02-04-2012, 11:08 AM
Identification of Fuse and Recloser Operations in a Radial Distribution System
Identification of Fuse and Recloser Operations in a Radial Distribution System.pdf (Size: 1 MB / Downloads: 51)
Abstract
Recloser and fuses are the main overcurrent protective
devices on distribution systems. Poor coordination adversely
impacts the overall power quality especially from the momentary
voltage interruption and voltage sag perspectives. Therefore, the
purpose of this paper is to develop a novel approach to detect and
identify which protective devices operate to clear a short-circuit
condition using power-quality waveform measurement data. The
proposed approach is intended to evaluate performance coordination
of overcurrent protective devices and help locate faults on the
feeder. The effectiveness of the techniques is demonstrated using
simulation and real-world data. Results indicate a promising potential
for real-world deployment and applications.
Index Terms—Overcurrent protection, power distribution,
power quality (PQ).
I. INTRODUCTION
Aradial distribution system is designed for unidirectional
electric power flow. In such a system, electric power is delivered
from a single source at the substation to multiple loads
through distribution feeders. Because of this configuration, a radial
distribution system requires only one fault interrupter to
clear a fault. Orchestrating the fault-clearing process is referred
to as the coordination of overcurrent protective devices. These
devices include circuit breakers, midline reclosers, and fuses
[1]. They appear in series along a feeder in order to sense a fault
current.
Since there is only one device that ultimately interrupts a
short-circuit condition, the overcurrent protection in distribution
systems involves the coordination between different levels
of protective devices, such as fuse-to-fuse, recloser-to-fuse, and
recloser-to-recloser [2], [3]. Poor coordination adversely impacts
the overall power quality especially from the momentary
voltage interruption and voltage sag perspectives. For example,
improper coordination between a midline recloser and downstream
fuses in a fuse-saving scheme can cause unnecessary momentary
interruptions and voltage sags downstream from the recloser.
Customer facility outages may result [4]. In many cases,
improper coordination can go undetected for a long time until a
major disturbance occurs.
Manuscript received April 3, 2006; revised September 17, 2006. Paper no.
TPWRD-00191-2006.
S. Santoso is with the Department of Electrical and Computer Engineering,
The University of Texas at Austin, Austin, TX 78712-1024 USA (e-mail: ssantoso@
mail.utexas.edu).
T. A. Short is with EPRI, Burnt Hills, NY 12027 USA (e-mail: t.short@ieee.
org).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TPWRD.2007.905386
Therefore, the purpose of this paper is to develop a novel approach
to detect and identify what protective device operated
to clear a short-circuit condition using power-quality waveform
data. For example, it determines whether a recloser or a fuse
operated to clear the fault, and determines the fuse type (e.g.,
65 K, 25 T, etc.). Knowledge of which protective device operated
is used to help identify improper overcurrent coordination,
identify incorrect settings of reclosers, and locate fault locations
since the fault is downstream from the protective device.
The detection and identification of protective device operation
requires power-quality waveform data collected at the distribution
substation. The proposed detection algorithm works by
analyzing voltage and current waveforms during fault events,
extracting and matching their characteristics to the time–current
characteristic curves of devices on the feeder.
Very little work, if any, has been brought to bear on identifying
the operation of protective devices using power-quality
measurement data. An extensive literature search does not reveal
any prior effort. However, there are many articles proposing
techniques to optimize and better coordinate overcurrent protective
devices, such as [5] and [6].
The organization of this paper is as follows. Section II describes
the scope of the problems, Sections III–V present the
proposed algorithms in detail, and, finally, Section VI illustrates
the efficacy of the algorithm using simulation and actual powerquality
monitoring data.
II. PROBLEM DESCRIPTION AND OVERALL APPROACH
Consider a typical radial distribution system, such as the one
illustrated in Fig. 1. Let us make the following assumptions: 1)
fault-induced disturbance waveform data measured at the substation
are available and 2) time–current characteristic curves of
fuses and reclosers or other protective devices used in the feeder
are available.
The problem addressed in this paper can be then stated as follows.
Given three-phase voltage and current fault-induced disturbance
waveforms as illustrated in Fig. 2, determine which
protective device interrupts the fault current. While the approach
can be used to identify the operation of any protective device
with an inverse time–current characteristic, this paper concentrates
on identifying fuse and recloser operations.
If the utility employs a fuse-saving scheme, one can determine
the midline recloser that clears the fault and identifies
which downstream fuses coordinate with the recloser.
The approach for detecting fuse and recloser operations can
be described as follows. We have assumed that a fault-induced
disturbance event has been identified from a power-quality database
using methods described in our earlier work [7], [8]. The
0885-8977/$25.00 © 2007 IEEE
SANTOSO AND SHORT: IDENTIFICATION OF FUSE AND RECLOSER OPERATIONS 2371
Fig. 1. Typical radial distribution system with a power-quality monitor available
at the beginning of the feeder.
Fig. 2. Three-phase voltage and current fault-induced waveforms seen at the
substation.
three-phase voltage and current disturbance waveforms are the
input to the algorithm. Since they are measured at the substation,
the following quantities have to be estimated: 1) the magnitude
of the fault current seen by the overcurrent protective device,
, 2) the duration during which the fault current flows
in the protective device, and 3) the characteristic
of the fault event. These three parameters will then be compared
to fuse and recloser time–current characteristic curves. The protective
device empirical operating point
must be between the fuse’s minimum melting time and maximum
clearing time in the case of a fuse operation, or on the
recloser’s fast or delayed curve in the case of recloser operation.
In addition, the empirical of the fault event must be
higher than the specified by the device manufacturer. The
protective device which satisfies the above criteria is the one
that operated to clear the fault.
III. ESTIMATING PARAMETERS SEEN BY THE PROTECTIVE DEVICE
This section presents techniques to estimate ,
, and empirical , as well as techniques to match
these empirical parameters to the protective device characteristics.
Fig. 3. Voltage (a) and current (b) waveforms along with their wavelet transform-
domain signals.
A. Estimation of Fault Current Duration Seen by the Protective Device
The time-current characteristic (TCC) curves of a protective
device describe how fast the device responds to the overcurrent
condition. Most distribution protective devices have inverse
time–current curves; thus the higher current magnitude is, the
faster the device reacts. The time duration during which the fault
current flows in the protective device can be estimated
directly from the faulted voltage and current waveforms. It is
the duration of the voltage sag or the duration during the high
current magnitude. The exact duration is determined by transforming
the voltage and current waveforms into the wavelet domain.
Since the wavelet transform is sensitive to signal irregularities,
the starting and ending points of a voltage sag will have
high magnitudes in the wavelet domain compared to other parts
of the waveform. The wavelet domain signals are then squared
to amplify the high magnitude outputs. A more detailed description
of the detection technique is described in our earlier work
[8].
Fig. 3 shows voltage and current waveforms along with their
squared wavelet transform outputs. Notice that the beginning
and end of the sag or high current magnitudes are clearly indicated.
In our example, the fault current flows in the protective
device between 0.0249 and 0.0576 s, so s.
B. Estimation of Current Magnitude Seen by the Protective Device
A protective device operates based on the magnitude of the
current that flows through it. Since the voltage and currentwaveforms
are measured at the substation, the current seen by the
protective device must be estimated. If measurements
are taken at the bus level, the load current can be sizeable relative
to the fault current. Therefore, the load current must be separated
from the fault current. One way to estimate the current is
to assume: 1) the current seen by the protective device during
the fault is dominated by the fault current, 2) loads behave as
a constant impedance element, and 3) the load impedance is in
2372 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 22, NO. 4, OCTOBER 2007
Fig. 4. Equivalent circuit for estimating the current magnitude seen by the protective
recloser.
parallel with the impedance to the fault during the fault condition.
Fig. 4 illustrates the assumption for estimating .
The estimation begins with computations of voltage and current
phasors before and during the short-circuit condition. Let
us denote these parameters as , , ,
and . The phasor quantity can be obtained by simply
taking the Fourier transform of the portion of the waveform of
interest and extracting the fundamental frequency component.
Using the waveforms presented in Fig. 3, the time duration before
and during the short-circuit condition is between 0 and
0.0249 s, and between 0.0249 and 0.0576 s, respectively.
The load impedance seen from the power-quality monitor
before the fault occurs can be estimated as follows:
. During the short-circuit condition,
the impedance seen by the PQ monitor is the parallel
combination of the load and fault impedance, thus
. The fault impedance
can be estimated as follows:
(1)
Note that in estimating the fault impedance, we assume that
the load impedance remains the same as before the shortcircuit
condition. The fault current seen by the protective device
can be estimated as follows:
(2)
Although the above method is simple, it provides reasonable
accuracy, and it does not require circuit and load data.
C. Estimation of Empirical Minimum Melt of a Fuse
The minimum melt characteristic of a fuse is very
useful to estimate whether the fuse has melted. It quantifies
the amount of the thermal energy associated with the current
flowing through the fuse link. The empirical value can be
determined by the product of the square of the estimated fault
current and the time during which the fault passes through the
fuse, i.e.,
(3)
If the fuse empirical is higher than the minimum given
by the fuse’s manufacturer, it will likely melt. The manufacturer
minimum melt can be computed using the fuse minimum
melting curve. For example, for a 100-A K-link fuse with a minimum
melting time of 0.01 s at 4931.21 A will have a minimum
melt of 243 000 - [1].
IV. IDENTIFYING RECLOSER OPERATIONS
This section describes the proposed approach to identify a
recloser operation in a fuse-saving scheme. The scheme is intended
to save the fuse from melting in case of a temporary fault.
However, for a permanent fault, the fuse should melt to isolate
the fault.
A. Identifying Recloser Fast and Delayed Operations
Let parameters and be available
from the previous computation. The identification is carried
out by comparing the empirical recloser operating point
to the recloser fast and delayed TCC
curves. The comparison is done by determining the time corresponding
to the empirical fault current seen by the protective
device using an interpolation technique. Due to the inverse relationship
nature between the current magnitude and duration,
the TCC curve is approximated using an exponential function
where the argument of the function is a fourth-order polynomial
function of the natural logarithm of the current flowing in the
recloser. Mathematically, it is
(4)
where are the current magnitude and duration given by
the manufacturer TCC curve. As an example, the fast curve of
a recloser with a phase pickup current of 560 A can be approximated
as follows:
(5)
Using (5), let the time corresponding to be noted
as . The derived point - -
must be on the fast TCC curve since - - is obtained by
interpolating the curve. Fig. 5 illustrates the relative time separation
between and - - . The time difference
between these two points - indicates how close the
empirical point is to the TCC curve
- - - (6)
The time difference - is evaluated for all recloser
TCC curves. The smallest time difference indicates a match (i.e.,
the recloser whose TCC curve produces the smallest operating
time difference).
When the recloser operates in the delayed curve regime, the
time corresponding to the current magnitude seen by the recloser
- - is estimated in the same fashion as described
before. The time difference between the empirical and
derived operation points would be - , i.e.,
- - - (7)
B. Identifying Downstream Fuses That Coordinate Well With
the Recloser
Fuses that coordinate with the recloser are chosen based
on the identified recloser’s TCC curves. A fuse’s TCC curves
SANTOSO AND SHORT: IDENTIFICATION OF FUSE AND RECLOSER OPERATIONS 2373
Fig. 5. Matching recloser empirical operating point to its TCC curves.
Fig. 6. Recloser-fuse coordination: fuse TCC curves must be located between
the recloser’s fast and delayed curves. Note that MM and MC stands for minimum
melt and maximum clearing.
must be located between the recloser’s fast and delayed curves
as illustrated in Fig. 6. Since the recloser has been identified,
the derived and empirical operating points must be very close
if not exact. For this reason, it is reasonable to assume that
- - is nearly identical with . The times
on the fuse’s TCC minimum melting and maximum clearing
curves corresponding to are estimated using the same
approach described before. Let these times be - -
and - - - . The two times must be between the
recloser fast and delayed times, i.e.,
- - - - - -
- - - - - - -
(8)
A match on both times indicates that the corresponding fuse
coordinates with the recloser. In order to accommodate manufacturer
tolerances and other error estimates, error curves
are considered. A match within these curves is assumed to be
valid as well.
Fig. 7. Fuse empirical operating point must be located in between the minimum
melting and maximum clearing curves. Note that MM and MC stand for
minimum melt and maximum clearing.