Selectivity Analysis in

Low Voltage Power Distribution Systems

with Fuses & Circuit Breakers
Copyright Material IEEE
Paper No. 2008-PSPC-224

Marcelo Valdes, PE, Cindy Cline, Steve Hansen, and Tom Papallo

Mersen (formerly Ferraz Shawmut)

374 Merrimac Street
Newburyport, MA 01950-1998
USA

© 2009 IEEE. Reprinted from the record of the I & CPS Technical Conference in Calgary in May
2009.

This material is posted here with permission of the IEEE. Such permission of the IEEE does not in
any way imply IEEE endorsement of any of Ferraz Shawmut’s products or services. Personal use of
this material is permitted. However, permission to reprint/republish this material for advertising or
promotional purposes, or for creating new collective works for resale or redistribution to servers or
lists or to reuse any coprighted component of this work in other works must be obtained from the
IEEE by writing to [email protected].

By choosing to view this document, you agree to all provisions of the copyright laws protecting it.
Page 1 of 10 Preprint for Peer Review

SELECTIVITY ANALYSIS IN LOW VOLTAGE POWER DISTRIBUTION SYSTEMS WITH

FUSES AND CIRCUIT BREAKERS
Industrial and Commercial Power Systems Paper #2008-PSPC-224

Marcelo Valdes, PE Cindy Cline Steve Hansen Tom Papallo

GE Consumer & Industrial Ferraz Shawmut Ferraz Shawmut GE Consumer & Industrial
IEEE Senior Member IEEE member IEEE Member IEEE Member
[email protected] [email protected] [email protected] [email protected]

Abstract – Interest in complete overcurrent device selectivity selectivity. The more common solutions for complete selectivity
has increased due to the addition of selectivity requirements to are fully fused systems where circuit sizes and fuse types are
articles 700, 701, and 708 of the National Electrical Code selected to maximize selectivity, and the use of low-voltage
(NFPA70). Many users, both commercial and industrial, use power circuit breakers without instantaneous protection in low-
fuses and circuit breakers simultaneously. Traditional Time- voltage switchgear to achieve better selectivity at the main
Current Curve (TCC) analysis is known to not fully equipment level. Neither of these solutions may yield the most
communicate fuse selectivity; hence fuse manufacturers size- or cost-efficient initial-cost solution nor the best solution
publish device ratio guidelines for selection of fuse type and from a safety and maintenance perspective, but they may be the
sizes. Recent publications of selectivity tables by circuit only perceived solution for a user designing for maximum
manufacturers also demonstrate that traditional TCCs are often selectivity. Furthermore, over the years many existing facilities
insufficient to fully communicate circuit breaker selectivity. have accumulated a variety of device types. Increasing interest
Traditional TCCs can lead to incorrect conclusions regarding in arc-flash protection may drive facility engineers to closely
circuit breaker fuse selectivity, indicating more or less scrutinize the protection and selectivity achieved by their existing
selectivity than may be possible. The authors will describe distribution systems in order to achieve maximum possible
various methods for assessment of selectivity in systems using protection at the least sacrifice to system reliability.
both fuses and circuit breakers together, with either device on
the line side. The methods will demonstrate that selectivity B. Conventional Selectivity Assessment Using Time-Current
above what TCCs demonstrate may be possible if devices are Curves and Fuse Ratios
selected correctly, and that traditional TCC analysis, can also
incorrectly demonstrate more selectivity than a more thorough Traditional assessment of selectivity is based on the use of
analysis would predict. The methods lend themselves to time-current curve (TCC) overlays. These have proven to be a
analysis that a power system engineer can perform with useful tool to evaluate selectivity over the long-time and short-
published information or information that may be requested time operating ranges of the various types of overcurrent
from manufacturers. devices. For circuit breakers, the curves are also used to
Index Terms – Selectivity, Low-Voltage Fuses, Circuit document the operation of overcurrent devices in the
Breakers, Current Limiting. instantaneous range. However, when overcurrent devices
operate faster than about one cycle, the TCC is a limited tool for
I. INTRODUCTION accurately predicting device behavior. In systems where at least
one device can operate in less than one cycle or the devices
A. Impetus for Total Selectivity in Mixed Circuit Breaker and interact with each other, the RMS-drawn TCC may not be an
Fuse Systems accurate representation of device performance. When there is
device interaction, a time-current curve that predicts how one
In 2005 the NFPA added the following requirement to article device operates in isolation may no longer describe how the
700.27, Emergency Systems, of NFPA 70-National Electrical device operates as part of a system. This is one reason why
Code (NEC) [1]: “Emergency system(s) overcurrent devices TCCs are not usually drawn below 0.01 seconds and why
shall be selectively coordinated with all supply-side overcurrent coordination studies often reflect the peak or RMS equivalent of
protective devices.” The same requirement was added in the the fully asymmetric peak current on the time-current curve.
2005 edition to article 701.18[2], Legally Mandated Standby Molded-case circuit breakers [5] usually are drawn showing
Systems and in 2008 to the new article 708.54[3], Coordination. instantaneous clearing times of 1.5 cycles or less. Over much of
Though local and state jurisdictions are interpreting these the instantaneous range these devices may be significantly
requirements differently, many of the interpretations require that faster than 1.5 cycles and may exhibit current-limiting behavior
substantial portions of the power distribution system provide even if not marked as UL 489 current-limiting circuit breakers.
complete selectivity up to calculated bolted fault values for both Furthermore, though the TCC may be labeled in RMS amperes,
utility and emergency generations sources. Similar requirements the circuit breaker’s instantaneous trip system may be sensitive
have existed previously in NEC article 620.62[4]. to peak amperes. That implies that faults of equal RMS value
Traditional power distribution system design often ignored but different power factors or closing angles [6] will be sensed by
selective performance to such high levels of fault current due to the circuit breaker trip system differently. Fuses are energy-
considerations of safety, equipment, or conductor protection and based devices and hence they may also be affected by fault
the real or perceived difficulty in achieving such high levels of current asymmetry.
selective behavior. However, the stricter interpretations of the It is important to understand whether time-coordination
new NEC requirements do not allow for these considerations. studies are designed to determine selectivity or nonselectivity.
Also, in many industrial and critical commercial systems it is not Protective devices should reliably be as or more selective than
unusual for designers to desire and design for higher levels of indicated by analysis. However, a determination of lack of

978-1-4244-3399-5/08/$25.00 © IEEE 2009

 Preprint for Peer Review Page 2 of 10

selectivity need not be as reliable. In other words, devices that minimum level, not an average or maximum level. If published
seem not to be selective by analysis may be selective under data are to be used for selectivity analysis, it is important to
some circumstances, but devices determined to be selective by know if the data are minimum, maximum, or average for the
the same analysis should be reliably selective under all parameter considered. In time-current curves, tolerance in time
reasonably expected conditions. and current are demonstrated by the band’s width. When the
time-current curve is shown as a line, it must be labeled as
C. Fuse Operation either a maximum or minimum characteristic. Other data, such
as let-through tables or melt-energy tables, may not clearly
The UL 248-1 definition of current-limiting fuse is, “A fuse identify whether the data are average, minimum, or maximum.
that, within a specified overcurrent range, limits the clearing time All analyses must take tolerance into account.
at rated voltage to an interval equal to or less than the first major
or symmetrical current loop duration; and limits the peak current D. Fuse Peak Let-Through Current (Ip)
to a value less than the available peak current.” UL 248, Low
Voltage Fuses, defines fuse performance by class including the The di/dt at the initiation of the fault is the primary external
maximum allowable peak let-through current (Ip), maximum factor determining the peak let-through current, Ip, passed by the
2
allowable clearing I t, and the maximum allowable threshold fuse. Higher di/dt will result in higher peak current let-through.
ratio. Threshold current is defined by UL 248 as “The lowest The maximum possible fuse Ip occurs at the maximum
prospective RMS symmetrical current above which a fuse is prospective fault current. Fuse Ip graphs are readily available
current limiting.” UL 248 defines threshold ratio as “The from fuse manufacturers. Figure 2 represents the Ip of a 100 A
threshold current divided by the fuse current rating. “ Class J fuse (AJT100). The uppermost diagonal line, labeled as
Fuses are thermal energy–sensitive devices. If the fuse 2.3x RMS, represents the peak available current, assuming a
element reaches its design melting temperature, it will melt. For power factor of 15%. The red line represents the maximum Ip of
each fuse design, there is a minimum melting energy that is the AJT100 fuse.
determined by the element material (typically copper or silver)
2
and by the minimum element cross-section. I t is a measure of
thermal energy under fault conditions represented by equation
(1) and defined in section V. Figure 1 depicts a current-limiting
Let-Through Current, kA peak

fuse element.

Fig. 1. Fuse elements.

Fuse minimum melting energy is valid for very short events

in the range of 1 ms or less, when there is minimal heat loss to
the surrounding environment. The higher the available fault
current, the faster the fuse element will melt. At lower available
fault currents, more time is required to melt the element and Prospective Fault Current, kA RMS
hence more energy is required, due to the loss of some heat to
Fig. 2. Fuse peak let-through.
the environment surrounding the notch area.
2 2 2
Fuse clearing I t is equal to melting I t plus arcing I t. Fuse Assessing selectivity, or the lack of it, between upstream
2
arcing I t is dependent upon numerous external factors, and downstream fuses is common in the industry today. Time-
including the instantaneous voltage during the time of arcing, the current curves are compared to determine selectivity for events
instantaneous current at the initiation of the arc, and the x/r ratio lasting longer than 0.01 s. If a separation is maintained between
of the circuit. The type of fault is also a factor, as it will determine the total clearing curve of the downstream fuse and the
the number of fuses clearing the fault. For example, a single minimum melting curve for the upstream fuse, the fuses are
fuse clears a line-to-neutral fault, two fuses operating presumed to be selective.
simultaneously clear a line-to-line fault. Under the latter As noted earlier, fuses are capable of melting and clearing
conditions the two fuses share the line-to-line voltage and will in less than one-half cycle; i.e., less than 0.0083 s at 60 Hz.
2
yield a lower arcing I t than if a single fuse were clearing the 2
Fuse melting and clearing I t values must be compared to
fault at line-to-line voltage. assess selectivity for fuses operating in their current-limiting
2
Some fuse manufacturers may publish I t melting data. 2
range. The total I t of the downstream fuse must be less than the
However, that data may not be optimized for use in selectivity 2
melting I t of the upstream fuse for selectivity for events lasting
analysis. If the melting data are used to determine whether a less than 0.01 s. Fuse manufacturers provide guidelines
downstream device will allow enough energy for the upstream documenting the minimum ratio in terms of fuse ampere rating
fuse to melt, it is important that the melting data furnished be a

978-1-4244-3399-5/08/$25.00 © IEEE 2009

Page 3 of 10 Preprint for Peer Review

that must be maintained between upstream and downstream breaker will react differently than if the prospective fault current
fuses to assure selectivity under all overcurrent conditions. was not being affected by the current-limiting device. The same
effect that creates the dynamic system that impairs engineered
E. Circuit Breakers series-rated systems can create a combination of devices with
desirable selective behavior, not evident from traditional time-
Two different types of physical mechanisms may cause current curve analysis.
current-limiting behavior in circuit breakers. The more limited
behavior in most molded-case circuit breakers not designed to II. CURRENT-LIMITING FUSE ABOVE NONCURRENT-
optimize current-limiting behavior derives from traditional contact LIMITING CIRCUIT BREAKER
arm construction. Magnetic repulsion forces are created at the
point where the contacts touch due to the constriction of current. When a circuit breaker does not provide current-limiting
The constriction is caused by the contact material’s inherent behavior, an upstream fuse will be subject to the full magnitude
roughness, which leads to conduction at only a few spots on the of the fault current for the time shown on the circuit breaker’s
contact’s surface. As the current flows toward those spots on time-current curve. For this combination of devices the
both contact surfaces, repulsive forces are created across the traditional time-current curve is a suitable analytical tool to
contacts. Mechanism spring forces should keep the contacts determine selectivity. If the fuse curve crosses the instantaneous
closed for currents within the circuit breaker’s operating range. foot of the circuit breaker curve, it is likely that the pair is more
At currents above the circuit breaker’s maximum instantaneous selective than the curve overlay shows, due to the conservative
pickup, the repulsive forces may start to overcome the spring manner in which most circuit breaker curves are drawn. This is
forces and the contacts may part temporarily, causing an arc particularly true for circuit breakers with high withstand levels
voltage to develop. This action is called contact popping. The above the intersection of the fuse curve and the circuit breaker’s
popping will have a current- and energy-limiting effect prior to withstand rating. Popping behavior may provide some current-
the contact’s being driven to full opening by the magnetic or limiting behavior that may help provide additional selectivity.
electronic unlatching mechanism. Popping does not normally
cause contacts to latch open and is power factor and closing III. NONCURRENT-LIMITING CIRCUIT BREAKER ABOVE
angle dependent, so the limitation caused by the popping is not CURRENT-LIMITING FUSE
normally shown on circuit breaker let-through curves or time-
current curves. Many circuit breakers employ magnetic trips or simple
A second design common in circuit breakers specifically digital electronic trips. For this kind of sensing the instantaneous
designed to be current limiting is the reverse-current loop shown trip may be described as peak sensing. Because they are peak
in Figure 3. In this design, current is routed through parallel sensing, the trips are sensitive to the peak let-through of the
contact arms so that opposing magnetic forces are formed. overcurrent device below. Peak-sensing trips are set to the
During high fault-current conditions, the magnetic repulsion nominal RMS current setting times √2. This comes from the ratio
forces quickly climb to values that force the contacts to of peak to RMS for a symmetrical sine wave. The analysis to
overcome the spring forces holding them together, so they part determine selectivity is based on a comparison of the peak let-
from each other very quickly. This is described as blowing the through current of the downstream device versus the pickup
contacts open. setting of the upstream device in peak amperes, for a given
value of available RMS fault current.
Figure 4 shows a simple system composed of a circuit
breaker above a fused switch. The minimum setting for the
upstream circuit breaker to reliably predict selective behavior at
maximum available fault current is determined from the
downstream fuse’s peak let-through characteristics at the
expected maximum fault current. Figure 5 shows the peak let-
through current for several current-limiting fuses. The uppermost
diagonal line represents the prospective peak current available
at the fuse’s 15% test power factor. The lower diagonal line, √2
times RMS, is the range of available instantaneous pickup
settings for circuit breakers. If the available bolted fault current
Fig 3. Circuit breaker reverse-current loop. (Ibf) at the fuse is 50,000 A, the 200 A class J fuse shown in
Figure 4 will let through a peak current of ~14,000 A. Dividing
Before the magnetic trip or other instantaneous trip initiates
the peak let-through current by the square root of 2 provides a
action to unlatch the circuit breaker, the repulsion forces may
value of ~10,000 A RMS. If the circuit breaker is set above
cause significant popping. The combination of forces acting
10,000 A, the pair will be reliably selective. For a 601 Class L
directly on the contact arms and the instantaneous trip
fuse with 62,000 A available, the circuit breaker’s trip setting
mechanism creates the circuit breaker’s instantaneous and
must be above 22,000 A to reliably maintain full selectivity.
current-limiting characteristics.
These derivations are shown on Figure 5 by the dashed line
Because of these various mechanisms, circuit breakers are
pairs drawn vertically up to the peak let-through curves and
sensitive to the peak current and peak energy delivered over the
down from the √2 diagonal line.
first few milliseconds of a fault. Circuit breakers also limit the
energy they allow to flow through the first few milliseconds of a
fault and up to complete interruption. This creates the possibility
that overcurrent devices above or below a current-limiting circuit

978-1-4244-3399-5/08/$25.00 © IEEE 2009

 Preprint for Peer Review Page 4 of 10

Fig. 4. Circuit breaker above a current-limiting fuse with 50 kA

prospective fault current.

Fig. 6. Circuit breaker and fuse with nonselective setting.

Fig. 5. Peak current let-through for several current-limiting

fuses.

Figures 6 and 7 show time-current curves for a 200 A class

J fuse under an 800 A circuit breaker. The pair of devices shown
is reliably selective up to 50 kA prospective fault current, as
shown in Figure 7. This level of selectivity is based on the peak
let-through analysis in Figure 5. A simple overlay, such as in
Figure 6, may lead to the conclusion that setting the circuit
breaker so that the instantaneous trip is higher than the RMS
current at which the fuse crosses the 0.01 s axis on the TCC is
enough to achieve selectivity. That ignores the fact that the Fig. 7. Circuit breaker and fuse with selective setting.
circuit breaker considers peak, not RMS, current and may
IV. CURRENT-LIMITING FUSE ABOVE CURRENT-
require very little peak current above threshold to trip. The pair
of devices shown in Figure 6 is not reliably selective. The time- LIMITING BREAKER
current curve is not sufficient to determine selectivity for this pair
A. Traditional Time-Current Curve Analysis
of devices. An understanding of the interaction between the
sensing of the upstream circuit breaker and the downstream The ability of current-limiting circuit breakers and fuses to
fuse’s current-limiting behavior is required. Testing comparing reduce thermal and mechanical stress as well as incident energy
current-limiting and noncurrent-limiting devices has during an arc-flash event is well known. However, what is not
demonstrated that the peak let-through analytic technique is well known is the selectivity improvement that the current- and
valid for determining selectivity between a current-limiting energy-limiting performance enables. Efforts to express this
branch and a noncurrent-limiting peak sensing main. have used selectivity tables for circuit breakers and fuse-ratio

978-1-4244-3399-5/08/$25.00 © IEEE 2009

Page 5 of 10 Preprint for Peer Review

guidelines for current-limiting fuses. However, there is little current. This also shows that the clearing time is also reduced to
information in the industry that indicates what selectivity is far less that the 0.025 s of the static TCC.
possible between these two types of current-limiting devices,
other than what may be shown by traditional time-current curve
analysis. This section presents the reasons why this selectivity is
possible and a technique to evaluate it for circuit breaker above
fuse combinations.
Figure 8 shows an upstream fuse and downstream current-
limiting circuit breaker. Overlaying the time-current curve of the
current-limiting circuit breaker and the melting time of the fuse is
the traditional way to analyze these devices. Figure 8 also
shows the time-current curve overlay for a 1600 A class L fuse
and a 250 A current-limiting circuit breaker.

2
Fig. 9. Let-through peak and I t energy waverform.

The thermal energy of this waveform, the area under the

2
curve, is measured by the I t and is calculated as
t
Energy = I t = ∫ i 2 dt
2
(1)
0

where I is in RMS terms and i is the instantaneous current.

As Equation 1 shows, the let-through energy is a function of
the circuit breaker’s ability to limit peak current and its ability to
limit the length of time the current flows. The energy limitation is
a more significant contribution because it is a second-order
term. The actual waveforms for three-phase devices interrupting
a three-phase fault are more varied and complex. However, they
Fig. 8. Fuse above a current-limiting circuit breaker. will limit the peak current to values equal to or below that shown
on published let-through curves. The actual interruption time
However, this type of evaluation treats the devices as static may vary significantly and may be slightly longer than one-half of
and independent; there are three dynamic characteristics of the the power cycle. For any one phase, if the current lasts longer
2
combination that are not considered: the peak will be smaller, if the I t term does not exceed the
2
1) It is a series circuit, so any current- and energy-limiting maximum I t defined by the device’s published curve. The
by either device will affect both. current is significantly reduced and, hence, the let-through
2) The device with the lowest current-limiting threshold energy remains low regardless of the interrupting time.
and the fastest response will affect the current
magnitude available to operate the less sensitive and C. Accounting for the Current Let-Through of the Downstream
slower device. The assumption is that the more Circuit Breaker
sensitive and faster device is the downstream device.
3) The faster current-limiting device limits the let-through Because the opening responses of fuses and circuit
energy in addition to the let-through current. Because breakers respond to different system parameters, they are
fuses require thermal energy to melt, the limitation difficult to analyze comparatively. Circuit breaker response is
caused by the downstream device has a major effect primarily a function of current, while fuse response is primarily a
on the response of the fuse. function of thermal energy. Evaluating the device combination
requires a technique that includes both variables and their
B. Current and Energy Limitations interaction across the spectrum of prospective fault currents.
The middle line and table in Figure 10 are the peak current
Figure 9 shows the prospective current and the actual let- let-through curve and values for a 250 A current-limiting circuit
through current of the circuit breaker during a fault. As this figure breaker as a function of the system’s prospective fault current.
shows, the let-through current and the clearing time are Because the upstream fuse responds to the reduced current
dramatically reduced from the fault’s prospective current. allowed to flow by the downstream circuit breaker, it effectively
Because the two devices are in series, it is this let-through operates on a smaller prospective fault current than the system’s
current that is seen by the fuse, not the full available bolted fault prospective fault current above the circuit beaker. This is
analogous to the way a larger and slower fuse responds to

978-1-4244-3399-5/08/$25.00 © IEEE 2009

 Preprint for Peer Review Page 6 of 10

achieve selectivity above a smaller and faster fuse. This reduced in the series circuit, the series circuit available current, Isf, has to
let-through current becomes the prospective fault current for the be 90 kA.
upstream fuse, shown by the lower darker line in Figure 10. After
converting the let-through current to a RMS value by dividing by
the square root of 2, we may refer to it as the effective RMS
current available to the fuse, Ie. A third designation, Isf, is used in
the analysis of the devices in series. Isf is the prospective fault
current that is required to generate the effective RMS of the
series combination.

2
Fig. 11. Half-cycle available I t based on prospective fault
current, Ibf, and effective fault current.
2
D. Fuse Response to I t
2
Fuses respond to the I t thermal energy flowing through the
2
Fig. 10. Peak and effective let-through current. fuse element. When the I t thermal energy is sufficient to melt
the current-carrying element, the fuse starts to interrupt the fault
The data for Isf are generated, with Isf as the dependent current. The energy required to accomplish this is called the pre-
variable and Ie as the independent variable. This may also be arc energy or melting energy. Figure 12 shows the minimum
2 2
considered a reverse let-through curve, with both terms melting I t of the fuse as a function of Ibf and Isf. The I t melt
expressed in RMS current, where the RMS prospective is a values are unchanged but are shifted to the effective RMS let-
function of the peak-current let-through by the smaller current- through current Isf. This shifts the melt curve from the bolted fault
limiting device. The data may be curve-fitted to create the current to the series fault current.
equation for Isf = f(Ie). Equation 2 is the fit of the data. The fuse will melt at a specific level of energy based on
prospective fault current. By shifting the prospective fault current
I sf = .746 − 0.149 I e + 0.0967 I e2 + 3 * 10−16 e I e from Ibf to Isf, the “apparent” energy required increases. The
(2) fuse’s melt-energy characteristic, inclusive of the current-limiting
Equation 2 is the system’s available fault current shifted by effect of the downstream circuit breaker, is represented by the
the peak let-through characteristics of the smaller downstream fuse-melting energy as a function of Isf. The graph demonstrates
current-limiting circuit breaker. Equation 2 is used to calculate that in a system able to deliver a 50 kA bolted fault current, the
2 2
the larger system fault current needed to produce the RMS fuse alone will melt at an I t of 2.2 million A s. But in the series
prospective current that determines the upstream fuse’s combination, which is arrived at by the Isf transform of the
2 2
performance. This equation can be used to shift the current axis current, the fuse apparent I t melt energy is 2.8 million A s. This
2
in an I t melting curve of the upstream fuse to properly is because for 50 kA available fault current, the downstream
demonstrate the current the upstream fuse will see. It shifts the circuit breaker will only let through the equivalent of a 23 kA
current of a circuit’s characteristic from the Ibf current to the Isf fault.
current needed to create the same let-through. Figure 11 shows
2
the half-cycle I t as a function of Ibf and Isf. For example, for the
2 2
half-cycle I t to reach a value of 5 million A s, the bolted fault
2
current has to be 14 kA. But for the I t to reach the same value

978-1-4244-3399-5/08/$25.00 © IEEE 2009

Page 7 of 10 Preprint for Peer Review

1600 A Class L Fuse Melt Energy as used is the current- and energy-limiting characteristics of the
function of Ibf and shifted to Isf circuit breaker and the pre-arc melt energy of the fuse. Fuse
manufacturers do not commonly publish the pre-arc melt-energy
curves for their fuses, but it may be available upon request.
I2t Melt as f(Isf)
Fuse Melt Energy in Million A2s

I2t Melt as f(Ibf) 1600 A Class L Fuse Melt Energy as

function of Ibf and shifted to Isf
Shifted by downstream CB’s
Current Limitation (Isf)
I2t Melt as f(Isf)
I2t Melt as f(Ibf)

Fuse Melt Energy in Million A2s

Unshifted CB’s I2t let-through

Available fault current, kA CB’s I2t let-through

2
Fig. 12. A fuse’s required melt I t as a function of Ibf and Isf.

This analysis is very conservative because the test power

factor of the fuse is ignored. Fuses are tested at 15% power
factor hence the peak prospective current is 2.31 times the RMS
value of prospective current. Dividing the downstream circuit Available fault current, kA
breaker’s peak let-through current by 2.31 instead of 1.41 yields
a bigger Isf shift. This may be more in line with the device’s true Fig. 13. A fuse’s required melt energy as a function of Ibf and
performance, however it would be a less conservative Ipf and the downstream circuit breaker’s let-through energy.
conclusion.
V. CURRENT-LIMITING CIRCUIT BREAKER ABOVE
E. Circuit Breaker Let-Through Energy and Selectivity CURRENT-LIMITING FUSE
Determination:
Advanced current-limiting circuit breakers may have three
Because the fuse responds to energy, the circuit breaker regions to their instantaneous trip. The leftmost region may be
let-through energy is used to evaluate the selectivity, not the composed of an adjustable electronic trip with an advanced
circuit breaker’s clearing time. Figure 13 shows the let-through algorithm able to filter narrow-peak let-through currents. The
energy of the current-limiting circuit breaker superimposed on rightmost region is where the circuit breaker contact assembly
the fuse’s shifted and unshifted melt energy. Selectivity may be has enough energy from the fault current to quickly blow the
determined by comparing the circuit breaker’s let-through energy contacts open and keep them open while the trip catches up and
with the shifted melt energy required by the fuse. latches the mechanism in the open position. The middle
2
The I t let-through of the circuit breaker is not shifted transition region is where the circuit breaker contacts may pop or
because the let-through energy is a function of Ibf as perceived start to open due to magnetic forces, but the circuit breaker still
by the faster downstream limiting device, in this case the relies on an electronic trip, magnetic trip, or other mechanical
current-limiting circuit breaker. Figure 13 demonstrates this trip to fully open and unlatch the circuit breaker. Figure 14
analysis for a specific combination of a 250 A current-limiting demonstrates the three regions in a 600 A circuit breaker with an
circuit breaker and a 1600 A current-limiting fuse. The analysis adjustable advanced-algorithm electronic trip. The device shown
demonstrates that these devices should be selective for more uses an algorithm designed to filter narrow-peak let-through
than 90 kA prospective fault current. The same two devices currents and hence may be set below the peak let-through of a
demonstrated a potential selectivity of 15 kA based on traditional downstream current-limiting device. The filtering algorithm
curve overlay, as shown in Figure 6. Based on a simple section is identified by the gap below the curve. The flat-topped
comparison of let-through energy and melt energy, selectivity up instantaneous portion between the adjustable section and the
to 65 kA may be expected, but when the effect of the effective beginning of the sloped portion includes a region where the
RMS shift is taken into account, predicted selectivity is over circuit breaker may trip because of the electronics or the
90 kA. mechanical trip mechanism. Which mechanism causes the
Combining the shifted fuse melt curve with the circuit circuit breaker to open depends on the closing angle, voltage,
breaker let-through curve shows the energy-based selectivity of fault-current X/R ratio, and let-through characteristics of a
the combination, including the current-limiting effect of the downstream device that may be limiting fault current. The sloped
downstream circuit breaker and the upstream fuse’s response to portion to the right is the truly current-limiting portion of the
the limited prospective current it has available. This method curve. The clearing time is not material, as the circuit breaker
provides a more accurate prediction of the selective behavior may allow minimal current to flow for a few milliseconds, but the
between a larger upstream current-limiting fuse and a smaller peak current and energy are limited regardless.
downstream current-limiting circuit breaker. The information

978-1-4244-3399-5/08/$25.00 © IEEE 2009

 Preprint for Peer Review Page 8 of 10

required analytical tool. Note that the time the fuse takes to open
is not part of the analysis. It is the energy the fuse lets through in
the process that matters.

Electronic Filtering Algorithm Region

Transition Region

Fully Current-
Limiting Region

Fig. 14. TCC for a current-limiting molded-case circuit breaker

showing separate regions for current limiting and filtering
electronic tripping.
2 2
Fig. 15. Fuse let-through I t and the circuit breaker’s I t
The trip system in the circuit breaker responds to both peak
requirement.
current and energy. Once the peak current is over the threshold,
there has to be enough energy to move the trip mechanically to Before the circuit breaker’s energy-based current-limiting
unlatch the breaker mechanism. In some breaker designs, the region can be considered, the circuit breaker’s instantaneous trip
mechanical system is intentionally damped to reduce the must be set above where the fuse is reliably current and energy
sensitivity of the trip. This creates a portion of the fault current limiting for a three-phase event. The three horizontal lines
range for which the circuit breaker will not commit to a trip for a 2
represent the let-through I t for three different sizes of class J
limited amount of time. With an understanding of how the trip fuses in a 480 V system. All three let less energy through than
operates, the circuit breaker can be analyzed as an energy- the circuit breaker’s mechanical system needs to commit. For
driven device over the range of fault currents. The shape of the high-level faults, all three fuses are probably selective with the
curve drawn in Figure 14 is intended to alert the user that the circuit breaker’s mechanical system. However, in this case the
circuit breaker behaves this way, but does not provide sufficient potential overlap in the curves is at lower current levels. The
information for a complete selectivity analysis to be made. three-phase behavior of the fuses in this region is typically not
However the manufacturer will have sufficient information to fully defined. Traditionally, the data of interest were for the
perform selectivity analysis and generate selectivity tables for highest available fault. Only the 200 A fuse, which is energy
specific pairs of current-limiting devices where the downstream limiting at ~5000 A, meets the criterion of being reliably current
device may be a fuse or a circuit breaker, regardless of how the limiting under the 6000 A threshold of the electronic trip. This
curve is drawn. energy-limiting threshold varies based on system voltage and is
The analytical technique for determining selectivity for higher at 600 V. Fuse manufacturers commonly publish peak
current-limiting circuit breakers above fuses is similar to that for current let-through curves for their circuit breakers. However,
fuses above circuit breakers. The analysis must be divided into this analysis requires the fuse manufacturer to provide the
two regions. In the leftmost region, the electronic trip filters the energy let-through value for the fuse at the application voltage
single peak allowed to flow by a current-limiting downstream and over a range of fault currents. This is typically constant
device. The rightmost region represents the mechanical portion energy after some prospective current level. Fuse manufacturers
of the trip that may be analyzed as a pure energy device. may be able to provide these data upon request.
Figure 15 shows the commit energy representation for a
molded-case circuit breaker with a waveform-recognition VI. TEST RESULTS
electronic trip and a mechanical trip. The flat section to the left is
equivalent to two half-cycle sine waves at the threshold peak. A. Upstream Fuse, Downstream Current-Limiting Circuit
This can be used to represent the peak filter algorithm in energy Breaker
terms. The rising slope is a representation of the circuit breaker
mechanical trip’s required commit energy. This is a simplification Tests were designed to confirm that a current-limiting fuse
of the actual required energy, but it is sufficient to provide the provides selective protection above a current-limiting circuit

978-1-4244-3399-5/08/$25.00 © IEEE 2009

Page 9 of 10 Preprint for Peer Review

breaker, as shown in Figure 13. The example analysis shown incorrect. Second, nontraditional measures, such as peak
2
previously indicated that a 250 A current-limiting molded-case currents and I t, can reliably be used to perform selectivity
circuit breaker should be selective with a 1600 A class L fuse. studies rather than solely using time. Third and most important,
The overlay of characteristics shown in Figure 13 demonstrates the limiting and trip-commit behavior of devices developed
that potential lack of selective performance occurs at high fault individually can be analytically combined for series
currents in the range of 90 to 100 kA. Without shifting the fuse combinations. This enables analysis of the series performance
melt-energy curve by the transform of the circuit breaker’s let- of the near-infinite combinations of upstream and downstream
through, peak current selectivity may be limited to 65 kA. With devices. The demonstrated methods can provide the industry
the curve shift, selectivity should be at least 95 kA. Three-phase with provable techniques to improve analysis of system reliability
short-circuit tests where performed at 100 kA with a 20% power and protection using devices and information available to the
factor. Ten tests were done at various closing angles (closing industry today.
angle is a measurement of the angular difference between when
the fault is initiated and the voltage on phase A of the test Main Advanced
circuit). In all ten cases the circuit breaker interrupted with no Current Conventional Current
apparent damage to the fuse. Impedance tests on the fuses pre Limiting Fuse Circuit Breaker Limiting
and post testing indicated no changes in fuse resistance. Feeder Circuit Breaker

B. Upstream Circuit Breaker, Downstream Fuse Fuse I Peak Let-

2
through vs CB Fuse Peak I t
A second set of tests was performed with a 600 A current- Current Published
Peak Let-through vs
limiting circuit breaker on the line side of 200, 300, and 400 A Limiting fuse Tables
Instantaneous CB I2t commit
class J time-delay, current-limiting fuses. This combination is Pickup
shown in Figure 15. These particular combinations of devices
show potential lack of selectivity at relatively low fault currents,
Conventional
where the fuse’s let-through current must be filtered by the Time Current Time Current Time Current
Circuit
circuit breaker’s electronic instantaneous trip. Sufficient data Curves Curves Curves
Breaker
were not available from the fuse manufacturer to model fuse
performance in the area around the fuse’s current-limiting
threshold at 480 V. Hence, testing in this range of fault current CB I Peak Let-
Current
was required. A total of 14 different tests were performed with CB I Peak through vs CB
Limiting 2 Published
the 200 A fuse at fault currents from 5 kA to 100 kA, at 20%– Shifted I t Let- Peak
Circuit Tables
50% power factor, and at various closing angles. In all cases, through Instantaneous
Breaker
two or more fuses cleared properly and the circuit breaker did Pickup
not trip. Some limited additional testing was performed with 300
and 400 A fuses at various low and high fault values. In all cases Fig. 16 Table of suggested assessment method versus
the fuses cleared properly and the circuit breaker did not trip. line and load side device type
Though insufficient tests were performed with the larger fuses to
make a definite determination, it is anticipated that the circuit Manufacturers may have access to the detailed information
breaker would be selective with fuses as large as 400 A. and may be able to provide it to the interested user, or may be
able to perform the analysis directly for the user interested in
VII.CONCLUSIONS specific combinations. In either case, there are methods to move
the industry past traditional analytical techniques that are not
The various techniques described provides methods for reliable in every case. And new techniques can be used to
analyzing the selective capability of fuses and circuit breakers in provide more reliable analyses, resulting in better protected, and
systems in which either may be used above the other. more reliable power distribution systems.
Traditional time-current curve analysis is not sufficient for some
of the combinations of devices, but other analyses based on an VIII. BIBLIOGRAPHY
understanding of the let-through characteristics of the
downstream device and the commit behavior of the upstream [1] Marcelo Valdes, Tim Richter, Mike Tobin & John Hill,
device, regardless whether either is a fuse or circuit breaker, “Enhanced Selectivity and protection Via Modern Current-
allows insight into how the system of devices will operate. Limiting Circuit Breakers”, IEEE Industrial and Commercial
Understanding the system operation allows selection of the Power Systems Conference Record, May 2005.
optimum assessment methods for every combination. Some of [2] Draft IEC/TR 61912-2, Ed.1.0: Low-voltage switchgear and
the analyses may be performed with published information, control gear – Overcurrent protective devices – Selectivity under
while others require more detailed understanding of the overcurrent conditions, International Electrotechnical
operation of both fuses and circuit breakers. Commission, March 23, 2007.
The analysis techniques presented here and the preliminary [3] Marcelo Valdes, Andrew Crabtree & Tom Papallo, “Method
test validation of these techniques illustrate three important for Determining Selective Capability of Current-Limiting
advancements in selectivity evaluation. First, that the limiting overcurrent Devices Using Peak-Let-Through Current, What
performance of downstream devices can be included analytically traditional time current curves will not tell you”, IEEE Industrial
in selectivity studies. The traditional static time evaluation and Commercial Power Systems Conference Record, May
excludes this dynamic downstream limiting characteristic 2009.
resulting in perceived selective or unselective results that are

978-1-4244-3399-5/08/$25.00 © IEEE 2009

 Preprint for Peer Review Page 10 of 10

[4] Ed Larsen, “A New Approach to Low-Voltage Circuit in mechanical engineering. He started with GE in 1986 in the
Breaker short-circuit selective coordination”, IEEE Industrial and New Product Development department. He has also worked on
Commercial Power Systems Conference Record, May 2008. circuit breaker and electrical distribution system projects for
several other major manufacturers, returning to GE in 1997. He
IX. VITAE is currently manager of the Innovation Engineering organization
for GE’s Power Distribution and control products business. He is
Cindy Cline graduated from Michigan Technological the holder of 35 US and international patents and has co-
University in 1983 with a BSEE in power systems and authored several papers for IEEE forums on selectivity, power
machinery. She has held engineering and management systems protection, and arc flash related topics.
positions with Ferraz Shawmut and formerly for LCI. Cindy is Marcelo E. Valdes graduated from Cornell University in1977
currently Manager, Applications Engineering for Ferraz with a BS in electrical engineering. He has been with GE for
Shawmut. Her responsibilities include application engineering, over 31 years, in field engineering, sales, marketing, and
finding solutions to difficult protection applications, technical application engineering. He is currently the manager of
support, technical seminars and training, product safety, and Application Engineering for GE’s Electrical Distribution Business
leading Ferraz Shawmut’s Applications Engineering team. She in Plainville, Connecticut, where he provides application
is currently a member of IEEE. Cindy has published several engineering and strategic product planning leadership. Mr.
papers and articles on overcurrent protection, ac/dc power- Valdes is past chair of the IEEE Power and Industrial
conversion systems, and fuse performance. Applications Engineering chapter in San Jose, CA, and the
Steve Hansen graduated from Iowa State University in 1973 Industrial Applications chapter in San Francisco, CA. He is a
with a BS in engineering operations. He has held field and registered Professional Electrical Engineer in California. Mr.
management positions with Ferraz Shawmut and Rockwell Valdes has authored and co-authored several technical papers
Automation. Steve is presently senior field engineer for Ferraz for IEEE and other engineering forums, and has several pending
Shawmut. His responsibilities include application engineering, patents in the field of power systems protection and circuit
technical training, product safety, major account development, breaker trip systems. Currently Mr. Valdes is a member of
and standards development. An active member of NFPA and several IEEE standard working groups, including the working
IEEE, Mr. Hansen serves on various IEEE working groups group editing the revision of the IEEE Color Books into a new
involving safety and arc flash, including IEEE P1683 and IEEE format, a new recommended practice for safe power distribution
1584. Mr. Hansen has been a member and chair of various system design. He is currently vice chair of IEEE P1683,
NEMA and UL committees in the area of fuse standards, Standard for Safe Motor Control Centers and leader of the
including NEMA 5FU, UL STP198(currently 248), and STP347. project to write the “Recommended Practice for Bus and
Mr. Hansen has published several papers in the area of Switchgear Protection” (P3004.11) derived from the present
overcurrent protection, safety, and overcurrent protective device IEEE Buff book. Mr. Valdes is also involved in various other
coordination. IEEE working groups.
Tom Papallo graduated from the University of Connecticut
in 1986 with a BS degree and in 1989 with an MS degree, both

978-1-4244-3399-5/08/$25.00 © IEEE 2009