Professional Documents
Culture Documents
magazine
on the
cover
37
©SHUTTERSTOCK.COM/ALBERTO MASNOVO
TURBINES & SOLAR PANELS: ©SHUTTERSTOCK/VIKTORIYA
features
contents
20 Consumer-Led Transition 64 Back in the Race
By Naomi Stringer, Anna Bruce, Iain MacGill, By Ricardo Guerrero Lemus
Navid Haghdadi, Peter Kilby, Jacqui Mills,
Taru Veijalainen, Matt Armitage, 75 On Good Behavior
and Nigel Wilmot By Tom Key, Gefei “Derek” Kou,
and Michael Jensen
37 Autonomous Energy Grids
By Benjamin Kroposki, Andrey Bernstein, 86 Predictive-Maintenance Practices
Jennifer King, Deepthi Vaidhynathan, By Richard Fioravanti, Kiran Kumar,
Xinyang Zhou, Chin-Yao Chang, Shinobu Nakata, Babu Chalamala,
and Emiliano Dall’Anese and Yuliya Preger
columns &
departments
4 From the Editor
10 Leader’s Corner
16 Guest Editorial
98 Society News
99
99 History
108 Book Review
109 Calendar
112 In My View
SFI-01681
Different network model sources such as GIS, PowerFactory database or other calculation tools
Hosting Capacity Analysis module of the well-proven calculation software PowerFactory
Wide range of constraints for capacity calculation such as loading and voltage limits, reverse power
flow, power quality
Results for base case and n-1 (feeder switch-over)
Available for all voltage levels and network model sizes
Modular integration into existing business application due to separation of front- and back-end
Our consulting team uses its expertise to provide a customised approach, integrating the calculation
module into your own systems.
a global perspective
integrating distributed energy resources
A
AS I REFLECT ON THE YEAR 2020
for this final issue of IEEE Power & En-
ergy Magazine, I am quickly reminded
of a lyric from a song by an American
rock band: “Lately, it occurs to me what
a long, strange trip it’s been.” As I write
©SHUTTERSTOCK.COM/ELENABSL
this column, it occurs to me that we are
not yet at our final destination on this
long, strange trip called COVID-19. As
we started the year, COVID-19 was just
entering our lexicon. And now, at the
end of 2020, the ways we do business,
educate, travel, hold meetings and con- from around the globe discussing the es in other industries, notes the
ferences, and interact with each other following topics: key components of an effective
have changed—all due to COVID-19. ✔✔ the challenges being addressed approach, and describes method-
We have not yet achieved an un- as Australia integrates very high ologies to identify leading fault
derstanding of what the “new normal” levels of distributed photovoltaic indicators.
will be. Many jobs remain out of the (PV) systems The “In My View” column provides a
office and at home locations; numerous ✔✔ an autonomous energy grid con- perspective on grid services from DERs.
universities continue with remote edu- cept for the control and opera-
cation; social distancing continues to be tion of grids with high levels of History
an active practice in our communities, DERs of all types We have explored a number of inter-
factories, and field work spaces; utility ✔✔ a detailed view from Japan of grid esting aspects of electrical power as-
load profiles have changed; and many operation challenges related to sociated with the Niagara Falls area
global travel restrictions persist. Please high levels of DERs in the form of through the years in the pages of the
continue to make wise choices and be distributed PV and wind energy “History” column. In this issue, Rob-
safe as we migrate through this long, ✔ ✔ the challenges of reaching a ert D. Barnett presents a view of the
strange trip during the COVID-19 era. 100% renewable energy target “loads” consumed by customers of the
by 2040 in the Canary Islands early hydroelectric generation plants
In This Issue ✔✔ system protection impacts re- near Niagara Falls. The article con-
The overall theme of this issue is the sulting from the integration of centrates on early electrochemical fac-
integration of distributed energy inverter-based DERs. tories and the distribution methods to
resources (DERs). We have tackled Along with these articles on DERs, serve those loads.
this topic in IEEE Power & Energy we have one more article to this issue:
Magazine through the years, and this ✔✔ “Predictive-Maintenance Prac- Leader’s Corner
latest issue represents the next pro- tices,” by Richard Fioravanti, Ki- In the “Leader’s Corner” column, Vijay
gression in this important area affect- ran Kumar, Shinobu Nakata, Babu Vittal, IEEE Power & Energy Society
ing our industry. We have five articles Chalamala, and Yuliya Preger, ad- (PES) vice president (VP) of Technical
that represent authors and DER topics dresses current safety practices in Activities, shares his thoughts on PES
battery energy storage system de- technical activities. As a VP on the
Digital Object Identifier 10.1109/MPE.2020.3016409
velopment, provides examples of PES Governing Board, he also serves
Date of current version: 16 October 2020 predictive maintenance approach- as chair of the PES Technical Council,
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
based on the ANSI/IEEE C37 Standards. Separate solutions for low, DC SYSTEMS ANALYSIS Battery Sizing, Load Flow, & Short Circuit
different cases very easy.” medium and high voltage systems and for symmetrical, momentary Analysis. Evaluate all loading conditions for DC duty cycle loads and
and interrupting calculations. (Unbalanced Studies also available) AC emergency loads. Complies with IEEE std. 485, 1115, 399, 946,
and IEC std. 61660.
EQUIPMENT EVALUATION Equipment Evaluation Report. Automatically
compares short-circuit ratings, withstand ratings. Applies de-rating GROUND MAT Substation Ground Grid Design and Analysis. Optimizes
adjustments and user defined pass/marginal/failed criteria. Includes grid design using general purpose finite element algorithm for
error checking for input data and topology. potential analysis and graphical facilities to validate grounding
systems efficiency.
TMS Transient Motor Starting Simulation. Time-based motor starting
simulation with graphical output. Includes reduced voltage and Data Exchange Seamlessly share project data between
capacitor starting, graphical motor and load models. SKM projects and other third-party software. Exchange
format includes Tab Delimited, CSV, and SKM XML. Also includes
HI_WAVE Power Quality Investigation and Filter Design. Frequency
an interface to exchange data with Autodesk Revit and Intergraph
Scan, Harmonic Current, Voltage Distortion, Harmonic Load Flow and
SmartPlant Electrical.
Interactive Filter Design. Evaluate and correct harmonic distortions.
T
THE TECHNICAL ACTIVITIES OF ✔✔ Power System Operation, Plan- explore, and provide solutions to emerg-
the IEEE Power & Energy Society ning, and Economics Commit- ing topics of interest within their subdis-
(PES) are coordinated by the PES Tech- tee—Fran Li, chair ciplines. TCs conduct various technical
nical Council. The current officers of ✔ ✔ Power System Relaying and activities through subcommittees, work-
the PES Technical Council are Vijay Control (PSRC) Committee— ing groups, and task forces.
Vittal, chair; Hong Chen, vice chair; Russ Patterson, chair A detailed description of the various
Diane Watkins, secretary; and Farnoosh ✔✔ Smart Buildings, Loads, and Cus- TCs and their activities can be viewed by
Rahmatian, past chair. The chair of the tomer Systems—Ron Melton, chair navigating to each committee’s website
PES Technical Council also serves as ✔ ✔ Substations Committee—Joe from the PES Technical Council web-
the PES vice president for Technical Gravelle, chair site, https://www.ieee-pes.org/tech
Activities. The PES Technical Council ✔✔ Surge Protective Devices Com- nical-activities. These websites clearly
coordinates the activities of 17 technical mittee—Steven Hensley, chair state the scope of each committee’s ac-
committees (TCs) and four coordinat- ✔✔ Switchgear Committee—Keith tivities and include a flyer that provides
ing committees: Flowers, chair a succinct description of the activities
✔✔ Analytic Methods for Power Sys- ✔ ✔ Transfor mers Com m ittee — of each TC. Many of the committees
tems Committee—Kevin Schnei- Bruce Forsyth, chair have meetings during the PES General
der, chair ✔✔ Transmission and Distribution Meeting, the PES T&D Conference and
✔✔ Electric Machinery Committee (T&D) Committee—Gary Chang, Exposition, and the Joint Technical
(EMC)—John Yagielski, chair chair Committee Meeting. Additionally,
✔✔ Energy Development and Power ✔✔ Energy Internet Coordinating many hold multiple meetings during the
Generation (EDPG) Committee— Committee—Hongbin Sun, chair year to conduct their activities. The
John B. Yale, chair ✔✔ Intelligent Grid and Emerging PES Technical Council website pro-
✔ ✔ Energy Storage and Station- Technologies Coordinating Com- vides updated information on com-
ary Battery Committee—Curtis mittee—Doug Houseman, chair mittee meetings held outside the PES
Ashton, chair ✔✔ Marine Systems Coordinating General Meeting and other conferences.
✔✔ Insulated Conductors Commit- Committee—Dwight Alexan- An important responsibility of the
tee—Henk Geene, chair der, chair PES Technical Council is the design and
✔✔ Nuclear Power Engineering Com- ✔ ✔ Renewable Systems Integra- planning of the technical program for
mittee—Daryl Harmon, chair tion Coordinating Committee— the annual PES General Meeting. This
✔✔ Power System Communications Andrew Leon, chair. includes planning the super sessions and
and Cybersecurity Committee— These committees, which were reor- technical paper and poster sessions. The
Ken Fodero, chair ganized in July 2016, are responsible for TCs coordinate the conference paper re-
✔✔ Power System Dynamic Perfor- the technical activities in each of their views and plan the panel sessions.
mance (PSDP) Committee— subdisciplines within the broad power The IEEE Industry Technical Sup-
Costas Vournas, chair and energy discipline. They conduct port Leadership Committee (ITSLC),
✔✔ Power System Instrumentation technical activities including the devel- formed by the governing board, provides
and Measurements Committee— opment of standards, technical reports, support to and cooperates with govern-
Jim McBride, chair and technical sessions for various PES ments, regulators, and other industry
Digital Object Identifier 10.1109/MPE.2020.3014552
technical meetings—paper, poster, organizations globally on technical is-
Date of current version: 16 October 2020 panel, and super sessions—and identify, sues related to power, energy, and policy
TSAT SSAT
Contact us:
dsainfo@powertechlabs.com
(through the annual IEEE membership had a significant number of downloads.
table 1. A summary of PES
renewal process) or institutional access A summary of the downloads through the Resource Center downloads.
may be required. end of June 2020 is shown in Table 1.
Rank TC
The Resource Center also contains The COVID-19 pandemic and its re-
several technical reports prepared by sulting consequences, coupled with the TCs With the Most Products
various TCs on topics of contemporary evolving explosion of renewable resourc- 1 PSRC: 49
interest and relevance. PES members can es in the grid, have produced unique
download these reports at no cost using technical challenges that have gener- 2 T&D: 46
their IEEE login credentials. Revenue is ated a significant interest in technical 3 EDPG: 44
earned when these reports and documents inquiries to produce solutions that will 4 Power System Analysis,
are downloaded by non-PES members. guarantee the expected reliability and Computing, and Economics: 27
Some of the more recent reports have economy of the electricity infrastructure
5 PSDP: 25
TCs With the Most Product
Downloads
1 ITSLC: 6,931
2 PSDP: 4,740
3 T&D: 3,320
4 PSRC: 2,446
5 EDPG: 1,155
TCs With the Most Revenue
1 T&D: US$1,962
2 ITSLC: US$1,726
3 PSDP: US$1,366
4 PSRC: US$1,296
5 EMC: US$790
Products With the Most
Downloads
1 “Impact of IEEE 1547 Standard
on Smart Inverters” (ITSLC):
1,439
2 “Impact of Inverter Based
Generation on Bulk Power
System Dynamics and Short-
Circuit Performance” (ITSLC):
1,277
3 “Microgrid Stability Definitions,
Analysis, and Modeling”
(PSDP): 1,219
4 “The Definition and
Quantification of Resilience”
(ITSLC): 1,059
5 Tutorial course: Transient
Analysis of Power Systems:
Solution Techniques, Tools,
and Applications (T&D):
1029
W
WHEN THIS COLUMN WAS WRIT- lowed by sustained deployment until (EV) charging stations, energy storage
ten, it is fair to say that most IEEE existing technical limits are met. Given devices, and energy-efficient buildings
Power & Energy Society members have the obvious need for highly reliable and is also increasing steadily. Managing
experienced, or continue to experience, resilient power delivery and the chang- many small resources embedded in the
a significant level of disruption to their ing landscape of not only the “type” of power system is a fundamental challenge
normal work practices. Yet, during this generation but also where it is located compared to managing the traditionally
extended period of wide-scale work- in the overall power system, how can centralized grid. Additional challenges
from-home reality that many of us are we effectively plan and operate the with DERs are the variability and un-
experiencing, the continued availability power system with ever higher levels of certainty in generation and load due to
of critical services has been impressive. DERs? Luckily, there is strong support unforeseen and uncontrolled events. But
Most impressive has been the reliable for high-quality research focused on DERs can also offer many benefits to
supply of power to our homes-turned- integrating distributed PVs and other power systems: they are more scalable,
offices. The integration of high levels DERs into the electric grid worldwide. the generation is closer to the load, and
of distributed energy resources (DERs) For instance, in the United States, the they are more adaptive. As the power sys-
into our power systems, the focus of this U.S. Department of Energy (DOE) has tem relies more and more on networked
issue, is perhaps even more topical in been supporting research in integrating communications and digital technolo-
our current working environment, where distributed PVs and other DERs into the gies and natural disasters such as hurri-
our working style is decentralized—just electric grid with millions of dollars canes, flooding, and wildfires increase in
like DERs compared to centralized gen- invested each year. These research severity and frequency, the exposure to
erators—and overall power system reli- programs address system-level issues cyberattacks and physical hazards also
ability and resilience is more impactful. in DER integration to meet customer increases. Localized DER technologies
With the continued installation of needs while maintaining grid reliability, can help the power system be more resil-
approximately 100 GW per year glob- resilience, and security. These activities ient when a disruption does occur. During
ally, the role of grid-connected photo- are also organized through DOE initia- hazards, the power system can be recon-
voltaics (PVs) in the power system con- tives such as the Grid Modernization figured into independent secure segments
tinues to increase. In the United States, Initiative and the Energy Storage Grand that each contain load and generation.
more than half of the grid-connected Challenge. The DOE collaborates with This can enhance grid resilience to keep
PV capacity is interconnected at the 14 national laboratories and more than critical services online and restore service
distribution level. Although other coun- 100 industry partners to develop the faster with help from the fast-responding
tries have different penetration levels knowledge base and technical capabili- power electronics in solar generation and
of DERs, all are generally experienc- ties needed to modernize the U.S. elec- energy storage systems.
ing higher levels of DER interest and tric infrastructure. This research has This is the seventh issue of IEEE
deployment. A common international helped to accelerate the development Power & Energy Magazine, since 2009,
theme is the rapid growth of DERs and application of DER technologies that has now focused on the challenges
driven by customer demand once the that advance the modernization of en- of integrating DERs into the power
capability to interconnect generation ergy systems in the United States. system at scale. Each of the five theme
on the distribution level is realized, fol- There are more than 2.3 million solar articles included in this issue focus on
generators on the U.S. distribution system multiple elements of DER integration,
Digital Object Identifier 10.1109/MPE.2020.3016100
today, with steady growth expected in the where the DER-centric technology, plan-
Date of current version: 16 October 2020 future. The integration of electric vehicle ning, operating methods, and so on play
EasyPower
ingly higher levels of capable DERs into
the world’s electric power systems.
The evolution of DERs, dominated
As Easy As
Although many challenges remain,
it is interesting to observe how the
most critical concerns of industry and
the highest focus by researchers have
morphed as most utilities gained expe-
rience with DER integration firsthand.
Without a doubt, issues such as voltage
regulation on the distribution circuit are
critical requirements for interconnect-
ed DERs, but other issues have also be-
come critical as DERs have increased
to be a sizeable generation source and
are impactful to system operations. Re-
cently, several studies were completed
that focused on the impact of DERs
CURRENT IN AMPERES X 100 AT 480 VOLTS
.5 .6 .8 1 2 3 4 5 6 7 8 9 10 2 3 4 5 6 7 8 9 100 2 3 4 5 6 7 8 9 1000 2 3 4 5 6 7 8 9 10000
1000 1000
800 TX-2 800
FLA
600 600
500 500
0.15
400 400
100 BL-1 100 Upstream Upstream Electrode Est Arc Flash Working Incident
GE AKR-50 Arc Fault Arc Fault Electrode
80
TX-2 80
Bus Name Bus kV Trip Device Trip Device Equip Type Configuration Gap Boundary Distance Energy
60 1000 / 1288 kVA 1600/1760 60 Name Function (mm) (inches) (inches) (cal/cm2)
50 6% 50
3
BUS-3 13.8 R-6 51/50 Open Air VOA 152 32.1 + 26 1.7
.1
SWG-4
24
40 40
MAIN SWG 3.8 R-18 51/50 Switchgear VCB + HCB 152 189.3 + 18 59.3
BL-1
10 GE MVT-Plus 10 REFINER 2.4 R-7 51/50 Switchgear VCB + HCB 104 261.9 + 18 122.8
17.58
8 Sensor = 1600 8
TIME IN SECONDS
TIME IN SECONDS
Plug = 1600 SWG-4 0.48 R-6 51/50 Switchgear VCB + HCB 32 213.2 + 18 135.1
6 6
Cur
Cur Set = 1.1 (1760A)
LT Band = 1 M-1
19
4 4
R R 3
STPU = 2.5 (4400A)
ST Delay = Int
BL-5
3
MCC-2 0.48 BL-3 MCC VCB 25 53.1 + 18 6.8
C-H HFD
ST Delay I²t = Out PNL-1 0.208 FS-2 Panel VCB + HBB 25 48.2 + 18 7.1
2 Override = 50000A 225/150 2
PNL-2 0.208 PNL-1 Panel VCB 25 57.8 + 18 7.7
1 1
.8 .8
Risk Hazard
.4 Sensor = 800 .4
and Shock
Plug = 800 BL-1
.3 Cur
C ur Set = 0.5 (400A) .3
d
Arc Flash riate PPE Require
18627A
LT Band = 1 C-6
Inst
In st = 4 (3200A)
Arc Flash and Shock
.2 .2
Approp
1 - 400 kcmil CU
TX-2
Appro Risk Hazard
Incident Energ coverallpriate PPE Requi
y
.1 1000 / 1288 kVA .1
Boundary
· Easy to Learn
and frequency events). As you peruse
the articles in this issue, we hope that
you marvel at the comprehensiveness
· Easy to Use
of the integration of DERs and their
respective studies but also identify new
areas of R&D to continue the integra-
· Fast Results
tion of DERs at higher levels.
The first article, “Consumer-Led
Transition: Australia’s World-Leading
Distributed Energy Resource Integra-
tion Efforts,” was prepared by a team Try instantly online or download a free demo copy at:
of authors from across many power www.EasyPower.com/demo
system domains. Naomi Stringer, from
the University of New South Wales,
and her coauthors present a compre- ®
hensive view of the challenges being
Power made easy.
Large Amounts of Distributed Energy nation between the grid cells is structured ushu Electric Power Company, respec-
Resources.” It describes a new concept to reduce communications requirements tively, offer “Making Renewables Work:
for the control and operation of grids with while still seeking to realize optimal op- Operational Practices and Future Chal-
high levels of DERs of all types. The au- eration of the entire power system. The lenges for Renewable Energy as a Major
tonomous energy grid concept realizes a scalability of the proposed concept is dis- Power Source in Japan,” which provides
cellular structure for the grid, with mul- cussed in detail, and examples of large- a detailed view of grid operation chal-
tilayer distributed control implemented scale, lab-based, and real-world imple- lenges related to the high levels of DERs
to enable highly flexible, resilient, and mentations are presented. in the form of distributed PVs and wind.
reliable operation of the grid cells in an Kazuhiko Ogimoto and Hiroshi Wani The article focuses specifically on the is-
optimal manner. Required control coordi- from the University of Tokyo and Ky- land of Kyushu, which has experienced
significantly higher levels of renewable
penetration compared to load than much
of the rest of Japan. A detailed view of
operating practices, both new and old, is
provided, along with an account of how
those practices, combined with large fore-
casting errors, led to an emergency oper-
ating condition in March 2019. Lessons
learned from this event are expertly pre-
sented, and new operating methods, now
employed, are described in detail. Further-
more, a view of future improvements in
power system operation with high levels
of renewables is given via a discussion of
grid constraints and the potential for ad-
ditional flexibility from virtual power
plants and managed EV charging.
Continuing the theme of islands with
high levels of distributed PVs, the fourth
article, “Back in the Race: Achieving
100% Renewable Energy in the Canary
Islands,” by Ricardo Guerrero Lemus
from the Universidad de La Laguna,
Spain, describes the challenges of reach-
ing a 100% renewable energy target by
2040. The article starts with a compre-
hensive overview of the current state
of many types of renewable generation
across the Canary Islands. Specific opera-
tional challenges are highlighted, such as
a highly seasonal, tourist-driven load pro-
file and relatively deep channels between
islands, which make the interconnection
of island systems difficult. Additionally,
the article provides a more detailed dis-
cussion of the operational challenges and
the amazingly high renewable operating
levels realized on both the El Hierro and
Tenerife islands. The integration of a ma-
jor pumped-storage hydropower system
on El Hierro and the lack of any signifi-
cant amount of storage on Tenerife pro-
vides a good comparison of how system
operation requirements can differ, even
within the same region.
The fifth and final theme article is an example of a currently implemented couragement, advice, direction, and edi-
“On Good Behavior: Inverter-Grid Pro- utility solution that increases the visibil- torial prowess while pulling this issue
tections for Integrating Distributed Pho- ity of protection-related issues and com- together. The final development of this
tovoltaics” and was authored by a broad munications-enabled control to support issue occurred during the transition of
team, including Tom Key of the Electric the appropriate mitigation. IEEE Power & Energy Magazine editor-
Power Research Institute, Gefei “Derek” The “In My View” column, “Dis- in-chief leadership, which was efficient
Kou of Dominion Energy, and Michael tributed Generation: Leveraging for and seamless. Thanks to Robert Hen-
Jensen of Pacific Gas and Electric Com- Grid Services,” authored by Tom Bi- derson for his excellent review of the
pany. This expert team takes on the dif- alek from San Diego Gas and Electric feature articles. Thanks also to the entire
ficult topic of system-protection impacts Company, rounds out this DER issue. IEEE Power & Energy Magazine edito-
resulting from the integration of invert- Bialek provides an excellent perspective rial board for their continued support of
er-based DERs. The article starts with a on the often-discussed and researched, this DER issue. Additional thanks are
description of the types of impacts and but not often-realized, provision of grid due to the 23 authors from 13 organi-
the resulting protection concerns that services from DERs. Starting with ob- zations in four countries who supplied
can occur, particularly at high penetra- served grid challenges, implementation articles for this issue. Their hard work
tion levels. Through the presentation of difficulties, DER adoption history in preparing the articles to provide a clear
example protection impacts and utility California, and the recognition of the and understandable picture of their re-
surveys focused on protection concerns, tremendous effort already exerted for search is evidence of their professional-
the article addresses issues such as an effective DER integration, the argument ism and dedication to properly integrat-
open phase on a three-phase line, direct- is made that much more work is needed. ing DERs. Finally, thanks to the readers
transfer trip requirements, and preferred Without the support of many people, of IEEE Power & Energy Magazine. We
interconnection transformer construc- this issue simply would not have been hope you find this issue enjoyable, in-
tion. Further, the solutions for many of possible. Thanks first to Michael Hen- formative, and thought-provoking.
p&e
these concerns are also given, including derson and John Paserba for their en-
800-722-8078 • pr-tech.com
6874
I
IT’S A SUNNY SPRING DAY
in South Australia. A light breeze
is cooling the coastal state capital of Ade-
laide as approximately 260,000 distributed
solar photovoltaic (D-PV) systems on residen-
tial and business rooftops generate electricity,
setting a new state record for the lowest mini-
mum electrical demand for the third time this
season. All looks peaceful throughout the suburbs.
D-PVs form an irregular yet persistent patchwork
across one-third of the houses in the state. Most
consumers are going about their day without
a glance at the D-PV panels on these roofs.
However, the mild temperature and bountiful
sunshine lead the D-PV generation to climb, ©SHUTTERSTOCK.COM/ARTHIMEDES
State of Play
The Australian National Electricity Market (NEM) serves approximately 90% of the population through one of
the longest interconnected power systems in the world (Figure 1). The electricity sector was restructured during
the late 1990s and features a regional wholesale spot market and retail competition. Transmission network service
providers and distribution network service providers (DNSPs) are regulated monopoly businesses, separate from
the generators and suppliers that purchase electricity from the spot market and sell power to consumers through
deregulated retail market arrangements. AEMO is both the independent system operator and the market operator
across the east and west coasts of Australia.
D-PV Uptake
Low PV system prices, excellent solar resources, and a significant proportion of
stand-alone housing in Australia, coupled with high retail electricity prices, have
led to a notable D-PV uptake during the past decade. Australia now has the high-
est proportion of houses with D-PVs worldwide, with approximately one in every
four houses (all dwellings excluding apartments) having installed D-PV equipment
and more than 2.2 million systems in total. The state of Queensland has the high-
est penetration of D-PVs, closely followed by South Australia (Figure 1). However,
the relative contribution of D-PVs to demand is much greater in South Australia
compared to Queensland, given South Australia’s smaller industrial load and mild
climate. The South Australian DNSP is SA Power Networks (SAPN). In Queensland,
there are two DNSPs. Energex serves the metropolitan southeast, and Ergon Energy
Network serves the more rural northern and western consumers; both are part of
the Energy Queensland Group.
While the majority of D-PV systems in Australia are small (with capacities
of lower than 30 kW ac), there is an emerging market in the 30–200-kW range
for commercial and industrial sites. Between 2018 and 2019, the number of new
D-PV applications approved in South Australia alone increased by 50%, and
the high D-PV uptake is widely expected to continue. Across NEM, AEMO has
forecast D-PV generation to more than double by 2039–2040 in all NEM market
regions under a central uptake scenario and to at least triple under a high-uptake sce-
nario. While D-PVs are the most common of the distributed energy resources (DERs)
in Australia, other forms of DERs, including BESSs, are also increasingly prevalent.
However, there remains a limited visibility of D-PV operation and minimal, if any,
remote interoperability.
South Australia
35% 3.2 GW
D-PV Penetration Peak Demand New South Wales
20% 13.8 GW
D-PV Penetration Peak Demand
Tasmania
Victoria 15%
18% 9.3 GW D-PV Penetration
D-PV Penetration Peak Demand 1.7 GW
Peak Demand
figure 1. The Australian NEM regions. The penetration percentages indicate the proportion of stand-alone households
with D-PV generation. The graph shows the D-PV uptake and the installed generation volume according to “legacy” (pre-
2015) and “current” (post-2016) inverter interconnection standards across NEM. Note that Australian and New Zealand
Standard AS/NZS4777.2:2015 for inverters is currently under review.
1,600 80
1,000 50
(MW)
800 40
600 30
400 20
200 10
0 0
12:00 a.m.
12:45 a.m.
1:30 a.m.
2:15 a.m.
3:00 a.m.
3:45 a.m.
4:30 a.m.
5:15 a.m.
5:45 a.m.
6:30 a.m.
7:15 a.m.
8:00 a.m.
8:45 a.m.
9:30 a.m.
10:15 a.m.
11:00 a.m.
11:45 a.m.
12:30 p.m.
1:15 p.m.
2:00 p.m.
2:45 p.m.
3:30 p.m.
4:15 p.m.
5:00 p.m.
5:45 p.m.
6:30 p.m.
7:15 p.m.
8:00 p.m.
8:45 p.m.
9:30 p.m.
10:15 p.m.
11:00 p.m.
11:45 p.m.
figure 2. The D-PV contribution to the minimum net load in South Australia on 10 November 2019.
1,200
(Excluding Major Customers) (MW)
1,000
2006, 953 MW
SAPN Minimum Loads
5 November 2006
800 3 November 2013
2013, 598 MW 2 November 2014
2014, 517 MW 22 November 2015
600
6 November 2016
2015, 507 MW 1 January 2018
400 2016, 454 MW 21 October 2018
January 2018, 361 MW 10 November 2019
2019, 62 MW
0
t
ht
gh
p.
p.
p.
p.
g
a.
a.
a.
ni
ni
00
00
00
id
id
00
00
00
0
M
M
:
3:
6:
9:
3:
9:
12
6:
figure 3. The SAPN minimum load days. Major customers are excluded, while major generation is included. D-PVs are
netted off as a negative load.
Scope
Expanded Scenario Analysis
Subtransmission
Feeder
Distribution
Transformer
Expansion in
LV Network Depth of Planning
to the LV Network
Customer
figure 4. SAPN’s initiative to increase the depth and scope of network planning activities. For further details, please refer
to “Future Network Strategy” in the “For Further Reading” section. LV: low voltage.
Demand (kW)
100
at the distribution transform-
ers appears quite high compared 0
17
17
17
17
17
17
of LV circuits (Figure 6), dem-
20
20
20
20
20
20
20
y
y
ar
ar
ar
ar
ar
ar
ar
onstrate a much lower V1% for
nu
nu
nu
nu
nu
nu
nu
Ja
Ja
Ja
Ja
Ja
Ja
Ja
customers that are distant from
8
10
11
12
13
14
the distribution transformer. The
average customer’s voltage distri- Underlying Demand Demand After D-PVs
Demand After D-PVs and BESSs Modeled VPP Export Event
bution in Queensland is likely to
sit in a narrower range than the
one shown in Figure 6, which is
figure 5. The South Australian Salisbury battery trial. a The existing thermal limit
illustrative of points at the end of
challenges during peak demand (orange). b The existing voltage challenges, with
the LV network. Given the limited high D-PV exports (green). c The reduced peak demand, with batteries, and the
LV network visibility, the samples observed peak during a VPP charging event (purple). d The modeled emerging
in Figures 6 and 7 are not repre- VPP-driven thermal limit challenges during VPP export events (dashed). For further
sentative of population-wide volt- details, please refer to “Low-Voltage Management Business Case” in the “For
age conditions. Further Reading” section.
500 100
216 V (Nominal –6%) 230 V (Nominal) 253 V (Nominal +10%)
450 90
94.9% of
400 V01% V99% ≤253 V 80 Cumulative Percentage of Sites
V50% (Median) 84.5% of
350 V99% 70
V50% ≤244 V
Number of Sites
300 60
250 50
200 40
150 30
100 20
0.7% of
50 V01% ≤216 V
10
0 0
216
218
220
222
224
226
228
230
232
234
236
238
240
242
244
246
248
250
252
254
256
258
260
Voltage
figure 6. Voltage measurements for the LV monitoring of Energex networks in spring (November) 2019. The data are from
~4,000 LV monitors (out of approximately 150,000 LV networks in Queensland). For further details, please refer to “Distri-
bution Annual Planning Report” in the “For Further Reading” section. V50%: 50th voltage percentile.
300 60
250 50
200 40
150 30
4.4% of
100 V01% ≤216 V 20
50 10
0 0
216
218
220
222
224
226
228
230
232
234
236
238
240
242
244
246
248
250
252
254
256
258
260
Voltage
figure 7. Voltage measurements for the LV monitoring of Energex networks during the summer peak week in Febru-
ary 2020. The data are from ~4,000 LV monitors (out of approximately 150,000 LV networks in Queensland). For further
details, please refer to “Distribution Annual Planning Report” in the “For Further Reading” section.
Customers at the ends of LV circuits are exposed only to numbers of nonlegacy D-PVs. The same standard also applies
the lowest voltages on a few days each year, when peak loads to BESSs and can prevent the systems from operating, even
happen. DNSPs are responsible for maintaining voltage when they are attempting to charge, which would otherwise
compliance during peak periods, when maximum demand assist with alleviating overvoltage conditions.
occurs, just as they must maintain compliance during the An analysis of more than 1,300 South Australian sites
spring and autumn months, when loads are modest and the observed by a solar-monitoring company, Solar Analytics,
reasonable solar irradiance results in minimum demand,
events that are likely to be observed significantly more fre-
quently. Figure 7 indicates that during the weeks when max- 6 270
imum demand occurs, there is very limited room for further
widescale reductions in voltage levels. 5 255
Average Power (kW)
4 240
Estimating PV Curtailment Voltage (V)
The curtailment of distributed PV generation due to high 3 225
voltages in the distribution system has received considerable
interest due to the potential lost revenue for consumers with 2 210
PV equipment. In addition, overvoltage curtailment may limit
1 195
opportunities for DER participation in the broader power sys-
tem, for instance, through VPPs. Typically, curtailment occurs 0 180
when local network voltages exceed inverter overvoltage set
.
.
m
a.
a.
p.
p.
p.
p.
30
00
30
00
30
00
30
9:
1:
3:
4:
10
12
Solutions
A ly
ril
e r
br y
ay
M ry
ch
pt ust
em r
r
ne
c r
ov e
ec e
be
O be
Fe uar
Ju
Ap
ua
N tob
D mb
M
ar
Ju
Se ug
em
n
300
Number of Enquiries
250
200
Generation Curtailed (%)
Summer Winter
80 Autumn Spring 150
During the worst-case day, 5% of 100
Estimated
60
the customers experienced the most
50
40 severe impact, with a generation loss of
at least 16%. 0
11 December
12 June
12 December
13 June
13 December
14 June
14 December
15 June
15 December
16 June
16 December
17 June
17 December
18 June
18 December
19 June
19 December
20
0
0 5 10 15 20
Proportion of Sites (All Sites) (%)
figure 9. A PV curtailment distribution, where each data figure 11. SAPN high-voltage/D-PV-related customer
point indicates an individual site that experienced curtail- enquiries per month. Spring is indicated in orange; all
ment in South Australia during 24 clear-sky days in 2018 other seasons are in purple. For further details, please refer
(there are 1,851 distinct sites in the data set, and 983 of to “Low-Voltage Management Business Case” in the “For
them experienced curtailment). Further Reading” section.
november/december 2020
Local: Affecting a Yes MV Augmentation
Single LV Circuit
Yes
Traditional or Traditional Reconductor:
Transformer Nonnetwork Solutions Cost Split 11-kV Feeder and/or
Set to Recommended Tap
on Recommended No Effective? Install MV Regulator
Position
Tap?
No Yes Use Optimal Incremental
Yes Nonnetwork Pricing to Seek Non-Network
Reverse Solutions
Loads and PV Power Likely to
Rebalance LV Network Exceed 50% (?) of Peak Yes
Balanced Across No Load?
Phases?
Yes No
Distributed Installation of
Matched LV Load Widespread Installation of
No VRDT Shunt Solutions
Profile? VRDTs, Regulators,
(e.g., STATCOMs, Stepped
STATCOMS, and BESSs
Capacitators, and BESSs)
Yes LV Augmentation
Traditional Isolated
High X/R?
Solution Practical/Cost No No to One Radial or No
(See Note 1)
Effective? Customer?
figure 12. Current and potential voltage management options framework. Note: X/R is the ratio of the source reactance to the source resistance. VRDT: voltage-regula-
29
to accommodate all additional DERs. This could be larger pilot to extend this capability to smart solar inverters in
achieved through a combination of network augmen- 2021. This will be undertaken in collaboration with another
tation (such as investment in voltage regulators) and DNSP, leading inverter manufacturers, and a gateway vendor.
non-network solutions, such as demand management. SAPN is planning to adopt Standard for Smart Energy Profile
2) It could start applying static connection limits to new Application Protocol, IEEE Standard 2030.5, which is also
DERs, which might result in “zero-export limits” for being applied in California as part of Rule 21. SAPN has lev-
DERs in parts of the network that are at their technical eraged the work undertaken by the standard-setting organi-
capacity (i.e., no exporting is allowed). zation SunSpec Alliance and is identifying which aspects of
3) It could apply dynamic export limits (“flexible export IEEE Standard 2030.5 should be applied in the South Austra-
connections”), meaning that DER export limits could lian context. A 24-h forecast series of export limits is expected
vary depending on actual conditions when there are to be communicated to solar inverters, VPPs, and aggregators
periods of export below the current 5-kW limit and every 5 min on a rolling basis, with the technical architecture
potentially during periods when the export exceeds appearing in Figure 14.
this threshold.
A techno-economic analysis indicated that enabling System Security
dynamic export limits would provide the greatest net eco-
nomic benefit to all consumers, and hence this is SAPN’s Challenges
proposed approach. The application is currently under A growing body of evidence indicates that D-PV generation
review by the Australian Energy Regulator as part of the poses significant system security challenges in NEM, due
revenue determination process conducted every five years (it to the response of inverters en masse during major power
is also referred to as a general rate case in some regions). If system disturbances and the continued “hollowing out” of
dynamic export limits are approved, SAPN plans to imple- the daytime system load. This section focuses on the D-PV
ment flexible export connections for all small and medium disturbance response. However, it is important to note that
DERs by 2023. reduced system load profiles have begun to impact operations
A key component of this solution lies in the establishment in South Australia due to the region’s high share of D-PV
of a standard web interface to communicate dynamic export generation relative to the local underlying demand and weak
limits to smart inverters, aggregators, and VPPs. SAPN is transmission interconnection with neighboring regions, mak-
already undertaking preliminary trials in this area in conjunc- ing the state more susceptible to islanding from NEM. For
tion with VPP operators in South Australia and planning a further details, please refer to “Renewable Integration Study
2018, Actual
Rural, Single Customer (2%) Current 5-kW
2025 Forecast Export Limit
Old Underground (2%)
(Neutral Uptake) Voltage Limit
Small Overhead (2%) PV Penetration
Thermal Limit
Rural Township, Small (3%)
0 1 2 3 4 5 6 7 8 9 10
D-PV Capacity per Customer (kWp)
figure 13. The modeling outputs for the SAPN hosting capacity. The orange bars indicate the D-PV capacity per custom-
er that networks of each type can accommodate before voltage limits are exceeded. The length of the bar reflects the sta-
tistical variability between individual networks of a given type. The purple bars represent the amount of D-PV generation
per customer before thermal limits are reached due to a reverse current flow. The green bars show 1) the average level of
D-PV penetration in 2018 and 2) the forecast level of penetration in 2025. For further details, please refer to “Low-Voltage
Management Business Case” in the “For Further Reading” section. kWp: kilowatts peak.
DER DER
Database
Proprietary
AusNet
DNSP API
DER
figure 14. The SAPN technical architecture for flexible export limit communication. Communication between the DNSP
and the server uses a Representational State Transfer (REST) software architecture, while exchanges between the server
and the DERs or the DER controllers employ IEEE Standard 2030.5. API: application programming interface; OEM: original
equipment manufacturer.
1.1 51
Frequency Peak: 50.866 Hz at 1:11:44 p.m.
Frequency: <50.15 Hz for 60 s
Beginning at 1:26:10 p.m. and
1 Frequency: >50.25 Hz at Ending at 1:27:10 p.m. 50.5
1:11:42 p.m.
(MW/MW at 1:10:55 p.m.)
Frequency (Hz)
0.9 50
Normalized Power
0.8 49.5
0.7 49
0.6 48.5
1:10 p.m.
1:12 p.m.
1:14 p.m.
1:16 p.m.
1:18 p.m.
1:20 p.m.
1:22 p.m.
1:24 p.m.
1:26 p.m.
1:28 p.m.
1:30 p.m.
1:32 p.m.
1:34 p.m.
figure 15. Queensland’s post-2016 inverter-over-frequency-droop response during the 25 August 2018 event. The aver-
age aggregate power is compared with the specified profile. For further details, please refer to “Technical Integration of
Distributed Energy Resources” in the “For Further Reading” section.
1.10 51
Frequency: >50.25 Hz at 1:11:49 p.m., With Frequency: <50.15 Hz for 60 s Beginning
an Initial Peak at 50.45 Hz at 1:11:51 p.m. 1:23:59 p.m. and Ending at 1:24:59 p.m.
Second Peak: 50.46 Hz
1.00 at 1:20:37 p.m. 50.5
(MW/MW at 1:10:55 p.m.)
Normalized Power
Frequency (Hz)
0.90 50
0.80 49.5
0.60 48.5
1:10 p.m.
1:12 p.m.
1:14 p.m.
1:16 p.m.
1:18 p.m.
1:20 p.m.
1:22 p.m.
1:24 p.m.
1:26 p.m.
1:28 p.m.
1:30 p.m.
1:32 p.m.
1:34 p.m.
figure 16. South Australia’s post-2016 inverter-over-frequency-droop response during the 25 August 2018 event. The
average aggregate power is compared with a specified profile. For further details, please refer to “Technical Integration of
Distributed Energy Resources” in the “For Further Reading” section.
400
Upscaled
200
0
.
.
m
Zone 3
p.
a.
a.
a.
a.
p.
p.
p.
p.
00
00
30
(942 km)
0
30
00
30
00
:0
:3
1:
7:
8:
2:
4:
5:
7:
10
11
60
Generation Reduction
Change (%)
40 Proportion Disconnected
t0–tnadir
Zone 2
20
(182 km)
0
Zone 1 Zone 2 Zone 3
Zone 1
(34 km)
0 700
(km)
figure 17. A South Australian voltage disturbance event map and D-PV response profile for 3 March 2017 (upscaled
generation indicates the estimated performance across the South Australian D-PV fleet, based on operational data). For
further details, please refer to Stringer et al. in the “For Further Reading” section.
0.2
Frequency (Hz)
0 50
–0.2
49.9
–0.4
49.8
–0.6
Power Draw to Suppress High Frequency
–0.8 49.7
1:30 p.m. 1:35 p.m. 1:40 p.m. 1:45 p.m. 1:50 p.m. 1:55 p.m. 2:00 p.m. 2:05 p.m. 2:10 p.m. 2:15 p.m.
Average Frequency per Site (Hz) VPP Frequency Response (MW) 49.85 and 50.15 Hz
figure 18. The FCAS response for South Australia’s VPPs on 10 December 2019, with Victoria and South Australia
regional separation. For further details, please refer to “Virtual Power Plant Demonstration: Knowledge Sharing Report
No. 1” in the “For Further Reading” section.
3 300
Battery Discharging During Periods of Elevated Energy Prices
2 200
1 100
Price (AUS$/MWh)
Power (MW)
0 0
–1 –100
–2 –200
Battery Precharging in Anticipation of Elevated Energy Prices
–3 –300
9 January 2020
10 January 2020
11 January 2020
12 January 2020
13 January 2020
14 January 2020
15 January 2020
figure 19. The energy response for South Australia’s VPPs on 9–15 January 2020, showing behavior during one week and
the price-versus-battery dispatch. For further details, please refer to “Virtual Power Plant Demonstration: Knowledge Shar-
ing Report No. 1” in the “For Further Reading” section.
T
THE DRASTIC PRICE REDUC-
tion in variable renewable energy,
such as wind and solar, coupled with
the ease of use of smart technolo-
gies at the consumer level, is driving
dramatic changes to the power sys-
tem that will significantly transform
how power is made, delivered, and
used. Distributed energy resources
(DERs)—which can include solar
photovoltaic (PV), fuel cells, micro- ©SHUTTERSTOCK.COM/ALBERTO MASNOVO
Real-Time Algorithm
Real (P) and Reactive (Q ) Set Points – Actual Value of P and Q at Specific
DER Location Subject to Voltage and
Current Constraints of Location
figure 4. The real-time algorithm tracks the optimal solution. (a) Green dots on the map indicate DER and measurement
locations. (b) The graph shows only the real power (P) set point in red and the actual DER power output in blue for one
DER location.
Additional
Levels
cc cc
Level 3 cc
(Multiple cc
Neighborhoods)
P, Q Setting
P, Q Response
Level 2
(Neighborhood)
P, Q Setting
P, Q Response
Level 1
(Home)
(cc) = Cell Controller
P = Real Power
Q = Reactive Power
figure 5. The communications architecture for distributed and real-time optimization of AEGs. In the figure, Level 1
would be at a home or business, Level 2 would be at a neighborhood, and Level 3 would be multiple neighborhoods,
all on a single distribution circuit.
1
Voltage Magnitude (p.u.)
0.95
0.9
0.85
figure 7. The voltages are controlled within 0.95–1.05 p.u. by the OPF algorithm (green dots).
Conclusion
AEGs of the future will need to control and optimize mil- figure 9. The inverters under test in the ARPA-E NODES
lions of controllable devices in real time. A traditional central experiment. (Source: NREL; used with permission.)
optimization approach to this problem is infeasible because
of the computational cost. Therefore, robust, scalable, and
predictive hierarchical and distributed control algorithms
with provable convergence are needed to optimize the grid
in real time. NREL has developed these scalable algorithms
to enable the proliferation of DERs on a massive scale.
A fundamental underpinning of AEGs is the ability to accu-
rately model the cellular building blocks and their interactions
with the rest of the systems so that control, optimization, and
forecasting methods might be applied in operation. NREL is
also building computational tools that can cosimulate multiple
technologies on the grid to design and evaluate these scalable,
distributed control and optimization algorithms using high-per-
formance computing. In addition, NREL has taken the prelimi-
nary steps of demonstrating these algorithms in real time for
real-world devices in the laboratory and now in smart homes. figure 10. The smart homes in Basalt, Colorado. (Source:
Additional work will be needed in controls, optimization, data NREL; used with permission.)
J
©SHUTTERSTOCK.COM/ANTON_MEDVEDEV
for
JAPAN HAS BEEN EXPERIENCING A RAPID DEPLOYMENT OF PHO-
tovoltaics (PVs) since the Feed-In Tariff (FIT) Scheme for Renewable Energy Renewable
was launched in July 2012. The increasing penetration levels of variable renew-
able generation, especially PVs and wind, have been affecting power system Energy as a
operations in each of Japan’s 10 balancing areas. To accommodate this rapid
growth in variable generation, the government established the Working Group Major Power
on Grid Connection of Renewable Energy in 2014 to continually discuss and
make timely decisions about these operational issues, including capacity con- Source in
nections, operation enhancement and asset improvement in each balancing
area, and inevitable curtailment procedures for renewable generation. Japan
The operation of a power system can be enhanced by the sophisticated dispatch
operation of thermal power plants, energy storage plants, and interconnection,
Renewables Work
Digital Object Identifier 10.1109/MPE.2020.3014744
Date of current version: 16 October 2020
System Operation
Challenges in Kyushu
Solar PV
Renewable Energy 120 Wind 18
Curtailment Geothermal 16%
Biomass
In Japan, the FIT program en 100 Hydro
14.6%
14.3% 15
10.4% 10%
the quality and stability of power
system operations in each balanc- 60 9
ing area, when the supply and
demand balance is in danger, the
40 6
generation of renewable energy
can be curtailed. Renewable gener-
ation can be curtailed in a balanc- 20 3
ing area only after all the thermal
generation, PSH plants, and inter- 0 0
connections have been dispatched 2011 2012 2013 2014 2015 2016 2017
to minimize the restriction under (Fiscal Year)
stability and quality constraints.
Among the renewable energy sys- figure 1. Renewable generation deployment in Japan. (Source: METI; used with
tems, PVs and wind can be cur- permission.)
Okinawa EPCO
(Prefecture Area) Okinawa Standard Hokkaido EPCO
Frequency
Standard 50 Hz
Frequency Tohoku EPCO
60 Hz
Hokuriku EPCO
Chubu EPCO
figure 2. The 10 balancing areas in Japan. Kyushu EPCO (circled) is the focus of this article.
10
11
12
13
14
15
16
17
18
20 ber
20
20
20
20
20
20
20
20
20
20
19
em
Population
October 2016
Electricity Demand
Fiscal Year 2016
PV–Wind
September 2017
1) Weather Satellite Data 2) Insolation Predictions 3) Past (Same Month) 4) Output Prediction by Site
From 47 Sites (Increased Insolation Data
From Eight Sites by and Latest Changes Past Results (Based on Eight Sites)
Satellite Prefecture) Improved Results (Based on 47 Sites)
Correction
Output Prediction
Insolation
figure 6. An improved PV generation output prediction. (Source: Kyushu EPCO; used with permission.)
(a) (b)
(d)
figure 7. An improvement of the PV generation forecast. (a) Satellite insulation data. (b) Historical pattern matching. (c) The
weights for each model. (d) The historical (red line) and 3-h (yellow line) predictions. JST: Japan Standard Time. (Source:
Kyushu EPCO; used with permission.)
times/days that restrictions are imposed on each unit. When sary, using the latest available intraday forecast. Through the
automated, the limitation orders are sent to on-site PV con- improvement, in terms of economy, the expected amount of
trol systems, or the systems download published curtailment curtailment decreases by the reduction of the target restric-
schedules. If the control of a PV unit is not automated, the tion. In terms of security, potential undercurtailment, as a
curtailment orders and schedules are sent to an on-site system result of the day-ahead orders, is treated, if necessary, by the
manager who is responsible for the
oversight of the PV unit, according
to the law.
Dispatch Order
(2 h Before) Parallel In
Curtailment by System
Operation Using the Fourth Axis
Dispatch Order
Average Forecast Error (2 h Before) Parallel In
Since autumn 2019, the renewable Third Axis
Generation Output
3 Stop Signal
3 Stop Signal
Wind
Power Plant
Kyushu
figure 9. The grid stabilization system that guards against failures of the Kanmon interconnection line. (Source: Kyushu
EPCO; used with permission.)
Bank 2
(~March 2016)
PCS–Transformer Bank 1
Package (~February 2016)
6-kV Power-Receiving
Package
Containerized NaS
Battery Unit (252 Units)
66/6-kV Interconnection
100 m • 200 kW/Unit
Transformer
• 50,000 kW Total
(30 MVA, Two Units)
figure 10. The utility-scale BESS in Kyushu (50 MW, 300 MWh). PCS: power conversion system; NaS: sodium–sulfur.
(Source: Kyushu EPCO; used with permission.)
figure 11. The renewable generation management system in the load-dispatching room. (Source: Kyushu EPCO; used
with permission.)
november/december 2020
Adjustment Orders)
Automatic
Load- Intraday: Chief Engineer
Interconnection Operation Confirming Power
Dispatching
Data, Long-Cycle Generation Output
Server
Cross-Regional Frequency
(Three Systems)
Control, Power Source
Integration Intraday:
Sending Control
Orders
Load-Dispatching
Information Server
(Two Systems)
Forecast Renewable Energy Power Plant
Insolation (Higher and Lower Voltages)
Renewable
Data Acquiring
Energy
Adjustment Schedules
Renewable Energy Control Server Renewable Internet
Power Generation (Two Systems) Energy
Curtailment Adjustment Schedule Information
Data Control Server Communication Server
Renewable Energy
(Two Systems) (Two Systems)
Operation Support
Servers (Two Systems)
Chief Engineer
Power Generation
Curtailment Grid Information Output Control
Information Communications Information
Monitoring Renewable Server (Two Systems) Previous Day
Energy Power Operational Information Automatic Telephone and
Generation Communication Server Email Server (Two Systems)
(Two Systems)
55
During the afternoon, a 0.7-GW gas plant was synchronized and less irradiation. To modify the system operation during
3 h after the operator’s order. During the evening, there was the intraday and in near real time when there are PV output
almost no remaining energy stored in PSH. forecast errors, PSH plants with a minimal start-up time of
It is crucial to understand that a reduction in PV genera- several minutes and thermal plants that have relatively short
tion often coincides with increased demand during winter start-up times are useful; however, when the PV capacity
and early spring because of a common root cause: increas- increases, the storage level of PSH plants should be carefully
ing cloud cover resulting in a lower atmospheric temperature monitored and managed.
Operation of
Distribution System
Forecast Error
Online
Curtailment, If Necessary tion increases and the power sup-
Average Day-Ahead Curtailment ply from traditional, large-scale,
Error Offline Curtailment Target centralized power plants decreases,
Nonrenewable Renewable
Energy Generation Generation
12
8
(GW, GWh)
0
.
2 .
3 .
4 .
5 .
6 .
7 .
8 .
9 .
10 m.
11 m.
.
m
m
a.
a.
a.
a.
a.
a.
a.
a.
a.
a.
a.
a.
p.
p.
p.
p.
p.
p.
p.
p.
p.
p.
p.
p.
12
9
10
11
12
2
PHS Generation Demand, Actual Demand, Forecast Thermal
T Power
PV, Actual PV Forecast at PV Forecast at P
PHS Storage
D -Ahead 1,000
Day
Day-Ahead 1 000 D -Ahead 1,600
Day
Day-Ahead 1 600
figure 14. The demand and supply operation on 2 March 2019. (Source: Kyushu EPCO; used with permission.)
Voltage Drop
220/66-kV 66/6-kV
Grid Transformer Demand
Distribution Transformer
Usual Power Flow
Thermal and
Other Power
Plants Power Flow Accompanying Renewable Energy Connection
(In Clear Weather and Similar Conditions)
Voltage Rise
figure 15. The difficulty of voltage management in a distribution grid. (Source: Kyushu EPCO; used with permission.)
6,800 6,800
6,700 6,700
6,600 6,600
6,500 6,500
(V)
(V)
10 30 51
Number : Switch With a Sensor
Load: 2,089 kW Load: 435 kW Load: 304 kW
10 30 51 60 62
56 57 69 4.85 km
LRT 3.96 km 5.62 km 1.72 km 0.93 km
52
1,990 kW
1,260 kW 3.87 km 1,500 kW
11.3 km
15.2 km
17.1 km
(c)
figure 16. The voltage deviation in a distribution system. (a) A sunny day with significant PV generation. (b) A rainy day with
little PV generation. (c) The line diagram. LRT: load ratio control transformer. (Source: Kyushu EPCO; used with permission.)
Biomass (Wood Biomass, ~20 MW) Solar (<10 kw) the curtailment of electricity demand
Price Target for whose production is storable, such
Wind (Onshore) 2025: Market Price
20 Price Target for as EV chargers and HPWHs, might
2030: ¥8–9 provide a good system service
resource to protect against inevitable
15 extreme forecast errors.
Solar (≥10 kw)
We made an 8,760-h produc-
Price Target for tion cost analysis to predict the
10 2025: ¥7 effectiveness of smart charging
by 8.86 million EVs in 10 bal-
ancing areas of Japan in 2030,
5
according to the 2015 long-term
demand and supply outlook.
EVs were categorized into five
0
2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 clusters based on their usage
(Fiscal Year) patterns. Figure 24 gives the
results of the demand and sup-
figure 21. The historical reduction of renewable energy FIT tariffs (US$1 = ¥107). ply analysis to minimize opera-
(Source: METI; used with permission.) tional costs for smart charging
(a) (b)
figure 22. The line capacity allocation under the Japanese Connect and Manage process. (a) The previous grid opera-
tion. (b) The new grid operation. Step 1: Perform a probabilistic evaluation of each generator. Step 2: Release the capacity
through intertripping in emergency conditions. Step 3: Obtain grid access without grid enhancement through the condi-
tions of curtailment. (Source: METI; used with permission.)
0 0
1 5 9 13 17 21
–10 1 5 9 13 17 21
–10
–20 –20
Discharge Limit
Discharge Discharge Discharge Amount
–30 –30
(a) (b)
140 140
120 120
Demand (GWh/h)
0 0
1 5 9 13 17 21 1 5 9 13 17 21
(c) (d)
figure 24. The flexibility provided by the smart charging/discharging of EVs (1 April, 10 balancing areas in Japan).
(a) Charge and discharge of uncontrolled charging. (b) Charge and discharge of controlled charging and discharging in
homes and offices. (c) Load and curtailment of uncontrolled charging. (d) Load and curtailment of controlled charging
and discharging in homes and offices. Clusters (CLs) 1–5: The charge and discharge capacity limits of five EV clusters
according to usage patterns.
8 Oil
Combined Heat and Power
6 Gas
HPWH
4
Capacity (GW)
Coal
EVs
2 Bioenergy
Hydro
0 Pumped Storage
Pumped Generation
–2 Wind
PV
–4
Request for Upward Flexibility
1
3
5
7
9
11
13
15
17
19
21
23
figure 25. An example of an optimum demand and supply analysis, including a balance of tertiary–slow flexibility against
an extreme forecast error in Kyushu in April.
: Ordinary Flow
SVR
: Reverse Flow
Substation
Telecommunication
Link
Voltage Storage
Management Utility-Scale SVC
PVs (Including EVs) Small-Scale PVs
Server
figure 26. A next-generation distribution system. (Source: Kyushu EPCO; used with permission.)
supply. Areas of innovation include renewable generation METI, “The second interim report of the subcommit-
forecasting, renewable generation management, demand and tee on Massive Integration of RE and Next-Generation
supply operations, distribution operations, online PV con- Electric Power Network,” (in Japanese), Tokyo, 2019. [On-
trol, and staff training. Although the issue was not described line]. Available: https://www.meti.go.jp/shingikai/enecho/
much in this article, under the progressing deregulation of denryoku_gas/saisei_kano/20190128_report.html
the business structure of the power industry, the penetration METI, “23th Working Group on Grid Connection of Re-
of renewable generation is also causing many institutional newable Energy: Improvements of RE curtailment process,”
challenges, including frequent ¥0 prices in the wholesale (in Japanese), Tokyo, 2019. [Online]. Available: https://www
market. The increasing penetration levels of renewable gen- .meti.go.jp/shingikai/enecho/shoene_shinene/shin_energy/
eration also bring new issues regarding the reduced opera- keito_wg/pdf/023_07_00.pdf
tion of synchronous machines and exacerbate the problems Y. Iwafune, K. Ogimoto, and H. Azuma, “Integration of
related to supply adequacy. We need to make every effort EV into the electric power system based on results of road
to realize a stable, affordable, sustainable, and safe future trac census, energies-12-01849,” Energies, vol. 12, no. 10,
energy system through efficient planning, investment, and pp. 1–21, 2019. [Online]. Available: https://econpapers.repec
operation by enhancing technologies and institutions. .org/article/gamjeners/
K. Ogimoto, J. Fonseca Jr., K. Kataoka, H. Azuma, and
Acknowledgments A. Isonaga, “General flexibility model analysis for mass PV
The authors are grateful to METI for the provision of materi- deployment, solar integration,” in Proc. 9th Solar and Stor-
als about relevant government council meetings, figures, and age Integration Workshop, Dublin, Ireland, 2019. [Online].
other information. Available: https://solarintegrationworkshop.org/dublin2019/
program/
For Further Reading
K. Ogimoto, K. Ohbayashi, and K. Asano, “Progress and fu- Biographies
ture of Japan’s PV deployment,” in Proc. Solar Integration Kazuhiko Ogimoto is with the University of Tokyo, Japan.
Workshop, No. 272, 2016. Hiroshi Wani is with Kyushu Electric Power Co., Inc.,
Ministry of Economy, Trade and Industry (METI), Fukuoka, Japan.
“Long-term energy supply and demand outlook,” Tokyo,
2015. [Online]. Available: http://www.meti.go.jp/english/
p&e
press/2015/0716_01.html
T
THE CANARY ISLANDS ARE AN ARCHIPELAGO
of 7,447 km2 and 2,127,685 inhabitants (2018), situated
to the northwest of Africa (Figure 1). The islands are an
autonomous community of Spain. There are eight inhab-
ited islands and six isolated power grids, as Fuerteven-
tura, Lanzarote, and La Graciosa are an interconnected
system. The available amount of solar energy is quite
high, reaching values of approximately 1,900 kWh/kilo-
watt peak (kwp)/year in favorable locations. Wind is also
an important renewable resource, mostly during summer.
In 2018, the electricity generated in the Canar-
ies reached 8,841 GWh, with a peak power demand of
1,403.6 MW. Of the electricity generated, 10.5% was
renewable, a 32.2% increase from the value in 2017. Of
the 3,007 MW of capacity installed in the Canaries, 168 MW industry accounts for up to 35% of the electricity consump-
are from photovoltaics (PVs), 429 MW are from wind, and tion in the Canaries’ insular power systems (Figure 1). In
13 MW are from hydro. Conventional power plants are mostly contrast to other popular tourism islands in the EU (e.g.,
combined-cycle units (865 MW), but there are also steam tur- the Balearic and Greek islands), visitors to the Canaries are
bines (483 MW), diesel engines (487 MW), and gas turbines well dispersed throughout the year. Thus, the load profile is
(557 MW). These units are fueled with diesel and heavy oil. relatively flat, and conventional generation is generally con-
Natural gas was approved to be introduced in the islands in stant year round, showing working-day hourly power peaks
1989, but it is opposed by many residents. Further, the only between 1,170 and 1,400 MW in 2018 (Figure 2).
refinery located in the Canaries definitively stopped produc- The average wholesale price of the electricity generated
tion in 2018, mainly for compliance with European Union by the conventional power units greatly depends on the size
(EU) environmental regulations. of the insular system (Table 1). This price includes a fixed
The Canaries are a very important tourist destina- cost, which is related to the availability of each unit to sup-
tion for many Europeans and others. In fact, the Canaries ply electricity during the time considered (the firm capacity),
hosted more than 15.6 million visitors in 2018. The tourism and the variable cost, which is mostly related to the price of
figure 1. The Canary Islands are located in the Atlantic Ocean. (Source: Uwe Dedering, https://es.m.wikipedia.org/wiki/
Archivo:North_Atlantic_Ocean_laea_relief_location_map.jpg.)
2 × 66 kV
50 MVA
578 MW
: 220 kV
3,514 GWh 549 MW
: 66 kV
196 MW 3,403 GWh
106 MW 173 MW
200 MW 36 MW
8 MW
200 MW
42 WWh
11.5 MW
0.031 MW
11.32 MW
figure 3. The transmission circuits and main parameters of the Canaries’ power grids. The box with the dark-red outline
is the key to the information conveyed by the boxes with light-red borders; the red text indicates pumping systems that are
being studied. The yellow boxes give the main parameters of interconnections that are in service or under review. Only
the repowering of the interconnection between Fuerteventura and Lanzarote is expected during the short term. (Source:
REE and the Spanish government; author’s elaboration.)
714
45.8
50
664
700
39.2
553
560
600 40
500
Total (MW)
Total (MW) 30
400
303
260
256
252
17.1
300
227
210
11.8
20
11.3
12.2
12.1
144
11.8
200 11.3
122
110
71
5.33
100 10
42
8
4.6
2.38
0
0.4
0.4
Lanzarote Gran Tenerife 0
Fuerteventura Canaria La Palma Gomera El Hierro
(a) (b)
figure 4. The main PV and wind capacity indicators for the power systems in (a) Lanzarote, Fuerteventura, Gran
Canaria, and Tenerife and (b) La Palma, Gomera, and El Hierro. (Source: REE and the Canary Islands government;
author’s elaboration.)
–3,700 m
–30 m
–1,022 m
–2,272 m
–1,555 m
–2,668 m
–3,065 m
figure 5. A bathymetric map of the Canary Islands, including the minimum depth between islands. (Source: Canary
Islands government.)
300
59.97% 0.75
200
PV Capacity:
100 107.1 MW
0.7
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
Time (h) Time (h)
Wind Capacity:
198 MW (a) (b)
figure 6. (a) Generating curves for wind, PV, and total energy produced on 19 May 2019 (record instanteous genera-
tion at 4:20 p.m.); (b) evolution of carbon emissions per MWh produced; and (c) detail of the generating curves for each
conventional power technology on 19 May 2019.
Tenerife: The Record for Renewable table 4. The conventional power units’ average
Electricity Injected Into the System flexibility in Tenerife.
As of October 2019, Tenerife held the record for the high-
Ramp-Up to Full
est instantaneous penetration of renewable energy (59.97%) Technology Hot Start (min) Load (min)
for a power system without any storage capacity (no pump-
Steam turbine 285 46
ing, batteries, or any other storage system serving the power
grid) in the Canaries. This was recorded on 19 May 2019 Diesel engine 23–68 4–11
at 4:20 p.m. (Figure 6). As expected, the CO2 emissions Combined cycle 330 66
per megawatt hour generated by conventional power units
Gas turbine 24 9
increase when the renewable supply reaches its highest val-
ues, but they grow by only a few percentage points com-
pared to other time intervals when the generating activity community self-consumption, and the connection between
is mostly supported by conventional power units. That is production and consumption sites can be used in an internal
partly because of the time intervals required by conventional grid as well as the distribution grid, considering some very
generating units to reach hot starts and ramp-ups (Table 4). innovative rules about the distance between feeders and con-
Lower CO2 emissions per kilowatt hour produced can be sumption locations. The regulation is very simple for distrib-
expected if new dispatchable power units with faster starts uted generation that has a nameplate capacity in kilowatts
and ramps replace the current conventional power units. In peak of fewer than 10, 15, or 100.
fact, EU Directive 2010/75, which covers industrial emis-
sions, limits the operating hours of many important conven- El Hierro: The Highest Penetration of
tional power units. Some of these are retrofitted, but others Renewable Electricity in the Canary Islands
will be decommissioned and replaced by new firm capacity. There is general consensus in the Canary Islands that large
In addition, the future of renewable energy in the Canary amounts of storage capacity are needed to increase the share
Islands will be connected to the development of Royal of renewable energy and reach a 100% renewable power
Decree 244/2019 about energy self-consumption (including system. At this time, however, there is discussion about the
net metering and selling surplus energy to the power sys- introduction of pumping systems and/or batteries (station-
tem). This regulation covers all of Spain, but it is particularly ary and those used by electric vehicles). The Gorona del
important for the Canary Islands, which constitute the Span- Viento hydro-pumping system on El Hierro island, which
ish region with the greatest solar irradiation rates and the has been operating since mid-2014 (Figure 7), plays a cen-
highest cost of electricity. The decree includes individual and tral role for the first option. The system includes five wind
Valves
Wind Farm
Pipelines
Downhill Reservoir
(a)
70
Annual Electricity Supplied to El Hierro (%)
60 56.5
50 46.5
40.68
40
30
19.39
20
10
2.32
0
2014 2015 2016 2017 2018
Year
(b)
figure 7. (a) The Gorona del Viento power plant at El Hierro. (b) The percentage of electricity produced by Gorona del
Viento per year. (Source: Gorona del Viento.)
Substation
figure 8. The location of two all-sky imagers and (b) the substation in the southeast of Tenerife. The monitored multi-
megawatt PV plants are circled in red, the all-sky imagers are indicated by red arrows, and the substation is circled in
purple. (Source of map image: Canary Islands government, www.idecanarias.es.)
precise PV forecasting since it is very difficult to obtain low “Plan Energético de Canarias (PECAN),” Gobierno de
error rates on partly cloudy days but easy to serve cautionary Canarias. Canary Islands, Spain. Accessed: Aug. 31, 2020.
information, including quantitative analyses, to conventional [Online]. Available: https://www.gobiernodecanarias.org/
power units, which is helpful for saving fuel costs. energia/temas/planificacion/pecan/index.html
“Estrategia Energética de Canarias 2015-2025,” Cabildo de
Conclusions Lanzarote, 2017. [Online]. Available: http://www.datosdelan
The Canary Islands are back in the race to reach 100% zarote.com/itemDetalles.asp?idFamilia=18&idItem=7848
renewable energy because solar and wind resources are “Declaración de Emergencia Climática en la Comu-
abundant. The new government plans to reach the total nidad Autónoma de Canarias,” Gobierno de Canarias,
decarbonization of the economy by 2040. The most impor- Spain, 2019. Accessed: Aug. 31, 2020. [Online]. Available:
tant challenge—a robust regional energy plan backed https://www3.gobiernodecanarias.org/noticias/wp-content/
by the Spanish government that describes the necessary uploads/2019/08/190830_Acuerdo_Declaracion_Emergencia_
steps to reach this goal—is being confronted. Important Climatica.pdf location: Canary Islands date: 08/31/2020
questions about the optimal ratio of distributed-versus-util- “Desarrollo de las redes de transporte de electricidad 2015-
ity-scale generation, how to guarantee a 24/7 energy supply, 2020, Spain.” Accessed: Aug. 31, 2020. [Online]. Available:
how to store energy in the different islands, where to place https://energia.gob.es/planificacion/Planificacionelectricid
new energy infrastructure, how to modify the current regula- adygas/desarrollo2015-2020/Paginas/desarrollo.aspx Min-
tions to accelerate the permit processes, how to attract private isterio para la Transición Ecológica y el Reto Demográfico
investment, and how to create permanent local technology A. Marrero et al., “Gorona del Viento Wind-Hydro Power
and business sectors that support this transition and expand it Plant: Results, improvement actuations and next steps,” in Proc.
to neighboring countries in Africa will be answered shortly. 3rd Int. Hybrid Power Systems Workshop, Tenerife, Spain,
May 8–9 2018. [Online]. Available: https://hybridpowersystems
For Further Reading .org/wp-content/uploads/sites/9/2018/05/1_2_TENE18_036
I. Nuez and J. Osorio, “Calculation of tourist sector electric- _paper_Marrero_Agustin.pdf
ity consumption and its cost in subsidised insular electrical
systems: The case of the Canary Islands, Spain,” Energy Biography
Policy, vol. 132, pp. 839–853, Sept. 2019. doi: 10.1016/j.en- Ricardo Guerrero Lemus is with Universidad de La Lagu-
pol.2019.06.032. na, Tenerife, Canary Islands. p&e
On Good Inverter-Grid
Protections
for Integrating
Behavior Distributed
Photovoltaics
T
THE INTEGRATION OF INCREASING NUMBERS OF to feeder open-phase events, and islanding-prevention and
larger photovoltaic (PV) power plants brings several chal- reclosing practices.
lenges to traditional distribution system protection. Two We draw on field experiences to identify typical circum-
major considerations are managing back-feed and coordinat- stances and issues as well as suggest practices to avoid some
ing the inverter response to abnormal conditions. problems and options to mitigate others. In addition, we
With the 2018 updates to IEEE Standard 1547, distrib- discuss DER power monitoring and data collection both for
uted energy resources (DERs) must be able to support the visibility and as modeling input to predict performance. We
electric grid during both steady-state and abnormal condi- also identify the need for analytic tools to screen islanding
tions. The need to continue operating during voltage and risk, verify settings/responses, and evaluate grounding.
frequency events has focused attention on protection coordi-
nation. In this article, we address several practical protection Overvoltage and PV Plants
issues with inverter-connected PV plants, including tempo- When islanded from the primary system ground, DERs can
rary overvoltage, fault contribution and arc flash, response contribute to a ground fault overvoltage (GFO) on unfaulted
Power & Energy Society Technical Report 21. This new face
0
on an old issue is expected to continue evolving with more
inverter-interfaced PV and energy storage.
Another overvoltage concern for inverters is their response
–50 to a sudden loss of load. Load-rejection overvoltage (LRO) hap-
13.5 13.55 13.6 13.65 13.7 13.75 13.8 13.85 pens when a grid-following inverter is disconnected, e.g.,
Time (s) due to breaker operation. Overvoltages occur because there is a
temporary surplus of generation. A recent study by Dominion
figure 2. The LRO measured at a utility-scale solar power Energy on utility-scale solar power plants recorded elevated
plant. Grid disconnection happened at 13.5 s. voltage levels upon the sudden disconnection of PV plants.
Figure 2 shows an example captured at a typical utility-
phases in a four-wire system. Effective grounding is intended scale, 5-MW plant at 34.5 kV. The transient overvoltage at
to limit the line-to-ground overvoltage to 0.8 Vl–l. Maintain- this plant reached 159% of the nominal voltage and lasted
ing this coefficient of grounding is a classic design objective 12 cycles. Such a phenomenon is not uncommon with rapid
in three-phase power systems and defined in IEEE/ANSI load rejection. Among 1,128 load-rejection events captured,
Standard C62.92. 12% exhibited root mean square voltage magnitudes above
For distributed generation (DG), the concern for GFO is 140% of the nominal, and most events lasted between one
a key factor in selecting the grid connection transformer. In and six cycles.
the case of synchronous machine DERs, connections are nor- Overvoltages exceeding equipment withstand capabil-
mally required to provide supplemental system grounding. For ity could lead to medium-voltage (MV) arrester and instru-
inverter-connected DERs, the need is less clear because of ment transformer failures as well as load equipment damage.
Equipment failure incidents associated with PV plant LRO generators for all three sequence components. This illus-
have been reported, and choosing plant interconnection trates why inverters need to be treated differently in any
equipment capable of handling transient overvoltages is a rec- short circuit modeling and simulation platform.
ommended practice. For example, heavy-duty-rated surge The collective efforts of the industry have greatly
arresters and potential transformers with higher basic insu- imp roved the modeling techniques of inverter-based
lation levels are more likely to tolerate LRO. generation. Currently, most short circuit simulation plat-
In the meantime, inverter overvoltage limits were applied in forms—including ASPEN, CAPE, and ETAP—provide
Hawaii and have been added to IEEE Standard 1547-2018. for inverter modeling. Solar inverters are modeled as a
These were confirmed by National Renewable Energy Labo- voltage-controlled current source. By populating a table
ratory and Southern California Edison lab testing of invert- listing fault-current magnitude and angles at various ter-
ers. EPRI is also conducting research for the New York minal voltages, fault characteristics are incorporated into
State Energy Research and Development Authority to bet- the analysis.
ter understand how LRO combines with GFO during unbal- Still, precautions are required for solar inverter modeling.
anced faults. This work included the role of island load and In addition to inherent current limits, solar inverters respond
MV surge arresters. to faults by preprogrammed control modes. For example,
Looking ahead, inverter LRO and GFO are expected to unity power factor control mode will behave differently than
be largely mitigated with the application of the new IEEE reactive support control mode. Therefore, it is recommended
Standard 1547-2018. An inverter certified under the standard that utilities verify inverter settings during the commission-
“shall not cause fundamental frequency line-to-ground volt- ing process and validate model performance through ongo-
age to exceed 138% for a duration exceeding one cycle.” ing field measurements.
worker practices. VC
20
15
PV Plant Fault Contribution
10
An accurate short circuit representation is key to protection
5
coordination and relay settings. Unlike synchronous genera-
0
tors, solar inverters act as a positive-sequence source and, 0 0.05 0.1 0.15 0.2
depending on design, have insignificant negative- and zero- Time (s)
sequence fault current. Figure 3 shows a single-phase-to- (a)
ground fault near a 20-MW solar PV power plant. The faulted 350
phase dropped to 35% of the nominal voltage. In response, 300 I1
the inverters contributed to the location fault current. 250 I2
Current (A)
Single-Phasing
Happens
Area Exposed to Open Phase
Given three-phase feeders, what
Connected Single-Phase role will DERs play? Incidents of
Load Downstream of F01 open phase (or single phasing on
three-phase distribution circuits)
F01 can happen. This brings several
concerns for PV plants. Depending
C Phase
Opened PV POI Protection on the transformer configuration
Water Treatment R (No Customer and DER output to load, voltage
Facility Loads) on the open phase might be regen-
erated. During the open-phase
Substation
incident, this leads to poor regula-
tion or elevated voltages and can
potentially expose the public to a
live downed conductor.
These events are not rare in
a distribution system with inde-
pendent phase protection, such
Two of Five 1-MVA PV as single-phase reclosers. They
Sites Energized at
Time of Events can happen in a single-phase fault
clearing if a single switch or fuse
is open or when a conductor is
figure 5. An example of an open-phase event upstream from a PV plant. broken. Figure 5 shows the case of
F01: phase C line fuse opened; R: site recloser. a blown fuse.
V Interconnection
Inverter ac Line Transformer
UG Cable
Filter Capacitor Impedance Impedance BK1, Phase C
BK3 Open
BK2
Closed Breaker
V Voltage Meter GT
figure 6. A plant layout, from the feeder main line on the right to the PV plant on the left, with its protection devices (BK)
and underground (UG) cable connection.
Island
Opens on
Detection
DER
Breaker/
Fuses
Open Phase
Substation
figure 7. An open-phase incident (midfeeder) detected by the downstream PV plant. (Source: EPRI.)
Ineg
Inverter
Trips Inverter
Trips
Vneg_trip
td
t Vneg
Vneg_trip
Open Phase Event
Vneg_db
(a) (b)
figure 8. Open-phase passive detection by (a) monitoring the reduction in negative-sequence voltage or (b) actively
injecting negative-sequence current.
and are limited to voltage and/or frequency measurements table 3. The effectiveness before and after IEEE
without positive feedback. Standard 1547-2018 depends on the transformer.
Yg/Yg Yg/Y
Island-Prevention Practices
The common practice has been to augment local detection Open-Phase
Voltage Transformer
with DTT when an upstream protective device opens. DTT Cases Type 5 leg 5 leg Yg/T T/Yg
(or transfer-trip) systems are used to quickly disconnect
Passive IEEE Good Good Poor Good
DERs when a feeder breaker opens, usually at the substa- detection 1547A-2014
tion; see Figure 9.
Island prevention using utility-controlled DTT is reli- IEEE Poor Poor Poor Poor
1547-2018
able and effective, but it is typically expensive for individual
sites. It can be seriously impractical for high numbers of Active IEEE Better Better Good Better
detection 1547A-2014
DER sites. Communications alternatives to copper and fiber
are being considered, such as power line carrier and cellular. IEEE Good Good Good Good
Also, practices extending the time for reclosing a main-line 1547-2018
device are used in some areas with high DER penetration.
Current practices to address DER islanding risk vary among auto-reclosing is blocked if the voltage is ≥15% of the
utilities, as shown in Figure 10. nominal voltage during the auto-reclosing attempt.
Pacific Gas and Electric Company (PG&E) and Dominion PG&E also evaluates aggregate DERs at substa-
Energy participated in the survey, and the authors provided tions and transmission line sections to ensure that the
these additional details on their practices. total does not present an islanding risk. This has been
✔✔ In California, all investor-owned utilities follow Cali- a primary concern because of high aggregate DER
fornia’s Rule 21 for screening islanding concerns, penetration and back-feeding into the transmission
even though risk varies among utilities, plant types, system (with related GFO concerns). DTT is normally
and feeder characteristics. For PG&E, supervisory required for all synchronous machines. These are com-
control and data acquisition (SCADA) system-con- mon to power plants in the San Joaquin Valley and
trolled reclosers have been used in cases of larger small hydropower units in the Sierra Nevada. For PV
individual PV plants connected to a 12.5-kV distri- plants with onboard islanding-detection systems and
bution system. Although this approach is slower than certified to UL 1741SA, DTT is not required in most
conventional DTT, it provides cost-effective telemetry cases. These criteria may be updated based on ongo-
and plant disconnection for islanding. Reclosing is de- ing review efforts as described in the Working Group 4
layed for 10 s to coordinate with the SCADA system Final Report (see “For Further Reading”).
and reduce the risk of an out-of-phase connection. In ✔✔ In U.S. states Virginia and North Carolina, Domin-
addition, voltage reclose blocking might be applied on ion Energy has experienced the deployment of many
some devices. In this case, the line-voltage magnitude large PV plant connections on its 34.5-kV distribution
is checked to verify that it is de-energized, and the feeder system. Islanding prevention is a high priority,
Relay
DER
Substation
Island
Laboratories (United States) have considered IEEE Stan- protection-grade DTT scheme. It is important that relays are
dard 1547-2018 grid support functions and fault ride- set to respect DER plant ride-through requirements.
through requirements. In another recent survey, we asked for current practices/
These studies found that some detection methods work relay settings at the DER plants. As shown in Figure 11,
better than others. The key outcomes were the definitions of there were two main responses.
generic islanding-detection methods. The studies addressed ✔✔ The recloser is used for overvoltage/undervoltage and
the performance evaluation of each method, including sen- frequency trips, disconnection for momentary power
sitivity to DER mix, system loading, and feeder characteris- loss, monitoring plant status, and ensuring time delay
tics. Ride-through and grid support functions were found to for re-energizing (48%).
have a limited impact on the performance of active islanding ✔✔ There is limited use of the recloser, mainly for on/off
detection methods. by the operator, not tripping in the case of grid momen-
Ride-through requirements have made previously devel- tary events or power loss, and to allow for immediate
oped Sandia guidelines for the assessment of unintended energization with the restoration of grid voltage (30%).
islanding risk less applicable with increasing DER penetra- A good example of coordinating site recloser relays with
tion levels. In the past, 67% of generation/load and 1% var DER site intertie protection functions is provided in Figure 12,
(reactive power) unbalance were reasonable criteria. With courtesy of Southern Company Service. Tie-line reclosers
IEEE 1547-2018 category III ride-through, simple screen- are required by Southern for any plant of 1 MW or more (a
ing would require a conservative 25% of generation/load very common threshold). When the DERs are connected via
and 7% var unbalance criteria. a low-side delta transformer or a zigzag GT, the DTT might
This work indicates that two things need to happen also be integrated with the site recloser. For all cases, the
for the acceptance of onboard protection as an alternative settings allow for overvoltage/undervoltage and frequency
to DTT. The first is for inverter manufacturers to identify ride-through.
which generic detection methods they are using. The second Finally, when DERs are placed into grid service, they
is for utilities to update island risk-assessment methods to become a functioning part of the grid. This implies that
account for new ride-through requirements. Certifica- the utility system depends on the DERs. The accuracy and
tion is necessary but not suffi-
cient. Assessments need to con-
sider plant location and DER type, Pick a Description That Best Describes Your Utility’s
load/generation balance, number Practices for Controlling Reclosers Connecting DER Plants
and type of grid loss events, and
detection method details. Recloser Is Used for O/U V and
F Trip; Disconnect for Feeder Outages A 48%
and Time Delay on Re-Energization
Coordination of Plant Site
Recloser Practices Recloser Is Used for O/U V and F Trip;
B 30%
Directly related to DTT and DER Disconnect for Feeder Outages
ride-through, there is the need
to coordinate relays that con- Typically Limit the Use of Recloser to
C 11%
trol plant-to-grid interconnection On/Off Switch for Operator and/or DTT
reclosers. Plant site multifunction
relays installed with reclosers Only Require Recloser If DTT D
have the full range of protection Is Specified
setting options. Typically, these
reclosers are required for opera- Avoid DER Reclosers Requiring
E 11%
tional on/off control of larger Utility Control
DER facilities. We found that, in
many cases, they are also used
as the disconnection means for a figure 11. The practices applied to DER site reclosers, with 76 utilities responding.
PCC
Feeder 1 N.O. Feeder 2
B B
Utility PoC
IEEE 1547-2018 DER Default
DER Site
Trip Limits (by ANSI):
? ? Draw-Out CB/Fused Switch
27 – 2 = 0.5 p.u. + 9.6 Cycles
27 – 1 = 0.88 p.u. + 120 Cycles
27
79
81 O/U = 60.5/57 Hz + 12 Cycles
DER
Site 79 = 300 s
81
AAI
Load
figure 12. The relays per ANSI Standard C37.2. Note that site relays might be integrated in inverter DERs. (Source: South-
ern Company Services.) AAI: active anti-islanding; PCC: point of common coupling with utility feeder; PoC: point of DER
connection; N.O.: normally open; M: meter; CB: circuit breaker; O/U: over/under; LL: live line; DB: dead bus; B: breaker.
C
CHANGES IN THE DEMAND PROFILE
and a growing role for renewable and
distributed generation are leading to
rapid evolution in the electric grid. These
changes are beginning to considerably
strain the transmission and distribu-
tion infrastructure. Utilities are increas-
ingly recognizing that the integration of
energy storage in the grid infrastructure
will help manage intermittency and improve
grid reliability. This recognition, coupled with
the proliferation of state-level renewable portfolio
standards and rapidly declining lithium-ion (Li-ion)
battery costs, has led to a surge in the deployment of
battery energy storage systems (BESSs). Although BESSs
represented less than 1% of grid-scale energy storage in the
United States in 2019, they are the preferred technology to meet grow-
ing demand because they are modular, scalable, and easy to deploy across
diverse use cases and geographic locations.
As the number of BESS installations has increased, system integrators, utilities, govern-
ment bodies, and professional organizations have put considerable effort into developing safety standards
and best practices for engineering and commissioning. However, some recent safety incidents in the field have
led to total BESS destruction and posed risks to first responders. Despite the efforts of the energy stor-
age industry to improve system safety, recent incidents show the need for a greater recognition of the
limitations of current practices. For example, much of the effort has focused on improving safety at
the cell and battery pack level. Additionally, risks that manifest during operation and the catastrophic
failures arising from operator error or component failures have not received as much attention as the
initial factory testing and commissioning of BESSs.
This article advocates the use of predictive maintenance of operational BESSs as the next step in
safely managing ESSs. Predictive maintenance involves monitoring the components of a system for
changes in operating parameters, which may be indicative of a pending fault. These changes signal
the need for maintenance while the fault is still recoverable. Many industries, including
utilities, use this maintenance approach for assets such as power plants, wind
turbines, oil pipelines, and photovoltaic (PV) systems. However, this
approach has yet to be fully explored and utilized for BESSs.
Predictive monitoring is complementary to and should not
replace safer system designs, which are essential for the
real-time mitigation of catastrophic failures. How-
ever, when applied to BESSs, predictive moni-
toring can initiate actions that potentially pre-
vent catastrophic failures from occurring.
This article reviews current safety prac-
tices in BESS development, provides
examples of predictive maintenance
approaches in other industries, notes
the key components of an effective
approach, and describes the meth-
odologies used to identify leading
fault indicators.
Current
Recommendations
and Standards
for Energy Storage Safety
Between 2011 and 2013, several major
grid energy storage installations experi-
enced fires (see Figure 1). As a result, lead-
ing energy storage industry experts recognized
that technologies and installations were beginning
to outpace existing standards. In addition, although
several energy storage technologies were available in the
marketplace, Li-ion-based storage systems made up an increas-
ing number of the installations. Of even greater importance was that
the deployments were beginning to grow faster, especially in behind-the-meter
©SHUTTERSTOCK.COM/PETOVARGA residential and commercial applications. As such, a stronger focus on the safety of
Li-ion-based storage systems took hold due to the fire potential of the batteries.
In 2014, the U.S. Department of Energy (DOE), in collaboration with utilities and first responders,
created the Energy Storage Safety Initiative. The focus of the initiative included “coordinating DOE
Energy Storage Systems Safety Working Groups with over 150 stakeholders from industries such
as electric utilities, standards organizations, and manufacturing companies.” These working groups
September August January February July January July April January February
Tsukuba, Japan Oahu, HI USA DOE OE Energy ANSI UL 9540 DNV GL IEC TS Publication of South Korea IEC 62932-2-2
Sodium-Sulfur Battery Room Boeing Storage Safety First-Edition GRIDSTOR 62933-5-1:2017 “Energy Storage South Korean Published (Safety
Government of Flow Battery
figure 1. A timeline of grid energy storage safety, including incidents, codes and standards, and other safety guidance. DOE: U.S. Department of Energy; FAA: Federal
Aviation Administration; OE: Office of Electricity; PAR: project authorization request; NFPA: National Fire Protection Association.
november/december 2020
and National Fire Protection
United States are UL 9540
most notable standards in the
tive maintenance. Two of the
nificant guidance on whole-
ing company) which, in 2018,
Other efforts included a col-
ing safety agencies and first
for Advanced Energy Storage Technologies (NLAB) of the 3) Faulty installations: Human error during installation
National Institute of Technology and Evaluation in Japan. This could have led to system faults, resulting in ESS fires.
NLAB Large Chamber is used to test containers up to 53 ft 4) A lack of ESS integrated control and protection systems:
(16 m) in length under controlled thermal and wind velocity Gaps in the integration of the BMS and energy man-
conditions (the first facility in the world to do so). agement system (EMS) may have caused the fires.
The guidelines under development include IEEE Stan- The conclusions of the investigation raise the following
dard P2686, Recommended Practice for Battery Manage- question: When it comes to the next stage of failure analysis
ment Systems in Energy Storage Applications (set for balloting for ESSs, how can the industry further improve operations to
in 2022). This recommended practice includes information reduce incidents in the field? Some of the issues noted in the
on the design, installation, and configuration of battery man- South Korea investigation were not captured by standards,
agement systems (BMSs) in stationary applications. The and there was no mechanism for identifying and fixing prob-
document also covers battery management hardware lems or design issues after the installation.
(e.g., grounding and isolation), software (e.g., algorithms Currently, the industry certifies ESSs based on defined
for optimal control), and configuration. More recently, the sets of codes and standards. This certification focuses on the
Modular Energy Storage Architecture (MESA) Standards overall design review of the core ESS, testing for adherence
Alliance, consisting of electric utilities and energy stor- to standards before shipment, and commissioning once the
age technology providers, has worked to encourage the use unit is installed in the field. Ideally, the certification process
of communication standards and advanced interoperability as ensures that the overall system design is sound, the factory
well as the reduction of the engineering effort required to inte- testing makes certain that the unit was constructed correctly,
grate an ESS into a utility. MESA is developing two standards: and the commissioning test confirms that there were no faults
one that defines the communication between ESS components created or discovered immediately after the unit was installed
and another that defines the communication requirements for at the site. Nevertheless, gaps remain in maintaining the unit
utility-scale ESSs. These standards include parameters for after installation and identifying potential failures that may
inverters, meters, general ESSs, BESSs, and Li-ion BESSs occur in the longer term. In short, there is not much guidance
under various operations. on what to do on day 2 after a project is completed.
Continuous monitoring of the system post-installation is dictive maintenance is carried out when needed based on the
needed to facilitate maintenance and ensure that problems actual condition of the equipment. Components are monitored
are identified and addressed before they can lead to cascading for changes in the operating parameters that may be indicative
failures. Systems can be monitored by a BMS, but designs are of a pending fault, and these changes prompt intervention.
not standardized, and owners/operators may not have ready Some organizations have offered general guidance on
access to critical information. Also, the inability of manage- preventive maintenance for BESSs. For example, an Energy
ment systems to “connect the dots” among large quantities Storage Safety 101 presentation during a May 2020 meet-
of data may cause systems to fail. IEEE P2686 may address ing of the California Energy Storage Alliance recommended
some of these gaps. Still, current failure-response mecha- semiannual steps such as visually inspecting the overall
nisms usually lead to total BESS destruction. ESS hazard- system; examining heating, ventilation, and air condition-
mitigation techniques are primarily designed to protect ing (cooling); and checking ESS software control and com-
human safety, which certainly needs to be the focus. These munications. An annual process similar to commissioning
responses (e.g., water quenching) will often render the system was also proposed. A 2019 Energy Storage News report on
unrecoverable, making the mitigation just as catastrophic (in operations and maintenance noted that the Smarter Network
a technical sense) as the initial event. Thus, we advocate the Storage Project, a 6-MW/10-MWh battery system, receives
development of a framework for the predictive maintenance a six-month checkup to ensure optimal performance (includ-
of operational BESSs as the next critical step in the safe ing identifying battery degradation levels, pushing software
deployment of ESSs. upgrades, and inspecting the power conversion system). In
the same report, a representative from an ESS integrator
Improving Operations Through noted that much of their maintenance involved software
Predictive Maintenance updates. BMSs implement safety functions and controls that
Preventive and predictive maintenance are mature concepts depend on algorithms, sensor data, and system parameters.
for operational systems in industry. Operators complete pre- Furthermore, BMSs and inverters must communicate to
ventive maintenance on a routine or timed schedule (weekly, coordinate control actions and responses to faults and warn-
monthly, annually, and so forth) based on the average or ings. Therefore, any software or firmware update glitches
expected lifetime statistics for equipment. By contrast, pre- in either of those components can impact the effectiveness
of safety features, leading to potential BMS malfunction
and damage to batteries. Periodic software patching also
ensures that systems are protected from known cybersecu-
rity vulnerabilities.
Although helpful, preventive maintenance may be an
oversimplification of what is required for maintaining com-
plex systems and preventing failures. Here we define a com-
plex system as one with many interacting components where
Oil Pipeline PV Partial it is difficult to comprehensively model all of the behaviors
Corrosion Shading Faults due to the dependencies, relationships, and all other interac-
tions between these components. In complex systems, faults
are less apparent and may not be visually identified or fixed
by a routine procedure. Hence, complex operations for other
systems (see Figure 3) often rely on predictive techniques,
which are yet to be fully explored for BESSs.
Predictive analysis involves understanding how all of
Power Plant Wind Turbine the components in a system fail and then actively moni-
Upkeep Gearbox Repair toring the components for failure criteria. A 2019 report
by GlobalData, “Predictive Maintenance in Power,” noted
figure 3. Applications of predictive maintenance for other several successful implementations of this approach in the
systems. utilities sector.
✔✔ The monitoring and diagnostics center at American Elec- and may not have the throughput to process all the data. Pre-
tric Power identified warning signs of failure and initiated dictive analysis must also depend on data from the EMS to
repair work of a gas turbine blade before breakdown, re- understand the system’s behavior. Current EMSs are often
sulting in savings of approximately US$19 million. intended for dispatching/controlling multiple grid resources
✔✔ Duke Energy used predictive analytics for the early and do not include the necessary monitoring and safety func-
detection of a crack in a turbine rotor. This resulted in tions to manage single or multiple BESSs. EMSs often lack
savings of over US$7.5 million. direct communication with BMSs, and any fault detection by
✔✔ Southern Company applied predictive analytics to pow- the BMS may not get communicated to the EMS, limiting
er station models to decrease unexpected maintenance preventive actions from system operators. Current standards
and maintain data quality reliability, resulting in sav- have not addressed this issue.
ings of approximately US$4.5 million. A properly designed monitoring approach for operational
✔✔ Many wind turbine operators now use predictive analyt- ESSs will create indicators that can provide characteristics
ics to monitor the health of gearboxes. The cost of gear- such as those listed in Table 3. The overall goal is clear:
box failure can be upward of US$350,000 per incident. identify the indicators of potential faults and preemptively
Despite these and other successes, both business and intervene on an operational ESS without making the inter-
technical challenges hinder a broader adoption of predictive vention itself a problem. However, the links and causal rela-
maintenance in BESSs. From a business perspective, the tionships between faults and indicators and the potential
energy storage industry is relatively young. Thus, business of those early faults to lead to larger faults are not read-
priorities and budgets do not always motivate investment in ily apparent at this early stage of the BESS industry. Ulti-
“soft” engineering methods such as data analytics and arti- mately, stakeholders must establish a methodology for iden-
ficial intelligence-based services. The nascent state of the tifying the indicator-fault relationships that can be tracked
industry also means reduced data on fault patterns, especially and monitored in these systems.
due to limited public knowledge exchange, making data ana-
lytics more difficult. Furthermore, the pressure to keep the Creating a Predictive Maintenance
cost per kilowatt hour as low as possible means less invest- Approach for BESSs
ment in the sensors and infrastructure needed to process The sophistication of approaches for identifying useful
large volumes of data. “flags” or fault indicators has evolved substantially. In the
In terms of technical challenges, predictive maintenance most basic, reactive approach, these indicators are based on
techniques tend to be used for mechanical systems where near misses reported by employees. All data are significant
factors such as “wear and tear” can be readily measured
and monitored. For electronic systems, predictive practices
may be more difficult to apply. Rather than wear out, elec- table 3. The key characteristics of indicators
for predictive monitoring.
tronic components are more likely to have a binary failure
profile. An issue with one component may manifest itself Element Description
downstream and result in the failure of another component,
Time • Days of warning rather than minutes or hours
requiring data collection from multiple points to identify a element • Not all faults will have long lead times, but
pending failure. anything that can extend the timeframe can
Implementing predictive monitoring in conventional BESS minimize destructive failures
hardware is also difficult due to limitations in communica- Actionable • Point to the components that need to be
tion channel availability and the processing power of battery/ warnings replaced
energy management devices. In a BESS, predictive monitor- • Allow time for the examination of areas
causing the warning to occur
ing would involve processing data from battery racks and
the overall system to identify failure indicators. Ideally, Recoverable • Safety measures intended to prevent
actions catastrophic failure and threats to human
the BMS of an energy storage device should have the abil- safety can ultimately destroy the unit
ity to assist in this area; however, not all current BMSs are (unrecoverable)
designed to recognize faults occurring outside of the imme- • Recoverable action must have minimal
impact on the system
diate impacts on the battery itself (e.g., cells and modules)
and can be useful in preventing future failures. Hence, we processes created from a system-safety perspective. These
recommend a culture where the reporting of near misses processes are rooted in historical data, where the mechanism
is encouraged. More rigorous approaches involve 1) lever- of a past failure is identified to improve designs and prevent
aging indicator-fault links established during the system- a similar failure from occurring in the future. To contribute
design phase and 2) combining with additional indicator- substantially to predictive maintenance, however, the sys-
fault links from the analysis of operational data on fielded tem-design process also needs to establish the monitoring
systems. The identification of these links is an iterative pro- criteria that can be used in maintaining device operation.
cess. During the design phase, system integrators develop PRA, built from a foundation of risk management, is
the product based on institutional or historical knowledge; the most widely used safety engineering method. A PRA
however, the use cases the system encounters in the field approach identifies hazards, their deterministic causes and
could lead to new fault indicators. Thus, the predictive consequences, and provides a method of describing uncer-
maintenance approach should be scalable to adapt to new tainty. The process enables the calculation of expected risk
“patterns” with minimal impact on the overall system cost (defined as the probability of an event multiplied by the rela-
and availability. The following section elaborates on this tive severity of its consequences) so that a developer can
two-layer approach for identifying indicator-fault relation- compare different design options. PRA uses fault and event
ships during the design phase and based on the data analyt- tree analysis to break a complicated system into subsystems
ics for fielded systems. and components when there is insufficient data to directly
predict behavior. Risk is then increased or decreased based
Identifying Indicator-Fault Relationships on how failures in components and subsystems operate
During the Design Phase together to generate accidents.
It is expensive to retrofit a fielded system. Thus, the first step Additionally, failure modes and effects analysis (FMEA)
during the design phase is to make a deliberate decision to is a systematic procedure for assessing reliability and how
sense critical information and obtain as much data as pos- component failures can impact system safety. Developers
sible to provide insight into various failure modes. Next, the begin an FMEA by compiling a list of each component or
process requires type of component in a system. Then they calculate the prob-
1) the creation of a comprehensive listing of recoverable ability of each component failing in a variety of ways based
battery system faults and the linking of faults to lead- on historical data. Table 4 shows a brief list of typical FMEA
ing indicators. This begins with thoughtful engineer- calculations for a BESS (adapted from an Electric Power
ing consideration of the system design; however, the Research Institute report on ESS safety). The probability and
designer should complement this consideration with a severity each receive a score of 1–10, with 10 correspond-
collection of historical data from key developers, op- ing to a more probable or severe event. Each failure mode is
erators, and manufacturers. linked to a hazard effect, consequence, method of prevention,
2) the determination of whether indicators are already and method of detection. The identification of the detection
being tracked through current BMSs, EMSs, or any method lays the foundation for predictive maintenance. It is
plant controllers. apparent, however, that this conventional FMEA approaches
3) the finalization of a list of indicators and criteria to system design from a safety perspective (preventing cata-
be monitored to reduce the field failures of BESS strophic failure) rather than detecting faults while they are
equipment. still recoverable. The process creates a probability but does
This is a beneficial process to leverage, although there not provide the leading indicators necessary to flag pending
may be gaps when new failure modes are identified or the failures of the areas. Still, these processes are beneficial in
process does not account for the design errors or field/envi- understanding what areas to focus on when creating indica-
ronmental degradation that could lead to failures. tor-fault relationships.
The industry has many well-established processes for sys- A more recently developed design tool, STPA, views
tem design, including various probabilistic risk assessment a system as a collection of interacting control loops. Acci-
(PRA) approaches (e.g., failure modes and effects analysis, dents happen when the component interactions in these
fault tree analysis, and so on) and systems-theoretic process loops violate safety constraints. Unlike PRA-based tools, it
analysis (STPA). It is important to note the nuances from does not rely on any component failure rate data. Thus, this
Predictive Analysis
Run Predictive-Analytics Algorithm
On Site Data to Get Misbehavior/Weak
Battery Indicators
Predictive-Analytics Algorithm
3.6
they could trigger safety hazards, negatively
The service team 3.5
impact system availability, or reduce the system
was alerted to this 3.4 VcellMax_Rack1 capacity. Furthermore, this approach helped
pattern, and they 3.3 VminCell_Rack1
VcellMax_Rack2
the service team plan maintenance visits more
reviewed the 3.2 VminCell_Rack2
in memoriam
Patrick Ryan
T
THE IEEE POWER & ENERGY ✔✔ the development of the IEEE
Society (PES) is saddened to announce Smart Village program, which
the passing of Patrick Ryan, IEEE PES integrates sustainable electricity,
executive director, on Tuesday, 8 Sep- education, and entrepreneurial
tember 2020. solutions to empower off-grid
Pat Ryan was a courageous indi- communities
vidual who fought very hard to beat his ✔✔ the establishment of four new pub-
battle with cancer. He faced this chal- lications to meet the growing de-
lenge with his usual tenacity and posi- mand of PES membership: IEEE
tive attitude, but he was sadly unable to Transactions on Smart Grid,
overcome it. IEEE Transactions on Sustain-
Pat became the third PES executive able Energy, IEEE Electrification
director in 2007 and was a driving force Magazine, and IEEE Open Access
in transforming PES into the second- Journal of Power and Energy
largest Society within the IEEE. He had ✔ ✔ the expansion of the portfolio
a relentless focus on the IEEE and PES of PES conferences to a global
mission and vision and a strong belief audience and inaugurated Inno-
that there needed to be significant vol- Patrick Ryan vative Smart Grid Technologies
unteer engagement to pursue anything Conferences worldwide
new. Pat was recently selected as the ing Society to the IEEE Power & ✔ ✔ the development of the IEEE
2020 recipient of the IEEE Eric Herz Energy Society to better reflect Smart Grid and IEEE Smart Cit-
Outstanding Staff Member Award “for the PES mission and vision ies programs.
leadership in creating a successful mod- ✔✔ the increase in PES membership Pat had a unique sense of humor.
el for IEEE and Society memberships from 22,000 in 2006 to more With his panache for making friends
by fostering member value and partner- than 40,000 by the end of 2019 and cultivating relationships, a main fo-
ships with volunteers.” ✔✔ the development of business pro- cus of his life, along with his cherished
No matter how good an idea was, cesses that helped increase PES wife, was to raise his three children with
Pat would not pursue it unless there was reserves, enabling the Society to strong values. He was so proud to watch
significant volunteer engagement. He create and establish humanitar- them each become positive, contribut-
clearly held the volunteers at the core of ian programs ing citizens of the community at large.
the organization. Under his leadership ✔ ✔ the foundation of the I EEE He also enjoyed hunting, fishing, and
as executive director, Pat drove several PES Scholarship Plus Initiative, anything involving the great outdoors.
significant achievements for PES, in- which has distributed more than Pat is survived by his wife, Patricia
cluding the following: 1,500 scholarships to undergrad- J. Hunkele Ryan, and his children Ro-
✔✔ the name change of the Society uate students man, Nicholas, and Bridget. For fur-
from the IEEE Power Engineer- ✔✔ the founding of IEEE PES Re- ther information, please see https://
source Center, providing prac- www.ieee-pes.org/about-pes/pes-staff/
Digital Object Identifier 10.1109/MPE.2020.3026223
tical content and educational in-memoriam-pat-ryan.
p&e
Date of current version: 16 October 2020 courses to industry professionals
electrochemical loads
served by Niagara Falls hydroelectric power
I
IN 1895, THE NIAGARA FALLS
Power Company installed an early large- In previous “History” columns, we have explored a number of aspects of elec-
scale polyphase generating station at trical power associated in the Niagara Falls area. In this issue, we present a
Niagara Falls, New York. Later named view from the perspective of the loads consumed by customers of the early
Adams Plant No. 1, power was first pro- hydroelectric generation plants in the Niagara Falls area. This article focuses
duced in 1895. Ten years later, a total of on early electrochemical factories and the distribution methods used to serve
six ac-generating stations were in opera-
these loads.
tion on both sides of the U.S.–Canadian
We welcome back Robert D. Barnett for a fifth time to the “History” pages
border. The rapid development of power
of our IEEE Power & Energy Magazine. A Life Senior Member of IEEE, Barnett
generation at Niagara Falls was in no
small part a result of the burgeoning graduated from the University of Waterloo and Niagara College. In 1982, he
electrochemical industry. formed the Niagara Society for Industrial History as a support group for a pro-
The Adams Plant generated two- posed museum in a former Niagara Falls power plant, and he has written on the
phase power. The industries that would history of the topic.
use this power required a mix of ac and John Paserba,
dc. For the ac plants, only a transform- Associate Editor, “History”
er was required to provide the process
voltage. For dc requirements, however,
rotary converters were needed. These U.S. Census Bureau said in 1910, “Niag- Falls Hydraulic Power & Manufactur-
machines, normally located in the ara Falls, New York, is the chief seat of ing Company had been in operation
manufacturing plant, were the usual the electrochemical industry, and it has since the 1880s, but because it was orig-
means for supplying high-current dc in held this position from the beginning.” inally a plant with small-size genera-
the days before the development of mer- The first electrochemical plant to tors, it was not suitable for supplying the
cury-arc rectifiers or, later, solid-state take power was the Pittsburgh Reduc- large amount of power required by the
power devices. tion Company. The plants of the Carbo- electrochemical processes. This article
rundum Company and Union Carbide focuses on the industries supplied by the
The Industries Company (now known as Union Car- Adams Plant, looks at three of the larger
Although Niagara Falls, New York, was bide Corporation) followed it closely. loads in some detail, and comments on
not the birthplace of the American elec- Several more plants were added in the their impact to the development of load-
trochemical industry, it certainly was its next decade (see Table 1 for a listing of serving power equipment.
nursery. Martha Moore Trescott, in her most of the large ones). Figure 1 shows
book The Rise of the American Electro- a 1904 map of the area near the Adams Local Distribution
chemicals Industry, said, “Most of the Plants in the part of the river just above Initially, Adams Power House No. 1 had
early advances in electrothermic elec- the falls. By the late 1890s, another gen- only a single two-phase alternator, but
trochemicals occurred in Niagara Falls erating station owned by Niagara Falls this alternator was 5,000 hp at a time
after 1895.” Trescott also noted that the Hydraulic Power & Manufacturing when a 500-hp machine was considered
Company and located below the falls large. In addition, the best practices gov-
Digital Object Identifier 10.1109/MPE.2020.3016101
began to supply power to electrochemi- erning the characteristics of ac circuits
Date of current version: 16 October 2020 cal industries in their area. Niagara were not well understood at the time.
figure 1. A 1904 map of the area near the Adams Plants in the portion of the Niagara River just above the falls, which
shows the local tenants on land owned by the Niagara Falls Power Company. (Source: The Niagara Falls Electrical Hand-
book, 1904.)
Carborundum
(a) The Carborundum Company began
operations in the fall of 1895. Edward
G. Acheson, one of the inventors (1891)
of the carborundum process, was also
the president of the company. The ma-
terial known as carborundum is the
compound silicon carbide (SiC). SiC
is a synthetic mineral, i.e., not found
in nature. It is formed using a mixture
of carbonaceous material (usually pe-
troleum coke) and a silica or quartz
sand, which react at high temperatures
to form SiC and CO. The energy for
the reaction was provided by pass-
ing an ac through the core. This core
was composed of very pure carbon
to offer lower resistance to the pas-
(b) sage of the electric current than that
of the surrounding material. It also
figure 7. The Carborundum Company. (a) A furnace ready for burning and (b) a had to be capable of being heated to
furnace during burning. (Source: The Niagara Falls Electrical Handbook, 1904.) an intense degree without destruction.
an adaptive text
suitable for academics and practitioners
T
THIS ISSUE’S “BOOK REVIEW” The text is organized into three
column discusses Power System Mod- parts after a brief Introductory chapter:
eling, Computation, and Control, by ✔✔ Part I: System Concepts
Joe H. Chow and Juan J. Sanchez-Gas- • Steady-State Power Flow
ca. The reviewer says that it is “a wel- • Steady-State Voltage Stability
come addition to the body of literature Analysis
on power system modeling, computa- • Power System Dynamics and
tion, and control.” Simulations
• Direct Transient Stability Analysis
Power System Modeling, • Linear Analysis and Small-
Computation, and Control Signal Stability.
By Joe H. Chow and ✔✔ Part II: Synchronous Machine
Juan J. Sanchez-Gasca Models and Their Control
Power System Modeling, Computa- • Steady-State Models and Oper-
tion, and Control (Wiley-IEEE Press, ation of Synchronous Machines
2020) is designed to make the ad- • Dynamic Models of Synchro-
vanced topics of modeling, computa- nous Machines
tion, and control more accessible to • Excitations Systems
graduate students, power engineers, • Power System Stabilizers
and control design professionals. The • Load and Induction Motor
initial use of this text was for teaching Power System Dynamic Performance Models
a two-course sequence on the topics at Committee and its various subcommit- • Turbine-Governor Models and
Rensselaer Polytechnic Institute as an tees, working groups, and task forces. Frequency Control.
evolution to the use of the well-known The organization and material in ✔✔ Part III: Advanced Power Sys-
text Power System Stability and Con- this text reflect the preferences of the tem Topics
trol, by Dr. Prabha Kundur. The new authors for 1) the topics covered, 2) • High-Voltage Direct Current
text offering chapters that address flex- the depth of the material included for Transmission Systems
ible ac transmission systems and wind- each topic, and 3) how the material is • Flexible AC Transmission Systems
turbine-generator technology. explained. In the Preface, the authors • Wind Power Generations and
The authors of Power System Mod- describe their contribution as “In ad- Modeling
eling, Computation, and Control have dition to being a modeling and control • Power System Coherency and
more than 40 years of experience work- book, it is also a systems and compu- Model Reduction.
ing in power systems, were colleagues tation book.” A distinctive feature of The book offers a comprehensive list
at General Electric Company, and have this text is that most of the topics are of 243 references and a thorough index
been long-time collaborators within illustrated with detailed numerical ex- as well as color versions of key figures.
the IEEE Power & Energy Society’s amples. The depth of their respective Each chapter ends with a set of prob-
industry experience is clearly shown lems, totaling 155 for the entire book.
Digital Object Identifier 10.1109/MPE.2020.3014544
throughout the nearly 600-page treat-
Date of current version: 16 October 2020 ment of the topics. (continued on p. 111)
PES meetings
for more information, www.ieee-pes.org
T
TH E I EEE POW ER & EN ERGY Ron Melton, ron.melton@pnnl.gov, June 2021
Society’s (PES’s) website (http://www https://ieee-tesc.org/ IEEE PowerTech Milan (PowerTech
.ieee-pes.org) features a meetings sec- 2021), 28 June–2 July, Madrid, Spain,
tion, which includes calls for papers IEEE International Conference on contact Tomas Gomez San-Roman,
and additional information about each Power Electronics, Drives and Energy tomas.gomez@comillas.edu
of the PES-sponsored meetings. Please System (PEDES 2020), 16–19 Decem-
check the website for the most current ber, Jaipur, India, virtual event, contact July 2021
information. Suryanarayana Doolla, doolla@gmail IEEE PES General Meeting (GM
.com, http://pedes2020.com 2021), 18–22 July, Washington, D.C.,
November 2020 United States, contact Roseanne Jones,
IEEE Electronic Power Grid (eGRID January 2021 roseanne.jones@ieee.org
2020), 2–4 November, Aachen, Ger- IEEE PES 2021 Joint Technical
many, virtual event, contact info@ Committee Meeting (JTCM 2021), September 2021
egrid2020.org, https://egrid2020.org/ 7–15 January, New Orleans, Louisiana, IEEE PES GT&D International
United States, contact Solveig Ward, Conference and Exposition, Istan-
IEEE International Future Energy sward@quanta-technology.com, www bul (GTD), 14–17 September, Istanbul,
Conference (IFEC 2020), 9–12 No- .pestechnical.org Turkey, contact Omer Usta, usta@ieee
vember, Aalborg, Denmark, contact .org, https://ieee-gtd.org/
Huai Wang, hwa@et.aau.dk, http:// February 2021
energycha llenge.weebly.com /ifec IEEE PES Innovative Smart Grid December 2021
-2020.html Technologies (ISGT 2020), 15–18 Feb- IEEE PES Innovative Smart Grid
ruary, Washington D.C., United States, Technologies Asia (ISGT Asia 2021),
IEEE Sustainable Power and Energy virtual event, contact Kathy Heilman, 5–8 December, Brisbane, Austral i a ,
Conference (iSPEC 2020), 23–25 No- kathy.heilman@ieee.org, https://ieee-isgt contact Tapan Saha, saha@itee.uq
vember, Chengdu, China, contact Min .org/ .edu.au
Liu, min-liu@csee.org.cn, http://2020
.ieee-spec.com/ May 2021 For more information on additional tech-
IEEE International Conference nical committee meetings, webinars,
December 2020 on Electrical Machines and Drives and events, please visit our IEEE PES
IEEE PES Transactive Energy Sys- (IEMDC 2021), 16–19 May, Hartford, calendar: https://www.ieee-pes.org/
tems Conference 2020 (TESC 2020), Connecticut, United States, contact Ali meetings-and-conferences/conference
8–10 December, virtual event, contact Bazzi, alibassi@ieee.org, http://iemdc -calendar.
-conference.org/
Digital Object Identifier 10.1109/MPE.2020.3014543
p&e
Date of current version: 16 October 2020
distr ibution management system to provide compensation for grid ser- dards was that the DG unit do no harm
(ADMS) Voices of Experience effort, vices. At the wholesale market level, under normal operating conditions and
DOE researchers found that a surpris- the requirements are clear, but DER disconnect as quickly as possible under
ingly low number—only a handful—of aggregators believe that they are too abnormal conditions. Given the cost of
utilities in the United States either had or burdensome, and there are distribu- a PV system at the time, the growth of
were considering deploying an ADMS. tion system operating constraints that installations was expected to be small
An ADMS provides a critical view of the impact the aggregation availability. At for the foreseeable future.
system and manages conventional utility the distribution level, without a gen- Just a decade later, at the end of
resources, such as capacitors, regulators, erator output meter, it is unclear what 2010, SDG&E had 11,644 PV instal-
transformers, switches, and so on, which the actual response to a request for lations totaling 84,765 kW. We were
is key to advancing a more resilient and service would be. (In 1999, California beginning to get a good sense of how
reliable energy future. All resources need investor-owned utilities asked policy DERs impact the grid when there is
to be managed concurrently to operate makers to require a generator output widespread deployment.
the grid both safely and reliably. meter, but the request was rejected due In 2011, on a rural circuit in San
Without an ADMS to provide visi- to concerns that the additional require- Diego County, an excellent example
bility into grid conditions, how do you ment would drive up the cost of solar of the impact of DERs on the grid was
leverage DERs? To forecast and deter- installations.) While it is possible to discovered. Power quality meters had
mine a service need and effectuate that estimate a PV system’s power output been set on the secondary of a primary
request to either the distribution system based on solar irradiance data and the metering station at the end of a long cir-
or wholesale market, a DERMS is a panel orientation, in California, solar cuit, which utilized a small copper con-
necessary software and control tool to panels are oriented in all directions, ductor for the last line segment. At the
integrate into distribution operations. including north, based on facility/ end of the circuit, a 1-MW PV system
ADMS deployment not only comes property siting constraints. was connected. The primary circuit
before the DERMS, but its functional- voltage was observed to fluctuate wild-
ity (power flow model, volt/var man- A Little Bit of History ly due to intermittent power production
agement, and so on) must be vetted In the late 1990s, DERs, or distributed from this PV system with a legacy in-
prior to DERMS implementation. The generation (DG) as it was known at verter as the morning fog burned off.
DERMS must be integrated into the the time, was deployed by larger en- SDG&E had identified the factors that
ADMS, allowing a single unified view tities with a significant heat load in determined the impact of a PV system
of the distribution system; the DERMS combined heat and power applications. on the grid and recognized, given the
guides the dispatch of all resources, However, during that time period, PV German experience (which California
subject to the total grid operating state systems and microturbines developed was emulating), that it was necessary
based on grid constraints, switch state, with an inverter (dc-to-ac converter) in- to update the Rule 21 interconnection
load, and DER capabilities. terface to the grid were also beginning requirements to incorporate smart in-
Based on two rounds of requests to be deployed. At the end of 2000, verter functionality to avoid the Ger-
for proposals at SDG&E several years SDG&E had 32 PV installations total- man 50.2-Hz retrofit experience.
apart, I learned that the distributed, ing 105 kW. After a 2013 CPUC-sponsored work-
federated, hierarchical control struc- In 1999, the California Public Utili- shop, an effort to update Rule 21 was au-
ture necessary to control large, scalable ties Commission (CPUC) opened a thorized, and the Smart Inverter Work-
quantities of DERs—tens of thousands rulemaking on DG interconnection ing Group began work on the technical
to millions—is not currently available with the goal of reducing barriers, low- requirements. This effort resulted in re-
commercially. Only pilot projects or ering costs, and accelerating adoption. quirements and timelines for implemen-
customized, one-off software solutions This inclusive stakeholder effort, codi- tation and the development of the UL
are being used to control small quan- fied under Rule 21, resulted in standard 1741-SA standard to allow for testing to
tities of DERs. The DOE is funding requirements for inverters. Utilities the new California Rule 21. An update to
DERMS R&D through its Grid Mod- were required to evaluate DG as an IEEE 1547 was initiated and approved,
ernization Lab Consortium as well as alternative to traditional “wires” solu- leveraging the California work. IEEE
its Office of Electricity. tions, with vendors offering to supply, 1547.1, the test standard for smart invert-
rent, or lease a generator to the utility. ers, has also been approved and will be
Lack of Measurement and This Rule 21 revision led the parties published soon. I expect this to set off
Verification Techniques involved and a broader group of stake- another round of work to harmonize all
Finally, certain measurement and veri- holders to develop IEEE 1547 and up- of the standards. The key point of the
fication techniques must be in place date UL 1741. A key tenet of the stan- 1547 standard was that the DER unit
distributed generation
leveraging for grid services
I
IN THE PAST TWO DECADES, SIG- Grid Challenges Lack of a Unique Identifier
nificant progress has been made to Associated With DERs For a DER to provide a service, a re-
deploy distributed energy resources quest must be made that uniquely
(DERs) in California. The Golden State Interoperability Issues identifies the location of the device ex-
(as California is nicknamed) now has San Diego Gas and Electric (SDG&E) pected to perform that service. IEEE
a much more open planning process engineers observed device interoper- Standard 2030.5 for common smart in-
where DERs can be utilized for grid ability issues while attempting to utilize verter profile implementation provides
services. With the tools and processes photovoltaic (PV) systems with smart a structure and method for locating de-
already instituted and given the quantity inverters to provide voltage regulation vices to service transformers; however,
and size of the aggregated deployment, service. In our testing lab, two different the provisioning of this identification
many people believe DERs should be inverters of the same model and manu- method has not yet been included as
providing these services and enabling facturer performed differently when part of new system interconnections.
utilities to defer large quantities of tra- tested. In another instance, different en-
ditional infrastructure investment. ergy storage devices responded to charge Legacy Inverters
The reality is that this is not hap- and discharge requests in very different Until 8 September 2017, all DER installa-
pening yet because leveraging DERs ways. When running predefined oper- tions had a legacy inverter designed with-
for grid services is not as simple as ating schedules, we also observed that out critical communications and testing
flipping a switch. There are many tech- some devices returned to a home state requirements in mind and were installed
nical challenges that we as an industry first before adjusting to the newly sched- under some version of a net energy me-
must first overcome. uled set points. Finally, customized in- tering (NEM) tariff. In California, the
During my quest for answers as to terfaces are required to communicate to phase 2 communication requirement that
why DERs have not yet reached their each device and control it. allows for manufacturer attestation, not
full potential, I identified some but an end-to-end test, was in place 22 June
not all of the technical gaps that must Cybersecurity 2020. This creates two issues for utiliz-
be filled: device interoperability, de- Based on a review of installed DER ing existing DERs for grid services. First,
vice and system cybersecurity, unique systems in SDG&E’s service territory with a NEM tariff that compensates en-
identifiers for DERs, communication where I live and work, 58% of all sys- ergy exported to the grid at the full retail
networks, participation decisions, the tems utilize an inverter from only two rate, there is little incentive for customers
development of DER management manufacturers. Both leverage the cloud to do anything else. Second, even if cus-
systems (DERMSs), and measurement and a customer’s Internet connection to tomers would, how would/could they ac-
and verification techniques. Certainly, monitor and provide firmware updates, cept a voltage regulation service request,
there are many other potential nontech- and the commands within the inverter for example, if they do not have a tested
nical challenges, such as developing utilize Modbus protocols. As the ag- smart inverter?
provider business models and restruc- gregate numbers increase, so do the
turing legacy compensation and incen- number of attack vectors. In Califor- Slow Adoption of Critical
tives to focus deployment on locational nia, there are approximately 8,000 MW Management Platforms
grid needs. of rooftop PVs installed. Device and While participating in the U.S. Depart-
system cybersecurity are critical as the ment of Energy’s (DOE’s) advanced
Digital Object Identifier 10.1109/MPE.2020.3014750
grid is operated in a high-penetration
Date of current version: 16 October 2020 DER environment. (continued on p. 110)
Varasset 14 varasset.com