Note: Descriptions are shown in the official language in which they were submitted.
CA 02075762 2001-07-23
-1-
TITLE: Load Measuring System for Refuse Trucks
FIELD OF THE INVENTION
This invention relates generally to the field
of load-weighing systems. In particular this invention
relates to the load-weighing systems and components of
such systems which may be used to determine the weight
of refuse emptied from a container into a refuse truck.
BACKGROUND OF THE INVENTION
Garbage is a major environmental problem,
especially now for urban centers. Typically garbage is
collected from various premises and taken to a disposal
site. Often such disposal sites are landfill sites,
where the garbage is dumped into a hole a.n the ground.
However, landfill disposal sites are expensive to
operate and tend to fill up over time, thus needing to
be replaced. Rising land costs have made replacement of
landfill sites expensive. Additionally, landfill sites
may be environmentally hazardous, with chemicals,
pollutants and the like leaching into the ground water
supply.
Faced with the rising costs and concerns about
landfill sites, the operators of landfill sites have
begun to dramatically increase the fee charged for
disposing garbage at such sites. Typically now the fee
charged is based on the weight of garbage disposed of.
Thus, each refuse truck seeking to off-load garbage is
weighed on the way into the landfill site and weighed on
the way out of the landfill site. The weight difference
is then calculated and a fee is paid in accordance with
that weight difference.
CA 02075762 2001-07-23
-1a-
Additionally, concerns have been raised about
the total weight of individual vehicles which travel
public roads. The greater the weight of the vehicle the
greater the Wear of the roadway. Weight restrictions
5 have
- 2 - ~ut~~~aø.1'~D
been imposed on many roadways and substantial fines may be
payable for overweight vehicles. The landfill site weight
records can be used as evidence of overweight violations
by the appropriate regulatory agency. Both of these
trends, namely increased costs of dumping and increased
overweight fines, increase the importance for refuse truck
operators being able to know the total weight of the
refuse being carried in their trucks, and the incremental
weight being added by each additional pickup.
In a typical refuse collection operation, a
. customer signs a contract which guarantees that a refuse
receptacle will be emptied on a periodic basis. Presently
the contract is typically based upon a flat fee and does
not take into account the actual weight of refuse to be
removed from the site. The weight of the refuse may vary
with demolished building materials being very heavy
whereas cardboard packing or paper packing products are
very bulky, but light. The refuse hauler pays to dump the
refuse at the landfill or other disposal site based upon
the weight of refuse being disposed. However, the customer
is charged on a flat fee for a given time, such as a week
or month, resulting in customers with light refuse being
over-charged, while customers with heavy refuse are under-
charged. What is desired therefore, is a way for the
refuse truck operator and the refuse-hauling companies to
determine the weight of refuse being hauled from each
customer's container each time. Preferably, such
measurement device would enable the operator of the refuse
vehicle to know also what the weight of the refuse of the
truck at any given time was to avoid overloading 'the truck
and thus being liable for over-weight fines.
There are several examples of past devices which
attempt to provide the dESired weight information.
However, these prior devices all suffer from various
unsuitabilities. For example, there are devices which
require that the lifting, emptying and return cycle for
emptying a container into the refuse truck be slowed down
"t~~,' ~ a WD
or stopped so that a weight reading can be made. The
stopping or slowing dcwn of the lifting and lowering
motion of the container is unsuitable for several reasons.
Firstly, there is a time loss associated with such a delay
which is unacceptable. Secondly, the machinery used to
effect the lifting is often quite powerful, but not very
nimble. Thus, requiring the cycle to be slowed or stopped
typically causes large vibrations, which may create
excessive wear on the equipment and premature breakdown.
Thus, devices which require the.operation to be stopped or
slowed down in order to effect measurement are not
suitable. Examples of such devices may be found in United
States Patent No. 4,645,081 to Garbade and German
Publication 33 32 058.
Other devices have been proposed which involve
vertical load cells upon which the container being weighed
is placed. However, these also have a number of
disadvantages. As indicated previously, the equipment
while powerful often vibrates upon beginning or ending its
motion. Thus, placing containers directly on vertical load
cells is unsuitable, because it is difficult to protect
the load cells from excessive wear by the rough use that
such equipment typically gets. An example of such a device
is again United States Patent No. 4,645,081 and United
States Patent No. 4,714,122.
Other systems have proposed for example
measuring the fluid pressure of the hydraulic cylinders,
which are used to effect the lifting of the refuse
containers. Examples of such systems include United States
Patent Nos. 4,771,837; 4,824,315; and 4,854,406. However,
in each of these devices the pressure sensed is remote
from the actual loading causing the pressure. Thus, other
factors affect the pressure reading and 'the results are
generally unreliable.
Another system is that shown in United States
patent 5,083,624 which shows mounting a special transducer
onto various parts of a vehicle axle or lift arm.
a/ ~~ :T ~..a r~ ~ m
- 4 -
However, the transducer needs to be specially machined, is
non-load bearing and thus is remote from the actual loads,
and does not take into account the variable positioning of
the center of gravity of the load. Thus, this prior
invention shares the same shortcomings as the others in
not being able to determine weight.
Finally, there is the type of system like that
proposed in U.S. patent 3,724,575, which teaches mounting
a pair of strain gauges inside a uniform beam, loading the
beam beyond the second strain gauge with an unknown
weight, and wiring the gauges together so that the
difference between gauge detections is in constant fixed
proportion to the magnitude of the weight. However this
prior invention requires special modification and
weakening of the fork lift arm, requires a gauge mounting
surface inside the beam parallel to the neutral axis, and
a beam of special characteristics, namely being of uniform
modulus of elasticity and moment area of inertia. Thus
this prior device is impractical fox application to
existing non--uniform mechanisms or mechanisms where
sensors cannot be mounted parallel to the neutral axis.
What is desired therefare, is a method of
accurately determining the weight emptied from a container
being lifted, emptied and lowered without requiring
slowing or stopping of the lifting and lowering cycle, or
relying on indirect readings such as from the hydraulic
system or non-load bearing sensors, and one which is
practical, and does not require hollowed-out or uniform
special beam sections.
SLLMMARY pF THE P1RES:ENT IIvTVEIV1TIQT~1
A method of determining a weight of material
being emptied from a container wherein said container is
lifted to and lowered from an emptying position in a lift,
empty and lower cycle by an arm means, and wherein mounted
on said arm means are at least-one front and at least one
rear first sensor having respective first outputs in an
~'~FJ.:d ~~
unknown proportion
to the weight
of said container,
said
method comprisings
a) selecting at least one measurement position;
b) calibrating said first sensors by recording
at said measurement position, said respective
first outputs from each of said first
sensors
fox both an empty arm means and for an
arm
means lifting and lowering a calibrating
weight;
c) subtracting from said respective first
outputs recorded from said first sensors
for
said calibrating weight, the respective
first
outputs recorded from said first sensors
for
said empty arm means, at said measurement
position, during both the lift and lower
portion of the cycle;
d) lifting, emptying and lowering an unknown
load and recording said first outputs
from
each of said front and rear first sensors
for
each of said lifting and lowering portions
of
said cycle at said measurement position;
e) subtracting from said output recorded
during
the lifting portion of said cycle, said
output recorded during the lowering portion
of said cycle, at said measurement position,
to determine a change in output far each
first sensor;
f) computing from said change of outpLlt
at said
fxont and rear first sensors for said
unknown
weight and said change of output recorded
for
said front and rear first sensors of
Jail
calibrating weight, at said measurement
position, the weight of material emptied
from
said container.
BRIEF DESDRIPTI~Id DF THE 1~'R~T~TIPIGS
Reference will now be made, by way of example
only, to preferred embodiments of the invention with
~~ ~ ;~~3~i~D s
- 6 -
reference to the following drawings in which:
Fig. 1 is a side view of a front end loading
refuse truck having a weighing system according to the
present invention;
Fig. 2 is a free body diagram of the weighing
system of Fig. 1;
Fig. 3 is a schematic view of a preferred strain
gauge used according to the present invention;
Fig. 4 is a schematic showing wiring according
to the present invention;
Fig. 5 is a schematic of circuitry according to
the present invention;
Fig. 6 is a flow chart of algorithm for a main
program for implementing the present invention;
Fig. 7 is a flow chart of one of the steps of
Fig.6;
Fig. 8a is a partial flow chart of one of the
steps of Fig.6;
Figure 8b is the remainder of the flow chart of
Fig. 8a;
Figure 9 is a flow chart of one of the steps of
Fig. 8a;
Figure 10 is a partial flow chart of one of the
steps of Fig. 8b;
Fig. lla is a flow chart for one of the steps of
Fig. 6; and
Fig. llb is the remainder of the flaw chart of
Fig. 11a.
DETAILED DESCRIPTION OF THE TNV'ENTION
Fig. 1 shows a typical front end loading .refuse
truck indicated generally with reference numeral 10 having
a load-weighing system or apparatus indicated generally as
12 for implementing the method according to the present
invention. The truck.l0 leas a cab 14 and a truck bady l6.
The truck body 16 carries a receptacle 18 into which
refuse containers l7 may be emptied as described below.
~~~~ ~~~~a
The receptacle 18 provided with a cover 20 which is
attached by hinges 22. A cover hydraulic piston 24 acts
between the cover 20 and the receptacle 18 to raise and
lower the cover 20 around the hinges 22. The cover
hydraulic cylinder 24 may be operable from the cab or may
be automatically operated upon operation of the front end
loading system, in a known manner. In other cases, the
cover 20 may slide, to be opened each day and closed at
night.
The load-weighing system 12 is preferably
comprised of a number of interacting elements as described
herein. The system 12 may be divided into three main
categories, namely, the lifting components, which comprise
the lift arms and hydraulic cylinders as described below;
the data, which comprises constants and measured variables
as described below, and a computational unit to translate
the data into useful information which comprises
circuitry, and a method of calculation embodied in
computational algorithm, again, as described below.
Lifting Components
Beginning at the front end Fig. 1 shows a fork
arm 26 which is pivotally attached to a generally n-shaped
lift arm 30 at pivot point A. In turn, the lift arm 30 is
pivotally attached to the main body 16 at a txunnion pivot
0. Acting between the fork arm 26 and the lift arm 30 is
a fork. hydraulic cylinder 34 which acts as a pivoter to
cause fork arm 26 to pivot relative to lift arm 30. The
fork hydraulic cylinder 34 is pivotally attached to the
fork arm 26 at a pivot F and to a pivot point B on a
gusset plate 38, which is in turn attached to the lift arm
30. The fork hydraulic cylinder expands and contracts in
the direction of double ended arrow 35. The fork arm 26
is releasably connected to the refuse container. 17,
usually by being insertable into a lifting slot 19 shown
in ghost outline in Fig. 1.
Acting between the lift arm 30 and the truck
body 16 is a main lift hydraulic cylinder 42 which also
~'~'~::~
_ g _
acts as a pivoter to cause lift arm 30 to pivot relative
to truck body 16. The main lift hydraulic cylinder 42 is
pivotally attached to the truck body 16 at a pivot point
D at one end and is pivotally attached to a gusset plate
46 at a pivot point C at the other end. The main lift
hydraulic cylinder expands and contracts as indicated by
double headed arrow 43. Each of the hydraulic cylinders
is similar in operation and are' well known in the art .
Also shown is G which is the center of gravity of 'the arm
30 and fork 26 mechanisms, and which is explained in
greater detail below, and R which is the center of gravity
of the weight being lifted, which is shown as w.
Data: Measured Variables
Also shown in Fig. 1 are some of the variables
which are measured and utilized as inputs in the instant
invention. According to a preferred embodiment of -the
instant invention strain readings are taken from at Least
one forward strain gauge 52 and at least one rear strain
gauge 53 which are mounted directly onto the lift arm 30
as shown in Fig. 1. These strain gauges 52, 53 axe
mounted preferably directly and even bonded onto a load
bearing portion of the lift arm 30 to become an integral
part of a loadbearing transducer comprised of the strain
gauge and the arm itself and are for the purpose of
measuring the variable forces in the lifting apparatus
which occur during a lifting, emptying and lowering cycle
of the refuse container 17.
It is also necessary to measure at least one
varying angle, and this angle is noted as, j3 which is
explained below in reference to Fig. 2. The angle ~ is
preferably measured with an inclinometer 56 or a position
encoder which provides an output corresponding to the
degree of rotation of the arm 30.
Data: Constants
Certain fixed constants are also used in the
present invention. The distance from the pivot point 0 to
- 9 -
r~ ~~ ~-,~
P~Sr'~, ~ ..o A .5 ' .s
each of gauges 52, 53 on the one arm 30 can be measured,
and the distance between the gauges 52, 53 can also be
measured. These measurements become the pre-input
constants.
Mounting the Gauges
It will be appreciated that there would
typically be two arms 30, one on each side of the cab 14.
In the preferred embodiment with two arms 30 the front and
rear strain gauges 52, 53, would be mounted onto both
arms. If there is only one arm 30, such as inside lift
trucks, then, of course, only the one arm 30 would have
gauges 52, 53
The present invention contemplates mounting the
strain gauges directly onto the outer surface of the lift
arms 30. The invention does not require that the cross
section of the arms be uniform. Further in the present
invention it is not necessary that the gauges 52, 53 be
mounted on a glane parallel to the neutral axis of the
beam, for the reasons described below. Also, while the
Figures indicate that the strain gauges 52, 53 may be
mounted on the topsides of the lift arms 30, they may also
be positioned around the corners of the more vertical
sections, or on the undersides of the lift arms 30. Tt
will also be understood that the present invention is not
restricted to, n-shaped arms, but that an arm means of any
shape which is,pivoted at one end and supports a load a~t
the other end can be used. What is required for the
present invention is that the force in the lifting
apparatus arising from a load on the arm means, be sensed
in two places at least.
Figure 3 shows a strain gauge model HBW made by
Hitec Products, Inc., of P.O. Box 790 - Ayer, MA 01432
U.S.A. which has yielded good results in the present
invention. Shown are an extension cable 60 which ends at
a cable transition 61: The gable transition 61 has a pair
of hold-down tabs 62, which are applied by tack-welds 63.
A transition tube or lead wire ribbon 64 extends to the
~~F:~ t.~ f,u~~A
- 10 -
gauge body 65. The gauge body 65 includes a thin film 66
which is spot welded in place as shown at 67. Alignment
marks 68 axe used to ensure proper alignment. A glue or
bonding agent is also preferred to ensure that the gauge
body 65 is integrally fastened to the outer face of the
lift arm 30. For good results, the manufacturers
installation instructions should be followed.
It is necessary for the two gauges 52, 53 to be
separated by a sufficient distance to ensure sufficient
difference in strain from front gauge 52 to back gauge 53
to allow any effect of a varying location of a center of
gravity of the load, -to be removed. Good results have
been achieved when the strain gauges 52, 53 are separated
by at least .5 meters. Further, to achieve good results
Z5 it is desirable to mount the gauges 52, 53 at relatively
more compliant portions of the arm 30, and not near or
adjacent stiffening plates or the like. Such a location
will enhance the output.
Fig.4 shows the pair of lift arms 30, from
above, with strain gauges, or first sensors 52, 53 mounted
thereupon. Each strain gauge has a separate wire lead,
shown as 57f and 57r (for front and rear) leading back to
a digital to analog converter 118. 57f and 57r correspond
to extension cables 60 in Fig.3. It will be appreciated
that the unit 118 is a means to convert the individual
sensor outputs into a form acceptable to the computational
unit.
Preferred Method
In its most general form, -the method of weighing
of the present invention consists of the following
elements, namely, the load arm 30, which may be considered
as a beam, fixed or in the preferred embodiment pivoting
at one end, at least two strain gauges 52, 53 placed along
the beam, and separated by some distance to measure strain
in the beam when a concentrated load is placed on the end
of the beam, arid a concentrated load occurring at any
point beyond the location of the two strain gauges 52, 53.
o-~, "r~d M'~ !'s"J '~
-- 11 -- vd <_, dw.lJ a ~ ~ A
With the measured variables and constants, it is possible
to determine the actual weight of waste from a waste
filled refuse container 17 which is emptied into the
refuse truck 10 by using the preferred method.
Prior to detailing the preferred method however,
it is necessary to understand some underlying assumptions
and geometry. Any load on the arms 30 will be related to
the strain measured at any given location of the arm 30.
In the preferred method certain assumptions are made about
the nature of the forces and moments in the arm 30 as
explained below.
In Figure 2 a portion of the arm 30 is isolated
and put in equilibrium by a cut 31. The location of the
cut 31 is identified with parameters LS and AS. In this
case LS is the distance from point O, the pivot point, to
the cut 31, and AS is 'the angle between ZS and a
projection line 54 joining 0 and A. The preferred section
of the arm 30 to make the cut 31 is between B and C along
arm 30. In this part of the arm 30 may be considered as
the sensing portion. For accurate results, it will be
appreciated that no portion of the forces created by the
load W, bypass this sensing portion of the arm.
Another variable necessary is the angle ~3
between the plane of the cut 31, and vertical. j3 is zero
in Figure 2, and for the purpose of the following
description is positive in a counterclockwise direction.
It is also necessary to include the weight of
the lift arms and fork arms outboard of the cut 31. 'hhus,
Gi is the weight of the arms and other components of the
lifting apparatus outboard of the cut 31, LGi is the
distance to the center of gravity of the outboard
apparatus from 0 and AGi is the angle that hGi makes with
a horizontal plane: Note that the subscript i is used as
a generality and that this analysis will apply in
particular to each strain gauge mounted on arm 30.
Therefore,
- 12 - ~w'ti ~ ;:~i~ ~~d
Moment i = - W[LW cos (AW) + LA cos (AA) - LSi cos (ASi)]
-- (Gi) (LGi) cos (AGi) (1)
Shear i - - [W + Gi] cos (~) (2)
Tensile i = - [W + Gi] sin (j3) (3)
where, LW equals the distance between R and pivot point A;
AW equals the angle between hW and a fixed plane, such as
horizontal (shown as H in Figure 2.);
LA equals the distance between point 0 and paint A in
Figure 2; and
AA equals the angle between OA and a fixed plane such as
horizontal (shown as H in Figure 2) and is positive in a
counterclockwise direction;
Equations (1), (2) and (3) show the portions of
the internal forces in arm 30 caused by the weight of the
refuse are a linear function of the applied loads or
weight of the refuse. Essentially three factors support
the linear relationship between strain and weight:
Firstly, shear loading does not affect the strain at the
surface of the arm; secondly, the moment loading, and
thus the moment stresses increase as the distance from the
load increases (and they will be relatively large); and
thirdly, the stresses due to tensile loading do not
increase with distance from the load and are small due to
the fact that measurements can be made through angles in
which these loads are small, as they are a function of Sin
(AA). Thus the present invention assumes that the
stresses in the arm are only due to moment loading in the
arm. Thus strain measured by gauges 52 and 53 will be
proportional to both the load and the horizontal distance
from the strain gauges to the center of the mass.
In the following, strain gauge 52 could be
represented by i - 1 and strain gauge 53 could be
represented by i = 2 for example.
Denote the horizontal distances from gauge i the
center of gravity of load W at R, as LG~ where i is the
number of the strain gauge. From equation (1):
L~~ = LW cos(AW) + LA cos(AA) - LSD cas(AS~) (4)
The equations relating the measured strain to
- 13 - ~;.~ ~> a .:~ a a .
the load and distance to the center of mass of 'the refuse
and to the weight of the arm can be written ass
K1w WLc9 ~' K1c G1 ( 5 )
E2 = K2w WLc2 + K2c G2 ( s )
Where E~ is the strain measured by gauge i, W is the weight
of the bin, Gi is the weight of the arm mechanism between
gauge i and the end of the arm, and the K values are
calibration constants.
To determine the weight of the refuse in the bin
the following method is preferred, which involves two
steps, namely, a calibration step and a measurement step.
(A) Calibration. In the preferred embodiment it is
necessary to calibrate the load weighing system, fox two
primary reasons. Firstly, the arm will produce different
strain readings at different angles of rotation about the
pivot point, and secondly, the different strain gauge
locations will give different linear calibration constants
with respect to W since the lengths are different and the
arm may be non~uniform (non-uniform arm structures can be
easily accommodated provided the strain is in some
proportional relationship to a variably W).
Calibration for these two effects can be-
accomplished by completing at least one lift and lower
cycle of a known weight and center of mass (a calibration
weight) and at least one lift and 7,owering cycle using no
weight respectively. The removal of values of these
calibrations outside normally acceptable statistical
limits further improves the reliability of the
calculations . Using sensors mounted on both lift arms 30
also increases the accuracy of the reada,ngs by providing
a comparison value which allows averaging of left an.d~
right outputs if they are sufficiently close.
It is necessary to identify at least one
measurement position of the arm 30 at which all recordings
of output are made so that they can be compared and used
to compute a weight. In the preferred emi~odiment a
continuous stream of readings are taken during cala.bration
;f'a r''9 ;~; r"7 ~_"~
iG~ z, .9 ~~ ~ 13 '-n
- 14 -
and measurement during both the lift and lower cycle. It
is preferred to restrict the range over which the output
is used for weight calculations however. Good results
have been achieved with the arm 30 in a range of 7° to 22°
to a horizontal plane. Continuously recording each gauges
output over a defined range of angles, as a function of
said angle, allows a plurality of readings to be obtained
which can increase the likelihood of accuracy of the
readings, by allowing spurious data points to be ignored.
After installing the strain gauges 52, 53 a
calibration procedure is performed to determine values fox
the constants KEW, KZW, K~~, Kz~ and each of these four
constants will. be functions of the angle of tilt of the
arm with respect to a fixed plane, like vertical (i.e.
functions of angle J3) . The absolute values of K~~, and Kzc
do not need to be explicitly determined. An empty arm
lift {with no extra weight) will determine the values of
K~~ * G1 and K2~ * G2 as functions of )3 angles far
calibration purposes. In order to find KEW and K2W a lift
of known weight with known center of mass position can be
performed. This will be referred to as the calibrating
weight or known weight. The results of the empty arm lift
and the calibrating weight lift are combined to complete
calibration, as explained below in more detail. The
location of the center of mass can either be calculated or
measured as desired. Subtracting these values from the
measurements for the lifting of the known weight will
yield in values ,of KEW * W * L~~ and KZw * W * L~2. Since the
weight W and the horizontal distances L~~ and L~2 are known
(or can be easily measured) the parameters KEW and KZ« as
functions of angle can be calculated. for greater
accuracy multiple tests using a range of W (around the
expected loads of operation) may be performed and the
results suitably combined.
{B) Measurement. The measurement procedure is similar to
the calibration procedure. The refuse bin l7 is lifted,
emptied into the truck and then lowered. During the
lifting and lowering motions the individual electrical
- 15 -
output from each strain gauge is separately measured and
.recorded as function of the angle J3. The difference
between the lifting and dropping values of each output is
proportional to the strain due to the weight of refuse
emptied from the bin and the horizontal distances from
each of the gauges to the center of mass of the load.
Since the exact locations of each of the strain gauges 52,
53 is knocvn, the horizontal distance between the gauges
can be determined.
L~~-2 - I'~~ ° Lc2 ( ? )
In this sense L~~-2 15 the horizontal distance between the
gauges. The * superscripts on the strains indicate that
these values are the differences between the lifting and
dropping values, so
E1 - E1, lifting ~1, dropping - KlwW~'a1
and
_ _
E2 ~2,lifting E2, dropping - K2wW~'c2
Solving the three equations (?), (8) and (9)
simultaneously we find that:
2 0 W = ~~2INE1 K1 k~2
K1WK2W'Z'C2-2
(where these values are noted as a function of angle (3).
Once W is found the distances from the two strain gauges
to the center of mass of the refuse could be calculated:
E1 KaWLoz-z
Lo2 -_
K2W 1 'K1 W~2
and
- 16 -
_ E2 -KlWfc1-2
La2
'K2WE1 'K1WE2
However, in the embodiment of the invention it will be
understood that determining the location of the center of
gravity of the load is an unnecessary step.
Figure 6 is a flow chart which depicts one
algorithm for implementing this invention. Shown in
Figure 6 are the steps of Initializations 100,
Calibrations 101, Compute Weights 102, and Get Sensor Data
103. As shown in the attached pseudo code in Appendix T
a number of tasks could operate simultaneously; through
mufti-tasking, such as Read Sensor task; Capture weight
data task; Compute calibration task; and Weight task.
The flow chart of Figure 6 reduces the number of tasks
down to two. The Read Sensor task and the Capture weight
data have been combined in the Get Sensor Data task 103.
This task 103 can be viewed as a timer interrupt routine.
The main program consisting of steps 100 to 102
incorporates the remaining tasks as shown in the pseudo
code.
The Get Sensor Data task 103 may be as
represented in a flow chart in Figure 7. Task 103
processes one set of sensor values at a time. In step
270, the sensors values are read. If these values are
read in analog form it may be necessary to convert to the
digital equivalent of the analog sensor outputs. If lift
arms 30 are moving up, the condition in step 272 is
satisfied and the angle of arm 30 is then checked to
determine whether it is within the preferred weighing zone
which may be defined as ~i being from 7 degrees to 22
degrees. If the latter condition is also true,_ these
sensors values are saved in a queue. Steps ?78 to 282 are
for handling the case when the arms are moving downward:
These are similar to steps 272 to 276. ~Iowever, this
time, the sensors data may be stored in a second queue.
~~'~ W,
- 17 -
Once a set of sensors values have been
processed, another set is fetched and the above operation
repeats.
The main program begins with the Initializations
routine 100. In this routine 100, system diagnostics and
variable initializations are performed.
In step 101, calibration of the load weighing
apparatus is performed. A more detailed flow chart of
this routine is shown in Figure 8a. After steps 201 and
202, this routine 101 can be divided into 2 sections. The
first section consisting of steps 204 to 215, is for
performing empty arm lift. This section creates 'two
tables K2~*G2 and K~~*G1. These two tables are used in the
second section to derive two other tables: KEW and K2w. The
latter two calibration tables are used by the Compute
Weights routine 102 to determine weight emptied from
container 17.
Most steps in the Calibration routine 101 are
self-explanatory, but some of these steps are explained
below. In step 205, sensors values are obtained fram the
first queue created by the Get Sensar Data task 103
described above. This step 205 fetches all the data from
the first queue consisting of sets of sensor values.
Steps 207 and 209 then operate on these sets to derive
tables KzG*G2 and .K~~*G1. The KZG*G2 and the K~~*G1 tables
are the calibration tables for the pairs of rear and front
sensors, respectively. Steps 211 to 215 are identical to
steps 205 to 209 except that the data are obtained from
the second queue created by the Get Sensor Data task 103.
A flow chart of step 207 is shown in Figure 9.
In step 300, the rear right and rear left sensors values
are summed to obtain a combined strain value, EZ,,~~fting~ for
a particular arm angle j3. If the table entry for this
angle f3 is cleared, the combined strain value is sa ed in
this table entxy in step 304. Otherwise, the current
value of this table entry is averaged with this combined
strain value to form the new value for this entry as shown
'L"' ~'' r'.~
~ 18 _ 6~s ,, ..3~~~'-A
in step 303.
The above operation repeats for the remaining
sets o~ sensors values.
The flow chart for step 213 is identical to the
flow chart for step 207 except that the combined strain
value corresponds to ~Z,dropping'
The flow chart for steps 209 and 215 are very
similar to steps 207 and 213 described above. For steps
209 and 215, the front sensors values are used and the
K1G*G1 table is being updated. In step 209, the combined
strain value corresponds to ~~ gifting' In step 215, the
combined strain value corresponds to e~ dropping'
The second section of the Calibration routine
101 consisting of steps 220 to 234 as shown in Figure 8b,
is to derive two more calibration tables when the lift is
lifting a 3cnown or calibrating weight W. Steps 220 to 234
are very similar to steps 203 to 215. However, the
calculations to derive these two tables found in steps
226, 228, 232, and 234 are somewhat different.
A flow chart of step 226 is shown in Figure 10.
In step 310, the rear right and rear left sensors values
are summed to obtain a combined strain value. This latter
value corresponds t0 Ez lifting' The Kz« value for a
particular arm angle is determined by the equation shaven
in step 311. If the KZW table entry for this angle J3 is
not cleared, step 313 will be executed. This step 313
averages the existing table entry value with the newly
calculated KZW value. Otherwise, the newly calculated KZ«
value is saved directly into the table.
The steps 310 to 315 repeat for any sets of
sensors values which remain to be processed,
Step 232 is identical to step 226 except that
the combined strain value there corresponds to ez,d~opping
instead. As shown in Figure 8b, steps 228 and 234,update
the KEW table and are similar to steps 226 and 232,
respectively. For steps 228 and 234 the font sensors
values, the K~~*Gl table and L~~ are used instead of 'the
~'r ~B
~~'' ~;9'"a'~v~
rear sensors values, the Kz~*G2 table, and Liz,
respectively.
A flow chart of the Compute Weights routine 102
is shown in Figure lla and 12b. Steps 240 to 250 compute
~1.liftingr ~l,dropping~ E2,Liftingr and E2~dropping for a range of
(through the preferred weighing zone) arm angle J3 values.
In step 205, s2,~ifting~ and E2~dropping are matched up according
to their corresponding arm angle j3. For instance, the
Ez,lifting for an angle of 15 degrees is paired with the
~z,dropping for an angle of 15 degrees . Step 254 operates on
these pairs to compute a set of e2. Steps 256 and 258
which calculate a set of si are analogous -to steps 252
and 254. In step 260, a weight W is calculated for each
ei and e2 using the equation shown there. For instance,
if the pair of si and s2 corresponds to an arm angle J3 of
15 degrees, the entries from the tables K1w and KZ~~ that
correspond to this angle j3 will be used in the
computation.
In step 262, the calculated set of weights W for
the range in values ~f arm angle ,ø will.be averaged to
determined to final weight value. A standard deviation is
also calculated. This deviation value is used to
determine whether the distribution of weight is within
acceptable limits in step 264. The error routine in 266
handles any unacceptable sets of W values.
Appendix I shows a preferred pseudo code for
implementing the preferred method generally described
above.
Qn-Board Circuitry
Turning now to Fig. 5, there is shown a
schematic diagram of -the on-board circuitry- according to
the present invention. Figure 5 shows a ground 110 and a
battery 112. Both positive and negatives leads from the
battery 112 pass through a noise filter 113 prior to
~''~ ~';:~'~
- 20 -
powering a computational unit or computer 114. The
computer has several input and output ports which can be
utilized to input and/or extract weight, sensor and other
information from the truck. One output port leads to a
mobile radio 116. Another output port leads to a
removable data storage device 147. An input port takes
input from a sensor junction box 118. The sensor junction
box in turn has a number of inputs as described herein.
Input 120 may be for example the vehicle speed, input 122
may be for example the vehicle fuel level. Dashed lines
124 indicate the measured inputs preferred for weight
reading system. The inputs 126 and 128 are the strain
readings on the lift and return portions of one of the
main arm 30 cycles from gauges 52, 53 respectively. The
measured inputs 130 and 132 are the strain readings taken
on the lift and return sides of the other main arm 30 from
gauges 52, 53. Inputs 136, 138 and 140 may be used as
needed for redundancy back-up. The measured input 134 is
J3, the measured angle of the main lift arms 30.
In the described system, the battery 112 is
preferably the vehicle battery. However, mobile radio 116
is preferably equipped with its own 12 V battery 142 to
avoid draining the vehicle battery 112. The mobile radio
116 is therefore provided with its own separate ground
144. The mobile radio 116 is also equipped with an antenna
146. The lines running into the sensor junction box axe
preferably two-channel lines, and 'the sensor junction box
is preferably attached to the computer 114 with a 28-
channel line. The connection with the mobile radio is
preferably a 5-channel line. A removable data storage
device 147 may also be used.
Some of the advantages of the instant invention
can now be appreciated. For example, in the preferred
embodiment, weight readings are not measured and
calculated on the lift cycle, nor are they calculated for
the lower cycle. To do so wou7.d require separate
calibration of the zero point of the strain ga~xges 52, 53
~n~ ~ ~.3~~ a
- 21 -
prior to every lift. Rather -this embodiment relies upon
recording a set of up cycle measurements from the gauges,
and a set of down cycle measurements from the same gauges,
and using the difference between the up and down sets of
measurements to derive a set of differential strain
readings. The set of differential strain readings are
then compared to a set of calibration curves obtained from
the very same equipment and combined with the measured and
derived (as a function of angle j3) distances between the
gauges to calculate the weight of the container contents
that was emptied into the truck.
It can also be understood that through the use
of the instant method, which incorporates a calibration
procedure and separate recordal of strain from each gauge,
the gauges can be mounted directly onto the outside of the
arm, without the need for special machining of a plane
parallel to the neutral axis of the arm, nor an arm with
an exactly. uniform response along its measuring length.
Separate measurement and recordal of strain, together with
calibration as aforesaid means that any local anomalies
can be accurately accommodated and compensated for with a
minimum of physical alteration to the lift arms 30.
It will now be appreciated that the present
invention provides a simple yet practical way to measure
loads from conventional equipment, without requiring
modification. While the foregoing discussion has focused
on n-shaped lift arms 30, it will be appreciated that any
shape of arm means can be used, providing that the arm
means is cantilevered, and for a portion of its travel,
contains a section with measurable strain in at least two
locations in response to an unknown load placed at its
end.
It will also be appreciated that the foregoing
description relates to strain gauges, but other sensors
will also be appropriate, providing that their output is
proportional to the load applied to the arm means. mhe
strain gauges of this invention measure the change in
r~ ~--rey ',~
22 _ ~'~~,~a~ ~
length over length of the arm means, when loaded. Thus,
these gauges measure the deformation of the arm means
under loading. Other devices for measuring such
deformation include, extensionmeters which may be defined
as devices which employ a system of levers to amplify
minute strains to a level where the strains, or
deformation can be read, linear variable differential
transformers which produce an output signal amplitude
proportional to strain, acoustical deflection sensing
devices where the time or phase delay in an acoustical
wave travelling along a body is proportional to strain,
optical deflection sensing devices that use interference
fringes produced by optical flats to measure strain,
vibrating beam devices where the frequency of vibration of
the sensor's beam is a function of the strain imparted to
the sensor's internal beam which is in turn affixed to the
body that is subjected to strain.
The basis of the preferred invention is to have
output from a sensor which is proportional to the force,
and indirectly to the weight, to which the measured body,
such as the arm means, is being subjected to.
It will be appreciated by those skilled in the
art that various modifications to the instant invention
can be made, which still fall within the scope of the
claims. Some of these variations have been suggested
above, and others will be apparent to those skilled in the
art. However, the important aspect of the instant
invention is the separate recording of output from at
least two sensors for at least one lift arm 30 and
preferably both lift arms 30, whereby reasonably accurate
weight measurement can be obtained, through proper
calibration and computational algorithms.
- 2 3 - y'.i~ s ~'a ~~
APPENDIx z
1) This appendix describes, through use of pseudo-
code, preferred algorithms for implementation of the
present invention. In the preferred embodiment referred
to in the pseudo code below there are four f9.rst sensors
in total, comprising two strain gauges 52, 53 (front,
rear) placed on each one of two arms 30 (left, right).
2) Pseudo Code
2.1) Sensor Readings
The following procedures preferably execute
continuously on the computational unit or computer.
This provides a constant stream of sensor readings for
use in calibration and weight computations. The GET
SENSOR READINGS procedure sums a set of successive
readings in order to average or "smooth" the data.
2.1.2) TASK READ SENSORS
FOR EVER
Select ARM ANGLE Analog Input
PERFORM GET SENSOR READING
Queue ARM ANGLE digital value
Select REAR LEFT Analog Input
PERFORM GET SENSOR READING
Queue REAR LEFT digital value
Select REAR RTGHT Analog Input
PERFORM GET SENSOR READTNG
Queue REAR RIGHT digital value
Select FRONT LEFT Analog Input
PERFORM GET SENSOR READING
Queue FRONT LEFT digital value
Select FRONT RIGHT Analog Input
PERFORM GET SENSOR READING
Queue FRONT RIGHT digital value
END FOR
2.2) Weight Data
The following procedure is preferably used to capture
a set of sensor readings, namely output, to be used
for either a calibration lift or a weight computation.
This procedure collects a set of readings or output
taken as 'the arm travels upwards through the vaeighing
zone and another set as the arm travels back downwards
through the weighing zone. The weighing zone is defined
as the range of angles of J3 which are preferred
measurement positions.
- 24 -
2.2.1) TASK CAPTURE WEIGHT DATA
Set Weigh Zone to be in the preferred embodiment in the
range of 7 to 22 degrees between a generally horizontal
part of the arm means 30 and a horizontal reference plane:
FOR EVER
Dequeue ARM ANGLE digital value
Dequeue REAR LEFT digital value
Dequeue REAR RIGHT digital value
Dequeue FRONT LEFT digital value
Dequeue FRONT RIGHT digital value
WHILE ARri ANGLE is travelling upwards 'through
Weigh Zone Collect ARM ANGLE, REAR LEFT, REAR
RIGHT, FRONT LEFT, FRONT RIGHT digital values
END WHILE
WHILE ARM ANGLE is travelling downwards through
Weigh Zone Collect ARM ANGLE, REAR LEFT, REAR
RIGHT, FRONT LEFT, FRONT RIGHT digital values
END WHTLE
Queue UP and DOWN digital values
END FOR
2.3) Calibration
The following procedures are preferred and are used to
calculate different calibration curves. These calibration
curves are used to identify the strains, or sensor
2S outputs, as functions of ARM ANGLE. Both the up and down
portions of the lift axe used to establish the curves. The
LEFT and RIGHT sensor. pairs are summed together to
provide separate REAR and FRONT calibration curves. An
empty arm lift is used to produce the REAR and FRONT EMPTY
ARM CURVES. A calibration or known weight is used to
produce the REAR and FRONT CALIBRATION curves: Multiple
lifts (minimum of 3) are preferred to "average" the data
in the curves.
2.3.1) PROCEDURE COMBINE STRAIN
FOR each reading in set
Sum LEFT and, RIGHT digital values
END FOR
2.3.2) PROCEDURE SET EMPTY ARM CURVE
FOR the set of LEFT and RIGHT digital values
PERFORM COMBINE STRAIN
END FOR
FOR the set of ARri ANGLE and combined strains
IF ARM ANGLE table entry clear THEN
Insert combined strains iwto table
ELSE
-- 25 -
Average combined strains into table
END IF
END FOR
2.3.3) PROCEDURE SET REAR CALIBRATION CURVE
FOR the set of LEFT and RIGHT digital values
PERFORM COMBINE STRAIN
SUBTRACT K1G*G1 curve value from combined strain
END FOR
FOR the set of ARM ANGLE and combined strains
Compute KlWi from LC1 and W
IF ARM ANGLE table entry clear THEN
Insert KlWi value into K1W table
ELSE
Average KlWi value into K1W table
END IF
END FOR
2.3.4) PROCEDURE SET FRONT CALIBRATION CURVE
FOR the set of LEFT and RIGHT digital values
PERFORri COMBINE STRAIN
SUBTRACT K2G*G2 curve value from combined strain
END FOR
FOR the set of ARM ANGLE and combined strains
Compute K2Wi from LC2 and W
IF ARM ANGLE table entry clear THEN
Insert K2Wi value into K2W table
ELSE
Average K26Ai value into K2W table
END IF
END FOR
2.3.5) TASK COMPUTE CALIBRATION
Enter measured values LC1 and LC2
Enter known calibration bin weight W
Compute LC12 as LC1 - LC2
Clear IC1G*Gl table
Clear K2G*G2 table
WHILE doing EMPTY ARM LIFTS
Dequeue UP digital values
USING REAR LEFT, REAR RIGHT, ARM ANGLE and
K1G*Gl table
PERFORM SET EMPTY ARM CURVE
USING FRONT LEFT, FRONT RTGHT, ARM ANGLE and
K2G*G2 table
PERFORM SET EMPTY ARM CURVE
Dequeue DOWN digital values
USING REAR LEFT, REAR RTGHT, ARM ANGLE arid
- 26 -
K1G*G1 table
PERFORM SET EMPTY ARM CURVE
USING FRONT LEFT, FRONT RIGHT, ARM ANGLE and
K2G*G2 table
PERFORM SET EMPTY ARM CURVE
END WHILE
Clear K1W table
Clear K2W table
WHILE doing calibration bin lifts
Dequeue UP digital values
USING REAR LEFT, REAR RIGHT, ARM ANGLE and K1W
table
PERFORM SET REAR CALIBRATION CURVE
USING FRONT LEFT, FRONT RIGHT, ARM ANGLE and K2W
table,
PERFORM SET FRONT CALTBRATION CURVE
Dequeue DOWN digital values
USING REAR LEFT, REAR RIGHT, ARM ANGLE and K1W
table
PERFORM SET REAR CALIBRATION CURVE
USING FRONT LEFT, FRONT RIGHT, ARM ANGLE and K2W
table
PERFORM SET FRONT CALIBRATION CURVE
END WHILE
END TASK
2.4) Weighing
The following procedure is used to compute a weight
emptied from a container.
2.4.1) TASK WEIGHT
FOR EVER
Dequeue UP digital values
FOR the set of REAR LEFT, REAR RTGHT, ARM ANGLE
readings
PERFORM COMBINE STRAIN to give combined
rear up values.
END FOR
FOR the set of FRONT LEFT, FRONT RIGHT, ARM
ANGLE readings
PERFORM COMBINE STRATN try give combined
front up values.
END FOR
Dequeue DOWN digital values
FOR the Set of REAR LEFT, REAR RIGHT, ARM ANGLE
readings
PERFORM COMBINE STRAIN to give combined
rear down val.ubs .
- 27 -
END FOR
FOR the set of FRONT LEFT, FRONT RIGHT, ARM
ANGLE .readings
PERFORM COMBINE STRAIN to give combined
S front down values.
END FOR
USING combined rear up and down values
PERFORM COMPUTE DELTA STRAIN to give E1
values
USING combined front up and down values
PERFORM COMPUTE DELTA STRAIN to give E2
values
USING the set of E1, E2, ARM ANGLE
PERFORM COMPUTE WEIGHTS
IF standard deviations outside of tolerance
Flag inaccurate weight
END IF
END FOR
2.4.2) PROCEDURE COIwIPUTE DELTA STRAIN
FOR each reading in set
Subtract down value from up value to give delta
END FOR
2.4.3) PROCEDURE COMPUTE WEIGHT
FOR each reading in set
COMPUTE W as (K2W k E1) - (K1W * E2) ~ K1W * K2W
* LC12
END FOR
USING the set of weights W
COMPUTE mean and standard deviation .
