PCJ Spring2009

PCJ Spring2009
A Paradigm Shift in Protection, Control and Substation Automation Strategy pg 41
Protection & Control Journal
8th Edition
Smart Grid
Page 15
The Road Ahead
7
Application of Phasor Measurement Units for
Disturbance Recording
15
Smart Grid: The Road Ahead
21
Developing the Smart Grid Business Case
25
An Enterprise Information Architecture Based on SOA
Enables a Smarter Grid
29
Key Smart Grid Applications
37
The Evolution of Distribution
41
A Paradigm Shift in Protection, Control and Substation
Automation Strategy
57
Implementing Smart Grid Communications
61
IEC 61850 Communication Networks and Systems in
Substations
69
Enhanced Security and Dependability in Process Bus
Protection Systems
85
AEP - Process Bus Replaces Copper
89
The Digital World and Electrical Power Supply
$19.95 USD
Stay Current
Get your FREE subscription today
To receive future publications of the Protection & Control Journal, sign up at:
www.GEMultilin.com/journal
Advertising Opportunities
Learn more about how to reach today’s protection and control professionals.
www.GEMultilin.com/journal
Submit Articles
To submit articles to be considered for inclusion in a future issue of this journal email:
[email protected]
The Protection and Control Journal is a unique publication that presents power system protection and control
engineers with a compilation of relevant and valuable documents. GE Digital Energy grants permission to educators
and academic libraries to photocopy articles from this journal for classroom purposes. There is no charge to educators
and academic libraries provided that they give credit to the Protection & Control Journal and GE Digital Energy. Others
must request reprint permission from GE Digital Energy.
The Protection & Control Journal is published by GE Digital Energy, 215 Anderson Avenue, Markham, Ontario, Canada,
L6E 1B3. All rights reserved. GE and g are trademarks of the General Electric Company.
Copyright © 2009 GE Digital Energy. All rights reserved.
www.GEDigitalEnergy.com
ISSN 1718-3189
Protection
& Control Journal
8th Edition
Contents
Smart Grid
The Road Ahead
Business Case for
the Smart Grid
15
7
Key Smart Grid
Applications
21
29
Application of Phasor Measurement Units for Disturbance Recording
M. Adamiak, R. Hunt
15
Smart Grid: The Road Ahead L. Sollecito
21
Leader or Follower: Developing the Smart Grid Business Case
J. McDonald
25
An Enterprise Information Architecture Based on SOA Enables a Smarter Grid
M. Van Helton
29
Key Smart Grid Applications B. Flynn
37
The Evolution of Distribution
J. Fan, S. Borlase
41
A Paradigm Shift in Protection, Control and Substation Automation Strategy
V. Madani, G. Duru, M. Adamiak
57
Implementing Smart Grid Communications
J. Cupp, M. Beehler
61
IEC 61850 Communication Networks and Systems in Substations: An Overview for Users
M. Adamiak, D. Baigent, R. Mackiewicz
69
Enhanced Security and Dependability in Process Bus Protection Systems
D. McGinn, V. Muthukrishnan, W. Wang
85
AEP - Process Bus Replaces Copper
J. Burger, D. Krummen, J. Abele
89
The Digital World and Electrical Power Supply
R. Kleger
Industry Innovations
Upcoming Events
105
108
Editorial
Richard Hunt
Market Development Leader
What is the Smart Grid Really?
Anyone working in the electric utility industry, and in fact most of
the general public, have heard something about “Smart Grid”.
Utilities are undertaking Smart Grid initiatives, and every vendor
in the electric utility area now seems to have a Smart Grid
“solution”. Even service suppliers that are traditionally outside the
utility industry are now, (if you believe the television commercials),
Smart Grid providers.
But what is Smart Grid really? Is it just an
Advanced Metering Infrastructure (AMI)
with “smart” communicating meters? Is
it integrating renewable or distributed
energy sources with the existing utility
grid? Is it real-time pricing for customers
to make informed decisions? Or is Smart
Grid really something else?
To define Smart Grid, lets look at some
challenges facing the present electric
utility grid. The grid is designed to be
operationally reliable while serving
peak demand. And with a reliability
of 99.97%, the electric grid is very
successful at meeting this goal. The
demand for electricity is such, however,
that an outage can easily result in losses
to customers of $1 million dollars a
minute. In the United States alone, the
demand for electricity is growing with an
anticipated growth of 2% per year. This
works out to a 50% increase in demand
over the next 20 years. And there is a
growing focus on the environmental
costs of this huge demand for electricity.
So, in short, the present utility system
must serve a growing demand, while
increasing reliability, with little or no
increase in environmental impact.
A cursory look shows that the present
model of the electric grid really can’t
meet these challenges. The traditional
centralized model for generation and
distribution of power requires huge
capital investments in generating plants,
transmission lines, and distribution
networks, to improve reliability and to
meet the requirements for new demand.
Let’s not forget, our current system
is also aging, with 60% to 70% of the
transformers, transmission lines, and
circuit breakers nearing the end of their
usable life. So clearly, something must
change for utilities to afford the capital
costs of new construction. So what is
Smart Grid, and how will it meet these
challenges? Smart Grid is a new model
for the utility grid, with a vision towards
operating the utility system as efficiently
as possible with connectivity to real-time
data through advanced communications,
while integrating more efficient methods
of producing electricity. A Smart Grid
has several characteristics, including the
ability to self-heal from disturbances,
being secure against attack, providing
power quality for future needs,
accommodating all power generation
and energy storage options, and
having an intelligent communications
infrastructure. Smart Grid solutions
can be implemented at any level of the
utility system: generation, transmission,
distribution, and power consumers.
From this definition of Smart Grid, one
can list many Smart Grid solutions. Much
talked about solutions include thin-film
solar panels to provide local generating
capacity, smart appliances to reduce
residential demand during peak load
periods, microgrids to efficiently integrate
distributed renewable resources, and
wide-area protection schemes that make
the transmission system self-healing
and secure. Which solution, or set of
solutions, to adopt is going to be based
on the operating needs, philosophy,
history, and regulatory considerations
of a specific utility. The best step a utility
Editorial
can take right now, is to define and
build a framework that supports future
use of real-time data and information
technology.
The framework for Smart Grid must
meet four major requirements. It must
be cost effective, both in terms of capital
costs, and in operating and maintenance
costs. It must provide useful information,
such as pricing of electricity, in realtime. It must provide this useful
information where it is needed, whether
at a substation, a control center, or at a
customer’s residence. Finally, whatever
steps are taken to develop a framework
for Smart Grid, they must respect the
traditional operating requirements for
the utility system: reliability, safety, and
the use of industry standards to support
open solutions.
While there is a lot of talk about what
Smart Grid is, and what a Smart Grid
framework could be, it’s good to know
about solutions that have been done.
Look at the real and practical solutions
described in this issue of the Protection
and Control Journal. Developing a
business case for Smart Grid solutions.
Designing
the
communications
infrastructure to transmit information
between utilities and customers. Making
sure your distribution management
system can actually operate a
distribution system that includes
distributed energy resources, intelligent
outage management information, and
self-healing capabilities. Using the new
data from synchrophasors to improve
the operation of the transmission system.
And the use of innovative technology
to replace copper wiring in protection
and control systems, making traditional
substations more efficient and reliable to
build and operate.
The more you know about the multifaceted Smart Grid, the better equipped
you will be as you journey towards
realizing your own smarter solutions.
5
Application of Phasor Measurement Units
for Disturbance Recording
Mark Adamiak, Rich Hunt
GE Digital Energy
1. Introduction
This paper looks at the specific application of Phasor Measurement
Units (PMUs) for disturbance recording, with a special emphasis
on wide area cross-triggering of recording PMUs during events.
Disturbance recording, or long-term recording of phasor data,
provides valuable information when analyzing wide area
disturbance and power swings in the utility system. The newly
approved NERC PRC-002 and PRC-018 standards require the
installation of disturbance recording equipment at strategic
points on the power system. The value of this equipment is only
realized when discrete records are captured simultaneously at
all points on the power system, to provide a complete snapshot
of a specific event. Traditional recorders rely on local triggers to
capture the data, however, an individual recorder may not trigger
for a specific event, or may trigger in a different time frame than
other recorders on the system, and not capture valuable data. A
practical challenge is adding the disturbance recording function
to existing substations and relay systems.
Ongoing projects, such as the Eastern Interconnect Phasor
Project, promote the installation of PMUs to provide real time
measurement of the state of the power system, by streaming
highly accurate synchrophasors at a high sampling rate. The
PMUs are generally installed at the same strategic substations
that require disturbance recording. In addition, today’s digital
relays (such as a line distance relay or current differential relay)
are capable of synchronous phasor measurements. In addition
to streaming data to a centralized database, PMUs may have the
ability to record data at the PMU based on local trigger conditions.
The record may include synchrophasor data as well as additional
analog values and digital status. This recorded data meets the
disturbance recording requirements set by NERC. The paper
discusses the applicability of synchrophasor data to disturbance
recording, and the capabilities of PMUs to capture the appropriate
data.
This paper also discusses practical aspects of using the IEEE
Synchrophasor standard communications in conjunction with
IEC61850 communications for wide area cross-triggering of
PMUs. Also discussed are communications channels requirements
and expected performance of cross-triggers. Other disturbance
recording applications exist in the industrial domain, such as
motor starting failure events on large motors. Synchronized
measurements provide the ability to correlate the failure with
other events in the industrial process. This paper will discuss
industrial applications of PMUs.
2. Phasor Measurement Units and
recording
In the context of this paper, disturbance recording is defined as
recording of phasor or RMS values of data over a long period of
time. Disturbance recording is intended to show the response
of the power system and equipment due to power system
faults, such as an out-of-step condition, as opposed to power
equipment faults, such as a short circuit. The time interval for
these “long term” events can range from 1 second (in the case of
a fault and high-speed reclose) to many minutes (in the case of
system oscillations). The fast sample rates (30 to 60 phasors per
second) of today’s synchrophasors-based disturbance recording
devices can be used to analyze both power system faults and
the more traditional power equipment faults. The term Dynamic
Swing Recorder (DSR) is also often used to describe a device that
captures disturbance data over a long period of time. A more
complete description of these terms is available in [1].
NERC has issued Standard PRC-002-1 entitled: Define Regional
Disturbance
Monitoring and Reporting Requirements. Section R3 specifically
addresses criteria for dynamic disturbance recording, including
Application of Phasor Measurement Units for Disturbance Recording
7
location of recorders, electrical quantities to record, recording
duration, and sampling rate. The NERC standard essentially states
that DSRs are to be situated at key locations, are to record voltage,
current, frequency, megawatts and megavars for monitored
elements, and are to record the RMS value of electrical quantities
at a rate of at least 6 records per second.[2]
The Regional Reliability Councils (RCCs) of NERC are responsible
for refining these standards for a specific operating region. By
reviewing the standards as interpreted by some of the RCCs, it
is possible to provide a good overview of disturbance monitoring
requirements.
Location of DSRs. DSRs are to be located at key substations for
the power system. Key substations are generally defined as
transmission substations with significant connected generation,
large transmission substations (containing 7 or more transmission
lines), transmission substations that interconnect to another
regional authority or company, at major load centers (such as
load centers greater than 2500 MW), or where undervoltage load
shedding schemes are implemented.
Electrical quantities to record. The NERC requirement is to record
voltage, current, and frequency, with the ability to derive or
record megawatts and megavars for each monitored element.
The minimum requirements defined by the RCCs are:
•
Bus voltages: at least one three-phase measurement per
voltage level, with two measurements per voltage level
recommended
•
Frequency: at least one frequency measurement for every
voltage measurement
•
Three-phase line currents for every critical line
•
Megawatts and megavars, three-phase, for each monitored
line.
Record length. Disturbance recording, and DSRs, are intended to
capture longer term power system faults. DSRs therefore require
longer record times. The recording length is typically specified as
90 to 180 seconds, including 30 seconds of pre-fault data. DSR
records may be require to automatically extend in length when
additional triggers occur during recording.
A second option for record length is to use continuous recording.
A DSR therefore always captures data for all analog channels,
and typically stores the last 30 days of data. The challenge with
continuous recording is to manage the large amounts of data.
Also, it is important to be able to retrieve the key pieces of the
data to analyze an event.
Triggers. Triggers are necessary to initiate recording for the
typical DSRs that have a discrete record length. For continuous
recording, triggers provide markers into the key pieces of data
during an event. The ability to “share” triggers between multiple
sites is also necessary in order to capture a wide-area view of an
event.
There are many types of triggers available in DSRs, including:
8
•
Magnitude triggers, on voltage, current, frequency, real
power, reactive power and apparent impedance
•
Rate-of-change triggers, on voltage, current, frequency, real
power, reactive power and apparent impedance
•
Harmonic content triggers, on a specific harmonic frequency,
or on total harmonic distortion
•
Delta frequency triggers
•
Contact triggers, such as breaker
communications channel operations
•
Symmetrical components trigger
operation
or
Frequency rate-of-change and voltage rate-of-change triggers
are the most commonly applied triggers. Previous papers at
Various conferences have suggested that real power rate-ofchange triggers also have the sensitivity and selectivity to trigger
recording for power system faults, without triggering recording for
power equipment faults.[1] Impedance triggers are an interesting
case for this paper. Impedance triggers will only operate when
the center of impedance of a power system fault is close to the
location of the DSR. However, there are some events, such as load
encroachment, or when the DSR is located close to the center of
impedance, where this trigger can capture valuable data.
Sampling rate. The minimum sampling rate required by NERC
is 6 Hz. However, a higher sampling rate, such as 30Hz or 60Hz,
provides a more accurate picture of the measured electrical
quantities during a power system event, providing frequency
responses up to 15 and 30 Hz respectively.
The requirements for disturbance recording as described in this
section are a synthesis of the requirements as defined by a few of
the Regional Coordinating Councils of NERC. For complete details
of an individual RCC, please see [3], [4], [5].
The term Dynamic Swing Recorder is a generic term to describe
any device capable of capturing RMS or phasor values of electrical
quantities. While typically a DSR is simply a function available in
a digital fault recorder, other devices may have the capability
to capture this type of data. One such device is the Phasor
Measurement Unit (PMU), a device that measures synchrophasors,
a highly accurate time-synchronized phasor measurement. The
typical PMU is designed to communicate these synchrophasors
to system operators for real-time control of the power system.
However, some PMUs have the ability to trigger on system
anomalies, and record synchrophasor data, to meet the
requirements of disturbance recording.
2.1 PMU as disturbance recorders
An AC waveform can be mathematically represented by the
equation:
where
= magnitude of the sinusoidal waveform,
=
where f is the instantaneous frequency
= Angular starting point for the waveform
Note that the synchrophasor is referenced to the cosine function.
In a phasor notation, this waveform is typically represented as:
Since in the synchrophasor definition, correlation with the
equivalent RMS quantity is desired, a scale factor of
Application of Phasor Measurement Units for Disturbance Recording
must be applied to the magnitude which results in the phasor
representation as:
Adding in the UTC-based absolute time mark, a synchrophasor is
defined as the magnitude and angle of a fundamental frequency
waveform as referenced to a cosine signal (Figure 1).
In Figure 1, time strobes are shown as UTC Time Reference 1 and UTC
Time Reference 2. At the instant that UTC Time Reference 1 occurs,
and, assuming a steadythere is an angle that is shown as
state sinusoid (i.e. – constant frequency), there is a magnitude of
the waveform of X1. Similarly, at UTC Time Reference 2, an angle,
with respect to the cosine wave, of
is measured along with
a magnitude or X2. The range of the measured angle is required to
be reported in the range of
. It should be emphasized that the
synchrophasor standard focuses on steady-state signals, that is,
a signal where the frequency of the waveform is constant over the
period of measurement.
In the real world, the power system seldom operates at exactly the
nominal frequency. As such, the calculation of the phase angle,
needs to take into account the frequency of the system at the
time of measurement. For example, if the nominal frequency of
operating at 59.5Hz on a 60Hz system, the period of the waveform
is 16.694ms instead of 16.666ms – a difference of 0.167%.
The captured phasors are to be time tagged based on the time of
the UTC Time Reference. The Time Stamp is an 8-byte message
consisting a 4 byte “Second Of Century – SOC”, a 3-byte Fraction
of Second and a 1-byte Time Quality indicator. The SOC time tag
counts the number of seconds that have occurred since January 1,
1970 as an unsigned 32-bit Integer. With 32 bits, the SOC counter
is good for 136 years or until the year 2106. With 3-bytes for the
Fraction Of Second, one second can be broken down into 16,
777,216 counts or about 59.6 nsec/count. If such resolution is not
required, the C37.118 standard allows for a user-definable base
over which the count will wrap (e.g. – a base of 1,000,000 would
tag a phasor to the nearest microsecond). Finally, the Time Quality
byte contains information about the status and relative accuracy
of the source clock as well as indication of pending leap seconds
and the direction (plus or minus). Note that leap seconds (plus or
minus) are not included in the 4-byte Second Of Century count.
2.2 Synchronized phasor reporting
The IEEE C37.118 revision of the IEEE 1344 Synchrophasor
standard mandates several reporting rates and reporting intervals
of synchrophasor reporting. Specifically, the proposed required
reporting rates are shown in Table 1 below.
System
Frequency
Reporting
Rates
50 Hz
10
60 Hz
25
10
12
15
20
30
Table 1.
Synchrophasor reporting rates
A given reporting rate must evenly divide a one second interval
into the specified number of sub-intervals. This is illustrated in
Figure 2 where the reporting rate is selected as 60 phasors per
second (beyond the maximum required value, which is allowed
by the standard). The first reporting interval is to be at the Top
of Second that is noted as reporting interval “0” in the figure. The
Fraction of Second for this reporting interval must be equal to
zero. The next reporting interval in the figure, labeled T0, must be
reported 1/60 of a second after Top of Second – with the Fraction
of Second reporting 279,620 counts on a base of 16,777,216.
Real Time Monitoring
>1 second time frame
Collect/
Decide
High-speed
Decisions
100 ms – 1S time frame
Collect/
Decide
Collect/
Decide
PMU
PMU
...
Very High-speed
Decisions
10-100 ms time frame
PMU
PMU
Collect/
Decide
PMU
...
Collect/
Decide
PMU
PMU
PMU
...
PMU
Figure 2.
Synchrophasor reporting hierarchy
Figure 1.
Synchrophasor definition
Application of Phasor Measurement Units for Disturbance Recording
9
2.3 PMU Distributed Architecture
The Synchrophasor standard and associated communication
protocol was designed to aggregate data from multiple locations.
As each dataset is transmitted synchronous to top of second and
as each transmitted dataset contains a precise absolute time
stamp, the data aggregation function becomes a simple matter of
combining sets of data with common time stamps. The “box” that
performs this function is known as a Phasor Data Concentrator
or PDC. In a “total” system, there will be a hierarchy of PDCs as
shown in Figure 2. The hierarch is designed to support different
performance criteria/data rates – depending on the application.
With the assumption that higher-level PDCs operate at lower
data rates, the data from the lower layer PDCs provides the most
frequency resolution. Depending on type and number of PMUs
installed in a substation, a substation-based PDC may or may not
be required as this function can be integrated into the PMU.
A major advantage of Synchrophasor measurements compared
to a normal DSR is that, as a result of standardization, data
from multiple manufacturers can be seamlessly integrated. This
is possible because the Synchrophasor standard requires that
magnitude and phase angle errors resulting from magnetic and
filter components be compensated in the final result.
Throughout North America, there exist today “pockets” of data
concentration. Specifically, the Eastern Interconnect Phasor
Project (now the North American SynchroPhasor Initiative – NASPI)
has created a network of PMUs that span most of the eastern half
of the continent. Data is being streamed at a rate of 30 phasors/
sec into a Super Phasor Data Concentrator as operated by TVA.
Communication bandwidths in the order of 64,000 to 128,000
bits per second will be required – depending on the number of
data items and the selected stream rate. At the receiving site,
real-time visualization of the data is available. Additionally, the
data is archived and can be retrieved to perform system dynamic
analysis as well as forensic analysis for larger system events.
In as much as remote communications may be disrupted by an
event, most PMUs/PMU Systems have the ability to locally store
synchrophasor data based on a range of event triggers. Typical
triggers include over/under frequency, rate of change of frequency,
over/under voltage, over current, over/under power, and status
change. Synchrophasor recording times in excess of 20 minutes
can be obtained within the confines of existing PMU memories.
3. Wide Area Recording
The benefits of disturbance recording, or power swing recording,
are already well established. The phasor data captured in these
records are used to validate system models of the power system,
validate the operation of system integrity protection schemes and
wide area protection schemes, and to provide root-cause analysis
of equipment operation during power system faults. Some typical
uses for the data include identifying the impact on the system due
to a loss of generation or loss of a significant transmission line.
Another use for this data is to analyze the performance of distance
relays due to power swings. [6],[7] In all of these cases, for proper
analysis, the phasor data must be measured simultaneously at
various points on the power system. By collecting and coordinating
records from multiple locations, the engineer can evaluate the
response of the system, and specific equipment, to a power
system or power equipment fault. The challenge is to capture
simultaneous recordings across the system.
10
The present method of disturbance recording is to use discrete
recording equipment, and local triggers. DSRs are placed at
key locations on the system. Each DSR is configured much like
a protective relay: trigger criteria are specific for the location of
the DSR. Therefore, a DSR will only create a record when a power
system fault is observable at the location of the DSR. Therefore,
the more remote a DSR is from the center of inertia of an event,
the less likely the DSR will capture a record for an event. Also,
local triggers are dependent on the propagation time of the event
across the system. A common trigger for DSRs is rate-of-change
of frequency. In one known case, full load rejection of a 1,100 MW
generating station took approximately ½ second to propagate
across the utility power system.[7] Local triggers will therefore be
problematic in such a case. With discrete DSRs, and local triggers,
records (such as for the load rejection example) may be created at
different instances in time. An engineer must identify, retrieve, and
combine the appropriate records from multiple devices. And this
assumes that all the DSRs in use are accurately time synchronized,
typically to Coordinated Universal Time using GPS clocks.
Wide area recording or wide area cross-triggering can solve some
of these issues. Wide area recording creates one synchronized
record across the power system when any local DSR triggers a
recording. The challenges in a wide area recording system are
similar to that of local recording, with the added complexity of
communications channel time delays. The only wide area recording
system presently available is a closed, proprietary solution. This
solution links DSRs as part of a client-server software system.
When one DSR triggers a recording, this DSR sends a message to
the server. The server then sends a message to trigger a recording,
with the same trigger time, on all other connected DSRs. This
system solves communications channel delay by using a rolling
data buffer to store data in the DSR. Once the recording is finished,
the server then retrieves the records from all the DSRs. This system
absolutely requires that each DSR is accurately time synchronized,
to ensure the data in the individual records are in phase.
Wide area cross-triggering sends a cross-trigger command to
other DSRs via communications when one DSR triggers for a power
system fault. Wide area cross triggering has not been used, in part
due to the challenges of communications, as the cross-trigger
signal must be sent to multiple DSR locations simultaneously.
Therefore, the complexity of communications is added to the
same challenges in creating simultaneous records. However, the
use of a PMU as the DSR can reduce these challenges.
In a typical DSR, although the records are time synchronized, there
is no agreement among manufacturers as to how and when a
measurement is made. However, when using a PMU as a DSR,
the measurement is standardized and time synchronized per
standard. Therefore, the trigger time of the record is not vitally
important. The data from records captured at two different PMUs
with different trigger times can be coordinated based only on
absolute time.
The other challenge in wide area cross-triggering is sending the
cross-trigger signal to multiple locations across the power system.
This assumes an intact communication channel. Speed is not
critical as long as the PMU can provide pre-trigger data memory.
By setting the pre-event memory to be longer than the trigger
and retrigger communication time, no data is ever lost. The IEEE
Synchrophasor standard has, as part of the message format, a
trigger signal that is typically sent as a PMU-to- PDC signal. Once
in the PDC, logic is needed to receive the trigger signal and then
to forward it to all PMUs connected to the detecting PDC. Once
Application of Phasor Measurement Units for Disturbance Recording
the signal is received by one PMU in a station, that PMU can issue
a GOOSE message to trigger other data captures or execute
controls in other devices in the substation. Figure 3 illustrates this
architecture.
PDC
Controller
Remote
Trigger
Network
Cross-Triggers/Controls
Per C37.118
“Control” Distribution via 61850
GOOSE
Relay
Relay
PMU
Relay
Relay
system shut down. It is very desirable to measure the effect of a
SIPS action on the electric power grid. This measurement is most
easily effected through the collection of synchrophasors across
the system. Using the cross-triggering methodology previously
described, the wide-ranging effects of a SIPS action can be
observed and used to validate system studies and models.
One such scheme protects large multi-generator power plants
against the severe disturbances that occur on transmission lines.
Based on the disturbance severity, the typical results are intensive
swings or loss of plant synchronism, which will lead into loss of
the entire generation complex either by out-of-step protection,
or unit shutdown by protective devices reacting to voltage dips
at auxiliary buses. Wide area recording of synchrophasors allows
the analysis of the power swing phenomena across the system, to
verify the operation of the SIPS scheme.
4.3 Capacitor Bank Performance
Figure 3.
Phasor measurement unit cross-triggering
3.1 The need for cross-triggering PMUs
PMU installations are normally designed to stream PMU data
via communications to a centralized database that stores
synchrophasors quantities for later analysis of the power system.
This seems to eliminate the need for cross-triggering recording,
as the data is readily available at a central location. However,
the data is not necessarily available. As more devices, such as
protective relays, can provide synchrophasors data, the less likely
these devices will continuously stream data to the centralized
database. The bandwidth of communications channels may
limit data transmission, and data storage requirements may limit
reception of data. Also, protection engineers may not have the
same easy access to stored synchrophasors data as the system
operations and system planning departments do.
In addition, for analysis of relatively local events, there may be the
need to capture additional data beyond synchrophasors, such as
power, power factor, and impedance. The cross-trigger signal can
also be used to initiate recording in a traditional DSR as well.
4. Applications of PMU data for analysis
4.1 Large motors
In the industrial environment, many processes have start-up and
shut-down times that are in the multi-second time frame and
sometimes, problems occur that either abort a startup or initiate
an undesired shut-down. Traditional oscillography, although
high-resolution, is typically set to record data only during fault
conditions and, as such, will not record the longer start-up or
shut-down events. Moreover, most industrials will own neither a
swing recorder nor an oscillograph. Synchrophasor capability in
motor protection can enable data capture in these instances and
can provide a high-resolution, long-term view of these events. In
addition, with proper trigger settings, the effects of power system
disturbances on plant processes can be observed.
4.2 AGC / SIPS analysis
System Integrity Protection Schemes (SIPS) is rapidly becoming
a common occurrence in many utilities around the world. A SIPS
event is usually a last ditch effort to prevent a complete power
Capacitor Banks are used to help maintain a flat voltage profile on
the transmission system. Capacitor bank installation typically use
some type of automatic control to switch in and switch out the
capacitor bank. This switching operates on some criteria involving
time of day, voltage magnitude, reactive power magnitude, or
power factor. The performance of the capacitor bank is monitored
at the system operations level by direct observation of the changes
in the system voltage. Direct recording of the changes to local
part of the system could provide some interesting insights into the
impact of capacitor bank operation on the system voltage.
The primary data necessary to analyze the performance of a
capacitor bank is the voltage magnitude and the reactive power
flow. PMUs directly record the voltage and current synchrophasors,
and can also record the real power, reactive power, power factor,
and system frequency. Consider the arrangement of Figure 4. Rich
Substation is a major load substation, with a switched capacitor
bank that operates on voltage magnitude. Recording PMUs are
installed at both Rich Substation and Mark Substation, a major
transmission substation. The PMU at Rich Substation is configured
to trigger a recording on operation of the capacitor bank controller.
Both PMUs are configured to send a cross-trigger command via
IEC61850 GOOSE messaging.
Rich Substation
Mark Substation
Load
Capacitor
Bank
PMU
52
52
Cross-trigger
Signal
52
52
52
52
52
52
52
52
52
52
PMU
Figure 4.
PMU cross-trigger for capacitor bank operation
A voltage profile may look something like that of Figure 5, where
increasing load drags the system voltage down. The voltage
recovers after the capacitor bank is switched in.
Recording the data at both PMUs can provide some valuable
information. The basic information includes the voltage magnitude
at each bus. Once the capacitor bank is switched in, the data
will show the impact on the voltage at each bus, the amount of
Application of Phasor Measurement Units for Disturbance Recording
11
overshoot on the voltage correction, and the time lag between
capacitor switching and voltage correction at the remote bus.
The end goal of using this type of data is to improve the efficiency
of capacitor bank switching, to ensure that bank switching
procedures result in the desired improvement in system voltage
level.
the performance system-wide. Recording synchrophasors in the
substation, along with power flow and device data, can verify the
local operation of the load shed devices, and the local impact
on load. Capturing this data across the system can verify the
performance of the load shed scheme system-wide. In addition,
this information can be used to determined the center of inertia of
the system during the event, and how close the system was to the
voltage instability point.
Rich Substation
Mark Substation
Transmission
Transmission
PMU
UFLS
52
52
52
52
UFLS
PMU
UFLS
UFLS
UFLS
UFLS
52
52
52
52
UFLS
UFLS
Cross-trigger
Signals
Craig Substation
Figure 5.
Transmission
Capacitor bank operation voltage profile
There are two advantages to using IEC61850 GOOSE messaging
as the cross-trigger signal. The first advantage is the GOOSE
message can be sent to one specific device or group of devices, or
it can be sent to all devices on the system. In this example, GOOSE
messages need only be sent between the two PMUs. The second
advantage is the non-proprietary nature of IEC61850. There are
two advantages to using IEC61850 GOOSE messaging as the
cross-trigger signal. The first advantage is the GOOSE message
can be sent to one specific device or group of devices, or it can be
sent to all devices on the system. In this example, GOOSE messages
need only be sent between the two PMUs. The second advantage
is the non-proprietary nature of IEC61850.
Rich Substation
Mark Substation
Load
Capacitor
Bank
52
Capacitor
Bank Relay
52
52
Cross-trigger
Signal
PMU
52
52
52
52
52
52
52
52
52
PMU
UFLS
52
UFLS
UFLS
4.5 Distance relay performance during small
disturbances
Not all disturbances need to be a system-wide phenomenon to
be of interest to study. Significant changes in voltage or current
may cause the operation of a distance relay. Of special concern
are distance relays that use a large over-reaching zone as remote
backup of lines from the next station. Even small disturbances,
such as the loss of a nearby generator, or heavy line loading, may
cause the operation of this distance element. PMUs can be used to
identify events where the apparent impedance of the line comes
close to a tripping zone of the relay.
Significant Generation
52
52
Cross-trigger
Signal
52
PMU
PMU
52
52
21 Relay
21 Relay
52
4.4 Capacitor Bank Performance
12
52
Cross-triggers for load shedding analysis
Load substation without PMU
Analysis of the performance of a load shedding scheme requires
both verifying the performance of local devices, and verifying
52
Figure 7.
Figure 6.
Underfrequency and undervoltage load shedding schemes are
used to prevent system collapse. The typical scheme uses a local
relay with a fixed threshold against voltage or current. A block
of load is shed when the frequency or voltage drops below this
threshold. Multiple thresholds are typically used to shed multiple
blocks of load. The power system phenomenon that predicates
the use of a load shedding scheme is a reduction in the system
frequency or system voltage due to a significant imbalance
between generation and load. At an individual device location, the
apparent impedance will fluctuate in response to the changes in
the system voltage and current.
52
UFLS
Rich Substation
Mark Substation
Figure 8.
Transmission line example
Consider the simple transmission system of Figure 8. There is
significant generation located one bus away from Mark Substation.
When this generation trips off, a small power swing occurs. This
power swing may encroach into the relay operating zone for the
relay at Rich Substation.
The data that is most interesting is the apparent impedance as
seen by the distance relays at both ends of the line. This requires
Application of Phasor Measurement Units for Disturbance Recording
the recording of the current and voltage by both PMUs. In terms
of the total power system, this disturbance may not be significant,
and may not trigger criteria. However, the local PMUs can be
configured to recognize the power swing conditions, and capture
a recording. The cross-trigger signal can be an IEC61850 GOOSE
message that is only received by these two PMUs. A big advantage
of PMU data, is the synchrophasors data is always synchronized.
6. References
[1] L. Swanson, J. Pond, R. Hunt, “An Examination of Possible
Criteria for Triggering Swing Recording in Disturbance
Recorders”, 2005 Fault and Disturbance Analysis Conference,
Georgia Tech, Atlanta, GA, April 25th – 26th, 2005.
[2] NERC Standard PRC-002-1 – Define Regional Disturbance
Monitoring and Reporting Requirements, North American
Electric Reliability Corporation, Washington, DC, August
2006.
[3] NPCC Document B-26 Guide for Application of Disturbance
Recording Equipiment, Northeast Power Coordinating
Council, New York, NY, September 2006.
[4] SERC Supplement Disturbance Monitoring Equipment (DME)
Requirements, SERC Reliability Corporation, Birmingham, AL,
August, 2006.
[5] DRAFT Standard RFD-PRC-002-1 Disturbance Monitoring and
Reporting Requirements, Reliability First Corporation, Canton,
OH.
[6] A. Klimek, R. Baldwin, “Benefits of Power Swing Recording”,
2204 Fault and Disturbance Analysis Conference, Georgia
Tech, Atlanta, GA, April 26th – 27th, 2004.
Figure 9.
Apparent impedance during disturbance
Figure 9 shows some results for a small-scale disturbance. The
apparent impedance seen by the relay came close to the largest
tripping zone of the distance relay. This small margin justifies a
contingency study to determine if the reach settings for this zone
are secure against local small-scale system disturbances.
[7] R. J. Murphy, “Disturbance Recorders Trigger Detection and
Protection”, IEEE Computer Applications in Power, Institute
of Electrical and Electronic Engineers, New York, NY, January
1996.
5. Conclusions
The value of disturbance recording to analyze the response of
the power system to power system faults is well established. For
this reason, the NERC guidelines for recording require utilities to
capture RMS or phasor values of voltage, current, frequency, and
power to analyze power system faults. Phasor measurements with
recording capabilities are ideal devices to provide disturbance
recording. The explicitly time-synchronized synchrophasors data
meets the accuracy requirements and time requirements of the
NERC guidelines.
The real strength of using PMUs for disturbance recording is
the ability to easily support wide area recording using existing
communications networks. Capturing data at various points
on the system provides better analysis of system performance
during power system faults. The challenges of synchronizing
data are eliminated, as each piece of data is explicitly time
synchronized. Cross-triggering signals are sent via nonproprietary
communications, such as defined in the IEEE Synchrophasor
standard and IEC 61850 standards.
070209-v3
Application of Phasor Measurement Units for Disturbance Recording
13
Switch Your
Frame of Mind
Reduce substation communication costs by up to 70%
A breakthrough in networking hardware that can reduce up to 70% of your total
communications costs, the Multilin UR Switch Module is a fully managed, embedded
Ethernet switch for the Universal Relay (UR) family. This advanced, 6-port Ethernet
Switch eliminates the need for external, rack-mounted switches and significantly reduces
the total costs associated with hardware, installation, wiring and troubleshooting,
required for today’s traditional substation communication architectures.
For a mere $200 more*, when compared to the cost of selecting the redundant Ethernet
option on a UR, the Multilin UR Switch Module delivers full station management,
monitoring, and control functionality, with complete communications redundancy.
* USD List Price
UR Ethernet Switch Module
g
Digital Energy
Multilin
GE Multilin
www.GEMultilin.com/URSwitch
[email protected]
Worldwide
Tel: 905-294-6222
North America
Tel: 1-800-547-8629
Europe/MiddleEast/Africa
Tel: +34 94 485 88 00
Smart Grid:
The Road Ahead
Larry Sollecito
GE Digital Energy
1. Introduction
From the time that Thomas Edison commissioned the world’s first
power system in 1882, the electric power industry has continually
moved forward – working to improve the functionality, efficiency,
and availability of electricity. Through evolutionary advancements
in technology, the electrical power industry has transformed the
way we generate, deliver, and consume power today.
As the electric power industry begins the next century, it is on
the verge of a revolutionary transformation as it works to develop
a Smart Grid to meet the needs of our digital society. Society’s
expectations and Utility Commission incentives/penalties are
driving changes to the industry where secure data is required
quickly, on-demand, and in an easy to search way. Customers are
demanding higher reliability and greater choice, and are willing to
examine and change their energy usage patterns.
To achieve the end-goals stated above, a unified vision of the
road to the Smart Grid is needed. Without a unified vision, the
issues currently facing the power system will be addressed
piecemeal by utilities, government agencies, and related power
system organizations. The result of isolated development
activities will be a power system that is plagued by islands of
separation. Subsequently, the power system of the future may
only be realized in limited areas or on a small scale. This article
presents a definition of the Smart Grid and examines the road
ahead to its development, which is only possible when power
system organizations work together to provide a more capable,
secure and manageable energy provisioning and delivery
system. (IntelliGrid Architecture Report: Volume 1, IntelliGrid
User Guidelines and Recommendations, EPRI, Palo Alto, CA and
Electricity Innovation Institute, Palo Alto, CA: 2002. 1012160.)
•
A power system that serves millions of customers and has
an intelligent communications infrastructure enabling the
timely, secure, and adaptable information flow needed to
provide power to the evolving digital economy
From this definition, we can conclude that the Smart Grid must
be:
•
Predictive (operationally and functionally) to preclude
emergencies
•
Self-healing to correct/bypass predicted/detected problems
•
Interactive with consumers and markets
•
Optimizable to make the best use of resources
2. The Smart Grid
•
Distributed in nature with both assets and information
According to the IntelliGrid Architecture Report cited above, the
Electric Power Research Institute (EPRI) defines the Smart Grid as:
•
Transformational to turn data into information
•
Secure from threats and hazards
•
A power system made up of numerous automated
transmission and distribution (T&D) systems, all operating in
a coordinated, efficient and reliable manner
•
A power system that handles emergency conditions with
‘self-healing’ actions and is responsive to energy-market and
utility needs
The Smart Grid must provide robust, reliable, and secure
communication as well as intelligent electronic devices (IEDs) and
algorithms to make the necessary system assessments when
needed. To achieve a Smart Grid, the industry must merge copper
and steel (electricity generation and delivery infrastructure) with
silicon and glass (computation and communication infrastructure).
We are currently at the crossroads in the coming of age of both
technology areas.
Smart Grid: The Road Ahead
15
Figure 1.
The Evolution of IED Communications
Over the past 35 years, we have seen communication speeds
increase dramatically from 300bps (bits per second) to digital
relays that today operate at 100Mbps - an increase of over 300,000
times! (Figure 1) Not only have the communication speeds
changed, but the communication protocols have migrated from
register-based solutions (e.g. - get the contents of Register 5) to
text based data object requests (e.g. - get the Marysville-Kammer
Line Voltages) as implemented through the IEC 61850 protocol.
In addition, the physical interfaces have transitioned from RS-232
serial over copper to Ethernet over fiber or wireless – both local
and wide-area. Interoperability has become a reality and today’s
devices are self-describing and programmable via a standardized
configuration language.
On the device side, the IED and its constituent components
have undergone an evolutionary convergence of multiple
functions and features into a single device (Figure 2). In the past,
protection, control, metering, oscillography, sequence of events
(SOE), annunciation, and programmable automation logic were
all separate functions. Today’s IEDs combine these functions and
more into a single platform that provides superior user interfaces
allowing visual access to metering, SOE, device diagnostics/
solution, and phasor views of data.
3. Enterprise Drivers
In migrating towards a Smart Grid, there are three clear enterprise
areas that are driving the industry forward: financial performance,
customer service, and organizational effectiveness.
3.1 Financial Performance
Financial performance is typically measured in terms of reduced
expenditures. In the case of utilities, cost reduction in capital
improvement, operations, and maintenance are the primary
areas of focus. In the capital improvement arena, advanced
communication technologies and global standards, such as IEC
61850, are enabling new architectures and implementations
such as Relay-to-Relay communication and Process Bus.
These technologies are already demonstrating a reduction in
engineering, implementation, and commissioning costs through
distributed architecture solutions and soft wiring. Standardized
long distance tripping via fiber or wirelessly and analog data
communication not only reduce wiring, but also enable new
applications such as distributed load shed.
3.2 Customer Service
Figure 2.
As described in the introduction, the digital society is demanding
higher availability and better quality of electrical energy delivery
from utilities. The Smart Grid provides the foundation for
addressing these requirements. Detection of incipient problems
is facilitated through monitoring and analysis of electrical
signatures from the transmission, distribution, and industrial
power systems. For example, analysis of electrical signatures
Convergence of Functionality in Today’s IEDs
16
Smart Grid: The Road Ahead
from the distribution system can indicate a downed conductor,
incipient cable fault, or failed capacitors. On transmission systems,
increased communication speeds enable real-time grid control for
enhanced stability. In the industrial realm, motor current signature
analysis can detect excessive motor vibration, motor turn faults,
and broken rotor bars.
•
Real-time pricing/hour ahead emergency pricing /automatic
home response
•
Direct load control
•
Energy usage/optimization display
In the area of power quality, much of today’s electronic equipment
demands 100% availability (or at least 99.999%) and preferably
clean (minimum harmonic content) power. To meet this demand,
the Smart Grid needs to look at new sources of auxiliary power
and the means to protect and dispatch power generation in small
localized areas. Dynamic filtering of power may be offered as a
fee-for-service option. Communication will be key in coordinating
this activity throughout the grid.
•
Load monitoring/sub-metering
•
Remote connect/disconnect
•
Outage detection and isolation/customer trouble call
management
•
Demand profiles
•
Security monitoring
•
Remote home control
•
Remote equipment diagnostics
3.3 Organizational Effectiveness
For an organization to be effective and drive improvement, its
people need operational information. Currently, data is gathered
through periodic polling of a limited set of measurements and
periodic manual inspection of assets.
The Smart Grid view of the substation has changed this paradigm
in several ways:
•
The IED is monitoring more of the assets in a substation,
collecting data, and converting the collected data into
information. Timely information about an asset enables
optimal use of that asset.
•
The IED is able to communicate, on exception, the semantics
of the situation. Semantic-based communication provides
a standard, unambiguous view of the information and
minimizes the documentation and configuration effort.
•
The automation aspect provides seamless information
aggregation, storage, and dissemination
The overall effect of the Smart Grid in this domain is to improve
manpower utilization through automation, and optimized asset
utilization through automatic monitoring and operation.
Home energy optimization will become automatic but the system
will also provide user interfaces to foster further energy use
optimization (Figure 3).
Figure 3.
Home Energy Dashboard
4.2 Distributed Generation / Microgrids
4. Application Domains
As the Smart Grid comes into existence, there are several
application domains that promise to drive its development.
4.1 Advanced Metering Infrastructure (AMI)
and Smart Home
The first domain, and the area where there is a lot of activity,
is advanced metering infrastructure (AMI). AMI represents the
next phase of automatic meter reading whereby the meter is
responsive to two-way communication, dynamically adjusting for
the price of electricity, and interacting with the various loads in
the metered facility. The various public utility commissions, who
strive to adequately balance the supply and demand aspects of
electricity, are driving the implementation of this domain.
The advanced functionality of AMI enables the creation of the
Smart Home. The combination of these domains brings to bare
functionality such as:
The second domain driving the Smart Grid’s development is
Distributed Generation/Microgrids. As the number of Distributed
Energy Resources (DER) increases throughout the electronic
enterprise, and communication to the various resources becomes
more pervasive, the ability to operate segments of the grid as
islands or Microgrids becomes reality. The drivers are clear the desire for high-availability and high-quality power for the
digital society. Figure 4 shows a vision of the evolving Microgrid
structure.
DER includes not only positive power generation such as solar, wind,
and micro-turbines, but also negative power generation through
demand response programs, controllable loads, and direct load
control. Renewable energy resources are becoming competitive
with existing generation resources. Many utility commissions are
incentivizing additional renewable generation sources (solar and
wind) on the grid. Battery and inverter technology is evolving such
that a utility can justify the capital cost of the installation based on
the difference in price from buying energy at night at a low price
and selling it back during peak daytime rates.
Smart Grid: The Road Ahead
17
Microgrids do present challenges to the utility from a protection,
control, and dispatch perspective. Traditional protection is based
on the fact that for a short circuit, significant overcurrents will flow
in a direction towards the fault. In a Microgrid environment, a
significant portion of the generation will be inverter-based which,
through design, is current limited. New protection philosophies
will need to be developed to protect these systems.
On the control side, it will be necessary for the Microgrid to be
seamlessly islanded and re-synchronized. It may be required that
the Microgrid be dispatched as a single load entity. This will mean
that a local controller will have to be able to communicate with
all of the DERs on the Microgrid, and dispatch both watts and
vars, to maintain a constant power flow at the Point of Common
Coupling (PCC). The amount of power dispatched will be set either
on a contracted value or optimized based on the dynamic price
of electricity.
Electric and Plug-in Hybrid Electric Vehicles (PHEVs) provide an
additional twist in the operation of the Microgrid because they
are mobile. As they connect and disconnect from a Microgrid,
the controller will have to be aware of their existence and include
their potential effect in the operation of the Microgrid. One of
the operational use-cases that evolves from this scenario is the
Microgrid controller sending buy/sell messages, that include
dynamic pricing information, to the owner of the vehicle.
4.3 Wide Area Measurement and Control
The third major domain area that is rapidly evolving is that of Wide
Area Measurement and Control Systems (WAMACS). These systems
have the ability to synchronously measure and communicate the
instantaneous state of the power system through a measurement
known as the Synchrophasor. The ability to dynamically view the
state of the power system is similar to being able to view a beating
heart. Normal and stressed system states can be assessed in real
time and acted upon to affect dynamic control. Today’s power
system operators take action in the multi-second to multi-minute
time frame, but WAMACS can make and execute decisions in the
100 millisecond time frame.
The development of the WAMACS infrastructure entails the
installation of measurement and data collection devices in
substations, a reliable wide area communication network,
and data concentration, visualization, and decision facilities.
Work is ongoing in all these areas. As utilities build out their
communication infrastructure, they must take notice of the
performance requirements dictated by real time synchrophasor
communication. Many of the projects will take five or more years
to come to fruition.
5. Smart Grid Architecture – Putting the
Pieces Together
As the operating scenarios of the above application domains
are fleshed out, it soon becomes apparent that there are
opportunities and needs for cross-application communication.
For example, the WAMACS function detects a system instability
and determines that it needs to shed load in the sub-second
time frame. It is clear that there needs to be a communication
interface between the WAMACS applications and the Smart
Home and Microgrid environments. In order to achieve this crossdomain communication, an architecture is needed to define the
parts of the communication system as well as to define how they
interact.
The architecture process defines a set of plausible scenarios
(Enterprise Activities) spanning the entire energy enterprise (utility,
industrial, commercial and residential). The scenarios then enable
analysis on the data and resulting communication requirements
needed to construct a complete, high-level set of functions for
the communications infrastructure to enable the envisioned
functionality. The requirements can be categorized as:
•
Communication configuration requirements, such as one-tomany, mobile, WAN, and LAN
•
Quality of service and performance requirements, such as
availability, response timing, and data accuracy
Figure 4.
Microgrid Vision
18
Smart Grid: The Road Ahead
Figure 5.
Smart Grid Architecture
•
Security requirements, such as authentication, access control,
data integrity, confidentiality, and non-repudiation
•
Data management requirements, such as large databases,
many databases across organizational boundaries, and
frequent updates
•
Constraints and concerns related to technologies, such
as media bandwidth, address space, system compute
constraints, and legacy interface
•
Network management requirements, such as health and
diagnostics of infrastructure and equipment, remote
configuration, monitoring, and control
Putting together the pieces, there are several guiding principles
that show the way to an architecture. According to the IntelliGrid
Architecture Report cited earlier, an architecture should be
technology independent, based on standard common services, a
common information model, and generic interfaces to connect it
together. Figure 5 is a conceptual view of such an architecture.
While adapters can accommodate the above heterogeneity,
to achieve interoperability using off the shelf components, we
need standards for what data is exchanged and how data is
exchanged. Furthermore, these standard information models
and interfaces must be applicable to a variety of utility services.
A standardized common information model solves what is
exchanged. A standardized set of abstract interfaces solves how
data is exchanged. A single technology for every environment will
never be agreed upon so adapters will still be required to convert
between different technologies.
It should be noted that the IEC 61850 communication protocol,
“Communication Networks and Systems for Utility Automation,”
meets these requirements today.
This protocol defines
internationally standardized models for protection, control,
metering, monitoring, and a wide range of other utility objects. In
addition, it defines a standard set of abstract services to read and
write to these models. Most importantly, it provides a mechanism
for self-description of the data models to any requesting client. This
feature becomes of paramount importance for auto configuration
as the number of devices in a domain becomes large, for example,
millions of electric meters.
6. Conclusion
The journey towards establishing a Smart Grid is underway.
The industry is working to meet the demands of the digital society.
Progress is being driven forward by demonstrating solid financial
performance through reduced expenditures, meeting customers’
demand for higher availability and better quality of delivery, and
increasing organizational effectiveness through high-quality,
timely operational information. Several application domains such
as Advanced Metering Infrastructure, Smart Home, Distributed
Generation/Microgrids, and Wide Area Measurement and Control
are also key in driving the Smart Grid’s development.
Traveling down the road to a Smart Grid will take an organized
effort to overcome isolated development and unite power system
organizations to provide a more capable, secure, and manageable
energy provisioning and delivery system.
070209-v2
Smart Grid: The Road Ahead
19
All your favorite channels
Looking for a proven, reliable integrated solution,
that covers the complete range of electric utility
applications; including teleprotection, SCADA,
video, voice, data, Ethernet WAN?
GE’s Lentronics JungleMUX substation hardened
multiplexers work seamlessly over all of your
network infrastructures, including T1 leased line,
microwave radio and SONET fiber optic systems.
Click on www.GEDigitalEnergy.com/JMUX and see
how solving your telecommunication challenges is
our business.
g
T1 Multiplexer
Digital Energy
Lentronics
GE Digital Energy - Lentronics
visit our website at www.GEDigitalEnergy.com/jmux
Worldwide
Tel: 905-294-6222
North America
Tel: 1-800-547-8629
Worldwide
Tel: 905-294-6222
Europe/MiddleEast/ Africa
Tel: +34 94 485 88 00
North America
Tel: 1-800-547-8629
SONET Multiplexer
Europe/MiddleEast/Africa
Tel: +34 94 485 88 00
Leader or Follower
Developing the Smart Grid Business Case
John McDonald
GE Energy
© 2008 IEEE. Reprinted, with permission, from IEEE Power & Energy Magazine
A world of energy developments are coming together right now
with consequences that we will all feel for generations to come.
World energy consumption is forecast to triple by 2050. Climate
change has become more fact than fiction. Our nation’s energy
security is vulnerable due to our current dependence on finite
natural resources for energy generation. We are in a time of
fundamental change in the utility industry, and we must make
decisions as significant as those wrestled over by Edison and
Westinghouse as they debated DC versus AC power distribution.
We have the technology, and market forces are converging to
fundamentally change the way the grid operates to improve
efficiency, reliability, and energy security.
After a decade of expectation that advanced grid and metering
technologies will provide a step change in system cost and
performance, momentum towards smart-grid implementation
is finally building. In practical terms, this means the industry will
see larger-scale commitments to not just advanced metering
infrastructure (AMI) deployment but investments in new and
improved grid technologies vastly expanding the scope of benefits
available to a utility and its customers.
A number of forces are aligning to drive momentum, particularly
as the industry is entering an era where expectations of what the
grid can do are being stretched to activities that include providing
instantaneous corrections to outages, enabling customers
to make energy consumption decisions, and facilitating the
smooth transfer of power flow from large wind farms. Supporting
these new expectations is an emerging environment of greater
regulatory acceptance and understanding of the real benefits
that smart grid can deliver.
Despite the changing environment, there are challenges
preventing utilities from fully understanding and developing
their smart grid strategies. Continually evolving technologies, silo
efforts within utilities, and cautious skepticism of benefits all lead
to slowing down decision-making and investment in smart grid
technologies. To address these challenges and to quickly move to
capture the benefits of smart grid that are accrued over a longer
duration, utilities need to get organized and act with urgency.
Utilities should develop a comprehensive fact-based assessment
of smart grid in the absence of developing the comprehensive
assessment of smart grid benefits; however, utilities are missing
out on an opportunity to capture significant benefits for customers
that are self-funding over the life of the investments.
This column discusses the key driving factors behind smart grid
momentum, the challenges facing utilities, and suggests how
utilities should think about developing their comprehensive smart
grid business cases.
Leader or Follower
1. Momentum Is Building Towards a
Smart Grid Vision
The past two years have been pivotal in the industry for
broadening the definition of smart grid from AMI to encompass
a larger set of customer facing and grid applications combining
communications, digital hardware, and decision making to
drive step-change improvements in cost, quality, and reliability
of service. This broadening represents a natural evolution of
improved technologies combined with expanding needs. In this
column, smart grid is defined as the set of advanced, digitally
based technologies that can be attached at the boundary
of generation and transmission and all the way through the
grid across the meter and into the home. This includes phasor
measurement, centralized and integrated voltage and VAR
controls, grid automation, advanced monitoring and diagnostics
as well as AMI. Home area network (HAN) applications based on
advanced telecommunications technologies are included in the
scope of smart grid.
This myriad of technologies represents an industry entering a
new era where the combination of environmental pressures
and tremendous investment needs are reshaping traditional
fundamentals. Large-scale grid investments are needed to
overhaul the U.S. transmission and distribution grids as historical
investment has lagged. Pressure on improved grid reliability
and power quality is increasing as more regulators think about
attaching penalties and rewards against performance. Customer
Developing the Smart Grid Business Case
21
satisfaction ratings are also viewed with greater scrutiny. Grid
security has emerged as a serious concern and much attention
is being paid to understanding how new investments will improve
grid resilience. New telecommunications technologies with
advanced encryption as well as remote inspection of assets
enable a stronger grid.
Against this backdrop, the environmental push to drive energy
efficiency, integrate large amounts of intermittent large-scale
renewables as well as distributed generation has stimulated
serious interest in ensuring the grid is well equipped to
accommodate this need. AMI-enabled demand response has
also assumed a critical role in the discussion of finding least-cost
alternatives to new build or expensive power purchases. While
direct load control without AMI will continue as an important
building block for demand response, introducing consumer
decisions to reduce usage will provide necessary additional load
reduction. Customer awareness will continue to grow if expected
rate increases materialize and share-of-wallet for electricity
continues to grow as a proportion of total household spending.
In parallel, technology functionality has grown significantly.
Basic SCADA systems have matured into advanced applications
of distribution management systems (DMSs). Geospatial
information systems (GISs) are normally integrated with outage
management systems (OMSs). There are opportunities for
significant improvement in service when this integration expands
to include DMSs and workforce deployment. More advanced
sensors are entering the market enabling real-time evaluation of
asset performance.
In addition to these factors, the industry has benefitted from a
dramatic reduction in technology costs. For example, disconnect
relay switch costs have declined from about US$120 to US$40
in just the past three years. While other core hardware prices
in the transmission and distribution space have been steady
or increasing as demand has outgrown supply, the smart grid
technology curve continues to push forward.
These trends are being matched with growing regulatory support
for smart grid. At the national level, the U.S. Congress passed the
Energy Independence and Security Act at the end of 2007 that
provides grants, sets policy direction for smart grid through a
national committee, and recommends accelerated depreciation
for technology investments. A number of states are also moving
to providing direction setting for smart grid investments. California
is pushing utilities to provide an integrated vision of smart grid.
In Illinois, Commonwealth Edison has filed for a surcharge to
rates that represent a broad set of smart grid investments. In
Ohio, the regulatory environment is pushing utilities to move
beyond AMI filings to broader grid improvements. In Florida and
Ontario, regulators have set precedence for faster depreciation of
potentially stranded metering assets.
2. The Case for Smart Grid
With the growing momentum for smart grid implementation, the
central question to developing the business case is how real is the
impact from the smart grid. Smart grids can help utilities deliver
significant benefits. Some of these benefits are quantifiable
reductions from the current baseline of costs (O&M and capital).
Other benefits will accrue by reducing expected future costs,
particularly in terms of deferred or cancelled capital investments
through peak reduction and advanced asset monitoring and
22
Leader or Follower
diagnostics. Not all benefits are monetary; reliability and customer
service improvements will ultimately result in better rate case
outcomes through generating goodwill for utilities, but may not
directly translate into monetary benefits.
Significant O&M cost reductions can be achieved through a
smart grid. Many of the AMI-related benefits have been discussed
across the industry over the last couple of years. Reductions
across metering, customer care, billing, and credit and collections
have been documented by utilities that have had AMI systems
implemented for several years. Additional O&M reductions in field
operations in both gas and electric have been more challenging
to capture but the opportunities are certainly visible.
Avoided capital spending is also a critical and key benefit supported
through demand response and asset optimization driven based
on advanced monitoring and diagnostics. For bundled utilities,
avoided capital covers both generation and grid investments and
for unbundled utilities this typically involves avoided purchases in
addition to grid investments.
There are significant reliability benefits driven by smart grid as
well. The combination of AMI and fault detection, isolation and
restoration (FDIR) creates a powerful driver to reduce SAIFI
(System Average Interruption Frequency Index) as well as SAIDI
(System Average Interruption Duration Index). On the other hand,
more accurate reporting of outages as opposed to traditional
estimation methods may actually increase CAIDI (Customer
Average Interruption Duration Index) and quicker automatic
restoration does result in shifting events from SAIFI into MAIFI
(Momentary Average Interruption Frequency Index).
A smart grid will also drive improvements in the customer service
experience. By delivering accurate and instantaneous meter
reads, bill estimation challenges will be eliminated. The ability to
complete the feedback loop to provide customers with up-todate service information becomes easier. Currently, many utilities
struggle with being able to provide accurate information back to
the customer on key areas such as billing and outage notification
and updates.
Smart grid also delivers significant control to the customer in
their ability to influence energy consumption decisions. Demand
response is a critical feature in the suite of smart grid capabilities.
Peak shifting as well as overall conservation (kilowatt-hour
consumption reduction) are both important impacts of a demand
response system. While utilities have been administering direct
load control programs for many years, the ability to leverage
two-way communications with greater control and accuracy, and
with customer input, can substantially enhance a utility’s demand
response capabilities.
In addition, safety on the grid is enhanced through smart grid
applications. Integration of mobile workforce applications with
advanced geospatial information systems and asset monitoring
and diagnostics provides frontline crews with accurate location
and state of sensitive equipment. The ability to diagnose outages
more comprehensively also helps in crew safety preparation.
In addition, smart grid is critical in preparing our grid infrastructure
for renewables and distributed generation. With intermittency
of larger-scale renewables (e.g., wind farms) and the technical
challenges of flowing power back and forth from distributed
generation, grid controls are increasingly important. Future
benefits of integrating plug-in hybrid vehicles (PHEVs) and more
advanced microgrids will also require smart grid infrastructure to
become reality.
Developing the Smart Grid Business Case
Understanding these benefits and their secondary and tertiary
impacts is important to establishing the case for the smart grid.
Recent research and work in how customer utilities are thinking
about smart grid suggests that most utilities are drawing a broad
and cautious brush over understanding the full capabilities the
smart grid can deliver.
Many of the benefits discussed above have been discussed
by utilities in recent filings and industry discussions. Recent
filings in California, for example, demonstrate the NPV positive
business cases that AMI can deliver. Announcements by Progress
Energy on integrated volt/var controls, PG&E’s distribution
automation program, AEP’s GridSmart initiatives and Duke
Energy’s development of broad smart grid applications are all
recent developments that indicate that momentum for a more
comprehensive smart grid definition is growing.
3. Typical Challenges
With the breadth of benefits the smart grid can deliver and
the improvements in technology capabilities and reduction in
technology costs, investing in smart grid technologies should be
a serious focus of utilities. Utilities, however, have been slow to
respond. A number of all-too-familiar factors contribute to the
stationary mode.
Technology obsolescence plays a critical role in utility caution in
investment. The fear new technologies will be outdated within a
short period results in a wait-and-see approach. This is particularly
the situation with communications technologies that form the
backbone from the meter end-point typically to the substation.
A decade ago, discussions centered around mobile versus
fixed AMR. Now the question is more complicated with options
spanning powerline, broadband, and radio frequency spectrumbased wireless technologies as well as emerging WiMAX, 3G and
4G or LTE (long-term evolution) applications that are fast maturing.
While communication technologies are certainly evolving, utilities
should feel encouraged that open standards—often IP based—are
fast becoming the expectation for emerging technologies. This
means the array of smart equipment required (e.g., meters,
sensors, capacitor bank controllers) is designed to be agnostic
to alternative communication environments. In addition, hybrid
solutions that combine technologies to fit specific terrain
characteristics of a utility are also becoming more common and
help mitigate the risk of single technology investments. Humayun
Tai, a partner with McKinsey who works with utilities on developing
their smart grid strategies, commented “Utilities need to work
backwards from understanding what specific grid and metering
functionalities they need, how this translates into bandwidth,
latency and other requirements and then determine what is the
optimal communications solutions. AMI and grid applications
have distinct needs and can be addressed differently.”
Smart grid efforts at utilities also struggle from being structured
within departmental silos. A comprehensive and realistic
assessment requires a cross-functional perspective merging
engineering, field operations, back office, IT, customer operations
as well as generation skills and expertise. Many smart grid
efforts find their genesis in AMI efforts that have typically been
ongoing within customer service organizations across many
utilities. Utilities that have had success moving their smart grid
vision forward have formed cross-functional groups with strong
executive mandates to explore and develop the business case for
smart grid. For example, AEP has a cross-functional group driving
Leader or Follower
smart grid development from both the corporate and operating
company perspectives.
In addition, there is significant confusion around the extent and
sequencing of IT integration and needs. The industry is beginning
to get its head around how engineering applications need to be
integrated as well as how they link up with back office applications.
Systems integration players are at early stages of understanding
the full picture and new products and solutions from IT platform
vendors are beginning to emerge. Despite this early state of IT
development, utilities should recognize smart grid deployments
will not happen overnight or within a year. Deployments will take
place typically over several years allowing sufficient time for
utilities to test and develop their integration plans.
The aggregation of these challenges leads to an environment
where utilities are overly cautious. The vision for the smart grid
becomes narrow rather than comprehensive. A prolonged period
of piloting multiple technologies prevails. While pilots are an
important step in confirming both the suitability and the benefits
of smart grid technologies, often pilots are designed to test specific
rather than a broad set of capabilities.
4. Moving Forward—and Moving with
Urgency
As industry conditions, technologies and utility needs mature,
there is urgency in understanding and developing a smart grid
vision. Utilities need to take advantage of the emerging regulatory
environment and the interest PUCs are showing in granting
accelerated depreciation or pilot financing costs.
Because benefits from smart grid accrue over time, investments
in smart grid need to be paced such that the cumulative benefits
correspond with the timing of needs. For example, developing
and launching a successful demand response program takes
time because it relies on educating and convincing customers. To
defer near-term, peak-based generation investments or to avoid
expected high-cost power purchases in constrained markets, a
utility will need to launch its demand response program several
years ahead of when it needs the additional capacity.
This implies utilities need to develop a perspective on smart grid
that is both comprehensive and fact-based. Developing what
constitutes the business case requires little investment. It requires
dedicating a cross-functional team of experts within the utility
backed by executive sponsors. The teams’ task would be to develop
the benefits available across a range of benefit possibilities, and
assessing the costs associated with these based on existing and
emerging technology options. Many utilities already have an AMI
case developed to some degree and in those instances, the smart
grid business case would build upon the AMI case.
In assessing the business case NPV values, utilities have
significant flexibility to shape payback and return profiles. Pacing
deployment quickly or over a longer time provides a trade-off
between spreading upfront investment costs and the payback
period. Targeting deployment with areas where there are greater
benefit opportunities (e.g., higher losses, uncollectibles, lower
reliability pockets) can also alter payback periods. The business
case can also be developed by evaluating smart grid capabilities
individually. For example, AMI, FDIR, Volt/VAR can be assessed as
three independent modules and their returns can be measured
separately although there will be cost interdependencies.
Developing the Smart Grid Business Case
23
In addition to the quantifiable benefits, the business case must
also ensure reliability, customer satisfaction and environmental
and distributed energy benefits are also articulated clearly along
side the financial information. The importance of the nonfinancial
benefits will depend on a utility’s specific performance and
regulatory situation.
Assessing costs can be challenging as technology evolves quickly.
The business case should be assessed initially with different
technology options for communications (e.g., RF, BPL, advanced
telecommunications). Different levels of investment also link to
corresponding levels of benefits.
4. Leader or Follower?
As momentum towards smart grid grows, utilities need to get
organized and systematically develop fact-based perspectives.
The gap between leader and follower in the industry will widen
as new implementations lead to more utilities establishing
precedence and standards for how smart grid deployments
should be defined.
Once a business case has been pulled together, the next step is
to understand how smart grid fits into the broader utility strategy,
especially developing the regulatory case for these investments.
Completing the financial and business regulatory cases helps the
utility build a “road map” for deployment which indicates where
the value is, what is the right sequencing of decisions and how
regulatory actions should progress (and with what step-gates).
Vendor assessment and technology selection should follow once
the deliverables on the smart grid are established.
062509-v1
24
Leader or Follower
Developing the Smart Grid Business Case
An Enterprise Information Architecture Based on
SOA Enables a Smarter Grid
Marcel Van Helton
GE Fanuc Intelligent Platforms
As part of President Obama’s stimulus program, an estimated
$19B will be invested in the Smart Grid to meet the needs of the
21st century—requiring a major overhaul of America’s electrical
grid. It will allow for better integration of renewable energy sources
and flexible trading of energy as well as improved reliability.
Providing better demand response, automatic recovery scenarios
and improved “pre-warning” systems of potential distribution
problems, the Smart Grid will enable the right measures to be
taken to avoid potential failures in the grid.
The acronym for SMART with respect to computer technology
stands for Self-Monitoring, Analysis, and Reporting Technology,
a monitoring system for computer hard disks that detects
and reports on various indicators of reliability to help anticipate
failures. Although originally developed for computer technology,
the description aligns well to one of the goals of the smart grid,
which is to enable more reliable operations. In addition, it aims for
increased efficiency and sustainability.
A smart grid requires leveraging multitudes of information flow
between different systems. For example, seamless integration of
information from a SCADA system on top of a wind farm into a
SCADA system of a larger utility may help determine short-term
generation capacity while other parts of the information, for
example, the transformer temperature profile, can concurrently
be leveraged by the maintenance management system.
It is critical to have the ability to leverage both operational data
and non-operational data. Operational data captures the “here
and now” and keeps power flowing over the grid, while nonoperational data spans a broader timeline. It provides insight
into why things happened, keeps track of how people fixed what
happened and predicts when problems may happen again. This
“memory” is a key element of the intelligence in the smart grid.
To make the smart grid a reality, utilities need to leverage
intelligence by ensuring the availability of critical information
needed for smarter, faster decision making. So how do you
collect, organize, distribute, integrate, secure and interpret the
vast sea of information in today’s grid environment to deliver this
intelligence?
All this information needs to be tied together to optimize the
entire electrical supply chain, which means utilities need a
comprehensive overview of the actual status of the suppliers
(generators), the warehouses (storage capacity), the trucks (the
grid) and the demand (consumers) to make intelligent decisions.
And not surprisingly, other industries also share this challenge of
developing IT structures that integrate multiple plant structures
and suppliers’ IT systems to make information available in “realtime.”
1. Abandoning the Traditional
Information Pyramid
2. Moving to a Service Oriented
Architecture (SOA)
Most utilities have implemented some type of Supervisory Control
and Data Acquisition (SCADA) system with a set of Remote
Telemetry Units (RTUs) or Intelligent Electronic Devices (IEDs),
primarily focused on key operational parameters in the case
of an outage or potential overload. Traditional pyramid-based
information models bring information up from these various RTUs
and IEDs using serial communication methods into a centralized
SCADA system, which has propriety interfaces or an integrated
means to communicate or provide Load Management System
(LMS)/Demand Management System (DMS)/Outage Management
System (OMS) functionality.
Companies can build intelligence into their systems with an
Industrial Service Oriented Architecture (SOA) such as GE
Fanuc Intelligent Platforms’ Proficy1 SOA, which improves
interoperability and enables composite applications that leverage
a cross-system, real-time data and services bus and repository.
Used in combination with appropriate industry standards, SOA
allows for a “plug and play” architecture for IT systems, whereby
functionality can quickly be added, changed or removed to meet
market demands.
These information models, even with point-to-point defined
communication of different applications, lead to many information
streams, resulting in an unmanageable web of information.
Furthermore, as more renewable generating assets are added to
the grid, the amount of information is expected to grow; some
organizations estimate an increase of up to 15 times the amount
of information in today’s systems.
The cornerstone of the SOA structure is the Enterprise Data and
Service Bus, which includes core components such as standardsbased data models, security, configuration, event management,
and user management. Other software functions can be
“plugged” into the SOA using and adding to the information and
1Trademark of GE Fanuc Intelligent Platforms, Inc.
All other brands or names are property of their respective holders.
An Enterprise Information Architecture Based on SOA Enables a Smarter Grid
25
services within the SOA. It is this information and service-enabling
architecture that becomes the foundation of the intelligence in
the smart grid as more and more value-added applications
are plugged into the SOA structure to optimize the different
components of the electrical grid.
The Industrial SOA structure with a standards-based data model
supports a transition from a “computer friendly” hierarchical
model for data flow to a more natural “user-friendly” approach
based on putting the data and applications in the context of
the real world. As a result, the information hierarchy is based on
the actual structure of Energy Enterprises, Operation Centers,
Substations, and Equipment—down to consumers’ homes.
In the structure of today’s smart grid IT solutions, functions like
SCADA, LMS, DMS or OMS are all components leveraging the
information available in the Enterprise Bus. The benefit is realized
when these components have access to the same data (real-time
and historical) in the same context of the real-world structure of
the grid. This approach avoids common pitfalls of current systems
where context is continually added and lost as data flows from
one system to the next.
Delivering a solid backbone for leveraging “real-time” and
historical information and providing an infrastructure to distribute
the information between different locations, an SOA also provides
developers with common architectural services like security,
event management and user services. They can design the
information structure based on specific industry requirements,
adding relevant intelligence to the system.
4. Basic Building Blocks of the Smart
Grid
4.1 Connecting to the Grid: The Collection
Function
The innovations of today’s communication networks are the
basis of the smart grid, without which a smart grid would be
impossible. We all remember the maximum baud-rate of 1200
bps (bits per second) where the SCADA systems could only poll
the most important operational data such as volt, amps and vars.
These systems were limited in bandwidth and would often only
flow operational data from its source to other systems, which left
a wealth of non-operational data islanded—never achieving its
full potential.
Today, these bandwidth issues no longer need to constrain
the systems and through the strength of a modern LAN/WAN
architecture with a modern SOA, non-operational data can be
included at all levels, bringing value to a whole new set of users.
For example, Marketing and Product Planning can monitor
equipment for efficiency in the actual physical environment and
plan more efficient product offerings for the future, and substation
maintenance can monitor conditions prior to failure and take
steps to proactively replace equipment before catastrophic
failures occur.
Figure 1.
Industrial SOA Structure.
3. A Structured Approach Toward
Information
Traditional SCADA systems are tag-based and when information
is transferred to other systems, the information structure has to
change, depending on the requirements of each system. However,
with an industrial SOA, information is available in a structured
manner and all software functions share the same structure, so
the data is maintained in a common context.
For example, operational and non-operational information such
as the actual primary and secondary power and overview of the
actual temperature and dissolved gasses is collected and stored
into the SOA. Having operational and nonoperational information
in a structured way in the same “repository” allows easier retrieval,
interpretation and management of information.
Aside from actual “real-time” information, it is crucial that the
system can store historical time series information and collect
information in the IED/RTU and store it in the same historical
system. As a part of the SOA Enterprise Architecture, a built-in
Process Historian stores information in the context provided by
26
the Enterprise Bus. There are two types of information stored:
normal time series operational data, and events and alarms.
Today’s IP-based communication through wireless (UMTS, radio
Modems) and fixed or leased lines allows the transfer of more
information on a continuous basis. Modern IEDs and RTUs provide
a wealth of non-operational data—for example, fault event logs—
whereby the more you know about the health of your systems,
the better and smarter your decisions can be.
The collection function needs to map the information from the
IED/RTU to the SOA structure, either through the traditional DNP
3.0 and the IEC 60870.5.101/103/104 protocols or through the
relatively recent IEC 61850 protocol. The key advantage of the
IEC 61850 is that it’s already modeled to map the most common
substation automation equipment.
4.2 Enabling Better Decision Making: The Data
Warehouse
Information has a tendency to creep up into personal Excel®
files located in different computer systems. To effectively put
information to work, data must be up-to-date, easily accessible,
and maintained in a common context. In this way, the data is
easy to correlate.
For example, the serial historical data of a transformer could be
correlated with sequence of event data of that same transformer.
An Enterprise Information Architecture Based on SOA Enables a Smarter Grid
To do this efficiently, a centralized information repository is
required. The SOA structure with a built-in historian database
allows you to store all information in a single logical location,
whereby all the data is in a common context so there is a single
version of the “truth.”
To satisfy standard and ad-hoc information requirements, the
user should have easy access to two types of reports. Standard
reports can be used for compliance purposes, for example, a
standardized emission report. Ad-hoc reports might be needed to
look at the performance of a single asset while troubleshooting a
grid problem.
Operators who can combine continuous and ad-hoc-based
operational and non- operational information can develop a
better understanding of the actual status and behavior of the
assets within the grid. Consequently, they can make better and
faster decisions in order to optimize performance.
4.3 Providing Operator Decision Support: The
Workflow Function
While the additional information collected and generated to run
a smart grid more efficiently allows operators and/or systems to
make better decisions faster, the vast amount of additional data
also requires an increase in the operators’ knowledge base. That’s
where an industrial workflow tool can help—ensuring certain
procedures are automatically triggered based on events arising
from the information stored in the SOA structures.
The workflow provides the correct procedures to the operator and
ensures appropriate sign off of those procedures. For example,
if the system detects a potential failure in a breaker based on
non-operational information, a workflow gets triggered, and the
operator initiates a maintenance order to repair the critical part
at the breaker. When the linesmen schedule the work, another
workflow gets triggered to ensure that the operator switches off
the segment and re-routes the power; once the work is complete,
the workflows are closed.
4.4 Functions Build on Workflow
Whereas workflow is an application that uses the SOA structure,
other applications are built using the workflow product. For
example, an outage management system can use the workflow
“foundation” to execute work orders for the linesmen.
5. The Smart SOA Foundation
The Industrial SOA is the basis of the Enterprise Information
Architecture, which meets the requirements of the smart grid. It
enables a structured information model based on standards that
combine “real-time” information with historical data, including
events-based and time-series information, and allows other
applications to seamlessly access the necessary information to
perform certain functions.
Furthermore, an SOA is not limited to a single location or server,
but easily spans over multiple locations with advanced search and
location capabilities to aggregate information across the grid. For
example, if a utility recently experienced an equipment failure, the
SOA would enable a quick search across the grid for all similar
equipment and collect its current operational and non-operational
information to assess the potential for similar failures.
Enabling companies to build intelligence into their systems by
leveraging all critical information, an Enterprise Information
Architecture powered by SOA can help make the smart
grid a reality—delivering improved reliability, efficiency and
sustainability.
Figure 2.
Proficy SOA Enables Composite Application.
062909-v1
An Enterprise Information Architecture Based on SOA Enables a Smarter Grid
27
Smart Grid Solution
Intelligent Power Management for
Operators, Managers & Engineers
Power System Monitoring & Simulation
Advanced Monitoring & Trending
Energy Accounting
Real-Time Simulation
Event Playback
Load Forecasting
Geographic Information System Map
Intelligent Substation
Substation Automation
Switching Management & Optimization
Load Management
Distribution / Energy Management System
Automatic Generation Control
Economic Dispatch
Supervisory Control
Interchange Scheduling
Reserve Management
Intelligent Load Shedding
Load Preservation
Load Curtailment & Restoration
Load Shedding Validation
GE-06-June-2009_b.indd 1
etap.com
OPERATION TECHNOLOGY, INC.
8 0 0 . 47 7. E TA P | 9 4 9 . 9 0 0 .10 0 0
6/26/2009 2:58:16 PM
Key Smart Grid Applications
B. R. Flynn, PE
GE Energy
1. Abstract
The culmination of attention by utilities, regulators, and society
for smart grid systems to address operational and electrical
efficiencies, improving system reliability, and reducing ecological
impacts, has resulted in a significant number of discussions
around the requirements and capabilities of a Smart Grid. This
session will explore three key Smart Grid functions with the
strongest business case justification.
•
Demand Optimization
•
Asset Optimization
In addition to technical functionality, the incremental costs and
benefits for each example will be presented. This session will
include a review of example system architectures and specific
functionality to provide improvements to performance metrics,
capital and O&M costs and reduction to environmental concerns.
2. Introduction
IMES are changing. More information about the many forces
behind these changes is being published every day.
•
It is expected that the demand for electric energy will triple
by 20501.
•
Digital-quality power which represents about 10% of the
total US electrical load will reach 30% by 2020.
•
Rolling power outages in developing countries, previously
just an unwelcome fact of life, have escalated to the level of
“national emergency”.
•
Distributed generation including renewable and sustainable
power has grown and is expected to double every three
years2.
•
The average age of Power Transformer in the US is estimated
to be 40 years old.
•
Generation of electricity accounts for 40% of the US’s CO2
emissions3.
1 Source: U.S. Army Corps of Engineers ERDC/CERL TR-05-21
2 REN21 2007 update + EER
3 EPRI, “Electric Sector CO2 Impacts, February 2007”, Carbon Dioxide
Emissions from the Generation of Electric Power in the United States,
July 2000, DOE
For several years, electric utilities have been turning to Smart Grid
technologies to help deal with these pressures. Currently utilities
are focusing their efforts on three major areas.
Delivery Optimization consists of the efforts by the electric utility
to improve the efficiency and reliability of the delivery systems.
Demand Optimization focuses on solutions to empower the end
consumer and to better manage the evolving demand and supply
equation along the distribution feeder
Asset Optimization is the application of monitoring and
diagnostic technologies to help manage the health, extend the
useful life and to reduce the risk of catastrophic failure of electrical
infrastructure.
A Smart Grid drives value with integration of open technologies.
Protect Critical
Information &
Infrastructure
Systems
& Process
Ensuring
Safety
Business Applications
Delivery
Optimization
Business Process
Delivery Optimization
Security & Safety
•
Protect
Consumer
Information
Demand
ptimization
Asset
Optimization
Drives Business &
Consumer Value
Computing & Information
Technology
Visualization,
Analytics &
Automation
Communications &
Information Infrastructure
Open Transfer of
Critical Information
Energy Infrastructure
Intelligent Sensors
& Controllers
Energy Consumer
Empowered
Consumers
Figure 1.
Smart Grid Technologies
Key Smart Grid Applications
29
Figure 1 highlights the relationship of the various major functions
involved with most Smart Grid solutions. The business applications
of Delivery, Demand and Asset Optimization are part of the highest
value smart grid business applications. They work on a structure
that includes computing and information technology designed
to support the applications and help manage the significant
amount of data required for each of the applications to function
appropriately.
Since the source for much of the data lies outside the utilities
data center, an open, secure and powerful communications and
information infrastructure is required to access the data. The
communications infrastructure touches portions of the entire
energy infrastructure from generations to the meter and in many
cases into the energy consumer’s premise. This system is designed
in a way that maintains a high level of safety and security.
3. Smart Grid Solutions
Figure 2.
Architectures for solutions that focus on each of the three areas
will be presented including a summary of functionality and value.
Station Based IVVC System
3.1 Delivery Optimization
Delivery Optimization includes two major areas which will be
reviewed separately, Efficiency and Reliability.
Utilities have deployed methods to improve the efficiency of their
electrical systems. This discussion will focus on recent efforts to
change the methods of controlling voltage and VARs on their
distribution systems.
Efficiency
There are two types of losses in electrical systems, restive and
reactive. Typically, capacitors are deployed in the station and
along the backbone of the feeder to help manage the reactive
losses and support the voltage. Distribution capacitors are usually
operated by local controllers based on Powerfactor, load current,
voltage, VAR flow, temperature, or the time (hour and day of
week). Through the use of communications to modern controllers,
coordinated VAR Optimization is available. This type of system
usually controls the line capacitors based on the Powerfactor
measured at the substation. Depending on the feeder’s load,
quantity of capacitors and existing type of control, the system
can reduce the electrical losses.
There are two example architectures for these functions as shown
in the following diagrams.
With the addition of communication and smart controllers to
various voltage regulators and transformer load tap changers,
utilities are taking finer control of their feeder’s voltage with a
technique called Conservation Voltage Reduction (CVR).
Controlling the regulator and LTC, the utility can reduce feeder
voltage levels and depending on the amount of resistive and linear
loads along the feeder, will reduce the load at the substation.
Utilities have seen a 1% reduction for a 4% reduction in voltage.
The architecture shown above illustrates a system where the
communications from the distribution devices communicate with
the substation. This system utilizes a station based system which
operates the line devices for VAR or voltage control based on the
measurements at the station and along the feeder.
30
Figure 3.
Centrally Based IVVC System
The next system illustrates a centrally based communications
system and a centrally located algorithm. This system adds the
ability to monitor the voltage from selected customers’ meters.
The customers selected for voltage monitoring are those at the
end or at low voltage points of various line segments. Depending
on the AMI system’s latency and bandwidth, selected meter
voltage reads are incorporated into the IVVC application to allow
for a wider operating range of the CVR application.
The capacitor communications can utilize a separate DA system
communications or can share the AMI communications network
and backhaul. Both systems require communications with the
substation for control and monitoring of station transformers and
feeders. Some utilities have estimated the aggregate payback
for an integrated volt/VAR control system to be less than two
years. These benefits can be significant and are typically shared
between the utility and the end consumer depending on the rate
structure. Regulating Commissions are becoming more and more
receptive to granting a portion of these benefits to the utility to
help compensate for the expenses of the systems. The balance
Key Smart Grid Applications
of the benefits often flow through to the end consumer in the rate
making process. The benefits typically include:
•
Improved distribution system efficiency
•
Reduced distribution line losses
•
Improved voltage profile along feeder
•
Improved system stability and capacity
•
Defered capital upgrades
•
Reduced energy demand
•
Reduction of environmentally harmful emissions
Reliability
Many utilities are turning to Smart Grid applications to provide
improvements to reliability metrics. Two typical architectures are
shown below. The first shows a station based automation system
with smart devices installed in the substation and in distribution
circuit reclosers, or switches for underground systems. Faults are
detected between switches and are isolated under control of the
automation system. Unfaulted sections of the feeder are restored
from alternate sources depending on available sources and their
capacity to carry the additional load.
Figure 5.
Centralized Distribution Automation
Many regulators do not provide the utility with direct incentives
for investment designed to reduce outages, however rely on
the threat of penalties to encourage investments. This makes it
prudent for the utility management to apply automaton to the
most troublesome circuits first. However, it is also necessary to
determine the appropriate level of automation to be applied. The
following chart compares the costs per improvement to customer
minutes interrupted compared to a typical non-automated base
case circuit.
Chart 1.
System Costs by Customer Minutes Interrupted (CMI)4
Figure 4.
Station Based Distribution Automation
The architecture shown below uses a centralized distribution
automation system with the addition of communication to smart
metering. The algorithm can utilize the information from the smart
meters to assist in detecting outages outside of the monitored
range of the automated devices. This includes fuses or other
non-automated fault interrupters. The meters can detect and
report an outage and can report successful restoration. Using the
restoration detection, the system can determine customers have
been properly restored as expected. After restoring customers
after a fault, any meters not responding with a restoration can
indicate a possible nested outage. This capability can significantly
improve the time to respond to these outages.
In this study, the Base Case consists of 2 Manual Operated
Disconnect Switches (MOD) and a shared tie switch with another
feeder. Case 1 changes the MODs to manual operated reclosers
on overhead circuits and customers on that circuit experience
a corresponding improvement to outage minutes. Case 2 adds
communications to the reclosers, allowing the dispatcher remote
manual control to of the switches. Cases 2-6 have a significant
improvement to reliability metrics from the ability to localize faults
and isolate and restore unfaulted segments of the feeder. Case
3 adds the ability for automation software for Fault Detection,
Isolation and Restoration (FDIR). Case 4 utilizes an expensive
but highly effective automation system where the normally
open tie switch is normally closed. This closed loop automation
4 Justifying Distribution Automation at OG&E, by Cristi Killigan and Byron
Flynn, prepared for DistribuTECH 2009
Key Smart Grid Applications
31
circuit requires expensive dedicated high speed communications
between the switches to prevent over-tripping. Case 5 adds the
ability to utilize a centralized automation system with the addition
more sophisticated electrical network model supporting the
automation software. Case 6 adds the ability to detect outages/
restoration from the meters.
3.2 Demand Optimization
In recent years, Demand Optimization has generated a significant
amount of interest. This solution has drawn the attention of
regulators and the US federal government. Often the benefits
from Demand Optimization is what is in mind when regulators
installation of smart metering. These are the benefits most directly
experienced by the end consumer.
Solutions around Demand Optimization are varied and range from
simple advanced metering systems to full home automation. The
figure below is designed to highlight the possibilities of a solution.
The following is a partial listing of example choices, programs,
and devices which consumers and utilities are electing to deploy
around Demand Optimization.
Empowering Customer Choice & Control:
•
Critical Peak Pricing
•
Time of Use Rate
•
Green Power Choices
•
CO2 Management Choices
•
Prepaid Metering
•
Voluntary or Automatic Control of Energy Demand
•
Usage Management – by Appliance
•
Home Energy Management
•
Net Metering, collecting KWH, KVARH, Voltage,
•
Power Quality
Providing Security & Safety Management:
Figure 6.
Energy Theft
•
Tamper Detection
•
Visual inspection during installation
•
Monitoring and baselining usage patterns
•
Interruptions and usage pattern changes
•
Detect load-side voltage with disconnect
Enabling Distributed Generation:
Demand Optimization System Architecture
This system usually consists of several different devices in the
consumer’s home, usually at a minimum it includes a smart
thermostat connected to a smart meter. Often the consumer
connects their PC to a web page containing information about
their usage and available programs the consumer can select.
Addition of a Smart Display allows continuous communications
between the utility and the end customer. These displays can be
simply a light which changes color based on time of use rates
or can be a complex full graphic color touch screen connected
to a full home automation system. New smart appliances, wall
switches and wall sockets are beginning to become available for
integration into a Home Automation Network (HAN). Some systems
integrate local generation into the home including Photo Voltaic,
solar panels, small wind generation, or more exotic systems such
as fuel cells.
The architecture above also includes an optional HAN gateway.
This gateway is intended to allow secure connection between the
utility systems and the consumer’s HAN when the communications
to the end meter is not robust enough to handle the utility-HAN
communications. In this option, the HAN gateway is connected to
a customer’s broadband Internet connection, providing a higher
bandwidth connection than may be possible through some of the
legacy slower speed AMI systems.
32
•
•
Photo Voltaic (Solar)
•
Wind
•
Biomass
•
Geothermal
Incorporating Distributed Storage:
•
Li-Ion Battery
•
Fuel Cells
•
Plug in Hybrid Electric Vehicle (PHEV as a storage device)
Facilitating New Programs and Capabilities:
•
Load Management Programs
•
Demand Response Program
•
Distributed Generation
•
Storage Management
•
Automatic Meter Reading
•
New Communications with Customer
Key Smart Grid Applications
•
Power Quality Management
•
Remote Service Switch
•
Cold Load Pickup
•
System Cyber Security
•
System Management
The monitoring and diagnostics of the station transformer can
include simply monitoring temperatures or continuous monitoring
of the oil for combustible gasses and moisture. Advanced
monitoring today can calculate internal hot spot temperature, the
transformer dynamic load ability and future capacity over time,
the insulation aging factors and data from many other models.
The value of a TOU and CPP program has been estimated for
a 6% peak load reduction as shown below.5
•
$26 MM/yr O&M and capital expenditure reduction
•
126 GWh generation reduction
•
Consumer’s savings: up to 10%
•
57K tons of CO2 reduction
•
Estimated capital cost of $12MM
•
Estimated O&M cost of $11MM/yr
Monitoring of the station or line protection relays provides valuable
information regarding the health of the breaker including operating
times, total interrupted fault current, and operation counts.
Data collected from the meters can help determine near realtime
loading on the distribution cables, especially the underground
cables, and loading on the local distribution transformers.
This can help distribution engineers improve the distribution
planning and design and help rebalance the load along the
phases.
3.3 Asset Optimization
Much of the modern electrical system was installed over 40 years
ago. Unfortunately, many devices in the system are frequently
being pushed to operate at overload conditions. One of the single
most expensive pieces of the distribution system is the station
power transformer. Given that the general life expectancy of
power transformers is around 40 years; this can result in a risky
and expensive challenge.
The architecture shown below illustrates how Asset Optimization
solutions can be added to a Smart Grid.
Figure 8.
Transformer reliability comparison6
The figure above contains the comparison of the power transformer
failure rates with and without monitoring. For transformers without
monitoring, the risk of catastrophic failure is approximately .07%
and for transformers with on-line monitoring and diagnostics it is
.028%, resulting in a reduced risk of failure of 2.5 times.
For example, if an average electric utility with 1 million customers,
installed an M&D system which monitored the dissolved gas,
temperature and load on their the transformerfleet, rated 20MVA
and above, the costs and benefits is determined as follows:
An annual capital expenditure savings of $12MM/yr with an
investment of $42MM in capital and $1.2MM/yr in additional
maintenance. These numbers would result in a net present value,
using 8.7% over a 15 year system life, of over $54MM.
Figure 7.
5 Based on avg. 1 MM customer utility, California Statewide Pricing Pilot,
Monitoring and Diagnostics (M&D) System Architecture
June 2006
The architecture includes monitoring and diagnostics on the
primary station transformer, station breaker, and distribution
feeder devices.
6 60% is an industry accepted effectiveness number for a quality
monitoring system. The failure reduction figure is based partly on a
CIGRE study. As an additional reference, a study conducted by one of
KEMA’s utility client shows that Distribution Automation projects can
reduce the costs of repairing substation transformers by 67%.
Key Smart Grid Applications
33
4. Summary
While each of these solutions can have significant benefits to
an electric utility and their consumers, elements of each can
be leveraged for uses outside those stated in the previous
discussions.
For instance, an IVVC system installed to control the VARs and volts
can coordinate with a DA system installed to improve reliability.
The rerouting of the distribution system to restore unfaulted
sections of a feeder can significantly change the voltage and VAR
profile of that feeder. If the IVVC system is operating on the same
system network model as the DA system, it can continue to control
the VARs and volts in the temporarily reconfigured distribution
system.
Coordinating the asset management system to work with the DA
system can result in new options to dynamically move load off of
overloaded equipment by moving the normally open tie between
different feeders to different locations. If the systems are working
on the same network model, the DA system can continue to
operate to improve reliability on the newly reconfigured network.
Furthermore, the CVR functions in the IVVC system should be
coordinated with the Demand Optimization systems to maximize
the benefits of improved load management. Depending on the a
number of factors such as rates, prices, and system loading, the
utilities’ operations staff can pick the most cost effective method
of reducing load by lowering voltage, issuing an automatic
demand response signal and/or changing the consumer’s time of
use rate.
leverage elements of the network model in the office and can
share data with other utility systems utilizing standards based
methods.
5. References
Papers from Conference Proceedings (Published, or pending
publication):
[1] Dalton, Howard, and Flynn, Byron, “SCADA Integrated with
Transformer Monitoring Systems over a Secure Wireless
Ethernet”, 2005 DistribuTECH Conference and Exhibition, San
Diego, California January 25-27, 2005.
[2] Flynn, Byron, “Case studies regarding the integration of
monitoring & diagnostic equipment on aging transformers
with communications for SCADA and maintenance”,
DistribuTECH 2008, Conference and Exhibition, Tampa
Convention Center, Tampa, FL, January 22-24, 2008.
[3] Flynn, Byron, “Secure Substation Automation Architectures for
Operations & Maintenance”, 2006 DistribuTECH Conference
and Exhibition, Tampa, Florida, February 7-9, 2006.
[4] Flynn, Byron, “What is the real potential of the Smart Grid?”
Autovation 2007, The AMRA International Symposium, Reno,
Nevada, September 30 – October 3, 2007.
[5] Lathrup, Steve, and Flynn, Byron, “Distribution Automation
Pilot at PacifiCorp”, Western Energy Institute, 2006 Operations
Conference, Costa Mesa, CA, April 5-7, 2006.
[6] Stewart, Robert, and Flynn, Byron, “Modeling DA Improvements
to Reliability Performance Metrics”, 2007 WPDAC, Spokane,
Washington, April 3 – 5, 2007.
Figure 9.
Smart Grid System Architecture
The architecture of a fully integrated system as shown above is
represented in the above figure. Properly designed using open
interoperable products and techniques, the various systems can
minimize the amount of duplication, can more easily integrate
with existing utility infrastructures, can more effectively share
communication infrastructure across the system, can share
070209-v2
34
Key Smart Grid Applications
Industrial Strength
Wireless AMI
MDS MercuryTM 3650
The MDS Mercury 3650 provides the high capacity and secure
throughput required for AMR/AMI backhaul, as well as distribution
automation, SCADA, mobile data, and video surveillance applications.
For the AMR/AMI market the versatile MDS Mercury 3650 is the ideal
choice featuring:
High capacity throughput (WiMAX platform)
• Up to 9 Mbps
• Supporting multiple services on the same infrastructure
Highly secure data handling
• Radius authentication and MAC address filtering
• AES 128-bit encryption & VLAN tagging to segregate
sensitive data
Rugged industrial grade construction
• Operational temperature ranges from -40C to +70C
• Class 1/Div 2 approval and IEEE 1613-compliant under industrial
MDS Mercury 3650
High speed industrial WiMAX networking
g
Digital Energy
MDS
GE Digital Energy
For more information visit www.gemds.com/mercury
Worldwide
Tel: 1-585-242-9600
North America
Tel: 1-800-474-0964
From Concept to Commissioning
Call us at 905-569-8030
E-mail: [email protected]
Website: www.virelec.com
Virelec Ltd., Mississauga, Ontario
The Evolution of Distribution
Jiyuan Fan, Stuart Borlase
GE Energy
© 2009 IEEE. Reprinted, with permission, from IEEE Power & Energy Magazine
With smart grids, confidence and expectations are high. To
various degrees, utilities are putting smart grid initiatives in place,
and many of the technologies that paraded under the smart
grid banner are currently implemented in utilities. The smart
grid initiative uses these building blocks to work toward a more
integrated and long-term infrastructure. If all goes as expected,
smart grids will provide tremendous operational benefits to
power utilities around the world because they provide a platform
for enterprise-wide solutions that deliver far-reaching benefits to
both utilities and their end customers.
The development of new technologies and applications in
distribution management can help drive optimization of the
distribution grid and assets. The seamless integration of smart
grid technologies is not the only challenge. Also challenging is the
development and implementation of the features and applications
required to support the operation of the grid under the new
environment introduced by the use of clean energy and distributed
generation as well as the smart consumption of electricity by
end users. Distribution management systems and distribution
automation applications have to meet new challenges, requiring
advances in the architecture and functionality of distribution
management, i.e., an advanced distribution management system
(DMS) for the smart grid.
feeder sections. It can reduce the service restoration time
from several hours to a few minutes, considerably improving
the distribution system reliability and service quality.
•
Integrated voltage/var control (IVVC) has three basic
objectives: reducing feeder network losses by energizing
or de-energizing the feeder capacitor banks, ensuring
that an optimum voltage profile is maintained along the
feeder during normal operating conditions, and reducing
peak load through feeder voltage reduction by controlling
the transformer tap positions in substations and voltage
regulators on feeder sections. Advanced algorithms are
employed to optimally coordinate the control of capacitor
banks, voltage regulators, and transformer tap positions.
•
The topology processor (TP ) is a background, offline processor
that accurately determines the distribution network topology
and connectivity for display colorization and to provide
accurate network data for other DMS applications. The TP
may also provide intelligent alarm processing to suppress
unnecessary alarms due to topology changes.
•
Distribution power flow (DPF) solves the three-phase
unbalanced load flow for both meshed and radial operation
scenarios of the distribution network. DPF is one of the core
modules in a DMS and the results are used by many other
DMS applications, such as FDIR and IVVC, for analyses.
•
Load modeling/load estimation (LM/LE) is a very important
base module in DMS . Dynamic LM/LE uses all the available
information from the distribution network—including
transformer capacities and customer monthly billings, if
available, combined with the real-time measurements along
the feeders—to accurately estimate the distribution network
loading for both individual loads and aggregated bulk
loads. The effectiveness of the entire DMS relies on the data
1. Current Distribution Management
Systems
Distribution Management Systems (DMSs) started with simple
extensions of supervisory control and data acquisition (SCADA)
from the transmission system down to the distribution network.
A large proportion of dispatch and system operation systems in
service today rely on manual and paper-based systems with little
real-time circuit and customer data. Operators have to contend
with several systems and interfaces on the control desk (“chair
rolls”) based on multiple network model representations. The
experience of operators is key to safe system operation. With
an increase in regulatory influence and the focus on smart grid
advanced technologies, there is a renewed interest in increasing
the investment in distribution networks to defer infrastructure
build-out and to reduce operating and maintenance costs
through improving grid efficiency, network reliability, and asset
management programs.
Various DMS applications are commonly used today.
•
Fault detection, isolation, and service restoration (FDIR) is
designed to improve system reliability. FDIR detects a fault on
a feeder section based on the remote measurements from
the feeder terminal units (FTUs), quickly isolates the faulted
feeder section, and then restores service to the unfaulted
The Evolution of Distribution
37
accuracy provided by LM/LE. If the load models and load
values are not accurate enough, all the solution results from
the DMS applications will be useless.
•
•
Optimal network reconfiguration (ONR) is a module that
recommends switching operations to reconfigure the
distribution network to minimize network energy losses,
maintain optimum voltage profiles, and to balance the
loading conditions among the substation transformers,
the distribution feeders, and the network phases. ONR can
also be utilized to develop outage plans for maintenance or
service expansion fieldwork.
Contingency analysis (CA) in the DMS is designed to analyze
potential switching and fault scenarios that would adversely
affect supply to customers or impact operational safety. With
the CA results, proactive or remedial actions can be taken by
changing the operating conditions or network configuration
to guarantee a minimal number of customer outages and
maximum network reliability.
•
Switch order management (SOM) is a very important tool
for system operators in real-time operation. Several of the
DMS applications and the system operators will generate
numerous switch plans that have to be well-managed,
verified, executed, or rejected. SOM provides advanced
analysis and execution features to better manage all switch
operations in the system.
•
Short-circuit analysis (SCA) is an offline function to calculate
the short-circuit current for hypothetical fault conditions
in order to evaluate the possible impacts of a fault on the
network. SCA then verifies the relay protection settings and
operation, and recommends more accurate relay settings or
network configuration.
•
Relay protection coordination (RPC) manages and verifies
the relay settings of the distribution feeders under different
operating conditions and network reconfigurations.
•
Optimal capacitor placement/optimal voltage regulator
placement (OCP /OVP) is an offline function used to determine
the optimal locations for capacitor banks and voltage
regulators in the distribution network for the most effective
control of the feeder vars and voltage profile.
•
The dispatcher training simulator (DTS) is employed to
simulate the effects of normal and abnormal operating
conditions and switching scenarios before they are applied
to the real system. In distribution grid operation, DTS is a very
important tool that can help operators evaluate the impacts
of an operation plan in advance or simulate historical
operation scenarios to obtain valuable training on the use
of the DMS. DTS is also used to simulate the conditions of
system expansions.
2. Transformation of the Grid:
Increasing Complexity
Distribution networks have not always been the focus of
operational effectiveness. As supply constraints continue,
however, there will be more focus on the distribution network
for cost reduction and capacity relief. Monitoring and control
requirements for the distribution system will increase, and the
integrated smart grid architecture will benefit from data exchange
between the DMS and other enterprise applications. The
emergence of widespread distributed generation and consumer
demand response programs also introduces considerable
impact to the DMS operation. Smart grid technologies will add a
tremendous amount of real-time and operational data with the
increase in sensors and the need for more information on the
operation of the system. Utility customers will be able to generate
and deliver electricity to the grid or consume the electricity from
the grid based on determined rules and schedules. This means
that the consumers are no longer pure consumers but sellers or
buyers, switching back and forth from time to time. It requires
that the grid operates with two-way power flows and is able
Energy Star
Appliances
Thermal,
Stability Limits
Automation
MW
Solar
Panels
Energy
Management
Smart
Lighting
Load
13,000
11,000
9,000
Transmission
Distribution
7,000
12 AM
12 PM
Hour of Day
12 AM
Load Center
Digital Meter
Plug in Hybrid (PHEVs)
POWER
Power
Generation
Figure 1.
Generation
Switchyard
Transmission
Substation
Distribution
Substation
Wind
PV
Transformation of the grid.
38
The Evolution of Distribution
End User
to monitor and control the generation and consumption points
on the distribution network. Figure 1 illustrates some expected
transformation of the grid.
The distributed generation will be from disparate sources and
subject to great uncertainty. The electricity consumption of
individual consumers is also of great uncertainty when they
respond to the real-time pricing and rewarding policies of power
utilities for economic benefits. The conventional methods of LM
and LE in the traditional DMS are no longer effective, rendering
other DMS applications ineffective or altogether useless. The
impact of demand response management (DRM) and consumer
behaviors may be modeled or predicted from the utility pricing
rules and rewarding policies for specified time periods, which can
be incorporated into the LM and LE algorithms; this requires a
direct link between the DMS and the DRM applications. When the
DRM application attempts to accomplish load relief in response
to a request from the independent system operator (ISO), it will
need to verify from the DMS that the DRM load relief will not result
in any distribution network connectivity, operation, or protection
violations. The high penetration of distributed generation will
require the load flow algorithm to deal with multiple, incremental,
and isolated supply sources with limited capacities, as well as a
network topology that is no longer radial or is weakly meshed. In
a faulted condition, the distributed generation will also contribute
to the short-circuit currents, adding to the complexity of the SCA,
RPC, and FDIR logic.
(GIS), an outage management system (OMS), or a meter
data management system (MDM) via a standard interface.
Standardized Web-based user interfaces will support
multiplatform architectures and ease of reporting. Data
exchange between the advanced DMS and other enterprise
applications will increase operational benefits, such as meter
data management and outage management.
Figure 2.
Advanced distribution management for the smart grid.
3. Advanced Distribution Management
Systems
•
FDIR will require a higher level of optimization and will need
to include optimization for closed-loop, parallel circuit, and
radial configurations. Multi-level feeder reconfiguration,
multi-objective restoration strategies, and forward-looking
network loading validation will be additional features with
FDIR.
•
IVVC will include operational and asset improvements-such
as identifying failed capacitor banks and tracking capacitor
bank, tap changer, and regulator operation to provide
sufficient statistics for opportunities to optimize capacitor
bank and regulator placement in the network. Regional
IVVC objectives may include operational or cost-based
optimization.
•
LM/LE will be significantly changed where customer
consumption behaviors are no longer predictable but more
smartly managed individually and affected by distribution
response management.
•
TP, DPF, ONR, CA, SCA, and RPC will be used on a more frequent
basis. They will need to include single-phase and three-phase
models and analysis, and they will have to be extended down
the network to individual customers. Distributed generation,
microgrids, and customer generation (such as plug-in hybrid
electric vehicles (PHEVs) will add many challenges to the
protection, operation, and maintenance of the distribution
network. Small generation loads at the customer interface
will complicate power flow analysis, contingency analysis,
and emergency control of the network. Protection and
control schemes will need to account for bi-directional power
flow and multiple fault sources. Protection settings and fault
restoration algorithms may need to be dynamically changed
to accommodate changes in the network configuration and
supply sources.
A number of smart grid advances in distribution management are
expected, as shown in Figure 2.
•
•
Monitoring, control, and data acquisition will extend further
down the network to the distribution pole-top transformer
and perhaps even to individual customers by means of an
advanced metering infrastructure (AMI) and/or demand
response and home energy management systems on the
home area network (HAN). More granular field data will help
increase operational efficiency and provide more data for
other smart grid applications, such as outage management.
Higher speed and increased bandwidth communications
for data acquisition and control will be needed. Sharing
communication networks with an AMI will help achieve
systemwide coverage for monitoring and control down the
distribution network and to individual consumers.
Integration, interfaces, standards, and open systems
will become a necessity. Ideally, the DMS will support an
architecture that allows advanced applications to be easily
added and integrated with the system. Open standards
databases and data exchange interfaces (such as CIM, SOAP
, XML, SOA , and enterprise service buses) will allow flexibility
in the implementation of the applications required by the
utility, without forcing a monolithic distribution management
solution. For example, the open architecture in the databases
and the applications could allow incremental distribution
management upgrades, starting with a database and a
monitoring and control application (SCADA ), then later adding
an IVVC application with minimal integration effort. As part of
the overall smart grid technology solution or roadmap, the
architecture could also allow interfacing with other enterprise
applications such as a geographic information system
The Evolution of Distribution
39
•
Databases and data exchange will need to facilitate the
integration of both geographical and network databases in
an advanced DMS. The geographical and network models
will need to provide a single-phase and three-phase
representation to support the advanced applications. Ideally,
any changes to the geographical data (from network changes
in the field) will automatically update the network models
in the database and user interface diagrams. More work is
required in the areas of distributed real-time databases,
high-speed data exchange, and data security. Take, for
example, the interfaces and applications required to support
roving PHEVs on the utility’s (or another utility’s) distribution
network. Point-of-use metering and energy charge or credit
must be managed and tracked on the distribution network.
This is a challenge in terms of not only the additional load
or potential supply (and related protection and control
issues), but also the tracking and accounting of energy use
or supply at various points on the distribution network or on
a neighboring utility’s distribution network. This will be a huge
challenge for utilities and will lead to a significant change in
data management and accounting—away from the once-amonth meter reading and billing of customers. The customer
interface challenge is illustrated in Figure 3.
Energy Star
Appliances
Analytics and visualization assimilate the tremendous
increase in data from the field devices and integration with
other applications, and they will necessitate advanced filtering
and analysis tools. Visualization of the data provides a detailed
but clear overview of the large amounts of data. Data filtering
and visualization will help quickly analyze network conditions
and improve the decision-making process. Visualization in an
advanced DMS would help display accurate, near real-time
information on network performance at each geospatially
referenced point on a regional or systemwide basis. For
example, analytics and visualization could show voltage
magnitudes by color contours on the grid, monitor and alarm
deviations from nominal voltage levels, or show line loading
through a contour display with colors corresponding to line
loading relative to capacity. System operators and enterprise
users will greatly benefit from analytic and visualization tools
in day-to-day operations and planning.
•
Enterprise integration is an essential component of the smart
grid architecture. To increase the value of an integrated
smart grid solution, the advanced DMS will need to interface
and share data with numerous other applications. For
example, building on the benefits of an AMI with extensive
communication coverage across the distribution system and
obtaining operational data from the customer point of delivery
(such as voltage, power factor, and loss of supply) helps to
improve outage management and IVVC implementation.
•
Enhanced security will be required for field communications,
application interfaces, and user access. The advanced DMS
will need to include data security servers to ensure secure
communications with field devices and secure data exchange
with other applications. The use of IP -based communication
protocols will allow utilities to take advantage of commercially
available and open-standard solutions for securing network
and interface communications.
Smart
Lighting
Load Center
Digital Meter
4. A New Way of Thinking
Plug in Hybrid
Electric Vehicles
(PHEVs)
The customer interface challenge.
•
•
Solar
Panels
Energy
Management
Figure 3.
Load forecasting and load management data will also help
with network planning and the optimization of network
operations.
Dashboard metrics, reporting, and historical data will be
essential tools for tracking performance of the distribution
network and related smart grid initiatives. For example,
advanced distribution management will need to measure and
report the effectiveness of grid efficiency programs, such as
var optimization or the system average interruption duration
index (SAIDI), the system average interruption frequency
index (SAIFI), and other reliability indices related to delivery
optimization smart grid technologies. Historical databases
will also allow verification of the capability of the smart
grid optimization and efficiency applications over time, and
these databases will allow a more accurate estimation of the
change in system conditions expected when the applications
are called upon to operate. Alarm analysis, disturbance, event
replay, and other power-quality metrics will add tremendous
value to the utility and improve relationships with customers.
Smart grids are not really about doing things a lot differently than
the way they are done today. Rather, they are about doing more
of what we already do—sharing communication infrastructures,
filling in product gaps, and leveraging existing technologies to a
greater extent while driving a higher level of integration to realize
the synergies across enterprise integration. A smart grid is not
an off-the-shelf product or something you install and turn on
the next day; it is an integrated solution of technologies driving
incremental benefits in capital expenditures, operation and
maintenance expenses, and customer and societal benefits. A
well-thought-out smart grid initiative builds long-term focus. It is
not a one-time solution but a change in how utilities look at a set
of benefits across applications and remove the barrier created by
silos of organizational thinking. While current smart grid initiatives
are typically driven by regulatory pressure and tend to focus more
on the meters as a direct impact on consumers, we are likely to
see more technology-rich initiatives after well-proven smart grid
evaluations (“staged deployments”). Expect to see traditional
distribution management evolve to include advanced application
the smart grid.
070209-v3
40
The Evolution of Distribution
A Paradigm Shift in Protection, Control
and Substation Automation Strategy
Vahid Madani, Godwin Duru
Pacific Gas and Electric Company
Mark Adamiak
General Electric Co.
Abstract
The electric power industry is experiencing significant changes
in protection, control, monitoring, recording and energy
management. The requirements for improved efficiency in
all phases of engineering, maintenance, and operation are
examined. The demand for reliability has placed emphasis on
service restoration and has highlighted the need for a more
precise exercise of the art and science of protection & control
both for local as well as the more complex schemes protecting
the integrity of the interconnected grid.
This paper is a transformation journey in substation integration
encompassing the components of new technology for a
major power company in North America which has resulted
in the implementation of tomorrow’s technology today. The
transformation blueprint is a blend of fundamentals using
new technology with the vision of the future. The markers
include elements of leadership and technical reasons, process
for capturing the scope, harmonization of different plans and
disciplines to a united vision, and justification strategies and cost
analysis for significant capital investments.
The article then leads to the details of protection and control
technology selection and application, design philosophy,
implementation strategies, establishment of partnerships for
protective relay feature developments. The paper then examines
the skill sets needed for a companywide rollout program and lays
the groundwork for the future directions. Both transmission and
distribution challenges and solutions are presented.
Integration of protection and control functions such as stub bus,
breaker failure interlocking, automatic restoration with voltage
supervision, loss of voltage during source switching, local /
remote access, and fail-safe operation mode are some of the
challenges of integration described in this paper. The flexibility of
the architecture to adopt the IEC 61850 process bus interface is
also described.
Pioneering designs and visions influenced decisions that
have propelled the journey from conception to reality are also
presented including innovative concepts for “drop-in plug and
play” control building that have become the hallmark of future
system protection.
The obstacles, struggles, and victories gave rise to a prototype
philosophy including the modularization and achievements in
new protection and control standards. This groundbreaking
work has led to the company’s commitment to upgrade over 200
substations companywide by 2014. The progress detailed in this
journey has ignited other professionals and companies to follow
the lead and to become an advocate of this pioneering path.
1. Introduction
The electrical grids are, in general, amongst the most reliable
systems worldwide. These large interconnected systems, however,
are subject to a host of challenges – aging infrastructure, need for
siting new generation near load centers, transmission expansion
to meet growing demand, distributed resources, dynamic reactive
compensation, congestion management, grid ownership vs.
system operation and reliability coordination, asset management
and life cycle planning, etc.
The aging infrastructure is at the center of the challenges facing
today’s expected grid reliability, availability, and demand for faster
restoration. The advancements in microprocessor and network
communication are amongst a host of technology enablers that
provide the opportunity for engineering a sustainable and reliable
infrastructure.
While microprocessor relays have gained full industry acceptance,
there is a large number of legacy electromechanical and solid
state devices still in operation resulting in increased maintenance
and failure costs as well as concerns that those costs will increase.
It is not trivial to determine end of life and probability of failure for
electromechanical relays. Spare parts availability for older relays
is only one issue. Experience shows that in some cases legacy
A Paradigm Shift in Protection, Control and Substation Automation Strategy
41
relays have long exceeded their life expectancy and continue to
perform reasonably well, particularly if well maintained. Reduction
in the skill set and knowledge base familiar with troubleshooting,
testing, and repairs of the old technology is across the power
industry spectrum from manufacturing to end-users [1], [2].
The probability of failure and replacement needs for equipment
can be approximated by the following criteria: age; maintenance
practices and records; obsolescence; visual inspection; industry
experience with certain devices; in-house knowledge; and
criticality of failure. However, it is not easy to economically
justify relay upgrades entirely based on probability of failure and
replacement needs. Reluctance to upgrade to microprocessor
relays is further emphasized by complexity associated with
increased functionality (e.g. settings), need for firmware upgrades,
short life span of microprocessor technology, and the overall
need to change the protection system philosophy and design. In
addition, while digital relays provide wealth of data, users may
be faced with data overload. It is often the case that even data
already available are not used or even collected.
Latest technological advancements in power system protection
and control (P&C) are important enablers to meet the challenges of
the electrical grid in the 21st century. Evolutions in protection and
restoration principles for interconnected power grids are made
possible by the wide-area measurement systems [3]. Real-time
adjustment of the protection system’s security-dependability
is within reach given the advancements in technology and
investments in communication systems infrastructures.
This paper is a summary overview of initiatives on transmission
and distribution substation automation and integration with
several different parallel strategies that were harmonized in mid
1990s to a united vision for a major power company in North
America. The paper describes the process from the initial stages
of identifying the scope to strategies for justification of such major
initiatives, details of selection, application; design philosophy,
implementation strategies, partnerships and resources, and
required skill sets are described. Many of the success and lessons
learned along the way are also presented.
equipment terminals that would need to be upgraded would take
several generations to accomplish. Given the rapid developments
in technology, it would not be feasible to maintain course both
from product manufacturing support of “out dated” equipment
and from the utility perspective of not making use of latest
technological advancements and the benefits offered to the
industry as a whole.
In the early 1990’s PG&E was facing difficult economic decisions
in its transmission business. Customer unrest over increasing
prices, greater demands on limited budgets and increasing new
business needs required a new approach to asset management.
In addition, device and panel replacements were exceedingly
costly. With multiple forces pulling in different directions, and
armed with the knowledge from the gradual panel migrations,
PG&E started to study a new paradigm for budgeting limited
dollars for aging control room equipment.
At the same time the industry was quickly developing new
automated systems for control and more modern flexible
protection and multifunction capabilities which were being
introduced to the market. The possibility of the new automation 1
and protection packages offering greater reliability and customer
service led to more focused evaluations. Various options were
explored and a new vision and strategy was initiated – the
Modular Protection, Automation and Control (MPAC) concept was
selected to be studied.
2.1 Vision
The vision was to create integrated systems and automated
applications that use data from transmission and distribution
facilities to effectively process and act on information to operate
and maintain the electric transmission and distribution systems,
Figure 1.
2. Background and Reasons for
Migration
In the early 1980s, Pacific Gas and Electric Co. (PG&E) started
to look at use of numerical devices on its 230kV and below
transmission systems. The approach started with replacement
of one level of protection with microprocessor based devices
and soon expanded to move towards replacement of protective
devices on per terminal (panel) basis. Reasons for the step-bystep conversion process are attributed to a number of factors
including the need for:
•
Field experience with the then new technology
•
Development of new standards to identify full range of
compatibility with the existing conventional practices
•
Determination of the financial merits
It was then identified that replacement by panel only results in
protection equipment upgrades yet does little to modernize the
overall aspects of power system operation and that a holistic
approach was needed. In addition, the relative number of
42
Figure 1.
MPAC Vision
1Since many functions in today’s substations are automated with regards
to system faults and restoration response, many believe that substations
are already automated. Automation used in this paper is in the larger
sense. Substation automation is the use of microprocessor based
devices for metering, monitoring, controlling and protection of substation
functions. Automation allows for the sharing of data between devices over
a local area network and accessing substation information remotely.
A Paradigm Shift in Protection, Control and Substation Automation Strategy
Developing a common vision for the substation of the future was
an important first step. The vision had three key drivers:
•
•
Safety
Another key step in the development of the strategy included:
•
Cost reduction
•
•
Improved reliability
Site visits to other power companies that had considered or
had implemented similar concepts
•
Discussions with manufacturers of protective devices and
system integrators to identify their offerings of products
and systems. Although the team was interested in what the
vendors already offered, PG&E was even more interested in
where the manufacturers were headed
With these drivers in mind, three areas of focus were defined.
These areas later became the major initiatives:
•
Information technology and management
•
Communications with different groups
•
Equipment performance; and work practices
The new systems needed to respond faster, remain pricecompetitive, improve customer reliability, and do it all with fewer
resources.
2.2 Strategy
In order to determine the feasibility and to identify road maps a
set of basic strategies would be needed. The strategy was based
on some fundamental objectives - To improve safety, efficiency,
service reliability, and manage the cost of transmission and
distribution Operations, Maintenance, Planning and Engineering
through the integration and the implementation of automation
applications, Figure 2.
Build a specification for the “Substation of the Future” that will
be the standard blueprint for all projects by the year 2002
Armed with better information in terms of manufacturing and
technology as well as the direction of others in the industry, PG&E
had a clear basis for its own strategy implementation, or migration
path. The strategy would be continually updated as time passes
and the future business and technology environments reveal
themselves.
2.3 Initial Stages of Expenditures on MPAC and
Prototypes
Cost justification for comprehensive upgrade programs comes
from combination of various benefits. Those benefits do not
result only from replacing protective relays, but deploying an
integrated system using enterprise Information Technology (IT)
architecture to enable collecting required data and processing
and distributing information when and where required for use by
various applications and users.
The initial MPAC program for PG&E came out of a budget process
which focused on single year expenditures for multiple projects of
different types competing against each other for funds. With this
approach, items like “new business” and distribution lines, and other
projects serving the immediate needs of the customer generally
received the bulk of the funding, while items like transmission or
distribution automation did not attract much funding. While the
promise of automation seemed hopeful and was being promoted
within the industry, the cost/benefit case could not be made to
justify the overall scope. Where power companies and vendors
had gone forward, they did so based on other reasons such as
aging workforce, equipment obsolescence, or reliability needs.
The traditional business cases by themselves were not sufficient.
Figure 2.
Integration and Automation Objectives
The initiative would require technical and human factor
considerations and many other related aspects for system-wide
implementation which are covered later in the paper.
Some of the factors in strategy explored included:
•
As equipment wears out, replace with the new generation of
devices, systems, or schemes
•
Build new substations in the new image
•
When opportunity knocks, follow the blueprint
Management of the above three initiative areas was intertwined
in order to:
•
Continue with incremental migration but proven technologies
to achieve gradual benefits and gain practical experience
Taking a longer term prospective and a program approach, a
budget plan was prepared for other major capital projects over
a five year timeframe; i.e. wood pole replacements. This budget
planning approach provided funding opportunity to focus on grid
infrastructure revitalization over time.
With funding realignments, multiple efforts had to come together
to make the MPAC concept a success. From a strict economic
perspective (one of the realities of budgeting in a large corporation
with competing priorities), there was not much incentive to replace
working electromechanical or solid state devices with newer
and more feature enhanced devices. However, with a program
approach to the larger needs of the substation, and an aging
infrastructure in existing control rooms, the need for determining
the overall reliability performance measures, and managing
infrastructure assets the idea caught hold. By combining
automation, aging equipment, and integration of functions using
the new microprocessor capabilities, a case could be made to
replace the equipment.
A Paradigm Shift in Protection, Control and Substation Automation Strategy
43
At the same time, many technical details needed to be worked
out. Development of the details such as the system architecture,
levels of integration, harmonizing PG&E’s protective relay
selection and application philosophy, automation and SCADA /
HMI (Human Machine Interface) hardware and practices, and
other requirements led to the concept of pilot projects in the mid
1990s.
The first pilot involved installation in an existing building of new
racks, equipped with then latest numerical devices. The cost,
however, to put in new racks in an existing control room was
extremely high. Again, the economics would not support the
approach. Attention was then turned to a more modular approach
of replacing the whole control room.
The first pilot project involving a completely new control building
showed promise from an economic cost benefit standpoint. In fact,
when compared to a more traditional PG&E approach of building
a custom building, a savings of 20-25% seemed achievable.
This pilot project then became the turning point for justification
of future modular buildings (MPAC). However, in order to have a
systemwide program, many factors had to be considered and
processes established or enhanced. For example, standardization
of equipment, levels of integration, establishment of policies for
network and information technology (IT) infrastructure expansions,
development of engineering standard schematics and setting
templates, automation diagrams and standards, and preparation
of test programs namely, Factory Acceptance tests (FAT) and Site
Acceptance Testing (SAT) for control building. Equally important
in the design and engineering development was maintenance
testing over the project life cycle that had to be incorporated in
the new paradigm.
By design, the first automation pilot project was somewhat limited
in functionality. The primary goal was to achieve small advances
with complete certainty and learn from the experience, rather
than try to solve all automation problems in a single stroke. Thus,
attention focused on establishing connectivity through modems
and collecting information, as well as to provide means for basic
SCADA functions. This project was completed successfully, and
has been performing well. The primary concerns were with the
accuracy of the data and false signals that were resolved after
further equipment calibrations.
Some information provided by the first pilot project demonstrated
the capability of new devices to provide information which
surpasses the typical data available from traditional SCADA. The
new system collected load information from redundant protective
devices which allowed for telemetry validity check. In traditional
SCADA, one set of transducers collected load information, and if
the transducers drifted there was no built-in check to diagnose
the problem. The newly installed devices also have the potential
to monitor breaker wear for example. Breaker operations, as
recorded by the relays, were not being entered directly into the
reliability calculations at this stage pending development of
databases.
The second pilot project was more ambitious and included
building a new control room and installing a completely new
integrated information and control system. The plan included
building a completely modular control room, built off-site and
delivered ready for operation. Traditional RTU (Remote Terminal
Unit) systems and numerical multi-function devices were used for
SCADA and for Protection & Control respectively. The idea included
44
minimizing, and if possible eliminating, traditional control switches
and providing a Human Machine Interface (HMI) for operational
needs.
3. Integration and Reliability
Considerations
While the modularization concept was being explored, the
evaluations in technology associated with numerical devices were
being conducted by another group at PG&E. The System Protection
and Controls Department was exploring the ideas of integrating
many protection functions and automatic reclosing features. The
ideas ranged from continuing with gradual migration of features
to delivering significant improvements in dependability, security,
and greature reduction in hardware. Several factors needed to be
examined including:
•
The architecture of integration.
•
Level of integration and associated complexities. For
example, would it be acceptable to integrate line protection,
breaker failure detection, stub bus protection, and automatic
reclosing into one device?
•
The capacity and processing speed of the hardware; e.g.
could a device perform its primary protection function while
it is also providing SCADA or HMI functions?
•
System reliability, will it be compromised compared to
traditional methods, and to what degree if any?
Since similar discussions were starting to take place in other
electric power companies, the approach considered was to discuss
the integration ideas and concepts with technical committees at
the Regional Council meetings. Exchange of ideas and reaching
for common grounds provided further support to PG&E.
4. Protection Reliability, Hidden Failures
& Unintended Operations Detection
Several key benefits for migrating to the new paradigm
are highlighted in other publications [4], [5]. Technological
advancements in modern protective devices, user selectable
supervisory elements, and monitoring features in modern devices
further supported the justification for proceeding with the
development of the blueprint.
4.1 Protection Reliability
The availability for proper operation of protective relays for
various system conditions is one of the most important factors
in improving system reliability. Detecting hidden failures is a
critical task that requires good understanding of the principles of
operation of the protective devices, their self-checking functions
and their limitations, as well as properly defined methods and
procedures.
Security and dependability are two measures of reliability.
Dependability is defined as the “Degree of certainty that a relay
system will operate correctly”, and Security: “Degree of certainty
that a relay will not operate incorrectly”, Figure 3.
A Paradigm Shift in Protection, Control and Substation Automation Strategy
NO-Operation
Operation
•
Equipment condition data that would help in asset
maintenance savings and prioritization of capital and O&M
investments as a part of the overall asset management
strategy.
•
Improved network security, as well as replacement energy and
congestion savings. The latter two are results of minimizing
equipment loss and, consequently, avoiding both a need to
replace energy on the spot market and congestion.
•
Enhanced system restoration.
•
Real-time system prognosis and capability for adaptive
protection and control.
•
Reduce down time.
•
Easily adaptable to advancing technology – Ease of upgrade
without impact to wiring.
•
Information as opposed to data rapidly available – Helps in
information broadcasting, automatic identification of trouble
spots, and client / customer satisfaction.
•
Cost savings in implementing condition based maintenance
- Improved asset management decisions based on solid
statistical data.
•
Identifying weak spots in the system.
•
Reliability improvements resulting from avoided failures due
to continuous monitoring and better T&D planning using
statistical data.
•
Cost savings in crew overtime.
•
Productivity improvements through integration of IT systems
and databases.
•
Better inventory control.
Reliable
Correct
Secure
Dependable
Unreliable
Incorrect
Undependable
Insecure
Figure 3.
Definition of Protection Reliability [4]
4.2 Monitoring the Protective System Identification of Hidden Failures
A failure of a protective device may be caused by many different
factors, including not only failure of the device itself, but also of
components of the overall substation protection, control and
monitoring system.
5. Integration of Protection and Control
for Improved Asset Management
The key to the asset management concept is three fold:
•
Replace the entire control building – replacing components
and racks is too costly and operationally disruptive.
•
Secure a program approach – develop a long term plan with
multi-year funding.
•
Standardization – Treat the whole control room like a
commodity, not a custom application. Develop a standard
design, procurement, and specification for the entire building
including the protection, integration, and automation
components to be able to bid out the specification to selected
venders.
The need to justify the preferred and perceived most cost effective
option resulted in preparing a comprehensive listing of benefits for
the standardization which included:
The comprehensive approach to standardization provided yet
another key pillar towards integration.
•
Utilizing the features of new devices and systems to reduce
the amount of equipment, floor and panel space, and
hardware failures while improving protection performance.
•
Reduced installation and maintenance cost for protection and
control equipment. The ultimate goal should be to eliminate
time-based maintenance while improving availability,
security, and dependability. For example, manual relay
testing could be significantly reduced.
•
Improved dependability and security, while drastically
reducing the number of units required, and complying with or
exceeding all requirements by the regulators.
•
Large scale strategy for systemwide roll out
•
•
Reduced operating costs. No or minimal human intervention
is needed to rapidly determine status, event triggers, event
reports, equipment performance, etc. Settings could be
securely managed remotely using well-defined procedures.
Development of system architecture and roadmaps for
testing and training new users
•
Development of Engineering design standards and supporting
documents such as logic drawings, test procedures, and
recommended maintenance steps
Accurate and timely fault location to the dispatch personnel.
It could also help with identifying the cause of both permanent
and momentary faults, weak spots in the network, and aid in
calculating tower footing resistance, and system parameters
such as the line impedance [7].
Before the system architecture could be developed, however,
other requirements had to be met. The equipment needed to be
evaluated for flexibility and performance in meeting the majority
of PG&E integration and automation applications. Different bus
•
6. Roadmap to the Development of
Architecture
The experiences gained from the trial projects provided many
insights and the lessons learned could be transferred to other
projects. The trial projects also shed light on some of the absolute
requirements in order to launch a full scale implementation. Some
of the requirements included:
A Paradigm Shift in Protection, Control and Substation Automation Strategy
45
configurations such as breaker and half, (BAAH), Ring, or Double
Bus Single Breaker, would have different interlocking, control, and
/ or automation requirements. Also, automation requirements had
to be defined and different substation automation hardware had
to be evaluated for best flexibility, performance, and overall cost.
6.1 MPAC Core Team
In order for the vision to come to fruition, a cross section of
expertise and interfaces with various departments was formed.
The MPAC core group members had to have vision, subject matter
expertise, and the dedication needed to make this company
initiative a success. These individuals had to think outside the
box, foster a paradigm shift, and come up with a plan that was
innovative, cost effective, reliable, and practical. The complexity of
the task required involvement from many affected departments.
Every group (Department) had to have ownership and contribute
in the development of the new plan. Each subject matter
expert (SME) was responsible for representing their respective
work group and providing the expertise from the respective
expertise. Therefore, each core team member was charged with
communicating milestones and acting as a conduit for feedback
to the MPAC core team.
6.2 Need for Partnership
Initial discussions focused on partnership with protection and
control device manufacturers and the possible need for an
integrating company, Figure 4.
•
Training of partners on each other products or standards
•
Better utilization of resources and resource sharing in peak
periods
•
Project Management support
•
Streamlined procurement process namely, request for
information (RFI) and request for proposal (RFP)
6.3 RFI Process for System Architecture
In order to avoid the need to obtain completely new and separate
bids every time, the PG&E MPAC team wanted to establish
standardized pricing as much as possible before projects were
initiated or procured. Bidders were required to propose pricing
that would remain effective for the respective year subject to
certain project defined modifications which may alter the final
price of a given project. The team also included an evaluation
process as part of the performance of the bidders for the year that
a project would be granted. The idea was focused on partnership
as opposed to a pre-bid evaluation based on price. To facilitate the
concerns and risks associated with initial configurations “targeted
price” proposals were accepted from the bidders as to their
recommendations and technical requirements for partnership.
Also, incentives were put in place for sharing the savings below
the “targeted price”.
7. Integration and Object Model
Development
One of the initial tasks undertaken by the core team was the
development of the major requirements such as:
•
Should integration be limited to consolidation of protection
features and components only? Should stub bus, breaker
failure, automatic reclosing with voltage supervision, etc. be
all integrated?
•
Could PG&E policy in the application of two levels of protection
with no potential for common mode failures in products,
scheme performance, or associated interlocking with
high-speed or slow speed reclosing be maintained, without
sacrificing reliability and security?
Figure 4.
•
Should metering and operational switches be integrated?
Would local / remote function be integrated and how?
Under this scenario, each component of the model would
contribute an equal amount of knowledge, skill sets, processes,
and development capability to the project as needed, Figure 4.
•
Should the auxiliary tripping relays be integrated or
maintained? Would the latching function during power loss
be available if the function is to be integrated?
Some of the key benefits of the partnership included:
•
How would the MPAC station interface with remote terminals
that would be not be immediately scheduled for conversion
to integrated protection and control (IPAC)?
•
How to maintain control functions similar to the existing
practice in functionality and in “Look and Feel” when the
control function is to be exercised from the front of the panel?
– e.g.: One control switch location, one switch to control circuit
breakers, one switch to control group settings, etc.
•
How to maintain visual information on the front of the panel
to complement any computerized HMI.
Partnership Model Considerations
46
•
Cost reduction and revenue sharing
•
Product and technology development
•
Efficiency and process improvement
•
Standardization to reduce costs
•
Knowledge sharing - Industry experience sharing
•
Quicker turnaround
•
Ability to influence suppliers
A Paradigm Shift in Protection, Control and Substation Automation Strategy
•
How to maintain “multi-directionality” of controls and
information between local panel information, the HMI at the
station, and the SCADA system at the control headquarters.
•
How would the MPAC station interface within an existing
established substation without significantly increasing the
project scope? For example, a two winding transmission
transformer bank (i.e.; 230/115kV) with two different MPAC
buildings for the respective bus voltages? If the MPAC was to
be installed at the 230kV site, and the 115kV control building
should not be required to become MPAC.
•
Interface with neighboring companies that do not have MPAC
substations or may opt to stay with their own standards.
•
Would SCADA be incorporated as part of consolidation and
integration of protection and control components?
Figure 6.
•
Would SCADA be a distributed SCADA or a centralized system
per station or per control building?
One of the initial tasks was the development of engineering
and protection design standards. They also recognized that the
best way to succeed would be to implement the development
of standards one item at a time while considering the various
elements of the object model.
Given the many challenges to face, and armed with the in-depth
knowledge of industry and internal practices, the development of
an object model was conceived to better visualize the interactions
of the various functions, Figure 5.
Protection Object Model
Figure 6 shows a simplified Protection object model for
components that may exist in a substation. Figure 6 also shows
that the initial task was development of integrated standards for
“line Protection”. As a component of the integrated system was
developed, it would be rolled out for implementation while other
components were being developed [5]. This concept of phased roll
out also allowed participation by a larger group of field personnel
or engineers as projects would start to be designed and fabricated
based on the new standards. Appropriate suggestions would be
incorporated both in the components already developed and also
for the components that were being developed.
Figure 5.
IED Object Mode
Use of the object model would allow consistency amongst
applications, suppliers, and partners. The high level object model
included components such as:
•
Logical Interface
•
Physical Interface
•
User Interface
•
CT/PT Source interface
•
Interface with non-MPAC
•
Reclosing – Both raid or high speed and time reclosing
Throughout the development, the standard / core team maintained
focus on design safety, reliability performance, internal practices,
and the overall cost.
The core team also recognized the human factors such as need
for knowledge transfer, training, and other related aspects for
systemwide implementation and strategies that needed to be
developed in parallel with the technical discussions.
Figure 7.
Simplified Object Model for Reclosing (both rapid and time reclosing)
Details of line protection object model and various types of
protection applications including direct transfer trip and third party
interconnections were then developed [5]. Other object models
such as Automatic reclosing and Human Machine Interface (HMI)
were also important in design of standards. Automatic reclosing
encompassed rapid reclosing (both with or without synchronization
A Paradigm Shift in Protection, Control and Substation Automation Strategy
47
supervision to meet existing practices for different applications)
and time reclosing and associated interlocking for situations that
required a lockout condition, Figure 7.
For the HMI development, the challenges varied between different
environments such as a brand new building (MPAC), or when
multiple buildings are needed due to station or bus configuration,
transitional projects in existing substations, and interactions with
SCADA, Figure 8.
Another key element in the design was maintaining reliability
in control functions. In the traditional control, highly reliable
control switches were used for controlling protection and also
elements. For example, only one setting group selector switch
would control the desired setting group for both levels of a line
protection. The new integrated design needed to maintain
conventional operational control concept to one device (one
virtual switch) while maintaining the reliability and operational
flexibility (“independence”) when the device assigned to perform
the group setting switch function was removed or became
unavailable. Also, the position of the control switch was visible in
the traditional applications, where control was performed locally.
These considerations led to the use of LED indications to reflect the
position of equivalent “Switch” functions. Also, the requirements
for reliability and availability for metering and control functions
drove the decision for distributed SCADA / computerized HMI
functions. Since in PG&E’s practice each piece of equipment has
redundant multifunction devices used for protection functions, it
would be possible to use the integrated concept for protection and
reclosing as well as for traditional physical controls and maintain
functional dependability for operational / switching needs. In the
new MPAC design, the traditional station metering and station
control switch function was assigned to one level of multifunction
devices (Set “A” protection and control), and the second level
multifunction protection and control device (Set “B”) was used as
the gateway for HMI / SCADA.
respective breaker. Likewise, the synchroscope display on the
HMI is designed to reflect the voltage vectors from each side of
a circuit breaker. When the vectors are in the same quadrant
and approaching each other, the operator can click on the close
command and the device synchronizer function will close the
circuit breaker when appropriate.
8. Development Roadmap and
Standards
The conventional protection and control schemes used many
discrete components and auxiliary devices. With the additional
capability of multi-function devices it was be possible to minimize
use of auxiliary equipment and include more of the features within
the devices. The new microprocessor devices contain more than
one protection scheme or feature. For example, backup, breaker
failure, reclosing, and stub bus protection.
Another major factor to improve overall performance and system
reliability was the goal to standardize on the design, while reducing
overall project cost over the life cycle. This concept required
standardizing on the size and type of building, type of equipment
installed within the building, and the various schemes.
The success of the concept would depend on standardizing on
many key factors such as:
•
The buildings
•
Protective schemes
•
Switchboard layout
•
Engineering
•
Training
•
Procurement process
•
Test and commissioning
Standardization would also lead to cost stabilization and
reduction while improving system and service reliability. This
concept led to forming partnerships with some manufacturers.
At the same time it was very important that the performance
and functionalities would not significantly change. Many of the
traditional functionalities have been enhanced based on many
years of practical experience and lessons learned and have been
reflected as part of fundamental practices.
Simplified Object Model for HMI
Development of such new standards required resources with
knowledge of internal standards, understanding the reasons for
previous practices, expertise in industry best practices, and vision
to develop a new generation of standards.
Also, all timed based automatic reclosing functions have been
programmed into the Set “A” device. Other functions have been
designed to be in a combination of both Set “A” and Set “B” devices.
For example, a breaker can be manually closed from the front of
the panel or via the station HMI. In the case of the front panel, the
Set “A” LED indications and the Set “A” device display have been
designed (programmed) to inform the operator of the voltage,
frequency, angle, and direction of voltages on the two sides of
the breaker, as well as a dedicated LED to inform the operator of
the closing angle being in the acceptable range for closing the
Process improvement was incorporated as part of the initial phases
of the vision development and roll-outs which allowed the core
team along with the respective building manufacturer to critique
the project and go over the successes and failures [9]. From the
discussions, the processes could be improved and further cost
reductions could be captured by both the builder and the company
without sacrificing performance, safety, or reliability. Then, the
core team’s job was to discuss the various issues and find ways to
improve the process itself. For example, the period from 2001 to
2003 was used to firm up the design and the supporting standards
for the MPAC buildings and systems. While the development of
Figure 8.
48
A Paradigm Shift in Protection, Control and Substation Automation Strategy
the interlocking logics between various elements of power system
equipment moved forward, some elements such as the interface
with the devices and the HMI architecture needed more time to be
reviewed and options examined.
•
First the SCADA program used for the HMI and at the switching
station was revised to provide the capability of DNP master
protocol. This change provided the capability to the SCADA
software to poll the IED devices directly.
The Remote Terminal Unit (RTU) was initially replaced with a Data
Concentrator that would offer the following:
•
The second change was the development of a secure
isolated wide area network with dedicated routers. This
new WAN is used solely for SCADA and operational data
and the new routers were selected to provide security and
user authentication. The additional new security features
provided by the new WAN routers eliminated the requirement
for providing access and security functions within the Data
concentrator.
•
A secure network interface for remote access by engineers
and others
•
Survive a rugged substation environment
•
An embedded processor and no hard disk
•
Communicate with one device using the device native
protocol (serial)
•
Communicate with the second level device using Modbus
TCP/IP protocol (LAN)
•
Communicate with the Substation HMI using in-house (PG&E)
protocol
After 2003, the Substation HMI software was converted from the
third party supplier to the software being used at the switching
centers. This approach offered consistency between the displays
used for operating the systems, and would eliminate the need to
develop and maintain two different databases. The database and
displays developed for the substation could be loaded directly into
the switching center systems. In this architecture, the substation
HMI would poll the Data Concentrator using DNP3 protocol and
would communicate to the switching center using a TCP/IP
protocol with three levels of security.
A system architectural redesign was later implemented to have the
substation HMI server communicate directly to the multifunction
devices. This alternative became available due to following two
changes in the system architecture.
The SCADA system direct IED poll concept was successfully tested
in the Engineering Laboratory. A second full scale test was also
run on a large MPAC building that was in a PG&E substation but
not connected to the power grid equipment at the time of the tests.
This test was also successful so the future designs will no longer
needed the data concentrator. Furthermore, the new architecture
reduced the site acceptance time from three weeks to less than
one week.
8.1 Integrated Data Management
Figure 9 shows a hierarchical information routing architecture
where the devices (redundant or non-redundant) are using
different protocols and communication at the substation level.
For example, Set “A” and Set “B” protection and integration devices
are using different communication cables and protocols. Likewise,
redundant System Integrity Protection (SIPS) devices “A” and “B”
have different topology in network connection.
This concept includes significant use of messaging (IEC 61850 for
protection and control) that have been applied to both transmission
and distribution switchgear. At the substation level today, the IEC
61850 is used for protection, control and interlocking [10]. A next
step at the Enterprise level would be implementation of IEC 61850
to migrate from the DNP protocol.
Figure 9.
Substation Network Showing Different Types of Equipment with Different Network Protocol
A Paradigm Shift in Protection, Control and Substation Automation Strategy
49
Similar to the protection and integration standardization efforts,
the next steps include development of standards for use of IEC
61850 for Automation and identification of migration path to a
complete communication based systems between the substation
and the Grid Control both at the transmission level for real-time
applications at transmission (i.e., link to EMS) or distribution
system. Synchronized measured values are also available at the
substation level and are part of the architecture of the future for
real-time measurements (as opposed to state estimation).
The Ethernet switch selection and standardization is also
important. In the era of rapidly advancing technology, clear
roadmaps should be developed to allow migration and gradual
transition while operating the grid over the life cycle. Equipment
tracking mechanisms should be applied for ease of identifying
brand, equipment location (substation), technology applied (i.e: 10
mbps vs. 100 mbps Ethernet switch).
9. Application of IEC 61850 for
Protection
One of the elements of the object model included the development
of standard applications for protection and control using UCA /
IEC 61850. Development of this object model requires listing
of benefits, identifying adverse functional impacts, training of
resources, and cost to name a few of the considerations.
For SCADA Data
-
Superior Asset Management options
-
Condition monitoring of primary equipment
-
Power Quality Information
Other large scale benefits include:
•
Exchange of Synchronized phasors data between PMU
devices via GOOSE which would better enable applications
such as wide-area out-of-step.
•
Potential for bringing synchronized phasor data via GOOSE to
phasor data concentrator (PDC)
•
Ease of use with optical sensor and merging unit technology
9.2 Challenges of Application IEC 61850 for
Protection
Below is a listing of some of the challenges in implementing the
IEC 61850 GOOSE for protection and control:
•
Adequate training of the protection engineers
•
Adoption to existing substation automation concepts and
changes in the specification and design process
•
Conformance Certification of Devices to IEC 61850
•
Tool for variety of purposes, for example:
-
Advance application development
-
Documentation of engineering design – Substitutes for
Wiring and Schematics
9.1 Benefits of IEC 61850 for Protection
Application
-
Mapping
-
Performance tracking
Benefits of the IEC 61850 for protection applications are described
in [5]. Highlights include Design, Monitoring, Application, Asset
management, etc. Below are some of the benefits in using IEC
61850 for protection interlocking and automation:
-
Configuration and testing
-
Troubleshooting and Maintenance
•
•
Self Monitoring and Alarm – Continuous real-time status of
control point communications from the source device to the
implementing / receiving device for failure anywhere in the
circuitry.
The protection engineer will need to develop some basic
understanding of the:
-
Engineering approach with the use of the configuration
language
-
Concepts of the object models and the basic
communication services
•
Improved flexibility in design
-
Ethernet technology with switches and priority tagging
•
Ease of applications when I/Os are limited - Minimizes use
of auxiliary devices – Less hardware, less components, less
monitoring, more reliability.
9.3 Practical Examples of IEC 61850 GOOSE for
Protection and Control Interlocking
•
Prevents potential for mixing DC circuits, for example, when
inputs or outputs are limited.
•
Reduced implementation and testing time and cost
•
Programmable timing and loss rate monitoring
Figure 10 shows a simple interlocking example used for automatic
reclosing in a breaker and half bus configuration. In this example,
each breaker has a multi-function breaker control and reclosing
device. Each breaker can be independently selected in advance,
by the operator, to be the designated breaker for line reclosing, or
automatic transformer restoration, when power system conditions
would allow the breaker to close [5]. Other examples include
breaker failure interlocking, breaker maintenance interlocking with
bus differential and middle breaker in the respective bay, etc.
Furthermore, deployment of this object model would also require
comprehensive and fully developed setting templates, logic
drawings, and tools that would allow protection and automation
engineers to assist technicians and control building contractors
during commission testing, FAT and SAT.
Data sharing benefits for automation, asset tracking, and
maintenance include:
•
50
•
Remote system monitoring
A Paradigm Shift in Protection, Control and Substation Automation Strategy
9.5 Architecture Adaptability to IEC 61850
Process Bus and Merging Unit
The modularity of the MPAC provides for opportunities to enhance
the architecture through the adoption of new technology. Similar
to other aspects of MPAC, the adoption of process bus is evaluated
to determine measurable and quantifiable overall benefits. Many
of the benefits are covered in earlier sections in this paper. Areas
for improvement in using the process bus include:
•
Reduced engineering
•
Simplification of electrical drawings – Some cabling diagrams
will need to be incorporated into the new design concepts
•
Minimization of copper wiring
•
Turn around time of installation and commissioning
•
Reduced maintenance cost
Interlocking Example using IEC-61850
•
Overall improvement in availability
Transformer protection interlocking between the high and low
side windings is another example where the devices on the high
side winding maybe in a different control building than the low
side winding devices. The buildings may, in some cases, be at
considerable distances (hundreds of feet) away from one another.
Application of the IEC 61850 makes the task of designing the
interlocking simple. It is important to properly address allocation
of network switches in multi-building stations where GOOSE
messaging is communicated between buildings. For example,
when protective devices for the high and low voltage windings of
a power transformer are in two different buildings and IEC 61850
GOOSE messaging is used for interlocking functions between the
two buildings. A simple solution is a fiber connection between
the network switches in the two protection and control buildings.
Other options include use of custom programmed network routers
to pass V-LAN Multicast traffic if the two buildings for some reason
(e.g: Cyber Asset allocations in each building) cannot be on the
same network.
The IEC 61850 Process Bus communication architecture and a “fit
for purpose” physical architecture can be advantaged to achieve
these goals.
Figure 10.
Another key advantage of the IEC 61850 application for
transformer protection interlocking is minimizing use of DC circuits
and running DC circuits between buildings when battery sources
between the high and low side windings are not the same and
protection DC sources cannot be mixed.
Other examples include breaker failure interlocking, breaker
maintenance interlocking with bus differential and middle breaker
in the respective bay, etc.
9.4 Design Considerations
In all of the above examples, it is recommended that the user
applications have a “fail-safe” mode of operation. In each case,
the use should consider factors such as:
•
Loss of source (i.e.: DC circuit) to the devices that are
communicating GOOSE messages?
•
Need for redundancy in communication equipment
•
Logic design and considerations – Example, upon DC source
loss, scheme performance should not be any different than
the conventional scheme.
The IEC 61850 Process Bus communication architecture provides
a standard configuration for communication messages from a
“Merging Unit (MU)” located in the field to a numerical relay located
in the MPAC building. The MU can be locally connected to monitor
CT, PT, Status Inputs, and Transducer Inputs. The MU can be
connected to operate process equipment (breakers, switches, tap
changers) in the field. From a physical architecture perspective,
the use of fiber optic cabling between the MU and the relays in the
MPAC building is the logical choice. A Relay-to-MU architecture
further simplifies the installation from several perspectives:
•
Cabling follows the existing wiring troughs
•
Precise MU Synchronization can be achieved through the
communication channel – eliminating the need for master
clock synchronization
•
Elimination of Ethernet switches in the path of this critical
cyber asset
•
Minimization of wiring connections and the associated
engineering, drawings, installation, and testing
•
Ease of implementation of redundant MUs
The Process Bus architecture allows connection of the MU
to process equipment as early as the factory manufacturing
stage. Implementation at this early stage allows for controlled,
repeatable, and factory-tested installations. Connection of the
MU to a relay in the control house requires the addition of a single
fiber jumper between the MU data source and the relay input data
card. A connection table now replaces wiring diagrams. For those
measurements where field installation of the MU is required, the
MU can be located anywhere in the yard and wiring is simplified to
a point-to-point exercise.
The choice of modular relays in the MPAC building has made
the adaptation to process bus straightforward. One Process
Bus interface card in each of the MPAC devices (system “A” or
“B” devices) replaces the respective analog, status input, the
transducer and output control cards.
A Paradigm Shift in Protection, Control and Substation Automation Strategy
51
In terms of device set points and already established object
models, all functions remain unchanged in the MPAC building. For
example, the pushbutton operation, LED mapping, interlocking
functions, and logic settings remain unchanged. The only
additional settings are those that map the data from the MU into
the appropriate locations in the relay (process bus interface card).
Note that the functions defined in the IED Object Model remain the
same – the only difference being that the data acquisition function
in the MPAC devices is transferred to the MUs in the field.
9.6 Training Philosophy
With the new paradigm shift, training the employees has taken
added importance, Figures 11-13. Training is a key component
in a major initiative as every aspect of design, engineering,
construction, maintenance and operation is affected.
While the basic control room appeared totally different, the
operating principles remained the same as the older control rooms.
Many of the functionalities from the conventional control room
were incorporated into the multi-function devices thus, giving the
new control rooms a different visual appeal. The challenge was
how to train all the employees so they felt comfortable operating
the new technology.
Figure 13.
MPAC Building – Minimized Wiring
The initial series of trainings sessions were conducted on-site
and consisted of hands-on training, just before the new buildings
went into service. The training manual contained the description
of operation, interaction with the HMI, Figure 14, and the multifunction devices [5].
9.7 Training Development and Technology Lab
The next step in the development of training involved a more
systematic approach for a centralized training facility with formal
classes and certified instructors that would provide training for
different types and classifications such as engineers, technicians,
and operating personnel. This facility would also be designed
such that further technology advancements could be evaluated
for systemwide roll out subject to successful testing. For example,
the synchronized phasor functionality was recently implemented
in the technology lab both for advance application studies and
also for training and education of resources. The training and
technology lab concept was studied and recommended to the
management along with the justification for such a facility.
Figure 11.
MPAC Building in Route to Station
Figure 14.
Typical MPAC Building Showing the HMI
Figure 12.
MPAC Building with Integrated Protection and Control
52
A Paradigm Shift in Protection, Control and Substation Automation Strategy
In parallel with the building of the simulator, a training matrix was
developed in order to deliver specific training modules to specific
groups based on the needs. Table 1, shows a simplified version of
the training modules for different functions. The intent was to
develop content such that each stand alone training module could
be presented as an individual class or given to certain groups of
employees that required the training. With the training modules in
place, types of tools and course contents were developed. For
example, the “troubleshooting logic” module required both hands
on and classroom type material. It was also anticipated that as
more MPAC buildings are constructed, the level of training for
different groups or individuals may vary and therefore the training
modules would need to be flexible to be enhanced as required.
One of the main differences and challenges with the use of
multifunction devices has been the scheme design, functionalities,
and interlocking embedded in the form of logic. Since the standards
and associated logic developments have been with the core
standards team, many aspects of training and troubleshooting
have been considered and developed during the initial design
phases. For training purposes, the logic related modules
concentrated on the new style of prints, logic symbols, and how
the logic interacted between the various discreet components.
This was a significant change compared with the conventional
standard drawings which normally contained device contacts
and auxiliary devices. The new designs incorporated many of the
functionalities of the inputs and outputs so the logic diagrams
were required to show the logic flow and the relationship between
the contacts and logic.
Another training module, the HMI module, was developed with
input from operating personnel, electricians, and first responders.
This module, as all the modules, contained situational problems
that had actual problems to solve during the training session.
The training modules also include examples of conventional and
the parallel functionalities based on the integrated schemes for
the respective disciplines. These training modules, along with
the other modules, provided the students with various hands-on
scenarios, training text books, and job aids. The various modules
would provide the company course flexibility when training the
employees. The development of the training modules utilized a
subject matter expert along with the training group. Standardized
designs implemented in the field allowed the company to
standardize on the training thus, capturing training costs and
efficiencies.
Module
10. Testing and Scheduled Maintenance
With the development of a partnership with the manufacturers,
one of the cost savings involved testing the components within
the buildings, including the multi-function devices. The core team
worked with the respective manufactures of the multi-function
devices and agreed to standard sets of tests based on the agreed
set points. The set points were developed based on the function
of the device. For example, a transformer device had a series
of set points which are by function different than a line current
differential device. While testing conventional protective relays
was a normal procedure at PG&E, having the manufacturers of
the respective devices test their various discreet components
within the MPAC building was a new concept.
Previously, the company would receive a new protective relay
from the manufacturer and spend the time to test every element
within the relay, proving the manufacturer specifications. Today,
the company’s practice is to accept the manufacturer’s testing
specifications and then only perform a functional test once the
microprocessor devices are installed in the scheme. One way
to capture efficiencies with the MPAC building was to have the
manufacturer perform some testing before shipping the building
on-site. Standardized testing was part of the requirement to save
costs and efficiencies.
10.1 Factory Acceptance Testing
The Factory Acceptance Testing has improved over a period of
time to capture efficiencies. During construction of the initial MPAC
buildings, representatives from PG&E participated in the Factory
Acceptance Testing (FAT) to assist and witness testing. FAT is a
series of step-by-step tests developed in advance and intended
to verify wiring, interlocking, the logic and functionality at the time
of building construction. The main purpose is to ensure the wiring
is correct between the discrete components and that the devices
operate as designed. A representative from the manufacturer
would communicate any problems or needed enhancements with
the company representative.
After the initial 3-4 projects, the PG&E core team did not attend
the FAT. Only the PG&E local inspectors participate in the building
inspection. The FAT results are submitted at the time of building
delivery on site.
Maintenance
Operations
Automation
Electrical Technicians
Engineering
X
X
X
X
X
X
Logic Diagrams
Building overview
Building Systems
X
X
Controls
X
X
X
X
Safety
X
X
X
X
Testing
X
X
X
Data & Communications
X
X
X
X
HMI
X
X
X
X
X
Table 1.
Sample Training Matrix
A Paradigm Shift in Protection, Control and Substation Automation Strategy
53
10.2 Site Acceptance Testing
12. Cycle Time and Measured Savings
The FAT also led to another standardized test procedure for when
the building was delivered and installed on site. Site Acceptance
Test (SAT) are intended to prove the functionality of the schemes
and the interaction among all the discrete components prior to
release of the building to operations. This procedure proved all
the relay logic without having the electrical technician follow
the actual logic diagrams. The SAT is preformed by company
technicians and start-up engineers.
To date, there are nearly 35 MPAC buildings in service and cycle
time and costs are continually measured and compared to the
conventional approaches. Project cycle time and the cost of the few
initial projects as anticipated were near the cost of conventional
projects until the complete process leading to product levels were
put in place and various functional groups and team members
became familiar with developments and implementation
process. The cost for the learning curve (including acceptance
of the concept) is measured and for the most part seems evenly
distributed across the various organizational teams.
The standardization of hardware, factory configuration of
equipment, direct interface HMI server to the multi-function
devices, and standardized test methods also reduce the SAT
duration by nearly 60%.
10.3 Routine Maintenance and Troubleshooting
Another consideration in the MPAC development roadmap was
the maintenance testing and troubleshooting. The SAT process
is designed and developed based on this particular need when
maintenance testing would be needed or required due to
regulatory requirements for example. Troubleshooting tools and
product training have been incorporated as part of the alliance and
procurement. Also, the use of the new standard test procedures
will also help the company capture labor costs and efficiencies.
The next step for the company in multi-function device testing is
to move towards a more condition based maintenance system.
Advances in technology, uses of a common database between
departments, and the information collected from the various
components within the new buildings allows this next step to
take place. These future steps will further assist in improving
maintenance effectiveness and efficiency.
11. Additional Considerations - System
Security at Corporate Level
The initial plan included one general agreement and requirement.
If the system could not be secure, the plan cannot move forward.
It was paramount that the system could not be hacked or
compromised or not be accessed by unauthorized personnel.
The architecture, by virtue of the design approach, had many
security levels and devices had multiple levels of alarms and
password controls. However, there was still the remote possibility
for someone knowledgeable to gain access into the system.
The design team looked to the Corporate Computer Systems
Department for direction, but few existing answers were available.
Up to that point it was primarily security through obscurity. Even
existing communications to the substations had security issues.
A totally new approach had to be developed. The team focus
shifted to the information technology and communications
department for developing a physically separate LAN / WAN
system for Substation Operations. This added to the complexity
of the overall approach, but with Homeland Security issues on the
forefront, greater support for the overall design was forthcoming.
Both physical and electronic security would be required for future
control buildings. Could we make the new control buildings more
secure through card swipe locks and video monitoring? This also
added to the specifications, communication capabilities and space
requirements putting pressure on valuable real estate within the
building. Finally a physically separate LAN / WAN architecture was
developed and implemented meeting both corporate requirements
and new emerging national requirements.
54
The costs for developing the standards, setting guidelines, and
templates, and engineering / maintenance tools, is a one time cost
and can be distributed amongst the many projects. The overall
project costs beyond the few initial projects has been steadily
declining – The “targeted” vs. “budgeted” costs of the projects
completed to date show a steady reduction in the 20% range.
The contingency funding allows for addressing complexities. For
example, if a station involves design and implementation of System
Integrity Protection Scheme (SIPS) or Remedial Action scheme.
13. Conclusion
This paper is the roadmap of the success and lessons learned
toward a systemwide revitalization of the assets and grid
infrastructure for a major power company. The paper concludes
with some measured savings based on installed systems. The
initiative started with a clear vision of the future and putting a long
term strategy in place. A set of trial migration sites were used
as the initial steps followed by the formation of a core team and
alliances with some protective relay and automation equipment
suppliers. With the core team in place and the development of the
object models, a “proof of concept” lab facility was established at
the engineering headquarters to refine the objectives, develop the
details for the architecture, troubleshoot the concepts, evaluate
overall performance, and to identify the requirements and
challenges. At the heart of the project was the commitment to
develop a comprehensive set of standards, tools such as setting
templates and logic drawings for various control features and
associated interlocking with protection elements, comprehensive
training program, establishment of asset maintenance process,
and the description of operations for various components of
modular building that led to the ground breaking concept for
building a technology and state-of-the-art training facilities. The
project success is credited to many hours of dedicated teamwork
and commitments by the partners to develop the required features
for implementation. The savings are approaching the original
vision and the anticipated cost reductions when the concepts for
MPAC were being developed. The vision has also laid the ground
work for the future direction once all substations are digitized. The
vision is now well accepted companywide and many companies
from around the world have come for information exchange and
site visits.
A Paradigm Shift in Protection, Control and Substation Automation Strategy
14. References
[1] V. Madani R. King, S. S. Venkata, W. Reder, The Challenges and
Opportunities to Meet the Workforce Demand in the Electric
Power and Energy Profession, IEEE PES, Munich, Germany:
November, 2007
[6] V. Madani “IEC 61850 Applications and Challenges for
Protection Engineers”, Pre-conference Day -Western
Protective Relay Conference, - October 2006
[7] V. Madani, D. Novosel, et al. “Taming the Power Grid”, IEEE
Spectrum Web. February 2005
[2] S. S. Venkata, V. Madani, et al - The Education and Training of
Future Protection Engineers: Challenges, Opportunities and
Solutions, IEEE Power System Relaying Committee, May 2006
[8] W. Peterson, D. Novosel, D. Hart, T.W. Cease, and J. Schneider,
“Tapping IED Data to Find Transmission Faults,” IEEE Computer
Applications in Power, April 1999
[3] M. Adamiak, V. Madani, et al. – “Wide Area Protection and
Control, Today and Tomorrow”, Proceedings of IEEE T&D, May
2006
[9] B. Tatera “Setting a Strategic Direction for Substation
Automation at Pacific Gas & Electric Co - Western Power
Delivery Automation Conference, Spokane, Washington April
2003
[4] V. Madani , D. Novosel, P. Zhang, S. Meliopoulos, R. King, “Vision
in Protection and Control Area Meeting the Challenges of 21st
Century”, IEEE PSCE, October 2006
[10] P. Myrda, D. Tates, E. Udren, and D. Novosel: “Optimal Strategies
for System-Wide Protection and Control Replacement
Programs” CIGRE, Paris, August 2006
[5] V. Madani, G. Duru, B. Tatera, M. Jones, B. Pace, A Paradigm
Shift to Meet the Protection and Control Challenges of the
21st Century, Georgia tech Conference, Atlanta, May 2006
070209-v3
A Paradigm Shift in Protection, Control and Substation Automation Strategy
55
GE Smart Grid ad FINAL:Layout 1
6/11/09
4:43 PM
Page 1
GE Energy
Empowering
smarter energy choices
GE has a clear vision for managing, distributing and consuming electricity more efficiently:
Holistic solutions powered by proven technologies help utilities integrate more green energy, improve energy productivity,
and empower consumers to manage energy usage and save money.
It’s the GE smart grid. And we’re building it today.
See for yourself at: itsyoursmartgrid.com
imagination at work
Implementing Smart Grid Communications
Managing Mountains of Data Opens Up New
Challenges for Electric Utilities
James G. Cupp, PE, and Mike E. Beehler, PE
Burns & McDonnell
Electric utilities faced with the prospect of increasing customer
rates are seeking solutions to challenges presented by rising
global energy demand, aging infrastructure, increasing fuel costs
and renewable portfolio standards in light of climate change.
Many consider Smart Grid to be one such solution.
If we define the Smart Grid as “the convergence of information
and operational technology applied to the electric grid, allowing
sustainable options to customers and improved security, reliability
and efficiency to utilities,” then we must focus on deployment in
ways that address rate and bill impacts. Here, we will outline the
technical implementation of technologies that enable Smart Grid
practices.
1. Communications for Data Transport
Electric utilities continue to be among the largest users of
privately owned and operated wide-area networks (WANs)
for communications. These networks include a hybrid mix of
technologies including fiber optics, power line carrier systems,
copper-wire line, and a variety of licensed and unlicensed
wireless technologies. The utility WAN is designed to support
applications vital to the safe and reliable operation of the electric
utility mission-critical infrastructure: protective relaying for high
voltage lines, SCADA/EMS, mobile fleet voice and data dispatch,
generating plant automation, distribution feeder automation and
physical security. Rather than relying on public communication
carriers (AT&T, Sprint, Verizon, et al), utilities justify the costs of
building and operating their own private WANs because of the
highly critical nature of these applications for maintaining a
reliable and secure power grid. Less-critical business applications
such as corporate voice and data networks are also supported,
but are not normally the driver for private WAN deployment.
A typical electric utility WAN consists of a high-bandwidth
transport backbone network that backhauls large numbers of
channels and applications from the utility service territory to the
control center(s). Lower-bandwidth segments, or spurs, connect
individual or small groups of facilities to the backbone. Fiber optics
and/or digital microwave radio are usually the technologies of
choice for backbone transport, whereas the spurs may combine
these technologies with less robust alternatives such as copper
twisted-pair wire lines, power line carrier, VHF and UHF radio
links, and unlicensed wireless systems. Common carrier leased
services are used only sparingly in most cases, for low-criticality
applications in locations where privately owned alternatives are
cost-prohibitive.
These utility WANs have served traditional applications
like SCADA/EMS, distribution automation (DA)/demandside management (DSM) and automatic meter reading
(AMR), now popularly encompassed as part of the
Smart Grid (see Figure 1). The number of locations
requiring communications service increases and the
criticality of each location to the integrity of the overall
grid decreases as these applications are pushed
deeper into the distribution system (i.e., farther out from
the primary substation and closer to the customer).
Historically, this combination of increasing costs and
decreasing benefits has been the primary obstacle to
deployment of more feeder-level and customer-level
applications such as DA/DSM and AMR/advanced
meter infrastructure (AMI). When such applications
were deployed, costs were controlled by limiting
communications to one-way systems like broadcast
radio signals or narrowband, highlatency systems such
as power line carriers or dial-up phone lines.
Figure 1.
A utility WAN under Smart Grid applications is required to handle more robust data
transport.
Implementing Smart Grid Communications
57
What is needed is a nearly ubiquitous IP transport
network operating at bandwidths robust enough to
handle traditional utility power delivery applications
along with vast amounts of new data from the Smart
Grid.
2. Networks for the Future
Today, the political and regulatory impetus for wider deployment
of Smart Grid applications, especially their deployment all the
way to the customer premises, has resulted in pressure on
utility engineers to solve the problem of establishing robust data
transport WANs to the distribution feeder and customer level. The
proliferation of information technology utilizing Internet protocol
(IP) transport over Ethernet has made IP the de facto standard
for data transport. What is needed is a nearly ubiquitous IP
transport network operating at bandwidths robust enough to
handle traditional utility power delivery applications along with
vast amounts of new data from the Smart Grid. These networks
need to be scalable enough to handle future applications as they
come.
Communications for Smart Grid data transport require that utilities
address both the backbone and the spur segments. Most electric
utility communications backbones today are based largely on
traditional time-division multiplexing (TDM) digital architectures.
TDM technology, while highly reliable, was originally developed
for the transport of point-to-point constant-bit-rate voice
communications and is not necessarily suited to cost-effective
transport of point-to-multipoint “bursty” data traffic required in an
IP environment. The Smart Grid will require that these backbones
be upgraded to backhaul Ethernet/IP data traffic at speeds ranging
from one to 10 gigabits per second in a highly reliable manner.
Rather than replacing their legacy TDM networks, many utilities
will opt initially to overlay these existing networks by overbuilding
gigabit Ethernets on unused fiber, and licensed or unlicensed
broadband wireless networks over existing microwave paths.
3. Last-Mile Challenges
The deployment of spur or last-mile communications for the Smart
Grid, typically from a backbone node to the customer premises,
offers additional challenges: First, the network must cover a
large area, especially if coverage of residential customers is to
be provided. This has prompted some utilities to take a phased
approach, deploying the Smart Grid to large-load industrial and
commercial customers initially, since the bulk of the benefits of
Smart Grid follow the bulk of the electrical load, while residential
applications may remain on the back burner, waiting for a clearer
quantification of benefits. This balanced approach may make
sense economically but may have broad ramifications politically
as rates rise and residential customers (voters) demand relief.
Second, the proper balancing of performance and cost is less
clear for these last-mile applications. Losing communications with
a small percentage of the DA or AMI for a time, while undesirable,
would pose no real threat to the safe and reliable operation of
the overall power grid. Communications with a single customer
or residence do not require the bandwidth and performance
needed in the backbone, so low-speed communication devices
with marginal signal strength that may require multiple
retransmissions to complete a message can be tolerated. These
58
issues raise questions like, “How reliable is reliable enough?” “How
fast is fast enough?” and “At what cost?”
The relaxed performance and reliability constraints in the last
mile also mean that the number of technology options available
for this portion of the WAN are more plentiful. Technologies like
meshed Wi-Fi, packet-based store and forward radio networks,
and broadband-over-power line (BPL), not considered reliable or
robust enough for the mission-critical infrastructure backbone,
are viable options for the last mile. Likewise, public carrier and
CATV-based services like broadband cable modem, digital
subscriber line (DSL) and cellular-based wireless data networks
may also make sense where utilities can negotiate bulk service
rates.
4. Data Integration and Management
Once the DA or AMI data is efficiently transported, a completely
new set of data integration and management issues will challenge
utilities technically and culturally. The Smart Grid will generate
billions of data points from thousands of system devices and
hundreds of thousands of customers. Data must be converted
to information through a knowledge-management life cycle in
which the data from meters and appliances or substations and
distribution systems are analyzed and integrated in a manner that
leads to action. A data-to-information-to-action plan will develop
as a better understanding of load factors, energy usage patterns,
equipment condition, voltage levels, etc. emerges through analysis
and is integrated as functional information into usable customer
programs and/or operation and maintenance algorithms that
identify, trend and alert operators to incipient failure.
The first phase of the knowledge management effort and a
key component in the system of information ecology is data
conservation in a data warehouse. Data storage needs will
explode. Data security will be important, but some of the best
system or customer programs may result by allowing engineers
and operators the opportunity to freely analyze some or all of
the data. IBM, Oracle and Microsoft recognize the huge growth
potential and are visibly promoting their solution concepts.
The Smart Grid is expected to be fully functional by 2030. Data
collected, analyzed, visualized and warehoused from the Smart
Grid will contribute to many new ideas and inventions that can
improve lives.
Dennis M. Klinger, vice president of information management
services for Florida Power & Light, calls this “moving at the speed
of value.” In an era of serial rate increases, customers will demand
value, and utilities must deliver that value. This will be the future
of electricity.
5. Customer Programs
The future of electricity begins with the customer. Integration and
management of system and customer data can lead to the ability
to analyze warehoused information in a manner that improves
operational efficiency and reliability, but most importantly,
provides sustainable options for customers. Sustainable options
will include demand response and demand-side management
programs for all customer classes that include a home area
network (HAN) plan for residential customers, allowing prices to
devices supported by ultra-simple rate plans. Data will become
information used for action.
Implementing Smart Grid Communications
6. Scheduling Savings
The HAN is a computer automation system for the home (or small
commercial business) that integrates devices through the Internet
and with the electric utility to allow the user to be proactive in
the use or generation of energy. The HAN will play a major role in
making the grid more efficient and in moderating rate impact for
the customer. The HAN begins on the customer side of the meter
and will be made up of plug-in hybrid electric vehicles, renewable
and/or distributed generation, HVAC systems, pool pumps,
intelligent appliances and consumer devices like MP3 players, cell
phones and iPods authenticated to the electric utility on a secure
network owned by the owner. The owner will have the ability to
control the operation of devices on the HAN from a computer to
maximize the advantages for demand response (DR) or DSM rate
structures offered by the electric utility.
7. Improving Load Factors
DR is a voluntary rate structure that typically lowers a customer’s
general rate per kilowatthour in return for the utility’s option to
curtail power as needed during system peak loading events. DSM
is the effort to incentivize customer use through simple timeof-use rates that generally correspond to the cost of producing
electricity. DR and DSM shift electric load and improve the electric
utility’s load factor and should not be confused with energy
efficiency programs that reduce load and, therefore, sales. The
current regulatory construct allows utilities a reasonable rate of
return, or profits, on prudent investments and the cost to operate
and maintain those investments. Some utilities seek to decouple
their sales from profits since energy-efficiency programs lower
sales as less electricity is consumed. Decoupled sales and profits
theoretically make the electric utility indifferent to energy efficiency
programs and distributed generation but remain a controversial
issue in the industry. Cyber security, ownership of customer
data, standardization of device protocols for low-power personal
networks, customer acceptance of DR and DSM programs and
other issues are also sources of controversy in HAN build-out.
Allowing customers to make sustainable decisions
on the use of electricity and, ultimately, satisfying
regulators will provide for full rate recovery and return
on investment.
8. Where Next?
Regulators and lawmakers are not passively waiting for utilities
to offer solutions to the serial rate increases that are coming.
Regulatory action is being taken and the desired result is clear.
The Smart Grid must provide sustainable options to customers.
Allowing customers to make sustainable decisions on the use of
electricity and, ultimately, satisfy regulators will provide for full
rate recovery and return on investment. That done, utilities can
move at the speed of value to confidently use the Smart Grid to
achieve other security, reliability and efficiency objectives. These
other objectives may include more efficient meter reads and
billing, better customer service, theft/tamper detection, turn-on/
turn-off service, advanced pay services, load forecasting, asset
management, transformer sizing, power quality improvements,
and a myriad of other efficiencies and services that will be
developed in the years to come, when today’s electric grid
becomes the Smart Grid.
Ü Originally published in TECH Briefs, 2008, No. 4
070209-v2
Implementing Smart Grid Communications
59
HardFiber System
TM
IEC 61850
Process Bus Solution
The Multilin HardFiber System was specifically designed
to break the bonds of copper cabling by reducing capital
costs, and freeing highly skilled utility workers from
performing labor-intensive activities.
The HardFiber System is a complete IEC 61850 system available
for a wide variety of protection applications, including generator,
transformer, transmission line, bus, feeder, capacitor bank, and
motor protection.
By eliminating the need to install and maintain thousands of
copper wires, utilities can save up to 50% of protection and control
installation and maintenance costs, while at the same time increasing
worker safety and power system reliability.
Faster Installation...
Better Reliability...
Lower Costs...
HardFiberTM Brick
Universal Relay Family
Hardened Switchyard Interface
IEC 61850 Compliant P&C Family
g
Digital Energy
Multilin
GE Multilin
www.GEMultilin.com/HardFiber
[email protected]
Worldwide
Tel: 905-294-6222
North America
Tel: 1-800-547-8629
Europe/MiddleEast/Africa
Tel: +34 94 485 88 00
IEC 61850 Communication Networks and
Systems In Substations:
An Overview for Users
Mark Adamiak , Drew Baigent
GE Digital Energy
Ralph Mackiewicz
SISCO
1. Abstract
Over the last decade, the “digitization” of the electron enterprise
has grown at exponential rates. Utility, industrial, commercial, and
even residential consumers are transforming all aspects of their
lives into the digital domain. Moving forward, it is expected that
every piece of equipment, every receptacle, every switch, and
even every light bulb will possess some type of setting, monitoring
and/or control. In order to be able to manage the large number
of devices and to enable the various devices to communicate
with one another, a new communication model was needed.
That model has been developed and standardized as IEC 61850
– Communication Networks and Systems in Substations [1]. This
paper looks at the needs of next generation communication
systems and provides an overview of the IEC 61850 protocol and
how it meets these needs.
2. Communication System Needs
Communication has always played a critical role in the real-time
operation of the power system. In the beginning, the telephone
was used to communicate line loadings back to the control
center as well as to dispatch operators to perform switching
operations at substations. Telephone-switching based remote
control units were available as early as the 1930’s and were
able to provide status and control for a few points. As digital
communications became a viable option in the 1960’s, data
acquisition systems (DAS) were installed to automatically collect
measurement data from the substations. Since bandwidth was
limited, DAS communication protocols were optimized to operate
over low-bandwidth communication channels. The “cost” of this
optimization was the time it took to configure, map, and document
the location of the various data bits received by the protocol.
As we move into the digital age, literally thousands of analog and
digital data points are available in a single Intelligent Electronic
Device (IED) and communication bandwidth is no longer a limiting
factor. Substation to master communication data paths operating
at 64,000 bits per second are becoming common-place with an
obvious migration path to much high rates. With this migration in
technology, the “cost” component of a data acquisition system has
now become the configuration and documentation component.
Consequently, a key component of a communication system is
the ability to describe themselves from both a data and services
(communication functions that an IED performs) perspective.
Other “key” requirements include:
•
High-speed IED to IED communication
•
Networkable throughout the utility enterprise
•
High-availability
•
Guaranteed delivery times
•
Standards based
•
Multi-vendor interoperability
•
Support for Voltage and Current samples data
•
Support for File Transfer
•
Auto-configurable / configuration support
•
Support for security
Given these requirements, work on a “next generation”
communication architecture began with the development of
the Utility Communication Architecture (UCA) in 1988. The result
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
61
of this work was a profile of “recommended” protocols for the
various layers of the International Standards Organization (ISO)
Open System Interconnect (OSI) communication system model.
This architecture resulted in the definition of a “profile” of protocols,
data models, and abstract service definitions that became known
as UCA. The concepts and fundamental work done in UCA became
the foundation for the work done in the IEC TC57 Working Groups
10, 11, and 12 which resulted in the International Standard – IEC
61850 – Communication Networks and Systems in Substations
[1].
3. Scope and Outline of IEC 61850
The stated scope of IEC 61850 was communications within the
substation. The document defines the various aspects of the
substation communication network in 10 major sections as
shown in Table 1 below.
Part #
Title
1
Introduction and Overview
2
Glossary of terms
3
General Requirements
4
System and Project Management
5
Communication Requirements for Functions and Device
Models
6
Configuration Description Language for Communication in
Electrical Substations Related to IEDs
7
Basic Communication Structure for Substation and Feeder
Equipment
7.1
- Principles and Models
7.2
- Abstract Communication Service Interface (ACSI)
7.3
- Common Data Classes (CDC)
7.4
- Compatible logical node classes and data classes
8
Specific Communication Service Mapping (SCSM)
8.1
9
- Mappings to MMS (ISO/IEC 9506 – Part 1 and Part 2) and
to ISO/IEC 8802-3
Given the data and services abstract definitions, the final step was
one of “mapping” the abstract services into an actual protocol.
Section 8.1 defines the mapping of the abstract data object
and services onto the Manufacturing Messaging Specification –
MMS2 and sections 9.1 and 9.2 define the mapping of the Sample
Measured Values (unidirectional point-to-point and bi-directional
multipoint accordingly) onto an Ethernet data frame. The 9.2
document defines what has become known as the Process Bus.
From a system perspective, there is a significant amount of
configuration that is required in order to put all the pieces together
and have them work. In order to facilitate this process and to
eliminate much of the human error component, an XML based
Substation Configuration Language (SCL) was defined in part
6. It allows the formal description of the relations between the
substation automation system and the substation (switchyard).
At the application level, the switchyard topology itself and the
relation of the switchyard structure to the SAS functions (logical
nodes) configured on the IEDs can be described. Each device must
provide an SCL file that describes the configuration of itself.
Although the scope of IEC 61850 was originally focused “inside”
the substation, discussions are underway to look at defining IEC
61850 for the Substation to Master communication protocol
(already in service in several installations). In addition, applications
are in service that uses various components of IEC 61850 for wide
area substation-to-substation communication.
Finally, part 10 of the document defines a testing methodology
in order to determine “conformance” with the numerous protocol
definitions and constraints defined in the document.
The rest of this paper goes into some focused details of the various
parts of the IEC 61850 standard.
Specific Communication Service Mapping (SCSM)
9.1
- Sampled Values over Serial Unidirectional Multidrop Pointto-Point Link
4. Modeling Approach
9.2
- Sampled Values over ISO/IEC 8802-3
Legacy protocols have typically defined how bytes are transmitted
on the wire. However, they did not specify how data should be
organized in devices in terms of the application. This approach
requires power system engineers to manually configure
objects and map them to power system variables and low-level
register numbers, index numbers, I/O modules, etc. IEC 61850 is
unique. In addition to the specification of the protocol elements
(how bytes are transmitted on the wire), IEC 61850 provides a
comprehensive model for how power system devices should
organize data in a manner that is consistent across all types and
brands of devices. This eliminates much of the tedious non-power
system configuration effort because the devices can configure
themselves. For instance, if you put a CT/VT input into an IEC
61850 relay, the relay can detect this module and automatically
assign it to a measurement unit without user interaction. Some
devices use an SCL file to configure the objects and the engineer
need only import the SCL file into the device to configure it. Then,
the IEC 61850 client application can extract the object definitions
from the device over the network. The result is a very large savings
in the cost and effort to configure an IEC 61850 device.
10
Conformance Testing
Table 1.
Parts 3, 4, and 5 of the standard start by identifying the general
and specific functional requirements for communications in a
substation (key requirements stated above). These requirements
are then used as forcing functions to aid in the identification of the
services and data models needed, application protocol required,
and the underlying transport, network, data link, and physical
layers that will meet the overall requirements.
The major architectural construct that 61850 adopts is that of
“abstracting” the definition of the data items and the services,
that is, creating data items/objects and services that are
independent of any underlying protocols. The abstract definitions
then allow “mapping” of the data objects and services to any
other protocol that can meet the data and service requirements.
The definition of the abstract services is found in part 7.2 of the
62
standard and the abstraction of the data objects (referred to as
Logical Nodes) is found in part 7.4. In as much as many of the
data objects are made up of common pieces (such as Status,
Control, Measurement, Substitution), the concept of “Common
Data Classes” or “CDC” was developed which defined common
building blocks for creating the larger data objects. The CDC
elements are defined in part 7.3.
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
The IEC 61850 device model begins with a physical device. A
physical device is the device that connects to the network. The
physical device is typically defined by its network address. Within
each physical device, there may be one or more logical devices.
The IEC 61850 logical device model allows a single physical device
to act as a proxy or gateway for multiple devices thus providing a
standard representation of a data concentrator.
Each logical device contains one or more logical nodes. A logical
node (see Figure 1) is a named grouping of data and associated
services that is logically related to some power system function.
XCBR Class
Data Name
Common
Data
Class
Description
LNName
Secure
Shall be inherited from Logical-Node
Class (see IEC 61850-7-2)
T
Mandatory/
Optional
DATA
Common Logical Node Information
LN shell inherit all Mandatory Data from
Common Logical Node Class
Mandatory
Mandatory
Each logical node contains one or more elements of Data. Each
element of data has a unique name. These Data Names are
determined by the standard and are functionally related to the
power system purpose. For instance, a circuit breaker is modeled
as an XCBR logical node. It contains a variety of Data including
Loc for determining if operation is remote or local, OpCnt for an
operations count, Pos for the position, BlkOpn block breaker open
commands, BlkCls block breaker close commands, and CBOpCap
for the circuit breaker operating capability.
Each element of data within the logical node conforms to the
specification of a common data class (CDC) per IEC 61850-7-3. Each
CDC describes the type and structure of the data within the logical
node. For instance, there are CDCs for status information, measured
information, controllable status information, controllable analog
set point information, status settings, and analog settings. Each
CDC has a defined name and a set of CDC attributes each with a
defined name, defined type, and specific purpose. Each individual
attribute of a CDC belongs to a set of functional constraints (FC)
that groups the attributes into categories. For instance, in the
Single Point Status (SPS) CDC described in Figure 2, there are
functional constraints for status (ST) attributes, substituted value
(SV) attributes, description (DC) attributes, and extended definition
(EX) attributes. In this example the status attributes of the SPS
class consists of a status value (stVal), a quality flag (q), and a time
stamp (t).
Loc
SPS
Local operation (local means without
substation automation communication,
hardwired direct control)
EE Health
INS
External equipment health
EE Name
DPL
External equipment name plate
OpCnt
INS
Operation counter
Mandatory
Pos
DPC
Switch position
Mandatory
Attribute
Name
BlkOpn
SPC
Block opening
Mandatory
DataName
BlkCls
SPC
Block closing
Mandatory
DATA Attribute
ChaMotEna
SPC
Charger motor enabled
Optional
Sum o Switched Amperes, resetable
Optional
Optional
Optional
SPS Class
Controls
Metered Values
SumSwARs
BCR
Status Information
Attribute Type
Functional
Constraint
TrgOp
Value /
Value
Range
Mandatory/
Optional
true|false
Mandatory
Inherited from Data Class (see IEC 61850-7-2)
Status
stVal
BOOLEN
ST
dchg
q
Quality
ST
qchg
t
TimeStamp
ST
Mandatory
SV
PICS_SUBST
Mandatory
CBOpCap
INS
Circuit breaker operating capability
Mandatory
POWCap
INS
Point on Wave switching capability
Optional
subEna
BOOLEN
MaxOpCap
INS
Circuit breaker operating capability when
fully charged
Optional
subVal
BOOLEN
SV
subQ
Quality
SV
PICS_SUBST
subID
VISIBLE STRING64
SV
PICS_SUBST
Figure 1.
Anatomy of Circuit Breaker (XCBR)
Logical Node in IEC 61850-7-4
There are logical nodes for automatic control the names of which
all begin with the letter “A”. There are logical nodes for metering
and measurement the names of which all begin with the letter “M”.
Likewise there are logical nodes for Supervisory Control (C), Generic
Functions (G), Interfacing/Archiving (I), System logical nodes
(L), Protection (P), Protection Related (R), Sensors (S), Instrument
Transformers (T), Switchgear (X), Power Transformers (Y), and Other
Equipment (Z). Each logical node has an LN-Instance-ID as a suffix
to the logical node name. For instance, suppose there were two
measurement inputs in a device to measure two 3-phase feeders.
The standard name of the logical node for a Measurement Unit for
3-phase power is MMXU. To delineate between the measurements
for these 2 feeders the IEC 61850 logical node names of MMXU1
and MMXU2 would be used. Each logical node may also use
an optional application specific LN-prefix to provide further
identification of the purpose of a logical node.
Substitution
true|false
PICS_SUBST
Configuration, description and extension
d
VISIBLE STRING255
DC
Text
Optional
dU
UNICODE STRING255
DC
Optional
cdcNs
VISIBLE STRING255
EX
AC_DLNDA_M
cdcName
VISIBLE STRING255
EX
AC_DLNDA_M
dataNs
VISIBLE STRING255
EX
AC_DLN_M
Figure 2.
Anatomy of the Single Point Status (SPS)
Common Data Class in IEC 61850-7-3
The IEC 61850 model of a device is a virtualized model that begins
with an abstract view of the device and its objects and is defined
in IEC 61850 part 7. Then, this abstract model is mapped to a
specific protocol stack in section IEC 61850-8-1 based on MMS
(ISO9506), TCP/IP, and Ethernet. In the process of mapping the
IEC 61850 objects to MMS, IEC 61850-8-1 specifies a method of
transforming the model information into a named MMS variable
object that results in a unique and unambiguous reference for
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
63
each element of data in the model. For instance, suppose that you
have a logical device named “Relay1” consisting of a single circuit
breaker logical node XCBR1 for which you want to determine if the
breaker is in the remote or local mode of operation. To determine
this you would read the object shown in Figure 3.
Relay1/XCBR1$ST$Loc$stVal
Attribute
Data
Functional Constraint
Logical Node
Logical Device
Figure 3.
Anatomy of an IEC 61850-8-1 Object Name
5. Mapping to Real Protocols
The abstract data and object models of IEC 61850 define a
standardized method of describing power system devices that
enables all IEDs to present data using identical structures that
are directly related to their power system function. The Abstract
Communication Service Interface (ACSI) models of IEC 61850 define
a set of services and the responses to those services that enables
all IEDs to behave in an identical manner from the network behavior
perspective. While the abstract model is critical to achieving this
level of interoperability, these models need to be operated over
a real set of protocols that are practical to implement and that
can operate within the computing environments commonly found
in the power industry. IEC 61850-8-1 maps the abstract objects
and services to the Manufacturing Message Specification (MMS)
protocols of ISO9506. Why was a protocol originally designed
for manufacturing used? Because MMS is the only public (ISO
standard) protocol that has a proven implementation track record
that can easily support the complex naming and service models
of IEC 61850. While you can theoretically map IEC 61850 to any
protocol, this mapping can get very complex and cumbersome
when trying to map IEC 61850 objects and services to a protocol
that only provides read/write/report services for simple variables
that are accessed by register numbers or index numbers. This
was the reason that MMS was chosen for UCA in 1991 and is
the reason that it was kept for IEC 61850. MMS is a very good
choice because it supports complex named objects and a rich set
of flexible services that supports the mapping to IEC 61850 in a
straightforward manner.
The mapping of IEC 61850 object and service models to MMS is
based on a service mapping where a specific MMS service/services
are chosen as the means to implement the various services of
ACSI. For instance, the control model of ACSI is mapped to MMS
read and write services. Then the various object models of IEC
61850 are mapped to specific MMS objects. For instance, the IEC
61850 logical device object is mapped to an MMS domain. Table
2 summarizes the mapping of IEC 61850 objects and Table 3 the
ACSI mapping to MMS.
In addition to the mapping to the application layer, Part 8.1 defines
profiles for the “other” layers of the communication stack that
are dependent on the service provided (as shown in Figure 4).
Of note on the various profiles: the Sampled Values and GOOSE
applications map directly into the Ethernet data frame thereby
eliminating processing of any middle layers; the MMS Connection
64
IEC61850 Objects
MMS Object
SERVER class
Virtual Manufacturing
Device (VMD)
LOGICAL DEVICE class
Domain
LOGICAL NODE class
Named Variable
DATA class
Named Variable
DATA-SET class
Named Variable List
SETTING-GROUP-CONTROL-BLOCK class
Named Variable
REPORT-CONTROL-BLOCK class
Named Variable
LOG class
Journal
LOG-CONTROL-BLOCK class
Named Variable
GOOSE-CONTROL-BLOCK class
Named Variable
GSSE-CONTROL-BLOCK class
Named Variable
CONTROL class
Named Variable
Files
Files
Table 2.
IEC 61850 to MMS object mapping
IEC 61850 Services
MMS Services
LogicalDeviceDirectory
GetNameList
GetAllDataValues
Read
GetDataValues
Read
SetDataValues
Write
GetDataDirectory
GetNameList
GetDataDefinition
GetVariableAccessAttributes
GetDataSetValues
Read
SetDataSetValues
Write
CreateDataSet
CreateNamedVariableList
DeleteDataSet
DeleteNamedVariableList
GetDataSetDirectory
GetNameList
Report (Buffered and Unbuffered)
InformationReport
GetBRCBValues/GetURCBValues
Read
SetBRCBValues/SetURCBValues
Write
GetLCBValues
Read
SetLCBValues
Write
QueryLogByTime
ReadJournal
QueryLogAfter
ReadJournal
GetLogStatusValues
GetJournalStatus
Select
Read/Write
SelectWithValue
Read/Write
Cancel
Write
Operate
Write
Command-Termination
Write
TimeActivated-Operate
Write
GetFile
FileOpen/FileRead/FileClose
SetFile
ObtainFile
DeleteFile
FileDelete
GetFileAttributeValues
FileDirectory
Table 3.
IEC 61850 services mapping (partial)
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
Oriented layer can operate over TCP/IP or ISO; the Generic
Substation Status Event (GSSE) is the identical implementation as
the UCA GOOSE and operates over connectionless ISO services; all
data maps onto an Ethernet data frame using either the data type
“Ethertype” in the case of Sampled Values, GOOSE, TimeSync, and
TCP/IP or “802.3” data type for the ISO and GSSE messages.
Part 9.1 specifies a pre-configured or “universal” dataset as
defined in IEC 60044-8. This dataset includes 3-phase voltage, bus
voltage, neutral voltage, 3-phase currents for protection, 3-phase
currents for measurement and two 16-bit status words. Note that
the analog data values are mapped into 16-bit registers in this
mapping.
6. Process Bus
As technology migrates to “next generation” low-energy voltage
and current sensors, the ability to digitize the base quantities
at the source and transmit the resulting sample values back to
the substation becomes a need. In addition to Sampled Values,
the ability to remotely acquire status information as well as set
output controls is very desirable. IEC 61850 addresses this need
through the definition of Sampled Measured Values services and
the implementation of a Process Bus. The Process layer of the
substation is related to gathering information, such as Voltage,
Current, and status information, from the transformers and
transducers connected to the primary power system process – the
transmission of electricity. IEC 61850 defines the collection of this
data via two different protocol definitions, namely, Part 9.1 which
defines a Unidirectional Multidrop Point-to-Point fixed link carrying
a fixed dataset and Part 9.2 which defines a “configurable” dataset
that can be transmitted on a multi-cast basis from one publisher
to multiple subscribers.
Figure 5, above, shows the basic concept of the Process Bus.
Signals from voltage and current sources (low or high energy) as
well as status information are input into a “Merging Unit” (MU).
The Merging Units in a station sample the signals at an agreed,
synchronized rate. In this manner, any IED can input data from
multiple MUs and automatically align and process the data. At
this time, there is an implementation agreement that defines a
base sample rate of 80 samples per power system cycle for basic
protection and monitoring and a “high” rate of 256 samples per
power system cycle for high-frequency applications such as
power quality and high-resolution oscillography.
Figure 5.
Sample Measured Value Concept
Part 9.2 is a more generalized implementation of Sampled
Measured Values (SMV) data transfer. In 9.2, the dataset or
“payload” is user-defined using the SCL. As a dataset, data values
of various sizes and types can be integrated together. Note that
the existing implementation agreement proposed a data value
size of 32 bits with a scale factor of 1 count = 1ma.
Both 9.1 and 9.2 specify mapping directly onto an Ethernet
transport (see Figure 4). Depending on the sample data rate,
anywhere from 1 to 5 devices can be mapped onto a single 100MB
Ethernet link. Multiple 100MB Ethernet data streams can then be
Figure 4.
Overview of IEC 61850 Functionality and Associated Communication Profiles
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
65
combined into a single Ethernet switch with a 1GB backbone.
In this configuration, 50 or more datasets can be published to
multiple subscribers.
7. Substation Configuration Language
IEC 61850-6-1 specifies a Substation Configuration Language (SCL)
that is based on the eXtensible Markup Language (XML) to describe
the configuration of IEC 61850 based systems. SCL specifies a
hierarchy of configuration files that enable multiple levels of the
system to be described in unambiguous and standardized XML
files. The various SCL files include system specification description
(SSD), IED capability description (ICD), substation configuration
description (SCD), and configured IED description (CID) files. All
these files are constructed in the same methods and format but
have different scopes depending on the need.
VLAN allows the Ethernet switch to deliver datasets to only those
switch ports/IEDs that have subscribed to the data. In migrating to
Process Bus implementations, manufacturers will need to provide
the ability to integrate data from existing CTs and PTs with the
data from the newer Optical/Electronic sensors. A redundant
synchronization clock architecture will also have to be addressed.
In this architecture, upon detection of failure of Clock 1, Clock 2
will have to automatically come on line and continue providing
sampling synchronization.
Even though an IEC 61850 client can extract an IED’s configuration
from the IED when it is connected to that IED over a network,
there are several scenarios where the availability of a formal
off-line description language can bring very large benefits to
users outside of configuring IEC 61850 client applications. These
benefits include:
•
SCL enables off-line system development tools to generate
the files needed for IED configuration automatically from
the power system design significantly reducing the cost and
effort of IED configuration by eliminating most, if not all,
manual configuration tasks.
•
SCL enables the sharing of IED configuration among users
and suppliers to reduce or eliminate inconsistencies and
misunderstandings in system configuration and system
requirements. Users can provide their own SCL files to ensure
that IEDs are delivered to them properly configured.
•
SCL allows IEC 61850 applications to be configured off-line
without requiring a network connection to the IED for client
configuration.
SCL can be used as best fits each user’s requirements. A user
can decide to use CID files to provide help in IED configuration
using its existing system design processes. Or SCL can be used to
restructure the entire power system design process to eliminate
manual configuration, eliminate manual data entry errors, reduce
misunderstanding between system capabilities and requirements,
enhance the interoperability of the end system, and greatly
increase the productivity and effectiveness of power system
engineers.
8. IEC Substation Model
Putting the pieces together results in the substation architecture
shown in Figure 6.
At the “process” layer, data from Optical/Electronic Voltage and
Current sensors as well as status information will be collected
and digitized by the Merging Units (MUs). MUs could be physically
located either in the field or in the control house. Data from the
MUs will be collected through redundant 100MB fiber optic
Ethernet connections. The collection points will be redundant
Ethernet switches with 1GB internal data buses and 1GB uplinks
that support Ethernet priority and Ethernet Virtual LAN (VLAN).
66
Figure 6.
IEC 61850 Substation Model
At the substation level, a Station Bus will exist. Again, this bus
will be based today on 10MB Ethernet with a clear migration
path to 100MB Ethernet. The Station Bus will provide primary
communications between the various Logical Nodes, which provide
the various station protection, control, monitoring, and logging
functions. Communications will operate on either a connection
oriented basis (e.g. – request of information, configuration, etc.) or
a connection-less basis (IEC Generic Object Oriented Substation
Event - GOOSE). Again, a redundant communication architecture
is recommended as application of IED to IED data transmission
puts the communication system on the critical path in case of a
failure.
Finally, this architecture supports remote network access for all
types of data reads and writes. As all communication is network
enabled, multiple remote “clients” will desire access the wide
variety of available information. Typical clients would include
local HMI, operations, maintenance, engineering, and planning.
The remote access point is one logical location to implement
security functions such as encryption and authentication. This
implementation un-burdens the individual IEDs from performing
encryption on internal data transfers but still provide security on
all external transactions.
9. Application Software
A variety of commercial products supporting IEC 61850 are already
available and the future holds promise for many new innovations
that will greatly benefit users. Of particular significance are
products that support both the IEC 61850 communications
standard and the OLE for Process Control (OPC see http://www.
opcfoundation.org) application program interface (API) standard of
the OPC Foundation. The combination of a standardized protocol
and a standardized API is a powerful tool that allows users to
dramatically lower their costs to build substation automation
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
systems by enabling products from different vendors to plug
together into a complete solution.
is able to install and commission systems quicker with less effort
and fewer errors resulting in lower costs.
The OPC Data Access (DA) specification is an API that enables
an OPC Client application, such as a SCADA or Human Machine
Interface (HMI) application, to provide a generic interface to
outside data that is independent of any specific protocol (Figure 7).
This enables third parties to develop OPC Servers to interface with
a wide variety of protocols, including IEC 61850, Modbus, DNP3,
and hundreds of other protocols. There is a wide availability of
both client and server applications that provide users choice and
flexibility. For instance, interfaces to many different applications like
relational data base management systems (RDBMS), spreadsheets,
data historians, trending systems, etc. are available that support
OPC and provide a large choice of options to implement complex
systems at a low cost (see Figure 8).
Figure 9.
IEC 61850 Data centrator Architecture
9.1 Interface with Legacy Protocols
Figure 7.
Using IEC 61850 with OPC
Electric power systems are designed to last for many years. For
any new technology to be successfully applied into a modern
power system, there must be some way to accommodate the
use of legacy IEDs and protocols from the past. IEC 61850 is no
different and there are several methods for accommodating legacy
protocols in an IEC 61850 system. IEC 61850 itself is well suited to
accommodate legacy protocols with its logical device model. The
ability to support multiple logical devices within a single physical
device allows IEC 61850 to directly support the modeling of a data
concentrator or multi-device gateway inherently without resorting
to techniques outside the scope of the standard. Data concentrator
devices (Figure 9) supporting the IEC 61850 logical device model
are available with new products under development. In addition
to the use of separate data concentrators, OPC technology also
offers a way to incorporate simple gateway functionality into a
substation SCADA system (Figure 10). In this case, the roles of
OPC client and server are reversed from the previous example
illustrating a substation SCADA application by building an OPC
client application on top of an IEC 61850 server. The OPC client
is then mapped to an OPC server supporting any legacy or
proprietary protocol. This enables data from legacy devices to
be accessed as IEC 61850 data simplifying the client application
development by providing a consistent standardized mechanism
for data access across the entire substation.
Figure 8.
OPC Enables IEC 61850 Interface to IT
In addition to providing access to data in IEDs, OPC interfaces
support an important feature called browsing. The OPC browse
interface enables the client to retrieve the list of data items
defined in a server instead of having to be pre-configured. This
works especially well with IEC 61850 devices because of built-in
support for object discovery. By combining OPC with IEC 61850
the substation engineer avoids many hours of configuration and
Figure 10.
IEC 61850 Data Gateways using OPC
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
67
10. Conclusions
11. References
IEC 61850 is now released to the industry. Nine out of ten parts of
the standard are now International Standards (part 10 on testing is
in the CDV stage). This standard addresses most of the issues that
migration to the digital world entails, especially, standardization
of data names, creation of a comprehensive set of services,
implementation over standard protocols and hardware, and
definition of a process bus. Multi-vendor interoperability has been
demonstrated and compliance certification processes are being
established. Discussions are underway to utilize IEC 61850 as the
substation to control center communication protocol. IEC 61850
will become the protocol of choice as utilities migrate to network
solutions for the substations and beyond.
[1] IEC 61850 - Communication Networks and Systems in
Substations; http://domino.iec.ch/webstore/webstore.nsf/se
archview/?SearchView=&SearchOrder=4&SearchWV=TRUE&
SearchMax=1000&Query=61850&submit=OK
[2] Manufacturing Messaging Specification; ISO 9506-1&2:2003;
Part 1 – Service Definition: Part 2 – Protocol Specification
070209-v3
68
IEC 61850 Communication Networks and Systems In Substations: An Overview for Users
Enhanced Security and Dependability in
Process Bus Protection Systems
David McGinn, Vijay Muthukrishnan, Wei Wang
GE Digital Energy
1. Introduction
Power system protection is a mission-critical application with
demanding requirements in terms of security, dependability and
availability.
This paper reports on a new generation protection and control
platform designed for enhanced security, dependability and
availability. This system is based on traditionally defined protective
relays with data acquisition through fully duplicated remote I/O
devices (merging units). The I/O devices are deployed directly on or
in close proximity to the primary equipment. They communicate
their input signals including ac sampled values, contact statuses,
and a variety of transducer inputs and execute the trip and control
commands using process bus. Message formats are as defined in
IEC 61850 [1]. To improve overall system reliability, availability and
performance for the task of protection and control, the system
does not include any active (switched) communication networks,
but is based on simple, dedicated point-to-point Ethernet
connections with extensive self-monitoring and inherent security.
In particular, by not having active Ethernet infrastructure, the
system is not vulnerable to cyber security threats through its
process bus.
This paper reviews requirements and practical design solutions
for electronics acting as an I/O structure of a protection system
when installed in a high voltage switchyard environment, using
the above mentioned new protection platform to illustrate
the design challenges and solutions. These challenges include
temperature requirements, weather proofing, electrical and
magnetic fields, conducted and radiated interference, mechanical
shock and vibration, and similar conditions. These requirements
are applied to mechanical packaging, fiber connectivity, copper
wiring connectivity, printed circuit boards, electronic components,
internal data buses, and software architecture.
Traditionally, microprocessor-based relays incorporate a certain
degree of internal self-diagnostics and checking to guard against
internal problems that could potentially result in a false operation
or a failure to trip. Extra dependability is achieved in applications
using multiple relays in parallel redundancy, while additional
security may be implemented by using multiple relays in a voting
scheme. The paper describes the implementation of self-testing
and redundancy in the presented system, which makes it secure
and dependable well beyond the realm of traditional protection
and control devices.
In summary, the system reported on features greatly improved
immunity to failures of its components, considerably reducing
the danger of false tripping while enhancing dependability. The
subject of this paper is not just a concept but also a new technical
solution now available commercially, and backed up by significant
development and testing efforts [2].
2. System Aspects for Reliability,
Security and Dependability
In general, the function of a protection system is to limit the severity
and extent of system disturbances and possible damage to system
equipment. These objectives can be met only if protection systems
have a high degree of dependability and security. In this context
dependability relates to the degree of certainty that a protection
system will operate correctly when required to operate. Security
relates to the degree of certainty that a protection system will not
operate when not required to operate. The relative effect on the
bulk power system of a failure of a protection system to operate
when desired versus an unintended operation should be weighed
carefully in selecting protection system design parameters. Often
increased dependability (fewer failures to operate) results in
decreased security (more unintended operations), and vice versa
[3].
Another important reliability index for protection and control
systems is the attendance rate, the expectation of the number
of instances where the system needs to be repaired, re-verified,
upgraded, etc. per unit time. With each such instance, in addition
to the cost of labour to schedule, prepare for and perform the
work, and the cost of any required power element outage, there is
a finite probability that an error will occur resulting in an undesired
trip operation during restoration to service or afterward. Trips are
of special concern where generation is tripped or load curtailed.
Infrequent but still of concern is the possibility that a widespread
blackout can result, such as occurred on February 26 2007 in
Florida [4]. Increased redundancy means an increased number
of components, and as each component failure requires a repair
attendance, the failure rate of each component contributes to the
total attendance rate.
Dependability and security can be obtained through the
appropriate use of redundancy. For instance, a voting scheme
may be proposed wherein tripping occurs only when at least two
of three independent protection systems indicate that tripping is
required. However, voting schemes complicate the application
in terms of engineering, implementation, commissioning and
testing, potentially increasing the risk of unexpected operations
due to procedural or site crew mistakes. The extra hardware
Enhanced Security and Dependability in Process Bus Protection Systems
69
alone yields a higher attendance rate. Care must also be taken
that the vote counting system does not compromise the reliability
of the system or the independence of the protection systems.
Previous studies on the impact of IEC 61850 [5], in particular
of so-called process bus systems, on protection and control
reliability have shown that a generic system architecture based
on merging units explicitly synchronized via an external standalone source and communicating via Ethernet networks would
drastically reduce overall protection system reliability by an order
of magnitude compared with today’s microprocessor-based
relays. This is because of substantial increase in the total part
count and complexity of such a distributed system as compared
with today’s integrated microprocessor-based relays.
A successful process bus system architecture, where individual
copper wires are replaced with fiber optic communications, must
be driven by the needs of the underlying application – in other
words the architecture must be fit for purpose of protection and
control. The system architecture would therefore have to keep
the total part count and complexity at the level of today’s relays
in order to maintain the current expectation for overall system
reliability.
3. A Protection and Control System with
Enhanced Security and Dependability
3.1 Overview
HardFiber™, the protection and control system presented
in this paper, is based on an architecture that incorporates
application-driven requirements for performance, maintainability,
expandability and reliability. This is achieved through the use of
remote I/O devices to collect CT/VT and status signals, and to
output CB/process control signals [2,5]. In the presented system,
these remote I/O devices (Bricks), fulfill the role of IEC 61850
Figure 2.
Brick - rugged outdoor merging unit
Figure 1.
HardFiber process bus architecture
70
Enhanced Security and Dependability in Process Bus Protection Systems
extremes, shock and vibration, electromagnetic compatibility, sun
exposure, pressure washing and exposure to salt and other harsh
chemicals (Figure 2).
Each Brick contains four independent digital cores, each composed
of a microcontroller with individual bi-directional (bi-di) fiber links.
Each core provides dedicated point-to-point communications
with a single relay using messages conforming to IEC 61850-8-1
(GOOSE) and IEC 61850-9-2 (Sampled Values). These digital
cores share the analog core’s common input/output hardware,
implementing a fail-safe hardware design strategy that ensures
total signal isolation and independence of the digital cores.
Cross connect panels are used to land and organize the fiber
cables to the relays and Bricks, and to distribute and individually
fuse the DC power to the Bricks (Figure 3). Standard patch cords
are used to accomplish “hard-fibering”, making on a one-to-one
basis all the necessary connections between the relay ports and
Brick cores as dictated by the station’s physical configuration,
without the use of switched network communications (Figure 3).
Figure 3.
Fiber communication cross connect panel
merging units [1]. The IEC 61850-9-2 sampled value output of
each Brick and the IEC 61850-8-1 control for each Brick are
communicated via pre-terminated fiber cables to a cross connect
panel that connects the Bricks to the appropriate relays.
Readers already familiar with the HardFiber system may wish to
advance over this overview and the following architecture section,
to the Design Considerations for Process Bus Device Reliability
section.
The system is currently implemented on the existing GE Multilin
Universal Relay platform, which supports all typically required
applications. An option module provides each relay with eight
optical fiber ports so the relay can directly communicate with
up to eight Bricks (Figure 4). These maximum connectivity
numbers have been selected upon careful analysis of substation
topologies and required data traffic patterns [6]. As such, the 8/4
connectivity (each relay can communicate with up to 8 Bricks,
and each Brick can communicate with up to 4 relays) covers most
typical applications. Each relay provides protection for one zone,
conforming to established protection philosophies. It receives the
signals to perform its function over secure and dedicated direct
hard-fibered links to each of the associated Bricks. The completely
deterministic data traffic on these dedicated links allows the use
of a simple and robust method for sampling synchronization
whereby each relay controls the sampling of the connected Brick
digital cores over the link without relying on an external clock or
time distribution network.
In reference to Figure 1, the system includes Bricks mounted at
the primary apparatus, relays mounted in the control house, preterminated cables, and fiber cross connect panels for patching
fiber connections from Bricks to relays.
The Bricks are designed to interface all signals typically used for
substation automation and protection as close to their respective
origins as practical, including AC currents and voltages from
instrument transformers, breaker status and alarms, breaker trip/
close control, disconnect switch status and control, temperature
and pressure readings, and so on. The Bricks are designed for the
harsh environments encountered there, including temperature
Brick order code
Figure 4.
Connections on a UR-series relay
Brick inputs and outputs
Connector D
AC currents
Connector C
Connector B
Contact outputs
AC
voltages
Contact
inputs
RTD/TDR
inputs
SSR
Latching
Form-C
---
---
3
18
4
1
2
---
8
---
3
18
4
1
2
4
---
4
3
18
4
1
2
---
4
4
3
18
4
1
2
1A
5A
BRICK-4-HI-CC11
8
BRICK-4-HI-CC55
BRICK-4-HI-CV10
BRICK-4-HI-CV50
Table 1.
Brick process I/O capacity
Enhanced Security and Dependability in Process Bus Protection Systems
71
To improve reliability and to facilitate design, construction, testing
and maintenance, the system is designed to be as simple and
modular as possible. Bricks are designed such that they have
no stand-alone firmware, individual configuration files, or data
processing algorithms; their sole function is to be a high-speed
robust IEC 61850 interface to the switchyard, a media and
protocol converter. The system “configuration”, in this case the
specific mapping of relays to their associated Brick digital cores, is
handled purely in the physical domain through the provisioning of
individual dedicated fiber optic connections.
each Brick to all of the connected relays, and all valid commands
are accepted from the connected relays. The relays themselves
determine which subset of the received collection of signals will be
consumed internally for protection algorithms and logic schemes.
Similarly the relays determine which specific commands are sent
to which Bricks. Firmware management is exactly the same as
with relays today; the Brick digital cores inherit whatever firmware
is required from the connected relay.
All architectural decisions have been made based on recognizing
present technology and its current momentum as well as making
practical tradeoffs. For example, the cost of implementing four
independent cores in a Brick is negligible compared with the gain
of simplicity and independency of relays in the system. Similarly,
the cost of point-to-point connectivity is acceptable given the gain
of avoiding active network devices and ability to perform system
maintenance and isolation. [7]
The example in Figure 5 illustrates the architecture of the
system. A second system not shown provides a completely
redundant protection system, and may or may not use a different
technology.
The point-to-point communications architecture provides a major
dependability and security advantage over packet switched
network architectures. The lack of switches, and their associated
failure mechanisms provides the dependability advantage.
Although the system dependability problems associated with
switches may be largely overcome through switch redundancy,
the redundancy adds problems in terms of system testing, and
increases the number of failures that do not impair dependability
but must be attended to nonetheless. It is important to note that
the total number of transceivers in the presented and in a switched
architecture is comparable due to the limited number of Bricks
a relay needs to interface to in a practical switchgear topology.
The direct relay to Brick communications architecture, without
intermediate switches, makes this process bus architecture
essentially immune to cyber security threats as there is neither
need nor mechanism for external access.
The configuration for individual protection applications is relaycentric, exactly as it is today. All process inputs are always sent from
3.2 Architecture
In this example, duplicate Bricks are employed on each circuit
breaker and on each bank of voltage transformers. Each circuit
breaker Brick (numbers 1, 2, 5, 6, 9 and 10 in the figure) acquires
the three-phase bushing CT signals on each side of the breaker,
breaker position and alarm contacts, as well as outputting breaker
trip and close contacts. The Voltage Transformer Bricks (numbers
3, 4, 7 and 8 in the figure) acquire the three phase VT signals and
line disconnect positions, as well as outputting line disconnect
open and close contacts.
As is apparent from this figure that to perform their protection
function, the relays need to interface with several Bricks installed
at different locations within the switchyard. For instance, the
D60 line distance protection relays need to communicate with
Bricks on two breakers and on one VT. For this reason, the relay
has eight optical fiber ports, allowing each to connect to up to
eight Bricks. Conversely, Bricks will need to interface with several
different relays. For instance Brick 5 on the center breaker needs
to communicate with the zone protection relay on each side of the
breaker and the breaker failure relay. Thus Bricks have four digital
cores, each of which can communicate exclusively with one relay.
Fiber connections to all the process bus ports of all the relays and
all the digital cores of all the Bricks are brought by indoor and
Figure 5.
Example system illustrating the architecture
72
Enhanced Security and Dependability in Process Bus Protection Systems
outdoor multi-fiber cables to cross connect panels. At the cross
connect panels, each fiber of each cable is broken out to an LC
type optical coupler. Patch cords then interconnect Brick digital
cores to relay ports according to the functional requirements
and physical configuration of the station’s power apparatus.
Thus continuous and dedicated point-to-point optical paths are
created between relays and Bricks, without switches or other
active components. This patching or “hard fibering” is what gives
the HardFiber system its name. This hard fibering approach takes
advantage of the fact that a relay needs to talk to only the few
Bricks that have input or outputs related to that relay’s function,
that only a few relays need the I/O of any given Brick, and that the
necessary relay-Brick connections change rarely, only when the
station one-line changes. For those few instances where more
than four Brick digital cores are required, for instance for VTs on
a large bus, additional Bricks can be installed sharing the same
copper interface to the primary apparatus.
Figure 6 provides an expanded view of a portion of the example
system. In this example, digital cores from Bricks 1, 3, and 5 are
connected to the D60. A single digital core in Brick 5 is connected
to the C60, and digital cores from Bricks 5 and 9 are connected to
the L90. Note that the choice of specific cores and specific relay
ports is arbitrary – Brick cores are functionally identical, as are the
relay HardFiber ports.
The various relay protection and measuring elements that use AC
data from multiple Bricks must have the currents and voltages
at various locations sampled at the same instant. The existing
method for determining the time of the samples is maintained.
Each relay contains a sampling clock that determines when it
needs samples to be taken. In the case of the UR this clock is phase
and frequency locked to the power system quantities measured
by that relay, although other sampling schemes are possible by
other relays. At each tick of the sample clock, a GOOSE Ethernet
frame is sent by the relay to each Brick digital core connected to
that relay. Digital cores sample the measured quantities on receipt
of each frame. As the digital cores are fully independent, different
relays may sample at different rates or with different phase, but
as each is connected to different and independent digital cores,
there is no conflict. Thus each relay is able to sample in a fashion
optimal for its application, independently from other relays, and
no external clocks are required.
Figure 7.
Brick digital cores sampling asynchronously
4. Design Considerations for Reliability
of Process Bus Devices
4.1 Component Quality and Environmental
Compatibility
As in the case of any device, proven design rules, tempered with
practical in-service experience, should be followed when designing
protection devices and systems. This includes component selection,
circuit synthesis and analysis, thermal modeling, mechanical
design and so on. In this respect, designers are bounded by the
commercially available components and tools. Reliance on special
components or tools is not recommended – such solutions are
typically cost-prohibitive, have limited selection and number of
suppliers, and face availability and life cycle management issues.
Figure 6.
Hard-fibered cross-connection of Bricks and relays
Enhanced Security and Dependability in Process Bus Protection Systems
73
4.2 Fail-Safe Design
One important aspect of the security of protection is a fail-safe
response to internal or external relay problems. With respect to
internal problems the ideal outcome is that a single component
failure should not cause the protection to issue a spurious
command. Moreover, such a failure should be detected by
internal diagnostics and alarmed so that corrective actions can
be taken. With respect to external factors, protection systems
are expected to withstand certain environmental conditions from
electromagnetic compatibility requirements, through weather
conditions, to mechanical stresses. Furthermore, if a protection
system fails due to external factors, it is expected to fail-safe: not
to issue spurious commands, or to report false values or status.
From the technical point of view, the fail-safe response of protection
systems to avoid unwanted outcomes is arguably the most
important factor in the design of such systems. The task is not to
attempt to design electronics that will never fail, as all electronics
will eventually fail – rather the task is to design a system with good
overall reliability meeting established reliability targets that will be
able to detect component failures and fail in a safe way.
The fail-safe aspect of protection security is achieved by building
internal monitoring and redundancy. Internal relay parameters are
measured and compared against their expected values. Monitoring
a supply voltage to operational amplifiers comprising analog filters
in a relay is a good example. Similarly, some subsystems can be
duplicated or some operations performed multiple times in order
to detect discrepancies and respond to internal failures. Using two
duplicated AC measurement chains within a single relay as a good
example of internal redundancy.
Care must be taken when adding monitoring circuits or duplicated
subsystems, not to complicate the device and impair other
aspects of its performance. Simply adding more monitoring
circuits will in fact adversely impact the reliability of the device
and the availability of the overall protection. The monitoring
systems introduce additional components and therefore new
failure modes, while driving up the attendance rate. The adverse
impact of extra hardware to provide for monitoring or redundancy
can be avoided by designing systems that are inherently simple.
Keeping the design simple with minimum number of components
and using good engineering design fundamentals optimizes
performance for dependability and security, the initial design is
better, manufacturing is less prone to errors and quality issues,
and reliability is higher due to lower component count.
One important observation with respect to fail-safe design is the
distinctly different response of analog versus digital systems to
internal failures. Digital systems tend to fail gracefully in that they
yield a solution that either works or it stops functioning in a selfevident way. Monitoring is based on a finite set of cases that can
be explicitly tested through such mechanisms as watchdogs and
data integrity checks where pass/fail criteria are very clear and
testing can be built into the system itself.
Analog circuits on the other hand are prone to malfunction
in ways that are not explicitly clear. A change in an analog to
digital converter (ADC) reference supply for example will appear
as a change in the signal being sampled. If this signal is used by
a sensitive protection element such as current differential, this
may result in an unwarranted protection operation. It is therefore
necessary to provide additional monitoring for such analog
systems.
74
It is important to note that at the lowest level, all circuits are
analog. For example, if the supply voltage to a microcontroller
drops for a short period, the microcontroller may enter a state
where it will not behave as designed. Therefore, it is prudent to
monitor parameters of digital circuits to make sure they behave as
digital circuits, and count on inherent properties of digital circuits
to ensure fail-safe operation.
Yet another aspect of design for reliability and fail-safe operation
is the proper identification of working conditions: temperature,
EMC levels, mechanical factors, humidity, and so on. With
respect to protective relays, well-established standards exist
and are followed. These standards are verified by the significant
installed base and performance history of microprocessor-based
protective relays. However, with respect to devices mounted in
close proximity to the primary apparatus the industry experience
and installed base is limited, so environmental standards for these
types of systems will likely need to evolve to take into account new
in-service experience.
The following sections will explain how the above principles have
been applied to the new protection and control system described
in this paper.
5. Design Implementation for
Component Reliability
5.1 Outdoor Communications Cable
Mechanical Design
The optical fiber is packaged in a rugged cable to ensure
mechanical survival. The cable is designed to meet United States
Department of Defence (DoD) standard MIL-PRF-85045 for ground
tactical cable, and is suitable for indoor and outdoor installation in
direct burial trenches, common use cable pans/raceways/ducts,
and when exposed to direct sun and weather. Figure 8 shows the
cable cross section, with its multiple layers of protection. Figure 9
shows an actual direct burial installation in progress.
Outer jacket
Ripcord
Central strength member
Inner jacket
Strength member
Rodent barrier
1
C
2
C
4
Drain wire
Aluminum/Mylar tape
Copper conductor
Ripcord
Buffered fiber
Sub-unit strength member
Sub-unit jacket
Figure 8.
Cross section of outdoor communications cable
To ensure dependability of the connections and reduce the Mean
Time to Replace (MTTR), standard military MIL DTL 38999 [8] type
connectors are employed. This particular technology has a long
history of dependable performance in challenging environments
such as aerospace and naval applications. As such, they are an
ideal fit to be deployed in the intended installation environment
for the Bricks.
Optical Fiber
To achieve high dependability of the digital communications
between the Bricks in the electrically noisy switchyard and the
relays in the control house, optical fiber media is employed. The
Enhanced Security and Dependability in Process Bus Protection Systems
immunity of optical signals to electrical and magnetic interference
is well known. The integrity of the data, and thus system security,
is further protected by the 32-bit cyclic redundancy check code
(CRC) that is a standard part of Ethernet communications. The
same relay-Brick link is used for process data acquisition, time
synchronization, control (i.e. trip and close), and diagnostic
reporting.
contribution to failure rates. On the other hand, each link has half
as may optical fibers and transceivers as an equivalent dual fiber
Ethernet link, so overall higher reliability can be expected.
Multimode fiber with a 50µm diameter core rather than single
mode 9µm fiber is used in the fiber cables. This large core makes
the connectors much more tolerant to contamination and
misalignment problems. A misalignment a minor impact on a
50µm core could totally obscure a 9µm core.
The optical fiber cables are fully pre-terminated within a factory
environment using special purpose jigs, automated testing, and
personnel particularly trained with the necessary skills, so that
consistent, highly dependable fiber terminations are obtained.
Conversely, fiber splicing/termination performed on-site in the
field would be done in an environment that may be non-ideal in
many respects, leading to instances of poor fiber dependability.
Embedded Copper
Figure 9.
Outdoor communications cable in direct burial application
To further increase system reliability, a single bi-directional (bi-di)
fiber is used for each link, as one fiber using a single transceiver
at each end will be more dependable than two fibers and two
transmitters and two receivers. Course Wavelength Division
Multiplexing (CWDM) is used for full duplex communications on a
single fiber, such that the signals in one direction use a different
wavelength (frequency) from that of the signals travelling in the
opposite direction, so that they may be discriminated from each
other. As shown in Figure 10, wavelength selective mirrors in each
transceiver reflect the incoming light to the receiver, while allowing
the outgoing light to pass straight through and out. The physical
Ethernet interface is implemented according to standard IEEE
802.3 2005 100Base BX bi-directional fiber optic communications
[9].
The outdoor communications cable also embeds a pair of copper
wires along with the fiber optic cables. These copper wires are
individually fused at the control building end and are used to
provide a reliable power supply to each Brick. From a reliability
standpoint, each cable becomes a single unit connection for
each Brick rather than relying on two separate connections to
independently provide power and communications access. The
end result is improved reliability of the overall system. In other
words, if the Brick is powered, it will be able to communicate and
conversely if the Brick communications are connected, the Brick
will be powered.
There is an additional rationale to include the power connection
for the Brick along with the communication cable, in that there
may not be a convenient uninterruptible supply where a Brick
is to be installed. For example, a Brick located on a VT structure
will have no local DC station battery supply. Self-powering a Brick
from the connected AC signals is not advisable, as outage or fault
conditions will render the Brick, and potentially the upstream
protection application, unavailable.
5.2 Brick
Bricks are intended for direct installation on or in proximity to
primary equipment without requiring additional environmental
protection. The solution is to provide a common chassis to clad
the electronics, provide EMC shielding, provide weatherproofing,
and act as an overall heat sink for the heat-generating electronics.
This opens new design opportunities in terms of mechanical and
thermal design.
Mechanical Design
Figure 10.
CWDM bi-di optical link
This bi-directional fiber technology has had substantial successful
experience in FTTX (fiber to the premises) applications, particularly
in Japan, so the technology is well established and designed for
high reliability in potentially harsh environments. The transceivers
used to implement the bi-di links have passed all the shock,
vibration, temperature cycling and accelerated aging tests
described elsewhere, and are thus expected have a negligible
Shock and vibration requirements for installation on breaker
structures and in close proximity to power transformers call for a
sturdy and heavy chassis. Cast aluminium is a rational choice for
both mechanical and thermal reasons. EMC and environmental
requirements call for a cage-like design with minimum number
of openings with overall aperture areas minimized. The openings
need to be hermetically sealed to guarantee performance at IP66
(dust tight, pressure washing), and limit the air gap for better EMC
immunity.
Heavy components, such as the AC input isolating transformers
and the integrated power supply need to be mechanically secured
by encasing them as separate integrated components and
mounting them directly to the Brick chassis.
Enhanced Security and Dependability in Process Bus Protection Systems
75
Thermal design considerations for critical heat-generating
components are accounted for by mounting these components
such that they use the chassis itself as a heat sink. At the same
time, Bricks mounted outdoors are exposed to so-called insolation
or “sunloading” effects – heating due to absorbed solar energy.
This calls for striking a balance between the ability to transport
the heat out of the chassis from the hot components inside,
and minimizing the absorption of the radiated solar energy and
heating effect on the components inside the chassis. A black
matte finish maximizes heat radiation out of the Brick, but also
maximizes absorption of solar radiation. A bright reflective finish
minimizes absorption, but also minimizes radiation. This effect is
not new and in fact standards such as IEC 60068-2-9 [10] and
MIL-STD-810F [11] exist to address these issues. These standards
call for over 1,000 W/m2 of solar radiation under an ambient air
temperature of 50 degrees C. A relatively small amount of heat
needs to be to dissipated, so a reflective finish is used. In order to
ensure adequate sunload immunity under worst-case scenarios
where the Brick chassis is dirty or otherwise non-reflective, solar
testing was also performed with the specimen Brick painted matte
black.
The Brick needs to be designed for both hot and cold temperatures.
Under cold temperatures mechanical properties of fiber and cold
temperature start-up sequencing of the electronics become a
concern. Necessary design considerations include avoiding fiber
connections by using pigtails embedded in transceivers, securing
the fiber mechanically and testing under combined temperature
and mechanical conditions.
The tests shown in Figures 11, 12 and 13 show environmental
tests for dust ingress, water ingress and sunload effect.
Figure 13.
Solar loading tests at the solar laboratory: Bricks mounted next to light
source (left), solar radiation test in progress (right)
Connectorized Interfaces
Terminations of the remote I/O device (Brick) are pre-connectorized
using standard off-the-shelf MIL-DTL-38999 connectors. These
connectors are available from multiple vendors and are used
in naval, military, avionics and industrial applications because
of their high performance and reliability including shock and
vibration, IP rating, proper electrical parameters and extremes of
temperature.
By selecting an extremely rugged and reliable connector
technology, a number of design challenges for implementing
reliable remote I/O devices are solved simply with a widely
accepted hardware solution. The Mean Time to Replace (MTTR)
is reduced by reducing Brick replacement time without the need
to re-commission physical connections. Adoption of standard,
reliable and proven connectors benefits from millions of unit years
of field experience, lessons learned, independent verification tests
and other aspects relevant to the overall reliability of the system.
An AC input connector brings the secondary currents into the
Brick. A danger exists of opening the connector with the CTs live.
Development of a self-shorting mechanism embedded in the
connector would prevent the use of standard connectors, and
would jeopardize reliability of the connector by adding moving
parts. Following the principle of a simple design, the presented
system uses a simple mechanical feature to prevent accidental
disconnection of a live CT cable: the yellow collar shown in Figure
14 below. This is a good example of a design to avoid adverse
impact on reliability of the system.
Figure 11.
Dust ingress testing for IP6x: pre-test (left) and post-test (right)
Figure 12.
Water ingress testing for IPx6: pre-inspection (left) and post-inspection
(right)
Figure 14.
Safety device protecting against accidental opening of a live CT secondary
circuit
76
Enhanced Security and Dependability in Process Bus Protection Systems
Electronics
The Brick power supply is a good example of how a simple design
can improve reliability by eliminating the need for electronics
that have an adverse impact on reliability. A common source
of failures in microprocessor relays is their power supply, in
particular electrolytic capacitors used in the power supply to
provide “hold-up” so a momentary interruption of the DC station
battery does not force the relay to have to restart, a process that
may take many seconds. Regardless of the grade of component
selected (automotive, industrial), electrolytic capacitors have been
repeatedly shown to be a reliability concern.
The simplicity of the Brick allows it to be fully functional within
milliseconds of being energized as opposed to several seconds
as with relays. By having the Brick start so quickly, the availability
of protection from being momentarily powered down is virtually
unchanged so there is no need to provide hold up in the power
supply. This eliminates the electrolytic capacitors from the power
supply, and reliability issues that they introduce.
Contact Inputs
The Brick contact inputs use dry contacts, with a 24 VDC sensing
voltage provided by an internal isolated wetting supply generated
by the Brick. The low wetting voltage allows a low sensing circuit
impedance and a high wetting current. Low input impedance
makes the Bricks highly immune to incorrect status indication due
to induced transients or insulation degradation in the external
contact wiring. High wetting current assists contact “wipe”
action in obtaining a clean contact by burning off contact surface
contamination. Having wetting supplies isolated independently
for each Brick prevents station battery grounds and grounds in
other Brick’s contact input circuits from causing incorrect status
indication. The Brick contact input supplies are also designed so
that two Bricks may be paralleled across a single dry contact,
buffered such that the failure of one Brick will not adversely impact
the operation of the remaining Brick.
Control Outputs
The Brick contains four solid-state relay (SSR) outputs, based
on an existing highly tested and field-proven design, to directly
interrupt typical breaker trip and close circuit currents. The SSR
outputs were chosen with no moving parts so that mechanical
x1
+
_
xK
vibrations caused by a breaker mechanically operating could not
cause a spurious or undesirable contact closure. The SSR outputs
themselves are thermally bonded to the cast aluminium shell of
the Brick, so the entire chassis acts as a heat sink improving the
life expectation for the SSR outputs.
Analog Inputs
The self-testing involved with AC and transducer inputs is worth
further discussion.
Two ADCs are employed on each AC input, as shown in Figure 15.
The input to the high range ADC is scaled to accept high currents
without saturating. The input to the low range ADC is scaled
for accuracy at low currents, but clips for currents much above
nominal. At each sample instant, the programmable logic device
(PLD) starts a conversion on both ADCs, and coordinates sending
both conversion results to the digital cores. Microcontrollers in
the digital cores use the low range ADC value if not saturated,
otherwise they use the high range ADC value. Part of the reason
this is done is that available monolithic analog to digital converters
(ADCs) with sufficient speed to sample at the rates demanded by
today’s relays cannot provide the necessary metering accuracy
without saturating at high fault currents.
The other reason for dual ADCs on each input is that it provides
a unique opportunity for continuous self-testing of the AC input
hardware.
Phase CT current inputs are at most times at a level where the
waveforms are continuously in the low range. However, except
near zero crossings the high level ADC retains sufficient accuracy
that there ought to be no significant difference between the high
and low range values after accounting for the design scaling
differences. Any significant difference indicates a failure in one
range or the other. The comparison is made on a sample-bysample basis, so protection can be disabled before the invalid
data is consumed. Disabling protection on detection of invalid AC
input data is an immediate security benefit. It also has a positive
impact on dependability, as it alarms triggering rapid corrective
action.
In addition to dual ADCs, dual anti-alias filters and dual input
conditioning and gain stages are provided so that problems
in these areas are also detected by the same comparison. The
Anti-alias filter
High Range Channel
LPF
16-bit
ADC
digital
core
Anti-alias filter
Low Range Channel
digital
core
LPF
16-bit
ADC
PLD
Figure 15.
Typical Brick AC input hardware block diagram
Enhanced Security and Dependability in Process Bus Protection Systems
digital
core
digital
core
77
low range uses the entire secondary winding of the isolating
transformer, while the high range uses only half of the secondary.
Thus any trouble in the isolating transformer secondary will be
detected as well. The comparison is made in the digital cores,
providing detection of troubles getting the data through the PLD
and into the microcontrollers. Thus the entire AC input hardware
circuit is covered, with the exception of the isolating transformer
primary, which for CT inputs consists simply of either one or five
turns of heavy gauge wire.
Figure 16 illustrates the comparison process. The low (blue) and
high (red) range scaled readings are shown in the top graph; the
selected signal in the middle graph, and the error flag in the bottom
graph, for the case of a 1.2 times nominal current experiencing a
negative 20% error in the high range channel.
5.3 Relay
In the presented system, the relay (IED) portion transforms from
a mixed analog and digital device to an almost exclusively digital
device as all analog inputs and outputs are eliminated and replaced
with digital communications. As a result, the relay portion of the
system becomes much more reliable and intrinsically fail-safe. The
reliability gain is achieved by reducing the individual part counts,
and by using digital components that tend to be more reliable due
to their highly integrated packaging and pre-determined failure
mechanisms. The fail-safe improvement is caused by the nature of
digital systems as explained in the previous section. The input and
output signals are moved between different digital subsystems of
the relay (receiving transceivers, communication processor, digital
signal processors, main logic processor) as digital packages with
embedded data integrity protection (CRC, check-sum). All digital
components are properly engineered (monitored rails, watchdog,
code integrity checks, etc.) yielding a very robust system.
Communication-based protection and control systems require
more design attention in terms of dealing with permutations of
various conditions related to distributed architectures, multiple
devices and communication traffic and impairments. Development
of distributed architectures is not a new enterprise - there is a
great deal of experience accumulated from engineering internal
architectures of modular relays and engineering and application
of digital line current differential or distributed bus differential
relays.
Figure 16.
Simulation of a high range gain error on a CT input
When the input signal is so high that the low range clips the peaks,
as happens with normal values of VT inputs and on CT inputs
during faults, the comparison can still be made in the vicinity of
zero crossings. Figure 17 shows a case of a 70V voltage losing
reference in the high range channel. The trouble is detected, but
after a short delay of up to ½ cycle.
In the case of the particular design presented in this paper, the
relays are built using an existing relay platform with a decadelong field record, the maturity of a large portion of hardware and
firmware being carried over to the new system. This modular relay
series uses just one new hardware module to interface with the
remote I/O (Bricks) and keeps the existing power supply, CPU and
teleprotection modules intact. Even the existing firmware is used –
the I/O data are seamlessly integrated with the rest of the relay.
In this way some 80% of the relay hardware is not changed, and
some 95% of the relay firmware is not changed. Moreover, the
new relay is not a new model or variant of the existing relay, but
an option on the existing platform. In this way even the impact of
regression during development of the code has been drastically
reduced as only a small amount of new code was added to the
existing proven relay firmware.
In this way a significant portion of the field experience in terms
of hundreds of thousands of unit years of run time, independent
testing of hardware and algorithms, and exposure to actual system
conditions can be instantly assigned to the relay and application
portion of the new protection and control system.
5.4 Extensive self-testing
Figure 17.
Simulation of a loss of high range reference supply on a CT input
The described protection system implements extensive selftesting in order to reduce the repair response delay thus
increasing availability and dependability, and to flag unreliable
data so that it is not used to make operational decisions thus
increasing protection system security. It is critical that internal
self-monitoring within the system overlaps adequately to ensure
complete end-to-end self-testing. The concept of overlapping test
zones is shown in Figure 18.
It can be seen that there are a number of self-test zones within
the system, each individual self-test zone overlaps with at least
one other zone.
78
Enhanced Security and Dependability in Process Bus Protection Systems
Remote I/O (Brick)
Fiber
Comms
•
Relay
The relay itself also runs a complete set of internal self-tests as
per established relay design philosophy, including processor
watchdogs, program execution and internal data integrity
checks.
Digital
Core
Process Bus Interface
Hardware Buffering
Common I/O Hardware
Digital
Core
Digital
Core
CPU
Digital Systems
Figure 18.
Overlapping test zones within the system
The specific self-tests within the Brick include the following:
•
On each AC and each transducer/RTD input, correspondence
of the high range analog to digital converter output with that
of the low range analog to digital converter.
•
Proper sequencing of the analog to digital converters through
monitoring of their Busy status output.
•
Temperature of the analog to digital converter.
•
Output levels of each of the internal power supplies.
•
Watchdog timers, one external to each microcontroller, which
together with firmware in the microcontroller verifies all tasks
are running as designed.
•
Microcontroller flash memory integrity through CRC-16
checking.
•
Communications between the analog core and each digital
core in each direction are uninterrupted and each frame’s
CRC-16 shows data is uncorrupted.
•
Match in clock frequencies between the analog core clock
and each digital core clock.
•
Correspondence of each contact output driver to the received
command for that contact.
•
Correspondence of an auxiliary contact on the latching
output relay to the latching output command.
•
Monitoring of the voltage across and current through each
SSR contact output, used by the relay to monitor the health
of the external control circuit.
The tests of the fiber optic communications include:
•
Also, it is important to note the dividing line between analog and
digital systems. The amount of the system that is based on analog
circuitry is limited to the interface to the primary power system
process. The rest of the system is entirely digital, allowing for this
high degree of built-in error checking and diagnostics.
5.5 Continuous Monitoring
Digital
Core
Analogue
Systems
Integrity of each frame sent and received on the optical fiber
links through CRC-32 checking, lost frames are detected with
sequence numbers.
Signal quality of each of the optical fiber links, including the
send and receive optical light levels at each end, transmitter
bias current, and transceiver temperature.
In order to ensure the security of protection, each Brick core
continuously monitors its key internal subsystems including the
common hardware (ADCs, output relay circuits) and the status
of the core itself. Each core includes this diagnostic information
with each set of samples transmitted to the connected relay. In
the event of a failure of an internal diagnostic test, the connected
relays are made aware instantly and can then ensure that the
overall protection system will fail to a safe state.
Additionally, each core continuously monitors the optical transmit
and receive power from the associated transceiver and sends
this data to the connected relay. Each transceiver in the relay
measures the same quantity and then calculates the respective
power link budget for the connection. In this case, troubles related
to the degradation of the optical communications path can be
determined early and explicitly.
5.6 Duplicated I/O Hardware
The protection system presented in this paper provides the user
with the ability to control the dependability and security of the
system by supporting duplicate Bricks (the remote I/O modules),
as illustrated in Figure 5. The primary sensors for signals critical to
system reliability can also be duplicated, the signals transmitted
thought independent Bricks and independent optical fibers to
the relay, where a variety of options exists for reacting to loss of
communications with the Bricks, Brick self-test error conditions,
and inconsistency of the received values. The status inputs and
control outputs may also be provisioned to provide a high degree
of dependability or security.
Data Crosschecking
For instance, it is possible where CT reliability is of special concern,
to use two independent CT cores per phase, one to the CT inputs
of each of two duplicated Bricks. Alternatively, a single CT core
can drive both Bricks. The elimination of long runs of CT wiring
back to the control house results in virtually zero external burden
on the CT, reducing voltage stress on the CT secondary, and
thus increasing the already high CT reliability. This of course also
decreases the CTs propensity to flux saturation. Each of the two
Bricks independently samples its input CT signal, converts the
samples to digital form, and sends the digital samples back to the
relay over independent optical fibers.
VT inputs are handled in the same fashion, so the term AC input is
used here to indicate either a CT or VT input.
At the relay, user settings control how the two streams of samples
from the two AC inputs are combined into a virtual AC bank, which
Enhanced Security and Dependability in Process Bus Protection Systems
79
X2
Freeze
Discrepancy
Detection
Freeze
Discrepancy
Detection
A1
d
Declare
Discrepancy
Va
lu
e
s
A3
te
This simple auto-transfer scheme is illustrated in Figure 19, where
the top group of traces are the samples from origin 1, the next
group are from origin 2, and the bottom group are that of the AC
bank. It is important to note that the AC bank is unaffected by all
except the loss of both origins.
the output of the two CT cores may differ to some degree due
to manufacturing tolerances or unequal burdens. To prevent
unwarranted discrepancy declarations, a restraint characteristic
is applied under normal (i.e. load) conditions, and discrepancy
checking is suspended when currents are much greater than the CT
nominal rating. The discrepancy checking algorithm characteristic
is illustrated in Figure 20. The feature of suspending discrepancy
check is not applied for VT inputs, but for CT inputs only.
A2
Ex
pe
c
is used by the relay’s internal functions operating on that CT/VT
signal. Each AC bank has two settings that select the three-phase
Brick AC inputs used for that bank. The Origin 1 setting selects the
primary source for AC inputs that are to be used provided that the
corresponding Brick is enabled, communications with the Brick
are intact and correct, and the Brick reports no internal self-tests
errors. The Origin 2 setting selects the AC inputs that are to be
used when Origin 2 AC inputs are available and Origin 1 AC inputs
are not available. If neither is available, the AC bank samples are
forced to zero, as this corresponds to the normal failsafe state in
traditional protection and control.
A2
-A1
-A3
-A1
A2
A1
A3
X1
Declare
Discrepancy
Ex
pe
ct
ed
Va
lu
e
s
A2
-A3
Freeze
Discrepancy
Detection
Freeze
Discrepancy
Detection
Figure 20.
X1 and X2 are the sample values from origin 1 and origin 2 respectively.
A1, A2 and A3 are parameters, values here exaggerated for clarity.
The discrepancy check algorithm can be enabled in one of two
modes: Dependability Biased, or Security Biased. These differ in
the reaction of the relay to the unavailability of just one of the
duplicated Bricks.
Figure 19.
Oscillography illustrating seamless auto-transfer on Brick trouble
To allow user control of the dependability/security trade-off when
duplicated Bricks are provided, an AC bank crosschecking setting
is available. This setting allows the user to enable comparison of
the samples from the two origins, and to control the behavior of
the protection elements in the event that the relay detects one of
the two origins is unavailable.
When enabled, the discrepancy detection algorithm compares
each sample from the two duplicated sets. If the two sets of
samples are available, and are substantially different without
clear declaration of a Brick self-test error, execution of the relay
protection elements is suspended, as at least one set of samples
is incorrect and the relay is unable to explicitly determine the valid
source with which to continue operation. Note that due to the
extensive self-checking, in virtually 100% of cases of HardFiber
equipment failure, the relay is able to determine the invalid source,
and thus can continue to operate using the other source. It is
only when it is unable to determine this that operation need be
suspended.
In applications where the duplicated CT inputs are from different
CT cores, there is the possibility that under fault conditions
80
In the Security Biased mode, protection will function only if both
duplicate Bricks are available and both sets of samples are in
agreement. The Security Biased mode is used in applications
where the probability of a fault is relatively low and the system
impact is very large, so it is important that protection not operate
incorrectly due to AC measurement error – in other words it
is justified to require two independent measurements be in
agreement to operate.
In the Dependability Biased mode, discrepancy checking is only
performed when sets of samples from both duplicate Bricks
are available. The discrepancy check is therefore declared and
protection blocked only if both Bricks are available and the sets
of samples are substantially different. The clear unavailability of
a single Brick (loss of communications, internal self-test failure)
does not block protection. The Dependability Biased mode is used
in applications where the probability of fault occurrence is higher
and the system impact is lower, or in instances where a failure to
trip may result in unacceptable damage to a major power system
element like a generator. It is important therefore that protection
is highly available – in other words it is more desirable to allow
protection to continue to operate without two independent
measurements.
Enhanced Security and Dependability in Process Bus Protection Systems
It is also possible to use duplicated Bricks without running the
discrepancy checking. By not choosing either the Dependability
or Security Biased mode, the relay will use the samples from
origin 1 exclusively. In the event that the origin 1 Brick becomes
clearly unavailable, the relay will simply switch to using the origin
2 samples.
Origin 1
status
Origin 2
status
Discrepancy
check
Available
Available
Available
Crosschecking setting
Dependability
biased
Security
biased
None
OK
Protection
available
Protection
available
Protection
available
Available
Discrepant
Protection
suspended
Protection
suspended
Protection
available
Unavailable
Available
Not relevant
Protection
available
Protection
suspended
Protection
available
Available
Unavailable
Not relevant
Protection
available
Protection
suspended
Protection
available
Unavailable
Unavailable
Not relevant
Protection
suspended
Protection
suspended
Protection
suspended
Table 1.
Effect of dependability biased/security biased setting on duplicated Bricks
The effects of Dependability Biased verses Security Biased
crosschecking modes are illustrated in Figure 21. In both graphs,
the top two groups of traces are the samples from the two Bricks,
the next group is the samples from the AC bank, and the bottom
two traces are the pickup of an instantaneous and a timed
overcurrent protection element respectively. In each graph, the
origin 1 Brick samples become discrepant shortly after the start
of a simulated fault.
the origin 1 samples, then timed protection times out on origin 2
samples; protection operation is dependable in spite of the Brick
failure. The Security Biased mode was used for the bottom graph
(b), so the instantaneous protection picks up as above, but then
protection is suspended when the Brick fails, halting timeout of the
timed protection, thus securing the protection.
Although in this particular case it is clear from the traces that
the discrepancy was most likely in the origin 1 samples so the
auto-transfer to origin 2 and subsequent tripping was the correct
decision, in other cases the choice might not be so clear even to
an experienced engineer looking at the post-mortem data, and
especially difficult for a relay in a real-time operation.
Note that discrepancy checking is only done when both samples
are received, and neither of the Bricks involved indicate any self-test
error that calls into question the validity of its AC input data. The
Brick self-tests are intended to detect internal Brick troubles
that could corrupt AC input data, and so two available samples
should only be found discrepant when the Bricks are otherwise
functioning normally. Thus assuming failure independence, in
the Dependability Biased mode, the unavailability of protection
approaches the square of the unavailability of a single Brick. In the
Security Biased mode, the availability of protection approaches
the square of the availability of a single Brick.
Duplicate Input/Output Hardware
In addition to supporting duplication of AC inputs, the described
system also supports duplicated contact inputs and outputs.
As there are many different schemes to make use of duplicated
contacts, user programmable logic is used.
Redundancy schemes for contact inputs include:
•
OR the two contact inputs – dependable
•
AND the two contact inputs – secure
•
AND a form A contact input with the inverse of a form B
contact input
•
Main/backup contact inputs
•
Instantaneous on both contact inputs closed, delayed on
single contact input closed
•
Last state where both contact inputs agreed
•
Two of three, majority logic
Implementation logic for a main/backup scheme is shown in
Figure 22. Here a main and a backup contact of a transformer gas
relay are connected to main and backup Bricks. The protection
(not shown) uses the main contact input provided it is available
(Main Brick On), otherwise uses the backup contact. The scheme
annunciates trouble should the main and backup states be
available and discrepant. The alarm is delayed by 4.0ms to
allow for unequal sensor contact operating times, yet still ensure
alarming for discrepant operations.
Redundancy schemes for contact outputs include:
Figure 21.
Oscillography illustrating effect of (a) dependability biased vs. (b) security
biased modes
The Dependability Biased mode is shown in the top graph (a), so
the instantaneous protection and the timed protection picks up on
•
Two output contacts in parallel - dependable
•
Two output contacts in series - secure
•
Four output contacts in “H” configuration – dependable and
secure
Enhanced Security and Dependability in Process Bus Protection Systems
81
Figure 22.
Main/backup contact input redundancy scheme
One example of the H configuration is shown in Figure 23. This
arrangement is dependable in that it will continue to operate
should either Brick fail or any single contact output fail open. It is
secure in that it will not operate should any single contact output
fail short.
Brick 1/OUT1
Brick 2/OUT2
Brick 2/OUT1
Brick 1/OUT2
“H” Configuration
Figure 23.
A significant amount of detail is given regarding the design of this
new solution as related to security and dependability. Special
attention has been paid to the fail-safe aspect of the design. By
relying more on digital interfaces and subsystems, the system is
made more fail-safe: it either works or it stops functioning in a selfevident way, with a greatly reduced probability of a subsystem
being in a faulted yet undetected state.
The system also supports the use of duplicated I/O devices for
protection applications to achieve a high degree of reliability, and
supports applications requirements for dependability or security
improved over existing protection technology. The remote I/O
devices are shared between multiple relays and therefore the
solution is not cost-prohibitive compared to providing multiple
independent remote I/O devices for each relay.
It can be demonstrated that the system is more reliable,
dependable and secure when compared with existing solutions.
The following key elements contribute to the enhancements.
•
The total amount of hardware in the system is less compared
with traditional solutions. This is due to the shared I/O devices.
Systems with fewer components are generally more reliable.
•
The I/O devices themselves incorporate a substantial amount
of self-monitoring to detect internal problems.
•
Duplicated I/O devices are used with constant crosschecking
of input data, while supporting the use of inputs and outputs
in redundant applications. The principle of duplication can be
extended to instrument transformers providing for an extra
layer of redundancy within each of the protection systems.
•
Data is moved digitally secured with data integrity
mechanisms. Digital systems are continuously monitored
and will fail in predictable ways and self demonstrating ways
compared with analog subsystems.
Highly dependable and secure contact output scheme
6. Conclusion
This paper outlined the high-level design principles for protection
and control systems. These principles are illustrated using a new
practical solution for implementing ultra-critical protection and
control IEC 61850 process bus applications. Several applications
of these principles are explained. Moreover, the system allows
eliminating significant amount of labour and therefore reduces
costs and shortens the required deployment time.
82
Enhanced Security and Dependability in Process Bus Protection Systems
The continuous internal self-testing, crosschecking and data
integrity mechanisms will detect problems instantly allowing
the field crews to rectify the problem quickly and precisely.
This eliminates a considerable number of failures that may
remain latent in traditional protection systems as well as
failures that require a great deal of labour effort to diagnose
and repair (for example DC battery grounds).
7. References
The proposed system is easily testable and maintainable [12].
The physical provisioning of communication links using fiber
patch cords provides a clear maintenance boundary that
does not require relay maintenance personnel to deal with
potentially lethal high-energy signals.
[2] HardFiber Process Bus System Reference Manual, GE
Publication GEK-113500.
•
The continuous monitoring of the digital communications
links and the overall architecture greatly reduces the potential
for human errors to result in undesired protection operations
during testing and maintenance activities.
•
The system is free from cyber-security concerns. The point-topoint, non-routable process bus network makes it inherently
secure with no need for external monitoring mechanisms
that would otherwise create extra cost and complexity and
expose the system to external threats.
[4] FRCC System Disturbance and Underfrequency Load
Shedding Event Report February 26th, 2008 at 1:09 pm Final
Report Issued by: FRCC Event Analysis Team October 30, 2008
https://www.frcc.com/OC/Shared%20Documents/FEAT%20
Report%20-%20Final.pdf
•
•
Also, the presented system is composed of only few highly
connectorized standard devices in a modular, scalable architecture.
This makes the solution very attractive from the point of view of
initial installation as well as repair and/or reconstruction in the
event of a catastrophic event such as a natural disaster.
[1] IEC International Standard “Communication networks and
systems in substations - Part 9-2: Specific Communication
Service Mapping (SCSM) – Sampled values over ISO/IEC
8802-3”, (IEC Reference number IEC/TR 61850-9-2:2004(E),
IEC, Geneva, Switzerland).
[3] Northeast Power Coordinating Council, “Bulk Power System
Protection Criteria”, Revised January 30, 2006
[5] B. Kasztenny, J. Whatley, E. A. Udren, J. Burger, D. Finney, M.
Adamiak, “Unanswered Questions about IEC 61850: What
needs to happen to realize the vision?” (Proceedings of the
32nd Annual Western Protective Relay Conference, Spokane,
WA, October 25-27, 2005)
[6] B. Kasztenny, D. McGinn, S. Hodder, D. Ma, J. Mazereeuw, M.
Goraj, “A Practical IEC61850-9-2 Process Bus Architecture
Driven by Topology of the Primary Equipment” (42 CIGRE
Session, Paris, August 24-29, 2008, paper B5-105).
The presented system allows protection relays to be made internally
redundant and safe, making it very attractive in traditional as well
as ultra-critical applications including nuclear power plants, and
naval installations.
[7] M. Adamiak, B. Kasztenny, J. Mazereeuw, D. McGinn, S.
Hodder “Considerations for Process Bus deployment in realworld protection and control systems: a business analysis”
(42 CIGRE Session, Paris, August 24-29, 2008, paper B5-102).
•
Observing the data received by the relay over the link is
reasonable and matches other indicators. For example
indicated current/voltage magnitude and phase matches
other indicators of these same quantities.
•
Causing some change of state and observing its correct
communication over the link. For example, observe the
reported effects of initiating a breaker operation or a tap
change. Initiation may be from the operator’s HMI where it
uses the same fiber link.
[8] Detail Specification MIL-DTL-38999L “Connectors, Electrical,
Circular, Miniature, High Density, Quick Disconnect (Bayonet,
Threaded, and Breech Coupling), Environment Resistant,
Removable Crimp and Hermetic Solder Contacts, General
Specification for”, United States of America Department of
Defense, 30 May 2008.
•
The relays are designed such that when normally in-service,
they alarm and reject data on a port when the HardFiber Brick
serial number that is included with the data fails to match the
relay setting for that HardFiber Brick’s serial number. The relay
serial number value is included with outgoing commands,
and the HardFiber Bricks are designed to accept commands
only when the accompanying serial number matches its own
serial number. Thus, once the HardFiber Brick serial numbers
are correctly entered into the relay settings, the fact of normal
communications establishes that the link is correct. The serial
number setting in the relay can be manually checked against
the serial number on the HardFiber Brick’s nameplate.
Thus it can be seen that testing of the passive interconnection
system is quite simple, and that after commissioning is complete,
it can be entirely automatic.
[9] “IEEE
Standard
for
Information
technology—
Telecommunications and information exchange between
systems— Local and metropolitan area networks—Specific
requirements Part 3: Carrier sense multiple access with
collision detection (CSMA/CD) access method and physical
layer specifications” IEEE Std 802.3™-2005, clause “58.
Physical Medium Dependent (PMD) sublayer and medium,
type 100BASE-LX10 (Long Wavelength) and 100BASE-BX10
(BiDirectional Long Wavelength)”
[10] IEC International Standard “Environmental testing - Part 2:
Tests. Guidance for solar radiation testing”, (IEC Reference
number IEC 60068-2-9:1975, IEC, Geneva, Switzerland).
[11] Test Method Standard MIL-STD-810F “Department of Defense
Test Method Standard for Environmental Engineering
Considerations and Laboratory Tests”, United States of
America Department of Defense, 1 January 2000.
[12] D. McGinn, S. Hodder, B. Kasztenny, D. Ma “Constraints and
Solutions in Testing IEC 61850 Process Bus Protection and
Control Systems” (42 CIGRE Session, Paris, August 24-29,
2008, paper B5-206).
062609-v2
Enhanced Security and Dependability in Process Bus Protection Systems
83
Aim for Savings &
Energy Efficiency
Do you know how much energy your mission critical
facility loses each year?
Your current UPS may not be efficiently utilizing the power you are purchasing. GE Power
Quality has solved this problem with the introduction of the Digital Energy™ SG Series 750kVA
UPS for data center and mission critical facilities, an extension of our best in class SG Series UPS
platform. One of the most efficient and reliable double conversion UPS systems available today,
the 750kVA UPS builds upon GE’s proven SG Series product technology to stop your energy
investment from leaving the building.
The SG Series 750kVA UPS hits the target.
> Energy Savings – maximum efficiency greater than 94%
> Installed Savings – smaller footprint & lower weight
> High Reliability – high MTBF and no single point of failure
Highly efficient 750kVA UPS
g
Digital Energy
Power Quality
GE Power Quality
www.geindustrial.com/750
www.geelectrical.com/datacenters
[email protected]
Tel: 1-800 637 1738
AEP Process Bus Replaces Copper
Innovations in stations translate to more savings
in material, time and manpower
John F. Burger, Dale A. Krummen and Jack R. Abele
American Electric Power
A huge opportunity for material, time and manpower savings
exists in the reduction or elimination of substation copper control
cables. This has inspired American Electric Power’s (Columbus,
Ohio, U.S.) interest in the process bus technology.
AEP decided to evaluate a next-generation distributed protection
and control system with all process interfaces located in the
switchyard, thus taking control cabling, with its associated
material and labor costs, out of the design and replacing it with
fiber-optic communication.
1. At the Forefront
For decades, AEP has been at the forefront of power system
protection and control technologies. In the 1970s, the utility
participated in the introduction of digital relaying. In the 1980s, it
took part in early research into optical instrument transformers. In
the 1990s, it was an early participant of the Utility Communication
Architecture (UCA) group, subsequently the UCA International
Users Group, and the IEC61850 standard.
As the concept of a microprocessor-based relay matured and
turned into practical products, AEP led the way with widespread
adoption of the technology. Major improvements have been
achieved in the areas of material cost savings, operational
efficiency through remote access, control capabilities, multifunctionality, availability of data, simplification through integration
of protection and control functions, and elimination of some
auxiliary devices with associated panel wiring.
AEP envisions a possible future generation of protection and
control systems with interface devices dispersed throughout the
switchyard. The dispersed devices would provide the required
input/output structure for the existing apparatus: a simple,
robust standard communication architecture and interoperable
intelligent electronic devices performing traditional functions,
working exclusively with communication-based inputs and
outputs.
AEP encouraged the vendor community to pursue this vision. In
2008, GE Digital Energy (Markham, Ontario, Canada) developed
the HardFiber system, a complete and commercialized solution
designed to eliminate copper control cables from the switchyard.
In the second half of 2008, AEP completed the installation phase
of an evaluation retrofit project with the HardFiber product.
Figure 1.
A demonstration installation at the AEP Corridor substation used the HardFiber process bus system, shown dispersed around the station, for
communications interface.
AEP Process Bus Replaces Copper
85
2. Technology as a Brick
The HardFiber process bus system is a remote I/O architecture
for protection, control, monitoring and metering that allows
designing out copper wiring for protection and control signaling
within substations, replacing it with standardized optical fiberbased communications. The system includes relays and fiber
cross-connect panels, factory pre-terminated fiber cables and
switchyard I/O interface devices known as bricks.
The bricks implement the distributed concept of an IEC61850
merging unit, expanded to optically connect relays with all types
of I/O signals in the switchyard, not just instrument transformers.
The bricks are interconnected to the relays in a simple point-topoint arrangement that does not involve other active components
such as Ethernet switches.
In those early days, AEP envisioned architectures where a single
digital data source could be shared by multiple protection units.
As technology improved, AEP continued to track and evaluate
what was available. In the mid-1980s, AEP evaluated a Delle
Alsthom digital current transformer, whereby measurements
made in the head of the current transformer were digitized and
transmitted to ground through fiber-optic cables. The digitization
was attractive, but at the time, there were no digital devices that
could accept such a data stream. In addition, the concept of
having active electronics at line potential was thought to be too
revolutionary. In the late 1980s, companies such as ABB, Square
D and 3M developed optical voltage and current measurement
devices. The measurement technology was desirable, but the lack
of integrated and complete solutions impaired AEP’s use of the
technology.
In the mid-1990s, work began on the development of a standard
low-energy analog interface between measurement sources and
protection, control and metering devices. During this time, AEP
installed and evaluated an ABB 345-kV optical current transformer
for metering. In a subsequent demonstration in 2003, AEP installed
a NxtPhase 345-kV combined optical current transformer/voltage
transformer and successfully integrated conventional and lowenergy analog-output signals into GE and Schweitzer Engineering
Laboratories protective relays and a Power Measurement revenue
meter. The data from this demonstration also was used in a
research project of Power Systems Engineering Research Center
to evaluate the performance of a digital protection system.
Figure 2.
HardFiber protection panel with three relays and two patch panels (top
and bottom). All I/O signals interfaced via fiber optic communications.
The relays are GE Universal Relay series devices. The relay’s
current transformer/voltage transformer and contact I/O plug-in
modules are replaced with an IEC61850 process card to allow
optical rather than copper signal interface. The balance of the
relay hardware, firmware, functionality, configuration software,
documentation and user-setting templates are unchanged.
3. Evolution of the Digital Substation
Early on in this Digital Age, American Electric Power recognized
the applicability of digital technology for the protection, control
and monitoring of the power system. As early as 1971, AEP began
taking steps to foster this technology by funding research into
digital architectures and algorithms. AEP teamed with IBM to
develop and install a prototype of the world’s first communicating
digital relay. The device sampled voltages and currents, performed
basic protection functions and communicated the resulting data
and events to a mainframe at AEP headquarters.
86
Figure 3.
AEP digital substation architecture concept, circa 1980.
Standardized communications were seen as the necessary link
between the optical/electronic measurement devices and the end
users of this data. AEP had played a lead role in the development
of standard intelligent electronic device communications
and interfaces — specifically with its support of the Utility
Communications Architecture (UCA) protocol. The work with UCA
provided relay-to-relay and relay-to-master communication, but
did not address interfaces between measurement devices and
relays.
Parallel to the development of UCA, IEC Technical Committee
57 began work on what has become known as a process bus
and is now codified in the IEC61850-9-2 document. While some
experimentation and implementation of process bus has taken
place in conjunction with optical instrument transformers, it has
not been widespread, mainly due to the lack of a consensus on the
implementation approach and a solid fit-for-purpose architecture
that would provide real-world benefits at low risk.
AEP Process Bus Replaces Copper
4. Demonstration Installation
The HardFiber demonstration installation is in AEP’s Corridor
Substation, a 345/138-kV transformer and switching station near
Columbus that has been used for other new-technology trials.
The HardFiber installation provides distance protection for the
Conesville and Hyatt 345-kV line terminals and breaker-failure
protection for breaker 302N, which connects these two lines in a
breaker-and-a-half-like arrangement.
This portion of the station was considered typical and of
sufficient size and diversity to demonstrate the HardFiber system
technology. In addition, the lines already had existing Universal
Relays installed. So the existence of these devices enabled event
and oscillography records to be easily compared to those from the
HardFiber system. The trip/close control outputs of the HardFiber
system are not connected at this stage of the evaluation.
A site survey was conducted early in the project with the
manufacturer. The survey confirmed the viability of the scope
described previously, the quantity and location of the equipment,
and the lengths of the required pre-terminated fiber optic cables.
Twelve bricks were necessary to provide fully redundant coverage:
two bricks on each of the three circuit breakers, two on each of the
two-line current-voltage transformers and two on the one freestanding current transformer in the zone. In each case, no space
was found for mounting bricks inside the mechanism/marshalling
boxes, so brick-mounting locations were selected either on the
outside surface of the power equipment or on a supporting steel
structure.
The fiber-cable routing for the 12 cables consists of a 200-ft (61-m)
section in 6-inch (15-cm) duct, a section of up to 400 ft (122 m) in
a pre-cast cable trench shared with conventional copper control
cables, a direct-bury section of up to 150 ft (46 m) and an exposed
section from grade to brick level. The factory-terminated cables
required accurate cable-length measurements; a cable that was
too short would have to be replaced and excess length would
present slack management problems. Several length-measuring
methods were tried, including use of site plans, timedomain
reflectometry on existing spare conductors, a pulling tape with
numbered foot markings and a measurement wheel. In the end,
a surveyor’s tape produced the best results. The cables were
ordered with a 2% margin over the measured length.
Consistent with AEP’s standard design practices, FT-style test
switches were installed in the brick current-transformer circuits
shared with in-service protection and the brick voltage-transformer
circuits were fused.
5. On-Site Installation
Installation of the HardFiber equipment proceeded smoothly
and did not reveal any obstacle to future deployments. Since the
outdoor fiber cables were installed before the bricks were available,
slack was left in the section between grade and the ultimate brick
location. If sufficient slack was available, then a loop could be
created in free space under the brick. This loop, not likely to be
repeated in future installations, will increase the damage exposure
in the evaluation installation, making the demonstration a more
sensitive indicator of cable ruggedness. The bulk of the fiber cable
slack was in the control house, where it was accommodated in an
under-floor trench.
Figure 4.
HardFiber bricks installed on a bus support structure (left) and a breaker
marshaling box (right).
A transcription error made in transferring the measured cable
lengths to the ordering system resulted in several of the outdoor
fiber being made shorter than intended, but they could still be
used by relocating the relay panel within the control house. A
manufacturer’s engineer visited the site to correct a minor patch
panel problem, but otherwise installation and commissioning was
completed entirely by AEP field staff.
6. In-Service Experience
The HardFiber relays are connected to the Corridor Station local
area network and thus to a station data-retrieval system, making
the event records and oscillography of both the HardFiber and
conventional relays available for remote access and analysis.
Figure 5.
On Sept. 30, 2008 event – Hyatt line currents: hardwired relay (blue) and
measured by HardFiber bricks (red). The resulting pink color is due to exact
overlay of traces.
The conventional and HardFiber relays are set up to cross-trigger
oscillography through generic object-oriented substation event
messaging over the local area network and force an oscillography
record weekly in the absence of grid-generated events. Since
the HardFiber systems for the Hyatt and Conesville lines were
commissioned in June and December 2008, respectively, the
corresponding records have been reviewed. Tens of external faults
and switching events have been captured by both the traditional
and HardFiber protection systems. All the records and responses
of the relays are in full agreement.
AEP Process Bus Replaces Copper
87
The system operation meets expectations to date; not a single
error or failure has been recorded. It is also worth noting that,
through analysis of the HardFiber system oscillography files, a
failed coupling capacitor voltage-transformer fuse was found.
7. Observations and Lessons Learned
Installation of AEP’s first HardFiber system was successful and
uneventful. The following factors contributed to this success:
•
The HardFiber system is straightforward and practical. All
obvious challenges are addressed “under the hood” and the
user is not burdened with solving new problems. For example,
once connected, the bricks and corresponding universal
relays self-configure to establish communications.
•
The system was engineered, installed and commissioned
using AEP’s existing workforce, procedures and tools.
•
Early and continual involvement of the field personnel made
the demonstration more efficient and successful.
•
The manufacturer’s initial site survey, field measurements
and subsystem prefabrication shifted much responsibility for
project success to the vendor.
•
Reliance on a familiar product for the relay part of the
HardFiber system made the integration easy.
•
The plug-and-play nature of the system, with all components
prefabricated, is an important component of next-generation
protection systems.
•
The factory-acceptance test, performed with the complete
Corridor HardFiber system, reduced the time and effort to
confidently commission the system on site.
Figure 6.
Before (left) and after (right) — the amount of cabling at relay panels is
greatly reduced.
Continual development and commercialization of new technologies
are required to address the problems of a shrinking workforce,
rising costs, the volume of green field and retrofit projects, and
the integration of new generation to the grid. If these technologies
incorporate the latest standards, the utility industry can expect
to build on the value of systems like HardFiber to arrive more
quickly at functionally equivalent and interoperable multi-vendor
solutions.
The installation phase of the HardFiber system accomplished
the early objectives of this demonstration. In particular, the
system proved easy to engineer, install and commission, and is
compatible with the existing workforce. Distributed I/O, process
bus and replacing copper with fiber cables are seen as a stepwise
evolution of traditional solutions.
Based on the evaluation project to date, the system seems to
offer opportunities in shortening the construction times and
labor required, standard designs for bricks, cables and panel
building blocks, easier on-site integration of physical components
and reduced complexity in the control building. A more formal
comparison of performance and cost is planned in 2009.
The system still needs to prove, through wider field experience,
the longevity of its outdoor components and overall performance.
Given its simplicity and the rugged design of the bricks, it seems
the required maturity is already there and any minor issues can be
addressed. As a result, this fourth generation of digital protective
relays, with input and output interfaces placed directly at the
power apparatus, appears to provide a viable and practical option
for utility engineers and designers.
Ü Originally published in Transmission & Distribution World Magazine, March 2009
062909-v2
88
AEP Process Bus Replaces Copper
The Digital World and Electrical Power Supply
A Hypersensitive Imbalance
Raymond Kleger, BSc.
Editor of the Swiss periodical ˝Elektrotechnik˝
A complete failure of utility power is a normal occurrence.
A company’s power supply may be interrupted briefly or for
longer periods of time because of lightning, building work, or
network overloading. Electrical power supply companies (EPSCs)
cannot guarantee an uninterrupted supply. Many power supply
companies promise a reliability rate of 99.997%. This alone means
interruptions amounting to 14 minutes a year! It is high time to
think about how to survive the next power failure unharmed.
1. UPS systems yesterday and today
Faults in utility power supply cause half of all inexplicable
computer problems, whether hardware damage, loss of data or
complete system failure. A UPS unit eliminates the problem – but
careful, they are not all of the same quality. It is only when the
correct system is combined with other important measures that
system administrators can keep cool in the face of lightning; self
produced power spikes, long or short power failures and other
electrical faults.
The nominal voltage of our utility power network is 400/230 V.
Most equipment is designed to tolerate an under- or over-voltage
of about 15%. Anything higher or lower than this may result in
unforeseen malfunction. Light bulbs are an exception; their life
span is reduced rapidly if operated with over-voltage. When
operated with an under-voltage of 5% they produce only about
half the amount of light. Statistics show that most power failures
are shorter than 300 ms. Light bulbs flicker briefly but computer
systems often suffer considerable damage because data can
inexplicably be lost or changed. Such short power failures are
especially detrimental to database systems.
Furthermore, it is important to know that the most frequent
power problems are not long-lasting power failures but overand under-voltage, frequency deviations, short sags, or extreme
spikes. What many computer specialists don’t realize is that these
problems are often not caused by the EPSC that is delivering the
energy but by other equipment or systems that adversely affect
the power network. This paper will firstly explain the causes and
effects of faults in the utility power and secondly the operation
of Uninterruptible Power Supplies (UPS). The advantages and
disadvantages of different UPS systems will be shown. An important
topic will be that of redundant systems – because even UPS units
can fail. Finally day and night maintenance needs to be looked at
more carefully and here it is possible to use intelligently designed
software that can monitor any UPS unit via the Internet.
Originally the use of UPS units was limited to the protection of
larger computer network systems. In the last twenty years
however the information technology world has undergone
farreaching changes. The emergence and spread of the personal
computer gave rise among critical users to a boom in the use
of small single phase UPS units. During the last 20 years big
computers have been replaced by efficient PC networks. At this
point many UPS manufacturers became seriously concerned
about the future of medium and large sized UPS units. However,
this fear was unfounded. In no time it wasn’t enough to have just
one computer in the office. Each office worker in the firm would be
equipped with a PC and almost every industrial process would be
controlled by computer. Hardly any computers would be working
independently as they would all be part of a network. It was quickly
realized that it wasn’t enough to have a UPS unit for each working
station and the server. When PCs communicate over a network,
files are open. If one computer has a crash the file which is open
can be lost even if the power supply of the server is backed up by
a UPS unit. It isn’t enough to make sure that the server is provided
with an uninterruptible power supply. For important work all the
computers in the network must be backed up by UPS units. Of
course with many PC systems a small UPS unit isn’t enough. It
also isn’t sufficient to provide an uninterrupted power supply
for each PC in the network. A computer network also comprises
components such as routers and switches. These also need to be
included in the UPS system because switches won’t work without
electricity. Incidentally, switches are also a problem because their
network part only bridges minute disruptions in the utility power.
The Digital World and Electrical Power Supply
89
Today we can’t imagine our world without computer technology.
The sending of data over the internet is important everywhere
where computers are being used. More and more firms do their
business over the internet. E-Commerce is no longer only for the
few big firms. Many smaller ones are also using this new platform,
which provides the customer with extra conveniences such as
downloading of information and leaflets. If a firm uses e-commerce,
it mustn’t upset any of its customers through computer crashes.
A firm offering quick service in retail, consultation or problem
solution can’t afford loss of image through computer crashes.
1.1 Deregulation of the power market
It isn’t long since many European countries completely deregulated
their electrical power markets. The intention was to move from
the cumbersome state run companies to free enterprise in the
production and distribution of electrical energy. Since then,
however, disillusionment has set in. These fundamental changes
in the distribution of electrical energy have meant that the
production of surplus energy has disappeared. There is also hardly
any investment in the development of the utility power systems
since that only generates cost – not profit. In spite of economic
stagnation the amount of electricity used rises annually. Because
of this it will be impossible to avoid bottlenecks in the coming
years. Many countries have also decided to stop using nuclear
energy. However, no one knows how this energy source will be
replaced.
In the year 2003 New York experienced a total power failure lasting
several hours. The cost of this blackout was estimated at over a
billion dollars. The loss of perishable foods alone amounted to 250
million dollars. The blackout in Italy wasn’t any less dramatic! How
did it happen? On that unfortunate day Italy, the biggest importer
of electricity in Europe, had to import 750 MW more electricity
than it had foreseen. France was the supplier. The electricity
was delivered through the Swiss mountains because there the
resistance was lower. The high voltage transmission lines were
so overloaded that the heat caused the cables to stretch. They
sagged deeper and deeper until they discharged onto a tree
thus causing the immediate automatic shutdown of the line. It
wasn’t possible to reconnect because the Italian power system
was now in disarray (acute phase displacement). All the power
was now automatically transferred to the remaining high voltage
transmission line leading to Italy.
However here too due to the overheating there was a disruptive
discharge on to a tree leading to the shutdown of that line. After
the first breakdown the Italians should have immediately shut
down all the pumps of their hydroelectric power stations. Because
this didn’t happen a chain reaction was unavoidable. The Italian
power stations weren’t able to meet the demand and so there
was a complete breakdown in the power supply. A power system
that has completely collapsed isn’t so easy to resurrect and so it
took many hours until the whole of Italy was again supplied with
power.
Why this detailed account? When the deregulation of the electricity
market began many years ago the EPSCs didn’t really show any
interest in redundancy, that is investment in extra capacities of
production and infrastructure. However, power consumption has
increased during the recent years of economic stagnation. Aside
from this it takes years or even decades to obtain permission for
new power stations and high-tension power lines. All these factors
will lead to the quality of electrical energy supply deteriorating
in the coming years. America has already sunk to the level of
90
third world countries in terms of reliability of electrical supply.
Experts had long been convinced of this and a study made by
the government a short time ago confirmed this situation. In the
coming years Europe will also inevitably have bottlenecks and
hence more frequent electricity cuts.
The expectation of cheaper energy through deregulation was
made attractive to voters and it was cheaper at first. Since
then, however, the trend has clearly changed. The most striking
example is Sweden where power prices have doubled. In Germany
prices have risen by 9.6% per year following deregulation. New
firms requiring a lot of power to operate are now unable to
set up business in Amsterdam because the EPCS is unable to
increase its supply of power. It’s a crazy situation when the job
opportunities that everyone is demanding are already out of
the question because of the power situation. In recent years in
China there has been amazing economic growth of almost 10%.
There have been power bottlenecks in 24 of the 31 provinces.
However China has the advantage of being able to build, at short
notice, power stations using fossil fuel. If the demand for power
in Europe continues to rise at the present rate over the next 15
years, power stations with a total capacity of nearly 300 GW
will have to be built. That would mean building a hundred large
nuclear power stations. We should also not forget that some old
power stations will have to be closed down. In Germany and other
European countries it is practically impossible to get permission
even for hydroelectric power stations. Environmental and other
organizations campaign against every possible development of
the electrical industry.
Figure 1.
Lightning is a frequent cause of computer crashes.
1.2 Attacks from heaven
Lightning is a beautiful natural phenomenon– at least for those of
us who aren’t afraid of it (Figure 1). In a powerful storm, lightning
strikes with half the speed of light and heats the air up to 20,000°C.
This heat is four times greater than that on the surface of the sun.
The frightening noise – the thunder– is caused by the explosive
expansion of the air around the arc. Lightning is a major threat to
much modern electrical equipment. Only private people, if they
The Digital World and Electrical Power Supply
happen to be at home, can pull the plugs out at the first signs
of the approaching storm and sit back to enjoy the spectacle.
Employees at work should, with clear consciences, be able to
work on during the storm. Quite apart from that why should an
Italian businessman be affected by a storm in Germany when
he is just in the middle of e-commerce business? Meteorologists
count approximately 750,000 occurrences of lightning a year in
Germany, most of which occur in the months of July and August.
This sort of natural phenomenon causes enormous financial
loss. UPS units may help to avoid damage generally but not that
caused by lightning. For this, lightning and surge protection must
be included in the electrical installation. This needs to be done for
both the power supply of the UPS and any cables transmitting
data to the outside world. Finally there must be good potential
(voltage) equalisation in the building combined with an effective
lightning trap around the building (Figure 2).
•
Loss of image
•
Loss of contracts
•
Loss of a customer
•
Breakdown of customer service
•
Backlog in production
•
Loss of operational data
Very important! No insurance covers loss of image or loss of
contracts. Also don’t forget the loss of working hours caused by a
crash of the communication system. Companies asked indicated
costs of 13000 € per hour in the case of system failures and the
figure is even higher in the service sector. Many firms who have
experienced a serious computer failure complained of months of
increased economical difficulties.
1.5 To sum up
Before the deregulation of the electricity market, experts warned
of difficult times ahead in terms of availability and quality of
electrical energy. Their predictions have proved correct here in
Europe more quickly than we would have liked. The requirements
of the digital world have increased dramatically in the last two
decades with regard to the availability and quality of electrical
energy. There is a growing imbalance between the need for stable
electrical energy and the situation as we have it on the energy
market. A disquieting prospect – not for the manufacturers of
UPS units though! They can reckon on a growing demand for their
products.
Figure 2.
Causes of problems with computers and other equipment in offices and
industry.
1.3 Industry – without electricity nothing works
In order to remain competitive, firms in all western industrial
countries must computerise their manufacturing processes. The
level of automation has constantly increased and become more
efficient. PLC and PC systems are increasingly the order of the day.
Industry is of course dependant on a top quality power supply.
Critical processes are therefore dependant on a UPS because
even the shortest power cut can have fatal effects on the process.
There will be rejects and the machine will possibly be damaged if
half-finished products get stuck in it. The more sophisticated the
automation the more the process depends on an uninterrupted
supply of electricity. Complicated networks used to control
lighting, blinds, air conditioning, and entry and security systems
in buildings also all depend on a continuous supply of electricity.
Here too centralised UPS systems which keep at least the essential
processes going are increasingly being used.
1.4 The consequences of a computer crash
It is surprising to see how careless even large firms are in securing
a reliable supply of utility power. They should know how much the
success of a company is dependant upon it. A power failure of
only some few minutes may have fatal consequences such as:
2. UPS technology –separating the
wheat from the chaff
There is a wide variety of UPS system architecture. There are
simple systems which are capable of providing power until the
computer network has been shut down. There are more costly
systems which offer a complete galvanic separation from the utility
power supply and guarantee that ‘spikes’ never get through to the
computer network and its components. For installations where
power interruptions even of milliseconds must never occur – even
if the UPS unit breaks down – redundant systems are important.
This chapter looks at UPS technology more closely.
2.1 Typical problems in the utility power
network
Problems are not only caused by power failures. Short interruptions
that do not even cause a light bulb to flicker can have treacherous
consequences for different sorts of equipment. In computers,
network components and telecommunication systems,
overvoltage can cause the electronics to become defective.
Hidden effects are the most treacherous. In such cases a sensitive
electronic device still functions but its power consumption rises,
leading to overheating of the element and finally to failure. Figure
3 shows the typical problems in utility power.
Low voltage (brown out)
Approximately 60% of the disruptions. This is the most frequent
problem and is usually caused by large consumers of electrical
power, not by the user or the supplier.
The Digital World and Electrical Power Supply
91
Power failure
It is common to distinguish between those failures lasting
milliseconds and those lasting minutes or hours. The latter are
much less frequent in Northern Europe than the former. Every UPS
system must be able to cope with both types of power failure.
2.2 What types of UPS are there and how do
they function?
What does a user expect of a UPS? It must:
•
Bridge power failures for minutes and even hours.
•
Protect from over- and under-voltage.
•
Keep transients in the utility power from reaching the load.
•
Provide failure-free and stable voltage for any kind of load.
•
Provide careful recharging of batteries and protection from
low discharge.
An important characteristic of a UPS is the signal form it produces.
The power plant delivers a pure sine voltage with an effective
mean value of 220–230 V and 50 Hz in Europe. America and other
countries have 110–120 V and 60 Hz. The operation of simple
loads, for example a light bulb, depends only on the mean value
and the resulting power. But modern switching power supplies
for computers are more demanding and require a much closer
approximation to a pure sine voltage. Simple inverters deliver a
square wave voltage with peak and mean values that are identical.
This can cause a power supply unit to malfunction. An acceptable
approximation is a trapezium where the peak and mean values
correspond approximately to a sinusoid. The ideal is of course that
a real sinusoidal voltage is generated, which is what high quality
UPS do.
2.3 Off-line mode
Figure 3.
Frequent faults in the utility power.
Over-voltage
Approximately 20% of the disruptions. It stems from switching
operations performed by large consumers and can lead to
hardware failure.
Approximately 8% of the disruptions. Transients (spikes) are
extremely short occurrences of over-voltage. They can be several
times higher than the rated voltage and get through the power
supply units to the equipment, causing faulty transmission of data
or leading to hardware failure.
During normal operation (utility power voltage present) this
type of equipment does not provide voltage regulation. The
consequence of this is that in the case of fluctuation, the UPS
must switch into battery operation in order to compensate for the
fluctuation. This simple type of UPS is useful for single workplaces
especially in private usage, but it should not be used to supply
telecommunication equipment, network components or even
server systems. The autonomy time ranges from 3 to 10 minutes,
and the power range goes up to approximately 3 kVA.
Sags
Advantages:
These considerably distort the ideal sine waves of the utility power.
The consequences can be ‘inexplicable’ system failures or faulty
transmission of data. These problems are caused by pieces of
equipment that do not draw a cleanly sinusoidal current (light
controllers with phase shifting control or utility power command
guiding systems).
•
Reasonable in price.
•
High degree of efficiency.
Transients
92
This is also called standby mode (Figure 4). This is the simplest type
of UPS. It has two paths for the current. The concept of off-line
technology is that when the UPS has utility power, the load is
directly supplied with utility power voltage. The inverter remains in
standby mode stepping into operation only when there is a power
failure.
Disadvantages:
•
No suppression of non-sinusoidal injection back into the
utility power.
The Digital World and Electrical Power Supply
•
Short voltage fluctuations have no problem getting through
to the load.
•
Switching to inverter mode takes several milliseconds.
•
Low-cost equipment does not provide sine form voltage
when in inverter mode.
Disadvantages:
•
Switching into inverter mode operation takes several
milliseconds.
•
Faulty frequencies can only be eliminated in battery mode.
•
Low-cost equipment does not provide sinusoidal voltage in
inverter mode operation.
2.5 Double conversion technology
This type of UPS has two elements (Figure 6). On the input side,
the alternating current is rectified to direct current, which in turn
charges the battery. An inverter which is on the output side of the
UPS uses this direct current to produce an alternating current with
the frequency of 50 or 60 Hz (depending on the user’s network).
The inverter permanently produces the alternating current. Filters
at both input and output end successfully eliminate practically all
faults coming from the utility power. Single phase equipment is
available up to 10 kVA, three phase equipment up to 1000 kVA.
Higher power can be achieved by connecting several UPS units
in parallel.
Figure 4.
UPS with off-line or standby mode.
Advantages:
•
Steady sinusoidal voltage and frequency at the output.
•
A secure protection from over-voltage because of the
continuous conversion.
Figure 5.
•
No substantial switching delay in the case of a power failure.
This is very important if sensitive equipment is being used in
the area of network and telecommunication technology.
2.4 Active standby mode
•
Well defined and constant conditions throughout the
network.
UPS with active-standby or line interactive mode.
Also called line-interactive (Figure 5). This is a refinement and
improvement of the off-line mode (see above). During normal
operation, the load is supplied directly with utility power voltage
from an auto transformer at the output of the UPS. In many
countries, the utility power fluctuates considerably depending
on the load. The power supply units of computers and also other
power supply equipment cannot take a fluctuation of more
than±15%. Active standby technology functions in such a way that
a switching device at the level of the auto transformer can make
the voltage increase or decrease depending on what is needed. If
the supply voltage is too low, the switch turns into ‘boost’ mode, if
the voltage is too high, it switches to ‘buck’ mode. This correction
of the utility power voltage is not very finely tuned and thus – as
will easily be understood – not particularly effective. But it has the
advantage that the inverter does not need to step into UPS mode
with every slight fluctuation of the utility power voltage, thus
conserving the batteries.
In the case of a power failure, the switch changes the UPS to
inverter mode operation. Then the load is completely supplied by
the battery. This technology is often used for small networks and
equipment that is not too sensitive. Certain loads, however, do not
tolerate the switching time (reaction time) of this type of UPS unit.
The autonomy time lies between 6 and 10 minutes, the power
range reaches up to 3 kVA.
Advantages:
•
The batteries are spared because the UPS does not switch
into battery mode until the voltage goes beyond over- or
under-voltage.
Disadvantages:
•
The degree of efficiency of the whole system is low.
•
The technology is more complex and therefore more
expensive.
Figure 6.
UPS with VFI or double conversion technology.
2.6 Is a complete galvanic separation
necessary?
Online and double-conversion technology can be divided up
in another way, namely into UPS units with or without galvanic
separation. Many users repeatedly ask themselves which
The Digital World and Electrical Power Supply
93
technology is the best for their application. Figure 6 shows double
conversion technology. The inverter is available with or without
galvanic separation. Both types have their advantages and
disadvantages.
2.7 With galvanic separation
IGBT (Insulated-Gate Bipolar Transistor) transistors generate direct
current impulses from the battery voltage (Figure 7). One complete
sine wave (one full period) is comprised of about 500 individual
direct current impulses. Assuming that the output frequency is 50
Hz (duration of one period = 20 ms), it follows that the inverter
frequency is 25 kHz.
IGBT transistors. Power-MOSFET transistors are used for small
power applications. The UPS manufacturer optimises the inverter
according to technology and cost. In the area of high power
applications it is possible to work with frequencies around 12 kHz
nowadays, and in systems up to 1 kVA with up to 30 kHz. The critical
aspect of this optimization is that the IGBT transistors must be
operated in such a way that they reach a high degree of efficiency
on the one hand and maintain a sinusoidal output (despite the
non-linear load) on the other. A special quality characteristic of
an inverter is its capacity to react to load jumps quickly. Another
advantage of high inverter frequency is the fact that frequencies
over 15 kHz are practically inaudible for the human ear.
Advantages of the transformer technology:
•
The load is isolated from the utility power.
•
No direct current in the load.
•
Good dynamic characteristics even when non-linear load is
connected.
•
Good short circuit performance.
•
Only one battery voltage needed, the level of which can vary
greatly.
•
Power capacity up to 1 MVA possible.
•
The static bypass can be directly mounted thanks to galvanic
separation.
•
Fewer components needed.
Disadvantages:
Figure 7.
UPS with double conversion and transformer in the inverter.
The width of the direct current impulses varies in such a way that
the linear mean value of the envelope corresponds to a sinusoid.
This sinusoidal envelope comes from the filtering in the secondary
winding in the transformer in combination with the capacitor at
the UPS output.
The higher the inverter frequency, the smaller is the physical size
of the transformer. If the power stays the same, the size of the
transformer diminishes almost in proportion to the frequency. The
difficulty consists in finding materials for the core of the transformer
which operate with minimal loss due to magnetic hysteresis and
eddy currents. Laminated iron cores are problematic above 5 kHz.
Ferrite cores are much better but also much more expensive.
Furthermore the capacity of the filtering capacitors at the
output end of the UPS drops with higher frequencies. But a high
frequency of operation results in better dynamic characteristics of
the inverter. The narrower the direct current impulses, the quicker
the reaction of the inverter to load jumps. Power supply units of
computers, networks, telecommunication equipment and other
electronic equipment draw a non-sinusoidal current. In simple
inverters the consequence of this is that the curve deteriorates
to a trapezoid.
IGBT transistors have the disadvantage of considerable energy
loss when operated at high inverter frequencies. In recent years,
the frequency has been increased due to the higher quality of
94
•
The transformer is fairly large and heavy.
•
Expensive in comparison to a design without a transformer.
•
Especially in the low power range the energy loss is high
when compared to a design without transformer.
•
The realization of Power Factor Correction (PFC, the current
which the rectifier draws at the input is sinusoidal) in the
rectifier is more difficult than with the solution without the
transformer.
2.8 Without galvanic separation
This approach requires the use of two sets of batteries because a
double supply of direct current is needed (Figure 8). Single phase
technology used this approach in the early nineties. The positive
half-sine wave at the output is produced by the upper three
transistors for all three phases. This is achieved in the same way
as in a design with a transformer. The half-sine wave is comprised
of direct current impulses of variable width.
It is clear that there is no galvanic separation between input
and output. The negative side to this is that the output can be
‘contaminated’ with a direct current component. This component
is caused by the fact that the linear mean value of the positive
halfsine wave is different to that of the negative half-sine wave.
It is very difficult to rectify this problem and consequently there is
always a possibility of a direct current component at the output.
This will not harm switched power supply units of computers,
networks or telecommunication equipment, but it will harm
The Digital World and Electrical Power Supply
supply units of toroidal transformers and AC motors. In this design
the inverter frequency is higher than in the one with a transformer,
lying between 15 and 30 kHz, because it only requires filter
chokes.
When to use the transformer technology?
•
When the load has to be separated from the utility power
supply and the battery.
•
When the load is sensitive to direct current.
•
When the dynamic characteristics of the UPS have to be
very good (also in case of high load jumps and short-circuit
behavior).
•
When the application does not allow high currents on the
neutral output line.
When to use a unit without a transformer?
Figure 8.
UPS with double conversion and without transformer in the inverter.
Advantages of the technology without a transformer:
•
When the efficiency of the unit is important.
•
When the price plays an important role.
•
When the level of noise needs to be as low as possible (inverter
frequency inaudible to the human ear).
•
Smaller in size
•
When galvanic separation is not mandatory.
•
Less expensive
•
When the load is not sensitive to direct current.
•
Higher degree of efficiency
2.10 Super Eco Mode
•
Less noisy
Disadvantages:
•
No separation of the load from the power supply.
•
Direct current component at output.
•
If badly regulated, the neutral output line can carry a high
current.
•
Twice the number of boosters and batteries are needed.
•
Limited to units up to 120 kVA.
Figure 9.
Output transformer magnetised by the output voltage, efficiency of 98%,
switchover time only 2 ms patented by GE (Super Eco Mode).
2.9 Which technology is better?
It is pointless to ask this question because both technologies have
their advantages and disadvantages. One or the other solution
is more appropriate depending on the requirements. This is why
leading UPS manufacturers build units with or without galvanic
separation in order to meet the varying requirements of customers.
However, the customer should make sure he knows which design
is being offered to him.
As mentioned previously, UPS units operating according to the
off-line system are the most efficient because the utility power is
normally led directly to the output via a static by-pass. The U I
current on the silicon-controlled rectifier and the energy needed
to sustain the charge of the battery are the only losses. The
disadvantages of this system were given previously.
UPS units with transformers are naturally less efficient. The
transformer alone absorbs approximately 3% of the power at
full load and 1% at zero load. For safety reasons, however, UPS
with transformers continue to be in use. The pros and cons are
provided above.
In many countries in Northern Europe the tension and frequency
of the utility power is remarkably stable. Hence it is obvious that
the off-line system is the ideal UPS from various points of view.
Unfortunately, however, this system does not satisfy the highest
requirements regarding safety and stability. For very sensitive
loads which do not tolerate direct current at all, users may have
to resort to UPS with an output transformer. But such a unit
cannot switch instantly from the ‘dormant’ mode of operation to
the active one. The problem is the transformer, not the inverter.
Even if the inverter can supply the output transformer instantly
with the required tension, the load will not receive it instantly. The
transformer has inductivity and this delays instant transmission.
The delay mounts quickly to 20ms and in addition to that surges
of tension are to be expected.
GE’s research division has provided a solution with the Super ECO
mode (patent pending). In this system the load current flows in the
ECO mode via a bypass as with the off-line system (Figure 9), but
of course only as long as the tension and frequency of the utility
power lie within specified limits. The innovation is that the output
transformer of the UPS is being ‘kept primed’ all the time by the
load, i.e. kept magnetized. The benefits are as follows:
•
If there is a power failure there is merely a drop in tension at
the output, lasting less than 2 ms. There is no complete loss
of tension.
The Digital World and Electrical Power Supply
95
•
The output transformer and the filtering capacitors of the
UPS act in the ECO mode as passive filters and thus improve
the load parameters with regard to the utility.
•
The inverter is turned off and the rectifier turned on
occasionally in order to charge the battery.
•
The user is sure that even with the worst case scenario of the
UPS there will never be any direct current on the output side,
since the unit has an output transformer.
•
The efficiency of the unit is around 98% (the additional loss
of 1% in comparison with the simple off-line system stems
from the energy used for the permanent magnetization of
the transformer)
2.11 DSP technology
Analogue technology is nowadays used by well-known UPS
manufacturers for the generation of the signal. The processing
of the signal, as well as the controlling of the transistors of the
rectifier and the inverter, is achieved through a DSP (Digital Signal
Processor). DSP technology drastically reduces the number of
parts needed since it measures the tension of the three phases
as well as the current directly and calculates the width of the
impulses of the direct current by way of a complex algorithm
transmitted to the IGTB transistors. In this way the reliability of the
electronic system is improved, i.e. the MTBF (Mean Time Between
Failure). DSP is nowadays so efficient that at GE, for example, the
total control of a UPS system is reduced to one single print-out
and this print can be used for systems from 10 to 500 kVA. DSP
also makes the application of a new technology for UPS possible,
the so-called Space Vector Modulation (SVM). This technology
makes it possible for only minute drops or rises in tension to occur
when the load changes. DSP generates three sinusoid tensions as
well as the modulated impulses for the three phases, taking into
account the current required by the load in order to control the
IGBT transistors. DSP and SVM greatly improve the performance
of the UPS, also in the most demanding case of redundant and
parallel circuiting of several installations.
2.12 The myth of a high degree of power
efficiency
A high degree of power efficiency is often used as a sales argument.
A UPS in off-line mode, operating normally and directing the
current from the utility power supply directly to the ‘protected’
load, obviously has the best power efficiency of approximately
98%. For line-interactive units the figure is 97%. For UPS in online
mode the efficiency is 89–96% depending on the size of the unit,
this being the real online-degree of efficiency. In systems with
transformers the efficiency can reach a maximum of 95%, in
those without 96%. The brochures of UPS manufacturers always
indicate the highest degree of efficiency. However, the efficiency
is dependent on the utilisation of the unit, the power factor of
the load, and the input voltage. For example, a UPS unit using
the so-called Delta Technology (not explained in this article) has a
definitely lower efficiency with non-linear loads, i.e. in the case of
poor power factors.
The decision to purchase a certain model must be well thought
out because the protection of the equipment connected to the
UPS must be given highest priority. The question therefore is: is it
worth taking a higher risk in order to save expenditure on energy?
The money saved by buying a cheaper unit and saving energy
never justifies the risk of damaging, for example, a sensitive server
system. Just one single unnecessary shutdown per year (together
with the resulting costs) exceeds any saving in energy costs.
To illustrate this, take the example of a large network with
approximately 200 computers, including the monitors, printers
and other network components, all of which are dependant on
a UPS for their energy supply. Let’s assume that the network is
used for 12 hours per day. For this network a UPS supplying 100
kVA is necessary. At full utilisation and assuming 3% less degree
in efficiency there is an additional loss of power of 3 kW. Assuming
further that these 3 kW are lost for 12 hours during 365 days per
year – which is unrealistically high – the additional cost of energy
amounts to 1000 €, if one kWh costs 0.10 €. This then would be
the amount saved by choosing a system with a lower level of
protection. It would never cover the cost of rebooting a network
of 200 computers after a crash. One aren’t think of the cost of the
time needed to reboot it.
2.13 GE’s LP and SG Series units –
examples of the most up-to-date
technology
Taking the example of these ultramodern units, both
designed for a power range of 80–100 kVA, let us
examine the differences in the realm of technology.
The following data is given for a 120 kVA system.
Figure 10 shows the schematic diagram of the LP
Series Model which incorporates the feature of
galvanic separation. Both models satisfy the highest
level of UPS standards. They are VFI units, i.e. units
operating in online mode. Both can be switched to
redundant parallel function by means of RPA (see
below) and both are equipped with the Super ECO
mode. The LP Series model, however, is a unit without
transformer. The SG Series model has a transformer.
Figure 10.
Schematic diagram of LP 33 Series, trimmed for high degree of protection and highest
degree of effectiveness.
96
The first difference is a purely external one: the LP
Series requires 0.52 m2 ground surface, the SG Series
almost double – 0.96 m2 – which still makes it one of
The Digital World and Electrical Power Supply
the most compact units on the market. The difference in size is
somewhat balanced out when we take into account the number
of batteries needed. The LP Series needs forty 12 Volt units, the SG
Series needs only thirty.
more in demand. On the UPS market GE offers a unique failuretolerance of N+1 redundancy.
The values of the output tension are almost identical. Both units
overcome a load jump of 100% with less than 2% fluctuation of
the tension. Both have a total harmonic distortion of less than
3% with 100% non-linear load. There is a small difference in the
efficiency of the units: the LP Series reaches 93% in VFI mode and
99% in ECO mode, the SG Series unit reaches 92–98%. This small
difference is the price paid for galvanic separation.
Important business activities and hightech control systems require
an uninterrupted power supply. The installation of a UPS unit makes
this possible. Such a unit is made up of electronic components,
batteries and mechanical parts, all of which can break down.
Hence it is clear that you cannot depend on a single UPS unit only
when supplying computing centres or other important systems
with electrical power.
At a first glance the two models are similar, both fulfilling the
standard range of expectations (Figure 11). Why then does GE
offer two models? Just as big car producers offer different ranges
of power and comfort, GE has produced two models for the
important power range up to 120 kVA. One can view the LP Series
model as the open sports car, and the SG Series as the luxury
four-door four wheel drive. With the LP Series unit you can add up
to four elements in parallel redundant setting – with the SG Series
(thanks to SVM) you can add up to eight. Furthermore the SG Series
is equipped with a whole range of additional conveniences and
safety devices, e.g. redundancy ventilator, optimized soft start,
greater capacity of the batteries and a convenient control panel.
A redundant system can cope with the failure of one of its parts
without normal operation being impeded. An example will help to
explain this. An aeroplane used for the transport of passengers
has at least two engines. If one of the two fails to operate, the
aeroplane must nevertheless be able to fly on to the next airport
and land there. In an N+x redundancy, the N stands for the number
of units operating in parallel. X stands for the number of units that
can fail to operate without affecting the operability of the entire
system. In our example we have an N+ 1 redundancy. In the case
of a plane that has three engines it can lose two and still continue
its flight; here we would talk about an N+2 redundancy.
3.1 What is a redundant system?
However, this does not mean that the aeroplane with three engines
necessarily has an N+2 redundancy. If there is one element in the
plane upon which all three engines depend then this plane does
not have redundancy at all! Here is the problem with most of the
global UPS manufacturers. Individual UPS units in parallel may
provide redundancy – or may not, depending on whether there
is an additional component in the system upon which all the UPS
depend. There must not be the possibility of a single point of failure
in the system. Because of the importance of this point, several
concepts of parallel architecture will be presented, highlighting
their advantages and disadvantages.
3.2 Hot-standby systems
Here, as figure 12 shows, two UPS units are connected in series. In
normal operation, unit 2 takes over the supply of the critical load.
If unit 2 fails to operate, it switches to bypass mode and after 2–8
ms unit 1 automatically takes over the supply of the load.
Figure 11.
LP Series on the right – an ultra-modern UPS without transformer, SG
Series on the left – units with transformers using very little floor space, 120
kVA each.
3. Assess your partner firm – avoid
being left in the lurch
The need for upgradable UPS systems with highest fail-safe
record has increased enormously. Computing centres of internet
providers, banks, telecommunication companies and all those
using large computer networks demand a high degree of
availability. Many customers are now less cost conscious when
it comes to power supplies. Yet some firms still entrust their
complex server systems to a low-cost UPS. Upgradeability, that is
the possibility of adding power and autonomous time, is more and
Figure 12.
Cascaded UPS system.
Advantages:
•
Reasonable in price since no additional components are
necessary.
Disadvantages:
•
There are many single points of failure.
The Digital World and Electrical Power Supply
97
•
There is no distribution of the load. If one UPS fails then the
other one has to take on the whole load. This means that it
must be able to cope with an increase of power supply from
0–100% within approximately 8 ms.
•
Overload is limited to the capacity of one single unit.
•
The MTBF of the whole system is lower than the MTBF of a
single unit.
•
The loss of energy is relatively high because one unit is only
’idling along’.
of several UPS units in parallel (Figure 14). The advantage of this
configuration is that there is a distribution of the load and losses
are minimal due to the fact that the load current does not run
through two switches. Yet it is also clear that the external switch
is the critical element in this configuration. If it fails, there is no
redundancy. This configuration is comparable to a jet plane with
a central hydraulic system. With regard to the UPS units that are
aligned in parallel, the system is redundant, but not with regard to
the central external switch.
3.3 Parallel system with automatic switching
mechanism
This architecture operates with two or more UPS units as well
as an automatic transfer switch (STS static transfer switch). The
sensor in the STS monitors the output voltage of each unit and
immediately switches to a different UPS (or several) as soon as
a failure is registered. Figure 13 shows that the system is not
redundant because of the STS. If this component fails then the
UPS are of no use.
Figure 14.
Parallel system with with external switch.
Advantages:
•
When one UPS unit fails to operate, a different one takes on
its load.
Parallel system with automatic switch (STS).
•
There is a distribution of the load.
Advantages:
Disadvantages:
•
•
Figure 13.
When one UPS fails to operate, a different one can take on
the load.
Disadvantages:
3.5 Parallel architecture with master and slave
•
There is no distribution of the load.
•
Additional expenses: an STS costs roughly what a UPS unit
without batteries costs.
•
Additional loss of 1% of the energy.
•
If the STS fails then the whole system collapses (a single point
of failure). The UPS units which actually still work are of no
use.
3.4 Parallel system with external switching
mechanism
UPS normally have an internal switch to change from inverter
mode to bypass mode. Some manufacturers situate this switch
externally for manual operation in order to facilitate the use 11
98
If the external switch (ES) fails (a single point of failure) then
the whole system fails. The UPS units which actually still work
are of no use.
In this configuration, one UPS (or special circuitry) takes on the role
of the Master and the other units take on the role of Slaves. The
Master is responsible for the load to be evenly distributed among
the UPS which are aligned in parallel. If one of the UPS units fails
to operate the Master automatically redistributes the load to the
other Slave UPS. It is clear from figure 15 that this configuration
also has its weak points. If the Master UPS fails to operate, then
it switches into bypass mode, but if the circuitry that controls the
Master unit fails, then the whole set-up is without a Master and
therefore no longer controlled. This configuration has at least two
single points of failure.
Advantages:
•
No casing with an external switch necessary.
The Digital World and Electrical Power Supply
Disadvantages:
•
If the circuitry that controls the Master UPS fails, the whole
configuration is out of control.
•
The data bus is not redundant. If it fails, the whole system
collapses.
the whole system requiring a new distribution of the load. In
the designs offered by other manufacturers, a change in load
distribution necessitates a change into bypass mode. This means
that the critical load is connected to the utility power without
protection in case of power failure.
Figure 16.
GE real, redundant parallel system, without a single point of failure.
Figure 15.
Parallel system without external controls or switch, but with at least two
single points of failure.
3.6 Real redundancy parallel architecture
without a single point of failure
GE is one of few manufacturers on the market producing UPS units
with real redundancy. The system is called Redundant Parallel
Architecture (RPA). In it there is no need for external electronics
or switches to control the UPS in parallel arrangement (Figure
16). With RPA, using so-called Active-Active-Technology, one of
the UPS in the system temporarily takes on the role of the Master
and the others follow, as in a democracy in which one person
takes on the role of leadership. However, all UPS have access to
all control parameters. The system is equipped with a redundant
bus (featured twice) which ensures constant distribution of the
load. If one UPS unit fails to operate, the load is automatically
redistributed among the other units. If the Master UPS fails to
operate, then a different UPS automatically takes on the role of
the Master. If necessary, any of the UPS in this democracy can take
on the role of leadership.
If there is a need for more protected power, it is possible to simply
add further UPS in parallel to the existing ones. Furthermore, a
UPS can easily be switched off or another one switched on. And
here is the unique feature of GE’s RPA: your critical load is always
protected. As soon as a UPS is switched off, its load is taken over
by the other units without the load even ‘noticing’ a change of
voltage. The addition of a UPS unit is a more complex operation.
The new unit must first be synchronised with the load voltage,
and then the Master UPS must take care of the integration into
The critical point of this GE technology is the exact synchronisation
of all the UPS that are aligned in parallel. The reference value is the
load voltage, and depending on the load, individual UPS will need
to provide more or less current. The supply of the necessary load
current is the control-condition reference value. The distribution of
the load with RPA technology is so precise that all the UPS in the
system provide substantially the same current (it varies by only
a few amperes). Finally, the excellent dynamic behavior of this
architecture needs to be mentioned. It guarantees a negligible
fluctuation of voltage even in the case of a sudden, big load jump
e.g. a short circuit.
3.7 The cost advantage of a completely
modular system
The need for upgradable UPS systems with highest fail-safe
operation has increased dramatically. Few companies operate
without having their own server, quite apart from all the internet
service providers, banks, telecom companies etc. All of these need
UPS systems. They are ready to invest capital in order to obtain
a maximum of operability. Important criteria in the choice of a
system are the questions of the expansion of the system later and
the length of time that the UPS system can provide power during
a power failure. It makes little sense to install an over-specified
UPS system just because there might be a need for more power in
the future. An over-specified UPS system produces unwanted heat
and costs more.
For the above reasons the upgradeability and failure-tolerating
N+1 redundancy is important for sensitive loads. The modular
approach also makes for cheaper production and running costs.
The Digital World and Electrical Power Supply
99
There are different versions of the modular system on the market
such as strict separation of the UPS into individual casings or into
drawers in a main casing. Both versions have their advantages
and disadvantages, which are not discussed in this brochure.
The customer has to make sure that he is given the pertinent
information.
3.8 UPS – only a part of a safe power supply
In a high-standard set-up the UPS with its set of batteries represents
only part of the whole. The quality and correct installation of other
parts such as the connection to the mains, the connections to the
output, the type of fuses used and the selectivity of the different
circuits all play a vital role.
The above points highlight the fact that the UPS is by no means
the only element in getting a secure supply of electricity to your
critical load. The experience of recent years has shown that
integrated solutions are called for. Market research shows that
manufacturers of UPS should offer integrated systems in the future.
In other words, offering a high quality UPS is no longer sufficient.
An overall solution is what is called for. This solution comprises all
the parts between the connection to the utility network and the
load.
3.9 The future: advantages through integrated
solutions
If a UPS manufacturer offers an approved overall concept, a whole
range of problems is eliminated. Here is an illustration: if you ask
an architect to build your house you will get a house that is unique.
If you chose your house from the catalogue you’ll get something
which is well-planned but off the shelf. In this house everything
should work. Teething troubles, which one would have to expect in
the case of a ‘prototype house’, should be eliminated. It may be that
for your own special house you’ll put up with ‘teething troubles’.
With regard to a secure power supply it is completely different. Here
it does not make sense to develop the whole system from scratch
again for each new application. For an every-day application this
is less important. The parts simply have to be wired up correctly
according to an existing diagram. It is rather like putting railway
carriages in the right order on the rails. For a complex application,
as required in the office or telecommunication sector, the design
of an individual power supply system would be a demanding and
time consuming job for a highly qualified ‘architect’. To go for an
individual special solution obviously does not make sense in this
case. Here one must go for the proven solution off the shelf –
one comprised of tested elements which can be put together in
different ways.
Such a system should not only include switch cabinets with
excess tension protection, power switches, automatic fuses, input
and output connectors, distributors, the obvious UPS unit with
its set of batteries, but also a comprehensive control and service
programme. Ideally a comfortable design programme greatly
simplifies the work of the system designer. The advantages are
striking:
100
•
Achieving greater reliability of the whole system.
•
Saving through the use of standard modular elements.
•
Testing of the entire system beforehand made possible.
•
Shortening of the installation time.
•
Dealing with only one partner.
•
Permanent monitoring, making visual checks unnecessary.
•
Discovering potential problems in advance.
3.10 Critical loads that overtax an ordinary
UPS system
There are extraordinary demands on UPS systems in the area
of telecommunication, medicine and also industry. Particularly
critical are loads that must under all circumstances be supplied
with electricity. Here a redundant system is imperative. But there
are also other kinds of critical loads, namely those that draw power
in an extremely pulsating manner (Figure 17). Medical scanners,
computer tomography and x-ray machines fall into this category
along with certain machines in industry. These consume large
amounts of electricity for short periods of time. On the right in the
picture the effect on the utility power can be seen clearly i.e. sags
and peaks of tension. These adversely affect the operation of the
instruments. It need hardly be said that a power failure in a hospital
is extremely critical if some diagnostic process or op- Near-by grid
distortion Dynamic current absorption Medical Scanner Figure 17
Pulsating power consumption of a medical appliance. 13 eration is
in progress. Which patient would be pleased about a power failure
while he is being examined or while the computer is evaluating
his data?
Figure 17.
Pulsating power consumption of a medical appliance.
What is the problem in this case? Ordinary UPS systems have
great difficulty in supplying loads that use electrical power in a
fluctuating manner. Here are the resulting problems: the tension
drops sharply and this in turn causes the instrument to work
improperly. The simple sort of UPS system may not even be able
to supply a current that is higher than the nominal one, not even
for a short time. What can the customer do? He could opt for
a (highly) over-rated UPS so that it will be able to cope with the
fluctuations. However, the problem is not yet solved. For a short
interval of time his unit is likely to deliver undertension, because
it cannot cope with the dynamics of fluctuation. When a load is
switched on, i.e. starts operating, the initial current causes the
tension of the UPS to drop, the switching off of the load causes
the opposite, i.e. a rise in tension. The other loads connected to
the UPS may suffer harm. The European standard EN 50091 for
UPS defines variation in tension in terms of dynamic changes of
the load. The UPS units of manufacturers which guarantee this
standard are, however, far from being able to supply critical
medical machines adequately. Class 1 standard allows a drop of
tension of 30% during 6 ms while the load rises from 0 to 100%.
The same is true for a drop of the load: the tension may rise up to
The Digital World and Electrical Power Supply
30% during 6 ms. 30% fluctuation in tension during 6 ms is a lot for
average electronic equipment. GE systems offer ten times better
values. With a change of load of 100% the tension fluctuates a
mere 3%. It is to be noted that greater changes of the load occur
with certain medical instruments. It is worth the purchaser’s while
to make a very careful comparison. A cost comparison alone
with a standard unit (which would have to be an over-rated one)
simply isn’t sufficient. In the medical sector the difference in price
is a relatively insignificant one when a high quality UPS is being
purchased – it amounts to 3–30%.
How does GE achieve these superior values (compared to EN
50091)? GE achieves them through the use of an inverter with
highly dynamic characteristics in combination with special output
wiring. In particular through the following three features:
•
The output transformer has zigzag windings, thus distributing
a change of load on one phase to two phases of the output
inverter. For very critical loads an output transformer is usually
installed in the UPS unit for safety reasons so that even in the
worst case no direct current gets through to the load.
•
The output tension as well as the output current are monitored
directly and are evaluated with regard to the tendency to
change (i.e. in a differentiating manner).
•
The highly dynamic SVM technology (Space-VectorModulation) is used for the control of the inverter. Most
equipment on the market still uses PWM technology (Pulse
Width Modulation).
These three measures allow the unit to discern changes of
the load ahead of time, and react through the inverter within
microseconds. The three features together with elaborate DSP
technology (DSP=Digital Signal Processor) have resulted in GE
producing a UPS unit with 10 times better dynamic values than
those of the European specification for Class One UPS.
Installing a UPS correctly is one thing. Shutting down all the
computers in case of a power failure is another – especially if no
one is around. Programmes that are running have to be stopped,
open files have to be shut down, and unsupervised systems have
to be shut down in a controlled way. When the utility power is
on again, the UPS software takes care of rebooting the system
again. This is the straightforward task of the so-called shutdown
programmes which most UPS manufacturers offer free of charge.
The operation becomes more difficult where a complex IT-system
with different operating systems and hardware from different
suppliers is in use (Multi-Platform and Multi-Vendor). The most
difficult case is a decentralised system that has to be monitored
by remote control.
4.1 Functionality of the UPS software
This is best explained by giving an example. Figure 18 shows in a
simplified way the computer/UPS system together with the server
and various customers online. By means of SNMP the UPS units
are integrated into the IT-network (Simple Network Management
Protocol, a worldwide standard language for the communication
between components of IT networks). The UPS jumps into action
as soon as the utility power fails.
Taking the example of the GE software ‘Suite’ the comprehensive
functionality is easily understood. The software packet consists of
two parts, which can be used individually or together.
JUMP (Java Universal Management Platform) affords high flexibility
and wide ranging independence from the differences of the IT
operating systems. JUMP jumps into action when the UPS are
integrated into the IT network and when an orderly shut-down of
the whole system has to be carried out, e.g. open communication
links having to be closed.
4. Intelligent software makes the
difference
IRIS (Internet Remote Information System) makes, as the name
says, remote monitoring of UPS possible. Wireless systems such
as GPRS and UMTS are used to supplement the link through the
internet. IRIS has its place when an independent structure for
monitoring is needed and when people who are not part of the
internal network have to be contacted, e.g. service technicians of
the UPS manufacturer.
Worldwide there are hundreds of UPS manufacturers, but most
offer only a very limited range of units. If you start looking for
manufacturers who produce units for a few hundred watts up to
the megawatt range, the number shrinks to a mere dozen. And
if you start looking for companies that offer efficient monitoring
and maintenance software, the figure drops to a mere handful of
names.
JUMP will call on the person responsible for the software, IRIS on the
person responsible for the technical infrastructure. Both software
users can adapt the programme and receive the information in
their own language. The person responsible for the software will
receive it by means of SNMP right into his network management
system (NMS) – the person responsible for the technical side of the
set-up will receive it in the form of Volt and kW information.
UPS LANPro
Server
Client
Client
SNMP Manager
TCP/IP
LAN/WAN
TCP/IP LAN
TCP/IP LAN
Figure 18.
Typical network configuration of UPS units and computers.
The Digital World and Electrical Power Supply
101
4.2 Big supermarket chains – maintenance of
1000 UPS
Thanks to maintenance by remote control and a warning system
at the main service centre it seldom happens that one of the UPS
in a branch of this chain breaks down.
A chain of supermarkets may have 1250 shops and 160
restaurants. The entire supply of goods is computer controlled. In
total this chain may have 6600 cash points and 2600 computers
to support the cashiers in their task. In most of the markets no
UPS are planned for the server, the network and the scanner
cassettes. If a UPS does not work properly no one is likely to notice
it. After all you cannot expect the personnel, in the rush of serving
customers, to take notice of a warning signal coming from a UPS.
In an individual branch of a supermarket chain it’s highly unlikely
that someone is responsible for the computers and the software.
Therefore no-one will do anything about the problem with the UPS.
UPS are there to bridge power failures and these don’t happen
according to plan – but a faulty UPS will in that situation not be
of any use. This calls for a solution. If there is a fault in a UPS or if
there is a utility power failure the service centre of the chain would
like know what has happened and where (Figure 19). Then, when
the case arises, the service centre will be able to act, i.e. send
instructions to the particular branch. In this way the problem can
be solved or the damage limited. Causes for faults in UPS can be
high temperature or unauthorized connecting of new equipment.
UPS units do require maintenance! The batteries at least have to
be checked regularly. To do this for 1250 supermarkets is a major
job, requiring personnel. Therefore GE has worked on a system
to monitor the state of UPS by remote control thus reducing the
number of breakdowns.
5. Other things that users of UPS
systems need to know
All the UPS are connected to the local area network (LAN) by means
of a SNMP card. This card gives access to remote maintenance.
The UPS reports any fault automatically. The report goes to the
regional centre as well as the headquarters of the supermarket
chain. In addition, the technical service centre of GE is notified.
Depending on the kind of fault reported it will be decided
whether a GE technician has to go to site or not. At the moment
this particular market chain has 300 UPS of 600 VA–400 kVA in
operation, all supplied by GE. In addition to this it has 500 UPS
in operation from various other suppliers, all of which will shortly
be replaced. The new UPS are fitted with a device which calls the
service technician of the market through a visual or audible signal.
How can I calculate the correct power requirement for a UPS
system? How can the UPS be tested to see whether it actually
works when there is a power failure? What maintenance work is
required? And last but not least, what set-up suits the needs of my
enterprise?
5.1 The right choice
Where you have one server with network stations (probably not
backed-up by UPS) a unit providing a long autonomy time in the
case of power failure doesn’t make sense – continuing work is out
of the question anyway. What is important is that the components
of the network can terminate their application in a controlled
fashion before the battery power is gone. Normally a server should
be shut down within five minutes of a power failure.
In order to provide a fully automatic shutdown of the above
system the right kind of UPS needs to be combined with the right
sort of software. This software must be installed on the server and
possibly on other computers in the data network. When the server
is booted up after a power cut, the batteries are still not recharged.
A second power failure could then lead to a tragic collapse of the
system. A feature of sophisticated UPS systems therefore is that
the load can only be turned on again after the batteries have been
recharged.
In computer centres in hospitals, where shutting down the
system is not an option, the UPS is mainly there to bridge the time
between the beginning of a power failure and the emergency
power generator (petrol or diesel) providing (synchronised) power.
In this case a few minutes of autonomy time are sufficient, but
it must be kept in mind that the back-up system has to provide
several hundred kVA of power because the entire installation is
running on the UPS.
Shop sends detailed info to 2 servers.
Server sees individual UPS
Alarm
Start
Data
Save
Redunant Server in the control
center of the supermarket chain
Shop sends Ticket to NMS
NMS sees UPS as one
system
Shop Server
JUMP Manager
installed on Servers
NMS sends Ticket to
regional supermarket
chain support (manned
724) as redundant info to
JUMP server
Both servers send e-mail message
to GE Service & reg. support
WAN
USV
Figure 19.
USV
USV
Network Management
System
Supermarket chain with 1000 UPS units, monitored by central remote control.
102
Support GE
Support
Supermarket chain
The Digital World and Electrical Power Supply
UPS are available from 300 VA rated power up to 500 kVA. In the
case of more complex projects it is advisable to consult a UPS
expert. As already stated, computer systems and industrial plants
that require a high level of energy availability need a redundant
UPS system. The electrical planning of such a project together
with software that is needed on different computers requires
considerable know how. In simple situations the customer can
chose the UPS himself. In order to do that he first needs to calculate
the power requirement.
Level of
security
1
2
5.2 What power does a UPS need?
1. The first step is to define all the different loads that will be
connected to the UPS. It is important to remember that it is
not enough to only connect the server. An operating system
such as Windows NT requires that the computers in the
network be shut down in a specific sequence if the rebooting
after the power failure is not to last hours or even days. It is
also clear that the monitors and external accessories such as
hard drives, tape drives and active network components like
switches or routers must be taken into account.
3
2. The next step is to determine whether any subsystems need
special protection.
4
3. The complex power S* (VA value) of each piece of equipment
is to be determined.
4. Is any system or network expansion foreseen in the not too
distant future?
5. The UPS is to be rated according to the addition of all the VA
values.
5.3 Is the UPS functioning correctly?
Modern UPS units are equipped with an automatic battery test
that is performed once or twice a month, but the customer should
perform his own test once a year in order to verify the capacity
of the batteries. Customers of GE can leave this to GE specialists
who do the test via the internet. A battery test monitors how the
battery voltage behaves with time and load. If the voltage sinks too
quickly, the batteries are defective. The time taken for the batteries
to discharge must not be lower than the specified autonomy time.
If this is the case, all the batteries must be replaced.
5.4 Is the UPS system adapted to the level of
protection needed?
UPS systems must fulfil very different needs depending on the
application. In practice different security levels are distinguished
in UPS systems. For example there are computer users who
simply want to protect their computer from crashing in the case
of a power failure. Such users are at the lowest level of security.
A highly complex computer network with a central server is at a
high security level. This network must remain in operation 365
days a year and round the clock in spite of power failures, and any
other problems.
Type of UPS?
of the UPS?
Characteristics
Off-line:
Dependent on utility power with
regard to tension and frequency,
switchover time 2–8 ms. Battery
operation needed in case of undervoltage. Square, trapezium or
sine wave voltage, shutdown and
diagnosis software optional.
Active Standby or Line-Interactive:
Dependent on utility power with
regard to frequency, over- and
under-voltage is adjusted in
steps using the auto-transformer.
Switchover time 2–8 ms. Square,
trapezium or sine wave voltage,
shutdown and diagnosis software
optional.
Double conversion (VFI):
Double voltage conversion, loadstable sine voltage independent
of utility power, quartz-controlled
frequency, no switchover time,
galvanic separation from utility
power.
Double conversion (VFI) with RPA:
Double voltage conversion,
load-stable sinusoidal voltage
independent of utility power, quartzcontrolled frequency, no switchover
time, galvanic separation from the
utility power, parallel redundancy,
upgradeable with regard to power
and autonomy, system exchange
without shutdown.
Protection from what?
Used for what?
Protection from power failures,
NO protection from overvoltage. Single work place.
Protection from power failures,
and from over-voltage. Single
workplace and small multipleworkplace situations with
network.
Protection from power
failures and all faults in the
utility power, such as overand under-voltage, distorted
frequency and spikes. Server
and multiple workplace with
network.
Protection as in 3 with
additional system availability
thanks to Redundant Parallel
Architecture (RPA). Computing
centres, internet providers,
mainframes, demanding
multiple workplace computer
systems, very sensitive
industrial applications.
5.5 Protection from lightning
A direct impact by lightning is very rare, but the spin-off effects of
lightning are devastating none the less. Once lightning current has
found its way into the building, the computer network is going to
suffer major damage. However, as just mentioned, a direct lightning
impact is very rare. Most cases of damage stem from indirect
impact, i.e. when lightning strikes somewhere in the vicinity. These
can be prevented, because an electrical installation or a computer
system can be protected, even from a direct impact. But this is
not possible simply by using a UPS. Over-voltage because of
lightning (far or near) can reach the building via the utility power
network, the telephone line or data-cables. In order to protect a
building from this, several options are available: a good potential
equalisation arrangement, Surge Protection Devices (SPD), and of
course a good lightning trap.
5.6 And finally
Not all UPS are of the same kind and quality. Many users have
had to learn this the hard way. In the case of complex systems
in the computer sector and industry in general, basic electrical
knowledge is no longer sufficient to chose a reliable – and at the
same time adaptable – solution or even to evaluate it. Reliability,
professionalism and performance of the manufacturer are at least
as important as the technical specification of the system itself.
* Where the active power P is indicated, S is approximately 1,4 x P. The
result is the approximate VA value. Or, where the voltage and current
consumption is known, then S = U x I.
070209-v2
The Digital World and Electrical Power Supply
103
Industry Innovations
Featured Innovation
Intuitive and Intelligent Feeder Protection
Multilin 350 Feeder Protection Relay
GE Digital Energy – Multilin
www.GEMultilin.com/350
The Multilin 350 Feeder Protection System provides the utility power distribution & industrial
industries with a technologically advanced, easy-to-use, and intuitive overcurrent (50/51)
protection relay. The Multilin 350 performs primary circuit protection of medium voltage
distribution feeders. The robust Multilin 350 streamlines user workflow processes and simplifies
engineering tasks such as configuration, wiring, testing, commissioning, and maintenance via
advanced communications and enhanced diagnostics. Mechanically designed for effortless
draw-out, to eliminate re-wiring, the Multilin 350 enables fast installation, simplified retrofit and
reduced lifecycle cost.
Measure Low Resistance from Anywhere
Heavy Duty Low Resistance Ohmmeter
Megger
www.megger.com
Equally at home in the laboratory, the workshop or in the field, on the bench or on the ground,
Megger´s new heavy duty DLRO10HD low resistance ohmmeter combines rugged construction
with accuracy and ease of use. It features an internal rechargeable battery and can also operate
from a mains supply, even if the battery is completely flat. Like all models in the DLRO10 range,
the DLRO10HD, when used with optional terminal insulating test leads, is rated CATIII, 300 V in line
with IEC61010.
Uninterruptible Power & Energy Efficiency for Data Centers
750 kVA UPS System
GE Digital Energy
www.GEDigitalEnergy.com/pq
GE Digital Energy’s 750kVA Uninterruptible Power Supply (UPS) has been ecomagination certified
after completing GE’s rigorous ecomagination certification Product Review (EPR) process. GE’s
750kVA UPS system helps customers reduce their energy consumption and costs. At a typical 50
percent load, GE’s 750kVA UPS system achieves efficiency of more than 94 percent, delivering
uninterruptible power to the booming global data center industry at higher efficiency than
competitors’ systems.
Increase Safety with Arc Proofing Tape
Scotch® Fire-Retardant Electric Arc Proofing Tape 77 & 77W
3M
www.3M.com/electrical
Scotch® Fire-Retardant Electric Arc Proofing Tape 77 and 77W are designed to protect all
types of electrical cables where exposed to potential failures of other energy cables. Its unique
formulation expands in fire to form a thick char buildup. This insulating shield protects the cables
and accessories from fault arc generated heat and flames.
105
Industry Innovations
Extend your Network’s Reach
MultiLink MC-E Media Converters
GE Digital Energy - Multilin
www.GEMultilin.com
The MultiLink MC-E series of media converters uses the power of optical fiber to extend your
Ethernet network to remote locations. Designed for harsh utility and industrial applications the
MultiLink MC-E provides robust and reliable connectivity.
Tested. Proven. Trusted. Cable Bus
Cable Bus Systems
MP Husky
www.mphusky.com
MP Husky engineers and manufactures Cable Bus systems that utilize continuous runs of cable
from termination to termination, eliminating the need for potentially faulty splices found in nonsegregated phase bus duct systems. And since MP Husky Cable Bus utilizes highly reliable outdoor
rated cables, moisture and weather do not damage the system. In fact, the system is designed to
allow air, moisture and other elements to flow through the system, which eliminates the need for
moisture canceling components such as filter breathers and heater strips. Optimal balance of the
conductors within the system, achieved by MP Husky’s exclusive Inductive Reactance Program,
results in the most balanced load on each conductor which prevents overloading and overheating
of the conductors. No splices, optimal ampere balance, air-cooled conductors, and fewer parts to
fail equal a more reliable electrical feeder system.
Enabling AMI with Industrial WiMAX Communications
MDS Mercury™ 3650
GE Digital Energy – MDS
www.GEMDS.com/mercury
The MDS Mercury™ 3650 is a highly secure, industrial-grade WiMAX platform for creating wireless
communications to support utility’s Advanced Metering Infrastructures (AMI) that are designed
to give electrical consumers the ability to monitor electrical usage and manage costs in near
real-time. The WiMAX enabled MDS Mercury is also ideal for mission critical, industrial and public
safety applications including SCADA, Distributed Automation devices, video, VoIP, mobile data, and
Intranet applications. With up to 9 Mbps of aggregate Ethernet throughput (or 800 kbps for nomadic
mobile deployments), and a choice of frequencies, MDS Mercury has the capacity and deployment
flexibility to facilitate your immediate and long-term requirements.
Submersible Underground Connectors
MUCI Series
SICAME Australia
www.sicame.com.au
Designed for water-proof, underground connection of 2, 3 or 4 cables, the MUCI Series of Submersible
Connectors provides reliable, secure connections. Suitable for copper or aluminum insulated
cables, single or double insulated from 6 to 50 sq. mm. This solution is ideal for connecting cables
in pillar boxes, buried connection pits and street light columns.
106
Industry Innovations
Universal tool for working with IEC 61850 IEDs
IEDScout
OMICRON electronics Corp. USA
www.omicronusa.com
IEDScout is a universal client to IEC 61850 servers (such as substation IEDs) and a publisher/
subscriber for GOOSE messages. It provides numerous useful functions needed in the substation or
the laboratory. As an IEC 61850 client it supports many functions, from generic reading/writing of
data attributes to using the self description of the IED and producing SCL files from it. It detects GOOSE
messages on the network and monitors them. The IEDScout also simulates GOOSE messages. With
IEDScout, the protection engineer has new options to enhance the depth and quality of testing.
Compact, affordable Thermal Imaging
MikroSHOT
LumaSense
www.lumasenseinc.com
The MikroSHOT is the latest offering from LumaSense Technologies’ Mikron Infrared thermal imaging
product line. This fully radiometric thermal imager allows for affordable, pocket-sized portability with
capabilities normally found in larger, more expensive thermal imagers. The MikroSHOT’s Thermalon-Visible mode allows radiometric temperature data to be displayed directly on the visible image.
The MikroSHOT is lightweight (10.5 ounces) and uses off-the-shelf batteries (AC adapter included). Its
2.7-inch display and 160x120 pixel image resolution allow easy viewing of images. The MikroSHOT
has a measuring range of minus 4 F to 662 F, operating temperature range of 5 F to 122 F, color
alarm and autofocus at distances 1.4 yards to infinity. The SD card, USB and video output capability
allow for convenient, quick analysis of the JPEG-format data on a laptop or other mobile device
using common software. MikroSpec 4.0 software is included for image analysis and reporting.
Rugged Power Sensing for Utilities
ITI Outdoor MV Instrument Transformers
GE Digital Energy – ITI
www.GEDigitalEnergy.com/ITI
MV outdoor current & voltage transformers are now available at 60 Hz from 15-35kV. HCEP Epoxy
and Automated Pressure Gelation are combined to produce the highest quality and reliability for
outdoor applications. GE combines superior design and advanced testing with HCEP resin and APG
process technology, to produce reliable Outdoor Medium Voltage Instrument Transformers that
meet IEEE C57.13-1993 and CSA CAN3-C13 insulation level standards, and IEEE C12.11 dimension
standards.
Expanded Design tools for Infrastructure Modeling
Autodesk 2010 line of 2-D & 3-D Design & Engineering Software
Autodesk
www.autodesk.com
Autodesk announces its new 2010 line of 2-D and 3-D design and engineering software. More
than 50 new products offer new features and functionality, as well as improved tools for Digital
Prototyping, Building Information Modeling (BMI), Infrastructure Modeling, sustainable design and
analysis, which will help architects, engineers and designers meet increasing commercial and public
sector demand for more energy-efficient buildings, products and infrastructure. The 2010 software
for Infrastructure Modeling includes new features that enable suers to more easily aggregate
multiple sources of data, improving the design of smart electric utility grids, making planning city
projects easier, and enabling more efficient design and repair of highways.
107
Upcoming Events
DataCenterDynamics
Jul 17 – 18
San Francisco, California, United States
As a focal point for end-users, consultants and solution providers, the DatacenterDynamics Conference
& Expo is the Bay Area’s largest gathering of professionals involved in the design, build and operational
management of 24/7 mission critical IT facilities. It is an unrivalled education & networking opportunity for
the industry, where the regular audience is characterized by senior representatives of the of the financial,
medical, media, service provider, outsourcing/managed services, and other Fortune 500 companies
operating throughout the Bay Area, Silicon Valley and the Western US.
Hilton San Fransico
www.datacenterdynamics.com
Visit GE Digital Energy – Power Quality, in the Exhibition Area
ITE
Aug 9 – 12
San Antonio, Texas, United States
Join nearly 1,000 transportation professionals as they exchange ideas on transportation issues. Highlights
include numerous technical sessions where you can share your perspective in the open dialogue between
the audience and panelist(s) in the conversation circle sessions.
Attending this growing annual event will give you access to dynamic paper presentations, technical tours
and professional development seminars. Stay abreast of the newest technologies and services at ITE’s
Transportation Products and Services Exhibit where displays from the public and private sectors.
Henry B. Gonzalez Convention Center
www.ite.org
Visit GE Digital Energy – MDS, in booth #121
IMSA
Aug 18 – 26
Orlando, Florida, United States
The 11th Annual IMSA Conference and 32nd Annual School is dedicated to providing quality certification
programs for the safe installation, operation, and maintenance of public safety systems; delivering value
for members by providing the latest information and education in the industry.
Omni Resort at Champions Gate
http://www.imsasafety.org/2009conf/2009conf.html
Visit GE Digital Energy – MDS, in booth #615
108
Upcoming Events
Protection & Automation
Sep 7 – 10
Moscow, Russia
The Actual Trends in Development of Power System Protection and Automation conference discusses the
current state and prospects of architecture development, of design principles and operation algorithms
of the relay protection and emergency control. Emphasis will be placed on main tools and techniques to
promote efficiency and reliability of the relay protection, automation, and emergency control systems.
The conference program will include a plenary session, paper sessions, poster session, round table and
manufacturers’ special presentations.
Hotel Borodino
http://www.relayprotect.ru/en/index.htm
GE Digital Energy – Multilin, Papers to be Presented:
• Problems and Solutions of Line Differential Application in Cable Transformer Protection
• New Approach in Functionality and Testing for HV Capacitor Bank Protection
• Evaluation of the Ground Operating Current in Industrial Systems with Network Distribution through
MV Cables
• Reducing Conventional Copper Signaling in High Voltage Substations with IEC 61850 Process Bus
System
• Improvements in Power System Integrity Protection Schemes
Visit GE Digital Energy – Multilin, in the exhibition area
UTC Canada
Sep 14 - 17
Halifax, Nova Scotia, Canada
UTC Canada is a trade association focused on addressing the critical telecommunications issues for
utilities and energy companies in Canada and the providers of telecommunications infrastructure or
information technology services. UTC Canada works to bridge the issues that impact both the parent
utilities and their competitive telecom subsidiaries as an advocate for the telecom and IT interests of all
Canadian electric, gas, and water utilities, and oil and gas pipelines.
Westin Nova Scotian
www.utccanada.org
Visit GE Digital Energy, MDS & Lentronics, in booth #200, 202
IEEE IAS PCIC
Sep 14 - 16
Anaheim, California, United States
The IEEE IAS PCIC is an international forum for the exchange of electrical applications and technology
related to the petroleum and chemical industry. The annual conference is rotated across North American
locations of industry strength to attract national and international participation. User, manufacturer,
consultant and contractor participation is encouraged to strengthen the conference technical base.
Hilton Anaheim
www.ieee-pcic.org
GE Digital Energy – Multilin, Papers to be Presented
• Reducing Arc Flash Risk with the Application of Protective Relays
• Ground Fault Protection for MV Bus Connected Generators
Other GE Papers to be Presented:
•
•
•
•
Arc Flash Mitigation by Fast Energy Capture
Revisions to IEEE 1068: Standard for the Repair of AC Motors in the Process Industries
A User’s Guide for Factory Testing of Large Motors: What Should Your Witness Expect?
Use of Thermal Network on Determining the Temperature Distribution Inside Electric Motors in Steady
State and Dynamic Conditions
• Core Loss Testing: A Good Procedure Gone Astray?
• Predicting Let-Through Arc Flash Energy for Current Limited Circuit Breakers
Visit us in the GE Hospitality Suite – Hilton Anaheim
• Nightly from Sunday, September 13th through Tuesday, September 15th - Lainai Delux Suite
109
Upcoming Events
ASGMT
Sep 21 - 24
Houston, Texas, United States
The American School of Gas Measurement Technology (ASGMT) is the largest gas measurement school
in the United States devoted to natural gas measurement, pressure regulation, flow control, and
measurement related arenas. The purpose of the ASGMT, the sponsoring associations, and the operating
companies within the petroleum and natural gas industry, is to provide instruction on technical subjects
for people in the industry.
In addition to the classes, leading industry manufacturers will exhibit the latest in products and services.
This offers an exceptional opportunity to see the latest solutions available to the natural gas industry.
Marriott Houston Westchase
www.ASGMT.com
Visit GE Digital Energy – MDS, in booth #105
ISA Expo
Oct 6 - 8
Houston, Texas, United States
ISA Expo is an exhibition for promoting Automation and Control technology. ISA is a leading, global, nonprofit organization that is setting the standard for automation by helping over 30,000 worldwide members
and other professionals solve difficult technical problems. Based on research, ISA develops standards;
certifies industry professionals; provides education and training; publishes books and technical articles;
and hosts the largest conference and exhibition for automation professionals in the western hemisphere.
ISA is the founding sponsor of the Automation Federation.
Reliant Center
www.isa.org
Visit GE Digital Energy, in booth #1243
GCC Power 09
Oct 19 - 21
Riyadh, Kingdom of Saudi Arabia
The conference will offer delegates the opportunity to discuss the latest trends, challenges, developments
and strategies to meet the region’s rapidly expanding energy needs through a series of panel and
technical sessions. Government decision-makers, leading regional and international power companies
and respected industry advisors will attend to share past experiences, exchange ideas and get up to date
on the latest scientific research. The conference will also provide delegates with a unique opportunity to
enhance cooperation and collaboration throughout the region and will promote competitiveness and
efficiency in the electricity industry.
Marriott Hotel
http://www.gcc-cigre-power.com/
GE Digital Energy – Multilin, Papers to be Presented:
• Reducing Conventional Copper Signaling in High Voltage Substations with
IEC 61850 Process Bus System
Visit GE Digital Energy – Multilin, in the exhibition area
110
Upcoming Events
APAP 2009 & CIGRE B5
Oct 18 - 24
Jeju, Korea
The purpose of Advanced Power System Automation and Protection (APAP) 2009 is to invite the researchers,
engineers, and experts in power system automation and protection filed, and to provide an opportunity
to share their experiences and knowledge. APAP2009 is specially devoted to the advanced protection and
automation technology in power systems, but not limited to. I believe that the conference will establish a
clear goal and direction of the researches in this field and make a contribution to develop the protection
and automation technology for next-generation power systems.
The mission of the Study Committee B5 is to facilitate and promote the progress of engineering and the
international exchange of information and knowledge in the field of protection and automation. SC B5
annual meeting and colloquium 2009 in Jeju, Korea will provide a platform for the experts, scholars and
engineers to exchange experience and share knowledge in three preferential subjects - Strategies for
the Lifetime Maintenance of SAS Systems, Protection & Control of FACT devices and impact on Protection
Systems and Wide Area Monitoring, Control & Protection Technologies.
Lotte Hotel Jeju
www.apap2009.org
www.cigreb5korea.org/english/portal.php
GE Digital Energy – Multilin Papers to be Presented at APAP 2009:
• Design and Implementation of an Industrial Facility Islanding and Load Shed System
• Impact of Frequency Deviation on Protection Functions
• Designing Copper Wiring out of High Voltage Substations: A Practical Solution and Actual Installation
GE Digital Energy – Multilin Papers to be Presented at CIGRE B5 Annual Meeting:
• Improvements in Power System Integrity Protection Schemes
• Reducing the Costs of Periodic Maintenance of Secondary Systems in High Voltage Substations with
the use of Process Bus
• The Impact of Digital Technology on the Maintenance of Substation Automation Systems
• Lifetime Management of Relay Settings
WPRC 2009
Oct 20 - 22
Spokane, Washington, United States
The Western Protective Relay Conference (WPRC) is an education forum for the presentation and discussion
of broad and detailed technical aspects of protective relaying and relayed subjects. This forum allows
participants to learn and apply advanced technologies to prevents electrical power failures.
Spokane Convention Center
http://capps.wsu.edu/conferences/wprc/
GE Digital Energy – Multilin Papers to be Presented:
• Reliability of Protection Systems: What are the Real Concerns
• Enhanced Algorithm for Motor Rotor Broken Bar Detection
• Designing Copper Control Wiring Out of High Voltage Substations: A Practical Solution and Actual
Installation
• Fault Locator Based on Line Current Differential Relay Sychronized Measurements
• Fully Utilizing IED Capability to Reduce Wiring
Visit us in the GE Hospitality Suite – Red Lion Hotel at the Park – Room
5009/5010
• Nightly October 20th through 22nd – 6:00 pm to 10:00 pm
111
Upcoming Events
MIPSYCON
Nov 3-5
Minneapolis, Minnesota, United States
This conference provides electric utility engineers and consultants the opportunity to stay abreast of
today’s power system technology. The conference emphasizes the unique challenges faced by electric
utilities in the Midwest United States. The conference also serves as a forum for power engineers to
meet their colleagues from other utilities to discuss mutual concerns. Topics include substations, utility
industry futures, delivery systems, project management, relaying, distribution automation and distributed
resources.
Earle Brown Heritage Center
http://www.cce.umn.edu/conferences/mpsc/index.html
Visit GE Digital Energy – Multilin in our Hospitality Suite
EUTC 2009
Nov 3 – 6
Budapest, Hungary
Technology is rapidly changing the role of telecommunications in Europe’s electric, gas and water utilities,
energy companies, and other critical infrastructure companies. Many are using their vast experience in
building and managing sophisticated telecommunications networks to enter Europe’s new competitive
telecoms markets. Many are also facing issues introducing new wireless communications systems and
managing internal telecoms businesses in a shared services environment. To meet this need, the Utilities
Telecommunications Council created a uniquely European program that builds on UTC’s 60 years of
experience, existing strengths and services. UTC’s programs in Europe is led by Europeans, designed for
Europeans, and is uniquely European in focus.
Corinthia Hotel Budapest
http://www.eutc2009.utc.org/content/eutc-2009
Visit GE Digital Energy – Multilin in the exhibition area
g
Digital Energy
Multilin
The right career is waiting for you
We protect and connect the world’s critical equipment to
ensure safe, reliable power
With operations in over 160 countries worldwide, GE employees have an
unparalleled foundation on which to build their careers and abilities.
We are seeking innovative, self-motivated, imaginative and entrepreneurial
candidates to join our fast-paced, technology-forward environment in:
• Protection Engineering
• Product & Program Management
• Hardware & Firmware Design
• System & Application Engineering
• Sales Application Engineering
To join our dynamic organization please forwards your resume to [email protected]
112
070209-v1
Advanced Training
GE Multilin 2009/2010 Course Calendar
Comprehensive Training Solutions for
Protection, Control and Automation
SCHEDULED COURSES IN NORTH AMERICA
Courses for 2009/2010
CEU JUL
Tuition* Credits
Smart Grid: From Basics to
Practical Applications
(Rochester, NY)
$2,400
2.1
Fundamentals of Modern
Protective Relaying
$2,400
2.8
Introduction to the IEC61850
Protocol
$2,400
2.1
Distribution Protection Principles
& Relaying
$1,800
2.1
Motor Protection Principles &
Relaying
$1,800
2.1
UR Platform
$1,800
2.1
UR Advanced Applications
$3,000
3.5
Enervista Software Suite
Integration
$600
0.7
MM300 2 Days Hands-on
$1,200
1.4
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
JUN
22-24
20-23
14-17
9-12
18-21
5-7
19-22
17-19
9-11
6-8
17-19
14-16
17-19
1-3
9-11
8-10
17-19
12-14
26-30
18
21-23
10-14
13
18
16-17
15-16
All North American courses are located in Markham, Ontario, Canada unless otherwise stated
*Tuition quoted in US dollars
SCHEDULED COURSES IN EUROPE
Courses for 2009/2010
CEU JUL
Tuition* Credits
AUG
SEP
OCT
NOV
DEC
JAN
FEB
MAR
APR
MAY
UR Advanced Applications
$2,400
2.8
18-22
21-25
15-19
17-21
UR Platform
$1,800
2.1
13-15
16-18
10-12
12-14
Distribution Protection Principles
& Relaying
$1,800
2.1
Fundamentals of Modern
Protective Relaying
$2,400
2.8
Motor Management Relays
$1,800
2.1
F650 Platform
$1,800
2.1
20-22
Introduction to the IEC61850
Protocol
$1,800
2.1
23-24
10-12
7-9
9-11
8-12
24-26
7-10
21-23
23-25
JUN
23-25
19-21
21-23
22-23
All European courses are located in Bilbao, Spain unless otherwise stated
*Tuition quoted in US dollars
Course dates are subject to change. Please visit our website at www.GEMultilin.com/training for the most up-to-date schedule.
113
Content Index
Articles
Advertiser Listings
Application of Phasor Measurement Units for Disturbance
Recording
MP Husky
7
M. Adamiak, R. Hunt
Smart Grid: The Road Ahead
Megger
15
L. Sollecito
Developing the Smart Grid Business Case
GE Digital Energy - Lentronics
An Enterprise Information Architecture Based on SOA Enables a
Smarter Grid
25
www.GEDigitalEnergy.com/Lentronics
Key Smart Grid Applications
29
37
41
57
GE Energy
GE Digital Energy - Power Quality
System Control
J. Cupp, M. Beehler
IEC 61850 Communication Networks and Systems in
Substations
Omicron
61
Enhanced Security and Dependability in Process Bus Protection
Systems
69
D. McGinn, V. Muthukrishnan, W. Wang
AEP - Process Bus Replaces Copper
85
J. Burger, D. Krummen, J. Abele
The Digital World and Electrical Power Supply
R. Kleger
36
56
89
84
www.GEDigitalEnergy.com/PQ
www.systemcontrol.com
M. Adamiak, D. Baigent, R. Mackiewicz
114
VIRELEC Ltd.
www.GEEnergy.com
V. Madani, G. Duru, M. Adamiak
Implementing Smart Grid Communications
35
www.virelec.com
J. Fan, S. Borlase
A Paradigm Shift in Protection, Control and Substation
Automation Strategy
GE Digital Energy - MDS
20
www.GEDigitalEnergy.com/MDS
B. Flynn
The Evolution of Distribution
14, 60, IBC
www.GEDigitalEnergy.com/Multilin
J. McDonald
M. Van Helton
6
www.Megger.com
GE Digital Energy - Multilin
21
IFC
www.mphusky.com/ge
www.omicronusa.com
104
BC
Unparalleled Control
Whether your application requires advanced substation automation, complete bay
protection and control or multi-stage load shedding capabilities, the Multilin C90Plus
Controller is simply the most powerful solution available for your utility substation or
industrial power system applications. As a single-platform custom engineering tool set,
the Multilin C90Plus Controller features true convergence of functions, including advanced
automation and control, digital fault recording, comprehensive communications and
extensive local HMI capabilities, delivering unparalleled flexibility for the design of your
custom applications.
To Learn more visit us at: www.gemultilin.com/C90P
C90Plus Controller
g
Digital Energy
Multilin
GE Multilin
www.GEMultilin.com/C90P
[email protected]
Worldwide
Tel: 905-294-6222
Worldwide
Tel: 905-294-6222
North America
Tel: 1-800-547-8629
Europe/MiddleEast/ Africa
Tel: +34 94 485 88 00
North America
Tel: 1-800-547-8629
Europe/MiddleEast/Africa
Tel: +34 94 485 88 00
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement