Background of the Invention
The present invention relates generally to a method of analysing faults
and a load analysis apparatus for use with an electrical utility power system,
and more particularly to a method and an apparatus for distinguishing high impedance,
low current faults from other normal events and activities on the power system.
High impedance faults may be caused by, for example, downed, broken, tangled or
dangling power lines, trees contacting the power lines, and various overcurrent
High impedance, low current faults are more difficult to detect than
permanent overcurrent faults, such as when a transformer fails. Conventional overcurrent
protection devices have time delays which allow a temporary fault to clear, and
if the overcurrent fault persists only then does the device deenergize the power
line. High impedance, low current faults may initialize the timing circuits of
the overcurrent protection devices but, by the end of the delay, the high impedance
nature of the fault limits the fault current to a low value. The overcurrent protection
devices cannot distinguish this low fault current from the levels of current ordinarily
drawn by customers, so the lines may remain energized even though a conductor has
US Patent No. 5,103,365 discloses a device for use in a three-phase
four-wire multi-grounded distribution system that detects downed conductors to
protect the distribution line from arcing ground faults and overcurrents. US Patent
No. 4,466,071 discloses a high impedance fault detection apparatus and method
that monitors the high frequency components for detecting the presence of high
impedance arcing fault on a high voltage power line.
Other methods of detecting high impedance faults have focused on
detecting third harmonics generated by the arcing behaviour of the high impedance
faults. These earlier methods use detection algorithms having variations in harmonic
current as the detection parameter. For instance, U.S. Patent No. 4,851,782 to
Jeerings detects high impedance, low current faults by analyzing third harmonic
currents on the power lines.
By relying only on the arcing behavior for detection, high impedance
fault detection systems using such methods experience significant reliability
problems. These systems lack security against a false trip, causing unnecessary
blackouts for the utility's customers. For example, a tree limb momentarily touching
a power line may cause a momentary fault, which is cleared when the tree limb moves
away from the power line. These earlier systems may misinterpret this momentary
tree contact as a permanent high impedance fault, and in response cause breakers
to trip to deenergize the line. Such systems may also interpret normal switching
actions of the power system protection equipment as a permanent high impedance
fault and cause unnecessary trips.
A primary goal of electrical utilities is to minimize such false
fault detections. Most utilities need a load analysis system which detects and
deenergizes a power line only for hazardous faults, such as when a broken conductor
is on the ground. During other minor fault conditions, such as a broken power
line dangling out of the reach of the public, it may be desirable to leave the
power line energized. Although a dangling power line is hazardous, service interruptions
to the electric utility's customers can also pose significant safety problems.
Thus, a need exists for an improved high impedance fault detection
system for electrical power utilities which is directed toward overcoming, and
not susceptible to, the above limitations and disadvantages.
Summary of the Invention
Thus, there is a clear need for a load analysis system having fault
detection techniques which accurately identify hazardous faults requiring line
deenergization, and which accurately discriminates, or distinguishes, a hazardous
fault from other events for which the line should remain energized. Embodiments
of the present invention encompass such a load analysis system which minimizes
unnecessary power service interruptions and outages.
An aspect of the present invention provides a method of analyzing
faults occurring on a distribution circuit coupled to an AC power source, comprising
the steps of monitoring a load current flowing over the distribution circuit;
analyzing the load current over time; and distinguishing, based on the analysis
of the load current, the difference between a first type of fault requiring deenergization
of the distribution circuit, and a second type of fault for which the distribution
circuit remains energized; characterised by classifying faults by detecting the
presence of arcing, changes in load level indicative of load loss, and the occurrence
of overcurrent level in the load current on the distribution circuit.
In accordance with an illustrated embodiment of the present invention,
a method of analyzing faults occurring on a distribution circuit coupled to an
AC power system, includes the step of monitoring a load current flowing over the
distribution circuit. In an analyzing step, the load current is analyzed over
time. In an identifying step, either the occurrence of a normal system event, a
hazardous fault or a minor fault are identified from the analyzed load current.
As used herein, hazardous faults are those requiring deengerization of the distribution
circuit, and minor faults are those for which the distribution circuit may remain
energized. The distribution line should also remain energized during normal system
events, such as during switching events.
Another aspect of the invention provides a load analysis apparatus
for analyzing faults occurring on a distribution circuit coupled to an AC power
source, comprising a monitor for monitoring a load current flowing over the distribution
circuit and in response thereto, generating a load signal; and a controller responsive
to the monitor for analyzing the load current over time, and for distinguishing
from the analyzed load current the occurrence of a first type of fault requiring
de-energization of the distribution circuit, and of a second type of fault for
which the distribution circuit remains energized; characterised in that the controller
is arranged to classify faults by detecting the presence of arcing, changes in
load level indicative of load loss, and the occurrence of overcurrent level of
the load current on the distribution circuit.
In accordance with another illustrated embodiment of the present
invention, a load analysis apparatus is provided for analyzing faults occurring
on a distribution circuit coupled to an AC power system. The apparatus includes
a monitor for monitoring a load current flowing over the distribution circuit and
in response thereto, generating a load signal. The apparatus also has a controller
responsive to the monitor for analyzing the load current over time, and for identifying
the occurrence of hazardous and minor faults from the analyzed load current.
An overall object of embodiments of the present invention is to provide
a load analysis high impedance fault detection method and apparatus for minimizing
unnecessary power service interruptions and outages.
A further object of embodiments of the present invention is to provide
a load analysis apparatus and method for accurately identifying and discriminating
selected high impedance faults which require a power outage to clear, from other
power system events and activites during which it is preferable that the power
line remains energized.
Still another object of embodiments of the present invention is to
provide a load analysis system which is more reliable than the earlier systems.
Embodiments of the present invention relate to the above features
and objects individually as well as collectively. These and other objects, features
and advantages of embodiments of the present invention will become apparent to
those skilled in the art from the following description and drawings.
Brief Description of the Drawings
Detailed Description of a Preferred Embodiment
- Fig. 1 is a schematic block diagram of one form of a load analysis apparatus
of an embodiment of the present invention;
- Figs. 2 and 3 are portions of a flow chart illustrating a first embodiment
for operating the load analysis apparatus of Fig. 1; and
- Figs. 4 and 5 are portions of a flow chart illustrating a second embodiment
for operating the load analysis apparatus of Fig. 1.
Fig. 1 illustrates an embodiment of a load analyzer 10 constructed
in accordance with the present invention. The analyzer 10 distinguishes a high
impedance fault from other occurrences on a distribution system conductor, power
line, or feeder 12. The feeder 12 receives power from an AC power source, such
as a generating station 14, through a substation 16. Other feeder lines (not shown)
may also receive power from the generating station 14 and exit the substation
16. The feeder line 12 delivers power from the substation 16 to a variety of customers,
such as customer 18. Altogether, the generating station 14, substation 16, and
feeder 12 delivering power to the customer 18 illustrates a portion of an electrical
utility's power system 20.
Between the substation 16 and the customer 18, the feeder line 12
may be subjected to a variety of different types of events, activities and faults,
such as: a downed conductor 22, a dangling conductor 24, momentary contact of
a tree or other object 25 with the feeder 12, an overcurrent event 26, or a switching
event 28 performed by a conventional recloser or the like. An overcurrent event
26 may be caused by a variety of events, such as a customer overload, the touching
or tangling of two or more phase conductors, a lightning strike, or a broken conductor
hitting grounded object or a conductor. These grounded objects may include a feeder
pole guy wire, an under-built neutral conductor, or the like.
Although utility engineers distinguish between recloser operation
and switching events, with recloser operating automatically and switches being
manually operated, the operation of both are collectively referred to herein as
"switching events" unless otherwise indicated. Regarding the various faults and
normal operations of the power system 20 described herein, the following terms
are used herein interchangeably: situation, occurrence, operation, event, and
The load analyzer 10 includes a monitoring device such as a transducer
30 coupled to the feeder 12 as indicated schematically by line 32. Monitoring device
is defined broadly herein to include sensing devices, detecting devices, and any
other form thereof known to be interchangeable by those skilled in the art. The
illustrated transducer 30 senses or monitors a load current IL flowing
through feeder 12. In response to the load current IL, the transducer
30 produces a load current signal 34 that indicates the magnitude and waveform
of current flowing in feeder 12. The transducer 30 may be a conventional transducer
or equivalent device, such as multiple current transformers typically with one
current transformer per phase plus one on the neutral.
The load analyzer 10 also includes surge protection, for example,
a surge suppressor or protector 36. The surge protection 36 which may be supplied
with the transducer 30, as illustrated, or as a separate component. The surge protector
36 protects the load analyzer 10 from power surges on the feeder 12, such as those
caused by lightning strikes or the like.
A controller 35 receives the load current signal 34 from the transducer
30. The controller 35 includes a signal conditioner 38 for filtering and amplifying
the load current signal 34 to provide a clean conditioned load current signal
40. Preferably, the signal conditioner 38 includes a low pass filter for satisfying
the Nyquist criteria of sampling known to those skilled in the art. The signal
conditioner 38 also amplifies the load current signal 34 for the appropriate gain
required by an analog to digital (A/D) converter 42. For example, the dynamic range
of signals received on a power system 20 range from 10 Amps to 10,000 Amps, so
the signal conditioner 38 appropriately scales these signals for conversion by
the A/D converter 42 from an analog signal 40 into a digital load current signal
44. The controller 35 may include a digital signal processor for determining a
frequency spectra of the digital signal 44 to generate a frequency spectra signal.
The controller 35 includes a discrete A/D converter 42 when transducer
30 is an analog device. The transducer 30 may also be implemented in a digital
device which incorporates the signal conditioning function of conditioner 38 and
the analog-to-digital conversion function of the A/D converter 42.
The load analyzer 10 also includes a line current sampling device
or sampler 45 which samples the digitized current signal 44 at selected intervals
to provide an accurate representation of the load level during rapidly changing
conditions, such as during overcurrent faults. For example, the sampler 45 may
measure the line current and determine either the fundamental frequency component
or the rms (root-mean-square) current component. In a preferred embodiment, one
rms value is calculated for each one or two cycles of the fundamental power system,
such as sixty or thirty values per second for a 60 Hz nominal power system frequency.
The sampler 45 provides a sampled current signal 46 corresponding to the sampled
line current values. The sampled current signal 46 is supplied via a microcomputer
bus 47 to a computing device, such as a microcomputer system 48. The microcomputer
system 48 has a computer, such as a single board computer 50, coupled with a memory
device, such as a random access memory 52, and a data storage device, such as
a hard disk 54. A suitable microcomputer system 48 may include a conventional
personal computer or any other form thereof known to be interchangeable by those
skilled in the art.
The controller 35 includes a circuit breaker interface 60 for receiving
a trip command signal 62 from the computer 50 via the bus 47. In response to the
trip command signal 62, the interface 60 sends a trip signal 64 to a circuit breaker
trip circuit 66. The trip circuit 66 drives a circuit breaker (not shown) located
at substation 16 to deenergize the feeder 12. The controller 35 may include an
optional serial interface 68, such as a modem for sending and receiving a peripheral
device signal 72 over a telephone network. The interface 68 may communicate with
an external peripheral device 70, such as a remotely located power distribution
control center. In some systems, the peripheral device 70 may provide a remote
input to the load analyzer 10 via the serial interface 68, for example to override
previous programming of the load analyzer, such as the initial settings, sensitivity
settings, operational delays, etc.
The controller 35 may also include an output device, such as a visual
display device 74 or a printer. Preferably, the output display provides a visual
indication of the status of the load analyzer 10, the feeder line 12, and previous
operating conditions of the feeder. The controller 35 may also provide an alarm
signal 76 via bus 47 to an alarm 78 which may be visual, audible, or both.
As an overview, the operational philosophy of the system has the
load analyzer 10 looking for arcing over a long period of time, on the order of
several seconds to several tens of seconds, or even minutes, before deciding that
a downed conductor condition 22 indeed exists. To reach this downed conductor decision,
the arcing is accompanied by either a significant loss of load, signifying the
beginning of the event, or by an overcurrent fault. If prolonged arcing is detected
that is not accompanied by a loss of load or by an overcurrent condition, the
load analyzer 10 interprets the situation as being associated with other types
of events, such as tree contact 25 or an insulator failure. Preferably, the load
analyzer 10 recognizes a significant loss of load because it is precipitous and
precedes any overcurrent or open breaker conditions. Preferably these other types
of arcing conditions activate an alarm separate from a downed conductor alarm
because utility practices may dictate different responses to the two types of arcing
The presence of arcing may be determined by other methods known to
those skilled in the art. Other methods known to those skilled in the art may be
used to detect a significant loss of load, and to recognize overcurrent and open
Preferably, the load analyzer 10 allows conventional overcurrent
protection, such as fuses, reclosers, and conventional overcurrent relays to operate
first. To accomplish this objective, the load analyzer 10 preferably will delay
issuing an output trip signal 64 until a sufficient time period has elapsed after
the beginning of the event. Thus, the load analyzer 10 acts as a last resort protection
device when conventional overcurrent protection devices have not yet operated.
Preferably, this minimum operating time of analyzer 10 may be programmed by the
user, to accommodate system protection philosophies which vary from one utility
to the next.
By recognizing and distinguishing the different reactions of the
power flow through feeder 12 when subjected to the various events 22-28, the load
analyzer 10 determines what type of event has occurred, whether it is a hazardous
fault or a minor fault, and what response is appropriate. For example, the downed
conductor 22 usually shows a loss of load, unless it is located far from the substation.
The downed conductor 22 touches and arcs to a high impedance object, such as the
ground. The arcing behavior of the downed conductor 22 lasts for a significantly
long period of time, on the order of minutes as opposed to fractions of a second
A broken and dangling conductor 24 may also show a loss of load.
However, no arcing behavior is exhibited because the end of the dangling conductor
24 does not touch any high impedance objects, as does the downed conductor 22.
Momentary contact of a grounded object, such as the tree contact 25, with the feeder
12 may exhibit an arcing phenomenon without any loss or increase of load. Overcurrent
activities 26 may involve any combination of phases and the neutral exhibiting
an overcurrent which exceeds a specified level, with or without arcing behavior.
A switching event 28 may or may not show any significant load change,
and does not show prolonged arcing behavior. Any arcing from a switching event
28 typically lasts less than one second. For example, a recloser action switching
event 28 may show significant load increase and load loss, with or without arcing
behavior. This load increase and load loss response to the recloser action 28
is repeated in a similar pattern at intervals dictated by the established practice
of recloser operation.
A downed conductor 22 may initially look like several different types
of faults which complicates the diagnosis. When a downed conductor first breaks
and starts falling to the ground, it may hit a guy wire or an under-built neutral
conductor, and appear as an overcurrent activity 26. This overcurrent activity
causes an overcurrent breaker (not shown) to trip and deenergize the line. A short
time later, a recloser operates to reenergize the line, which appears as a switching
event 28. By the time the line has been reenergized, the downed conductor 22 may
have slipped off the guy wire or the under-built neutral and made arcing contact
with the ground. The low level current flowing between the downed conductor 22
and the high impedance ground is usually not high enough to cause further operation
of the overcurrent breaker (not shown); hence the need for the load analyzer 10.
Generally, most electrical utilities only want a feeder breaker to
trip for a downed conductor 22. The load analyzer 10 identifies downed conductors
22, broken and dangling conductors 24, tree contact 25, and other events, such
as overcurrent (Ovc) fault situations 26. To initially distinguish between normal
operating conditions and a disrupting event, the load analyzer 10 continually
monitors a value of a parameter of the power flowing through the feeder 12, here,
the rms of the load current IL. Once a disrupting event is detected,
the load analyzer 10 then identifies what type of event has occurred. This event
identification is accomplished within the microcomputer system 48 by analyzing
the load current patterns in two stages of event classification logic (see Figs.
2 and 3).
Three important elements of this load pattern analysis scheme are
detecting and analyzing:
- 1) the presence of arcing;
- 2) changes in the load level; and
- 3) occurrence of overcurrent activity.
For example, whenever a power line breaks, some percentage of the
total feeder load current IL is lost causing an initial disruption in
the rms data monitored by transducer 30. This conductor breakage initiates a sequence
of events in which there is a sudden loss of load current IL. If this
sudden loss is followed by an indication of arcing, then a downed conductor situation
22 most likely exists. This sequence of events provides a good indicator of the
downed conductor scenario, which calls for tripping the feeder. If no arcing follows
the sudden load current loss, then it is likely that a dangling conductor situation
24 exists for which many utilities allow the feeder 12 to remain energized. If
an overcurrent fault occurs and then vanishes, but is followed by a period of sustained
arcing, this is also an indication of a downed conductor (with or without loss
The load analyzer 10 classifies the events by using these three elements,
arcing, load level changes, and overcurrent level excursions, to define three
analysis variables, flags or signals for use in the logic analysis:
The load analyzer 10 uses these three variables to generate pre-action outputs
or commands, action outputs or commands, and several diagnosis outputs. The pre-action
- 1) Arc,
- 2) Load-Loss, and
- 3) Overcurrent Level.
The action commands include:
- 1) Trip-Ready,
- 2) Alarm, and
- 3) Normal.
The action command Trip is initiated by the circuit breaker interface 60 sending
signal 64 to the trip circuit 66. The Alarm pre-action and action commands are
sent to the alarm 78 by signal 76. The Trip-Ready, Trip, Alarm, Wait and Normal
commands are preferably shown on the display 74 along with the diagnosis outputs.
The diagnosis outputs characterize the particular conditions of the feeder status,
such as normal or one of the five situations 22, 24, 25, 26 or 28. The feeder
status and commands may also be sent via signal 72 to a peripheral device 70.
- 1) Trip,
- 2) Alarm, and
- 3) Wait.
Referring Figs. 2 and 3, a flow chart 100 illustrates a method of
detecting, analyzing and discriminating between faults and normal events on a
power line with reference to the operation of the illustrated load analyzer 10.
The flow chart 100 shows one manner of operating the microcomputer system 48.
First, data and variable preparation operations of the load analyzer 10 are described.
The Analysis Unit
To determine the presence of arcing and load level change, the rms
value of the load current IL is monitored by transducer 30, conditioned
by the signal conditioner 38, converted from an analog to a digital signal by
converter 42. The digital signal 44 is sampled at the selected interval by sampler
45 to determine the sampled current signal 46 which is an input to the computer
50. To check for permanent variations in the load level and persist t arcing behavior,
a sufficient time or number of rms data values, referred to herein as an analysis
unit, is used. In the illustrated embodiment, this analysis unit is set at five
seconds, which is equivalent to 300 rms data values. However, it is apparent that
the analysis unit may be set at other values depending upon the particular implementation,
such as by analyzing the energy of the load current IL over every two
Preferably, the analysis unit has a duration long enough to allow
operation of conventional automatic protection devices, such as overcurrent breakers
(not shown), installed on the feeder 12. Most utilities want the load analyzer
10 to operate only after operation of the conventional overcurrent protective
devices (not shown), so the analysis unit is selected to a value on the order of
30 seconds. During the time analysis unit, the load analyzer 10 analyzes the pattern
of the load to determine values for the variables Arc, Load-Loss, and Overcurrent
Data Management and Storage
The normal load level situation exists when there are no abnormal
indications during an analysis unit. The feeder line 12 is continually monitored,
and the normal load level is updated over time analysis unit by analysis unit.
Each succeeding analysis unit is compared with the previous analysis unit to provide
a long term load comparison. The RAM 52 has two data storage locations designated
herein as Packet #1 or P1, and Packet #2 or P2. The newly
updated normal data is stored in the Packet #1 location in the RAM, and the new
incoming data fills Packet #2.
If Packet #2 is found to be normal data, then the 300 data values
in Packet #2 are moved or swapped into Packet #1, leaving Packet #2 empty. As the
data is read by the computer 50, if an abnormal rms value is found in Packet #2,
then a data pointer device, index or software control is frozen at the abnormal
rms value. This abnormal rms value is then remembered and stored during the analysis,
with the location of the abnormal within the data pack indicated by the pointer.
All of the data values in Packet #2 which were entered before freezing of the
pointer are moved into and stored in the Packet #1 located, eliminating the first
portion of the data which was previously stored in Packet #1. Thus, Packet #1
contains and stores the newest normal data, which is a combination of the data
from the previous analysis unit and the current analysis unit where the abnormal
value is encountered. Packet #2 then stores only the post-disruption rms data
Abnormality or Disruption Check
To check for the presence of arcing and load level changes for each
phase and the neutral, two data differencing checking routines are used:
These data diffferencing routines are useful to find any trend in this time series
of data. The short term differencing routine (DF) is used to set the
Arc flag, and the long term differencing routine (DL) is used to set
the Overcurrent Level flag and the Load-Loss flag.
Short Term Trend Determination
- 1) the short term or first difference DF, and
- 2) the long-term difference DL.
The short term differencing routine (DF) looks for changes
in the incoming rms data stored in Packet #2 to recognize any random behavior of
the data. As used herein, "short term variations" refers to variations occurring
within about one second or so. The first difference DF is found by determining
the difference between the neighboring rms values, indicated as X, in the incoming
DF(i) = X(i) - X(i-1)
When the value of the first difference DF is out of a
range of certain values defined by an abnormality threshold TAB, a difference
count CF is incremented by one. In the analysis unit, the difference
count CF is accumulated and compared with a preset arc threshold value
TARC. If the difference count CF is greater than the arc
threshold value TARC, then the flag of the variable Arc is set to Y
for yes, and otherwise is set to N for normal as a default. The random activity
of arcing causes these short term difference trends. In addition to this method
of arc detection, the load analysis process may also use other equivalent methods
of arc detection known to those skilled in the art.
Long Term Trend Determination
The long term differencing routine (DL) looks for changes
between the incoming rms data stored in Packet #2 and the data from the previous
analysis unit stored in Packet #1 to recognize any changes in the trend of the
load level. The long term difference DL is found from calculating the
long term load level trend from the rms data values as follows:
DL(i) = XP2(i) - XP1(i)
where the subscript "P2" refers to the incoming data stored in Packet #2, and the
subscript "P1" refers to data from the previous stream stored in Packet #1.
When the long term difference DL is above a level
of overcurrent threshold TOVC, then a flag for the variable Overcurrent
Level is set to Y for yes, and otherwise as a default remains at N for normal.
Once the Overcurrent Level flag is set to Y during an analysis unit, it remains
unchanged for the remainder of the analysis unit.
When the long term difference DL is above a level
of significant load loss threshold TLOSS, then the flag of the variable
Load-Loss is set to Y for yes, and otherwise remains as a default N for normal.
For example, if the load drops percipitously, e.g., from 200A to 20A within a
few cycles, then the Load-Loss flag is set to Y for yes to indicate such a significant
loss of load. Once the Load-Loss flag is set to Y during an analysis unit, it
remains unchanged for the remainder of the analysis unit.
Load Pattern Analysis Logic
Having described the manner in which data is handled by the load
analyzer 10 and the manner in which the data is checked for abnormalities, the
manner of discriminating among the conditions to determine what type of fault
or event, such as events 22, 24, 25, 26 or 28, have occurred is described next.
With the flags set to Y for yes or N for normal for the three variables, Arc,
Load-Loss, and Overcurrent Level, the load pattern analysis proceeds in two stages
of analysis and classification.
In the first stage, the flags of the three variables are analyzed
to generate three pre-action commands (Trip-Ready, Alarm, and Normal) and the
diagnosis status conditions or states. The six diagnosis status conditions with
their various combinations of flag settings are shown in Table 1. If the pre-action
command is either Alarm or Normal it is the final analysis output, and in the illustrated
embodiment, no further analysis occurs. If the pre-action command is Trip-Ready,
the second stage of load pattern analysis begins.
First Stage Load Pattern Analysis
Load Analyzer Outputs
Diagnosis Status-Condition / State
Broken & Dangling
In the second stage of load pattern analysis, only two of the three
variables (Arc, Load-Loss, and Overcurrent Level) are used, depending upon which
diagnosis status condition is encountered in the first stage. For example, for
the downed conductor 22 of case #3 in Table 1, the two variables considered are
the Arc and the Load-Loss as shown in Table 2. Depending upon the status of the
flags, the action commands are either Trip, Alarm, or Wait.
Second Stage Load Pattern Analysis for a Downed Conductor
Case #4 of Table 1 diagnoses an overcurrent event 26 comprising a
broken conductor hitting a neutral. This case uses another second stage of analysis
based on consideration of the two variables the Arc and Overcurrent Level, as shown
in Table 3. Depending upon the status of the flags, the action commands are either
Trip, Alarm, or Wait.
Second Stage Load Pattern Analysis For a Broken Conductor Hitting Neutral
Although several cases in Table 1 have a first stage pre-action command
of Alarm, it is apparent that some utilities may prefer a Trip-Ready pre-action
command for some of these situations. For example, a Trip-Ready pre-action command
may be preferred for tree contact 25 on feeders through certain forested areas,
such as in a National forest, park or monument area. The second stage analysis
would then be conducted for any tree contact events 25. As another example, in
urban settings a dangling conductor 24 may be considered a safety hazard of the
same degree as a downed conductor 22. For a dangling conductor 24, the second
stage of analysis considers the states of the Load-Loss and Overcurrent Level flags
to provide action commands of Wait and Trip, as shown in Table 4. Individual utility
operators may determine and set these truth tables to other values (yes or no)
as they deem appropriate for particular implementations.
Second Stage Load Pattern Analysis for a Dangling Conductor
The analysis is divided into first and second stages with the first
stage having a pre-action Trip-Ready command to allow more certainty and security
in the identification of the event as a particularly hazardous one requiring trip,
such as the downed conductor 22. The Wait action refers to waiting a few seconds
before sampling again to allow time for operation of the coordinated protection
devices (not shown) installed on the feeder line 22. Such coordinated protection
devices include overcurrent relays, fuses and reclosers. For example, if the action
command is Wait, the illustrated load analyzer 10 waits for about five seconds,
and then aborts all the flags and settings, then repeats the first stage load
pattern analysis of Table 1.
To further enhance security of the load analyzer 10 and allow operation
of the coordinated protection devices, an initial delay period is inserted before
the first stage analysis begins. The two stage load pattern analysis is not invoked
until there is an indication of an abnormal event which may lead to a trip or
alarm decision. To accomplish this initial delay, the load analyzer 10 has an initialization
timer (not shown) which may be internal to the computer 50, for setting an initial
delay period. In the illustrated embodiment, the initial delay is set for 30 seconds,
although duration selected depends upon the particular implementation because feeders
typically have unique operational characteristics and parameters.
Upon detecting an abnormal event, the initialization timer begins
to run to allow the conventional protection devices (not shown) installed on the
feeder 12 to operate before the first stage analysis begins. If the conventional
protection devices have not worked by the end of the 30 second delay period, before
issuing a command, the load analyzer 10 delays an additional five seconds while
the data storage Packet #2 is filled. The second stage analysis may immediately
follow completion of the first stage analysis. However, in some implementations
it may be preferable to insert an interstage delay period between the two analysis
stages, for example by emptying and refilling Packet #2 before commencing the
second stage analysis.
The Wait command of the second stage analysis also increases confidence
in the diagnosis that indeed a fault situation requiring a trip exists to prevent
unnecessary trips. For example, many utilities want the load analyzer 10 to trip
only when arcing is accompanied by a broken conductor per Table 2 (Y for Arc and
Load-Loss). If the arcing is intermittent and goes away for seconds at a time,
this indicates a conductor on the ground. The load analyzer 10 looks for immediately
present arcing, then waits if there are no indications of arcing or load loss,
or alarms if only Load-Loss is detected.
The case #2 situation may be caused by a variety of overcurrent events
26, such as a customer overload, the touching or tangling of two or more phase
conductors, or a lightning strike which occurred within the initial delay period,
here, within the preceding illustrated 30 second period. Tangled conductors may
appear as either case #2 or #7 conditions, depending upon whether or not arcing
exists. If the conductors remain tangled for more than the illustrated 30 second
initial delay period, then the conductors are probably permanently tangled. To
prevent melting of the tangled conductors, the overcurrent protection devices (not
shown) operate to deenergize the feeder 12 well before the end of the 30 second
delay period. Momentary touching or contacting of two or more conductors may appear
as case #2, #4, #6 or #7, each of which have an Overcurrent Level flag set to Y.
While the exact diagnosis of some of the overcurrent events 26 may elude the load
analyzer 10, a reasonable conjecture based on strong possibilities may be made
as to which overcurrent event has occurred. Operator inspection of the feeder
12 may be ultimately required to attempt to determine which type of overcurrent
event 26 indeed occurred.
Threshold Value Settings
The first difference DF is compared to the abnormality
threshold TAB, the difference count CF is compared with the
preset arc threshold TARC, and the long term difference DL
is compared to the level of overcurrent threshold TOVC and to the load
loss threshold TLOSS. These threshold values TAB, TARC,
TOVC and TLOSS, may be established by studying the normal
load data in a statistical manner. The load level change is determined by comparing
the normal or predisruption load level stored in Packet #1 with the new post-disruption
load level stored in Packet #2 of the RAM 52.
Load Analysis Flow Chart
Referring to Figs. 2 and 3, the flow chart 100 for the illustrated
software embodiment of the load analyzer 10 shows one method of analyzing faults
occurring on the feeder 12. Before beginning the actual load analysis routine
an initialization routine or device 102 performs an initialization process. The
initialization routine 102 sets several flags and indices. A disruption index
is set to zero to indicate a normal current status, with no disruption thus far
being found. The difference counter CF is set at zero because no arcing
behavior is found under normal conditions. The flags of Arc, Load-Loss and Overcurrent
Level are set to N for normal as a default. A stage index is set at zero to indicate
that the data analysis is not yet in second stage. The data Packet #1 in RAM 52
is filled with 300 normal rms data values representing the load current IL.
From the rms data in Packet #1, the first difference DF is calculated,
and the initialization process is then complete.
After initialization a read data and increment pointer device or
routine 104, sequentially reads each value of rms data, and advances a data pointer
by one as each data point is read. In a stage one checking routine or device 106,
the stage is determined by checking the stage index. A stage index of "one" indicates
the data analysis is in the second stage, and zero indicates that it is not. In
a first disruption checking device or routine 108, the disruption index is checked
to see if the rms data value received differs from that previously received to
indicate the occurrence of one of the events 22, 24, 25, 26 or 28.
In the illustrated embodiment, if both the stage index and the disruption
index are zero, then a first difference calculating device or routine 110 calculates
the first difference DF according to the equation described above. The
first difference DF is then compared to the abnormality threshold value
TAB by an abnormality threshold comparison device or routine 112. If
the first difference DF is above the abnormality threshold TAB
or below the negative value of the abnormality threshold (-TAB), the
disruption index is set to "one".
In a second disruption checking routine or device 114, if the disruption
index remains zero and no abnormality has been found, a pointer status checking
routine or device 116 determines whether the analysis unit is complete. In the
illustrated embodiment, the analysis unit was chosen as five seconds, or 300 normal
rms data units. If the data pointer is indeed at 300, the analysis unit is complete
and the data in Packet #2 is swapped into Packet #1 by a data packet swap routine
or device 118. Then, in a pointer reinitialization routine or device 120, the
data pointer is reset to zero. If the illustrated checking routine 116 indicates
the data pointer is at a value less than 300, the rms values of the next data point
are read by the data read and increment pointer routine 104, and the cycle continues.
If the disruption status routine 114 determines that the disruption
index is "one," a pointer freeze and data packet rearrangement routine or device
122 freezes the data pointer to remember the value and location of the rms data
at the disruption time. In the illustrated embodiment, the data packet rearrangement
portion of routine 122 attaches the predisruption data from Packet #2 to the end
of the data stream stored in Packet #1, which displaces the earliest data from
Packet #1. Following this data packet rearrangement, the pointer is reset to zero
by a another reset pointer routine or device 124, and the next rms data value
is read by routine 104.
Although the stage index is still set at zero, the disruption status
check routine 108 notes the disruption index is now set at "one." When the disruption
index is "one," a difference calculation device or routine 126 calculates the first
and the long term difference DL according to
the equations discussed above. In a first difference comparison routine or device
128, if the first difference DF is found to be greater than the abnormality
threshold TAB, or smaller than the negative value of the abnormality
threshold (-TAB), then the difference counter CF is incremented
In a long term difference comparison routine or device 130, the long
term difference DL is compared to two different thresholds. If the value
of the long term difference DL is greater than the level of overcurrent
threshold TOVC, then the Overcurrent Level flag is set to Y. If the
long term difference DL is greater than the load loss threshold TLOSS,
the Load-Loss flag is set to Y. These flag settings remain unchanged until the
data Packet #2 is full as determined by a second pointer status routine or device
132. When the illustrated pointer status routine 132 determines that data Packet
#2 is full, a difference count comparison device or routine 134 compares the difference
with the arc threshold level TARC to set the Arc flag.
If the difference count CF is greater than the arc threshold level
TARC, the Arc flag is set to Y, and if CF
is less than or equal
to TARC, then the Arc flag is set to N. The flag settings and status
of the indices are carried as a signal 135 which links together the portions of
the flow chart 100 shown in Figs. 2 and 3.
With the flag settings of the three variables, Load-Loss, Overcurrent
Level, and Arc, set as described above, the event classification logic is applied
in two stages, as described above with reference to Tables 1-4. Referring now
to Fig. 3, another stage status checking device or routine 136 determines whether
the load pattern analysis is in the first or second stage, with the first stage
being indicated by a stage index of zero, and the second stage being indicated
by a stage index of "one." In a first stage of event classification routine or
device 138, the logic illustrated in Table 1 is performed. A trip-ready output
status routine or device 140 determines whether the pre-action command of the first
stage logic is Trip-Ready or not.
If the pre-action command is indeed Trip-Ready, a stage advancing
device or routine 142 changes the stage index to "one," indicating commencement
of the second stage analysis event classification. Whether the analysis is in the
first or second stage, an output and resetting device or routine 144 generates
a signal 145 to empty the data stored in Packet #2 of the RAM 52. The signal 145
links together the portions of the flow chart 100 shown in Figs. 2 and 3. The
load analyzer 10 begins reading data using the read data routine 104. The new incoming
data stream is stored in Packet #2 for the second stage of event classification.
The second stage of analysis is noted by routine 106, and routine 126 calculates
and DL. Then routine 130 sets the Overcurrent Level flag
and the Load-Loss flag. When routine 132 indicates that Packet #2 is full, the
Arc flag is set by routine 134. The stage status routine 136 determines the load
analyzer 10 is in the second stage of analysis, and a second stage event classification
device or routine 146 interprets the flag settings in accordance with the logic
of Tables 2, 3 and 4 (if used), as applicable.
If the illustrated trip-ready routine 140 determines that the pre-action
command of the first stage event classification is either Alarm or Normal, this
is the final action performed by the load analyzer 10 and the status determined
with reference to Table 1 is the final diagnosis status condition. Output status
signals are sent to the display 74 and any peripheral device 70 to indicate the
status of the system. For an event classification generating an Alarm output, the
alarm signal 76 activates the alarm 78. If the second stage of event classification
reaches the action command of Trip, then the breaker interface 60 sends a trip
signal 64 to the trip circuit 66 to deenergize the feeder line 12 by disconnecting
it from the substation 16.
Advantageously, the load analyzer 10 and method of analyzing faults
described herein identify load patterns to recognize and distinguish situations
in which a power line is broken or intact. In this manner, security against false
trips is provided in the classification and detection of high impedance faults.
Thus the overall security and reliability of downed conductor fault detection
is greatly enhanced beyond the capabilities of the earlier systems described in
the background portion above.
Referring Figs. 4 and 5, a flow chart 200 illustrates a method of
detecting, analyzing and discriminating between faults and normal events on a
power line with reference to the operation of the illustrated load analyzer 10.
This second embodiment may be used alone, or in conjunction with the first embodiment
of Figs. 2 and 3. The flow chart 200 shows an alternate preferred manner of operating
the microcomputer system 48. The variables used in flow chart 200 are defined
Overcurrent Flag Set
High Rate-of-Change Flag Set
Significant Loss-of-Load Flag Set
Arcing Flag Set
Three Phase Event Flag Set
Breaker Open Flag Set
LOL Flag was the flag which initially caused the transition
into the non-normal state.
Overcurrent flag was set at some time while the algorithm was
in the non-normal state.
Conventional Protection Coordination Timer
Conventional Protection Malfunction Timer
Breaker Reset Timer
When the load analyzer 10 is operated in accordance with flow chart
200, upon receiving a start command 202 from an operator, an initialization routine
or device 204 performs an initialization process to set flags and indices to initial
values as described below. A flag setting routine or device 206 uses conventional
checking routines to determine from the sample current signal 46 whether one or
more of the following events has occurred:
- 1. A significant, precipitous loss of load (LOL);
- 2. An overcurrent level is detected (OCF);
- 3. A high rate of change in current (ROC);
- 4. Significant arcing is detected (ARC); or
- 5. A breaker-open condition is detected (BRKR).
When a loss of load occurs, the flag setting device produces a loss
of load flag set (LOL) signal 208. When an overcurrent level is detected, the flag
setting routine 206 generates an overcurrent flag set (OCF) signal 210. When the
flag setting routine 206 determines that a high rate of change in current has
occurred, a high rate of change flag set (ROC) signal 212 is generated. When the
flag setting routine detects significant arcing, it generates an arcing flag set
(ARC) signal 214. When the circuit breaker coupled to trip circuit 66 is open,
the flag setting routine 206 produces a breaker-open flag set (BRKR) signal 216.
After initialization by routine 204, a line status checking device
or routine 218 constantly monitors signals 208-216 from the flag setting routine
206. If none of the flags are set, the system is considered to be in a normal
state, and the checking routine 218 issues a NO signal 220 to initialize the next
sequence of the checking routine.
When the checking routine 218 determines one or more of the flag
setting signals 208-216 indicate that an event has occurred, the analyzer 10 enters
a triggered state and the checking routine 218 issues a YES signal 222. In response
to the YES signal 222, two timers are set with a set TIMER1 device or routine 224
setting the first timer 226 and a set TIMER2 routine or device 228 setting the
second timer 230. The first timer 226 has a time duration selected to coordinate
with conventional overcurrent protection on the transmission line 12. The second
timer 230 detects whether the conventional overcurrent protection is not operating
as intended when it continues to operate beyond a reasonably period of time.
For instance, depending upon the conventional protection reset times
used by a utility, certain patterns of load current may cause the conventional
protection to start through its sequence toward lockout. If the fault experiences
periods of inactivity which are sufficiently long to cause the conventional protection
to reset itself, the conventional protection must start again from the beginning
of its sequence when the fault current returns to a high level. If this situation
occurs repetitively, the conventional breaker may never reach a lockout state.
In severe cases, such repetitive operation of a load breaking device,
such as a breaker, can cause extensive, and perhaps even catastrophic, damage to
the device. This continual resetting phenomenon may be caused by improperly coordinated
trip settings with respect to downstream devices, or it may be caused by the intermittent
nature of a downed conductor, high impedance fault. For either cause, the second
timer initiates a conventional protection malfunction alarm 232 when such conditions
After setting timers 226 and 230, the analyzer 10 waits for the first
timer 226 to time out. While waiting, a second event checking device or routine
234 monitors the OCF signal 210 for overcurrent conditions, the ROC signal 212
for high rates of change in the phase and/or residual currents, the BRKR signal
216 for breaker openings, and a three phase event flag set (3&phis;E) signal 236
generated by the flag setting device 206. The 3&phis;E signal 236 is generated
when the flag setting routine 206 determines the occurrence of a multi-phase fault
event. When routine 234 detects the occurrence of one or more of these four events
(OCF, ROC, BRKR, or 3&phis;E) a YES signal 238 is generated.
When a second timer monitoring device or routine 240 receives the
YES signal 238, it checks to see whether the count of the second timer 230 has
expired. If the second timer count has indeed expired, the monitoring device 240
issues a YES signal 242 to a second event output alarm 244. If not, the monitoring
device 240 issues a NO signal 246. A first timer reset is generated by a NO signal
246 from the second timer monitoring device 240. Upon receiving the either NO
signal 245 or 246, the set TIMER1 routine 224 resets the first timer 226.
If the second event checking routine 234 detects none of the four
events (OCF, ROC, BRKR, or 3&phis;E), then a NO signal 248 is generated. When a
first timer monitoring device or routine 250 receives the NO signal 248, it checks
to see whether the count for the first timer 226 has expired. If the count of timer
226 has not expired, the monitoring device 250 generates a NO signal 252 which
is returned to initialize the second event checking routine 234. If the monitoring
device 250 determines the first timer count has expired, it generates a YES signal
254. When a set TIMER3 initialization routine or device 256 receives the YES signal
254, a third timer 258 is initiated.
Referring also to FIG. 5, while the third timer 258 counts, the proper
output to be sent by the controller 35 to the peripheral device 70 or the alarm
78 is determined. Specifically, a downed conductor checking routine or device
260 receives a TIMER3 initiation signal 261 from initialization device 256. Upon
receiving signal 261, the device 260 checks if and when the LOL signal 208, the
ARC signal 214, and the OCF signal 210 indicated the occurrence of a significant
loss of load, arcing, and overcurrent conditions, respectively. If arcing is present
on the transmission line 12 as indicated by the ARC signal 214, and a significant
loss of load indicated by the LOL signal 208 initially caused the analyzer 10 to
enter a triggered state (abbreviated as "LOLI" in Fig. 5) then the checking routine
260 issues a YES signal 262.
Upon receiving the YES signal 262, a downed conductor output device
264 provides an output to the peripheral device 70. The checking routine 260 also
monitors for arcing by checking the ARC signal 214. Routine 260 monitors for the
combination of this arcing and an overcurrent condition which occurred during the
triggered state. This is the same overcurrent condition which occurs when the
second event checking routine 234 receives the OCF signal 210. When these two
conditions are encountered, the checking routine 260 issues the YES signal 262
to the downed conductor output device 264.
If the checking routine 260 does not find that the LOL signal 208
initiated the triggered state ("LOLI"), or does not find that the OCF 210 was
monitored by routine 234, and if arcing is still present, routine 260 issues a
NO signal 266. When an arc checking routine 268 receives the NO signal 266 and
the arc signal 214 still indicates the presence of an arc, the checking routine
268 issues a YES signal 270. When an arcing detected output device 272 receives
the YES signal 270, an arcing detected output is provided, for example to the
peripheral device 70. Also, when the output device 272 receives the YES signal
270, a continuation signal 274 is supplied to permit the checking routine 260
to continue to monitor for a downed conductor condition.
If the arcing condition checking routine 268 determines that the
arc signal 214 is absent, a NO signal 275 is generated. When a third timer monitoring
device or routine 276 receives the NO signal 275, it checks to see whether the
count of the third timer 258 has expired. If not, the monitoring device 276 issues
a NO signal 278 which is provided as a continuation signal to the downed conductor
checking routine 260. If the monitoring routine 276 determines that the count
of the third timer 258 has indeed expired, a YES signal 280 is generated and delivered
to the initial line status checking routine 218 to return the load analyzer 10
to its "normal" state.
The condition of the breaker controlled by the breaker trip circuit
66, as indicated by the BRKR signal 216, is monitored by a breaker monitoring device
282 for monitoring whether the breaker is closed (conducting state) or open (non-conducting
state). When the monitoring device 282 determines the breaker is closed, it issues
two YES signals 284 and 286. The first YES signal 284 is provided to the set TIMER3
device 256 to reset the third timer 258. The second YES signal 286 is supplied
to a stop TIMER2 device 288. The stop TIMER2 device 288 suspends the count of the
second timer 230 but does not reset the timer whenever the breaker is closed to
Having illustrated and described the principles of our invention
with respect to a preferred embodiment, it should be apparent to those skilled
in the art that our invention may be modified in arrangement and detail without
departing from such principles. For example, while the illustrated embodiment
has been implemented in software, or discussed in terms of devices in some instances
structural equivalents of the various hardware components and devices may be substituted
as known to those skilled in the art to perform the same functions. Furthermore,
while various hardware devices, such as the transducer and microcomputer are illustrated,
it is apparent that other devices known to be interchangeable by those skilled
in the art may be substituted. We claim all such modifications falling within the
scope of the following claims.