Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
L7
1 - 20-LC-1496
AUTO~TIC GROUND FAULT P~OTECTION
FOR AN ELECTRIC POWER SYSTEM
Background of the Invention
This invention relates generally to an electric
power system in which a variable amount of electric power is
supplied to an electric load circuit from a controllable
source of power, and it relates more particularly to
improved means for protecting such a system in the event of
abnormally high magnitudes of ground leakage current in the
system.
The invention is described in the context of a
propulsion system for a large self-propelled traction
vehicle, such as a locomotive, wherein a thermal prime mover
~typically a 16-cylinder turbocharged diesel engine) is used
to drive an electrical transmission comprising generating
means for supplying electric current to a plurality of
direct current (d-c) traction motors whose rotors are
drivingly coupled through speed-reducing gearing to the
respective axle-wheel sets of the vehicle. The generating
means typically comprises a main 3-phase traction alternator
whose rotor is mechanically coupled to the output shaft of
the engine. When excitation current is supplied to field
windings on the rotating rotor, alternating voltages are
generated in the 3-phase stator windings of the alternator.
These voltages are rectified and applied to the armature
and/or field windings of the traction motors.
During the "motoring" or propulsion mode of
operation, a locomotive diesel engine tends to deliver
constant power, depending on throttle setting and ambient
conditions, regardless of locomotive speed. Historicall~,
locomotive control systems have been designed so that the
operator can select the desired level of kraction power, in
discrete steps between zero and ma~imum, and so that the
engine develops whatever level of power the traction and
auxiliary loads demand.
'7 20LC 1'196
--2--
Engine horsepower is proportional to the produc~ o~ the
angular velocity at which the cranksh~ft turns and the torque
opposing such motion. For the purpose of varying and regulating
the amount of available po~er, it is CommGn practice to equip a
locomotive engine with a speed regulating governor which adjusts
the quantity of pressurized diesel fuel (i.e., fuel oil) injected
into each of the engine cylinders so that the actual speed (RPM)
of the crankshaft corresponds to a desired speed. The desired
speed is set, within permissible limits, by a manually operated
lever or handle of a thro~tle that can be selectively moved in
eight steps or "notches" between a low power position (Nl) and a
maximum power position (N8). The throttle handle is part of the
control console located in the operator's cab of the locomotive.
The position of the throttle handle determines the engine speed
setting of the governor.
For each of its eight different speed settings, the engine
is capable of developing a corresponding constant amount of
horsepower (assuming maximum output torque). When the throttle
notch g is selected, maximum speed (e.g., 1,050 rnm) and maximum
rated gross horsepower (e.g., 4,000) are realized. Under normal
conditions the engine power at each notch equals the power
demanded by the electric propulsion system which is supplied by
the engine-driven main alternator plus power consumed by certain
electrically and mechanically driven auxiliary equipments.
~5 The output power (KYA) of the main alternator is .
proportional to the product of the rms magnitudes of generated
voltage and load current. The voltage magnitude varies with the
rotational speed of the engine, and it is also a function of the
amount of current in the alternator armature and field windings,
respectively. For the purpose of accurately controlling and
regulating its power output, it ~s common practice to adjust the
field strength of the traction alternator to compensate for load
changes and to minimize the error between actual and desired K~.
The desired power depends on the specific speed setting of the
. .
20LC 1496
-3--
engine. Such excitation control will establish a balanced
steady-state condition which results in a substantially constant,
optimum electrical power output for each position of the throttle
handle. The alternator output regulating function is performed
by an associated controller which is responsive to the throttle
position and to a plurality of feedback signals representative,
respectively, OT certain parameters or quantities (such as the
magnitudes of the alternator output voltage and current) of the
electric propulsion system.
In an electric propulsion system, all of the power
components (alternator, recti~ier, traction motors, and their
interconnecting contactors and cables) need to be well insulated
to avoid harmful shcrt circuits between the electrically
energized parts of these components and ground. The insulation
has to withstand very harsh conditions on a locomotive, including
constant vibration, frequent mechanical shocks, infrequent
maintenance, occasional electrical overloads3 a wide range of
ambient temperatures, and an atmosphere that can be very wet
and/or dirty If the insulation of a component were damaged, or
if its dielectric strength deteriorates, or if moisture or an
accumulation of dirt were to provide a relatively low resistance
path through or on the surface of the insulation, then
undesirably high leakage current can flow between the component
and the locomotive frame which is at ground potential. Such an
~5 insulation b.eakdown can be accompanied by ionization discharges
or flashovers. The discharge will start before the voltage level
reaches its ultimate breakdown value. The dirtier and wetter the
insulation, the lower the discharge starting voltage relative to
the actual breakdown value. Without proper detection and timely
protection, there is a real danger that an initially harmless
electrical discharge will soon grow or propagate to an extent
that causes serious or irreparable damage to the insulation
system and possibly to the eq~ipment itsel~.
20I,C 1D,96
3L7
--4--
It is convent onal practice to provide ground fault
protection Tor locornotive propulsion systems. In a typical prior
art practice, the operating coil of a ground relay is connected
between the lncomotive ~rame (ground) and a point between the
armature and series field windings of one of the d-c traction
motors of the propulsion system. This is the only point of the
system that is intentionally grounded, and normally the ground
leakage currer,t in the relay coil has a negligible magnitude.
However, in the event of a ground fault, the leakage current
magnitude increases above the "pickup" point (e.g., 0.~5 ampere)
of the relay, whereupon the ground relay initiates the opening of
a contactor in series with the alternator field and thereby shuts
down the electrical propulsion system. At the same time, an
alarm bell is sounded and an appropriate light on an annunciator
1~ is turned on. The locomotive operator can then manually reset
the propulsion system and restore traction power. The reset
mechanism is arranged to lockout after three tries. This prior
art ground relay is sensitive enough to respond to any
potentially harmful degradation of the insulation system. But a
~o propulsion outage due to ground relay action may sometimes be
unnecessary, as when the increase in leakage current is due
primarily to moisture in the insulation system, and any such
outage will undesirably reduce the productivity of the
locomotive.
Summary of the Inver,tion
A general objective of the present invention is to provide
improved means for protecting an electric power system in
automatic response to the detection of actual or incipient ground
faults.
A more specific objective is the provision, for an electric
power system including means for varying the amount of power
supplied from a controllable source to an electric load circuit~
of ground leakage current responsive means that automatically
initiates a series of power limiting and restoring measures to
20L,C 1496
--5--
protect the power components of the system from serious damage
due to ground faults without unnecessarily disrupting normal
operation thereof.
In carrying out the invention in one form, electric power is
supplied to an electric load circuit from a suitable source of
power, and the amount of power is varied as a function o~ the
value of a variable control signal that is provided by associated
cGntrol means. Normally the control signal value is determined
by a give~ command signal in combination with other selected
input signals to the control means. A representative ~eedback
signal is derived from ground leakage current in either the power
source or the load circuit. The control means includes ground
fault responsive means activated when the feedback signal
i~dica~es that the magnitude of leakage current is abnormally
high to mndify the value cf the control signal in the following
manner:
(1) If the leakage current rises to a magnitude higher than
a predetermined deration threshold level but not higher than
a predetermined maximum permissible limit~ the control
signal is limited so that the power output of the source is
reduced to a fraction of its normally desired amount (which
fraction is inversely proportional to the leakage current
magnitude in excess of the threshold level), and
(2) If the leakage current magnitude rises above the
~S maximum limit9 (a) the control signal is limited so that the
power output is restricted to zero for at least a
predetermined time interval (e.g., 15 seconds) and (b) at
the e~d of that interval the zero-power restriction is
automatically removed if the leakage current magnitude is
then below a certain reset point (which is appreciably lower
than the maximum limit).
The ground fault responsive means is so arranged that it will not
automatically remove the zero-power restriction after the leakage
current magnitude has remained continuously above the aforesaid
20LC 1'196
-6--
reset point for a predetermined span of time after increasing
above thP maximum limit, or after the ground fault responsive
means has a history of being repeatedly activated "n" different
times within a predetermined period (e.g., 30 minutes)
immediately preceding the time at which leakage current again
increases above the maximum limit. In either case, a "permanent"
rather than a temporary fault in the ground insulation is
assumed, and the electric power system remains shut down until an
authorized maintainer finds and corrects the problem and then
manually resets the ground fault protection means. But a
temporary ground faul~ (which t~pically is caused by excesslve
moisture) is allowed to cure itself (as by drying out so that the
normal dielectric strength of the insulating medium is restored)
during the time power is fractionally reduced, or during the
short interval (e.g., 15 seconds) of zero power before this
restriction is automatically removed, thereby avoiding an
unnecessary or prolonged loss of power.
The invention will be better understood and its various
objects and advantages will be more fully appreciated from the
following description taken in conjunction with the accompanying
drawings.
Brief_Description of the Drawings
Fig. 1 is a schematic diagram of an electrical propulsion
system for a traction vehicle, including a thermal prime mover
~5 ~such as a diesel engine), a traction alternatora a plurality of
traction motors, and a controller;
Fig. 2 is an expanded block diagram of the controller (shown
as a single block in Fig. 1) which produces output signals for
controlling the field excitation of the alternator and the
rotational speed of the engine;
Fig. 3 is a diagram of an "equivalent circuit" that is used
~o illustrate the manner in which the controller normally
produces the alternator field excitation control signal and also
20~,C 1496
-7-
to illustrate its significan~ inter~aces with the systern ground
fault protection means of the present invention,
Fig. 4 is a flow chart that explains the preferred manner of
providing automatic ground fault protec~ion in accordance with
the present invention; and
Figs. 5, 6, 7 and 8 are flow charts that explain the
operations of the preferred embodiments of the four subroutines
that are shown as single steps in Fig. 4.
Descr.ption of the Preferred Embodiment
The propulsion system shown in Fig. 1 includes a
variable-speed prime mover l1 mechanically coupled to the rotor
of a dynamoelectric machine 12 comprisins a 3-phase alternating
current (a-c) synchronous generator, also refPrred to as the main
traction alternator. The main alternator has a set o~ three
star-connected armature windings on its stator. In operation it
generates 3-phase voltages in these windings, which voltages are
applied to a-c input terminals of at least one 3-phase,
double-way uncontrolled power rectifier bridge 13. The rectified
electric power output of the bridge 13 is supplied, via a d-c bus
14 and individual contactors (15C, 16C), to an electric load
circuit comprising parallel-connected armature windings of a
plurality of variable speed d-c traction motors, only two of
which (15,16) are identified in Fig. 1. The described power
components 11-16 are all located on board a self-propelled
~5 traction vehicle such as a locomotive. In practice each traction
motor is hung on a different axle of the locomotive, and its
shaft is coupled to the associated axle by speed-reduction
gearing (not shown). There are usually two or three axles per
truck, and there are two trucks per locomotive.
The traction motors have non-rotating ~ield windings (not
shown) that are respectively connected in series with the
windings on their rotatable armatures during the motoring or
propulsion mode of operation. However, for braking or retarding
the locomotive the armature windings of the traction motors are
20LC 1~96
--8--
disconnected from the power rectifier 13 and connected to a
conventional dynamic braking resistor grid (not shown), and the
motor field windings are rPconnected in series with each other
for energlzation by the rectified output of the main alternator
12. (Alternatively, a-c traction motors could be used, in which
case suitably controlled electric power inverters would be
connected between the respective motors and the d-c bus 14.)
Field windings 12F on the rotor ~f the main alternator 12
are connected for energization through a contactor 12C to the
output of a suitable source 17 of regulated excitation current.
Preferably the so~rce 17 comprises a 3-phase controlled rectifier
bridge the input terminals 18 of which receive alternating
voltages from a prime mover-driven auxiliary alternator that can
actually comprise an auxiliary set of 3-phase armature windings
on the same frame as the main alternator 12. This source
includes conventional means for varying the magnitude of the
direct current that it supplies to the alternator field as
necessary to minimize any magnitude difference between a variable
control signal on ~n input line 1~ and a feedback signal which
~o during motoring is representative of the average magnitude V of
the output voltage of the power rectifier 13. The latter voltage
magnitude is a known function of the magnitude of excitation
current in the field windings 12F and of the magnitude of output
current in the armature windings of the main alternator 12,
~5 respectively, and it also varies with the speed of the prime
mover 11. It is sensed by a conventional voltage sensing module
connected across the d~c output terminals of the po~ler rectifier.
A current detecting module 22 of relatively low resistance
(e.g., approximately 125 ohms) is connected between the neutral S
of the alternator stator windings and the grounded chassis or
frame of the locomotive, and it provides on an output line 23 a
feedback signal representative of the magnitude (IGND) of ground
leakage current in the electric propulsion system. It will be
apparent that IGN~ is a measure of current flowing, via the
20LC 1~96
_g_
module 22, between the neutral S and any ground fault in the
stator windings of the main alternator 12, in the po~er rectifier
13, or in the electric load circuit that is connected to the
power rectifier. The latter circuit includes the field windings
of the traction motors 1~, 16, etc. and, in the motoring mode of
operation, the motor armature windings as well.
The prime mover 11 that drives the alternator field 12F is a
thermal or internal-combustion engine or equivalent. On a
diesel-electric locomotive, the motive power is typically
provided by a high-horsepower9 turbocharged, 4-stroke,
16-cylinder diesel engine. Such an engine has a number of
ancillary systems, some of which are represented by labeled
blocks in Fig. 1. A diesel engine fuel system 24 conventionally
includes a fuel tank, fuel pumps and nozzles for injecting fuel
oil into the respective power cylinders which are arranged in two
rows or banks on opposite sides of the engine, tappet rods
cooperating with fuel cams on a pair of camshafts for actuating
the respective injectors at the proper times during each full
turn of the crankshaft, and a pair o~ fuel pump racks for
controlling how much fuel oil flows into a cylinder each time the
associated injector is actuated. The position of each fuel p~mp
rack, and hence the quantity of fuel that is being supplied to
the engine, is controlled by an output piston o, an engine speed
governor system 25 to which both racks are linked. The governor
2~ regulates engine speed by automaticall~ displacing the racks,
within predetermined limits, in a direction and by an amount that
minimizes any difference between actual and desired speeds of the
engine crankshaft. The desired speed is set by a variable speed
control signal received from an associated controller 26, which
signal is herein called the speed command signal or the speed
call signal. An engine speed signal RPM indicates the actual
rotational speed of the engine crankshaft and hence of the
alternator field.
~OLC 1'196
~ L~ 7
The speed command signal for the engine governor system 25
and the excitation control signal for the alternator field
regulator 17 are provided by the controller 26. In a normal
motoring or propulsion ~ode of operation, the values of these
signals are determined by the value of a command signal that is
givell to the controller by a manually operated throttle 27 to
which the controller is coupled. A locomotive throttle
conventionally has eight power positions or notches (N), plus
idle and shutdown. Nl corresponds to a minimum desired engine
1~ speed (power), while N8 corresponds to maximum speed and full
power. When dynamic braking of a moving locomotive is desired,
~he operator moves the throttle handle to its idle position and
manipulates a manually operated lever of a conventional brake
controller 28 so that the main controller 26 is now supplied with
1~ a variable "brake call" signal that will determine the value of
the alternator excitation control signal. (In the braking mode,
a feedback sigr,a1 which is representative of the magnitude of the
current being supplied to the motor field windings from the
rectified output of the main alternator 12 will be supplied to
~0 the alternator field regulator 17 and there subtracted from the
control signal on line 19 to determine the difference or error
signal to which the regulator responds.) In a consist of two or
more locomotives, only the lead unit is usually attended, and the
main controller on board each trail unit will receive, over
trainlines, encoded signals that indicate the throttle position
or brake call selected by the operator in the lead unit.
For each power level of the engine there is a corresponding
desired load. The controller 26 is suitably arranged to
translate the notch information from the throttle 27 into a
control signal of appropriate magnitude on the input line 19 of
the alternator field regulator 17, whereby in motoring the
traction power is regulated to match the called-for power so long
as the alternator output voltage and load current are both within
predetermined limits. For this purpose, and for the purpose of
L'7 2 0 L~C 14 9 6
deration (i.e., unloading the engine) in the event of certain
abnormal conditions3 i~ is necessary to supply the controller 26
with information about various operating conditions and
parameters of the propulsion system.
More particularly, the controller 26 typically receives the
above-mentioned engine speed signal RPM, the voltage feedback
signal V, and current feedback signals Il, I2, etc. which are
representative, respectively, of the magnitude of current in the
armature windings of the individual traction motors. It also
receives a load control signal issued by the governor system 25
if the engine cannot develop the power demanded and sti?l
maintain the called-for speed. (The load control signal is
effective, when issued, to reduce the magnitude of the control
signal on the line 19 so as to weaken the alternator field until
1~ a new balance point is reached.) As is illustrated in Fig. 1,
the controller is supplied with additional data including: "VOLT
MAX" and "CUR MAX" data that establish absolute maximum limits
for the alternator output voltage and current, respectively;
"CRANK" data indicating whether or not an engine starting (i.e.,
cranking) rou~ine is being executed; and relevant inputs from
other selected sources, as represented by the block labeled
"CTHER." The alternator field regulator 17 communicates with the
controller via a multiline serial data link or bus 21. The
controller 26 also communicates with "CONTACTOR DRIVERS" (block
~5 29) which are suitably constructed and arranged to actuate the
alternator field contactor 12C and the individual traction motor
contactors l~C, 16C, etc. in accordance with commands from the
controller.
For the purpose of responding to ground faults in the
propu7sion system, the controller 26 is supplied, via the output
line 23 of the current detecting module 22, with the aforesaid
feedback signal whose value varies with the magnitude IGND of
ground leakage currentO If this signal indicates that IGND is
abnormally high~ the controller executes certain protective
20LC 1~6
~ 7
functions that will soon be described, and at the same time it
sends appropriate messages or alarm sîgnals to a display module
30 in the cab of the locomotive.
In the presently pre~erred embodiment of the inYention, the
controller 26 comprises a microcomputer. Persons skilled in the
art will understand that a microcomputer is actually a
coordinated system of commercially available components and
associa~ed electrical circuits and elements that can be
prograrruned to perform a variety of desired functions. In a
typical microcomputer, which is illustrated in Fig. 2, a central
processing unit ~CP") executPs an operating program stored in an
erasable and electrically reprogrammable read only memory (EPROM)
which also stores ~ables and data utilized in the program~
Contained within the CPU are conventional counters, registers,
accumulators, flip flops (f1ags), etc., along with a precision
oscillator which provides a high-~requency clock signal. The
microcomputer also includes a random access memory (RAM) into
which data may be temporarily stored and from which data may be
read at various address locations deterrnined by the program
~0 stored in the EPROM. These components are interconnected by
appropriate address, data, and control buses. In one practical
embodiment of the invention, an Intel 8086 microprocessor is
used.
' The other blocks shown in Fig. 2 represent conventional
peripheral and interface components that interconnect the
microcomputer and the external circuits. More particularly, the
block labeled "I/O" is an input/output circui~ for supplying the
microcomputer with data representative of the selected throttle
position or the brake command and with digital signals
representatiYe of the readings of various voltage, current and
other feedback sensing modules associated with the locomotive
propulsion system. The latter signals are derived from an
analog-to-digital converter 31 connected via a conventional
multiplexer 32 to a plurality of signal conditioners to which the
- 13 - 20LC 1496
sensor outputs are respectively applied. The signal
conditioners serve the conventional dual purposes of
buffering and biasing the analog sensor output signals. As
is indicated in Fig. 2, the input/output circuit also
interconnects the microcomputer with the alternator field
regulator via the multiline bus 21, with the Qngine speed
governor, with the display module, with the contactor
drivers, and with a digital to-analog signal converter 33
whose output is connected to the line 19.
Tha controller 26 is programmed to produce, on the line
19, a control signal having a magnitude that depends on
either the throttle position selected by the locomotive
operator (in the normal motoring mode of operation) or the
brake command selected by the operator (in the dynamic
braking mode). The presently preferred manner in which this
is accomplished during motoring is described in a Canadian
Application S.N. 511,600, filed June 13, 1986, Balch et al
and is assigned to General Electric Company. As simplified
block diagram of some of its presently signiflcant functions
is shown in Fig. 3 which will now be described.
As is explained in the referenced application, the
alternator excitation control programs (reference No. 41 in
~ig. 3) include routines for providing, on two channels
labeled "PWR" and "V & I," respectively, numbers
representing reference values of traction power and of a
common voltage and current limit. Both of these values are
on a per axle basis. Normally they vary with the throttle
position, being highest at notch 8. But the normal values
we appropriately modified by a rate limit function in the
event of a step change in the call data and by a deration
function in response to a wheelslip or certain other
temporary abnormal conditions.
As is illustrated in Fig. 3, the modified power reference
value from the excitation control programs is fed to an "un-
balance correction" function 42 which is also supplied, on a
L ~ 2 0 LC 1 4 9 6
-14-
line labeled "K~A(AV)," with datum representative o~ the actual
kilowatts of power o~tput (per axle) of the traction alternator
12. (Suitable signal processing programs are included in the
block 41 for the purpose o~ deriving the latter datum from the
feedback signals V, Il, I2, etc.) The output of the unbalance
correction function represents the desired value of the
alternator output. It will di~fer ~rom the modified power
reference input when necessary to correct for any appreciable
power unbalance among the various traction motors of the
locomotive.
The modi~ied V & I reference value is fed to a "reference
limits" function 43 which is also supplied with data
representative of the maximum current and voltage limits
established by CUR-MAX and VOLT MAX, respectively. The latter
limit is hereinafter called VMAX. In the function 43 the common
Y & I reference input is deployed to provide a limited current
reference value (I) that varies with the input up to the maximum
limit of current and to provide a separate limited voltage
reference value that also varies with the input up to the limit
established by VMAX.
The limited voltage reference value is compared with the
actual value of the alternator voltage feedback signal V to
derive a voltage error value equal to their difference. This
error is then processed in accordance~ with a progral~med
~5 compensation routine to derive a voltage control value that is
representative of the voltage error value. The compensation
routine introduces a proportional plus integral transfer function
(see reference No. 44 in Fig. 3), the gain of which is determined
by datum that depends on the throttle position and other
parameters of the locomotive and its controls. Thus the voltage
control value varies as a function of the time integral of its
associated error value. Similar routines (not shown in Fig. 3)
are provided for comparing the limited current reference value
with the actual value of the motor current feedback signal I(MAX)
L3~ ~ 20LC 149 6
from the most loaded traction motor to derive a current error
value equal to their difference, and for comparing the desired
power value with the actual power demand of the most loaded
motor, as found by multiplying V by I(MAX), to derive a power
error value equal to their difference, if any. The latter two
error values are then processed in accordance with programmed
compensation routines similar to the transfer function 44 to
derive curren~ and power control values that are respectively
representative of the current and power error values. All three
of the control values are supplied to a gate 45 that selects the
least value for passing to a limit function 46 from which an
output signal VC is derived, and accordingly the value of YC
corresponds to the smallest control value. Means 47 for clamping
the value of the output signal VC to zero is prov;ded between the
limit function 46 and the digital-to-analog converter 33.
The value of VC determines the magnitude of the analog
control signal that the controller 26 supplies, via the line 19,
to the alternator field regulator 17 (Fig. 1). In the motoring
mode of operation the field regulator will respond to the latter
signal by varying the field strength of the traction alternator
as necessary to minimize any difference between the value of the
voltage feedback signal V and the value of the output signal VC.
So long as both V and I(MAX~ are within a limit that varies with
the throttle position and are not above their respective maximum
limits as imposed by the function 43, the value of VC is
determ;ned by the power control value which will now be smaller
than either the voltage or current control value. Consequently
the alternator output voltage is maintained at whatever level
results in essentially zero error between actual and desired
traction power. But if V (or IMAX) tends to exceed its limited
reference value, the voltage (or current) control value is driven
lower than the power control value and the value of VC
accordingly decreases, whereby the alternator output is adiusted
to whatever level results in zero voltage (or current) error.
20LC 1496
-16-
In accordance with the preserlt invention, the value of the
control signal on line i9 is reduced to zero, thereby restricting
the power output of the main alternator 12 to zero, by activat,ng
the clamping means 47 fcr at least a predetermined minimum
interval of time (e.g., approximate1y 15 seconds) if the
magnitude of ground lea~age current in the propulsion system
rises above a predetermined maximum permissible limit.
Ooncurrently with the start of this clamping action, contactor
opening commands are given to the contactor drivers 29, the
normal excitation contrcl programs 41 are disabled or turned off9
and the occurrence of the ground fault is logged in the display
module 30. For the purpose of detecting and responding to such a
ground fault, the controller 26 includes sround fault responsive
means which in Fig. 3 is symbolized by a block 48 labeled "GND
FAULT PROTECTION PROGRAM" and by a block 49 labeled "DISABLE EXC
PRQGRAMS, MAKE VC=O, & OPEN PWR CONTACTORS."
While a zero-power restriction is in effect, any excessive
moisture that may have been the cause of the ground fault can dry
out, in which case the fault will be self-curing. At the end of
~o the aforesaid minimum interval of time, the ground fault
protection program 48 will automatically remove the zero-power
restriction by deactivating the clamping means 47, turning on the
excitation control programs, and issuing contactor closing
commands, unless (1) IGNn did not decrease below a predetermined
reset limit !e.g., le_s than approximately 70% of the maximum
limit~ within a predetermined span of time (e.g., approximately
nine seconds) measured from the start of this interval, or (2)
the ground fault responsive means has repeatedly activated the
clamping means "n" different times within a predeter~ined period
(e.g., approximately 30 minutes) immediately preceding the time
at which IGND increases above its maximum permissible limit~
where n is a predetermined whole number (e.g., 3). If either one
of the latter conditions is true, a "permanent" or
non-self-curing ground fault is assumed, and the ground fault
~OI,C 1496
-17-
p,otection ~unction must be manually reset ~o remove the
zero-power restriction.
The ground fault protection program 48 is also efFective, i~
and when the feedback signal on the output line 23 of the ground
current detectin~ module 22 (Fiq. 1) indicates that IGND is
higher than a predetermined deration threshold level but has not
exceeded the aforesaid maximum limit, to modify the value of the
control signal in a manner that reduces the power output of the
alternator to a fraction of its normally desired amount. The
latter function is preferably accomplished by reducing VMAX if
the propulsion system is operating in its motoring mode or by
reducing the value of the brake call if the system is operating
in its dynamic braking mode, with the amount of reduction being
proportional to the magnitude of leakage current in excess of the
~eration threshold level. As was explained hereinbefore, ground
leakage current tends to increase, and the ionization discharge
starting voltage tends to decrease, as moisture increases. By
fractior,ally reducing the alternator output when the leakage
current is in a "medium" range (i.e., when IGND has increased to
an abnormally high level but is not above its maximum permissible
limit), potentially harmful discharges can be avoided or at least
minimized without a total loss of traction power, and the
alternator voltage amplitude is allowed to increase as the ground
insulation medium dehydrates and its dielectric strength
~5 gradually returns to normal.
Althouyh this ground fault protection function could be
implemented in a variety of different ways to obtain the results
summarized above9 the presently preferred way is to program the
microcomputer 26 to execute the routine tnat is illustrated in
Fig. 4. This routine is repeated once every 60 mill;seconds. It
starts with an initializing subroutine 51, the basic steps of
which are shown in Fig. 5. Ilhis subroutine begins at at inquiry
point 52 which determines whether or not the locomotive is in its
engine cranking mode. If it is, the next and final step 53 in
2 0 LC 1 'I 9 6
3~ 7
-18-
the ground fault protection routine will start a timer ~1 and
will then reset second and third timers and a pair oF counters.
~therwise the subroutine ~1 proceeds from point 52 t~ an inguiry
point 5~ where the status of the first timer is tested. So long
as this timer is active (i.e., not over), the ground fault
protection routine ends here. In effect, the ground fault
protection function is disabled while the engine 11 is being
started and thereafter for the period of time (e.g.,
approximately 10 seconds) that timer #1 is running.
Once the timer ~1 has timed out, the initializing subroutine
~1 will proceed from point 54 to another inquiry point 55 which
determines whether or not the propulsion system is operating in
its dynamic braking mode, and this is followed by a step which
presets the deration set point or threshold level and the maximum
permissible limit of ground leakage current. If the system is
not in a brake mode, the presetting step 56 will load a first
predetermined number (M1) corresponding to the desired threshold
level (e.g., 0.5 ampere) into a "Kl" register and will also load
a second predetermined number (M2) corresponding to the desired
~ maximum limit (e.g., 1.0 ampere) into a "K2" register.
Alternatively, in the dynamic brake mode only, the presetting
step ~7 will load a third predetermined number (B1) corresponding
to the desired threshold level (e.g., 0.25 ampere) into the Kl
register and will also load ar,other number (B2) corresponding to
~5 the desired maximum limit (e.g., 0.5 ampere) into the K2
register. In the example aiven, the number in the Kl register is
apprGximately one-half of the number in the K2 register.
Preferably Bl is a lower number than M1 and B2 is a lower number
than M2, whereby the sensitivity of the ground fault protection
mPans is increased when the propulsion system is switched to its
dynamic braking mode of operation. This is both desirable and
permissible because the normal amount of ground leakage current
that inherently exists in the electric power system and that is
represented by the feedback signal on line 23 will be much lower
2 ~) LC 1 ~ 9 6
-19-
during the braking mode (w~en the armature windings o~ the
traction motors are disconnected from the power rectifier) than
during the motoring mode.
After the presetting step 56 or .~7, the subroutine shown in
Fig. 5 proceeds to an inquiry point 58 where the status of a
ground counter ("GND CNTR") is tested. The operation of the
ground counter will soon be explained in connection with the
description of the grounded subroutine shown in Fig. 6. If the
count in this counter is not greater than 2, the control is
returned directly to the next step ~1 of the main ground fault
protection routine (Fig. 4). But otherwise, a step 59 is
executed to replace the number in the K2 register with a lower
number before returning to the step 61 in the main routine. This
lower number corresponds to the desired reset level of ground
leakage current. Preferably, the reset point is appreciably
lower than (e.g., approximately two-thirds of) the maximum limit.
In effect, step 59 introduces desirable "hysteresis" in the
operation of the ground fault protection means.
After completing the initializing subroutine 51, the ground
~0 fault protection routine executes the step 61 of reading and
saving the present value of the feedback signal on line 23, which
value corresponds to the magnitude (IGND) of ground leakage
current in the propulsion system. As is indicated in Fig. 4, the
routine then proceeds to an inquiry point 62 where the saved
~5 value is compared with the number stored in register K2 to
determine which one is greater. So long as the ground counter
has nat counted more than two consecutive passes through the
grounded subroutine (Fig. 6), the answer to inquiry 62 is
affirmative only if IGND is above its preset maximum permissible
limit, and thereafter the answer will continue to be affirmative
until IGND has decreased below the aforesaid reset level. If the
answer to the inquiry 62 is affirmative, the grounded subroutine
63 is called.
20LC 1496
-20-
The presently preferred embodiment of the ground2d
subroutine 63 wi11 now be described with reference to Fig. 6. It
begins by reset~ing a timer ~2 (step 64). Then, at step 65~ the
above-mentioned ground counter is automatically incremented
(i.e., whatever count is stored in a dedicated address of the
micr~computer memory is increased by 1). This is followed by
testing, at 67, the status of the ground counter. If the count
is 1, the grounded subroutine is aborted here; otherwise it
proceeds to another inquiry point 69 where the status of the
ground counter is retested to determine whether or not the count
is 2. If the answer is affirmative, a ground f2ult is assumed
and the grounded subroutine responds by executing a series of
steps 71-79. As ~ill soon be apparent, the count in the ground
counter will never reach 2 unless IGND is above the preset
maximum limit (K2) for two consecutive passes through the
grounded subroutine, and thus the steps 71-79 are delayed until
IGND has remained above this limit for 60 milliseconds after this
condition is initially detected. In effect, the grounded
subroutine ignores the first time IGND rises above the preset
maximum limit, and the rest of this subroutine will not be
executed if the initial indication of a ground fault were
actually caused by a harmless electrical transient ("noise") that
subsides in less than 60 milliseconds.
If the count in the grGund counter is 2, a temporary ground
fault counter ("TEMP FAULT C~TR") is automatically incremented by
1 (step 71), and then its status is tested at 72. So long as
the count in the temporary ground fault counter is not greater
than a predetermined whole number "n" (typically n=3)~ the answer
to the inquiry 72 is negative and the grounded subroutine
proceeds to another inquiry point 73 where the status of a timer
#3 is checked. If this timer is active (i.e., not over), the
subroutine proceeds directly from point 73 to the next step 74,
but otherwise the timer #3 is started (at step 75) before step 74
is executed. Timer ~3 is set to be active (i.e., to cont;nue
20LC 1496
-21-
running) ~or ~ predetermined period o~ ti~e "T" after being
started, and in a typical application of the invention T is
approYimately 30 minutes. Step 7~ will restrict the power output
of the main alternator 12 to zero by activating the
above-described function shown in block ~9 of Fig. .~, and in
addition it will enter a "~ILL PWR" message in the display module
30. This i5 followed by a step 76 which starts a timer #4 and
se~s a "ground" flag in its true state. Timer #4 is the
automatic reset timer; it is typically set to be active for an
interval of approximately 15 seconds after beins started. From
step 76 of the grounded subroutine 63, the control is returned to
the main routine which ends at this point.
As is indicated in Fig. 6, if the count stored in the
temporary ground fault counter were equal to n (e.g. 9 3) just
before the grounded subroutine is executed, step 71 will increase
the count to n ~ 1 (e.g., 4), and consequently the answer to the
inquiry 72 will now be affirmative rather than negative. In this
event, a "permanent" ground fault is assumed, and the subroutine
will proceed from the inquiry pcint 72 to a step 77 which resets
~0 timers ~2 and ~3 and also resets the temporary ground fault
counterO Step 77 is followed by steps 78 and 79. In step 78, a
predetermined number "N" is loaded into the ground counter,
thereby freezing its count at this number. Step 79 will restrict
the alternator power output to zero in the same manner as step
74, and in addition it will enter a "WONT LOAD" message in the
display module 30. From step 79 of the grounded subroutine 63,
the control is returned to the main routine which ends at this
point.
If the answer to inquiry 62 in the main routine (Fig. 4)
remains affirmative for 120 milliseconds or longer, the grounded
subroutine (Fig. 6) will be executed more than two consecutive
times. On the third consecutive pass through this subroutine,
the answer to inquiry 69 will be negative, and the subroutine
will then proceed from the inquiry point ~9 to yet another
20LC 1496
-22-
inquiry point 81 where the status of the ground counter is
retested tO determine whether or not the count is greater than N.
If not, the control is returned to the main routine which ends at
this point. Thereafter, the inquiry 81 is repeatedly e~ecuted at
60-millisecond intervals until IGND decreases below its reset
level and inquiry 62 yields a negative answer. However, once the
grounded subroutine 63 has been executed N consecutive times, on
the next pass through this subroutine the answer to inquiry 81
will be affirmative. In th,s event, a "permanent" ground fault
is assumed, and steps 77-79 are executed before er,ding the ground
fault protection routine. In a typical application of the
invention, N is 150 which is reached in a time span of nine
seconds. This length of time is appreciably shorter than the
aforesaid 15-second reset interval.
1~ Returning now to the description of the ground fault
protection routine shown in Fig. 4, it will be apparent that the
grounded subroutine 63 is not called (1) before IGND increases
above its predetermined maximum permissible limit or (2) after
IGND has exceeded this limit for at least 60 ms. and then
decreases below its predetermined reset level. In either case,
the answer to inquiry 62 is negative, and the main routine will
proceed from point 62 to an inquiry point 82 which checks the
state of the ground flag. If the ground flag is not in its true
state, the main routine will proceed directly from point 82 to
~5 the next step 83. Otherwise it proceeds to the step 83 via an
inquiry point 84 where the status of the automatic reset timer #4
is checked. So long as this timer is active, the routine wil?
proceed directly from point 84 to the next step 83. But once the
reset timer #4 times out at the end of the aforesaid 15-second
interval, an additional step 85 is executed before step 830 In
step 85 the power restrictions that were imposed by step 74 of
the grounded subroutine 63 are automatically reversed or removed,
the "KILL PWR" message in the display module is cancelled or
reset, and the ground flag is set in a false state. This permits
20LC 1496
-23-
the power output of the maln alternator I2 to be restored to
whatever level i5 determined by the normal operation of the
excitation control programs 41 (Fjg 3).
In step 83 of the main rout,ne, the count in the ground
counter is reset to zero. This is Followed by an inquiry point
86 where the status of the temporary ground fault counter is
tested to determine whether or not its count is zero. If the
answer is affirmative, the routine proceeds directly from point
86 to another inquiry point 87, but otherwise a ground forget
subroutine 88 is called before executing the inquiry 87. The
ground forget subroutine is shown in Fig. 8 which will be
described later. As is indicated in Fig. 4, the inquiry point 87
compares the saved value of the leakage current feedback signal
with the number stored in register K1 to determine which one is
greater. The answer to inquiry 87 is affirmative if IGND is
higher than its preset deration threshold level, in which case a
deration subroutine 89 is called.
The presently preferred embodiment of the deration
subro~tine 89 will now be described with reference to Fig. 7. It
begins at an inquiry point 90 which determines whether or not the
propulsion system is operating in a dynamic braking mode. If
not, the maximum voltage limit VMAX is fetched (step 91) and then
reduced by a step 9~ which calculates the product o, VMAX and a
fraction equal ~o one minus the ratio of actual to maximum
~5 deviations of IGND from the deration threshold 1evel. The actual
deviation corresponds to the saved value of the leakage current
feedback signal less the number stored in register Kl, and ~he
maximum deviation corresponds to the known difference between
these quantities if IGND were to increase to a predetermined
magnitude at which 100% deration (i.e.g VC = 0) is desired. The
latter magnitude can be approximately the same as, but preferably
is slightly more than, the previously mentioned maximum
permissible l;mit of ground leakage curren~ when the propulsion
20LC 1496
-24-
system is ~otoring. The produc~ of VMAX and the aforesaid
fractjon is saved ~or the excitation control program.
Alternatively, ,f the inquiry point 90 reveals that the
propulsion system is operating in a braking mode, the value of
the brake call signal is fetched by a step 93 of the deration
subroutine. Then, at a step 94, it is reduced by calculating the
product of the fetched value and a fraction equal to one minus
the ratio of actual to maximum d~viations of IGND from the
predetermined deration threshold level of ground leakage current
during dynamic braking~ The product of the brake call value and
this fraction is saved for the excitation control prGgram. After
executing either step 92 or step 94, the deration subroutine 89
returns to a final step 95 of the main ground fault protection
routine (Fig. 4) where a "LOAD LIMITED" message is entered in the
lS display module 30 and a "derate" flag is set. Once the derate
flag is set, the excitation control programs (41) will use the
reduced value of VMAX that was saved at step 92 of the deration
subroutine (or, in the dynamic brake mode, the reduced brake call
value that was saved at step 94).
~O So long as IGND is not higher than its predetermined
deration threshcld level, the answer to inquiry 87 will be
negative. In this event, as is shown in Fig. 4, the final step
97 of the ground fault protection routine is to clear the derate
flag. Now the excitation control programs (41) will use the
normal V~lAX and brake call values.
Anytime the count in the temporary ground fault counter is
greater than zero when the aforesaid inquiry 86 is executed, the
ground fault protection routine will proceed from inquiry point
86 to inquiry point 87 via the ground forget subroutine 88 which
will now be described. Fig. 8 illustrates the presently
preferred embodiment of this subroutine. It begins by checking~
at an inquiry point 98, the status of timer #3. If this timer is
active (indicating that IGND has increased above its maximum
permissible limit and step 71 of the grounded subroutine 63 has
3L~l7
20LC 1496
-25-
increm~nted the temporary ground fault counter at leas~ once
during the immediately preceding 30-minute period), the control
returns directly from point 98 of the ground forget subroutine to
the inquiry point 87 of the main ground fault protection routine
(Fig. 4). But once 30 minutes has elapsed since the last time
the timer #3 ~as started, the answer to inquiry 98 is affirmative
and the ground forget subroutine will proceed to another inquiry
point 99 where the status of timer 7r2 iS checked.
Noting that timer rr2 iS reset by step 64 each time the
grounded subroutine 63 is executed, it will be apparent that this
timer is initially inactive (i.e., over), and consequently the
ground forget subroutine shown in Fig. 8 proceeds from point 99
to a step 100 where the temporary ground fault counter is
automatically decremented by 1. Step 100 is followed by a step
101 which starts timer 7r2~ and the ground forget subroutine then
returns to the main ground fault protection routine. Timer #2 is
set to be active, after being started, for a predetermined
interval of time equal to approximately Tn (e.g., approximately 10
minutes). Once this timer is active, the ground forget
~0 subroutine 88 will return directly from inquiry point 99 to the
main ground fault protection routine, and its counter-
decrementing step 100 will not again be executed until the end of
the last-mentioned 10-minute interval (assuming no new ground
faults occur during this interval). In this manner, the
~5 temporary ground fault counter automatically forgets or loses one
count if there is no ground fault when timer #3 indicates the end
of the 30-minute period, and so long as timer #3 is not
restarted, this counter will thereafter forget one more count at
the end of each successive 10-minute interval until the count
therein is reduced to zero. It will now be apparent that the
temporary ground fault counter will accumulate a count of 4 if a
ground fault recurs four times during any 30-minute period, or if
ground faults are detected three different times during a
30-minute period, a fourth one is detected during the next 10
~2~ 7
-26-
minutes, and a fifth one occurs within 30 minutes after the
fourth.
While a preferred embodiment of the invention has been shown
and described by way of example, many modifications will
undoubtedly occur to persons skilled in the art. The concluding
claims are therefore intended to cover all such modifications as
fall within the true spirit and scope of the invention.
