Advanced Protection, Automation,

and Control Functions

Bogdan Kasztenny and Normann Fischer

Schweitzer Engineering Laboratories, Inc.

Revised edition released August 2010

Originally presented at the

1st Annual Protection, Automation and Control World Conference, June 2010
 1

Advanced Protection, Automation,

and Control Functions
Bogdan Kasztenny and Normann Fischer, Schweitzer Engineering Laboratories, Inc.

Abstract—This paper reviews the historical background, directional elements, capacitive coupling voltage transformer
present state, and future challenges and opportunities of (CCVT) transient detection logic for distance protection, load
microprocessor-based power system protection, control, and encroachment and blinders, and adaptive reactance elements
automation systems. The paper focuses on protection and control
functions in terms of improved performance, new protection
for ground distance protection are good examples of
principles, addressing new challenges, and simplifying protection enhancements that resulted from lifting the
application. It excludes many associated fields, such as the role of constraints of previous relay technologies.
communications, adaptive protection, or system integrity Modern relay implementations have access to more
protection schemes. information: higher sampling rates, more direct inputs, and
more auxiliary signals delivered via communications from
I. INTRODUCTION other devices. These relays have more processing power to
This paper reviews advancements and the state of the art in utilize the available information, yielding better performance
power system protection, automation, and control. We focus and facilitating new functions.
on protection first and foremost and proceed briefly to New microprocessor-based relays are designed for lower
automation and control. This paper covers progress in internal latencies, thus overcoming the inherent disadvantage
protection principles and ways of implementing them in of using sequential processing and incurring internal delays as
microprocessor-based relays. Our intent is to summarize the compared with analog relays. Modern relays are designed with
evolution of the last few decades, the state of the art today, a fair amount of parallel processing, using field-programmable
and some opportunities going forward. We narrow our gate arrays (FPGAs) in order to speed up processing and
discussion to just the protection and control (P&C) functions, provide for deterministic execution of computationally
leaving the many related topics such as IEC 61850 and the intensive functions. Internal relay architectures, while
role of communication, synchrophasors, wide-area control, designed for simplicity, are heavily optimized for
adaptive protection, and so on to the other reference papers of performance. Integrating protection, automation, recording,
the inaugural Protection, Automation and Control World metering, human-machine interface (HMI), and communi-
Conference. cations functions, these devices are designed for guaranteed
performance under a variety of activity patterns in any of the
II. PROTECTION respective functional areas.
Shaping the operating characteristics via device The change from the mechanical detection of problems to
construction substantially limited the electromechanical and calculations and storage of values led to the possibility to
static generations of P&C devices. Microprocessor-based perform self-tests, and the art of self-checking of
relays brought freedom in forming the operating character- microprocessor-based relays became a mature field of
istics by way of real-time calculations and lifted other engineering. The abilities to detect an internal failure, fail
construction-driven constraints, inviting designers to go back gracefully (“fail-safe”), and alert the user of the problem are
to first principles and improve on existing protection considerable benefits that have been available since the early
functions, as well as to invent new ways of detecting faults days of microprocessor-based relays. Avoiding undesired
and other abnormal conditions in power systems. operations and preventing hidden failures to operate have a
Microprocessor-based relays delivered a wealth of new significant and direct impact on the overall performance of the
functions—built-in digital fault recording, sequential event protection system. Today, a renewed attention is given to this
recording, metering, multiple settings groups, easier adaptivity aspect with the goal to improve and optimize periodic testing
to changing system conditions, fault location, and digital and maintenance activities for P&C systems. Relying on
communication, to name only a few. Remarkable cost and natural events, such as tests of current and voltage
space savings accelerated adoption of the new technology. transformers (CTs and VTs) and circuit breakers, the
Early implementations were limited by the small amount of enhanced relay self-monitoring can facilitate a “run-to-fail”
information available to the relays and their limited ability to maintenance strategy, where a P&C system is left operational
process this information. Designers skillfully worked around until it reports problems. Avoiding human activities and
the low sampling rates and scarce processing power in order to associated errors would yield a potentially better performance
deliver sophisticated algorithms with many performance than the periodic testing approach. This strategy is one of the
improvements over the older generations of relays. Adaptive new, yet to be fully realized advantages of microprocessor-
based protection technology.
 2

As human resources at user organizations become scarce, System stability margins continue to become tighter, and
new emphasis is given to simplicity. While integration of the this requires that faults be cleared as rapidly as possible. When
many traditional functions (from protection, through a circuit breaker fails to operate, all sources that contribute to
recording, metering, and remote terminal units [RTUs], to the fault have to be disconnected in order to maintain the
automation and control) in a single device brings size, cost, integrity of the power system. Because disconnecting sources
and performance benefits, the device itself appears more has a drastic effect on the system, it is best to delay this action
complex. Microprocessor-based technology opens new as long as possible, giving the circuit breaker as much time as
opportunities in this area. A power swing detection element possible to clear the fault. Therefore, if the primary breaker
that requires no settings to track the rate of change of the operates, we want to reset the breaker failure initiate
swing center voltage is a good example of a better protection commands as rapidly as possible. Using current magnitude to
principle that eliminates the need for some of the engineering reset the breaker failure typically introduces a delay in excess
required to apply the function. of one power cycle due to the latency of the filters. The reset
In the following sections, we elaborate on several key time can be reduced by 0.5 to 0.75 cycles in modern relays via
dimensions of the progress in microprocessor-based power fast processing of the raw current samples. The algorithm uses
system protection. the following principle to determine when the primary current
has been interrupted. As current flows through the breaker, the
A. More Information Enables Better Protection
current goes through a zero crossing every 0.5 cycle. The
Higher sampling rates with greater precision enable derivative of the current (di/dt) also has a zero crossing every
protection devices to obtain finer measurements more rapidly 0.5 cycle, but the zero crossing of the current derivative is
and more often. This leads to protection devices with faster offset by 0.25 cycle (90 degrees) to that of the zero crossing of
and more secure elements. The following are a few examples the current. Therefore, when current flows through a breaker,
of higher sampling rates leading to better protection a zero crossing occurs in either the current or the derivative of
algorithms. the current at least once every 0.25 cycle. If there is no zero
Consider a transformer differential relay. One of the crossing for 0.5 cycle, then the primary current has been
greatest challenges for any differential relay is securing the interrupted and the circuit breaker can be considered open.
differential element during external faults when one of the Using this method, it is possible to determine that a circuit
CTs goes into deep saturation. CTs that supply differential breaker is open at least 0.5 cycle faster than using the
elements do not go into saturation immediately after fault traditional current magnitude method.
inception but saturate sometime later. The first few In addition, higher sampling rates of the analog quantities
milliseconds (< 4 milliseconds) after a fault, CTs will and access to more digital information (digital inputs) at a
reproduce the primary current faithfully. A fault detection higher rate also create better and more flexible protection
algorithm that uses high-rate data together with the fact that functions. The following are some examples of how access to
CTs do not saturate immediately determines whether a fault is more digital inputs helps achieve better protection
internal or external to the protective zone within performance.
2 milliseconds after fault inception. The algorithm uses the Consider a multizone busbar system where zone
following principle to determine if the fault is outside the reconfiguration occurs frequently by switching feeders from
differential zone. If the fault is internal, both the differential one zone to the next while the feeder is under load. One of the
and the restraint quantities change instantaneously (both challenges for a busbar protective relay is to rapidly assign a
quantities experience an incremental change). For an external transferring feeder to the correct zone or zones during the
fault, only the restraint quantity changes instantaneously; the switching operation to maintain the integrity of the differential
differential quantity does not change. The differential quantity element. Earlier busbar protection schemes either disabled the
will experience a change only once one or more of the protection during the switching procedure or switched the CT
differential CTs go into saturation. When one or more CTs go secondary current through auxiliary relays. If the relay could
into saturation, the differential element begins to measure a not correctly determine which zone or zones to assign the
fictitious differential current. Therefore, it is essential that the feeder to during the switching operation, the protective relay
relay determine if the fault is external before one or more CTs was disabled. This is not ideal, because some busbar faults are
saturate. Once the fault detector detects an incremental change caused as a result of the switching. The idea behind switching
in the restraint quantity, it opens a 2-millisecond window and the CT secondary current through auxiliary relays is to
checks to see if the differential quantity changed during this replicate the busbar configuration using the status from the
time. If there is no incremental change in the differential isolator auxiliary switches. This is a good idea, except for the
current during this 2 millisecond window, the relay declares fact that this switches the CT secondary current. Should one of
the fault external and switches the relay into a high-security the auxiliary contacts fail, not only does the CT open-circuit
mode. This means that if the CTs saturate after this time, the but the bus relay also trips. The reason that these methods
differential element is in a secure mode and any fictitious were employed was because these earlier busbar protective
differential current will not negatively impact the security of relays did not have enough digital inputs to replicate the
the differential element. busbar configuration in the relay itself. The newer busbar
differential relays not only have enough digital inputs to
 3

monitor the status of the isolators, they also sample these Load current can significantly influence the performance of
inputs fast enough to replicate the busbar configuration in the faulted phase selection logic and directional elements, which,
protective relay in real time. This has the advantage that the in turn, influences the performance of the distance elements.
relay always assigns the transferring feeder to the correct zone Protection designers have known for some time that by using
or zones and maintains high-speed bus protection, and the CT incremental quantities, the influence of load on these elements
secondaries do not need to be switched. can be removed. To extract the pure fault current from the
Combining a higher sampling rate with a greater number of total fault current, the prefault current must be subtracted from
analog inputs allows for some very innovative protection the total fault current. In other words, the load current must be
schemes that not only offer greater flexibility but also provide removed from the total fault current. To accomplish this, the
system security while breakers are taken out of service and distance relay stores the prefault voltages and currents. These
tested. prefault data are stored in buffers approximately 2 to 3 cycles
A breaker-and-a-half scheme is commonly used in deep. The information in these buffers is updated every
transmission systems to provide greater feeder availability; processing interval so that the relay tracks the system load
however, it is often necessary to remove a breaker from dynamically. The relay detects a fault if the incremental torque
service and test the specific breaker without interfering with of any of the fault loops (produced by the product of the
the healthy breaker or the feeder. Newer protective relays no incremental voltage and current) exceeds a preset threshold.
longer require the CTs from the individual breakers to be Once the relay detects a fault condition, the information in the
paralleled outside of the relay but allow them to be fed into the buffers is frozen and stored as prefault values so that a
relay individually. These relays are equipped with source reference point is maintained. The faulted phase or phases are
selection logic, where the logic determines how the currents determined by comparing the incremental torques against one
from the two individual CTs are presented to the protection another. The faulted phase (or phases) is typically the one with
functions. In the combined mode, the logic combines the the greatest incremental torque. The fault direction is
individual currents and presents them as the line currents to determined by the sign of the incremental torque; if the
the relay. Should one of the breakers be taken out of service, incremental torque is negative, the fault is declared forward,
an external input informs the source selection within the relay and if it is positive, the fault is declared reverse. In this
that the line current is no longer a combination of the breaker manner, the fault direction and faulted phase(s) are readily
currents but rather a current from the remaining in-service identified, which results in a robust and faster protective relay.
breaker (this is a simple form of zone selection). The source With systems being pushed ever closer to their limits, a
selection logic then routes the current from the in-service loss of a major transmission line may result in an unbalance in
breaker to the line protection function in the relay without the power generated versus the power dissipated in a part of
interfering with the protection of the feeder. This allows the the power system, manifesting itself as a power swing (out of
user to perform maintenance on the out-of-service breaker step). It is critical at this time to keep the power system from
without interfering with the protection performance of the breaking apart and ending up with unstable islands. To
feeder. The user can inject current into the out-of-service CT accomplish this, protective relays employed on transmission
and verify the CT measurements by using the meter function systems are equipped with out-of-step protection that will
on the relay. Note the relay still monitors the current in the detect this power swing and prevent the distance element from
out-of-service breaker; it simply does not include it in the line inadvertently tripping during this condition and leading to a
protection functions. Once the maintenance is complete, the further demise of the stability of the power system. Setting the
breaker can be simply switched back into service, and the out-of-step protection properly requires engineers to run
source selection logic will once again combine the current extensive stability studies to determine the swing rate of the
from the two CTs and present it to the protection function. An power system under different operating conditions so that the
added benefit of not combining the current external to the out-of-step logic can clearly distinguish between a fault
relay is that the relay can monitor the wear and tear on each condition and a swing condition. This task proves particularly
individual breaker, as well as offer breaker failure protection challenging when long lines with heavy load are involved.
to each individual breaker. Protection engineers have challenged themselves to develop
new methods of detecting out-of-step conditions without the
B. More Processing Power Enables New Protection Methods
user having to run extensive system studies. Engineers came
Many original protection methods have been limited by up with several solutions to this problem. One solution is to
construction of protection devices. Early microprocessor- have the relay monitor the rate of change of the positive-
based relays carried forward principles derived under sequence impedance. Under normal conditions, the positive
limitations of electromechanical and static technologies. impedance is stationary or varies very little; for a fault
Today, increased processing power allows sophisticated condition, the impedance changes suddenly and then reaches a
protection principles, limited only by the laws of physics and steady-state value (the fault impedance). During a power
imagination of the designers. Examples of how the power in swing, the impedance changes constantly. So to determine a
newer microprocessor-based relays liberated the ideas of power swing, the new logic takes the difference between the
protective relay designers are discussed below. positive impedance now (Z1k) and subtracts it from the
positive-sequence impedance from the previous processing
 4

interval (Z1k–1). It then computes the rate of change of percentage (x) of actual voltage (Vk) and a certain percentage
impedance because the relay knows the time difference (1–x) of the memory voltage (VMEM). If a symmetrical fault
between processing intervals (Δt). If the rate of change is occurs and the voltage measured by the relay falls below a
below a factory-optimized threshold, the relay declares an out- lower limit, the relay switches to the second time constant
of-step condition and blocks the appropriate distance (long time constant) that almost exclusively uses memory
elements. The logic declares that the out-of-step condition has voltage (VMEM) to polarize the distance elements. This ensures
ended, once the rate of change drops below a threshold for a that the distance elements remain directionally stable for
period of time. three-phase faults close to the relay location. Similarly, when
Another method to detect a power swing is to monitor the the relay detects that a voltage inversion has occurred, it
swing center voltage. The swing center voltage is the voltage switches the memory voltage to the long time constant,
measured at the electrical center of a power system. Because a ensuring that the distance elements maintain their directional
relay cannot be positioned at the electrical center of a power integrity.
system (the swing center of a power system changes Protective relays depend on instrument transformers (VTs
constantly because of changing system conditions), a direct and CTs) for their data; therefore, any transient response by
relationship exists between the line current and the swing the instrument transformers impacts the performance of the
center voltage. Therefore, by monitoring the rate of change of protection device. CCVTs are known to not accurately
the angular difference between the voltage and the current, we replicate the primary voltage signal at the onset of a fault,
can monitor the rate of change of the swing center voltage (in because of the resonance between the stack capacitors (used
other words, we are monitoring the power factor). Under for voltage division) and the tuning inductor. This transient
steady-state conditions, the angle between the voltage and causes the voltage output to be of a lower frequency than the
current remains approximately constant. When a fault occurs system frequency, which ultimately translates into a voltage
on the power system, the angle between the voltage and with a lower magnitude. For the distance relay, this voltage
current approaches that of the line angle and remains there. reduction translates into an overreaching phenomenon,
When the system experiences a power swing, the angle meaning that a Zone 1 element may operate for an external
between the voltage and current changes constantly; the relay fault. Protection engineers know that there is a direct
uses this information to determine the presence of a power relationship between the strength of the power system and the
swing. The relay calculates the angular difference between the CCVT transient. In addition, they know that there is a direct
voltage and current for the present processing interval (Θk), relationship between the fault current and the fault voltage.
and using the angular difference from the previous processing For a protective relay to detect a CCVT transient, the relay
interval (Θk–1), it calculates the change in the angles (ΔΘ) looks at the relationship between the voltage and the current.
between the processing intervals. Knowing the time difference If the relay detects that the voltage is lower than anticipated
between the processing interval (Δt), the relay can calculate for a particular fault current, the relay sets the CCVT blocking
the rate of change of the angle. If this rate of angle change is signal, which holds back the Zone 1 element. However, the
above a threshold, an out-of-step condition is declared and the relay does not want to hold back the Zone 1 element
appropriate distance elements are blocked. The relay removes indefinitely; studies have shown that CCVT transients last a
the block condition once the rate of change falls below a maximum of 1.5 to 2 cycles. A transient may decay within
threshold for a given period of time. 0.5 to 0.75 cycles, so it is desirable to reset the CCVT block
Series-compensated lines are becoming more and more signal as rapidly as possible once the CCVT transient has
prevalent in modern power systems as systems are pushed decayed sufficiently. This is done by computing the
closer to their limits. One challenge for distance relays is how smoothness of the distance element calculation. The
to detect and deal with a voltage inversion. A voltage smoothness detector continually calculates the difference
inversion occurs when the impedance between the VT feeding between the present distance calculation (m_calck) and the
the relay and the fault is capacitive instead of inductive. Under previous distance calculation (m_calck–1). During the transient
this condition, the voltage is 180 degrees out of phase (inverts) condition, the difference is greater than 10 percent; once the
with the prefault voltage. One method of detecting that a transient decays sufficiently, this difference becomes less than
voltage inversion occurred is to compare the phase angle of 10 percent and tends towards zero (settles toward the steady
the present voltage (Vk) to that 1 cycle previous (Vk–1 cycle). fault value). Once the smoothness detector determines that the
If no voltage inversion occurs, the angular difference between difference between three consecutive values is less than
the two signals is small (typically a few degrees). If a voltage 10 percent, the block signal is removed and the Zone 1
inversion does occur, the angular difference between these element is allowed to operate.
signals is greater than 90 degrees. What happens once the C. Microprocessor-Based Technology Allows Easy
relay has detected a voltage inversion? Most modern distance Adaptation
relays use memory voltage to polarize distance elements. The Providing protection by means of calculations,
memory voltage in state-of-the-art distance relays typically microprocessor-based relays allow for adaptive behavior. The
employs more than one time constant. For normal following are examples of how, by continuous calculation of
nonsymmetrical faults, the relays use memory voltage with a certain parameters, protective relay sensitivity can be
short time constant, meaning the relay uses a certain enhanced.
 5

Detecting ground faults on an ungrounded power system is time (e.g., 5 minutes) and uses it as a threshold. Because the
a trivial task when the fault resistance is low; however, as the relay updates this threshold every processing interval, the
fault resistance increases, this task becomes more complex threshold adapts to loads. If an arcing fault occurs, the
because the unbalance caused by the fault begins to approach interharmonic content will be much larger than the normal
the standing unbalance of the power system. On an feeder content, and the algorithm will count the number of
ungrounded power system, the standing unbalance is due to times that the adaptive threshold is crossed. At the same time,
the nonsymmetrical capacitance per phase to ground. The the adaptive threshold is frozen, because the relay suspects the
standing unbalance of the power system varies as the system presence of a fault on the feeder and does not want to corrupt
load varies. This unbalance manifests itself in the form of a the threshold with fault data. Should the interharmonic content
neutral (residual) current. Most protective relays applied on remain above the threshold, the relay will declare a fault on
these power systems use a fixed, user-settable threshold to the feeder.
monitor the magnitude of the residual current on specific Ungrounded shunt capacitor banks are protected by
feeders. If the magnitude exceeds the threshold, the relay measuring the potential difference between the voltage at the
declares a fault and either sets a local fault indication flag or busbar and a tap point somewhere on the stack. With no fault
sends it to a central location to indicate the presence of a fault condition on the system, the two ratio-matched voltages are
on the system/feeder. However, if the load variation on the the same. However, because of the unbalance of the power
feeder is quite large, the sensitivity of the protection has to be system and the tolerance in the capacitor cans, a potential
compromised in order to prevent nuisance fault indications. difference exists between the two measuring points. Shunt
This method has a further disadvantage in that if the fault was capacitor bank protection detects the amount of capacitor cans
of such a nature that it caused the system to become balanced that are either open- or short-circuited. To do this detection as
(i.e., the fault current canceled the standing unbalance), the sensitively and accurately as possible, the inherent difference
relay would not detect the presence of the fault. Modern between the two measuring voltages needs to be nulled out.
microprocessor-based relays are capable of measuring the This is often done by means of an adjustment factor (k); this
actual unbalance at any instant in time and can determine the factor adjusts the magnitude and phase of the tap point
threshold by taking the average standing unbalance over a voltage. The adjustment factor is usually set when the
period of time. This means the relay has a dynamic threshold capacitor bank is taken into service and is a settable static
that adjusts to load and provides maximum sensitivity and value. The issue with this setting is that it was determined
security to the feeder at all times. On these systems, the load based on the system conditions at that time; should system
varies slowly and in a predictable manner (e.g., processes are conditions change (i.e., the voltage at the busbar experiences a
switched on and off). Therefore, if the relay detects a sudden change in magnitude and phase due to a shift in load), the
change in the neutral current (irrespective of if it increases or potential difference between the two VTs will now no longer
decreases), the relay can indicate the presence of a fault. be zero, even with the adjustment. So in order to prevent an
Therefore, the relay can not only detect faults that result in the incorrect indication of the number of cans lost in the stack, the
neutral current exceeding the dynamic threshold, it will also pickup threshold of the alarming or tripping function has to be
detect faults that create a large change in the neutral current. raised, compromising the overall sensitivity of the bank
In recent years, the detection of high-impedance ground protection. To overcome this deficiency in the previous
faults on medium-voltage, solidly grounded power systems method, a third VT is introduced that measures the potential
has become a priority. These faults do not cause a threat to the difference between the neutral point of the ungrounded
power system but do pose a large threat to human life and capacitor bank and ground. Using this information and the
livestock alike. Because these faults cause very low ground voltage at the busbar, a shunt capacitor bank protective relay
current to flow, they are virtually indistinguishable from can now dynamically calculate the adjustment factor by
currents that flow because of normal load unbalance. measuring the standing unbalance of the system and
Conventional algorithms based on the magnitude of the compensating for it. In this manner, a user does not need to set
ground fault current cannot be used to detect these faults. One this threshold, and the k factor calculation automatically
of the characteristics of these types of faults is that they arc. adjusts for any system unbalance. This not only makes it
Arcing faults result in interharmonic currents; however, many easier for the user but also increases the sensitivity of the
power electronic loads also generate interharmonic currents. capacitor bank protection without compromising its security.
Therefore, for a protection device to distinguish between a
D. Design for Low Latencies Improves Protection Speed
normal condition and a fault, the relay must learn the feeder’s
normal interharmonic or noise current. One method used to Early microprocessor-based relays were limited by the
learn the feeder’s background or normal noise pattern uses the available processing power, and having to process information
sum of the difference current. In essence, this method takes sequentially, they were inherently disadvantaged compared
the current at time t (Ik) and subtracts it from the current with analog relays. Modern relays apply parallel processing
1 cycle previous (Ik–1cycle); essentially, it is a delta quantity and utilize powerful microprocessors and FPGAs, all within
calculation. This calculation removes all load current and optimized internal data handling architectures, allowing for far
harmonics from the signal and leaves only the interharmonics. shorter latencies.
The relay then averages these interharmonics over a period of
 6

An internal latency can be measured as the response time Internal architectures of modern relays allow achieving
of the relay, not counting the data window required by the latency times of 1 to 3 milliseconds.
protection principle to maintain selectivity and security. For Because of these diminishing returns, relay designers look
example, we may configure a relay to operate a contact from a for new ways to reduce the trip times. For example:
highly sensitive disturbance detector and measure the response • A solid-state, trip-rated output can provide a response
time between the change in the input current and the closure that is 2 to 3 milliseconds faster compared with a
of a trip-rated output. This time delay is a good measure of the mechanical output and another 3 to 5 milliseconds
performance of the relay architecture in terms of processing faster if the interposing/lockout relays are eliminated.
latency. • Because the critical fault-clearing time is of
The length of the data window required for selective paramount importance, improving breaker failure reset
operation is a measure of the power of the applied algorithms time is a safe way to reduce the overall trip time,
and logic constituting a given protection method. The shorter compared with attempting to make the trip decision
the data window, the less information it contains about the faster in the first place.
event, challenging the ability of the device to make the correct • Moving critical signals via digital communication can
decision. be done faster compared with analog signals,
Today’s algorithms are able to make reach-unconstrained considering the same level of security—as a result,
decisions using between 0.25 and 0.5 cycles of data. These integrating breaker failure and tripping functions in
reach-unconstrained functions include directional elements, one relay or sending breaker failure initiate signals via
fault identification logic, starting elements and disturbance communication can bring extra savings in the overall
detectors (if applied), overreaching distance functions, response time.
external fault detectors, current detectors of a breaker failure Early research into numerical protection focused on phasor
element, and so on. Reach-constrained elements, meaning estimation algorithms. Hundreds, if not thousands, of papers
elements that require relatively precise reach accuracy, can be have been written in search of the “holy grail” of speed and
implemented using between 0.5 and 1 cycles of data. These accuracy. Today’s design activities are multidimensional,
reach-constrained elements include directly tripping under- involving analysis of power system phenomena and charac-
reaching distance Zone 1, directly tripping instantaneous teristics as a basis for new and improved protection principles,
overcurrent elements, or differential elements. sophisticated signal processing, hardware and firmware co-
Advancements in protection algorithms, such as external design, and relay application practices.
fault detection guarding against CT saturation errors in a
differential function, allow shortening the data windows for E. Enhanced Self-Monitoring Improves Security and
decision making to 0.25 cycle or so, without impacting Availability
security. Protective relays are designed and manufactured to high
More sophisticated protection algorithms require more standards of reliability. Mean time between failures (MTBF)
processing power. The ability of the relay to execute its reaches 300 to 400 years for best-in-class relays. Still, there is
algorithms within a short scan cycle is a measure of the relay always a non-zero probability of internal component failure.
processing power. Built-in self-monitoring is designed to maximize security and
Today’s powerful relays optimize all three elements: avoid unintended operation by detecting internal problems
internal architectures to limit latencies inherent in digital under practical component failure scenarios. Therefore, the
implementations, protection algorithms to respond correctly MTBF viewed from the security perspective, meaning
using shorter data windows, and processing power to enable considering only failures that lead to unintended operations, is
more sophisticated algorithms. considerably better than 400 years for best-in-class relays.
The overall response time of the relay is the sum of the Detection and alarming on internal relay failures maximize
processing latency and the effective length of the data window availability by reducing hidden failures to operate and
required to make secure trip decisions. Both of these times shortening the mean time to repair (MTTR). A combination of
asymptotically approach their limits with diminishing returns. high MTBF numbers, low MTTR numbers, and enhanced
For example, the amount of research and extra processing security (bias to fail gracefully) brings exceptional availability
power required to improve a 30-millisecond distance function and security to the field of microprocessor-based protection.
to become a 28-millisecond distance function is negligible. While early generations of microprocessor-based relays
But the amount of research and processing power required to operated more slowly because of processing latencies and
improve the same distance function from a 10-millisecond were considered to be at a relative disadvantage compared
operating time to an 8-millisecond trip time most likely with analog technologies, self-monitoring is an inherent
doubles; going from 8 to 6 milliseconds probably quadruples advantage of the microprocessor-based technology. Both
the effort and processing power. The shorter the response processing capability and self-monitoring capabilities have
time, the less guaranteed is the positive operation. Typically, improved dramatically.
short response times are achievable under best conditions, Digital systems fail gracefully by nature and by design.
calling for a parallel, slower, but fully dependable algorithm. Data and code integrity checks, watchdogs, and other standard
and optimized integrity monitors ensure fail-safe operation of
 7

the digital subsystems of a microprocessor-based relay. role in short-haul communication and access to pole-top
Internal data buses are protected with strong data integrity devices outside of the substation environment.
(redundancy) codes. Power supply rails are constantly The ability to communicate information over a wider area
monitored to ensure digital subsystems remain digital and enables sophisticated automation and control functions.
never brown-out into a nondeterministic state where the built- Distribution automation is the key application example.
in safety mechanisms could be defeated. Tripping and control
C. Increased Role of Time-Synchronized Data
outputs are actuated using digital techniques, ensuring fail-
safe behavior even if the driving subsystem misbehaves. New applications are possible when utilizing time-aligned
Communications ports are protected with data integrity checks data. Synchronized measurements (synchrophasors for ac
and often include monitoring circuitries for fiber transceivers, quantities and time-stamped measurements for other data)
if used. allow new applications.
The analog interface of a modern relay is designed for Fast load-shedding schemes for industrial facilities
maximum reliability, with clean design and low component measuring and controlling the actual power balance in real
counts. Some degree of redundant measurements is often time are good examples of acting upon time-synchronized
employed to ensure failures in this area can be detected in a measurements in automation and control functions.
timely fashion to prevent undesired operations. The ability to execute a time-coordinated control action is
Robust self-monitoring, combined with state-of-the-art another novel application of time. Instead of executing a series
MTBF characteristics and tracking of natural events to prove of manual and automated control actions in an uncoordinated
CTs, VTs, cabling, and circuit breakers, enables a run-to-fail fashion, an intelligent control system can precalculate an
maintenance strategy. Either the periodic testing or optimum control action, assuming all control commands will
maintenance is considerably relaxed or even eliminated, be executed simultaneously, preload those actions (“recipes”)
delivering considerable savings to the user organizations, via communications to the appropriate controllers, and launch
addressing the human resources gap, and improving protection them based on time. For example, taking a line out of service
system performance by avoiding human errors introduced manually causes overvoltage or undervoltage conditions.
while maintaining the system. Upon opening the breaker, tap changers and other controllers
will eventually respond to maintain the voltage at lower levels.
III. AUTOMATION AND CONTROL Subsequently, the operator may take action to rectify the
voltage excursions on the high-voltage level, causing the
Automation and control functions benefited from many of
reversal of the previous actions of tap changers, capacitor
the same advancements as protection (more information
banks, and so on. All this can be avoided by launching the
available to the devices, more processing power). The
circuit breaker open command and the associated control
fundamental progress, however, is in communication—the
actions simultaneously based on time. The net effect is lower
ability to share more information among multiple devices to
wear of tap changers and other controllers and better power
gain a better understanding of the situation when deriving
quality.
control actions.
D. Availability of Computing Platforms
A. Integration of Protection and Control
Powerful computing platforms allow deployment of
Modern multifunction devices integrate protection and
sophisticated control and automation schemes, such as
many control and automation functions. This allows multiple
distribution automation. A combination of information from
savings (device count, engineering, documentation, wiring,
multiple devices giving visibility to the system state,
and commissioning).
affordable communication, and computing capabilities opens
As far as performance is considered, integration brings
new opportunities.
extra advantages. The automation and control functions
Distributed schemes are also possible and used. In these
residing in a device that performs protection and meas-
schemes, individual devices that collect the data and execute
urements have natural access to a wealth of internally
commands run the automation and control algorithms,
generated information and can benefit from it. For example,
typically within their programmable logic. Modern program-
fault location information may aid restoration.
mable logic engines embedded in protective relays allow
Complexity and user acceptance are the primary factors
mathematical calculations, including unsynchronized and
deciding on the degree of integration. Setup and
time-synchronized data (precise time alignment based on time
commissioning tools address this concern.
stamps).
B. Cost-Effective Communication
E. System-Level Control
Local intersubstation communication is a norm, including
Special protection schemes are beginning to play bigger
both point-to-point serial and Ethernet networks. Cost-
roles as countermeasures to lowering the stability margins in
efficient intrasubstation means become available primarily
the power system. Following the trend of standardization of
through multiple (shared) usage of high-bandwidth channels.
hardware and software, these schemes are recently built on
Modern radio solutions (spread spectrum) play an increasing
standard protection devices instead of low volume, slowly
maturing, special products. Applications of varying sizes
 8

emerge between the strictly local and system-wide functions. technology and will continue into the future. Synchrophasors
Examples are station reconfiguration, distribution automation, recently joined a long list of integrated functions. Integration
industrial plant control (including islanded operation), and of more functions brings multiplied benefits, not only by
system integrity protection schemes. sharing the relay hardware and firmware to perform more
tasks, but also by eliminating duplication in the construction,
IV. THE FUTURE engineering, and maintenance areas.
A. Present Trends B. New Challenges and Opportunities From the Relay Design
The trends outlined earlier will continue—more powerful Perspective
relay platforms will emerge with more processing power, The trends of improving the performance of P&C functions
running more sophisticated algorithms and delivering better while approaching limits of what is physically possible,
protection performance (speed, sensitivity, security, and integrating more functions into a single device, and improving
dependability). Designers will more often go back to first reliability of the devices and schemes will continue.
principles in power engineering to extract more detailed There are new opportunities in the area of P&C system
characteristics of the protected and controlled apparatus to engineering and construction. Replacing labor-intensive
devise better P&C methods. This will bring improvements, copper wiring and switching to prefabricated components with
but with diminishing returns as we approach the physical more standardization in the physical domain, while moving
limits of what is possible when utilizing current and voltage variability into system configuration, have been recognized as
measurements. key opportunities to reduce cost, speed up retrofit schedules,
Protection functions geared toward speed will be more and deal with the shortage of skilled workers.
often implemented with parallel and complementary This architectural change involves communication for
algorithms. Two or more algorithms may operate in parallel, protection applications. Successful deployment of these
each responding rapidly under certain, but not all, conditions. systems will need to solve a number of practical consid-
Quite often, these algorithms will be engaged only for a erations, such as isolation for testing and rework, test
limited period of time to boost the speed of operation when it methodologies, and maintenance methodologies, in addition to
is safe to do that. As a result, a slower but dependable solving performance challenges for the high-bandwidth,
algorithm, typically based on the fundamental frequency critical networks for protection applications.
components, needs to run in the background to ensure Another opportunity is to address the maintenance
dependability and enforce the expected operating challenge by designing much stronger self-tests with some
characteristic of the function. degree of internal redundancy in order to virtually guarantee a
Protection functions geared toward sensitivity will employ fail-safe response of the P&C devices with a near 100 percent
a fair amount of adaptivity. For example, capacitor bank detection rate of internal failures. This would facilitate a run-
unbalance protection may self-calibrate as the final stage of to-fail maintenance strategy.
commissioning in order to null out all standing errors resulting
C. New Protection Principles
from the natural bank unbalance and instrumentation errors.
Disturbance detectors may monitor the level of the natural When utilizing information contained in the fundamental
variation in their operating signals and adjust their pickup frequency components of voltages and currents, generated by
thresholds in proportion to this standing background noise. the power system itself, and acquired via traditional VTs and
Existing power system challenges will not disappear. For CTs, we asymptotically approach what is possible in terms of
example, series compensation of transmission lines becomes speed and sensitivity of protection. Nontraditional VTs and
applied with compensation levels exceeding 100 percent (as a CTs did not seem to deliver on their promise of reducing cost
result of splitting existing lines to connect new generation); or and size, and besides, it is still problematic whether the
more phase-shifting transformers and power electronic-based higher-fidelity measurements promised by the nonconven-
devices will be installed to better control power flows in the tional instrument transformers can truly benefit protection
system. All of this amplifies the associated protection functions.
challenges. We cannot rule out the invention of novel operating
The trend of allowing relaxed CT ratings and compensating principles that would change the field of protection. One such
for the resulting errors with protection algorithms will also opportunity would be a widespread deployment of next
continue. Today, many differential functions for busbar, generation traveling wave devices. While the principle is
transformer, and line protection are stable on external faults known and applied in fault-locating devices, it has been
with as short as 2 to 3 milliseconds of saturation-free CT mostly dismissed in protection applications. However, early
operation. In metal-clad switchgear, poor CT dimensioning implementations used very old technology compared to what
creates dependability problems for overcurrent protection. is available today. Using modern analog-to-digital (A/D)
This too is being overcome by designing magnitude converters, precise timing, and abundant processing power
algorithms that work reasonably well under extreme saturation may lead to a rebirth of this ultra-fast line protection principle.
of low-ratio CTs. Another avenue is to deploy more instrumentation for P&C
The trend of integrating more functions in a protective compared with what is traditionally used today. For example,
relay started in the very early days of microprocessor-based wireless sensors embedded in rotors of large motors or
 9

generators improve thermal and short-circuit protection of the New primary devices or unusual system configurations
machines. Integrating electrical, thermal, mechanical, visual, have been introduced in the past with little or no consideration
and audio signatures will improve sensitivity and dependa- for the protection aspect (series compensation, phase-shifting
bility of protection as well as—to a degree—speed and transformers, HVDC [high-voltage direct current], reconfig-
security. urable busbars). These devices have been conceived and
Yet another way would be to apply active injection to driven by power system economics. Each time, the protection
monitor the health of protected primary devices, similar to the industry found ways to ensure proper protection of the new
active injection methods for stator and rotor ground fault devices and the devices in the vicinity of their installation.
protection of generators and motors. The idea is to generate This trend will only continue.
well-controlled signals and respond to them, instead of Multisource distribution networks can be protected with
reacting to the fault-generated power system voltages and communications-assisted schemes acting in concert with the
currents. This concept may not be practical for protection of isolation and restoration (autoreclose) schemes used for
transmission lines but may be useful for protection of decades in the higher voltage networks. Again, short-haul,
machines, transformers, and capacitor banks. low-latency, and cost-efficient communications solutions and
reclosers are key enablers.
D. Challenges of the Changing Power System
Islanded operation of a microgrid poses a challenge that is
Power systems are on a trajectory to evolve considerably. no different than running a bigger interconnection. State
The trend to connect renewable energy sources brings several estimation, load flow, frequency control, load shedding,
changes: restoration, synchronization, black start, and so on are all
• System configuration, power flows, and short-circuit applicable concepts with existing solutions. What is different
levels may change very quickly, depending on the is the scale, cost expectations, ease of use, and ability to run in
state of the nondispatchable generation. a more autonomous mode, because human operators cannot be
• Short-circuit response of the new sources is different deployed at the scale typical for large systems.
compared with traditional synchronous generators. At the transmission system level, new sources, rapidly
Very often, new sources are effectively connected via changing power flow patterns, and interaction of multitudes of
converters and controlled quickly and aggressively by actively controlled schemes would change the dynamics from
devices aimed at protecting the source. As a result, the high-inertia smooth operation of today to more hectic
traditional fault signatures used to design protection patterns with shorter time constants. Specific challenges and
algorithms for decades are violated. This poses new detailed system integrity algorithms are still to be defined for
challenges to the existing protection principles and this area, but it is self-evident that the solution will be built on
applications. For example, fault identification logic synchrophasors and high-speed, wide-area communication.
may be challenged when a power electronic source
actively controls the negative-sequence current, while V. SUMMARY
the zero-sequence current is driven passively by the
When looking at the past, present state, and future
grounding points in the network.
challenges and opportunities for P&C, we offer the following
• Distribution networks evolve toward multisource observations:
configurations through deployment of distributed
• Microprocessor-based relays are not limited by their
generation, defeating the classical time-graded
construction when it comes to the operating
overcurrent protection approach.
principles—we have enough processing power to run
• Islanded operation of networks with distributed sophisticated algorithms. Laws of physics and the
generation will most likely be allowed in the future, imagination of the designers are the only limiting
creating more challenges to control and automation factors.
schemes.
• Going back to first principles allows enhancing the
The above changes appear immense and will most likely
classical protection criteria and schemes. Running
consume considerable resources and investment, but from the
parallel algorithms brings an increase in speed and
technical perspective, they should not be seen as over-
sensitivity, but we approach an asymptotic limit of
whelming.
what is possible when responding to classical
Tightly coupled networks with large variability in short-
protection signals.
circuit levels, weak terminals, or aggressively controlled
• In addition to enhancing individual functions, we need
short-circuit sources can be reliably protected using
to look at complete applications and target
differential protection or simpler communications-assisted
performance of the entire scheme. For example,
schemes. Recent advancements in communications technol-
reducing the reset time of a breaker failure overcurrent
ogies for the power industry (Ethernet, spread-spectrum radio,
detector is much easier than speeding up a distance
WiMAX) facilitate wider applications of these premium
function by the equivalent margin—when considering
protection principles. Reliable, cost-efficient, and easy-to-use
a critical fault-clearing time, both have the same
communications technologies are key enablers.
impact.
 10

• Communications and time-synchronized Normann Fischer received a Higher Diploma in Technology, with honors,
from Witwatersrand Technikon, Johannesburg in 1988, a BSEE, with honors,
measurements open new opportunities for wide-area from the University of Cape Town in 1993, and an MSEE from the University
schemes in protection and automation. These of Idaho in 2005. He joined Eskom as a protection technician in 1984 and was
technologies, although not new, are being deployed at a senior design engineer in the protection design department at Eskom for
three years. He then joined IST Energy as a senior design engineer in 1996. In
a massive scale today, owing to a wealth of new, cost- 1999, he joined Schweitzer Engineering Laboratories, Inc. as a power
efficient products. It is fair to observe that the engineer in the research and development division. Normann was a registered
technical capacity (measure, communicate, calculate) professional engineer in South Africa and a member of the South Africa
Institute of Electrical Engineers. He is currently a member of IEEE and
of the existing devices exceeds their demonstrated
ASEE.
applications. The industry can do much more with the
new products.
• There are still some untapped opportunities to make
relays more fail-safe and use self-monitoring and
internal redundancy to eliminate all failure-related
undesired operations. This would enable a shift toward
a run-to-fail, optimized maintenance strategy.
• There are opportunities in the way P&C systems are
put together (replace low-density copper signaling
with all communications-based solutions), but
substantial challenges need to be overcome first
(maintenance, testing, determinism of the network,
upgrades).
• More functions yield more complexity. The industry
addresses this challenge with better configuration and
test tools.
• The anticipated evolution of the power system toward
distributed generation, intermittent (nondispatchable)
generation, narrower margins, new power devices, and
more autonomous distributed control creates new
challenges for P&C. However, solutions to many of
these challenges exist today as developed for high-
voltage applications. Migrating the “high-end”
solutions to lower voltage networks will required extra
investment, primarily in communication, but we do
have a solid starting point as far as the technology is
considered. Cost-efficiency and simplicity will play a
role when applying the high-end solution to
distribution and subtransmission levels.

VI. BIOGRAPHIES
Bogdan Kasztenny is a principal systems engineer in the research and
development division of Schweitzer Engineering Laboratories, Inc. He has 20
years of experience in protection and control, including his ten-year academic
career at Wroclaw University of Technology, Poland, Southern Illinois
University, and Texas A&M University. He also has ten years of industrial
experience with General Electric, where he developed, promoted, and
supported many protection and control products.
Bogdan is an IEEE Fellow, Senior Fulbright Fellow, Canadian member of
CIGRE Study Committee B5, and an Adjunct Professor at the University of
Western Ontario. He has authored about 200 technical papers and holds 16
patents. He is active in the Power System Relaying Committee of the IEEE
and is a registered professional engineer in the province of Ontario.

© 2010 by Schweitzer Engineering Laboratories, Inc.

All rights reserved.
20100811 • TP6440-01