U.S. patent number 3,581,071 [Application Number 04/815,075] was granted by the patent office on 1971-05-25 for train length measuring device.
This patent grant is currently assigned to The Baltimore and Ohio Railroad Company, The Chesapeake and Ohio Railway Company. Invention is credited to Lawrence A. Payseure.
United States Patent |
3,581,071 |
Payseure |
May 25, 1971 |
TRAIN LENGTH MEASURING DEVICE
Abstract
An apparatus and method for measuring the overall length of a
series of connected and moving articles each of which may be of
indeterminate length. The example to be described comprises
interconnected railroad cars forming a train. The total length of
the train is determined by employing two separated photocells and
two light sources, each source normally shining on a respective
photocell, so that each car of the moving train sequentially breaks
the lightbeams to the photocells causing the photocells to produce
electrical outputs. Since the distance between the two photocells
is known and the time which each car takes to travel from one to
the other is easily determined from the occurrences of output
signals from the photocells, the distance from the beginning of one
car to the beginning of the next is computed, the computations
being accumulated to indicate the total train length.
Inventors: |
Payseure; Lawrence A.
(Baltimore, MD) |
Assignee: |
The Chesapeake and Ohio Railway
Company (Cleveland, OH)
The Baltimore and Ohio Railroad Company (Baltimore,
MD)
|
Family
ID: |
25216780 |
Appl.
No.: |
04/815,075 |
Filed: |
April 10, 1969 |
Current U.S.
Class: |
702/158;
246/122R |
Current CPC
Class: |
G01B
11/043 (20130101) |
Current International
Class: |
G01B
11/04 (20060101); G01b 007/04 () |
Field of
Search: |
;235/151.32 ;246/182,1
;340/31,38 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Morrison; Malcolm A.
Assistant Examiner: Wise; Edward J.
Claims
What I claim is:
1. A method of determining the total length of a plurality of
moving articles, said method comprising the steps of:
detecting the leading edge of an article as it passes a first
sensing device to produce a first electrical signal,
detecting the leading edge of said article as it passes a second
sensing device separated from said first device by a given distance
to produce a second electrical signal,
determining a first time interval between the time when said first
signal is produced and the time when said second signal is
produced,
detecting the leading edge of the next article as it passes said
first sensing device to produce a third electrical signal,
determining a second time interval between the time when said
second signal is produced and the time when said third signal is
produced,
utilizing said first and second time intervals and said given
distance to calculate the distance from the leading edge of said
one article to the leading edge of the next article, and
summing all of the distances from the leading edge of one article
to the leading edge of the next article to determine the
approximate total length of said plurality of moving articles.
2. A system for measuring the total length of a chain of moving
articles, said system comprising:
first detecting means for detecting the leading edge of each said
article and for producing a first signal at the time said edge is
detected,
second detecting means separated from said first detecting means by
a given distance for detecting the leading edge of each said
article, after said edge is detected by said first detecting means,
and for producing a second signal at the time said edge is so
detected,
means operatively connected to said first and second detecting
means for determining a first time interval between the time when
said first signal is produced and the time when said second signal
is produced for any given article,
means operatively connected to said first and second detecting
means for determining a second time interval between the time when
said second signal is produced for said given article and the time
when said first signal is produced for the next succeeding
article,
computing means operatively connected to said means for determining
a first time interval and to said means for determining a second
time interval, said computing means being adapted to utilize said
first and second time intervals and said given distance to
calculate the distance from the leading edge of one article to the
leading edge of the next article, and
accumulator means for summing the distances from the leading edge
of one article to the leading edge of the next article to determine
the total length of said chain.
3. Apparatus as in claim 2 wherein each detecting means includes a
photocell and a source of light positioned so that a beam of light
is normally directed onto said photocell but is interrupted by said
article to cause the photocell to produce an output signal.
4. Apparatus as in claim 2, wherein:
said means for determining a first time interval comprises means
for producing a first number of pulses representing the time
between the leading edge of one article being detected by said
first detecting means and being detected by said second detecting
means,
said means for determining a second time interval comprises means
for producing a second number of pulses representing the time
between the leading edge of said one article being detected by the
second detecting means and the leading edge of the next article
being detected by said first detecting means,
said computing means comprising means for developing a given pulse
count representative of said given distance and means for
effectively dividing said second number by said first number and
for effectively multiplying the quotient by said given pulse count
to result in an additional count representing the distance between
leading article edges that is in excess of said given distance,
and
said accumulator means comprises means for adding said given pulse
counts and said additional counts for successive articles in said
chain.
5. Apparatus for measuring the length of a moving railroad train
made up of a plurality of connected cars comprising:
first means for detecting a given portion of each railroad car,
second means for detecting said portion of each railroad car after
that car is detected by said first detecting means, said first and
second means being separated by a given distance,
means for determining a first time interval between the detection
of the given portion of one car by said first detecting means and
by said second detecting means and for determining a second time
interval between the detection of the given portion of said one car
by the second detecting means and the given portion of the next car
by said first detecting means,
means responsive to said given distance and said first and second
time intervals for calculating the distance between the given
portion of said one car and the given portion of the next car.
6. Apparatus as in claim 5, wherein said given portion is the
leading edge of said car.
7. Apparatus as in claim 5, further including means for adding all
of the distances between said given portions to produce a result
approximating the total length of the train.
8. Apparatus as in claim 5, wherein said determining means
includes:
a source of clock pulses,
a first register,
a second register,
means for routing said pulses to said first register during said
first time interval so that said first register accumulates a
number of pulses which is a function of said given distance and the
duration of said first time interval,
means for routing said pulses to said second register during said
second time interval so that said second register accumulates a
number of pulses which is a function of the duration of said second
time interval,
and means for comparing the numbers accumulated in first and second
registers.
9. Apparatus as in claim 8, wherein said first detecting means
produces an electrical pulse when the given portion of a car is
detected to permit said routing means to route said clock pulses to
said first register and said second detecting means produces an
electrical pulse when said given portion is detected to route said
clock pulses to said second register.
10. Apparatus as in claim 8, wherein said first register
accumulates the total number of clock pulses produced during said
first time interval by said clock pulse source divided by said
given distance, said second register accumulates clock pulses
produced during said second time interval, and said comparing
means, during the second interval, divides the number accumulated
in said second register by the number accumulated in said first
register.
11. Apparatus as in claim 5, wherein said first and second
detecting means each comprises a photocell and a source of light
disposed in relation to said photocell such that said source
produces a lightbeam which shines on said photocell except when
interrupted by said cars.
12. Apparatus as in claim 5, including additional detecting means
for sensing the presence of said train to actuate said first and
second detecting means, said determining means, and said
calculating means.
13. A method of determining the length of a moving railroad train
made up of a plurality of cars comprising the steps of:
detecting the leading edge of a railroad car as it passes a first
sensing device to produce a first electrical signal,
detecting the leading edge of said railroad car as it passes a
second sensing device separated from said first device by a given
distance to produce a second electrical signal,
determining a first time interval between the time when said first
signal is produced and the time when said second signal is
produced,
detecting the leading edge of the next railroad car as it passes
said first sensing device to produce a third electrical signal,
determining a second time interval between the time when said
second signal is produced and the time when said third signal is
produced,
utilizing said first and second time intervals and said given
distance to calculate the distance from the leading edge of said
one car to the leading edge of the next car, and
summing all of the distances from the leading edge of one car to
the leading edge of the next car to determine the approximate total
length of said train.
14. A method as in claim 13, wherein said utilizing step includes
the further steps of :
dividing said first time interval by said given distance,
dividing said second time interval by said divided first time
interval, and
adding said second time interval divided by said divided first time
interval to said given distance to produce the approximate distance
from the leading edge of each car to the leading edge of the next.
Description
BRIEF DESCRIPTION OF THE PRIOR ART
AND SUMMARY OF THE INVENTION
The invention relates to an apparatus and method for measuring the
overall length of a chain of moving articles, each of indeterminate
length, such as the cars forming a railroad train.
It is often desirable to know the overall length of a long chain of
moving articles each of which may have a different and unknown
length, and the need for detecting the overall lengths of railroad
trains is particularly acute. Besides providing general records
which are of great value and interest, such train length
determinations are particularly desirable at locations where a
train may be routed to one or another siding since it is vital to
know in advance whether the train, which, of course, may stretch
for more than a mile or two, will fit within a particular
siding.
Many attempts have been made in the prior art to measure the
overall lengths of railroad trains, but all have proved either too
complex, expensive, or unreliable to be practical or too inaccurate
in their computations to be worthwhile. Although, it is simple
enough to measure the time between when the train reaches a given
point and when it leaves that or another point, it is much more
difficult to continually determine the velocity and acceleration of
the train as it passes the measuring point or points especially
when dealing with accelerating or decelerating trains of
extraordinarily long lengths. Without measuring velocity and
acceleration continuously, or without somehow minimizing the errors
resulting from changes in velocity and acceleration, it is
virtually impossible to accurately measure the length of a train
the velocity and acceleration of which are changing.
The present invention accomplishes such a measurement accurately by
individually determining the distances between the fronts, or
leading edges, of adjacent railroad cars which pass successively by
two separated sensing devices, each comprising a light source and
photocell, and then summing all of the front-to-front distances so
measured to determine the total length of the entire train. By
making such a car-by-car determination, the errors due to
acceleration and changes in acceleration are minimized and, as will
be seen hereinafter, for all practical purposes are eliminated.
The invention includes an arrangement of two separated sensing
devices mounted along a railroad track and a simple computing
device connected to the devices which can quickly and accurately
compute the train length and which does not require data to be
stored during computation. The arrangement can be adapted for use
regardless of the direction of travel of the train, and accurate
measurement over a wide range of speeds is possible. Furthermore,
accuracy is not reduced by variations in car lengths or dimensions.
Accordingly, the detection device of this invention is quite simple
and reliable while at the same time economical to install and
maintain.
In the embodiment set forth below, each of the two sensing devices
comprises a photocell and a light source mounted across the
railroad track from the photocell so as to project a lightbeam
across the track onto the photocell, and the two sensing devices
are separated by a distance which is less than the length of most
cars. As the moving train passes the first sensing device and
blocks the light beam which is directed towards the first
photocell, an electrical signal is produced causing the computing
network to begin calculating the length of the first car of the
train. The interruption of the beam to the first photocell, and the
electrical signal resulting therefrom, produces a pulse in the
computing network which is applied to logic circuitry to control
the gating of clock pulses through a divider to a first register
where the received pulses are counted. When the moving train breaks
the beam from the second light source, which has been striking the
second photocell, another electrical pulse is produced and this
pulse causes the logical network to shift the same free-running
clock pulses directly to a second register where the pulses are
counted. The two registers in effect record time. When another
electrical signal at the first photocell is produced indicating
that the moving car being measured has completely passed the first
sensing device and the next car has broken the beam, the clock
pulses are shifted back to the divider. During the period when
clock pulses are applied to the second register, the contents of
this register are compared with those of the first register. When
the pulse counts are equal, a pulse is generated to reset the
second register and to be applied to an accumulator in which is
already stored a number representing roughly the separation between
the sensing devices, this number having been entered at the time
that the first pulse was produced. As will become apparent from the
discussion below, the total information added to the accumulator
for each car represents the distance from the front of one car to
the front of the next. This calculation is then repeated for each
car until the entire train has passed at which time the total count
in the accumulator is roughly equal to the overall train
length.
Details of the invention will become clear from reading the
following description of an illustrative embodiment of the
invention.
A BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram illustrating a train approaching the
sensing devices;
FIG. 2 is a diagrammatic view of the position of a light source and
a photocell with respect to a railway car being sensed;
FIG. 3 is a block diagram of a logical and computation circuit for
determining the distance from the beginning of a car to the
beginning of the next car and for accumulating all of such
distances to determine the overall train length; and
FIG. 4 is a graphical representation of photocell outputs versus
time.
DETAILED DESCRIPTION OF THE INVENTION
Reference is now made to FIG. 1 which diagrammatically illustrates
this invention as used in measuring a conventional train 20, which
is shown as comprising three cars 22, 24 and 26, approaching the
measuring arrangement on a pair of conventional rails 30 and 32. Of
course, most railroad trains are of much greater length than three
cars and, in fact, can stretch a mile or more. It is with these
extremely long trains that this invention is particularly
useful.
As shown in FIG. 1, two sources of light 34 and 36 are disposed so
as to normally direct light onto conventional photocells 38 and 40,
respectively, which are both connected to computation circuitry 58.
One arrangement of the light source 34 and photocell 38 is shown in
greater detail in FIG. 2, wherein the lightbeam is shown just about
to be interrupted by the leading edge of a moving car 42. In FIG.
2, the light source 34 is disposed at roughly the level of the
railroad crossties 44, while the photocell 38 is disposed on the
opposite side of the rails 30 and 32 from source 34 and preferably
at an elevation of 10 to 15 feet above the ground. Thus, the
lightbeam from source 34 is projected upwardly at a substantial
angle. Interruption of the beam by a railroad car causes the
photocell 38 to produce an electrical signal as will be further
described hereinafter.
When the first car 22 of the train 20, which would ordinarily be
the engine pulling the other cars, breaks the beam projected by the
light source 34 onto the photocell 38, the photocell 38 responds by
changing its electrical output from one condition, arbitrarily
called "on," to another condition, arbitrarily called "off." This
is shown in FIG. 4 in which the change in output, indicated as 50,
occurs at the point in time S.sub.1 at which the beam from the
light source 34 is interrupted. The train 20 moves onward and at
time S.sub.2 car 22 interrupts the beam from the source 36 causing
a change in output, shown as 52 in FIG. 4, from the photocell 40.
As discussed in detail below, the separation between the photocells
38 and 40, which is labeled S in FIG. 1, is preferably sufficient
to minimize errors due to acceleration and yet less than the length
of most of the cars in the train, so that each car interrupts the
beam from source 36 before the following car breaks the beam from
source 34.
After car 22 has passed the beam from source 34, the output of
photocell 38 returns to its previous condition, as shown by 54 in
FIG. 4. Since it is desired to determine the total length of the
train 20 rather than the length of any individual car, the change
is output 56 produced at times S.sub.3 when car 24 breaks the beam
to the photocell 38 is employed to signify the end of the distance
measuring period rather than the pulse 54. Thus, the distance
between the fronts of adjacent cars is calculated rather than the
length of an individual car.
As shown in the graphs of FIG. 4, each moving car covers a distance
d.sub.1 in a time t.sub.1 between the output changes 50 at time
S.sub.1 and 52 at times S.sub.2 representing, respectively, the
time that a car breaks the beam from the source 34 and the time at
which it interrupts the beam from source 36. Since the distance
d.sub.1 is obviously the distance which separates the two sources
34 and 36, and the time t.sub.1 can be easily determined using
appropriate electrical circuitry responsive to changes 50 and 52,
the average velocity v.sub.1 at which the car being measured is
moving during this time t.sub.1 can be computed by dividing d.sub.1
by t.sub.1. If it is assumed that the car continues to move at this
same velocity (the accuracy of this assumption is discussed below),
then the distance d.sub.2 which the car moves during time t.sub.2
(which is the time from which the car first interrupts the beam to
the photocell 40 to the time at which the next car in line breaks
the beam to the photocell 38), can be calculated by multiplying the
velocity v.sub.1 times the second time interval t.sub.2 which is
determined by circuitry responsive to changes 52 and 56. Thus, the
total distance between the front of one moving car and the front of
the next moving car can be established by summing the distance
between the two photocells 38 and 40, d.sub.1, and the distance
d.sub.2.
The formulas for calculation are derived as follows:
d.sub.1 =v.sub.1 t.sub.1
d.sub.2 =v.sub.2 t.sub.2
and v.sub.1 roughly equals v.sub.2,
therefore, d.sub.1 /t.sub.1 =d.sub.2 /t.sub.2 ;
and since d.sub.1 =S, d.sub.2 =St.sub.2 /t.sub.1.
Assuming the distance S between photocells 38 and 40 is 35 feet,
the total distance D between the leading edges of adjacent cars is
as follows:
D=d.sub.1 +d.sub.2 =S+S t.sub.2 /t.sub.1
or D=35+35 t.sub.2 /t.sub.1
Thus, it can be seen that the distance D is the sum of a constant
and a variable, only the variable being subject to errors caused by
acceleration or deceleration of the train.
Reference is now made to FIG. 3 which shows an arrangement of
logical and other elements for performing the computations required
to calculate the total distance D from the front of a car to the
front of the next car and then to accumulate the distances so
calculated in an output register which can be simply read without
further computation or effort after the train 20 has passed to give
the overall train length. As shown in FIG. 3, a pulse forming
amplifier 60 senses the changes in electrical output from the
photocell 40 signaling the detection of the front of cars which, as
shown in the graph of FIG. 4, occur at times S.sub.2, S.sub.4,
S.sub.6, etc. The pulse forming circuit 62 similarly senses the
changes in electrical output from the photocell 38 also signaling
the detection of the front of cars, and as shown in FIG. 4, these
changes occur at times S.sub.1, S.sub.3, S.sub.5, S.sub.7, etc. The
pulse forming amplifiers 60 and 62 are intended to be of
conventional type and any appropriate devices which will perform in
the manner set forth below may be employed.
As shown in FIG. 1, two conventional detecting circuits 70 and 72
are associated with the rails 30 and 32 in the immediate vicinity
of the light source 34 and serve to produce appropriate signals
when the presence of the train wheels passing them is sensed. The
detecting circuit 70, which is connected to sources 34 and 36 and
photocells 38 and 40, is disposed ahead of, and insulated from, the
circuit 72 so that it detects the moving train 20 first, and the
detection signal produced by circuit 70 serves to apply power to
the photocells 38 and 40, the sources 34 and 36, and the
computation circuit 58, as well as resetting the accumulator 96 to
zero. As shown in FIG. 3, both circuits 70 and 72 are required to
be producing signals indicating the presence of a train before the
computing circuit 58 can actually calculate. More particularly, in
FIG. 3 the detection of a car by circuit 70 is represented as
closing switches 74 and 80 while the closing of switch 76 is shown
resulting from the detection of a car by circuit 72. The closing of
switch 80 applies power to the circuitry 58, the photocells 38 and
40, and sources 34 and 36 and resets the accumulator 96 to zero.
The subsequent closing of switch 76 readies the circuit 58 for
computation. Thus, the system cannot be inadvertently actuated, for
example by a person walking down the tracks.
After power supply and resetting has occurred due to detection of a
car by circuit 70, the first signal received by the arrangement
shown in FIG. 3 will, of course, be a change in electrical output
50 from photocell 38 at the time S.sub.1, and this change is
converted by the amplifier 62 to an appropriate pulse for causing
the registers 84 and 106 to be reset via the diodes 86 and 122,
respectively.
The output of amplifier 62 is also applied to the flip-flop 88 via
diode 89 to cause flip-flop 88 to shift from its "off" or zero to
its "on" or one condition, thereby enabling the gate 90 in
preparation for a signal from the pulse forming amplifier 60 in
response to the subsequent interruption of the beam to the
photocell 40.
Also, the pulse from the amplifier 62 is multiplied by 36 in the
appropriate circuitry 94, for the reasons set forth below, and the
output from 94, which may be in the form of 36 short pulses, is
applied to the accumulator 96.
The detection of the first car by circuit 72 permits computation of
distance to proceed as will now be described.
The pulse produced by the amplifier 62 is applied to a logic gate
100 thereby enabling gate 100 to pass a train of uniformly spaced
timing pulses from a free-running, pulse-producing clock 102 to a
circuit 104 which divides the number of pulses received by 35, so
that one pulse is produced at the output of circuit 104 for each 35
pulses received on its input. The pulses from circuit 104 are then
counted in a register 106, the output of which is applied to a
comparator circuit 108 as described below.
The output of the amplifier 62 is also applied to the "off" input
of the logic gate 110 to prevent signals from passing from the
clock 102 through the gate 110 to accumulate a count in the
register 84, which, as discussed hereinafter, is used in
determining the time t.sub.2 as shown in the graph in FIG. 4.
A short time after the change in signal 50 produced at time S.sub.1
occurs, another change 52 is sensed at the input of pulse amplifier
60 at time S.sub.2 as the car being measured interrupts the beam
produced by the source 36 and striking photocell 40. The pulse
output of amplifier 60 is then applied to the gate 90, and since
gate 90 was previously enabled by the output of the amplifier 62 at
time S.sub.1, the pulse passes gate 90 and is applied to the gate
110 to enable gate 110 thereby passing the output of clock 102
directly to register 84 which counts the clock pulses.
The output of the gate 90 is also applied to the "off" input of the
gate 100 to disable the gate and to prevent further pulses from
being applied to the register 106 by the clock 102. The output of
the pulse forming amplifier 60 is also applied through a delay
circuit 112 and a diode 114 to the flip-flop 88 to turn that
flip-flop off and disable the gate 90 until another pulse from the
amplifier 62 is received at time S.sub.3 by flip-flop 88 to again
enable the gate 90.
As the counts are stored in register 84, a comparison is made with
the counts in register 106 by the comparator 108. When the counts
in register 84 and 106 are equal, a pulse is generated by the
comparator 108 and is applied to the accumulator 96 via diode 120.
Also, the pulse from the comparator 108 is applied via diode 124 to
register 84, resetting this register to a zero count. This sequence
continues until a signal change is sensed at time S.sub.3 by pulse
forming amplifier 62 to disable the gate 110 and prevent further
pulses from being applied to the register 84. During the comparison
period, the comparator 108 actually performs the division of
t.sub.2 by t.sub.1 /35.
As pointed out above, the total distance from the beginning of one
car to the beginning of the next is equal to the distance between
the two sources, 35 feet in this embodiment, plus that distance
times the ratio of t.sub.2 to t.sub.1. This sum is calculated in
the illustrative embodiment, in the manner just explained, by first
adding 36 to the accumulator via multiplier 94. The number 36 is
used to compensate for errors resulting from roundoffs in the
division. Then added to the accumulator is the comparator output,
35 times the time t.sub.2 divided by the time t.sub.1.
Thus, as each individual car passes the photocells 38 and 40 in
turn, the distances from the front of each car to the front of the
next are calculated and successively added to the accumulator 96.
The last car of the train is sensed by photocell 38 to cause a
count of 36 to be entered into the accumulator 96. However, before
any further input to the accumulator can be generated by the
computation circuit 58, circuit 70 opens to terminate the
computation. At this time the accumulator holds a count
representative of the total length of the train.
Since a minimum car length equal to the distance between the
separated photocells 38 and 40 is assumed, and the acceleration and
changes in acceleration of the train during the computing periods
are ignored, the system discussed above is inherently subject to a
certain amount of error. This error is obviously largest when the
train is moving at a minimum speed and, for the embodiment shown in
FIG. 1 with a separation between photocells of 35 feet, the minimum
speed with which any accurate measurement can be made has been
found to be approximately 41/2 to 5 miles per hour. Below this
speed the measurement has been found to be grossly in error. Thus,
it is desirable to detect low train speed below a given minimum and
to eliminate any distance information which results from
measurement at or below such speed. This slow speed detection is
accomplished in the embodiment shown in FIG. 3 by detecting an
overflow of the register 106. This overflow is designed to occur
when the number of pulses received in register 106 exceeds a chosen
number. The overflow causes the computation to stop and the
accumulator 96 to be reset to zero. The circuit 58 then remains in
this suspended condition until the train 20 being measured clears
the circuits 70 and 72 at which time the circuit 58 is once again
made ready to respond to the next train and calculate its
speed.
On the other hand, excessive error in the train length calculation
may result if the train speed is so great as to cause the circuitry
to operate near its upper frequency limits. However, the proper
choice of conventional electronic components in a manner which is
well known will permit the maximum allowable train speed to be well
in excess of 60 miles an hour without resulting in substantial loss
of accuracy.
As pointed out above, the measurement system of FIG. 1 makes no
allowance for acceleration of the car after the average velocity
has been determined during the first 35 feet of travel. This
assumption obviously introduces errors into the calculated car
length if the train 20 is accelerating or decelerating, but these
errors have been found to be small and acceptable. The longest car
entering the track at minimum speed and maximum acceleration will
obviously produce the greatest error and for an acceleration of 1
foot per second per second, which is the maximum expected, a car
length of 94 feet, an initial speed of 4.1 mile per hour and an
overall train length of 7,144 feet in 76 cars, the errors due to
acceleration for each of the cars have been found to be roughly as
follows: ##SPC1##
It is thus apparent that the error at this low speed for the first
few cars is substantial, but as the speed increases the errors
decrease until, for the last 63 cars, they amount to roughly only a
foot per car. The total difference between the true overall train
length and the calculated length as shown is only 118 feet and this
amounts to an error of only about 1.65 percent. Thus, it can
normally be assumed that, for any train, the acceleration error
will be less than 120 feet and results can be examined on that
basis.
To determine the acceleration error for an average train it is
assumed that an acceleration of 0.4 feet per second per second is
an average value and a car length of 60 feet and a speed of 20.5
miles per hour are typical. Thus, with a train length of 9,000 feet
and 150 cars the errors due to acceleration are as follows:
##SPC2##
An average train with these characteristics would have therefore an
error of only about 31 feet out of almost 9,000 due to
acceleration. This is only 0.34 percent of the total train length
and is small enough to be almost negligible. Of course, in the
above examples the acceleration was assumed to be positive, that is
the train speed was increasing, but if deceleration is occurring,
the error would be approximately the same size but of opposite
direction.
If by chance a car having an overall length less than the distance
between the photocells is encountered, the car is computed as
having a length equal to said distance. This introduces a positive
error in the computation. To overcome such error, the distance
between photocells should not be longer than the shortest car
measured.
Another source error is introduced by the clock pulses produced by
the clock 102 shown in FIG. 3 since the output can vary by as much
as one pulse for any given time interval. This error can be
minimized by running the clock at a frequency of 10,000 cycles per
second, thus making the maximum error practically zero.
Another error results from the division operation within the
circuit 104 since it is contemplated that the fractional remainder
of the quotient will be discarded. However, with a train speed of
4.1 miles per hour, it has been shown that this discarding of
remainder results in an error of less than 0.06 percent of the
residual car length over 35 feet (no error in the first 35 feet of
the car length) and even when the car is travelling at 60 miles per
hour, the error is still only 0.88 percent of the residual car
length in excess of the first 35 feet. Similarly, the second
division of t.sub.2 by t.sub.1 also omits the remainder and thus
also produces a slight error. However, this error is compensated
for by increasing the constant value 35 which is fed into the
accumulator 96 to 36, and this on the average adds one-half foot to
the calculated length of the car.
Yet another source of error results from the fact that the
separation between the photocells 38 and 40 is chosen to be 35 feet
in this embodiment and any car with an overall length less than 35
feet is computed as having a length of 35 feet. However, the
acceleration error is decreased as the length of the separation
between the photocells 38 and 40 approaches the length of the cars
being measured. Thus, to minimize acceleration errors the photocell
separation should be equal to the length the car being measured,
while to minimize errors resulting from short cars, the length
between the photocells needs to be no longer than the shortest car.
A separation of 35 feet has been selected as a value which
substantially satisfies both of these mutually antagonistic
conditions since freight cars having coupler length less than 35
feet are rare and the acceleration error of the system using a 35
feet separation is well within acceptable limits.
Yet another small error is introduced by the train overhang beyond
the last pair of wheels which adds roughly 10 feet to the
calculation. The time delay of the electrical and mechanical
sensing devices and the errors resulting therefrom will be
ordinarily negligible. For the train in what is thought to be among
the worst possible situations, consisting of 76 cars of 94 feet in
overall length, the following tabulation gives all the errors
expected: ##SPC3##
The computed train length is then 7,095 feet for total error of
less than -49 feet. However, to insure that a train will always fit
into a sliding, it is preferable that any error should always be
positive and that calculated train length should never be less than
the true length. This can be accomplished by adding a fixed amount
to each computed train length. For realistic situations, the total
error, however, will probably be positive and accordingly it will
be unnecessary to add a positive safety factor to the total
calculated train length.
For the train in the average case consisting of 150 cars of 60 feet
overall length, the following tabulation gives all the expected
errors. Computed train length in this example would be 9,080 feet
for a total error of only plus 80 feet or approximately one car
length. ##SPC4##
The system as shown in FIG. 1 is designed for measuring the total
train length when the direction of the train travel is from the
light source 34 to the light source 36. To measure train length for
the reverse train direction, however, it is simple to provide means
whereby train direction can be detected and circuits to the
photocell detectors 38 and 40 reversed so that the photocell 40 is
applied to amplifier 62 while the output of the photocell 38 is
applied to the amplifier 60. With additional circuits of the type
70 and 72 FIGS. 1 and 3) positioned immediately adjacent light
source 36, these additional circuits being insulated from each
other and from circuits 70 and 72 adjacent source 34, precise
measurement of train length can be obtained in the same manner as
heretofore described. If less accuracy can be tolerated, only
original circuits 70 and 72 may be utilized. However, this would
require that circuit 72 have associated therewith a further contact
connecting a source of power to photocells 38 and 40, sources 34
and 36 and the computation circuit 58, as well as resetting
accumulator 96. By using only original circuits 70 and 72 for both
directions of train travel, an error approximating the separation
of the photocells (in this case 35 feet) would occur when the train
is moving from light source 36 to source 34.
The measuring system of this invention is a particularly simple,
accurate, economical and reliable device for determining the
overall length of a series of moving connected articles, such as
railroad cars. The above example of the invention is merely
intended to set forth an illustrative embodiment and many changes
and modifications are possible without departing from the spirit of
the invention. Accordingly, the above invention is intended to be
limited only by the scope of the appended claims.
* * * * *