| INTRODUCTION
A good place to start is with a problem statement. This may
best be expressed by relating experiences that you may commonly
encounter. Do you ever run into these conditions?
| SECTIONS |
 |
Upgrade of a
Service Entrance or Substation pushes the available
fault currents beyond your system limits. What are the
options to replacing all of the underrated switchgear? |
|
One could avoid a service entrance
upgrade and improve system regulation if you could simply
Close the Tie Breakers (now
open due to excessive fault current). |
|
A Cogenerator
Addition yields fault contributions beyond system
limits. Replacement of all the impacted gear or reactor
regulating voltage makes the cogenerator addition far
less attractive. |
|
The regulating voltage
across a Current-Limiting Reactor
prevents startup of a large new motor, or its startup
causes a voltage sag that disrupts other equipment on
the system. Also, what about the continual losses which
can be tens of thousands of dollars per year. |
|
Energy
Limitation is required to prevent catastrophic
tank rupture of a failed transformer or other equipment.
While current-limiting fuses are commonly applied on lower
rated units, how about the larger transformers? |
What if these typical problems did not
exist as a major barrier for system engineers? Solutions of
all are in the domain of G&W's CLiP® Current Limiting
Protector and PAF® Power Assisted Fuse. These will be
referred to hereinafter as Commutating Current Limiters (CCL),
sometimes called Triggered Current Limiters. They fill a unique
over-current protection role in the higher continuous current
range of medium voltage (1-38kV) equipment where traditional,
meltable element fuses generally do not exist. Most of this
discussion will be oriented toward the CLiP.
TYPICAL
SOLUTIONS
What are the more common options? What are the typical benefits
and drawbacks of each? It should be noted that there is a
role for each, and that any one may be the best solution for
a particular system configuration and user need.
- Replace or upgrade all of the circuit
breakers and upgrade system bracing throughout. This is
often the most expensive solution. It meets the overall
goals. If the speed of interruption is not improved, however,
it will not limit the greater fault energy level which is
related to the square of the current.
- Open the tie between multiple sources.
This may work well with double-ended substations as applied
by many plants. It is a simple and effective solution but
often undesirably affects system regulation. If the system
is not well balanced, one of the transformers may be overloaded.
If a cogenerator is added on one bus, its benefits can not
be transferred to the adjacent bus without closure of the
tie.
- Apply current limiting reactors to limit
the fault. This permits usage of the existing equipment
and will reduce the fault energy. It results in a continual
operating energy loss due to the resistance (not reactance)
of the reactor. This is often neglected and for larger reactor
banks may be tens of thousands of dollars per year. It imposes
a regulating voltage drop on the system which often prevents
startup of larger motors. It may also impose an unwanted
phase shift to the system or impede the supply of vars from
a cogenerator.
- Apply a Commutating Current Limiter.
This will limit faults within the momentary and interrupt
fault duty on the existing equipment. It will also permit
the equipment to interrupt lower level faults, within its
capabilities, without operating a Limiter. If it operates,
however, the Limiter's interrupter must be replaced. This
is the least well recognized option presented. The Commutating
Current Limiter will be explored in greater detail below.
CCL
CAPABILITIES
Lets explore the most common questions. Why do these devices
exist? Why will traditional meltable-element fuses not work?
What capabilities for system voltage, continuous current and
maximum interrupt are available?
These products were developed in the late
1970's as part of an EPRI-funded project. They cover a need
where meltable fuses reach their practical limit. For example,
at 15.5kV, the meltable element fuses are generally available
to 200A continuous (but are already in a double-barrel design)
with some manufacturers reaching up to 300A continuous (in
four-barrel designs). For these traditional fuses heat rejection
becomes a major consideration. Also, the very high let-thru
current may be in excess of the crests of many systems. In
other words, it may not limit current to a usable range if
it limits peaks at all for the corresponding available current.
The Comutating Current Limiters cover a
continuous current range from 200A
to 5000A. Available voltages are from 2.8kV
to 38kV. Higher voltages are projected for the future.
Interrupt ratings are 40kA rms, symmetrical
for all ratings with most having an optional 120kA
rms, symmetrical interrupt rating. While not common,
there are a surprising number of applications of what seem
to be unbelievable available fault currents of 65, 80, 95
and even 115kA rms, symmetrical. These
devices have literally interrupted 311kA rms, symmetrical
at 15.7kV. The formidable current-limiting capability
at extreme fault levels adds a whole realm of possibilities
to the overcurrent protection spectrum. Energy rejection is
minimal with a 15.5kV, 3000A unit rejecting approximately
140 watts per phase.
Another factor to consider with a CLiP®
type of Commutating Current Limiter is that the continuous
current is, for all practical purposes, completely independent
of the current-limiting performance of the device. These are
electronically sensed and triggered units whose operating
criteria is preset and not dependent on time versus temperature,
element size (or melting I2t) or preconditions. They are
catastrophic protection devices that allow circuit breakers
and other lower rated devices to clear faults within their
capabilities. The Commutating Current Limiters described in
this paper are intended to operate at more severe overcurrent
levels, typically beyond the limits of the existing equipment.
CCL
USERS
Users come in many forms. Utilities account for approximately
25% of the users, primarily applying the limiters as a substation
or generating station device. They have been used in the protection
scheme of nuclear plants. Government bases and research facilities
account for approximately 7%. Universities, office complexes
and medical centers account for about 8%. The mainstay of
the users, say 60%, are in the industrial sector.
CCLs are presently applied on 5 continents
and on oil well processing ships offshore. These devices may
be placed indoors or out, from -40 °C to +40 °C and
beyond. They do not require enclosures for use outdoors and
are often suspended from steelwork or crossarms. Units are
commonly enclosed in switchgear lineups and have been adapted
to busduct arrangements. They have also been adapted to circuit
breaker trucks for simple insertion/removal.
In the industrial sector the users vary
widely from pulp & paper, petroleum and chemical processors
to automotive plants, steel mills and gold mines. They provide
protection to the reportedly 1st and 2nd largest steel mills
in the world (in Korea and Brazil), the world's largest paper
pulp facility (in Brazil), newsprint mill (Canada) and highest
volume wellhead in the southern hemisphere (Indian Ocean).
They are applied to protect cogenerators on test by their
manufacturer as well as at installations in the field. Lets
look more specifically at how they are applied to resolve
those issues listed in the problem statement.
APPLICATION
AND PERFORMANCE Installation
Photo
The general philosophy of these Commutating
Current Limiters is to act as catastrophic protection for
the system. In other words, let the overloads and lower level
faults be cleared by downstream devices operating within their
capabilities. Clear the major faults with the Limiter.
Lets first look at the system limits. We
will have set the limiter so that the equipment will not be
overexposed from interrupt or long term thermal standpoints.
How do we determine whether or not we will exceed the short
term or momentary conditions?
If you originally have breakers rated for
20kA rms, symmetrical interrupt at 5 cycles (I will avoid
k-factors etc. in this work), the downstream breakers will
typically be able to handle 2.7 times this amount as a peak,
asymmetrical value on the first 1/2 cycle. Remember that the
instantaneous peak of a 20kA rms, symmetrical wave will be
Ö2 (or 1.414) times that value giving 28.28kA
inst. When fully asymmetrical or offset (due to a fault initiation
on that phase of zero degrees on the voltage wave), the crest
can theoretically be double that magnitude (2 x Ö2)
or 2.828 times that number for an infinite circuit X/R. From
a practical standpoint it is uncommon to find general substation
circuit X/R values greater than 30 which would yield a multiplier
of 2.69 for that system. For generator circuits, where it
is not uncommon for the circuit X/R to be 75 or higher, this
value approaches 2.8. [As a side note, CCLs perform very well
interrupting the very high X/R circuits since they introduce
a high resistance to the circuit during the interrupt process
which is discussed later.] This leads to the fact that the
peak asymmetrical crest that the circuit breaker is capable
of is typically in excess of 2.7 times the rms symmetrical
interrupt rating during the first peak. A maximum of 2.5 times
the rms symmetrical rating is typically applied as a limit
(which is in agreement with certain standards). The actual
value should be verified by the circuit breaker manufacturer.
One does not want to exceed this value as
the electro-mechanical forces resultant from the square of
the current may fracture insulating supports, bend bus, disengage
contacts or cause "popping" at the contact surfaces
due to the repelling forces overcoming the contact spring
force. Since most equipment on the system will have a one
second or three second rating equalling the rms symmetrical
interrupt rating of the breaker, there should be plenty of
thermal capability if we stay within its instantaneous peak
capability. In other words, the thermal limits are generally
not an issue since the fault duration will only be a fraction
of a cycle.
Let's consider the first problem, Upgrade
of the Service Entrance or Substation in Example
1. We will assume a dual transformer supply. We will also
assume that the original transformers are replaced by ones
which yield an available fault current of 15kA for a total
of 30kA rms symmetrical versus the 20kA rms symmetrical rating
of all downstream equipment.
Example
1 Circuit Diagram Example
1 Plot
The first
plot depicts the interruption of one transformer under
fully asymmetrical conditions. The operating point had been
selected as 14kA instantaneous amperes. We will not explore
the methodology of trigger level selection as that is a very
lengthy topic, but instead, will provide appropriate selections.
The point of peak let-thru (after triggering, commutation
and shunt fuse melt of the limiter) is 17.5kA instantaneous.
The available current is the full dashed outline. The shaded
area is the projected current through the Limiter. The
point of peak let-thru is well before the point of current
extinction. This characteristic is described in greater detail
in the operating sequence. A symmetrical wave would have reached
its point of limitation at a much earlier time and final extinction
would have occurred at 1/4 cycle. The fully asymmetrical fault
will be depicted throughout the examples as this is typically
the worst case for momentary duty as well as energy let-thru.
The removal of the one transformer source
will keep the circuit breakers within their interrupting range.
One can show that the instantaneous (momentary) current capabilities
of the gear will also not be exceeded. This can be accomplished
by a more traditional means of using let-thru curves and adding
to the other sources but with an accommodation for the timing
of the peaks. An improved method, by computer generation,
is depicted in the second
plot for the tie closure problem.
In Example
2 we depict the Closure of a Tie
within the system. We have also added a cogenerator as this
is increasingly common. Lets assume that a fault occurs at
the location shown. We have 16kA rms symmetrical available
thru the limiter for a total of 25 but with the equipment
limit at 20. Available currents on both sides of the tie are
shown on
the plot as well as the total to the fault.
Example
2 Circuit Diagram Example
2 Plot
The Limiter's current profile is depicted
by the shaded
area. If we add the instantaneous current of this shaded
area to the "residual current" from the faulted
side of the tie (which does not flow through the Limiter),
we can project the total instantaneous currents to the fault
with reasonable accuracy. You will note that the peak does
not occur at the time of peak fuse let-thru in this example.
Also, it is not equivalent to the peak let-thru of the Limiter
added to the crest of the "residual current" wave.
In some other cases the peak let-thru condition will occur
at the peak let-thru point of the Limiter. These conditions
are more difficult to project with traditional meltable element
fuses. With the aid of a computer these plots are readily
generated for virtually all users of the CLiP® Commutating
Current Limiter. It permits them to visualize what is occurring
without much guesswork.
Following CCL clearing, the buses are separated
and the downstream breaker can safely interrupt the "residual
fault," which is still supplied by the left source transformer.
While load shedding may be necessary as a temporary measure
to sustain the left bus, critical processes or operations
will usually not be dropped. Not shown on the Example
2 circuit diagram is a disconnect switch to the left side
of the Limiter. Opening this and the tie breaker will enable
replacement of the expended Limiter without de-energizing
either bus.
The Addition of a
Cogenerator has been depicted in Example
2 as well. In many cases the system engineer will decide
to place the Limiter in the cogenerator supply bus instead
of at the tie. This will (for
Example 2) permit the breakers and other equipment to
operate within their ratings while providing exceptional protection
to the generator itself. Should a fault in a generator winding
occur, the fault energy as supplied by the utility transformers
is greatly reduced. Let-thru I2t values (approximately proportional
to fault energy) are typically one-half % of that through
a five cycle breaker. A winding meltdown may be reduced to
little more than a puncture. Given the value of the cogenerator
unit, it is excellent insurance. From an operating perspective,
the triggering and let-thru characteristics are similar to
the tie-position application.
If we consider the reactor, this certainly
has a solid place in the protection world. A few drawbacks
were previously listed. These can, however, be mitigated by
Bypassing the Reactor with a
Commutating Current Limiter. The secret here is that this
combination maintains the benefits of the reactor without
the operational drawbacks. It is a more costly scheme than
some of the others but can avoid the operation shutdown since
the sources are only limited and not interrupted. Very simply,
the Limiter carries virtually all of the continuous current.
Therefore the continual losses and regulating voltage imposed
by the reactor are avoided.
If a fault should occur in Example
3, the Limiter operates and thereby commutates the reactor
into the fault circuit. This in-turn limits the magnitudes
within the ratings of the downstream equipment. The lessened
fault level is then cleared by that downstream equipment.
After the fault is cleared, the reactor continues to conduct
load current to the system such that the critical processes
are not shut down or otherwise compromised. Power is not lost
to the critical loads as would happen with the Limiter itself.
The bypass switches are then opened to isolate the Limiter
and change the expended interrupter. Following replacement,
the bypass switches are closed to re-bypass the reactor.
Example
3 Circuit Diagram Example
3 Plot
The Plot
3 projection is an approximation. The Limiter will commutate
the fault into the reactor well before the system's voltage
zero point due to the reactor voltage which will be far less
than system voltage. The actual time is highly dependent on
reactor characteristics, distance from the Limiter, cable/bus
runs etc. A typical range is 0.5 to 1.0 millisecond. Reactor
bypass accounts for approximately 25% of the Limiter applications.
The last problem suggested at the start
of this paper is Energy Limitation.
This was discussed in the cogeneration analysis above. Since
the Commutating Current Limiters are so effective at limiting
energy, we enter a realm where we begin to look at the possibilities
of keeping the lids on faulted transformers. The circuit breakers
may be capable of interrupting the circuit but not necessarily
in time to prevent the catastrophic tank rupture. This applies
to catastrophic failure of other equipment as well. Particularly
for the transformer, the sensing of the fault is often difficult.
The first indication may not come from current sensing relays,
but from instantaneous overpressure or change of pressure
relays. Protection becomes feasible from the standpoint of
remote triggering of the Limiter in response to these other
inputs such that the fault is cleared at an early stage. Discussion
of how this is performed at such speed follows.
CCL
OPERATION
The essence of the operating procedure will be given below.
It will not go in technical depth but rather into an overview,
seeking a generalized understanding of the process.
These devices can be characterized by a
primary conduction path which electrically parallels a special
current limiting fuse of very high energy absorption capability
and low melting I2t. Upon incident of a fault meeting the
triggering criteria, the primary current path is opened -
essentially a high-speed switching operation. This causes
commutation of the fault current into the current-limiting
fuse and its rapid interruption. The interrupt process of
this fuse is typical of the traditional current-limiting fuse
with 1/4 cycle interruption of symmetrical and 1/2 cycle interruption
of asymmetrical faults.
- The
Normal Operating Condition - The primary current path
is a busbar with precision-machined notches. This forms
the basis of the high-speed switching portion of the unit.
The current limiting fuse, being a higher impedance path,
carries a minuscule portion of this continuous current.
- Triggering
Logic Senses a Fault and Responds - As the fault is
initiated the current begins to rise. At the point where
the fault conditions meet the operating criteria, the triggering
process begins. These units use a threshold-level sensing
system, not a rate-of-rise sensing system. The current must
actually reach a preset level and be held for 80 microseconds.
The intention is to refrain from triggering in response
to transients and harmonics where the rate-of-rise of the
current may be quite high, but a fault does not exist. Also,
the maximum rate-of-rise of an asymmetrical fault will occur
half way to its crest. We will generally want to have operated
and be in the clearing mode before that time. With the threshold
level sensing system, these units are not limited to bus
and line-up protection. These CCLs are also effectively
applied in capacitor bank and harmonic filter switch protection
schemes. Upon meeting the triggering criteria the triggering
logic sends a pulse the actuator.
- The
Actuator is Initiated - This begins the interrupt process.
The actuator in-turn operates a high-speed cutting charge
which literally cuts through the notched section of busbar.
This is the basis of the high-speed switching process. Note
that while the drawings indicate a single switch, there
are actually a multiplicity of cuts in the bus. For example,
there are four cuts or switching points in the 15.5kV bus.
The switch opening time following the cutting process is
approximately 13 microseconds. The tube surrounding the
bus remains in tact, capturing the ionized gases. No blast
of these gases occurs outside of the tube.
- The
Switch Opens, Fuse is Commutated into Circuit - The
actuator opens the high-speed switch which draws an arc
and yields the corresponding arc voltage. This is multiplied
by the number of cuts. The effective commutating voltage
is accordingly higher for multiple cuts than with a single
cut. This provides for a more rapid commutation of the fault
current to the shunt fuse and improved dielectric withstand.
As commutation is completed, the current-limiting fuse carries
the full magnitude of fault current and is in the melting
process.
- The
Fuse Melts and Interrupts - While the fuse is in the
melting process the high-speed switch must recover dielectrically
so that it will withstand the arc voltage of the current-limiting
fuse. With the low arc times at the switch and therefore
a low ionization level, the dielectric recovery of the switch
is very rapid, in as little as 50 microseconds.
Upon completion of the melting process,
the arcing of the fuse begins. The arc voltage caused by
the inductively driven current through the resistance of
the fuse yields an instantaneous fuse voltage greater than
the system voltage. This is responsible for the rapid reduction
of fault current. Interruption is not instantaneous however.
Though the current falls quickly it will not be fully extinguished
until nearly at the voltage-zero point.
Note that this is not the current-zero, the point of interruption
for circuit breakers, expulsion type fuses and other interrupters.
The fuse is absorbing much energy in melting the element
and the surrounding quartz sand interrupt media. This is
a high resistance that is now being introduced into what
is typically a highly reactive fault circuit. Consider that
the current and voltage may initially be as much as 90 electrical
degrees out of phase. During the interrupt process a phase
shift occurs which brings the current and voltage back in
phase with each other. This yields an ultimate clearing
at the zero point for both. It occurs at 1/4 cycle for the
symmetrical fault and 1/2 cycle for the fully asymmetrical
fault (with varying asymmetries in between). The characteristics
of the current-limiting fuse in this paragraph are typical
of any current-limiting fuse clearing a heavy fault.
- The
fault has been cleared - The fuse quickly cools and
recovers into a non-conducting mode. The fault has been
cleared.
Conceptually, in simple terms, the device
contains a high-speed switch that carries the continuous current.
Upon sensing of a fault by the electronic triggering logic,
the switch is opened and the current is forced into a current-limiting
fuse that interrupts the circuit. In this way the Commutating
Current Limiter provides an effective means of protection
without costly equipment replacements or introducing unwanted
performance characteristics to the system.
Biography:
John S. Schaffer, an employee of the G&W Electric company
since 1982, is General Manager of the System Protection Division
and was previously employed by Allis-Chalmers Corporation.
He has BSEE, BSME and MBA degrees and is a registered PE.
John is a Senior Member of IEEE and is a member of the High
Voltage Fuse Subcommittee, the Pulp & Paper Industry Committee
and the NEMA High Voltage Fuse Technical Committee and TAPPI.
Example
1: Circuit Diagram
Example 1:
Plot
Example
2: Circuit Diagram
Example 2:
Plot
Example
3: Circuit Diagram
Example 3:
Plot
The Normal
Operating Condition
Triggering
Logic Senses a Fault and Responds
The Actuator
is Initiated
The Switch Opens,
Fuse is Commuted into Circuit
The Fuse Melts and
Interrupts
The Fault has
been Cleared
Installation -
38KV, 2500A Continuous
|
