Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
20~(~7C~
Description
Synchronous Wheel Slip strateqy
For A Locomotive Governor
Technical ~ield
The present invention relates to a system
for detecting and controlling locomotive wheel slip
and, more particularly, to a system for detecting and
controlling synchronous wheel slip in a locomotive
having at least one electric traction motor powered by
an engine-generator unit controlled by a field current
controller.
Backqround Art
In a typical modern locomotive, a diesel
engine is used to provide mechanical energy to a
electric generator. The generator converts this
mechanical energy into electrical power which is used
to operate a plurality of direct current (dc) traction
motors, each driving a separate drive axle having a
pair of drive wheels connected thereto.
Wheel slip usually occurs during
acceleration, and can take two forms. The first
type of wheel slip is referred to as differential
wheel slip which occurs when at least one set of drive
wheels maintains tractive contact with the rail while
at least one set of the remaining drive wheels slip.
A second type of wheel slip is synchronous slip which
occurs when none of the drive wheels maintains
tractive contact with the rail and all of the drive
wheels slip more or less simultaneously.
Wheel slip has long been a problem in
locomotives and many systems have been developed which
either reduce or completely eliminate wheel slip. A
common scheme is to compare speed signals from driven
~0:~7~4
and idler wheels or speed signals from each of several
driven wheels or highest and lowest speed signals from
traction motors. A slip condition is presumed to
exist if the compared speed signals differ by more
than a preselected magnitude~ In the above mentioned
systems, sensors, such as speed transducers, are used
to produce the speed signals and such sensors add
extra costs to wheel slip control systems.
Furthermore, such systems are usually directed towards
controlling only differential wheel slip and, as
mentioned above, synchronous wheel slip also degrades
locomotive performance.
A system for limiting both synchronous and
differential wheel slip is disclosed in U.S. patent
number 4,463,289 which issued on July 31, 1984 to
Young. seginning on line 35 of column 4, Young
discusses a rate circuit for controlling synchronous
wheel slip. The rate CirCUit provides a signal
representing the rate of acceleration of the wheel
with the greatest velocity. If the measured wheel
acceleration is greater than a predetermined value, a
rate signal is applied to the generator exciter which
reduces exciter ou~put and wheel speed. However, this
scheme still requires sensors to detect the speed of
each drive whael.
The present invention is directed towards
addressing the above mentioned problems by controlling
synchronous wheel slip in a locomotive without
requiring speed sensors. Other aspects, objects and
advantages can be obtained from a study of the
drawings, the disclosure, and the appended claims.
Disclosure Of The Invention
In accordance with one aspect of the
invention there is provided an apparatus for
-3~ 7~
controlling synchronous wheel slip of a locomotive
having at least one electric traction motor powered by
an engine-generator unit of the type having a field
current controller. A current sensor detects a
current produced by th~ engine-~enerator unit an~
produces a current signal responsive to the detected
current. A voltage sensor detects a voltage produced
by the engine~generator unit and produces a voltage
signal responsive to the detected voltage. A
processor processes the produced current and voltage
signals to produce a calculated wheel speed signal
indicative of the actual locomotive wheel speed,
processes the calculated speed signal to produce a
lagged wheel speed signal, derives a difference signal
in response to a difference between the calculated and
lagged wheel speed signals, and delivers a control
signal to the field current controller in response to
the difference signal being greater than a first
preselected reference signal.
Brief Descri~tion of The Drawinqs
Fig. 1 is a simplified block diagram of a
locomotive microprocessor governor incorporating an
embodiment of the immediate synchronous wheel slip
controller.
Fig. 2 is a graph of generator current
versus generator voltage divided by wheel speed for a
locomotive having a particular combination of
engine-generator unit and traction motor.
Fig. 3a is a graph of time versus wheel
speed illustrating a relationship between a calculated
wheel speed and a first-order-lag of the calculated
wheel speed during non-slipping acceleration.
Fig. 3b is a graph of time versus wheel
speed illustrating a relationship between a calculated
~U7~
--4--
wheel spe~d and a first-order-lag of the calculated
wheel speed during a synchronous slip condition.
Fig. 4 is a graph of engine speed versus
rack position for optimum operating efficiency of a
diesel engine.
Fig. 5 is a flowchart of certain functions
performed by an embodiment of the immediate
synchronous wheel slip controller.
Figs. 6a and 6b are flowcharts of certain
functions performed by a locomotive governor which
incorporates an embodiment of the immediate
synchronous slip controller.
Figs. 7a and 7b are flowcharts of certain
functions performed by another embodiment of the
immediate synchronous wheel slip controller.
Best Mode For Carrying out The Invention
Fig. 1 illustrates a locomotive
microprocessor governor control lo which incorporates
an embodiment of the immediate synchronous wheel slip
controller 12. The microprocessor governor control 10
is similar to one disclosed in U.S. patent number
4,498,016 issued on February 5, 1985 to Earleson et
al.; therefore, the governor control 10 will not be
described extensively herein. The microprocessor
governor control 10 is connected to an
engine-generator unit 14 which includes a diesel
engine 16 mechanically connected to drive a generator
18 by a drive shaft 20. The diesel engine 16 has a
rack controller 22 for controlling the rate of fuel
delivery to the engine 16. The generator 18 has a
field current controller 24 for controlling power
output thereof. The engine-generator unit 14 provides
electrical power to a plurality of dc traction motors
Z6a-26c for driving a plurality of drive axles and
~Q~
5--
drive wheels (not shown). The generator 18 produces
an alternating current (ac) which is passed through a
rectifier (not shown) to produce a dc current to power
the traction motors 26a-26c.
A power selector 28, typically advanced in
steps or ~'notches~ by a human operator, is used to
regulate the speed of the locomotive. In the
preferred embodiment the power selector 28 has an idle
notch and eight load notches. A selector sensor 30,
is connected to the power selector 2~ for producing a
power signal in response to the position of the power
selector 28. In the preferred embodiment, the
selector sensor 30 includes a system of electrical
switches which produce a mathematically encoded
four-bit output signal in response to the position of
the power selector 28. It is apparent to those
skilled in the art that this sensing function can also
be performed by any one of a number of devices such as
a potentiometer, a transducer, etc.
An idle sensor 32 is connected to the power
selector 28 for producing a load signal when the power
selector 28 is not in the idle position. In the
preferred embodiment the idle sensor 32 is a switch
that closes when the power selector 28 is positioned
in one of the eight load notches. It is forseeable
that the power signal could also be used for
determining when the power selector 28 is positioned
in a load notch. The load signal is applied to an
input terminal 33 of the synchronous wheel slip
controller 12.
A desired engine speed calculator 34 has an
input terminal 36 for receiving the power signal from
the selector sensor 30. The desired engine speed
calculator 34 accesses a first look-up table (not
shown), stored in a memory 38. The first look-up
-6- Z~070~
table relates each notch setting to a desired engine
speed and is used to pr~duce a desired engine speed
signal. The desired engine speed signal is applied to
one input terminal 40 of a speed summer 42.
A magnetic pick-up sensor ~4 is connected to
the drive shaft 20 and produces an actual engine speed
signal which is applied to a second input terminal 46
of the speed summer 42. It is apparent to those
skilled in the art that this function can also be
performed by a device such as a tachometer. The speed
summer 42 produces a speed error signal eN in response
to a difference between the actual and desired engine
speed signals.
A speed controller 48 has an input terminal
49 for receiving the speed error signal eN from the
speed summer 42. The speed controller 48 accesses a
fuel delivery rate formula stored in the memory 38 and
uses the formula to produce a fuel delivery rate
signal as a function of the speed error signal eN.
The fuel delivery rate siqnal is applied to an input
terminal 51 of the rack controller 22 to regulate
actual engine speed so as to reduce the engine speed
error eN signal to zero.
A current transformer 50 is connected to the
generator 18 and produces a current signal in response
to the ac generator current. In the preferred
embodiment the current signal is actually a dc voltage
signal which is proportional the the generator
current. More particularly, the current transformer
50 produces a current which is rectified and applied
to a burden resistor (not shown) to produce a dc
voltage proportional to the ac generator current. A
potential transformer 52 is connected to the generator
18 and produces a voltage signal in response to the ac
voltage potential produced by the generator 18. The
o~
--7--
voltage signal is rectified to produce a dc siynal
proportional to the ac generator voltage. It is
apparent to those skilled in the art that the current
and voltage signals could be produced directly by
monitoring the rectified dc generator current and
voltage with the appropriate circuitry.
The synchronous wheel slip controller 12 has
input terminals 54,56 for receiving the generator
current and generator voltage signals from the current
and potential transformers 50,52 respectively. The
synchronous wheel slip controller 12 accesses an
empirical wheel speed formula stored in the memory 88
and uses the formula to produce a calculated wheel
speed signal as a function of the current and voltage
signals. The synchronous wheel slip controller 12
then produces a lagged wheel speed signal by
calculating a first-order-lag of the calculated wheel
speed signal. If the lagged and calculated wheel
speed signals differ by more than a preselected
reference amount, the synchronous wheel slip
controller 12 restricts the magnitude of a wheel slip
voltage limit signal VLWS.
A desired rack calculator 58 forms part of a
generator control loop 60 which is used to regulate
the power produced by the generator 18. The desired
rack calculator 58 has input terminals 62, 64, 66, 68,
69 for receiving the wheel slip voltage limit VLWS,
actual engine speed, generator current, generator
voltage, and power signals from the synchronous wheel
slip controller 12, the engine speed sensor 44, the
current transformer 50, the potential transformer 52,
and the selector sensor 30, respectively. The desired
rack calculator 58 accesses formulas and tables stored
in memory 38 to calculate a desired rack signal as a
function of these input signals.
71~
The desired rack signal is applied to a
first input terminal 70 of a rack summer 72. A rack
sensor 74 detects actual rack position and produces an
actual rack signal which is applied to a second input
terminal 76 of the rack summer 72. The rack summer 72
produces a rack error signal eR in response to a
difference between the actual and desired rack
signals.
A control signal calculator 78 has an input
terminal 80 for receiving the rack error signal eR
from the rack summer 72. The control signal
calculator 78 accesses a control signal formula in the
memory 38 and uses the formula to produce a control
signal If as a function of rack error eR. The control
signal If is applied to an input terminal 82 of the
field current controller 24 to regulate generator
power output so as to reduce the rack error eR signal
to zero.
Industrial Applicability
Referring now to Fig. 5, a subroutine used
to control an embodiment of the synchronous wheel slip
controller 12 is illustrated by a flowchart. In the
block 220 the generator current and voltage IG,VG are
determined by monitoring the current and potential
transformers 50, 52 respectively. In the block 222
the empirical wheel speed formula is used to calculate
locomotive wheel speed as a function of the generator
current and voltage. In particular, the wheel speed
formula is as follows:
WNC = VG/ (K2 ~ Kl (IG) )
where WNC is the calculated wheel speed, VG is the
generator voltage, IG is the generator current, and Kl
-9- 2~G
and K2 are empirically determined constants for a
particular final drive ratio and wheel diameter
combination. An example of a curve generated by the
wheel speed formula is shown Fig. 2. The constants
K1, K2 are empirically determined by measuring
generator current and voltage IG, VG at various
locomotive speeds under controlled conditions.
A first-order-lag of calculated wheel speed
is determined in the block 224 using the following
Laplace transform equation:
WNL = WNC/(l + T(S))
where WNL is the lagged wheel speed, WNC is the
calculated wheel speed determined in the block 222, T
is an empirically derived time constant in seconds,
and s is the Laplace transform operator. The use of
software for implementing first-order-lags is commonly
known in microprocessor based control systems.
A relationship between the calculated wheel
speed WNC and the lagged wheel speed WNL during
locomotive acceleration without wheel slip is
illustrated in Fig. 3a. Initially, at A, the
locomotive is traveling at constant speed and the
calculated wheel speed WNC equals the lagged wheel
speed WNL. In response to an operator demand for
greater locomotive speed, the power produced by the
generator increases. As the locomotive accelerates,
the locomotive wheel speed WNC increases along the
solid line. It eventually levels off at B where the
power fed to the traction motor equals the power
required to operate the locomotive at the new speed.
The lagged wheel speed WNL follows the dashed line of
Fig. 3a and essentially lags the calculated speed WNC
by "X" mph during non-slipping acceleration.
-10- 2~ ~)7~
Furthermore, the difference "x" between the calculated
wheel speed WNC and the lagged wheel speed WNL is
related to, and can be used as a measure of, the
calculated wheel speed WNC.
continuing with the discussion of Fig. 5, a
slip flag is checked in the decision block 226.
Initially, the slip flag is not set; therefore,
control is passed to the block 228 where the lagged
wheel speed WNL is stored in the variable reference
wheel speed WNR. The reference wheel speed WNR is
used later in the routine to calculate the wheel slip
voltage li~it VLWS if synchronous wheel slip is
detected and to determine if synchronous wheel Slip
has stopped.
In the decision block 230 a difference
between the calculated wheel speed WNC and the lagged
wheel speed WNL is compared to a first empirically
determined reference Bl. If the compared speeds
differ by more than the first reference Bl, a
synchronous slip condition exists and the routine
continues to the block 232 where the slip flag is set.
A better understanding of the relationship
between the calculated wheel speed WNC and the lagged
wheel speed WNL during synchronous wheel slip can be
gained by referring to Fig 3b. When synchronous slip
occurs, the load on the traction motors is reduced
causing an instantaneous decrease in generator current
and increase in generator voltage. Therefore, if
synchronous wheel slip occurs, the calculated wheel
speed WNC will change rapidly as shown by the solid
line. However, the lagged wheel speed WNL will change
much slower due to the lag introduced by the Laplace
transform equation. The difference between the
calculated wheel speed WNC and lagged wheel speed WNL
is approximately proportional to the rate-of-change of
~0~(~7~4
the calculated wheel speed WNC. More partieularly, as
the rate-of-ehange of the ealeulated wheel speed
inereases, the differenee between the ealeulated wheel
speed WNC and lagged wheel speed WNL also increases.
Therefore, synehronous slip can be detected by
comparing the difference between the calculated wheel
speed WNC and the lagged wheel speed WNL to an
empirically determined value at which synchronous slip
occurs.
Continuing with the discussion of Fig. 5,
after setting the slip flag in the block 232 control
is subsequently passed to the block 234. In the block
234 the wheel slip voltage limit VLWS is calculated
using the following voltage limit formula:
v~wS = (WNR - B2) * (K2 ~ Kl(IG))
where B2 is an empirieally determined eonstant. The
wheel slip voltage limit VLWS is used later in the
ealeulation of the eontrol signal If. More
partieularly, when wheel slip is deteeted, the
referenee wheel speed WNR is redueed by a second
empirieally determined reference B2. This results in
a reduetion in the magnitude of the control signal If
and subsequently a reduetion of the locomotive wheel
speed.
If wheel slip is not detected in the
decision block 230, control is passed to the block 236
where the slip flag is eleared. Thereafter, control
is passed to the bloek 238 where a new wheel slip
voltage limit is ealculated using the following
equation:
VLWS = (WNR + B4~ * (K2 + Kl(IG))
-12 - ;~(~93.V70~
where B4 is an empirically determined constant. This
prevent~ the wheel slip voltage limit from limiting
generator power during a non-slipping condition.
The decision blocks 240 and 242 are used to
determine if a previously detected slip condition has
ended. More specifically, synchronous slip has
stopped if both the lagged wheel speed WNL and
calculated wheel speed WNC are less than the reference
wheel speed WNR plus a third empirically determined
reference B3. If wheel slip has ended, control is
passed to the block 236 where the slip flag is
cleared, and then to the block 238 where the wheel
slip voltage limit VLWS is increased. Similarly, if a
slip condition still exists, control is passed to the
block 232 where the slip flag is set, and then to the
block 234 where the wheel slip voltage limit VLWS is
reduced.
Referring now to Figs. 6a and 6b a flowchart
illustrative of software for controlling the generator
control loop 60 is described. In the block 260 the
actual engine speed NA is determined by monitoring the
magnetic pick-up sensor 44. Thereafter, in the block
262, a first preliminary desired rack R1 is calculated
as a function of the actual engine speed using a
second look-up table stored in the memory 38. The
second look-up table equates the actual engine speed
to a desired rack setting as illustrated in Fig. 4.
In the block 264 a measured rack position RM
is determined by monitoring the rack sensor 74.
Subsequently, in the block 266, selector voltage and
current limits VL, IL are calculated as a function of
the position of the power selector 28.
In the decision block 272 the selector
voltage limit VL is compared to the wheel slip voltage
limit VLWS. Thereafter, in the blocks 274, 276 a
070~
-13-
final voltage limit VLF is equated to the lesser of
the compared volta~e limits VLWS, VL.
In the block 278, the desired rack signal RD
is set to the lesser of three preliminary rack signals
Rl, R2, R3. The first preliminary rack signal Rl was
previously calculated in the block 262 as a function
of measured engine speed NA. The second and third
preliminary rack signals R2, R3 are limited by the
final voltage VLF and the current limit IL,
respectively. In the block 280 a rack ~rror eR is
produced in response to a difference between the
measured rack position RM and the desired rack signal
RD.
Finally in the block 284, the control signal
If is calculated using a transfer function of the PID
(proportional, integral, differential) type which is
consistent with known control theory. More
particularly the control signal is calculated using
the following control signal formula:
I f = K8 * eR + Kg * ~eR + Klo
where K8, Kg, and Klo are empirically determined
constants and ~B is a temporary software integrator.
The blocks 286 and 288 are optional and are
used to update the main software integrator ~A and
assure the control signal If is valid, respectively.
In summary, when synchronous slip occurs,
the synchronous wheel slip controller 12 produces a
wheel slip voltage limit VLWS which is used by the
desired rack calculator 58 to limit the desired rack
signal RD. As a result, a rack error eR occurs
causing the control signal calculator 78 to reduce the
control current If. Subsequently, the actual
locomotive wheel speed is incrementally reduced until
2~l(J70~
-14-
the synchronous slip condition is no longer detected
by the synchronous wheel slip controller 12.
Referring now to Figs. 7a and 7b, a
subroutine used to control an alternate embodiment of
the synchronous slip controller 12 is illustrated by a
flowchart. Fig. 7a is essentially the same as
previously discussed Fig. 5; therefore, only Fig. 7b
will be described.
In the block 338 a first-order-lag of the
measured generator voltage VG i5 calculated using the
following Laplace transform equation:
VGL = vG/(l + T(S))
where VGL is the lagged generator current, VG is the
measured generator current, T is an empirically
derived time constant in seconds, and S is the Laplace
operator. The Laplace transform equation effectively
filters noise from the generator voltage VG, thereby
providing a more stable signal with which to regulate
the wheel slip voltage limit VLWS.
In the block 340 the idle sensor 32 is
monitored to determine if the locomotive is in an idle
or load condition. If an idle condition exists, a
ramp voltage RV is set to zero in the block 342.
Under load conditions, the ramp voltage RV i5 used to
ramp the control signal If at a predetermined rate.
This strategy will be explained in greater detail
below.
Subsequently, the slip flag is checked in
the decision block 344. If the slip flag is set,
control is passed to the block 346 where a ramp
voltage RV is reset to the lagged generator voltage
VGL. Thus, when a slip condition ends, ramping
~0~1~7~
-15-
resumes at a level equal to the lagged generator
voltage VGL.
Thereafter, in the block 348, a preliminary
voltage limit VLP is set to the lagged generator
voltage VGL minus an empirically determined reference
B2B. Under a slip condition, the preliminary voltage
limit VLP is continuously reduced in this manner until
the slip condition ends.
If the slip flag is not set in the decision
block 344, control is passed to the block 350. In the
block 350 the preliminary voltage limit VLP is set to
the lagged generator voltage VGL plus an empirically
determined reference BlB. This prevents the
preliminary voltage limit VLP from being less than the
ramp voltage RV during a non-slipping condition.
Subsequently, in the block 352 the ramp
voltage RV is incrementally increased at a preselected
rate. In the preferred embodiment a rate of 125
volts/second is selected to allow the generator
voltage to reach the maximum rated voltage of 1250
volts over a ten second interval. As mentioned above,
the ramp voltage RV is used to ramp the control signal
at a predetermined rate. Therefore, the wheel torque
gradually increases under increasing load and the
ramping scheme achieves maximum load over a ten second
time interval.
Control is then passed to the block 354
where the ramp voltage RV is compared to the
preliminary voltage limit VLP. Subsequently, the
wheel slip voltage limit VLWS is set to the lesser of
the compared currents RV, VLP in the blocks 356, 358.
The wheel slip voltage limit VLWS is used in the
generator control loop 60, as previously described in
the discussion of Figs. 6a and 6B.
-16- 2~7~4
While the present invention is described for use
with the microprocessor governor control 10 disclosed
in Earleson et al., it is recognized that such a
synchronous wheel slip controller 12 could be used in
combination with numerous other locomotive gove~nors.
