, coverage, and access at a reasonable
low-cost networks tend to be less reliable. Understanding cost (notice that access is still a problem in countries such as
this cost-performance balance is critical in developing coun- Bolivia, Nicaragua, and Honduras, as indicated in Figure 2).
tries because higher reliability levels might require funding Remuneration should evolve, in our opinion, to enable dis-
increased levels of network investment and, hence, higher tribution networks to innovate and become more active, such
electricity bills. as by operating with smart grid technologies. This would
After nearly 40 years of distribution network rate mak- improve the optimization of CAPEX and OPEX in network
ing in Latin America, the quality of service is becoming companies, driving new operational measures and nonwire
more critical. In addition, policy makers are pushing regu- solutions to deliver reliability and other services more effec-
lators to include more sophisticated objectives in remunera- tively. With the increasing penetration of DERs, price sig-
tion and tariff regimes (apart from the historical ones focus- nals will become more relevant for them to be adequately
ing mainly on cost reductions). The goals include deployed and operated, potentially providing services to
✔✔ quality of service, even beyond reliability and voltage distribution and transmission networks. Increased levels of
quality, including resiliency against hazards and cus- coordination will be needed at the interface between trans-
tomer satisfaction mission and distribution systems. It is envisioned that net-
✔✔ equity and fairness in tariffs and quality work operators will be able to trade various security and

0.09 AT Austria
LU PT
Electricity Bill (Distribution Network Cost) (US$/kWh)

CL Chile
0.08 DE CZ Czech Republic
SE DE Germany
0.07 DK Denmark
NL EE Estonia
UK IE ES Spain
0.06 FR EE FR France
DK AT CZ
SI GR Greece
0.05 ES HR Croatia
HU HU Hungary
0.04 HR IE Ireland
CL IT Italy
LT LV RO LT Lithuania
0.03 IT GR LU Luxembourg
MT LV Latvia
0.02 MT Malta
NL The Netherlands
0.01 PT Portugal
RO Romania
SE Sweden
0
0 100 200 300 400 500 600 SI Slovenia
SAIDI Distribution Network (min) UK United Kingdom

figure 3. Distribution network costs in capital cities versus SAIDI in all distribution networks. Due to the lower voltage
levels in Chile’s distribution networks (up to 23 kV), we have considered part of the subtransmission systems (which cover
up to 110 kV) as distribution networks to properly compare values among countries.

may/june 2020 ieee power & energy magazine 37

 flexibility services between regimes to achieve, through decen-
tralized operation of various distribution systems and the
transmission grid, optimal solutions systemwide.
These considerations are critical in countries where costs
must be kept as low as possible while improving quality and
addressing other objectives of the new energy policy agenda.
To achieve those goals, new regulations must remunerate
and charge for distribution networks. Next, we present rel-
evant regulatory discussions from Chile and Brazil, whose
authorities are conducting rate making reforms. We focus
specifically on remuneration and tariffs.
Chile
The New Policy Landscape
In Chile, as in the rest of the world, new energy policies are
evident. Today’s agenda includes themes that go beyond the
historical debates about limiting costs and improving access
to electricity systems. Chile’s energy policy has been reason-
ably successful in regulating network development, access,
and costs (as shown in the previous sections) and progressed
to incorporate two major objectives: 1) improve poor reli-
ability by evolving toward a more resilient system and 2)
decarbonize the energy sector by integrating renewables.
There are other more specific objectives. In the context of
distribution networks, they include the following:
✔✔ more DG and DERs
✔✔ an increased role for demand-side response and man-
agement
✔✔ including more innovative smart grid solutions
✔✔ increasing participation from the transport and heat-
ing/cooling sectors
✔✔ more buildings and communities that have their own
generation, storage, energy-management systems and
controls (such as microgrids)
✔✔ growing the number of initiatives to foster smart cities.
To address these objectives, the government began to change
its approach to distribution network regulation by supporting
the transition toward a more reliable, resilient, affordable,
and sustainable energy sector. We now summarize the main
aspects of the debate in Chile, focusing on remuneration.
Aspects associated with tariffs will be discussed for Brazil.
The Remuneration Approach
The Philosophy
Distribution network remuneration in Chile, unchanged since
1982, is based on the idea of benchmarking the cost perfor-
mance of real firms against model ones. The so-called model
firm is a theoretical, virtual company created to ­calculate
remuneration. It is optimally designed in a greenfield fashion
at the beginning of each control period (such as every four
years) to provide distribution network services in the same
area as the corresponding real company. Because model
firms are designed in a greenfield fashion, their infrastruc-
38 ieee power & energy magazine may/june 2020
tures are valued at current market prices by using the replace-
ment-cost concept. Once model firms have been constructed
and valued, the final network tariffs to be applied to the real
companies are calculated so that the model firms feature a
10% cost of capital (before taxes and assuming there is no
leverage). The main idea behind model firms is to incentiv-
ize real companies to be economically efficient, since their
revenues and costs are decoupled. Under this philosophy, the
cost functions of real companies do not affect tariffs. The
method simulates competition between a regulated incum-
bent monopolistic firm and a virtual one that was cost-effec-
tively designed from scratch using the latest technologies.
Practical Implementation
In practice, model firms are determined by using character-
istics from reality that reflect exogenous variables that are
beyond the control of distribution companies, including
✔✔ locations and capacities of transmission-distribution
entry points and primary substations
✔✔ locations and loads of demand points (including the
current consumption and what is projected for the next
15 years)
✔✔ city road maps
✔✔ other constraints imposed by security standards and
municipalities (for example, the use of underground
rather than overhead lines).
Although today’s model firms are calculated based on
advanced computational mathematical models, during the
1980s and 1990s they were determined without significant
support from computers. The complexity associated with the
escalating number of companies to be regulated was sim-
plified by using a sample of representative firms selected
through clustering methods. More than 30 distribution com-
panies were divided into six clusters according to their den-
sity metrics (companies with similar $/km–kW belonged to
the same cluster). One firm per cluster was chosen as the
basis for the model-firm remuneration process. Six model
firms were determined, and each real company adopted the
tariffs for its cluster.
For each representative firm, two model firms are deter-
mined through simultaneous studies, one undertaken by the
regulatory authority and the other by the company. Hence,
the final cost valuation (including the CAPEX and OPEX)
of every representative firm is equal to the weighted average
valuation of the studies: two-thirds and one-third for each
authority’s and firm’s study, respectively. This mechanism
avoids long negotiations between the authority and each
company because each study is developed in isolation, and
the company and authority do not have to agree on input
parameters and results to determine their corresponding
model firms.
To account for the possibility that the costs of real com-
panies and model firms differ, a check is undertaken after
the tariffs have been determined to verify that the aggre-
gated profitability of the nationwide distribution sector is
between 6 and 14%. At the company level, however, profits
are allowed to rise above that range. To calculate profits, real
companies’ expenses are determined by considering their
actual infrastructure (not that of model firms) and valued at
current market prices by using the replacement-cost concept.
Arguments For and Against
Current Remuneration Practices
There are advantages and disadvantages to the model-firm
approach applied in Chile. Three of the chief benefits are 1)
firms’ perceived incentives to limit costs, 2) simplicity and
low-cost implementation, and 3) the ability to deal pragmati-
cally with significant information asymmetries between the
regulator and firms since there is no reliance on real com-
panies’ expenses. The approach was exported to other coun-
tries in Latin America, such as Argentina, Peru, and Bolivia,
and to Central America, demonstrating its attractiveness and
practicality at a time when limiting costs (while improving
coverage/access) for recently privatized companies was key.
However, experts pose several questions about the method,
since there is no evidence that it will deliver the objectives of
the present energy-policy agenda.
Concerns Regarding Supply Security
As mentioned, one of the primary objectives of today’s
energy policy is to enhance the supply security and resil-
iency. Chile wants to reach a 1-h SAIDI by 2050 (uni-
formly distributed across the country), with an interim
goal of 4 h in 2035. That would be a significant improve-
ment from the current SAIDI, which averages 15 h at the
national level and can climb higher than 50 h in regions
such as Tarapaca, Atacama, and Araucania. Table 3 gives
the geographical distribution of the SAIDI, which can
change significantly by year depending on exogenous fac-
tors, such as natural hazards.
There is significant debate regarding how to improve
the remuneration method to deliver efficient investments
to ensure a more secure, reliable, and resilient future. As
with any other incentive-based, network-remuneration
method, the model-firm approach needs additional incen-
tives (such as penalties and rewards) to discourage reli-
ability degradation, since delivering adequate reliability
usually requires spending more, which is clearly discouraged
by an approach that aims to reduce expenses. There are
security standards in Chile that establish limits for various
reliability metrics, including the SAIDI, with associated
penalties for distribution companies that fail to deliver.
However, they are unlikely to be sufficient because reli-
ability improvement requires network investments. Penal-
ties will not encourage companies to make investments
when there is insufficient funding or enough certainty
about future revenue streams.

table 3. The SAIDI in hours per region and per year in Chile.
Seven-Year
Region Population 2012 2013 2014 2015 2016 2017 2018 Average
Aysen 103,158 23.1 27.9 26.1 29.9 19.7 31.2 14 24.6
Magallanes y Antartica 166,533 6.2 8.8 8.2 9.2 5.1 6 6.9 7.2
Chilena
Arica y Parinacota 226,068 21 14.3 33.9 12.3 10.7 15.3 23.2 18.6
Atacama 286,168 25.9 19.5 22.8 53.6 11.1 22.6 16.4 24.6
Tarapaca 330,558 29.8 24.4 59.9 23.8 20.2 18.1 14.6 27.3
Los Rios 384,837 27.9 25.6 25.1 26.6 22.3 24.7 19.5 24.5
Antofagasta 607,534 18.6 14.9 25.2 22.9 15.9 16.3 11.7 17.9
Coquimbo 757,586 10 11.5 9.8 44 11.5 10.5 10.1 15.3
Los Lagos 828,708 30.1 24.2 25.9 23.9 18.4 22.3 17.2 23.2
O’Higgins 914,555 16.6 18.2 18 20.4 17.9 23.2 16.8 18.7
La Araucania 957,224 34.1 34.6 30.7 32.3 31.5 51 28.3 34.6
Maule 1,044,950 20.1 14.1 16.9 26 20.8 33.1 14.7 20.8
Valparaiso 1,815,902 12.4 9.3 10.1 17.2 9.4 10 7.1 10.8
Biobio y Nuble 2,037,414 28.6 19.3 20.6 19.3 16.9 20.5 13.2 19.8
Metropolitana de Santiago 7,112,808 8.9 7.7 8.4 8.8 8.2 13.6 8.5 9.2
Country level 17,574,003 16.7 13.9 15.7 18.1 13.4 18.6 12.1 15.5
Exceptional events are included. Values higher than 50 h are in red. Regions are sorted according to population.

may/june 2020 ieee power & energy magazine 39

 There is no easy solution to this problem within the
present paradigm because the model-firm philosophy is to
remunerate a virtual company. Built in a greenfield fashion,
model firms cannot properly capture the need for additional
funding to improve reliability in real companies. The mis-
match between the remuneration needed to cover the cost
of a greenfield model firm and that of a hypothetically effi-
cient real firm with legacy concerns can be mathematically
proved. In mathematical programming terms, the solution
to the optimization problem that determines the model
firm’s investments and assets from scratch will differ from
the solution to the optimization problem that determines
the efficient investments and assets of a firm subject to past
decisions. Although the mismatch is self-evident (given the
different building approaches of the real and model firms),
it must be emphasized since it has profound impacts on
remuneration adequacy.
Reliability, Fairness, and Affordability in Rural Networks
In the discussion of supply security and remuneration,
another important aspect arises: fairness and affordabil-
ity in rural networks, since the current system is ill-suited
to remote areas. It is well known that delivering reliable
power to consumers in rural areas is more costly than in
urban areas (for example, Frontel versus Enel in Figure 4)
because the number of connected customers per kilometer
in cities is significantly higher than in provincial locations.
Consequently, a cost-benefit analysis to balance investments
against their reliability gains will justify worse reliability
levels in rural areas and higher network tariffs. This techno-
economic result is fundamentally problematic from a public
policy perspective. This is particularly relevant in light of
the Chilean energy policy that seeks uniformly distributed
140
From Rural to Urban Networks
120
Density (Connections/km)
100
SAIDI (h/year)
80
60
40
20
0 0
Frontel SAESA CGED Chilquinta
Utility
Density SAIDI
Linear (Density) Linear (SAIDI)
figure 4. The density, residential electricity bill, and SAIDI for the main distribution
companies in Chile.
40 ieee power & energy magazine may/june 2020
SAIDI across the country (that is, reliability fairness) at
affordable costs.
Incurring higher costs to improve reliability in rural
areas is no trivial matter, since provincial consumers, who
already pay higher electricity bills (see Figure 4), are likely
to have lower incomes. Equalizing reliability levels across
a country may significantly raise network tariffs beyond
affordable levels in rural areas. That would be the case if
tariffs sought to be cost reflective (note that cost reflectiv-
ity is a key techno-economic principle and an important
objective in tariff design, especially in the future context
of DERs). Hence, cross subsidies between areas might be
needed to achieve reliability fairness at affordable costs for
rural consumers.
Finding the right balance between conflicting objectives
(in this case, cost reflectivity and fairness) and resolving
the political-economy conundrum becomes a complex task.
There have been efforts in Chile to develop a concept called
equidad tarifaria (tariff equity), which internalizes the cross
subsidies within tariffs, tending to equalize them. Ques-
tions regarding the efficiency of the current mechanism have
been posed, since the issue of balance can be alternatively
resolved (and potentially more effectively) through subsi-
dies in the form of separate payments that target consumers
who really need the financial support and preserve the origi-
nal cost-reflectivity levels in network tariffs. In light of the
escalating need for reliability and fairness, this debate will
become increasingly important during the coming years.
Concerns Regarding Decarbonization
Through Grid Modernization
Effectively managing a distribution network with increased
DER penetration and participation in new flexibility- and
ancillary-services markets at the
transmission level requires an
active (rather than the historical
0.3
passive) approach to its opera-
0.25 tion, which necessitates changes
in the way system infrastructure
Total Bill (US$/kWh)
0.2 is planned. Chile has an increas-
ing number of DG projects and EV
0.15 infrastructure, as shown in Fig-
ure  5. In the future, we expect to
0.1 see a larger array of dispatchable
resources at medium- and low-
0.05
voltage levels, including demand
response, energy storage (for in­­
Enel stance, batteries, EVs, and thermal
demand), DG [primarily photo-
voltaic (PV)], and equipment that
Bill
will efficiently adapt the network
Linear (Bill)
to changing operating conditions
on a minute-by-minute basis. This
will require distribution innovation
to promote cost-effective solutions
throughout the system level and during the long term. CAPEX
and OPEX tradeoffs will be critical for minimizing the total
expenditures of distribution companies that increasingly rely
on operational measures and nonwire solutions, following the
smart grid concept.
Modernizing distribution networks will require stra-
tegic short-term investments that must be remunerated.
Future systems will extend beyond the conventional network
infrastructure. For example, following ideas implemented
in other jurisdictions, distribution systems are expected to
become “market platforms,” necessitating significant IT
system enhancement (including communication, moni-
toring, processing, control, and so on). Incentivizing strate-
gic investments for modernizing electricity grids is a major
challenge for regulators. Under the model-firm paradigm,
where remuneration is fixed for four years, and revenue
streams beyond that point remain uncertain (and poorly cor-
related with real costs), it is difficult to foster strategic invest-
ments to reduce longer-term costs and remunerate advanced
and innovative technical solutions. Without the certainty

800

700
Number of EVs and Plug-In EVs

600

500

400

300

200

100

0
January
April
July
October
January
April
July
October
January
April
July
October
January
April
July
October
January
April
July
October
January
April
July
October
January
April
July
October
January
April
July
October
2012 2013 2014 2015 2016 2017 2018 2019

(a)
6,000
Number of PV Rooftop Installations

5,000

4,000

3,000

2,000

1,000

0
January
March
May
July
September
November
January
March
May
July
September
November
January
March
May
July
September
November
January
March
May
July
September
November
January
March
May
July
September
November

2015 2016 2017 2018 2019

(b)

figure 5. The number of (a) EVs and (b) DER connections.

may/june 2020 ieee power & energy magazine 41

 Recognizing uncertainty in the network-expansion-planning
problem increases the value of smart grid technologies, which
network planners and regulators must acknowledge.
that their investments will be adequately remunerated, com-
panies are likely to prefer conventional solutions that may
be less expensive during the short term but compromise the
distribution system’s economic performance (and the entire
electricity system) through the longer term.
The debate about modernization and innovation includes
the uncertainty related to planning network expansions.
This is becoming increasingly important due to the many
factors that must be considered, including the rapidly evolv-
ing costs of technologies, market prices, policies, and DER
deployment. The need to find technical solutions to network
uncertainty becomes a challenge, and regulators must prop-
erly scrutinize and incentivize proposals. With uncertainty,
innovative solutions (such as battery energy systems and
open soft points, which apply power-electronic devices to
normally open portions of the distribution system) become
more attractive as a means to more easily adapt to, cost-
effectively deal with, and hedge against a number of sce-
narios that might happen in the future. In other words, rec-
ognizing uncertainty in the network-expansion-planning
problem increases the value of smart grid technologies,
which network planners and regulators must acknowledge.
The second important aspect of planning under uncer-
tainty relates to regulatory scrutiny. It is important to rec-
ognize that an optimal-investment decision that was made
under uncertainty cannot be evaluated and justified ex post
facto, when complete information is available. From a sto-
chastic-programming viewpoint, it is clear that an ex ante
decision made under uncertainty may seem suboptimal at a
later date. This argument has special relevance for incentive-
based remuneration methods that face after-the-fact regu-
latory scrutiny and correction of the regulatory asset base
(RAB), where companies face the risk of a backlash if invest-
ments are deemed unjustified. This is also important in Chile,
where real companies’ infrastructure, in every tariff period,
is valued through a greenfield model firm that possesses per-
fect information, an approach that disregards legacies built
under uncertainty and the optimal infrastructure needed to
face future unknowns. Incorporating tools to deal with uncer-
tainty in network regulation is becoming a must.
Other Concerns About Remuneration
There are other important concerns that arise in the debate
surrounding the model-firm approach and are common to
other remuneration methods. Among them the following:
✔✔ Model firms minimize costs within the distribution
sector, in isolation from the rest of the power system.
42 ieee power & energy magazine may/june 2020
This relates to the discussion about the interface be-
tween transmission and distribution and the need to
have a whole-system view of the planning and regula-
tory problem.
✔✔ The set of prices used to calculate the model firms’ costs
follows the concept of the replacement cost, valuing
assets as new according to current market prices. This
can cause fluctuations in a company’s valuation and
produce higher risks for investors.
✔✔ The real costs of capital to investors remain hidden
during the regulatory process, since companies’
expenses are valued by the replacement cost. Most
importantly, this blindness includes the return that in-
vestors will ultimately receive and is a failure of mo-
nopoly regulation.
✔ ✔ A referential cost of capital equal to 10% (before
taxes and assuming that there is no leverage) that is
fixed by law (for model firms) may not appropriately
reflect 1) market conditions at the time of the control
and 2) real firms’ risks, including those imposed by
the remuneration approach. This problem should be
fixed by calculating appropriate costs of capital every
time tariffs are determined.
✔✔ Within the control period, there is no differentiation
between the costs that distribution companies can and
cannot control. Therefore, all variations in real costs,
even those that cannot be managed by the companies,
are passed to network owners. This increases the risk
to investors, without a clear benefit.
✔✔ Despite its complexity, the remuneration mechanism is,
in practical terms, too approximated (it relies on a small
sample of companies, averages results from different
sources in a two-thirds and one-third fashion, uses sim-
plistic greenfield model firms, and so forth). Even for
the same company, results can vary too much, depend-
ing on who determined them (the authority or the firm;
see the values and biases in Figure 6). This implies that
a network owner can realize a lower or higher cost of
capital depending on the inherent randomness of the
process, not on corporate efforts to reduce expenses.
✔✔ There is no coherence between the distribution of net-
work owners’ capital costs and efficiencies. Following
the principle of comparing similar companies through
yardstick regulation, firms with larger efficiencies
should receive larger returns. This does not occur under
Chile’s remuneration approach, which does not measure
firms’ relative inefficiencies to mimic real competition.
 ✔✔ A major issue is the absence of stakeholder engage- regulatory capture that could occur, since network compa-
ment, where parties contribute feedback and proposals nies have more technical expertise about network operation
through an advanced planning and remuneration pro- and planning.
cess. In the current framework, it is unclear how deci- Addressing the problem of information asymmetries
sions between companies and stakeholders to improve is key, potentially requiring a different approach than the
the distribution service (for example, installing new as- model-firm paradigm, where companies’ real costs are prac-
sets agreed among stakeholders to increase the supply tically ignored. There has been important progress to incen-
quality) could be adequately remunerated. tivize the quality of the information that companies provide
to regulators, achieved by offering a menu of contracts (see
The Path Forward the “For Further Reading” section), and this may inform the
Since the price-control regime in Chile has remained almost future debate in Chile.
untouched for nearly four decades, there is a consensus that
some of its features have to be updated. Despite this agree- Brazil
ment, there are heated debates about the level of change
that is necessary. Although the current approach is being The Regulatory Framework
questioned in light of the new paradigm, there is a group Brazil’s installed generation capacity is roughly 167 GW (as
of experts that defends it, supporting its arguments mostly of 2019), with a yearly energy consumption of 472 TWh (as
on the system’s simplicity and low-cost implementation. of 2018), serving 84.6 million users (as of 2019). Consumers
The group envisions a set of smaller changes in network in Brazil fall into two broad groups: 1) free consumers, who
remuneration. An opposite position supports a more radi- bilaterally negotiate for energy and pay regulated charges (or
cal change toward incentive-based remuneration methods, network tariffs) for access to distribution and transmission
such as Revenue = Incentives + Innovation + Outputs in networks, and 2) regulated consumers, who purchase energy
the United Kingdom, which recognizes and addresses the at a controlled price derived from supply-contract auctions
challenges associated with the
new paradigm. Such approaches,
however, would impose more 45
Distribution Charge (US$/kW/Month)

Company Regulator
complexity and require a greater 40
effort by regulators to understand 35
the costs and decisions of real
30
firms. Chile’s answer probably lies
25
between the two extremes. The
debate involves questions such as 20
the following: Up to which point 15
do real costs have to be recognized 10
in the remuneration process? What 5
would the new role, set of tasks, 0
and burden be for regulators? 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6
Some principles underlying 2004 2008 2012 2016
Chile’s remuneration method may Representative Company and Year
need to be maintained to smooth
any transition. Part of companies’ figure 6. The distribution charge per representative firm under the past four tariff
real costs may need to be rec- controls, according to company and regulator studies, before the weighting process.
ognized (for example, by using Representative firms are sorted by connection density (number 1 presents the highest
a “brownfield” model firm) to and corresponds to Santiago, Chile).
encourage investments in quality
and reliability and other desir- table 4. Tariff modalities in Brazil.
able distribution-service outputs
Time-Dependent Network
as well as to compensate investors Tariff User Structure Pricing Remuneration
for part of the risk they face. Rec-
ognizing companies’ real needs Conventional Low voltage kWh only No Volumetric
and costs would increase regula- Blue High voltage kWh and kW Peak/off peak Peak demand
tors’ burden of scrutiny and per-
Green High voltage kWh and kW Peak/off peak Peak demand
haps their responsibility in future
investment decisions. A major White Optional for kWh only Peak/intermediate/ Volumetric
low voltage off peak
concern involves the extent of the

may/june 2020 ieee power & energy magazine 43

 and pay the network tariff. There are large consumers con-
nected directly to the transmission network that do not pay
charges for distribution networks. Regulated consumers
account for approximately 70% of the energy consumption.
Brazil’s regulation and tariff-structure framework was
defined in 1968 with the objective of establishing general rules
for distribution companies’ charges. It was the responsibility
of the National Electric Energy Agency (ANEEL) to define
specific rules based on the legal framework’s guidelines. In
general, the regulatory framework deals with fixing and revis-
ing electricity tariffs, the general classification of consumers
according to voltage levels, tariff structures, and special sup-
ply conditions (for example, rural areas and public lightning).
There are two tariff structures that can be applied to con-
sumers connected to distribution networks:
1) binomial tariffs, which have two components, the first
representing charges for energy, losses, and other ex-
penses (for instance, the costs of governmental poli-
cies) in $/MWh, and the second representing trans-
mission and distribution charges in $/kW
2) monomial tariffs, in which all components are sum-
marized in a single volumetric tariff in $/kWh.
To apply the tariffs, consumers are classified according to
their voltage levels: group A (high voltage level) and group B
(low voltage level), which are broken into subgroups. Group
A is divided by voltage levels: A1 (≥230 kV), A2 (≥88 kV
and <230 kV), A3 (>44 kV and <88 kV), A3a (≥30 kV and
≤44 kV), A4 (≥2.3 kV and <30 kV), and AS (<2.3 kV in the
underground network). Group B is divided by consumer
classes: B1 (residential), B2 (rural), B3 (other classes, such
as commercial, industrial, and public power), and B4 (public
lighting). Binomial tariffs are compulsory for high-voltage
consumers, and monomial tariffs are mandatory for low-
voltage users. The tariff menu offered to consumers is sum-
marized in four modalities—conventional, blue, green, and
white—as illustrated in Table 4.
The conventional tariff modality is a constant, flat rate
used for almost all consumers connected to the low-voltage
network. Since 2018, users with an electricity demand higher
than 500 kWh have been able to opt for the white tariff, with
time granularity for peak, intermediate, and off-peak hours.
The peak period consists of the three consecutive working-
day hours that have the highest demand. It is defined by the
regulator as 5:30–8:30 p.m., which is not in line with the
actual peak load. As many consumers increase their use of
air conditioners, the peak load has shifted to 2–3 p.m. in
many regions. However, the regulator has not updated the
definition, which incentivizes consumers who opted for the
white tariff to increase their consumption during the real
peak time. Changes to this modality might not reduce the
system’s peak load, since many users will not necessarily (be
able to) change their consumption behavior. High-voltage
consumers can opt between the blue and green modalities.
The green one has a higher volumetric component for the
peak hour, which has motivated some consumers to build
44 ieee power & energy magazine may/june 2020
natural-gas cogeneration and diesel power plants to reduce
their peak load. For users who cannot shift their load, the
blue option may be more attractive.
The regulator periodically implements ratchets to revise
the tariffs. Readjustments are made annually on a percent-
age of the costs that network companies cannot control.
If the proportion of the uncontrollable costs increases or
decreases, companies’ revenues will be scaled to avoid
rewarding or penalizing firms. Tariff reviews are carried out
at the end of each control period. The regulated revenue is
determined and decoupled from the evolution of the actual
costs. The idea is for the regulated revenue to cover efficient
costs. The relative efficiency of the distribution companies is
measured through data-envelopment analysis. Thus, if a net-
work company can establish a more efficient cost structure,
incurring expenses that are lower than those included in the
tariffs, it can capture those savings within the control period.
During the periodic tariff review (which occurs, on average,
every four years), efficiencies are passed on to the consumer.
Distribution companies whose concession was renewed after
2015 have a revision every five years.
From the regulator’s point of view, passing efficiency
gains to companies encourages efforts to reduce the cost
of providing the service. On the other hand, if uncon-
trollable costs are substantially altered, companies will
face expenses that do not have tariff coverage during the
period between revisions. In such cases, companies may
request an extraordinary tariff revision to reestablish
their financial balance. The regulator is responsible for
judging the merits of this request and deciding whether to
approve the application.
Brazil’s price regulation incentivizes companies to save
investment costs during the first years of the control period,
toward the end of which the efficiency incentive becomes
weaker than other enticements in the policy framework.
Closer to the next ratchet, the incentive for companies to
increase their RAB during the long term becomes stron-
ger. This is due to the Averch–Johnson effect, which grows
during the last part of the control period. It occurs because
investors’ costs of capital (used by network owners to dis-
count future cash flows) may be lower than the regulator esti-
mated to determine the allowed revenue. Although this effect
is mostly associated with cost-of-service regulations, it may
occur under incentive-based rules. Because this is counterin-
tuitive, we illustrate the fundamentals in an example.
The Averch–Johnson Effect
on Incentive-Based Regulation
Table 5 lists the profits (in net present value) from three
investment options for a hypothetical distribution company
whose revenues are regulated through an incentive-based
approach, where income remains fixed during the four-
year control period regardless of the firm’s costs. In our
example, the network owner realizes (after the referential
network-expansion plan has been approved) that part of the
planned investment (with a cost of US$100) can be elimi-
nated without affecting reliability (for instance, there may
be no way to exercise operational measures, such as demand
control during peak hours, at a very low cost, which will
be assumed to be negligible for simplicity). The question
is whether the regulatory framework will incentivize the
network owner to realize the saving or opt a different, inef-
ficient result. Other data relevant to the problem include
the investor’s 7% cost of capital, the 8% regulated rate of
return (used to calculate the company’s revenues as the
remuneration of the RAB minus the asset depreciation, that
is, revenue t = RAB t -1 # rate of return - depreciation t), and
the assets’ 20-year life span. The investment options are (see
Table 5) as follows:
✔✔ Option 1: Undertake the investment (albeit not need-
ed) at the beginning of the control period (as planned),
which will be remunerated during the lifespan of the
asset (including the four-year control period).
✔✔ Option 2: Eliminate the investment and save $100.
Revenues will remain fixed as agreed during the con-
trol period (that is, assuming the initial investment of
$100) to incentivize efficiency. Beyond the control pe-
riod, revenues will be canceled since the investment
was not undertaken.

table 5. Yearly revenues, costs, and profits of three investment options (in generic currency).
      Option 1 Option 2 Option 3
Invest­ Profit in Invest­ Profit in Invest­ Profit in
Yearly RAB ment Present ment Present ment Present
Year Depreciation (Option 1) Revenue Cost Value Revenue Cost Value Revenue Cost Value
100 100 –100
1 5 95 13 12.15 13 12.15 13 12.15
2 5 90 12.6 11.01 12.6 11.01 12.6 11.01
3 5 85 12.2 9.96 12.2 9.96 12.2 9.96
4 5 80 11.8 9 11.8 9 11.8 100 –67.29
5 5 75 11.4 8.13 13 (as in 9.27
year 1)
6 5 70 11 7.33 12.6 8.4
7 5 65 10.6 6.6 12.2 7.6
8 5 60 10.2 5.94 11.8 6.87
9 5 55 9.8 5.33 11.4 6.2
10 5 50 9.4 4.78 11 5.59
11 5 45 9 4.28 10.6 5.04
12 5 40 8.6 3.82 10.2 4.53
13 5 35 8.2 3.4 9.8 4.07
14 5 30 7.8 3.02 9.4 3.65
15 5 25 7.4 2.68 9 3.26
16 5 20 7 2.37 8.6 2.91
17 5 15 6.6 2.09 8.2 2.6
18 5 10 6.2 1.83 7.8 2.31
19 5 5 5.8 1.6 7.4 2.05
20 5 5.4 1.4 7 1.81
21 6.6 1.59
22 6.2 1.4
23 5.8 1.22
24 5.4 1.06
Total 6.72 42.12 47.24

may/june 2020 ieee power & energy magazine 45

 ✔✔ Option 3: Delay the investment to the end of the four-
year control period, increasing future revenues in
comparison with option 1, since the asset will be newer
(less depreciated).
Table 5 demonstrates that although option 2 is, from the
network owner’s perspective, more attractive than option 1
(as expected, since this is aligned with the primary objec-
tive of incentive-based regulation), option 3 provides higher
long-term profits. This corresponds to an inefficient and
unintended outcome of price-control regulation, since, from
a central-planning perspective, the most efficient result is
option 2, creating a need to increase the efforts to scrutinize
investments (in volume and timing). This also incentivizes
network owners to submit inflated forecasts of their planned
network assets, as reported in other countries, such as the
United Kingdom during the early 2000s under the retail
price inflation minus expected efficiency improvements
approach. Note that options 2 and 3 will become equally
attractive if investors’ capital costs and regulated rates of
return are the same.
Concerns About the Low-Voltage
Tariff Structure
The tariff structure applied to low-voltage-level consum-
ers (group B) attempts to remunerate the electricity value
chain’s CAPEX and OPEX: generation, transmission,
distribution, and other charges. The final tariff paid by
each consumer is divided into two parts: 1) energy and
2) network. The energy tariff contains a large part of the
distribution companies’ unmanageable costs: energy pur-
chases, transmission network losses, and charges related
to consumption. The network tariff contains the remaining
unmanageable expenses (distribution network losses, trans-
mission costs, and charges related to peak loads) and the
manageable ones, which are related to distribution compa-
nies’ CAPEX and OPEX.
Although regulators recently allowed the group B elec-
tricity tariff to be multipart, the levy structure remains volu-
metric with no time dependency. Given the growth in DERs,
Total Installed Capacity (GW)
25
New Tariff Structure
20 Current Tariff (Volumetric)
15
10
5
0 connected), it would better balance the cost and benefits of
18
19
20
21
22
23
24
25
26
27
20
20
20
20
20
20
20
20
20
20
Year
figure 7. The binomial tariff’s impact on the penetration of
DG for low-voltage consumers. (Data source: EPE, 2018.)
46 ieee power & energy magazine may/june 2020
mainly solar DG by prosumers, the volumetric tariff will not
guarantee adequate remuneration because lower net energy
sales reduce distribution companies’ revenues. The growth
of solar DG could potentially decrease firms’ remunera-
tion below the level required to recover CAPEX and OPEX.
Therefore, there will be an increase in tariffs, originating
a cycle of incentives for consumers to become prosumers,
which is known as the death-spiral problem.
Concerns about the impact that low-voltage prosum-
ers’ volumetric tariff has on distribution companies are
becoming increasingly important in Brazil because of the
two net-metering mechanisms that have been in place since
2014. The first is called local DG, which is the standard net-
metering approach, where prosumers install a rooftop solar
panel on the place of consumption and use the distribution
network to balance their energy consumption and produc-
tion. In the second, called remote DG, prosumers install
rooftop solar away from their premises (and potentially very
far away), meaning that consumers use the distribution net-
work to meet 100% of their load. Sharing DG among differ-
ent consumers is also allowed under this approach. Hence, a
wide variety of different business models has been created
to take advantage of the situation, encompassing apartment
dwellers who benefit from net metering by installing roof-
top solar on beach houses to groups that lease land to share
the benefits associated with the remote DG mechanism.
In local and remote DG arrangements, the monomial
volumetric tariff is applied to the net monthly consumption
of the prosumer or group of consumers. Generated energy
that exceeds the monthly consumption is compensated with
energy credits to be used for reducing the cost of consump-
tion within the next five years. If the net consumption (gen-
eration minus demand plus past energy credits) is zero or
negative for a month, the prosumer pays a minimum tariff
for distribution network availability. However, the tariff for
network availability represents, on average, 29% of the elec-
tricity bill, while transmission and distribution networks’
CAPEX and OPEX account for 46%.
The regulator has been discussing new tariff structures
for the distribution segment as well as changes to the net-
metering mechanism for prosumers. Extending the binomial
tariff to low-voltage-level consumers is one of the alterna-
tives under consideration. According to the Energy Research
Company (EPE), doing so could significantly impact the
evolution of DG. Following the ten-year energy plan, apply-
ing the binomial tariff to all consumers could reduce the DG
installed capacity from 21 to 12 GW by 2027, as illustrated
in Figure 7. Although such a result could appear negative
from an energy-policy perspective (since less DG would be
system expansion.
Tariff-Structure Alternatives
During 2018 and 2019, ANEEL held two public hearings to
discuss new tariff structures for low-voltage consumers and
 table 6. Alternative tariff structures under discussion in Brazil.
Alternative Name Description
— Current Present condition, used as a comparison parameter

One New minimal payment Increase minimum consumption levels

Two Commercial cost Definition of fixed tariff without differentiation among consumers, charged in U.S. dollars
per consumer to recover commercial distribution costs (for example, billing and help desks)
Three Fixed cost Definition of fixed tariff without differentiation among consumers, charged in U.S.
dollars per consumer to recover commercial and distribution network costs
Four Differentiated fixed cost Definition of fixed tariff for different consumer sizes, charged in U.S. dollars per consumer
Five Load Definition of a tariff in U.S. dollars per kWh for distribution-system-availability costs

changes to the net-metering mechanism. For the low-voltage new meters would represent a 13% increase for a typical elec-
consumers, ANEEL defined five alternatives to the purely vol- tricity bill. In alternatives 3 and 4, a fixed payment to recover
umetric tariff, carrying out a regulatory analysis for representa- the distribution network costs would be applied. The differ-
tive distribution companies based on the following principles: ence is that, in alternative 4, the amount to be paid is based on
1) Revenue adequacy: guaranteeing sufficient service a consumer’s average consumption during the past 12 months,
remuneration while in alternative 3 the fixed payment does not vary with
2) Allocative efficiency: encouraging productive con- consumption. According to the regulator, alternative 4 pres-
sumer behavior during the short and long term ents the better tradeoff between costs and benefits (especially
3) Transparency and intelligibility: simplicity of adop- when considering the transaction cost), since there would be
tion and understanding no need to change the existing meters.
4) Justice and nondiscrimination: no competitive advan- At the second hearing, the regulator proposed five alter-
tages to any consumer. natives for net metering. The mechanism would continue
The alternatives under discussion are listed in Table 6. applying volumetric tariffs, but, contrary to today’s prac-
Alternative 1 proposes to increase the minimum amount of tice, rates would vary for withdrawals and injections, dif-
power prosumers must consume to use the distribution net- ferentiating prosumers’ production and consumption. While
work (even if they consume zero or negative kWh), from 100 consumption would continue paying for the entire electricity
to 219 kWh, which is necessary to recover distribution net- supply chain, production would pay for only part of it. Table 7
work costs. This would reduce the distribution companies’ presents the alternatives. The first, for example, corresponds
cost-recovery risk. Still, consumers without DG would sub- to the current practice, where the energy produced by DG is
sidize prosumers, since both would be exposed to volumetric remunerated at a price that contains all components. Instead,
tariffs. In alternative 2, the minimal consumption limit would alternative 5 remunerates DG production at the wholesale
be replaced by a fixed tariff per consumer. The analysis car- price (without including network and other sector charges),
ried out by the regulator showed that cross subsidies remained which is significantly lower than that of the first option.
in this approach, since all consumers would be charged by the The chosen alternative will function temporarily, until a
same tariff regardless of their
peak load levels.
Alternative 5, with a pay- table 7. Components included in the volumetric tariff to remunerate DG
ment in $/kW, proved to be the production.
most efficient in terms of cost Sector Sector
allocation, since all consum- Distribution Transmission Charges Distribution Charges
Network Network Applied to Network Applied to Energy
ers would be charged based Alternative Tariff Tariff Peak Load Losses Consumption Tariff
on their (approximate) use of
the network (although the time Status quo X X X X X X
element is still missing, that is, One X X X X X
when the peak load occurs).
Two X X X X
This approach’s main problem
concerns the cost of replac- Three X X X
ing all the low-voltage-level Four X X
consumers’ meters. Initial es-
Five X
timates show that the cost of

may/june 2020 ieee power & energy magazine 47

 prescribed DG-penetration threshold is reached, enabling
regulators to make policies that actively promote DG during
the short term and giving investors certainty about how long
such a program will last.
Final Remarks
The transition toward economic, reliable, and low-carbon
power networks in Latin America is not easy, since spe-
cial care is needed to avoid raising electricity bills beyond
affordable levels. Equity and fairness are highly relevant to
the variety of tariff levels and desirable network-services
outputs, such as quality and reliability levels, among net-
work users, making network regulation and rate making
particularly difficult. We illustrated the regulatory con-
cerns in Latin America, with a focus on Chile and Brazil,
related to finding the right distribution network price-con-
trol regulations to achieve energy policy goals in the most
effective manner.
Balancing distribution-sector costs and benefits will
require a more active network-operation approach to take
advantage of innovative smart grid technologies and cre-
ate opportunities for OPEX-based solutions (which are
more cost-effective) to displace CAPEX-based solutions.
Network users, especially DERs in the form of DG, stor-
age, flexible demand, and so forth, will need to cooperate
by deploying and managing their equipment properly. This
will require the use of smart meters and new IT technolo-
gies, price signals, and other control signals that must be
managed in real time.
Incentive-based regulations (which are preferred in most
of Latin America instead of cost-of-service regulations) and
tariff structures must evolve to align private and public pol-
icy objectives, incentivizing network companies and users
(producers, consumers, and prosumers) to act accordingly.
Under an evolved approach, regulators are expected to take
a more active role in scrutinizing the data of real companies
regarding costs, assets, and plans. Regulations would need
to provide appropriate guarantees, rewards, and penalties
for companies and users, motivating them to deploy efficient
solutions aligned with public policy, especially reliability.
They should also ensure that the resulting tariffs are afford-
able to all, including vulnerable consumers in rural areas.
This evolution will be challenging in Latin America as it will
require more regulatory resources, since the historical prac-
tice has relied on simpler approaches with a low burden and
implementation cost.
Acknowledgments
We would like to acknowledge financial support from
the National Research and Development Agency of Chile
(through the Complex Engineering Systems Institute CONI-
CYT PIA/BASAL under grant AFB180003; the National
Fund for Scientific and Technological Development, under
grants 1181928 and 1181136; the National Fund for Scien-
48 ieee power & energy magazine may/june 2020
tific and Technological Development/Iniciacion, under grant
11180875; the Fund for Financing Research Centers in Prior-
ity Areas, under grant 15110019); and FONDEF IDeA grant
number ID19I10158). We would also like to thank Paula
Valenzuela from PSR, Rio de Janeiro, Brazil, for her valu-
able contributions.
For Further Reading
T. Jamasb and M. Pollitt, “Reference models and incentive
regulation of electricity distribution networks: An evalua-
tion of Sweden’s Network Performance Assessment Model
(NPAM),” Energy Policy, vol. 36, no. 5, pp. 1788–1801,
2008. doi: 10.1016/j.enpol.2008.01.034.
P. L. Joskow, “Incentive regulation in theory and prac-
tice: Electricity distribution and transmission networks,”
in E
­ conomic Regulation and Its Reform: What Have We
Learned? N. L. Rose, Ed. Chicago, IL: Univ. of Chicago
Press, 2014, pp. 291–344.
G. Strbac, D. Kirschen, and R. Moreno, Reliability Stan-
dards for the Operation and Planning of Future Electric-
ity Networks (Foundations and Trends in Electric Energy
Systems, vol. 1). Hanover, MA: Now Publishers, 2016,
pp. 143–219.
R. Moreno, A. Street, J. M. Arroyo, and P. Mancarella,
“Planning low-carbon electricity systems under uncertainty
considering operational flexibility and smart grid technolo-
gies,” Philos. Trans. Roy. Soc. A, Math. Phys. Eng. Sci.,
vol. 375, no. 2100, Art. no. 20,160,305, 2017. doi: 10.1098/
rsta.2016.0305.
P. Djapic et al., “Taking an active approach,” IEEE Power
Energy Mag., vol. 5, no. 4, pp. 68–77, 2007. doi: 10.1109/
MPAE.2007.376582.
J. D. Jenkins and I. J. Pérez-Arriaga, “Improved regula-
tory approaches for the remuneration of electricity distribu-
tion utilities with high penetrations of distributed energy
resources,” Energy J., vol. 38, no. 3, pp. 63–91, 2017. doi:
10.5547/01956574.38.3.jjen.
Biographies
Rodrigo Moreno is with the University of Chile, Santiago;
Instituto Sistemas Complejos de Ingeniería, Santiago; and
Imperial College London.
Bernardo Bezerra is with PSR, Rio de Janeiro, Brazil.
Hugh Rudnick is with Pontificia Universidad Católica de
Chile, Santiago.
Carlos Suazo-Martinez is with SPEC, Santiago, Chile.
Martha Carvalho is with PSR, Rio de Janeiro, Brazil.
Alejandro Navarro is with the University of Chile, San-
tiago, and Systep, Santiago.
Carlos Silva is with Universidad Adolfo Ibáñez, San-
tiago, Chile.
Goran Strbac is with Imperial College London, United
Kingdom.
p&e

 China’s Solar
Subsidy Policy

©ISTOCKPHOTO.COM/GYN9038

Government Funding Yields to Open Markets

T
By Houqi Dong, THE CHINESE GOVERNMENT HAS ISSUED NUMEROUS REGULA-
tions that significantly affect the number of photovoltaic (PV) installa-
Bo Zeng, tions in the country and the subsidies for their use. This article summa-
Yuqing Wang, rizes the internal and external environment of China’s PV industry and
describes its future trends and prospects and also discusses a proposed
Yingxin Liu, rate-making process and renewable energy supervision.
In 2017, China added 53 GW of PV power generation, an increase of
and Ming Zeng 18.5 GW compared to 2016, reaching a new historical high year-on-year
growth of 53.62%. However, prior to 2018, the national PV policy had many
issues, such as high presubsidy prices, a lack of incentives for technological
innovation, and the absence of installation performance monitoring.
Digital Object Identifier 10.1109/MPE.2020.2971824
Against a backdrop of excess production of PV panels and a waning
Date of current version: 17 April 2020 desire to provide financial support to the manufacturers, the government

may/june 2020 1540-7977/20©2020IEEE ieee power & energy magazine 49

 issued the 531 New Policy in May 2018 to reduce the
intensity of subsidies and limit the scale of new produc-
tion. Implementing the policy led to higher prices and a
sharp drop in domestic demand for new PV installations.
During the first three quarters of 2018, 34.5 GW of PV
PV System-Application

PV Power-Generation
power were added, down 19.7% from a similar period in

PV Application
Downstream

2017. Driven by lower local demand, silicon wafer pro-

Products
Products

System
duction decreased by 48.9% during the third quarter, and
the output of batteries and associated components fell by
more than 20%. Downstream enterprises were directly
affected by the policy change and sluggish demand, and
their profit margins shrank sharply through the second
half of 2018.
To ensure the sustainable development of PV panel
production, the State Energy Administration established
reduced subsidies in June 2018, extending them until
2022 while also postponing a planned feed-in tariff
PV Modules

Solar Cell
Modules

on non-PV distributed resources. The government also

sought to stabilize market expectations by increasing the
goals for installed PV nameplate capacity from the origi-
nal 13th Five-Year Plan, which covered 2016–2020.
Midstream

The short-term fluctuation in the PV industry’s devel-

opment occurred in response to the initial rapid expan-
Monocrystalline

Polycrystalline

sion of production capacity, the 531 New Policy, and the

Solar Cells

Solar Cells
PV Cells

new subsidies. Since 2018, prices across the PV industry

supply chain have declined, impacting the profitability of
enterprises, and the prices of silicon and silicon wafers
continue to fall sharply, plunging 39% between May and
November 2018 and hurting the performance of upstream
PV enterprises. Due to the lower raw material costs and
Monocrystalline

Polycrystalline
Silicon Wafers

Silicon Wafers

prices, some manufacturers of batteries and associated

components have maintained adequate margins.
This article discusses the history of China’s PV
Ingots/Silicon Wafers
Silicon Rods/Silicon

industry and the drivers of each development stage,

from manufacturing through installation, and summa-
rizes the country’s PV electricity policies. The internal
Monocrystalline

Polycrystalline
Silicon Ingots
Silicon Rods

and external environments affecting the PV industry are

discussed as well as opportunities for and threats to the
sector’s development. Future trends and prospects for
national policies, technologies, industries, and markets
Upstream

are described in relation to electricity rates.

Silicon Material
Polycrystalline
Metallurgical-Grade Solar-Grade

China’s PV Industry
Crystalline Silicon

Development Status
figure 1. The PV industry chain.
Furnace
Feedstock

Silicon

The PV industry collects and processes silicon materials

to build solar electricity cells, modules, and generation
Monocrystalline

systems. The upstream PV industry gathers crystalline

Silicon

silicon feedstock that is manufactured into silicon rods,

ingots, and wafers. Midstream companies manufacture
PV cells, cell modules, and related solar production
equipment, and downstream firms integrate and operate
the PV power station system, as shown in Figure 1. The
industry’s specific situation and development status are
now discussed.

50 ieee power & energy magazine may/june 2020

 In 2017, China’s solar cell production exceeded 68 GW,
an annual increase of more than 30%, constituting nearly
70% of the global total.
Crystalline Silicon Material
The most upstream part of the industrial chain is the manu-
facturing of solar crystalline silicon. Owing to its high tech-
nological barrier, the core technology of polycrystalline-
silicon production was long dominated by companies in
the United States, Germany, Japan, and South Korea. Since
2005, China’s polycrystalline silicon industry has been pri-
marily promoted by the rapid development of its domestic
PV industry. The industry’s consolidation has increased
through mergers and acquisitions as well as the elimina-
tion of enterprises in response to overcapacity. By the end of
2017, the global polycrystalline silicon production capac-
ity was 0.52 million tons/year, and China was capable of
producing 0.276 million tons/year. Domestic production
was dominated by 10 enterprises, each having a production
capacity exceeding 10,000 tons/year, with a total production
capacity of 0.228 million tons/year, amounting to 82.6% of
the country’s polycrystalline silicon production capacity.
In recent years, driven by a strong domestic PV market,
national polycrystalline silicon enterprises have promoted
gradual technological improvements, maintained high load
productions, and constantly improved their process technol-
ogy. This resulted in a continuous increase in production
capacity and decrease in the energy consumption associ-
ated with manufacturing as well as other rapid reductions in
production costs. In 2017, global polycrystalline silicon pro-
duction reached 0.44 million tons, of which China contrib-
uted 0.242 million tons, accounting for more than one-half
(54.5%) of the world’s polycrystalline silicon output. Thus,
China’s dependency on foreign market supplies of polycrys-
talline silicon feedstock was reduced.
Silicon Wafers
Silicon wafer manufacturing is a capital-intensive down-
stream process of the polycrystalline silicon industry, with
its products classified as monocrystalline and polycrystal-
line silicon wafers. Due to different manufacturing pro-
cesses, monocrystalline silicon products are characterized
by a high conversion efficiency, operability in low operat-
ing temperatures, and good low-light performance; poly-
crystalline silicon products, however, have dominated the
PV application market in recent years because of their cost
advantages. But since 2016, the cost of monocrystalline
silicon products has greatly declined thanks to advances in
manufacturing technology.
China’s share of the world market for monocrystal-
line silicon wafers was approximately 18% in 2015, and it
may/june 2020 ieee power & energy magazine 51
increased to 27% in 2017, according to nationally developed
statistics. By 2017, a monocrystalline wafer cost approxi-
mately US$0.10 more than a polycrystalline one. Both prices
have fallen continuously during the past three years, with
polycrystalline-silicon wafers prices sliding from US$0.80/
piece (pc) at the beginning of 2016 to lower than US$0.30/pc
by the end of 2018.
In China, silicon wafer production reached 105.5 GW
in 2017, with a year-on-year growth rate of 39%. The wafer
industry is highly concentrated, with the top 10 companies
constituting 77% of the total national production. The coun-
try produces 83% of the world’s output and has eight of the
top 10 global silicon wafer manufacturers. With the increase
in the world market share of monocrystalline products,
China’s wafer manufacturers have expanded their produc-
tion capacity.
Solar Cells and Modules
The solar cell manufacturing process is capital and technol-
ogy intensive, requiring companies to use the latest tech-
niques to improve efficiency. Higher solar cell conversion
efficiency and lower production costs improve market con-
ditions for a manufacturer’s power systems. Thus, PV cells’
conversion efficiency is a key driver of price competition
among solar cell companies.
Under the 12th Five-Year Plan, from 2011 to 2015, the
nation accelerated the development of its PV cell manufac-
turing technology, and its product quality greatly improved,
with product-commercialization efficiency increasing by
approximately 0.3–0.4% per year. In 2017, China’s solar cell
production exceeded 68 GW, an annual increase of more
than 30%, constituting nearly 70% of the global total. Pro-
duction costs have continued to decline as manufacturing
technology improves.
Solar cell-module manufacturing has short plant-construc-
tion periods, high technical and capital thresholds, and close
access to domestic markets. In 2017, China’s solar module
production was approximately 75 GW, with an annual growth
rate of 43.4%, constituting 71% of the world’s output. Better
PV product quality and reduced prices have been achieved
through production-technology improvements; declining
manufacturing costs; and increased automation, digitization,
and networking. The average 2015 price for crystalline sili-
con modules of approximately US$0.60/W dropped to lower
than US$0.25/W in 2018. Figure 2 shows the national pro-
duction of polycrystalline silicon, silicon wafers, solar cells,
and solar cell modules from 2012 through 2017.
 The country’s PV manufacturing industry faced an antidumping
and antibribery investigation by the United States and Europe
during 2011 and 2012.

PV Application Systems Historical Review

As of 2019, large-scale ground power plants constitute more
than 80% of China’s installed PV capacity. Under the influ-
ence of “eliminating the use of and restricting the feed-in of
PV-generated electricity” and the reduction of feed-in tar-
iffs, the country has accelerated the construction of applica-
tion-demonstration zones for distributed PV power genera-
tion. As a result, investment in the PV industry has gradually
shifted to distributed projects, focusing on roof-based dis-
tributed PV power-generation systems in large-scale indus-
trial parks, economic-development zones, public facilities,
and residential buildings.
In 2017, when rapid growth occurred for all types of PV
systems, China’s new installed capacity was approximately
53 GW, a 57% increase from 2016, constituting more than
50% of the global total, resulting in the country’s distributed-
PV installed capacity reaching 19 GW. Since then, the cumu-
lative PV installed capacity has reached 130 GW, with the
annual increase and cumulative PV installed capacity rank-
ing first in the world.
Table 1 summarizes the characteristics of the development
phases in an overall expansion of China’s PV industry.
Before 2007, the industry was in a demonstration phase in
which the industry was primarily centered on the promo-
tion of demonstration projects; that phase was also known
as the initial exploration period. Between 2007 and 2010,
despite the global financial crisis, the international PV
industry developed rapidly, and China’s government sup-
ported and launched many PV projects. The national PV
manufacturing industry quickly increased in scale, and by
2007, the country had become the world’s largest producer
of solar cells.
The Golden Sun project was the first national implemen-
tation of PV subsidies. The country’s PV ­manufacturing
industry faced an antidumping and antibribery investiga-
tion by the United States and Europe during 2011 and 2012,
which caused a serious setback to its progress. From 2013
to May 2018, the PV industry adjusted its marketing direc-
tion and contributed significantly to developing its domestic

140

120

100
Production

80

60

40

20

0
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Polycrystalline
4.5 8.4 7.1 8.4 13.6 16.5 19.4 24.2 25.6 28
Silicon (10,000 Tons)
Silicon Wafers (GW) 11 20 26 29.5 38 48 64.8 91.7 109.2 120
Solar Cell Pieces (GW) 10.8 21 23 25.1 33 41 51 72 87.2 95
Modules (GW) 10.8 21 23 27.4 35.6 45.8 53.7 75 85.7 93

figure 2. China’s production of polycrystalline silicon, silicon wafers, solar cell pieces, and modules.

52 ieee power & energy magazine may/june 2020

market. The government formulated PV industry develop-
ment plans, subsidy policies, and construction projects; con-
sequently, China’s PV installed capacity became the high-
est globally, transitioning from a large PV manufacturing
country to a large PV application country. Moreover, the
initiation of a series of programs, such as the PV Forerun-
ner, promoted significant technological progress and lower
installed PV costs. With the issuance of relevant industry
standards and specifications as well as the implementation
of a PV feed-in tariff and PV project construction mandates,
the industry entered another period of rapid development,
expanding significantly faster than planned.
Industry overdevelopment and a subsidy shortfall led
China to enact the 531 New Policy, which took effect on
table 1. The development of China’s PV industry.
Development Period Related Policy
Initial demonstration • Brightness Program
(before 2007) • Delivering Electricity to Every
Village
Industry establishment • Golden Sun
(2007–2010) • PV Building Application
Demonstration
Setback resulting • The National Energy
from antidumping Administration’s 12th Five-Year
and antibribery Plan for Solar Electrical Energy
investigation Generation Development
(2011–2012) • The State Grid’s Opinion on Grid-
Connection Services of Distributed
PV Power Generation
• A PV power generation feed-in
tariff policy formulated by the
National Development and Reform
Commission
Recovery and rapid • The National Development and
development Reform Commission’s Notice
(2013–2018) on Promoting the Healthy
Development of the PV Industry by
Using Price Leverage
• The National Energy
Administration’s Notice on Further
Implementing Policies Related to
Distributed PV Power Generation
• PV Poverty Alleviation
• PV Forerunner
• The State Council’s Opinions
on Promoting the Healthy
Development of the PV Industry
Transition to PV grid • Notice on Matters Related to PV
parity (2018–present) Power Generation in 2018
• On Actively Promoting Grid Parity
for Wind Power and PV Power
Generation Without Subsidy
may/june 2020 ieee power & energy magazine 53
31 May 2018. The government aimed to reduce subsidies
and limit development, which significantly affected the PV
industry. Moreover, PV power plant construction was sus-
pended during the second half of 2018. At the beginning of
2019, however, the government proposed subsidy-free grid
parity for PV electricity. Figure 3 illustrates China’s PV
installed capacity through the years.
Evolution of China’s PV Electricity
Price Policy
China’s PV subsidies primarily include installation assis-
tance, tax relief, and feed-in tariff mitigation, which varies
by types (centralized versus distributed PV) and regions
(which have different resource types). Since the creation of
Macro Situation Solar PV Capacity
The domestic PV power-generation No more than
projects had little marketization, and 0.5 GW by 2007
most of them were demonstrations.
The PV manufacturing industry grew More than 1.5 GW
rapidly in foreign markets, improving
technology and making capital
investments. The global financial
crisis occurred, and the industry
suffered setbacks and market turmoil.
Simultaneously, a series of policies
was introduced, and the PV sector
was accepted as an emerging strategic
industry that promoted a new round of
investment.
The United States launched an More than 9 GW
antidumping and antibribery
investigation, later joined by Europe,
resulting in a serious setback to the
development of China’s PV industry.
The national government increased
its support for PV applications, and
the country’s PV companies quickly
adjusted their focus to the domestic
market, thereby realizing distributed PV
applications.
The government introduced intensive More than 126 GW
support policies for the PV industry (2016: 34 GW;
that were rapidly implemented. The 2017: 53 GW)
PV industry entered a period when the
technology developed quickly, domestic
feedstock and equipment became
mainstream, costs fell, and the power-
generation efficiency increased, with
continuous, fast-paced growth in the PV
installed capacity in China and other
primary markets.
The PV industry policy tightened, 2018: 44 GW
subsidies decreased, construction
decelerated, and the sector entered a
slow phase. Trade tariffs became more of
an issue for China, which was accused
worldwide of not respecting intellectual
property rights.
 its initial policies in 2011, the National Development and
Reform Commission has launched new policies seven times
that each changed subsidies in the PV feed-in tariff. Every
time a new subsidy policy is launched, it affects the PV
industry significantly, albeit to a varying degree, thus result-
ing in fluctuations in the capital market. The country’s PV
benchmark electricity prices are shown in Figure 4. Region 1
has more than 1,600 annual equivalent-utilization hours.
Region 2 has 1,400–1,600 and Region 3 has 1,200–1,400
annual equivalent-utilization hours.
In 2013, to mitigate the impact of the antidumping and
antibribery investigation of China’s PV companies by the
United States and European Union, the national govern-
ment decided to vigorously support the development of its
domestic PV industry, especially the PV downstream indus-
try. Therefore, after the initiation of the investigation, the
government proposed subsidies for the PV feed-in tariff by
dividing the country into three types of resource regions
drawn according to insolation levels, conditions, and bench-
mark electricity prices.
After the 2013 subsidy policy for the PV electricity price
had been implemented for more than two years, the PV
industry underwent rapid development because of the subsi-
dies, with a significant growth in newly added PV installed
capacity that constituted more than 50% of the global newly
installed PV capacity and resulted in a rapid reduction in PV
product costs. To promote the development of the PV indus-
try, the government sets the next year’s new benchmark elec-
tricity price for PV power generation at the end of each year,
with an average annual decline of 1.453–2.18 cents per kilo-
watt hour. Because governmental support for PV is subject to
annual change, PV power-generation systems are installed as
soon as possible so that higher subsidies can be received in
the early years.
On 31 May 2018, the Nationa l Development a nd
Reform Commission changed the long-established policy
of gradually reducing subsidies at the end of each year
and announced a sudden major decline in PV subsidies.
Despite this reduction, because of quota limits, new PV
power plant projects would not be approved. As a result
of this news, the PV industry immediately experienced a
severe setback, with companies’ share prices plummet-
ing. Subsequently, in early 2019, the National Develop-
ment and Reform Commission issued a notice regarding
the promotion of PV grid parity and proposed a series of
policies in favor of subsidy-free feed-in PV power, signal-
ing the government’s intention to further reduce or even
cancel the subsidies.
As shown by the changes in PV installed costs and
the adjustment of electricity prices during the past, the

120 450
425%
Setback 400
Phase Resulting 102
Initial Recovery and Rapid
100 From Antidumping and
Demonstration Development Phase 94.3 350
Antibribery Investigation
Phase
300
80 269.64%
73
250
206.72% (%)
60 200
53 53.06

166.67% 128.29% 44.26 150

43
40 38.4
34.54 100
100% 27.7 29.7
72.46% 42.74% 50
53.62%
20 16.63
15.13
10.95 10.6 0
2.39 5.56 7.26
2.07 3.57 –3.2% –16.6%
0.02 0.04 0.21 0.56
0 –50
2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Year

China’s New PV Installed Capacity (GW)

Global New PV Installed Capacity (GW)
Growth Rate of China’s PV Installed Capacity (%)

figure 3. China’s PV installed capacity.

54 ieee power & energy magazine may/june 2020

PV grid price typically falls as PV product costs decline,
with the former decrease sometimes occurring at a later
time. This is primarily attributed to policies that have not
been implemented as planned, thus severely affecting the
economy and benefits of a project. The price of electric-
ity generated from coal has remained low, preventing or
delaying PV grid parity.
PV Electricity Prices in China
In 2013, when the national PV subsidy policy was new, the feed-
in tariff was US$0.1308/kWh; in 2017, it was US$0.0944/kWh,
a five-year decrease of approximately one-third. In addi-
tion to national subsidies, the PV industry can also receive
explicit ones, such as local aid, project support, and enterprise
assistance, as well as implicit ones to help fund early stage
market d­ evelopment. Solar technology development reduces
PV costs, accelerating marketization and contributing to the
decline in subsidies. The government has introduced stopgap
measures until the subsidies are phased out, which is inevi-
table and the final stage in the industry’s maturity. As of 2019,
the cost-recovery period for PV systems in China was seven
to eight years.

0.16
Resource Region 1
0.14 US$0.1453 Resource Region 2
US$0.1380 US$0.1424 Resource Region 3
Benchmark Electricity Prices (US$/kWh)

US$0.1308 US$0.1278 Distributed Subsidy

0.12 US$0.1235
US$0.1162
US$0.1090 US$0.1090
0.1
US$0.1017
US$0.0944 US$0.0944
0.08 US$0.0872
US$0.0799
US$0.0726
0.06
US$0.061 US$0.061 US$0.061
US$0.0538
0.04 US$0.0465

0.02

0
2013–2015 2016 2017 2018 2018.6

Resource Region 1
Resource Region 2
Resource Region 3
Other Resource Region

figure 4. China’s PV benchmark electricity prices.

may/june 2020 ieee power & energy magazine 55

 Strengths–Weaknesses–Opportunities–
Threats Analysis of China’s PV Industry
The PV industry is strategic to China’s new energy develop-
ment. To understand the internal and external environments
the sector faces, the government conducted a strengths–
weaknesses–opportunities–threats (SWOT) analysis to for-
mulate development strategies, plans, and countermeasures.
The SWOT analysis appears in Figure 5.
Strengths
High Insolation Levels
China has abundant solar energy resources, with annual solar
radiation exceeding 5, 000 MJ/m 2 and more than two-thirds of
the country receiving better than 2,000 h of solar radiation each
year. China encompasses 1.08 million km 2 of deserts, primarily
distributed in the northwest region, with high insolation levels.
In 2017, it was estimated that the total radiation that a PV mod-
ule could receive at the optimal tilt angle was 1,725.3 kWh/m 2.
Low Cost
PV enterprises are growing to form industrial conglomerates,
thereby achieving economies of scale. As their shipments
increase and upstream, midstream, and downstream costs
decline, they benefit considerably.
Policy Support
Governments at all levels in China have provided subsidies
and made concessions to PV power-generation projects. They
have introduced encouragement and support policies in differ-
ent periods that provided higher subsidies than other countries,
including all aspects of the financing, construction, produc-
tion, and sale of PV projects. As the national supply-side struc-
tural reform and electric-power-system overhaul continue, the
functions of power-grid enterprises are changing, with gradu-
1 Widening of PV Subsidy Gap
ally increasing support for constructing distributed PV systems
and the local consumption of PV power.
Weaknesses
Overcapacity
With supportive policies, the domestic PV power-generation
market is growing rapidly. Stimulating downstream factors
have spurred PV manufacturing enterprises to expand their
production capacity to rank first in the world. However, with the
decline of PV subsidies, the construction of power-generation
projects has decelerated. Given insufficient domestic demand
and low export sales, the problem of manufacturing overcapac-
ity is becoming increasingly prominent.
Difficulties in Consumption
With the rapid increase in PV power generation, some areas
have experienced serious problems of reaching hosting-capac-
ity limits. The national PV generation rejection rate was 11%
in 2016 and 6% in 2017, and it eased to 3% in 2018 through
various efforts. However, the PV generation rejection rates are
still relatively high in some places, with Gansu and Xinjiang
Provinces reporting 2018 figures as high as 10.06 and 15.84%,
respectively. The main causes are as follows:
1) Solar energy resources are unevenly distributed, lo-
cated primarily in the “Three Norths” region, which
has a relatively backward economy and limited ability
to locally consume PV power.
2) The PV power station development speed far exceeds
the construction times for grid-network improvements.
3) The electricity market is an imperfect trading
mechanism.
Weak Industry Structure and Technology
China’s PV industry structure is relatively simple, and the
quality of its high-efficiency solar
cells and other products needs to
be improved.

1 Abundant Light Resources 2 Significant Capital Volatility

and Financing Difficulty Opportunities
2 Low Cost 3 Trade Risk
3 Policy Support 4 Missing Distributed- Good Macro Environment
Power Retail Market The government promotes elec-
tric-energy production and con-
sumption; strengthens the con-
struction of renewable energy
resources; proposes that nonfos-
sil energy sources supply 15% of
the primary power consumption
1 Overcapacity 1 Good Macro Environment by 2020; participates in the Paris
2 Difficulties in Consumption Agreement to limit carbon diox-
Weak Industrial 2 Promotion of Electric-
3 Power-System Reform ide emissions; and accelerates
Structure and Technology
3 Broad Global Market the establishment of a clean, low-
carbon, safe, and efficient modern
figure 5. The SWOT analysis of China’s PV industry. energy system. Corresponding

56 ieee power & energy magazine may/june 2020

policies provide the PV industry with a good external
macro environment.
Promotion of Electric Power System Reform
In March 2015, China promoted a new round of electric
power system reforms and proposed clear requirements for
transmission and distribution prices, construction of the
electricity market, establishment and operation of trading
institutions, and loosening of planning controls in power
generation and use. In particular, reforming market trad-
ing mechanisms and electricity prices provides an effective
method to solve the renewable energy consumption problem,
and it is conducive to renewable energy’s priority develop-
ment and fair participation in market transactions.
Broad Global Market
With the general trend of global energy systems undergoing
a low-carbon transformation and the increased competitive-
ness of China’s PV industry, broad international markets
have become available, with huge opportunities in develop-
ing countries, such as those in Latin America and Africa.
Threats
Widening PV Subsidy Gap
The need for renewable energy subsidy payments to be made is
increasing rapidly. The financial gap between the true costs and
revenues of PVs has widened, cumulatively exceeding roughly
US$15 billion by the end of 2017 due to delays in receiving
subsidy payments. This creates cash-flow issues for PV com-
panies, which recognize that future policy changes reducing
subsidies are potential threats to their business models.
Capital Volatility and Financing Difficulty
With the decline in China’s PV subsidies and the lower
limits on PV construction, the profits of some companies
will decline. Further declines could occur because of the
large subsidy gap and long cost-recovery period for fixed
investments in centralized PV power stations. Banks and
other investment institutions recognize that PV power
station projects pose an increased risk, rendering financ-
ing difficult.
Trade Risk
The 2012 antidumping and antibribery investigation and
subsequent Sino–U.S. trade war demonstrated the signifi-
cant risk in international trade.
Missing Distributed-Power Retail Market
Direct transaction is the primary mode whereby distributed
power-generation participates in the market. However, in
China it is still impossible to conduct electricity transactions
between a distributed power-generation project and users in
neighboring networks. The lack of such transactions hinders
the development of distributed PV systems.
may/june 2020 ieee power & energy magazine 57
Future Trends and
Development Prospects
Given the history of the worldwide PV industry and consider-
ing the advantages, disadvantages, opportunities, and threats
that China’s faces, a policy, enterprise, technology, and mar-
ket analysis is presented in this section to predict the sector’s
development path.
Policy
During the next few years, the government will strive to pro-
mote PV grid parity.
✔✔ Further decline in PV subsidies: Undoubtedly, direct
PV and electricity price subsidies will decline, and the
construction quota for centralized PV power ­stations
will be further limited, which is the most direct and
effective policy that the government can implement
to regulate domestic industrial development. The gov-
ernment may propose the requirements of PV grid
parity and initiate a number of pilot projects at some
future time.
✔✔ Promoting PV power consumptions: To ensure the
priority and effective utilization of PV-generated elec-
tricity, grid-parity, and low feed-in tariffs, provincial
power-grid enterprises have signed contracts with proj-
ect developers for the long-term purchase and sale of
electricity at fixed prices to reduce financial risks. The
government will strengthen the full-amount indemnif-
icatory-purchase policy by requiring grid companies to
guarantee power and energy payments to PV generation
stations. In addition, the government will require sup-
port for priority power generation in the case of power
rationing and allow compensatory incomes to be ob-
tained by trading (transferring) generation rights. The
government can also encourage power-grid enterprises
to develop interconnections that ensure that PV systems
can operate upon the completion of construction.
✔✔ Financial policy support: The government will in-
troduce policies to encourage financial institutions to
support the construction of unsubsidized PV power-
generation projects, appropriately increase the scale
of credit funds for the PV industry and promote in-
novative financial services, and support new power
generation for achieving grid parity. Moreover, it will
introduce policies to encourage and support the initia-
tion of eligible power-generation projects by facilitat-
ing the issuance of corporate bonds and accelerating
the bond-review process.
✔ ✔ Optimization of the assessment and supervision
mechanism: As long as a province fulfills the na-
tionally set energy-consumption goal, its renewable-
energy use exceeding its planned quota will not be
penalized when the central government assesses the
total energy consumption and energy-consumption
intensity. This policy will promote the construction
of subsidy-free PV power-generation projects. With
 respect to the risk that PV power stations pose, the
government will strengthen supervision and manage-
ment and formulate and improve the corresponding
policies and regulations.
✔✔ Continued implementation of the PV poverty-alleviation
policy: The government will introduce management
measures to increase PV development in economically
backward regions, support energy-development proj-
ects in poverty-stricken areas, and cooperate with de-
partments to provide concessions on feed-in tariffs for
PV projects and provide high-priority subsidy payments
to help poverty-stricken areas reduce electric rates.
Technology
✔✔ Increased market share of monocrystalline silicon:
With production technology advancements, mono-
crystalline silicon’s cost has declined, narrowing its
price difference against polycrystalline silicon and
increasing its market share.
✔✔ Application of passive emitter and rear cell (PERC)
technology: PERC is a compatibility technology that
can be superimposed on silicon wafers to improve so-
lar cell capacity. Further, monocrystalline silicon wa-
fers possess better superposition properties than poly-
crystalline silicon wafers, motivating major solar cell
manufacturers to gradually increase PERC production.
✔✔ Application of N-type double-sided modules: N-type
double-sided modules have better performance, pro-
viding significant technology and industrialization
improvements. This should help the production of N-
type modules to expand as well.
✔✔ Protection of intellectual property: China may have
greater difficulty incorporating foreign technologi-
cal advances.
Enterprise
✔✔ Enterprise overhaul and reorganization: Because of
the decline and delay in PV subsidies, some small
companies with high operating costs and fragile capi-
tal chains have been affected. Hence, enterprises and
capital will be restructured or withdrawn; consequent-
ly, market concentration will increase, with the oligar-
chic effect becoming increasingly prominent.
✔✔ Reduction in enterprise production costs: Given
technological advances and industry integration,
the PV companies’ production costs will be further
reduced. In addition, given the quota limits, the
enterprises face increased competition from indus-
trial peers, motivating them to continue reducing
their expenses.
✔✔ Cross-sector enterprise development: Considering the
high profits from producing single products and the de-
cline in PV subsidies, PV enterprises will vertically bun-
dle by further expanding their business scopes, becoming
integrated across the upstream, midstream, and down-
stream of the PV industry.

• Clear Targets
• Follow the Reform Principles of Electricity Price and Subsidy Rates
Ensure the Same Targets • Adopt a Gradual Price-Reform Implementation Path

• Identify What Needs to Be Improved

Propose Designs for Prices • Optimize the Subsidy Mode Based on the Existing One
and Subsidy Rates • Clear the Values and Costs of the Considered Renewable Energy

• Analyze Consumers’ Electricity Bills

• Defer the Investment Costs
Impact Analysis
• Consider the Sunk Costs

• Developers • Power Enterprises

Stakeholder Consultation • Representative Consumers • Independent Third Parties

• Regional Pilot Projects

Trial Implementation • Transition Between Market Access and the Market Exit

Extension

figure 6. The rate-making flowchart.

58 ieee power & energy magazine may/june 2020

Market
✔✔ Gradual enhancement in electricity-market transac-
tions: With the expansion of PV power generation,
China’s increasing electricity consumption will pro-
mote the construction of the retail power market.
✔✔ Huge distributed-PV market space: China’s distribut-
ed PV systems constitute a low percentage of its ener-
gy-production mix, and the country’s policy support is
gradually shifting to distributed PV systems that will
provide a broad domestic market in the future, espe-
cially for home-based systems. Hence, as the technol-
ogy continues to advance, distributed PV systems will
likely exhibit major growth.
✔✔ Opportunities and risks for overseas market: Al-
though overseas markets are large, international trade
is not encouraging for Chinese PV companies since
a trade war is beginning to spread and the free trade
concept is weakening; thus, some risks exist in the
international arena. Protection of foreign intellectual-
property rights could also slow the development of PV
technologies in China.
Subsidy-Rate Formulation and
Renewable Energy Regulation
Based on the previous analysis of the PV industry’s situation
and future trends, it can be seen that the design of tariff and
subsidy rates is important for determining the investment in
and operation of renewable energy power generation. In this
section, we discuss key issues concerning renewable energy
price and subsidy-rate formulation and regulation, including
the formulation process of price policy, to provide references
for the future design of renewable energy prices and subsidy
rate mechanisms.
Formulating Price and Subsidy Rates
Based on international experience, the design of electricity
prices and renewable energy subsidies sends different signals
to power consumers and related industries and affects their
investment decisions and operation modes, which impacts
renewable energy development. A comprehensive formula-
tion for tariffs and subsidy rates can be made to meet the
specific objectives of the policy.
The National Development and Reform Commission
and the National Energy Administration lead the process of
formulating prices and energy-subsidy rates. Although the
details vary by region, the overall process (as presented in
Figure 6) should be similar. The first two steps—reaching
consensus on the electricity-reform objective and selecting
the electricity-rate design to achieve that objective—form the
core and basis of the formulation process of renewable energy
generation price and subsidy rates. Although there are several
possible compensation strategies and methods, principles
for designing tariffs and subsidies can guide the effort and
ensure that it does not deviate from the expected results. An
important task for regulators is to decide how to follow the
may/june 2020 ieee power & energy magazine 59
principles while balancing the interests of all parties. Several
factors should be considered while recognizing the tradeoffs:
✔✔ cost analysis
✔✔ results orientation
✔✔ consumer orientation
✔✔ policy transparency
✔✔ decision making
✔✔ fairness
✔✔ stability
✔✔ gradual implementation.
Ultimately, the regulator must properly incorporate these
principles and formulate electricity prices and subsidies
that benefit distributed-generation users as well as the over-
all grid system. After setting targets for reforming renew-
able energy generation prices, regulators need to propose
potential tariff designs and conduct a deep analysis. The
ideal subsidy mechanisms should be based on calculations
of the current market situation and evolve continuously
through time. Rates can be decomposed according to dif-
ferent components: attributes (such as electricity, capacity,
and transmission and distribution costs), time (the electricity
price varies), and space (electricity demand changes across
locations on the grid). Therefore, before choosing specific
electricity price and subsidy mechanisms, regulators must
clearly understand which components should be optimized.
The U.S. National Renewable Energy Laboratory com-
piled a list of electricity price reform options that have been
implemented or are being considered by regulators, includ-
ing net billing, fixed charges, non-bypassable charges, mini-
mum bills, time-varying rates, residential demand charges,
value of solar tariff, customer self-supply, and so forth. It
made a detailed comparison of the application scope, advan-
tages, and challenges of those mechanisms to which regula-
tors in China and other countries can refer in light of their
respective national conditions.
With the targets set and rate designs selected, a good founda-
tion has been laid for the rate reform. Subsequent impact analy-
sis, consultation, and implementation stages can be designed
to validate and continually improve the rate design. Therefore,
these three stages are related to multiple iterative optimiza-
tions of the rate design. Soliciting opinions is an important step
throughout the process, and considering the involves parties’
views is essential for successfully implementing reforms.
Suggestions for Rate Making and Supervision
Adopt a Gradual Rather than a Comprehensive
Reform-Implementation Path
Without the necessary foundation, rashly choosing an overly
detailed electricity-bill mechanism may not achieve the
desired results. It is also challenging for project developers
to adjust their business models and user purchase methods
to respond to the rapid development and implementation of
new electric rates. Financing distributed energy resource
projects becomes even more difficult when the electricity
 fees and compensation mechanisms change too much, caus-
ing an uncertain impact on income and electricity costs. A
step-by-step transition to the ultimate goal is more practical
than immediately selecting a highly sophisticated electric-
ity-billing mechanism.
Gradual implementation can ensure the certainty and sustain-
ability required by developing industries, such as PVs, while con-
tinuously improving the precision of projected electricity prices.
Therefore, the ideal progressive electricity reform should set a
five-to-10-year target and plan, which helps related industries
develop steadily. At the same time, the gradual reform should also
allow enough time for testing and fault tolerance to help regula-
tors understand the electricity-pricing mechanism’s impact and
make timely adjustments before full implementation.
Improve Transparency
Increasing the transparency of the rate and compensa-
tion development process can achieve multiple objectives,
including providing more information about the price set-
ting, facilitating the work of regulators and third-party agen-
cies, and supporting reasonable consumer decisions about
investing in and using PVs. Transparency is about the open-
ness of data and methods as well as the development process.
Improve Regulatory Foresight and Stakeholder
Engagement
Regulators should take the initiative to raise and solve prob-
lems, anticipate and identify potential issues in advance, and
propose solutions through electricity-price-design analysis and
stakeholder participation. This forward-looking role enables
them to drive management transformation and achieve better
policy implementation in a systematic, strategic, and collab-
orative manner. In addition, the participation of all stakehold-
ers has always been the cornerstone of the regulatory process,
and it therefore requires more active, in-depth, and collabora-
tive participation.
Summary
China’s PV industry restructuring in 2018 resulted in the
sector’s transformation. The industry is currently under-
going an even greater transition from public subsidies to a
market orientation, although the government still enforces
provincial targets for renewable development.
After the implementation of the 531 New Policy, China’s PV
industry faced dramatic changes because the new policy strictly
controlled and weakened the extent of subsidies. Through anal-
ysis and elaboration, this policy will likely encourage develop-
ment and consolidation in the sector. As a result of the intro-
duction of fair feed-in tariffs, the PV industry will be initially
weakened by the subsidies policy, but industry development is
anticipated to gradually stabilize. During the next two years,
the annual growth of installed PV capacity in China will hold
at approximately 40 GW, the market will be further integrated,
the issues of manufacturing overcapacity and electrical-hosting
limitations will be addressed, and innovation will rise.
60 ieee power & energy magazine may/june 2020
Overall, the 531 New Policy will have a positive impact
on the long-term development of the PV industry. To enhance
the competitiveness of the country’s PV power generation and
achieve fair feed-in tariffs as soon as possible, the initiative
will help to improve PV manufacturing efficiency, lower elec-
tricity prices, and close outdated production facilities. Enter-
prises will accelerate the development of overseas markets,
where the PV upstream and midstream market demand will
improve. Through this reform, the transformation of China’s
PV industry from quantity to quality will be accelerated.
Acknowledgments
This work was supported by The National Social Science
Fund of China (grant 19ZDA081) and Fundamental Research
Funds for the Central Universities (grant 2018ZD13).
For Further Reading
“Renewable energy policies in a time of transition,” Interna-
tional Renewable Energy Agency, Abu Dhabi, United Arab
Emirates, 2018. [Online]. Available: https://www.irena.org/
publications/2018/Apr/Renewable-energy-policies-in-a-time
-of-transition
“Renewable energy statistics 2019,” International Renew-
able Energy Agency, Abu Dhabi, United Arab Emirates, 2019.
[Online]. Available: https://www.irena.org/publications/
2019/Jul/Renewable-energy-statistics-2019
“Renewable energy policies and measures of China,” Inter-
national Renewable Energy Agency, Abu Dhabi, United Arab
Emirates, 2019. [Online]. Available: https://www.iea.org/
­policiesandmeasures/renewableenergy/?country=China
“Renewable power generation costs in 2018,” Interna-
tional Renewable Energy Agency, Abu Dhabi, United Arab
Emirates, 2019. [Online]. Available: https://www.irena.org/
publications/2019/May/Renewable-power-generation-
costs-in-2018
“Evolving distributed generation support mechanisms: Case
studies from United States, Germany, United Kingdom, and
Australia,” National Renewable Energy Laboratory (NREL),
Washington, D.C., 2017. [Online]. Available: https://www.nrel.
gov/docs/fy17osti/67613.pdf
PV InfoLink. [Online]. Available: https://en.pvinfolink
.com/index.php
Biographies
Houqi Dong is with North China Electric Power University,
Beijing.
Bo Zeng is with North China Electric Power University,
Beijing.
Yuqing Wang is with North China Electric Power
University, Beijing.
Yingxin Liu is with the North China Electric Power
University, Beijing.
Ming Zeng is with North China Electric Power
University, Beijing.
p&e

H
HISTORICALLY, UTILITIES HAVE STUDIED HOW THEIR CUS-
tomers use electricity through load research, the practice of studying the
By Curt Puckett,
characteristics of electric loads. The outcomes of load research play a critical Craig Williamson,
role in cost-of-service studies and rate design as well as in many other activi-
ties, such as transformer sizing, demand-side management, load settlement, Claude Godin,
and load forecasting. Despite its existence for nearly a century (as described Will Gifford,
in the “History” column in this issue of IEEE Power & Energy Magazine),
load research is not a widely known subject in the power engineering com- Jonathan Farland,
munity. This article describes the current practices and explores the future of
load research and expansion of interval load analytics in the utility industry.
Tom Laing,
and Tao Hong

©ISTOCKPHOTO.COM/Z_WEI

Utility Load
Research
The Future of Load Research Is Now
Digital Object Identifier 10.1109/MPE.2020.2972668
Date of current version: 17 April 2020

may/june 2020 1540-7977/20©2020IEEE ieee power & energy magazine 61

 Sample stratification seeks to divide a population into groups that
are similar within the individual segments exhibiting characteristics
related to the measure being estimated in the study.
The Load Research Lifecycle
Figure 1 shows the load research lifecycle. We start in the upper
left corner with a “need for information.” Load researchers
should work with internal clients to develop a preliminary plan
that specifies the goals, objectives, and analytical approach for
fulfilling clients’ needs.
Next, load researchers are responsible for securing or
creating an appropriate population frame. For classical load
research, this comes in the form of the population billing
data known for all customers on the system. The customers
are typically segmented by rate class or other domains of
interest. For some domains, such as customers with rooftop
solar or electric vehicles (EVs), the construct of the popu-
lation can be more complex and may require combining
multiple internal and/or external sources of data to create an
appropriate framework for analysis.
After identifying the measures of customer usage to be
used as the focus of the research study and the populations
for which the measures need to be reported, load research-
ers develop an appropriate sample design, selecting the
minimum number of customers needed for estimating a
population-level measure constrained by a desired level of
precision of the estimate. Load researchers are responsible
for making decisions with respect to the stratification strat-
egy for the sample. Sample stratification seeks to divide a
population into groups that are similar within the individual
segments exhibiting characteristics related to the measure
being estimated in the study. The greater the similarity of
the customers within the strata, the smaller the number of
customers needed to be sampled. The particular features of
Need for Population
Data
Information Preliminary
Plan
Population
Model Load Research
Data
Lifecycle
Expand
Results Sample
Data
figure 1. The load research lifecycle.
62 ieee power & energy magazine may/june 2020
stratification include identifying the variables to use, such as
seasonal and annual use, peak demand, and load factor, and
the key sample design planning variables of interest, such as
system and class peaks.
Next, decisions are made with respect to the confidence
level, desired relative precision, and anticipated variability
of the target variables. These decisions dictate the sample
size requirements. This often becomes an iterative process,
with tradeoffs necessitated by the available budget. Finally,
having settled on an appropriate sample design, the load
researchers guide the selection of the sample.
For some projects, such as an analysis of demand
response, technology assessments, and pricing experiments,
the sampling plans will need to be developed inside an
experimental design paradigm. The load researchers provide
guidance and governance on the appropriate experimental
design to use, potentially randomized control trial, within-
sample comparison, test versus comparison, preparticipation
versus postparticipation, and so forth.
Historically, the next step has been to work with the
metering department to procure and install pulse-initiat-
ing meters with load recorders. This includes determining
an appropriate data collection strategy, such as handheld
devices, modems, or a fixed network. Due to the high cost
of equipment, equipment installation, and data collection,
load researchers had to rely on interval data collected from
a limited sample of customers in the population of interest.
However, in today’s fast-approaching advanced metering
infrastructure (AMI) world, these constraints are fading
away, as we are seeing utilities deploying “super samples” or
even “census” representations.
As we transition to full AMI, sample data present several
unique challenges, such as storage and how to conduct vali-
dation, editing, and estimation (VEE). Added to the chal-
lenge of storage is the communication bandwidth needed
to transmit interval data from the field. Utilities often find
the need to compromise load interval granularity (5 min,
15 min, hourly, and daily) to ensure that they have the capac-
ity to reliably transmit data while maintaining bandwidth
for other operational priorities. Data quality issues arise
from any number of sources, including communication
problems, database malfunctions, and human errors. These
issues can affect the data capture and its eventual analysis,
if not handled properly.
Some simple sanity checks such as ensuring consistency
with the correct time stamp, interval orientation (begin-
ning or end of period), units of measure, and frequency
of intervals are all important first steps in facilitating data the current roles played by the load research department in
integrity. The ability to compare the total of the interval load each of the functions listed in Table 1.
data to the corresponding billing month data will always
serve as the first line of defense. However, additional issues Cost of Service
such as zeros, valleys, gaps, and spikes still need to be moni- The load research department has been responsible for devel-
tored and corrected or eliminated appropriately. There are oping the hourly or subhourly (i.e., 15-min) load profiles for
a host of alternative strategies that the load researcher can each rate (or tariff) class of interest. The load profiles are
employ, but the automated VEE processes of meter data used to develop the required factors that distribute cost alloca-
management systems (MDMSs) may not be enough. Load tions to the various customer classes. The allocation factors
researchers would be well served to check and potentially can be as simple as the class or system peak demand of the
re-estimate data cleaned by an MDMS using more robust class or can be a mixture of variables. In the past, utilities
strategies such as regression-based analysis. have reported using various cost-allocation factors, such as
Once the sample data are of the highest quality possible, the 12 coincident peak (12-CP) demands, the four-summer
then the analysis, often referred to as the expansion of the CP (4-SCP) demands, the on- and off-peak energy use (utility
sample to the population, can begin. It has always been the defined on-peak periods), the class contribution to the top 50,
purview of the load researcher to ensure that proper weights 100, or 200 h of system load, and something as elaborate as a
are applied to the sample data to protect the integrity of the 28-h allocator, which uses the 4-SCP peak periods, including
results. Of course, load researchers should always calcu- the three hours before and after each system peak.
late and report variances, standard errors, and relative and
absolute precisions along with their estimates. These form
the very foundation of the “results” that can be summarized table 1. The summary statistics for a recent load
tabularly and graphically in a wide array of forms during the research survey on the importance of load research
(1 = “not at all important” to
reporting cycle. 5 = “critically important”).
The last step of our load research lifecycle is “model,”
where the parameters from the current study are retained Business Number Number
Functions Average Median of 5s of 4s
for future use. This includes the coefficients of variation
and/or error ratios for the various demand variables of inter- Cost of service 4.8 5 14 2
est and provides insight and guidance for use in planning Rate design 4.7 5 14 1
future studies.
Fulfilling internal 3.9 4 5 7
Historically, the load research lifecycle took 18–24 data requests
months to complete, assuming that a full year of data was
Fulfilling external 3.5 4 4 5
required, when sampling was used to develop an efficient data requests
representation of the populations of interest. With the use
of AMI, assuming an ongoing retention of interval data col- Load forecasting 3.4 4 5 6
lected, the lifecycle may shrink to 18–24 days or, perhaps, Special studies 3.2 4 1 8
18–24 h, and sampling will be unbounded. This does not Demand-response 3.2 3 3 4
mean that, as an industry, we should abandon the principles planning and
of the past, namely, the use of proper population identifica- evaluation
tion and statistically valid sampling approaches; however, Key account 2.9 3 1 7
today, our samples can be bigger and cover more domains, management
allowing us to explore and use the data in a much richer form. Energy efficiency 2.9 3 1 3
planning and
Interacting With Utility evaluation
Business Functions Marketing studies 2.8 3 1 3
In a recent survey conducted by DNV GL in 2018 covering 17
T&D planning 2.8 3 1 4
U.S.-based load research departments, the respondents were
asked to rank the importance that load research plays in servic- DER planning 2.6 3 1 3
ing several utility business functions. Table 1 summarizes the Renewable 2.5 3 0 4
responses. Not surprisingly, the cost of service and rate design integration
ranked the highest. These were followed by fulfilling internal Smart grid 2.4 3 0 2
and external data requests, load forecasting, special studies, initiatives
and demand-response planning and evaluation. Fulfilling inter-
IRP 2.2 2 0 2
nal and external data requests supports the notion of the load
research department as the internal load consultant to the rest T&D: transmission and distribution; DER: distributed energy
resource; IRP: integrated resource planning.
of the utility. The following sections briefly describe many of

may/june 2020 ieee power & energy magazine 63

 Rate Design
The load research department supports the rate design func-
tion by testing various rate designs under consideration and
providing information on elements such as structural win-
ners and losers based on an anticipated new rate design. The
analysis of structural winners and losers helps to differentiate
customers who would benefit from the rate from those that
would be hurt by the rate. The load research department pro-
vides other critical insight into the load characteristics of the
proposed rate, such as on- and off-peak energy and on-peak
demand as well as their current contribution to the revenue
associated with the rate design. In addition, load research-
ers are often asked to model how likely the customers are to
change their behavior under various time-of-use (TOU) sce-
narios to predict how that behavior change may affect future
revenue. Load researchers are often consulted to help develop
and conduct randomized control trials for pilot programs that
support the testing of various rate designs.
Fulfilling Internal Data Requests
Fulfilling internal data requests is simply responding to
ad hoc data and information requests from various inter-
nal departments related to the load characteristics of cus-
tomer segments or the load impact associated with spe-
cific activities. Countless examples of this exist, such as
low-income customer use compared to non-low-income
customer use, comparing the load characteristics of cus-
tomers based on their seasonal or annual consumption
(e.g., low summer use versus high summer use or low,
medium, and high annual use), and how the load during a
particular holiday season compares to the load during the
days surrounding holidays.
Fulfilling External Data Requests
The task of fulfilling external data requests is associated
with responding to data requests from external stakeholders,
including parties not working for the utility. In the United
States, this is dominated by fulfilling formal data requests
during rate cases. The commission staff, or outside interve-
nors in rate cases, may request that the load research depart-
ment examine several things, such as how a recommended
rate impacts low-income customers.
Load Forecasting
The load research department generally conducts an
annual analysis of all customer classes of interest. These
historical load profiles can be used to develop class-level
load forecasting models. Further, load researchers can be
called upon to provide information and insight for new
emerging technologies, such as EVs, which can be incor-
porated into the forecasts. Additionally, load researchers
can aid in the design, implementation, and analysis of an
end-use load research effort to develop end-use appliance
profiles, which can be used to support a bottom-up load
forecasting methodology.
64 ieee power & energy magazine may/june 2020
Special Studies
The load research skill set is particularly useful in supporting
special studies. This may include the design and execution of
a randomized control experiment focused on a special rate
design or an enabling technology, such as smart thermostats.
The list of special studies is endless and includes studying
the load characteristics for new and evolving appliances,
such as heat pump water heaters and high-efficiency air con-
ditioners, EVs, residential or commercial solar impacts, resi-
dential or grid-side battery storage, and conservation voltage
reduction (CVR) pilot programs.
Demand-Response Planning and Evaluation
Demand response includes economic or reliability programs
designed to reduce customer consumption for short periods
of time. Load researchers often assist in determining the
demand impacts associated with demand-response program
activities through a comparison with a baseline. This may
include innovations like high-frequency water-heating con-
trols or more traditional measures such as residential air-
conditioning switches or smart thermostats. Additionally,
the load research department often works with the indepen-
dent system operator to develop appropriate baseline strat-
egies for their demand-response program offerings. It can
also support planning efforts by helping to determine the
load available for reduction at various times of the day.
Key Account Management
The load research department normally has direct access to
the interval load data for individual meters/accounts. These
can be brought together, aggregated if appropriate, to repre-
sent key accounts/customers and provided to the other internal
departments, including key account representatives responsi-
ble for, or have direct contact with, customers. Usual customer
requests include daily peak values (if on a demand charge)
and the typical day graphs used to spot irregular use patterns,
such as lights kept on at night when they should be switched
off. Customers on tariffs that include penalties for excessive
reactive power consumption are interested in reactive power
load shapes. In some utilities, customers have direct access to
their own individual interval load data. Here, the load research
department may support the development of standard presen-
tation formats and reports that allow for meaningful visu-
alization and insights or show the savings that are available
from alternative rate designs or load management strategies.
Energy Efficiency Planning and Evaluation
Energy efficiency programs are defined as activities that
reduce overall end-use customer energy consumption via the
adoption of more efficient appliances, air conditioners, light-
ing, and other energy-reduction technologies. Load research-
ers often assist the energy efficiency team with determining
the demand and energy characteristics associated with spe-
cific measures or measure bundles. This may include a for-
malized randomized control experiment or simple/complex
modeling to estimate the load-shape changes that may occur
based on customer participation. Load researchers may also
support the development of building simulation models to esti-
mate measure impacts, such as the effect of high-efficiency
air conditioners replacing their low-efficiency counterparts.
Additionally, load researchers can also support experimental
designs that estimate the energy usage patterns of new energy
efficiency initiatives, such as behavioral programs.
Marketing Studies
The load research department generally supports marketing
the same way it supports energy efficiency, namely, devel-
oping the likely load shapes resulting from new marketing
efforts. From a marketing perspective, the load research
department can provide the data needed to calculate the
profitability of specific customer segments or individual
customers as well as target information for likely partici-
pants in programs or promotions. For U.S. utilities, market-
ing is now often referred to as beneficial electrification, and
load research is beginning to help estimate the load associ-
ated with specific activities, such as switching electric heat
to heat pumps, changing natural gas homes to all-electric
homes, and adding electric forklifts.
Transmission and Distribution Planning
This role tends to focus on modeling the effect that load has
on system assets such as transformers, feeders, circuits, and
substations. This may include running simulations to show
the load characteristics on existing or planned circuits, feed-
ers, and transformers and how the diversity of load changes
with the addition of customers to a transformer. These data
can be used to identify potential overloads of assets such
as transformers as well as support the appropriate sizing
of planned assets. Additionally, load researchers may help
determine the experimental design and subsequent analyses
needed to calculate the savings from CVR.
Distributed Energy Resource Planning
Distributed energy resource (DER) planning is associated
with distributed assets like solar generation and batteries
on the system. Load researchers can provide insight into the
loading of local areas to identify the siting of microgrids or
DERs. They can also help identify how additional distrib-
uted resources, such as the increased installation of rooftop
photovoltaics (PVs), can change customer load over time.
The load research department is the keeper of the current
class load profiles that provide the benchmark for measuring
the impact of customers with new distributed assets.
Renewables Integration
The load research department provides the baseline for esti-
mating impact changes on the system as a result of renew-
able energy technologies. Load researchers may also get
involved in studying the technology directly and contribute
to the modeling of the impact under certain weather condi-
may/june 2020 ieee power & energy magazine 65
tions, such as the effect on the net load of customers with
PVs behind the meter when cloud cover increases.
Smart Grid Initiatives
Smart grid initiatives involve load researchers in bench-
marking existing loads of various customer classes of inter-
est. Additionally, the load research department is part of
the greater smart grid team, which deals with issues like
meter data management, VEE, data analysis, and visu-
alization. Depending on the smart grid initiative, other
aspects previously discussed in this article may include
energy efficiency, demand response, DERs, and renew-
ables integration.
Integrated Resource Planning
The load researchers provide foundational information for the
integrated resource planning (IRP) function in terms of class-
level load shapes as well as load-shape changes that are likely to
occur as a result of demand response, energy efficiency, DERs,
or renewable technologies. For some utilities, load researchers
are located within the IRP department, while for others, load
researchers are an integral part of the IRP team.
In addition to the aforementioned business functions with
which load researchers interact regularly, the load research
department can also audit the billing system by tracking the
meter-to-bill process, thus helping to ensure the data’s integ-
rity. Additionally, load researchers can support the calcula-
tion of unbilled revenues based on the bill-cycle sales of the
company. If the company participates in wholesale electric-
ity markets, the load research department provides the load
profiling data needed for the initial billing settlements before
customer data are fully available as well as the final settlement
for customers without interval load information.
Emerging Analytical Tasks
Although today’s load research activities continue to focus
on creating efficient sampling plans, the high cost of meter-
ing, data collection, processing, and analysis, what about the
future? Table 2 summarizes the DNV GL survey responses on
the likelihood that load research departments would engage in
the identified analytical tasks in the near future. The survey
respondents rated their responses based on a scale of 1 (not
at all likely) to 5 (very likely). Big data analytics and big data
management were at the top of the list, followed by solar-
related issues, market segmentation, and the investigation into
customer load characteristics. In addition to the ones listed in
Table 2, a few utilities indicated that they anticipated work-
ing on tasks related to EVs, distributed generation, and battery
storage in 2019.
The following is a summary of the various analytical
activities:
✔✔ Big data analytics refers to the processing and analy-
sis of large blocks of data, such as using individual
load profiles aggregated to domains of interest. This is
different from taking a sample of customers from the
 domains of interest and extrapolating the sample and ✔✔ Transformer sizing is a classic analysis that develops
introducing third-party customer demographics into strategies for sizing transformers based on the diver-
the analysis of AMI data to segment customers into sity of customers served by a transformer.
domains of interest. ✔✔ Theft detection focuses on using interval load data to
✔✔ Big data management represents the challenges associ- identify unanticipated changes in a customer’s load
ated with managing the large amount of data coming pattern; they can be the result of utility failure or de-
from AMI. One challenge is how to relate all the vari- liberate action taken by the customer.
ous sources of large data sets into a manageable rela- ✔✔ Revenue protection is closely aligned with theft de-
tional model that allows for access by the various ana- tection and associated with any analyses that look to
lytical tools used today, such as SAS, Python, and R. improve the financial performance of the utility.
✔✔ Solar-related issues focus on understanding the effect ✔✔ M&V 2.0 is a growing area of demand-response mea-
of PVs on residential and commercial customer loads, surement and verification that uses interval load data
including forecasting behind-the-meter solar in real to estimate the savings associated with program inter-
time to better predict net loads. vention in real time.
✔✔ Market segmentation classifies customers based on ✔✔ Cryptocurrency identification recognizes customers
some observed attributes. The attributes could be con- running a server farm mining for cryptocurrency at
tained in the customer billing system, characteristics a time when utilities are beginning to explore what
of the interval load, or based on a third-party data drives residential customers with a high base load and/
source matched to utility billing records. or high load factor.
✔✔ Customer load characteristics use the interval load ✔✔ Indoor agriculture investigates the likely load growth
data to isolate customers with a high load factor or a resulting from indoor agriculture activities such as
high base load. The high load-factor customers have cannabis production using hydroponics.
a high value of the ratio of average demand to maxi- As an industry, we still have not advanced very far
mum demand. Typically, they can be put on a more beyond analyzing loads at the building or household level.
favorable rate because they provide a better utilization Although interest in end-use research persists, the business
of utility assets. The high base-load customers could case for collecting end-use data continues to be difficult to
be good candidates for some utility intervention, such justify due to the high cost of data collection. Some progress
as an energy efficiency program. is being made in the disaggregation of the whole building
or household load data into end-use components, such as
air conditioning, lighting, and refrigeration. However, with
table 2. The emerging analytical tasks
for load researchers. more and more utilities rolling out AMI, applications such
as the evaluation of direct load control seems to be mov-
Number Number
Tasks Average Median of 5s of 4s ing away from requiring end-use metering to being based on
AMI shapes. As an example, PJM requires utilities to pro-
Big data analytics 3.6 4 6 6 vide whole-facility metering data, where available, to prove
Big data 3.5 4 4 6 the value of the utility demand-response programs.
management
Solar-related issues 3.5 3 6 2 The Next 10 Years of Load Research
What will load researchers be spending their time on 10 years
Market 3.4 4 5 6 from now? Given the continuation of the utility/regulator
segmentation
model, the increasing penetration of AMI and DER, and the
Customer load 3.4 3 6 2 advancement of data technologies, we highlight some ever-
characteristics
expanding load research applications.
Transformer sizing 2.2 1 2 2
Theft detection 1.9 1 2 0 Cost of Service and Innovative Rate Design
As listed in Table 1, cost of service and rate design are cur-
Revenue 1.9 1 2 0 rently the two most important load research applications,
protection
and we expect this to continue in the context of increased
M&V 2.0 1.8 1 1 0 penetration of DERs and the availability of interval data.
Cryptocurrency 1.6 1 1 1 Collectively, load researchers will take fresh looks at these
identification core applications while getting used to the novel position
of having access to interval data from nearly all of their
Indoor agriculture 1.6 1 1 0
(cannabis) customers. Because of the increased volatility in customer
demand at the meter associated with high penetrations of
M&V: measurement and verification.
DERs, an important application of load research in the near

66 ieee power & energy magazine may/june 2020

term will be balancing the accurate accounting for cost of
service and meeting customer expectations of choice and
flexibility in rates.
If we asked utility professionals 10 years ago what they
expected rates to look like today, TOU and other innovative
rates would have been much more prevalent than what is cur-
rently in use. Innovative rates like TOU, critical peak pric-
ing, peak-time rebates, and even residential demand rates
were the answer for truing up cost of service with customer
demand in real time. Although there is still talk about man-
datory TOU pricing in some jurisdictions, the actual num-
ber of customers on these innovative rates continues to lag
behind the deployment of AMI meters in the United States,
in large part because they are often perceived as confusing,
scary, or punitive to the consumer. The big challenge here is
offering customers choices while accurately and equitably
allocating costs. Ultimately, economic efficiency requires
having all customers, especially those cross subsidized by
others, on a time-varying rate. This issue will continue to
plague rate designers, who will need data and analytical sup-
port from load researchers well into the future.
Assessing Emerging Technologies
The advent of new technologies will continue to pres-
ent challenges to the utility industry. A well-known one is
behind-the-meter solar generation. There are many issues
that arise with increased customer generation on a grid that
was designed for one-way delivery of power, but the load
researcher’s primary concern will be with the measurement
of customer load and generation. This presents a challenge,
as the actual customer energy use and load shape are a mys-
tery. The utility must deliver power, but the demand for
power varies not only based on customer usage, as it always
has, but also on customer generation. To complicate matters,
solar generation can be volatile. With more and more PVs
being installed, the accurate measuring and forecasting of
customer PV generation is crucial to maintaining the stabil-
ity of the electric grid.
Today, localized battery storage is expanding its foot-
print. As an example, the Glasgow Electric Plant Board
installed smart battery systems at 165 households in 2016
as part of a virtual power plant assessment. The battery
systems were charged during off-peak hours using grid
energy. During demand-response events, a discharge
command was issued to partially or fully carry the site’s
load. The battery system also served as an uninterruptible
power supply that provides power to the home within mil-
liseconds of a power outage. Figure 2 presents the impacts
on a July event day when the batteries were charged dur-
ing the early morning hours and discharged during the on-
peak period.
Another example is the electrification of the transporta-
tion industry. EVs are being viewed as both an opportunity
and a challenge by the electric utility industry. EVs will cer-
tainly increase electricity sales but they will also have a sig-
nificant impact on the distribution network, from the sizing
of local transformers to the build-out of a dc fast-charging
infrastructure. As consumer confidence and the EV range
extends, the popularity of EVs and the energy use per EV
will increase. An important element for utility planning and
a challenge for the load research community is the develop-
ment of a realistic and diversified load profile representing
the current and future state of EV utilization.
In a 2016 project at Dominion Energy, data loggers were
installed on in-home chargers to collect direct usage data
from the EV charging station. The charging stations were

26 July 2016
4
System Peak Hour
3.5
T1: Treatment Group
3 C1: Comparison Group
kW Per Customer

2.5

2
VPP Impact
1.5 Battery Charge

0.5

-
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Hour Ending
(a) (b)

figure 2. The residential grid-side battery storage impacts: (a) the storage battery and (b) the results of the impacts when
the batteries were charged in the early morning and discharged during the on-peak period. VPP: virtual power plant.

may/june 2020 ieee power & energy magazine 67

 a combination of level 1 and 2 chargers. Figure 3 shows the
EnergyPrint of the EV-only load following a super off-peak
TOU rate program. The vertical EnergyPrint shows the time
on the y-axis, the day of the year on the x-axis, and the load
as a color gradient, with low levels of load in the blue-to-
black spectrum and high levels of load in the yellow-to-white
spectrum. The super off-peak generation plus distribution
charge was US$0.00695/kWh compared to the on-peak rate
of US$0.133. The transmission charge was the same for both
time periods: US$0.00970/kWh. Most of the energy (81%)
was consumed during the 1–5 a.m. time period. The average
annual use shown in the figure was calculated to be just over
4,000 kWh, with a noncoincident peak demand of 5.08 kW.
Predictive Load Analytics for Distribution
Operations and Planning
Even though population- and class-level load shapes will still
be relevant, load researchers will increasingly provide analysis
on localized subsets of customers, such as individual circuits, to
their internal clients within the utility. Importantly, they will be
expected to provide analytics relating to present or future condi-
tions, rather than historical ones. These shifts will help distribu-
tion operators manage the grid more effectively by examining
future system conditions. To serve this need, load research-
ers will increasingly employ load forecasting models, either
at the individual customer level or at various ad hoc levels of
desired aggregation. This will allow for a better understanding
of loading issues in the distribution grid, proactively switching
loads for maintenance ahead of outages or after asset failures,
developing improved replacement and upgrade strategies, and
increasing transparency at the transmission and distribution
(T&D) interface. Figure 4 shows a localized distribution node
load forecast and its available capacity plots overlaid in green
and white, respectively. Offloading opportunities for portions
of the forecasted load from switching are shown in the bottom
right panel of the figure.
One planning strategy gaining popularity today is called
nonwires alternatives, namely, the grid investments or proj-
ects that use nontraditional T&D solutions such as distrib-
uted generation, energy storage, energy efficiency, demand
response, and grid controls that defer the need for distribu-
tion equipment upgrades in specific grid areas. It will be
paramount that the load research community be prepared to
0013: EV TOU Group, EV Load, EV
5.08 24
28
12
6
0.01 0
kW J F M A M J J A S O N D
figure 3. An EV-only charging station load on the super
off-peak TOU rate.
68 ieee power & energy magazine may/june 2020
provide the information and insight needed to support these
pending planning strategies.
Customer Targeting Using
Segmentation and Clustering
Utilities need to improve customer engagement through
analytics. Segmentation and clustering are two strategies
for identifying customers with similar attributes. Segmenta-
tion refers to a process that divides units into similar groups
and has been in use for years. With better customer over-
lays from improved customer records systems, the ability
to group customers and determine their hourly load varia-
tions from segment to segment are nearly unlimited. Some
common secondary sources include utility business intel-
ligence databases, energy efficiency program-tracking sys-
tems, demand-response systems, census tract data, and tax
records. Machine learning algorithms can be used to iden-
tify customers with specific load characteristics, helping
to target marketing, pricing, energy efficiency, or demand-
response programs. Alternatively, the clustering process
finds similarities, in that load researchers can use clustering
algorithms to drive segmentation, with the differences based
on customer energy use or load shapes.
Figure 5 illustrates a simple strategy based on billing
determinants, e.g., low, medium, and high seasonal usage.
Another frequently used algorithm in practice, k-means
clustering breaks a population into k groups where the
within-group variance is minimized. These algorithms can
be designed to focus on customer load characteristics dur-
ing specific periods, such as largest contributor during peak,
highest morning rise, and longest valley. They can also be
used to identify demand-response customers not contribut-
ing the anticipated load relief during event calls.
The Key Factors Needed for Load
Research to Continue to Thrive
As the primary study of customer demands over time, there
are several important foundational areas for load research to
ground itself as it advances to its future state.
Improving Data Quality
The future of load research will use considerable amounts
of interval load data from individual customers and super-
visory control and data acquisition meters positioned
throughout the grid. This does not mean that all AMI data
will be flawless. The experience to date has shown defi-
ciencies in automated systems used today and an almost
endless list of possible mistakes occurring throughout the
data-gathering process, ranging from initial AMI readings
to the formation of a data storage repository. The analyt-
ics process will continue to require careful management of
data inputs to ensure proper quality. Certainly, given the
volume of data being processed, we will need new, auto-
mated ways to track completeness and quality, positioning
data scientists to efficiently monitor and take appropriate
figure 4. Circuit-level load forecasts for proactive switching.

Residential A /C and Heating Loads

Are Very Clear in Some Segment Profiles Normalized Load-Duration Curves

All 100
HSMW HSHW Residential 90
80
0.47 5.37 0.65 9.63 0.36 5.14
% of Max kW

kW kW kW 70
0 6 12 18 24 0 6 12 18 24 0 6 12 18 24 60
M M M 50
40
A A A
30
20
J J J 100 80 60 40 20 0
A /C J J J Percent of Time Value Is Below

A A
S S
O O
N N
Heating J J
A
HSMW HSHW
S
O
Residential high winter load
N customers have a poorer load
factor than their medium winter
counterparts . . . this is an
J indication that they likely
have a higher cost to serve.

figure 5. Clustering employing seasonal billing use determinants. A/C: air conditioning; HSMW: high summer medium
winter; HSHW: high summer high winter.

may/june 2020 ieee power & energy magazine 69

 actions to automatically correct issues quickly. Forward-
looking utilities with high expectations for load research
and analytics will make significant investments in software
technology, hardware capabilities, and staffing to bolster
data quality so that the future advanced applications do not
fail due to a lack of reliable input data.
High-Value Research Questions
Leadership is the biggest determining factor among utilities
that make significant returns on their investment in analyt-
ics and those that fail to achieve tangible benefits. Utilities
should position managers who can pinpoint high-value goals
and challenge load research and other teams to meet those
goals using the appropriate technology and methodologies.
Importantly, these leaders do not always have to understand
the minutia of the solution, but they must lead the formulation
of the value proposition.
Do Not Forget Where We Came From
It is a cliché, but as we move toward increasingly available
smart meter data and the new technologies used to harness
it, we should not forget the lessons learned from the past. In
many applications, it will make more sense to work with a
robust, statistically designed sample for a population rather
than a census, given the potential for bias from unknown
data quality issues. An agile load research group of the future
will consider the efficient solutions used in the past when
determining the best approach to a problem. Using fully
deployed and perfectly functioning AMI, various class and
domain load profiles can be easily calculated directly from
the interval data collected for each account in the popula-
tion. In practice, however, the AMI data are incomplete and
contain errors. There will always be some accounts missing,
and many others may have only partial data. As a result, a
simple tally of all the available data will usually fall short
of the total hourly load of the entire class or domain. If the
accounts with usable AMI data are materially different than
those with missing AMI data, the estimation calculation will
not accurately result in the true total load of all accounts for
the class or domain of interest, which could introduce bias or
create misleading estimation results.
Take Advantage of Powerful Analytics Tools
Although we use the lessons of the past, moving forward,
we must also take full advantage of newer and more innova-
tive methods. The list of these, such as artificial intelligence,
machine learning, and advanced statistical techniques, is
familiar. Load researchers must keep abreast of develop-
ments in the data science realms in addition to the happen-
ings in the utility industry.
Attract Top Talent for Both Utility and
Supporting Vendor Communities
For the utility industry to be among the leaders for getting
high value from data-driven applications, we must attract and
70 ieee power & energy magazine may/june 2020
retain top talent. Data scientists, statisticians, electrical engi-
neers, computer scientists, and behavioral economists are
among the disciplines needed by utilities and vendors, and
the competition for this talent continues to be fierce. Bank-
ing, finance, and high-tech sectors are also vying for profes-
sionals with the experience and expertise to mine value from
big data.
For Further Reading
“Global best practices and trends in load research,” DNV
GL, Høvik, Norway, 2020. [Online]. Available: https://www
.dnvgl.com/publications/global-best-practices-and-trends
-in-load-research-166325
W. G. Cochran, Sampling Techniques, 3rd ed. New York:
Wiley, 1977.
Y. Wang, Q. Chen, T. Hong, and C. Kang, “Review of
smart meter data analytics: Applications, methodologies,
and challenges,” IEEE Trans. Smart Grid, vol. 10, no. 3, pp.
3125–3148, May 2019. doi: 10.1109/TSG.2018.2818167.
T. Hong, D. W. Gao, T. Laing, D. Kruchten, and J. Cal-
zada, “Training energy data scientists: Universities and in-
dustry need to work together to bridge the talent gap,” IEEE
Power Energy Mag., vol. 16, no. 3, pp. 66–73, May–June
2018. doi: 10.1109/MPE.2018.2798759.
C. Puckett, “Smart energy technologies: An application
using residential battery storage,” presented at the 2018 Int.
Energy Policy & Programme Evaluation Conf., Vienna,
Austria, 2018. [Online]. Available: https://energy-evaluation
.org/wp-content/uploads/2019/06/2018-puckett-paper
-vienna.pdf
“Report: Evaluation of the establishment of uniform pro-
tocols for measuring, verifying, validating, and reporting the
impacts of energy efficiency measures; a methodology for
estimating annual kilowatt savings; and a formula to calcu-
late the levelized cost of saved energy for energy efficiency
measures implemented by investor-owned electric utilities
in the Commonwealth,” Commonwealth of Virginia State
Corporation Commission, Richmond, Dec. 1, 2016. [Online].
Available: https://www.scc.virginia.gov/comm/reports/2016_
EM&V_Report.pdf
Biographies
Curt Puckett is with DNV GL, Høvik, Norway.
Craig Williamson is with DNV GL, Høvik, Norway.
Claude Godin is with DNV GL, Høvik, Norway.
Will Gifford is with TROVE Predictive Data Science,
Buffalo, New York.
Jonathan Farland is with TROVE Predictive Data Sci-
ence, Buffalo, New York.
Tom Laing is with North Carolina Association of Elec-
tric Cooperatives, Raleigh.
Tao Hong is with the University of North Carolina at
Charlotte.
 p&e
 Discover PES University
Propel your Professional Growth with
Resources from IEEE Power & Energy Society
Whether you wish to earn continuing education credits or simply advance your knowledge of a particular
topic, PES University has a program for you. Choose from online and in-person options:

BEGINNER INTERMEDIATE EXPERT

PLAIN TALK ABOUT WEBINARS TUTORIALS

THE ELECTRIC
POWER SYSTEM
This short course in power
system fundamentals is
taught in-person and online
and is designed as a rapid
introduction to the technical
aspects of the electric power
system. Geared towards
non-engineers and
engineers new to the power
industry, this course covers
power system basics as well
as transmission and
distribution systems.
These convenient 1-hour
online seminars – offered
both LIVE and
on-demand – are intended
to introduce attendees to a
new concept or to explore a
focused technical topic,
providing a springboard for
more in-depth learning.
These focused courses
explore the technical aspects
of a topic in greater depth
than is possible with a single
webinar. Whether attended
in-person at one of the
many PES conferences or
online via web technology,
technical tutorials provide
attendees with both the
background and practical
knowledge needed to apply
in their work.

Most PES University courses are eligible for continuing education credits.
Visit ieee-pes.org/pes-university to learn more

Digital Object Identifier 10.1109/MPE.2020.2983526

 history
U
UTILITIES STUDY HOW THEIR
customers use electricity across time
through load research. The outcomes of
load research play an important role in
cost-of-service studies and rate design
as well as many other activities, such as
transformer sizing, demand-side man-
agement, load settlement, load forecast-
ing, and distribution and resource plan-
ning. Despite its existence for nearly
a century, load research is not a well-
known subject in the power engineering
community. Here, we discuss the histo-
ry of load research and its current prac-
tice and applications. In another article
in this issue of IEEE Power & Energy
Magazine, “Utility Load Research: The
Future of Load Research Is Now,” we
provide a look into the future of load
research as utilities increase their in-
terval load data-collection capabilities
through the use of automated metering
infrastructure (AMI).
Association of Edison
Illuminating Companies
Load Research Committee
Load research got its start in the United
States in the late 1930s as a way to better
understand electric customers and their
contribution to an ever-expanding base
and unprecedented growth. Interesting-
ly, the Association of Edison Illuminat-
ing Companies (AEIC) has had a Spe-
cial Committee on Load Studies dating
back to 1938. Table 1 lists those who
were members of this committee from
Digital Object Identifier 10.1109/MPE.2020.2971845
Date of current version: 17 April 2020
72 ieee power & energy magazine may/june 2020
Curt Puckett and Craig Williamson
utility load research
a look at our past
Load research allows utilities to examine their customers’ use of electricity,
either in total or by individual end uses, and requires numerous disciplines in-
cluding engineering, statistics, computer programming, and marketing. In this
issue’s “History” column, we examine utility load research, with a look at the
past from the first days of load research in the 1930s up to the rapidly advancing
future with automated metering infrastructure.
We welcome Curt Puckett and Craig Williamson, both with DNV GL, to
these “History” pages of IEEE Power & Electronics Magazine.
John Paserba
Associate Editor, “History”
table 1. The members of the AEIC Special Committee on Load Studies
(1938–1943).
• S.W. Andrews, American Gas and Electric Service Corporation
• W.E. Barbour Jr., Boston Edison Company
• C.W. Bary, chairman, Philadelphia Electric Company
• A.D. Caskey, Public Service of Northern Illinois
• H.A. Enos, American Gas and Electric Service Corporation
• E.J. Fowler, Commonwealth Edison Company
• J.R. Gardner, Central Hudson Gas and Electric Corporation
• R.E. Ginna, Rochester Gas and Electric Corporation
• H.L. Harrington, Niagara Hudson Power Corporation
• L.V. Nelson, Union Electric Company
• E.H. Schmidtman, Wisconsin Electric Power Company
• H.A. Snow, Detroit Edison Company
• E.T. Steel, Potomac Electric Power Company
• A.H. Sweetnam, Boston Edison Company
• F.M. Terry, Consolidated Edison of New York
• C.M. Turner, Ebasco Services
• W.R. Waggoner, Commonwealth and Southern Corporation
• R.R. Hermann, correspondent member, Northern States Power Company
 Paper Chart Cartridge Recorders Electronic Meters The Internet of
Recorders (Magnetic Tape) Things
Open Reel Tape Solid-State AMR/AMI
Recorders Recorders

1950s 1960s 1970s 1980s 1990s 2000s Beyond

figure 1. The evolution of metering technologies since the 1950s. AMR: automatic meter reading.

1938 to 1943. The AEIC Load Research

Committee was established in October

Stop The
1943 and held its first organizational
meeting in 1944. Soon afterward, mem-

PAIN!
ber utilities began conducting load stud-
ies to develop a better understanding of
major energy-using appliances, heating
and cooling equipment, and commercial
and industrial processes.
Following World War II, the elec-
tric utility industry in the United States Get Back to Your
underwent an extended period of rapid Oneline Faster With
growth, and new facilities needed to be
built. Load forecasts projected that the Revit Integration
high rate of growth would continue into
the foreseeable future; however, U.S. ex- Get to your analysis faster. Use our efficient data
pansion slowed in the 1970s resulting in integration between Autodesk® Revit® and EasyPower
the utility industry overforecasting the power system analysis. Stop the pain of wasted time
need for capacity. In 1978, the Public in your day.
Utilities Regulatory Policy Act (PUR-
PA) established requirements for utili- This stand-alone module automatically creates a
ties to collect load research data, such one-line diagram in EasyPower with one mouse-click
as daily kilowatt (kW) demand load in Revit. Get back to what you want to work on.
curves, for large utility rate classes that
were at least 10% of the overall sales
· Map Revit families and properties to EasyPower
equipment
of the utility. PURPA helped establish
confidence (90%) and precision (±10%) · Bring EasyPower analysis results back into Revit
requirements on estimates of selected Autodesk EasyPower
cost-allocation factors, such as coin- Revit Model System One-line
cident (system) peak demand. In turn,
T-SCV
500 kVA
.48-0.48 kV
4%
kV

T-SCV-S
8
.4
3-#250, 1#250, 1-#3

these requirements laid the foundation

kV
8
.4
3-#350, 1#350, 1-#4

3-#1, 1#1, 1-#8

3-#1, 1#1, 1-#8

for future considerations with respect 3-#250, 1#250, 1-#6

PANEL

to confidence and precision needs for Analysis and

Auto-Design
CURRENT IN AMPERES X 100 AT 480 VOLTS

energy-efficiency and demand-response

.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

Automated
600 600
500 500
0.15

400 400

300 TX-2 300

1 / 1.288 MVA
200 13.8 - 0.48 kV 200
6%
18.71

100 BL-1 100

LV Momentary Report GE AKR-50

Protective
80 80
TX-2
-2
1000 / 1288 kVA 1600/1760

evaluations. PURPA gave rise to a re-

60 60

EasyPower 10.4.0.198 10/11/2018 10:00:07 AM C:\...\Protection-1..dez 50 6% 50

.13

SWG-4
24

40 40
EasyPower LLC 30 BL-3 30

Comments: GE AKR-30H
20 800/400 20
1.90

d
Risk Hazar
BL-1
GE MVT
MV T-Plus
MVT-Plus
and Shock

Device
10 10
17.58

Arc Flash priate PPE Required

Vpu = 1.00 Sensor = 1600
Arc Flash and Shock
8 8
TIME IN SECONDS

TIME IN SECONDS

Plug = 1600
6 6

Appropriate PPE Risk Hazard

Cur Set = 11.1 (1760A)
.48

Appro
5 5
LT
T Band = 1
3 PHASE Fault Total Fault Fault Equipment Duties M-1
19

4 4

Required
STPU
STPU = 2 2.5
.5 (4400A) BL-5
ST
S T Delay
Dela
elayy = Int
Incident Energy coverall
3 3
Boundary C-H HFD
Flash Hazard inches - Arc Flash4’ - 0”or arc rated
ST
S T Delayy I²t
I² = Out
Sym X/R Mult Asym

newed interest in load research and an

Flash Hazard Boundary Override
O verride = 50000A 225/150
Bus Name Bus kV Equi Type Duty Amps 4’ - 0” cal/cm at
2 18 pants
2 2

Amps Ratio Factor Amps arc rated 6.0 cal/cm2 at 18 inches

6.0 shirt and - Arc Flash Incident
Arc-rated

Coordination
Energy
M-1 0.480 19480.3 3.69 1.19 23188.3 LVPCB 19480.3 cover is removed 1 1

Hazard when 0.48

.8 .8
kV Shock kV Shock Hazard
MCC-1 0.480 20032.6 3.55 1.18 23640.9 LVPCB 20032.6 0.48
BL-3
Limited Approach when cover is removed .6 .6
3’ - 6” Limited Approach .5 MVT-9
GE MVT
MV T-9 .5
3’ - 6” Approach 1’ - 0” .4 Sensor = 800 .4
MCC-2 0.480 19779.7 3.43 1.17 23181.0 LVPCB 19779.7 Restricted Restricted Approach Plug = 800
1’ - 0” (Fed By: 27B) ONLY .3 Cur
C ur Set = 0.5
0 (400A)
BL-1
18627A
.3

PNL-1 0.480 8882.5 3.73 1.19 10597.4 LVPCB 8882.5 t Name: MCC-23A CONFIGURATION Equipment Name:
SWG-4 (Fed By: .2
LT Band = 1
IInst
nst = 4 (3200A)
C-6
.2
Equipmen SYSTEM VALID FOR NORMAL 18A) 1 - 400 kcmil CU
NORMAL
PNL-2 0.208 5961.4 2.11 1.07 6377.3 LVPCB 5961.4 VALID FOR SYSTEM CONFIGURATIO
N ONLY TX-2
.1 1000 / 1288 kVA .1

entire generation of load research de-

SWG-4 0.480 24127.6 6.18 1.33 31990.4 LVPCB 24127.6 .08
BL-5
INRUSH .08

.06 .06
C uttler H
Cutler am
m
Hammer Series C
.05 .05
HF
HFD
.04 BL-3 .04
Frame
Frame = 225A (150A
(150AT)
Trip = 150 29249A
.03 .03

Short Circuit
.02 .02
BL-5
20574A
.01 .01
.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

CURRENT IN AMPERES X 100 AT 480 VOLTS

partments and activities. The load re- Calculations

searchers that joined the industry dur-
Explore more online and request a free demo copy at:
ing this era were often referred to as
PURPA babies. www.EasyPower.com/Revit
In the early days, instrumentation
and data collection were extremely chal-
lenging. In Figure 1, we show the evolu- ®
tion of metering equipment, beginning
Power made easy.
with paper tape recorders in the
1950s. These were followed by
open reel tape recorders and
then cartridge (magnetic tape)
recorders and solid-state record-
ers. Today, we are quickly ap-
proaching ubiquitous interval
load information, not just on the
whole facility but on a multitude
of elements surrounding a cus-
tomer’s every day existence.
Load Research
Manual
Partially in response to PURPA,
AEIC, with the support of the
U.S. Department of Energy and
Argonne National Laboratory,
began developing the first load
research manual (Figure 2). The
first edition of the manual, driven
by a team of industry profes-
sionals, was published in No -
vember 1980. Volume 1, Load
Research Procedures, included
Transformer
Bags
an overview of load research pro-
cedures and a review of the PUR-
PA requirements. Volume 2 was
Fundamentals of Implementing
Load Research Procedures, and
Volume 3, L oad Research
for Advanced Technologies,
also included solar, wind, and
cogeneration technologies.
Key contributors to this ef-
fort were Lois Brandenberg,
the head of load research for
Detroit Edison Company, and
Clarence Grund Jr., the head of
load research at Southern Com-
pany Services.
The second edition was up-
dated in 2008 and compiled into
11 chapters that included sec-
tions on business applications of
load research, load research for
competitive market settlements,
and load research evaluation.
figure 2. The front cover of Load Research Manual Miriam Goldberg led the sec-
(first edition). ond edition update.
The third, and most recent,
edition was updated in 2017 by a
consortium of load research practitio-
ners from DNV GL and included two
new chapters focused on the applica-
tion of AMI data and the exploration
and presentation of load research data.
In the
early days,
instrumentation
and data
collection were
extremely
substation
liners challenging.
Workforce Development
Following the publication of the first edi-
tion of the manual, the Load Research
Committee launched an educational se-
ries built around it. The Fundamentals
of Load Research Workshop continues
as the cornerstone of the education se-
1.800.231.6074 ries, helping to educate and introduced
reefindustries.com
new professionals to the world of load
research. This course is held annually
in New Orleans, Louisiana.
In the late 1970s, Bill
Hines from the Georgia
Institute of Technology
The
worked as a consultant Fundamentals
to several Southeast-
ern utilities, ­supporting of Load
the implementation of
load research efforts as
Research
well as building the ini- Workshop
tial intermediate-level
training session for the continues as the
AEIC, along with Doug
Montgomery. H i n e s
cornerstone of
and Montgomery were the education
joined by Roger Wright,
forming the original in- series.
structor group for the
AEIC’s “advanced” course. This AEIC
course was a week-long immersion in the
mathematics of load research, including
BIRD DIVERTERS
FireFly® & BirdMark®
The most effective bird diverters
in independent studies.
may/june 2020 ieee power & energy magazine 75
sample design, regression, and advanced
analytics. The course was taught by this
team until late in the 1990s.
The early training was
directed by a team of in­­
dustry practitioners and
academics. Today, the
AEIC load research work-
shops and training are
led by long-term indus-
try professionals com-
prising consultants and
practitioners. The intro-
ductory (fundamentals)
course is taught each year
and was recently moved
to a fall offering. The in-
termediate course was
redesigned by Craig Wil-
liamson and Kelly Marin
and focuses on the mathematics of
load research. Similarly, the advanced
course was redesigned and includes a
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
comprehensive analytics review while
focusing on new and advanced appli-
cations of load research. Curt Puckett
and Tim Hennessy serve as the lead
instructors for the advanced course.
The intermediate and advanced cours-
es are taught every other year on a ro-
tating basis.
In addition, the AEIC load research
committee organizes an annual con-
ference where load researchers make
presentations and share their analyti-
cal experiences. In addition to load re-
searchers, the conference also attracted
many load forecasters from utilities
as well as load forecasting vendors
and consultants. Tao Hong, the for-
mer chair of the IEEE Working Group
on Energy Forecasting, is a frequent
speaker. Many attendees of the AEIC
load research conference are also ac-
tive participants in IEEE Power & En-
ergy Society conferences.
LOW LINE MARKERS
QuikMark TM & SpanMark TM
Saving lines & saving lives.
Snaps directly onto live lines.
6874
 In North America, a second load
research alliance, the Western Load
Research Association (WLRA), was
formed in the late 1970s. The WLRA is a
regional organization for the western part
of North America whose most important
function is information exchange. Some
utilities participate in both the WLRA
and the AEIC. Whereas the WLRA is a
regional association of primarily west-
ern U.S. utilities with two conferences
per year, the AEIC is a formal member
association that hosts workshops, train-
ing sessions, and conferences and main-
tains the load research manual. In recent
years, the WLRA fall conference and
AEIC annual load research conference
have been jointly held.
Analytical Methods
During the early years, the cost of in-
stalling and maintaining metering and
data collection equipment was both
IS YOUR REPUTATION FOR
RELIABLE SERVICE AT RISK?
BirdguarD™
Wildlife Power Outage Protection
BirdguarD™ Material Qualities:
UL94 V-0 Flammability Rating
Material Will NOT Support Combustion
IEEE 1656 & ASTM Tested to 72kV
UV Resistant
ISO 9001:2015 Certified
Women Owned Business - WBENC
Public Utility Commission Certified
Customer Support
P (801) 972-5400 | (888) 658-5003
www.KADDAS.com
76 ieee power & energy magazine
formidable and challenging. The cost- where the average for the sample was
per-sample point could easily exceed calculated and extrapolated based on
US$5,000, meaning that sample design the number of customers in the popula-
efficiency was a primary consideration. tion. MPU estimation could be used with
However, the minimi- a simple random sample
z a tion of the sa mple or within a stratification
conf l ict e d w it h t h e
W. Cochran’s framework. Stratification
reliability of the early book, Sampling was primarily done using
data collection networks, Dalenius and Hodges’
which were not depend- Techniques, minimum variance strat-
able. Load researchers ification with Neyman
would often make the
became the allocation. These were all
planning assumption go-to sampling well-established statistical
of a 20% data loss, in- methods used for estima-
creasing sample sizes and analysis tion and sample design
to account for missing and formed the original
data. The sample design
book of the statistical methods de-
strategy was tied directly load research ployed in load research.
to the sample extrapola- In the mid-1970s, Con-
tion approach. community. sumers Power Company
In the beginning, most was presented with a
companies were using a simple mean challenge by the Michigan Public Ser-
per unit (MPU) expansion strategy, vice Commission. Although the com-
pany had a long history of practicing
load research, the commission staff
requested that the samples be signifi-
cantly increased. As mentioned pre-
viously, adding sample points was a
formidable challenge and an expensive
endeavor. The company took the chal-
lenge to Roger Wright at the University of
Michigan. Wright examined the chal-
lenge and quickly realized that the load
factor seemed to be a very important
element in the analysis schema. He also
noticed that utilities had significant
amounts of useful information about
how their customers use energy, name-
ly, the monthly billed energy for its full
population of customers. Wright specu-
lated that if this information could be
used in the sample design, then the
sample size requirements could be re-
duced by leveraging the correlation
between the known billing information
and the variable(s) of interest, such as
INNOVATIONS IN PLASTIC
system peak demand or the 12-month
coincident peak demands. This gave
BirdguarD™ Delivers a Triple Win:
rise to the use of an approach called
1. Consumer Reliability model-based statistical sampling, which
2. Utility Asset Protection was merely a modified application of
3. Positive Environmental Impact stratified ratio estimation.
W. Cochran’s book, Sampling Tech-
Kaddas Enterprises, Inc. 2019. Not for Personnel Protection
niques, became the go-to sampling
and analysis book of the load research
may/june 2020
community. Wright developed a meth-
od for optimizing the sample design
for ratio estimation based on the un-
derlying relationship between each
account’s demand variable of interest
and the account’s corresponding billed
energy. The elegance of his approach is
that it stratified the sample based on
the anticipated standard deviation of
the variable of interest, allowing for
an equal number of sample points to
be allocated to each stratum. Howev-
er, it should be noted that Dalenius and
Hodges’ minimum variance stratifica-
tion with Neyman allocation is often
used with stratified ratio estimation
with little adverse effect. The methods
put forward by Wright convinced the
commission staff and have helped lay
the foundation for the application of
stratified ratio estimation in both the
load research and the energy-efficiency
evaluation communities.
may/june 2020
In the late nificant early end-use research proj-
ects included
1970s, the 1) the End-Use Load and Consum-
er Assessment Program, con-
industry started ducted under the direction of the
conducting Bonneville Power Association in
the mid-to-late 1980s
residential and 2) Sierra Pacific Power Company’s
End-Use Load Research Project,
commercial performed in the mid-1980s
direct load by t hei r Resou rce Pla n n i ng
Department
control 3) the New South Wales, Australia,
Residential End-Use Study, con-
experiments. ducted in the early 1990s.
These projects were comprehensive in
Emerging Topics their design, collecting data on a wide
Although most of the focus was on array of classes and end uses. The pri-
the entire facility’s load data, load re- mary goal of the end-use research was
search did spin off two derivative areas to provide information for demand-
of interest, namely, end-use research side planning, load forecasting, and
a nd demand-response evaluation. demand-side management program de-
Three of the largest and most sig- velopment and delivery.
IEEE power & energy magazine 77
 In addition to these selective and
comprehensive end-use research proj-
ects, it was common for the industry
to conduct one-off research on se-
lected appliances, particularly in the
residential market, including room and
central air conditioning, washers and
dryers, microwaves and
ranges, refrigerators and
freezers, ­d ishwashers, It is important
and water heaters. to note that
Around 1990, the Elec-
tric Power Research In- load research
stitute initiated a project
to build a clearinghouse studies,
for utility end-use data including end-
called the Center for
Electric End-Use Data use studies,
(CEED). Utilities with
data could contribute to have been
CEED and receive load conducted
shapes, and those without
any data could purchase on other
the end-use load shapes
they needed. The two commodities.
main purposes of the end-
use research were to 1) lay the foun-
dation for bottom-up forecasting and 2)
develop and inform demand-side manage-
ment programs.
In the late 1970s, the industry start-
ed conducting residential and com-
mercial direct load control experiments.
These projects included air-condi-
tioning and water-heating direct load
control, residential storage heating,
and a wide range of time-of-use rates.
These studies often included measur-
ing the affected end uses to better
isolate and understand the end-use
loads. Here again, load researchers
were the optimum choice to conduct
interval load research focused on plan-
ning for and quantifying the impacts
associated with these types of programs
and initiatives.
Although most of this discussion in
this article has been focused on electric
load research, it is important to note
that load research studies, including
end-use studies, have been conducted
on other commodities including natu-
ral gas, steam, and water. Consolidated
Edison of New York routinely conducts
load research studies on its natural
78 ieee power & energy magazine may/june 2020
gas and steam customers. East Ohio
Gas Company, Michigan Consolidat-
ed Gas Company, Consumers Power
Company, and Puget Sound Energy
Company have all conducted gas load
research. In the mid-1990s, Minne-
gasco conducted an end-use gas load
research project focused
on the impact of high-
efficiency residential
furnaces on its gas dis-
tribution network.
What’s Next?
It is a time of transition
for load research. Some
utilities are still living in
the past, with pulse-ini-
tiating meters and load
recorders procured and
deployed for data cap-
ture, while other utilities
are rapidly advancing
toward the future with
AMI, shifting the land-
scape to the forefront of
interval load analytics. The industry is
quickly moving from a focus on data
capture to a world where interval load
data are available on large portions of
the population, moving the focus toward
data access, handling, and analytics.
The AEIC Load Research Committee
acknowledged this transition by renam-
ing itself and including “Analytics”
in the committee name, i.e., the AEIC
Load Research and Analytics Commit-
tee. The current objectives of the com-
mittee include
✔✔ promoting improvements in the
collection, analysis, and applica-
tion of load research data and as-
sociated information
✔ ✔ educating the industry about
load research
✔✔ encouraging the use of load re-
search in utility management de-
cision making.
Even though load research’s past
was built upon the fortitude and fore-
sight of a handful of individuals, the
present is the product of many more
industry practitioners. Not only are
past and current members of the AEIC
and WLRA helping to guide and shape
our industry by actively promoting
the growth and advancement of load
research, we are starting to see load
research workshops and conferences
attract people interested in load fore-
casting, distributed energy resources,
and demand-side management evalu-
ation and planning. Presentations are
highlighting more advanced analytical
techniques like machine learning and
artificial intelligence, and the focus of
many of the studies include interesting
customer-side resources such as elec-
tric vehicles, photovoltaics, and battery
storage. For a look into the future of
load research, please continue reading
the feature article, “Utility Load Re-
search: The Future of Load Research
Is Now,” in this month’s IEEE Power
& Energy Magazine.
For Further Reading
L. J. Vogt, “The history of the AEIC Load
Research Committee: 1944–1985,” AEIC
Load Research Committee, Birming-
ham, AL, June 1985. Accessed on: Mar.
10, 2020. [Online]. Available: https://aeic
.org/wp-content /uploads/2017/05/
H istor y- of-A EIC-L oad-Resea rch
- Com m it tee-194 4 -1945-P repa red
-June-1985-1.pdf
L. Brandenburg et al., “Load Research
Manual: Load Research Procedures,” U.S.
Department of Energy, Argonne National
Laboratory, Argonne, IL, ANL/SPG-13,
vol. 1, Nov. 1980. Accessed on: Mar. 10,
2020. [Online]. Available: https://www
.osti.gov/servlets/purl/6705685
L. Brandenburg et al., “Load Re-
sea rch Ma nua l: Fu nd a ment a ls of
Implementing Load Research Proce-
dures,” U.S. Department of Energy, Ar-
gonne National Laboratory, Argonne,
IL, ANL/SPG-13, vol. 2, Nov. 1980.
Accessed on: Mar. 10, 2020. [Online].
Available: https://www.osti.gov/servlets/
purl/6705684
L. Brandenburg et al., “Load Re-
search Manual: Load Research for Ad-
vanced Technologies,” U.S. Department
of Energy, Argonne National Laboratory,
Argonne, IL, ANL/SPG-13, vol. 3, Nov.
1980. Accessed on: Mar. 10, 2020. [On-
line]. Available: https://www.osti.gov/
servlets/purl/6705679
 AEIC Load Research & Analytics
Committee, Load Research Manual,
3rd ed. Birmingham, AL: Associa-
tion of Edison Illuminating Companies,
2017. Accessed on: Mar. 10, 2020. [On-
line]. Available: https://www.techstreet
.com/standards/load-research-manual
-third-edition?product_id=1991775
W. G. Cochran, Sampling Techniques,
3rd ed. New York: Wiley, 1977.
R. Wright, “Model-based statis-
tical sampling: A statistical method-
ology for program impact evaluation and
other studies,” Model Based Sampling. Ac-
cessed on: Mar. 10, 2020. [Online]. Avail-
able: http://www.modelbasedsampling
.com/index.html
T. Dalenius and J. L. Hodges Jr., “Min-
imum variance stratification,” J. Am. Stat.
Assoc., vol. 54, no. 285, pp. 88–101, 1959.
doi: 10.1080/01621459.1959.10501501.
J. Neyman, “On the two different
aspects of the representative method:
The method of stratified sampling and
the method of purposive selection,”
J. Royal Stat. Soc., vol. 97, no. 4, pp.
558–625, 1934. doi: 10.2307/2342192.
Western Load Research Association,
“A chronology of the Western Load
Research Association (WLRA).” Ac-
cessed on: Mar. 10, 2020. [Online].
Available: http://www.wlra.net/history
.htm
R. L. Wright, “Chapter 12: Uncer-
tainty,” in “The California Evalua-
tion Framework,” California Public
Utilities Commission and the Project
Advisory Group, Proj. no. K2033910,
June 2004.
R. L. Wright, “Chapter 13: Sam-
pling,” in “The California Evaluation
Framework,” California Public Utili-
ties Commission and the Project Advisory
Group, Proj. no. K2033910, June 2004.
p&e

& IAS

2020 PowerAfrica Conference

25-28 August 2020 | Nairobi, Kenya
You are invited to attend the 7th Annual PowerAfrica Conference
(PAC2020), a powerful international forum where research
scientists, engineers, and practitioners present and discuss the
latest research findings, innovations, emerging technologies and
applications.

Come hear world-renowned speakers and share your own

experiences with industry colleagues from around the globe.

Co-sponsored by IEEE Power and Energy Society (PES) and

Industrial Applications Society (IAS), the conference will address
important issues facing the African continent, including:
• Power systems integrations
• Business models
• Technological advances
• Policies and regulatory frameworks

Digital Object Identifier 10.1109/MPE.2020.2981788

Learn More at: ieee-powerafrica.org

may/june 2020 ieee power & energy magazine 79

 calendar
T
TH E I EEE POW ER & EN ERGY
­Society’s (PES’s) website (http://www
.ieee-pes.org) features a meetings sec-
tion, which includes calls for papers
and additional information about each
of the PES-sponsored meetings. Please
check the website for the most current
information.
June 2020
IEEE Transportation Electrification
Conference & Expo (ITEC 2020), 24–26
June, Chicago, Illinois, United States, con-
tact Rebecca Krishnamurthy, rebecca.k@
rna-associates.com, https://itec-conf.com
August 2020
IEEE PES General Meeting (GM 2020),
2–6 August, Montréal, Canada, con-
tact Roseanne Jones, roseanne.jones@
ieee.org, http://www.pes-gm.org/2020/
IEEE PES/IAS PowerAfrica (Power-
Africa 2020), 25–28 August, Nairobi, Ken­
ya, contact Mercy Chelanget, chelangat
_ke@ieee.org, https://ieee-powerafrica.org/
September 2020
IEEE International Conference on
Power Systems Technology (Power-
Con 2020), 9–12 September, Banga-
lore, India, contact Kasi Viswanadha
Raju Gadiraju, kasiViswanadhaRaju
.Gadiraju@ge.com
IEEE International Forum on Smart
Grids for Smart Cities (SG4SC 2020),
16–18 September, Aachen, Germany,
Digital Object Identifier 10.1109/MPE.2020.2971823
Date of current version: 17 April 2020
80 ieee power & energy magazine may/june 2020
PES meetings
for more information, www.ieee-pes.org
contact Alina Schneiders, a.schneiders@
academy.rwth-aachen.de
IEEE PES Asia-Pacific Power &
Energy Engineering Conference
(APPEEC 2020), 20 –23 September,
Nanjing, China, contact Li Zhang,
zhanglinuaa@hhu.edu.cn, https://ieee
-appeec.org/
IEEE PES Transmission and Dis­
tribution Conference and Exposi-
tion Latin America (T&D LA 2020),
28 September–2 October, Montevideo,
Uruguay, contact Kathy Heilman, kathy
.heilman@ieee.org, https://www.ieee
-tdla2020.org/
October 2020
IEEE PES Transmission and Distri-
bution Conference and Exposition
(T&D 2020), 12–15 October, Chicago,
Illinois, United States, contact Carl
Segneri, csegneri@mjelectric.com,
http://www.ieeet-d.org/
IEEE PES Innovative Smart Grid
Technologies Europe (ISGT Europe
2020), 25–28 October, Delft, The
Netherlands, contact Kathy Heilman,
kathy.heilman@ieee.org, https://www
.isgt-europe-2020.nl/
November 2020
IEEE PES Innovative Smart Grid
Technologies Asia (ISGT Asia 2020),
23 –26 Novemb er, Per t h, West er n
Australia, contact Farhad ­Shahnia,
f.shahnia@murdoch.edu.au, https://
ieee-isgt-asia.org/
December 2020
IEEE International Conference on
Power Electronics, Drives, and En-
ergy System (PEDES 2020), 16–19
December, Jaipur, India, contact Sury-
anarayana Doolla, doolla@gmail.com,
http://pedes2020.com
January 2021
IEEE PES 2021 Joint Technical
Committee Meeting (JTCM 2021),
7–15 January, New Orleans, Louisiana,
United States, contact Solveig Ward,
sward@quanta-technology.com, www
.pestechnical.org
June 2021
I EEE PES GT&D International
Conference & Exposition, Istanbul
(GTD), 1– 4 June, Istanbul, Turkey,
contact Omer Usta, usta@ieee.org
IEEE PES Innovative Smart Grid
Technologies Asia (ISGT Asia 2021),
8–11 June, Brisbane, Australia, contact
Tapan Saha, saha@itee.uq.edu.au
IEEE PowerTech Milan (PowerTech
2021), 28 June–2 July, Madrid, Spain,
contact Tomas GomezSan-Roman,
tomas.gomez@comillas.edu
For more information on additional tech-
nical committee meetings, webinars, and
events, please visit our IEEE PES calen-
dar: https://www.ieee-pes.org/meetings
-and-conferences/conference-calendar.
p&e

in my view (continued from p. 84)
r­esidential ­customers to a default time-
of-use (TOU) rate. As another example,
Oklahoma Gas and Electric (OG&E)
offers residential and small commercial
customers a variable peak pricing (VPP)
rate, where the peak price changes daily
to better reflect real-time market con-
ditions. As of December 2016, OG&E
enrolled approximately 19% of its resi-
dential customers in the VPP rate.
The Development of Rates and
Programs to Promote Midday
Load Building
Utility system planners and system oper-
ators in some regions are anticipating, if
not already observing, steep declines in
net system loads (i.e., gross load minus
generation from renewable resources)
during the morning and steep inclines
in late-afternoon/early-evening periods,
due to solar PV resources. This often re-
sults in wholesale power costs dropping
precipitously during midday hours and
the curtailment of renewable generators.
One strategy utilities have implemented
to address these grid management chal-
lenges employs TOU rates with very
low-priced (“super off-peak”) periods
that coincide with low-cost midday
hours (sometimes referred to as matinee
pricing). California’s investor-owned
utilities are currently testing whether
residential customers will increase their
usage in response to such super off-peak
prices as well as reduce or shift usage
away from higher-priced peak periods
that primarily cover late-afternoon and
early-evening hours. Utilities are also
considering programs that provide di-
rect financial incentives for custom-
ers to increase electricity consumption
during periods with excess generation
on the system. Most notably, Arizona
Public Service (APS) filed a proposal to
implement a reverse demand-response
(DR) program to promote load building
during certain time periods to reduce
the curtailment of renewable generators.
The Increased Application of
Residential Three-Part Rates
Although three-part rates (i.e., de-
mand charges, along with volumetric
may/june 2020 ieee power & energy magazine 81
energy charges and fixed customer
charges) have traditionally been lim-
ited largely to commercial and indus-
trial customers, utilities have become
increasingly interested in extending
demand charges into the residential
sector as well, with the stated purpose
of better aligning rate design with un-
derlying cost causation and stabilizing
fixed-cost recovery. Within the resi-
dential sector, demand charges have
historically been offered on a voluntary
basis. For example, APS has more than
a decade of experience
with voluntary residen-
tial demand charge rates Emerging
and recently launched
a new set of tariff op- rate designs
tions, with 17% of cus-
tomers opting into one generally
of the demand charge encourage
rates. In contrast, other
utilities, including the load building.
Salt River Project, have
recently implemented
or proposed mandatory residential de-
mand charges, often targeted at custom-
ers with rooftop PVs or other DERs.
The Development of New
Net-Metering Alternatives
Many states have reformed existing
NEM tariffs, driven chiefly by en-
rollment caps, concerns about cost
shifting between NEM participants
and other ratepayers, and a desire to
incentivize customer DER invest-
ments that provide greater benefits to
the broader electric system. Of all the
alternatives to NEM, net billing has
been, by far, the most common ap-
proach—whereby customers contin-
ue to offset contemporaneous usage
with DERs but any exported energy
is compensated at some designated
grid-export rate. New York’s Value of
Distributed Energy Resources tariff
represents a relatively sophisticated
form of net billing, with grid-export
rates that vary by time and location
and a phased-implementation sched-
ule for different market segments.
Hawaii has also moved to net billing,
with a range of transitional tariff op-
tions, one of which promotes combi-
nations of solar and storage configu-
rations that qualify for more favorable
grid-export prices.
The Development of Rates
Specific to Electric Vehicles
The states and utilities with some of
the highest growth and interest in sup-
porting electric vehicle (EV) adoption
are introducing retail rates specific to
EVs. Some of these rates are designed
to encourage charging behaviors that
minimize grid impacts
or potentially benefit the
grid from such electric-
intensive end uses. EV-
specific rates primarily
differ from one another
in terms of whether they
include demand- or time-
based energy charges, a
potentially contentious
detail that may incentiv-
ize or deincentivize cer-
tain forms of EV charging (e.g., demand
charges may particularly impact public
charging by penalizing fast chargers,
which are demand intensive). Georgia
Power’s rate offers EV owners a TOU
energy charge applicable to the entire
household consumption. This contrasts
with Austin Energy’s residential fixed
monthly fee limiting EV charging to
off-peak hours only. San Diego Gas
and Electric’s rate for EV charging
at multiunit dwellings and workplaces
includes locational costs based on
California independent system operator
day-ahead market prices and distribu-
tion feeder load.
Implications for
DER Deployment
Each trend entails potentially signifi-
cant implications for solar and other
DERs, in terms of both the quantity
and type of deployment that may oc-
cur in the future. In considering how
these various rate reform trends may
impact DER deployment, several broad
themes emerge.
First, DER impacts depend critically
on the specifics of the tariff structure.
 Even if obvious, the importance of the
specifics of any particular rate design
in assessing the potential impacts on
DER deployment cannot be overstated.
This includes details such as the timing
and peak-to-off-peak pricing differen-
tial under time-based rates, the choice
between intermittent versus continuous
incentives to increase midday load, the
use of coincident versus noncoincident
demand charges, the specific price paid
for grid exports under net billing rates,
and whether or not EV-specific rates are
submetered versus applied on a whole-
house basis. Details such as these dic-
tate not only the magnitude but, in some
cases, also the directionality of the fi-
nancial effects for certain DERs.
Second, flexible DERs generally be­­
nefit more under emerging rate design
trends. DERs exist along a continuum of
flexibility, ranging from
those with largely un-
controlled load shapes The incentives
[energy efficiency (EE)
and PVs] to those with
for load
some level of discretion building
in how they are oper-
ated (EVs and certain are often
other forms of electrifi-
cation) to fully dispatch-
concentrated
able resources (storage during
and certain forms of DR
and electrification). Most particular times
emerging rate reforms
tend to support a greater
of the day.
deployment of flexible
DERs (e.g., storage and DR), while often
constraining the adoption of less flexible
resources (e.g., PVs and EE). This out-
come is driven by the general movement
toward rate structures with greater tem-
poral granularity, which naturally tends
to encourage price-responsive resources.
Third, emerging rate designs gener-
ally encourage load building (during
specific times of the day). Although only
one of the five rate design trends noted
in this article is explicitly intended as a
tool for load building (see the “The De-
velopment of Rates and Programs to
Promote Midday Load Building” sec-
tion), most of the other rate design trends
also incentivize load building, whether
82 ieee power & energy magazine may/june 2020
in the form of EVs, other types of elec-
trification, or energy storage (which in-
creases net electricity consumption due
to round-trip losses and ancillary loads).
The incentives for load building are of-
ten concentrated during particular times
of the day; depending on their design,
three-part rates may encourage load
building across a fairly broad range of
hours. In contrast, the emerging rate de-
signs discussed in this article generally
tend to constrain the growth of DERs
that reduce the consumption of grid-
supplied electricity (EE and, especially,
PVs). This outcome is driven partly by
the general movement toward greater
levels of attribute unbundling and tem-
poral granularity that better reflect
marginal costs. Load building is also a
natural response on the part of electric
utilities to slowing sales growth and
ongoing concerns about
revenue erosion from
EE and PVs. Electrifica-
tion is also a strategy that
some state policy mak-
ers and regulators have
endorsed for address-
ing their greenhouse gas
abatement goals.
In addition to DER
deployment, these rate
reforms have broad im-
plications: for utilities,
in terms of their opera-
tions, planning, and fi-
nancial health; for cus-
tomers, in terms of distributional bill
impacts within and across customer
classes and their opportunities to man-
age energy costs; and for society more
broadly, in terms of the economic and
environmental impacts associated with
the provision of energy services. Ac-
cordingly, regulators engaged in retail
rate reform efforts will need to weigh
these impacts against each other, while
balancing other considerations and
stakeholder perspectives as they estab-
lish utility rate structures.
Acknowledgments
This work was supported by the U.S.
Department of Energy Solar Energy
Technologies Office under Lawrence
Berkeley National Laboratory contract
DE-AC02-05CH11231. The U.S. Gov-
ernment retains, and the publisher, by
accepting the article for publication,
acknowledges that the U.S. Govern-
ment retains a nonexclusive, paid-up,
irrevocable, worldwide license to pub-
lish or reproduce the published form of
this manuscript, or allows others to do
so, for U.S. Government purposes.
For Further Reading
N. Da rghouth, G. Ba r b o s e, and A.
Mills, “Implications of rate design for
the customer-economics of behind-the-
meter storage,” Berkeley Lab Rep., CA,
2019. [Online]. Available: https://emp
.lbl.gov/publications/implications-rate
-design-customer
D. Glick, M. Lehrman, and O. Smith,
“Rate design for the distribution edge:
Electricity pricing for a distributed re-
source future,” Rocky Mountain Inst.,
Boulder, CO, 2014. [Online]. Available:
https://rmi.org/insight/rate-design-for-the
-distribution-edge-electricity-pricing
-for-a-distributed-resource-future/
R. Hledik, “Rediscovering resi-
dential demand charges,” Electricity
J., vol. 27, no. 7, pp. 82–96, 2014. doi:
10.1016/j.tej.2014.07.003.
A. Satchwell, P. Cappers, and G.
Barbose, “Current developments in re-
tail rate design: Implications for solar
and other distributed energy resourc-
es,” B e r k el ey Lab Rep., CA, 2019.
[Online]. Available: https://emp.lbl
.gov/publications/current-developments
-retail-rate
A. Satchwell and P. Cappers, “Evolv-
ing grid services, products, and mar-
ket opportunities for regulated elec-
tric utilities,” Berkeley Lab Rep., CA,
2018. [Online]. Available: https://emp
.lbl.gov/publications/evolving-grid
-services-products-and
T. Stanton, “Review of state net en-
ergy metering and successor rate de-
signs,” National Regulatory Res. Inst.,
Washington, D.C., 2019. [Online]. Avail-
able: https://www.naruc.org/nrri/nrri
-library/research-papers/electricity/
 p&e
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

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

DIgSILENT GmbH 3 www.digsilent.de /powerfactory

EasyPower LLC 73 www.EasyPower.com/revit

ETAP 9 etap.com/microgrid +1 949 900 1000

G&W Electric Company 77 gwelectric.com

Kaddas Enterprises, Inc. 76 www.KADDAS.com +1 801 972 5400

P & R Technologies 75 www.pr-tech.com +1 800 722 8078

Powertech Labs, Inc. 16 www.dsatools.com

PowerWorld Corporation 12 www.powerworld.com +1 217 384 6330

PSCAD a product of MHI 11 pscad.com

Reef Industries 74 reefindustries.com +1 800 231 6074

RTDS Technologies, Inc. 7 www.rtds.com

S&C Electrical Company 15 sandc.com/vacufuse19

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

YVI Labs 13 www.photonplusplus.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.2019.2954663

may/june 2020 ieee power & energy magazine 83

 in my view
R
RETAIL ELECTRICITY PRICING
is evolving in the context of broader
shifts in how electricity customers pay
for grid services and are compensated
for customer-sited generation. These
retail rate reforms are occurring most-
ly, though not exclusively, among resi-
dential customer classes and are driven
by a variety of technology- and policy-
related trends.
Key Drivers
of Retail Rate Reforms
Four particular drivers stand out. First,
many utilities across the United States
have deployed residential advanced
metering infrastructure (AMI), allow-
ing for two-way communication and
a more granular recording of energy
consumption. According to the Edison
Foundation Institute for Electric Inno-
vation, more than 55% of U.S. house-
holds now have advanced meters, the
majority of which were installed be-
tween 2010 and 2015. One of the key
rationales for justifying AMI deploy-
ment has been its potential to enable a
more widespread application of time-
based rate designs.
Second, residential and commercial
electricity customers continue to in-
crease their investment in distributed
energy resources (DERs). For exam-
ple, U.S. residential solar photovoltaic
(PV) capacity has grown by 50% per
year on average since 2010, spurred
by substantial installed-cost declines
Digital Object Identifier 10.1109/MPE.2020.2971844
Date of current version: 17 April 2020
84 ieee power & energy magazine may/june 2020
Andrew Satchwell, Peter Cappers, and Galen Barbose
retail rate designs
current developments in the industry
as well as by favorable incentive poli-
cies and electric utility rate designs,
such as net energy metering (NEM).
Increased DER adoption has put pres-
sure on regulators and utilities to re-
vise retail rate designs prompted by
NEM program caps and optimize the
size, location, and operation of DERs.
Third, utilities have increasingly
raised concerns about fixed-cost recov-
ery and revenue sufficiency, an issue
partly associated with increased DER
deployment but related more generally
to flat or declining load growth and
compounded by cost increases related
to grid modernization. This motivated
some utilities and states to reexamine
their historically heavy reliance on
volumetric energy charges to cover
both fixed and variable utility costs
and also prompted efforts to revise net-
metering tariffs.
Fourth, at the bulk power system
level, utility system planners and op-
erators are experiencing or anticipat-
ing significant changes to hourly net
load as a result of the broader deploy-
ment of variable renewable energy re-
sources, both utility and smaller scale
distributed resources. In particular,
the surplus generation of electricity
by these resources during the middle
of the day is causing significant de-
clines in energy supply prices. This
creates an opportunity for customers
to capture value in shifting load to
those periods and away from morn-
ing and/or afternoon peaks if rate de-
signs reflective of these price profiles
are implemented.
Designing Retail Rates for
a Changing Grid
Although numerous changes to retail
rates have occurred across the United
States in recent years, five particularly
salient trends have emerged, at least
in part, in response to these drivers.
The depth and longevity of experience
across these five rate designs vary sig-
nificantly, but each represents some
incremental progression from fully
bundled rates based on average util-
ity costs toward rates with a greater
unbundling of electricity services and
temporal differentiation and location-
al granularity.
The Increased Pursuit of
Residential Time-Based Rates
Despite more than four decades of ex-
perience, only roughly 3% of residen-
tial electricity customers were on time-
based rates in 2017. As AMI deployment
progresses, broader and stronger regula-
tory support for time-based rate options
is beginning to emerge. The implemen-
tation of time-based rates is primarily
motivated by economic efficiency (i.e.,
to better align electricity prices with
marginal system costs) and the need to
lower utility costs by reducing peak de-
mand and improving system load fac-
tors by encouraging customers to shift
electricity consumption from peak to
off-peak periods. Recent and notable
examples of this include the Sacra-
mento Municipal Utility District in Cali-
fornia, which is transitioning all of its
(continued on p. 81)
 Get Published in the New
IEEE Open Access Journal of
Power and Energy

Submit a paper today to the

premier new open access journal
in Power & Energy.
Your research will benefit from
the IEEE marketing launch and
5 million unique monthly users
of the IEEE Xplore® Digital Library.

Submit your paper today!

Visit ieee-pes.org/oa-journal
to learn more.

Digital Object Identifier 10.1109/MPE.2020.2983527