Abstrict The present invention is a linearized flow meter apparatus. A turbine
flow meter measures a liquid flow and a processor corrects measurement
errors and provides an adjusted number of output pulses. The processor
uses an endpoint displacement method to provide the corrected output
pulse signals.
Claims We claim:
1. In a fuel dispensing apparatus having a flow meter and a liquid
flow line, an apparatus for adjusting the measured flow rate of
said flow meter to correct for the non-linear output signals of
said flow meter, said apparatus comprising:
a flow meter having a rotor, said flow meter adapted to measure
the flow rate of a liquid flow line as a function of the rotational
velocity of said rotor;
an output means, coupled to said flow meter, for generating signals
at a rate which is proportional to the rotational velocity of said
rotor, and providing an output frequency;
memory means for storing a plurality of calibration values, each
one of said calibration values being associated with a frequency
which is an endpoint of a prefined interval, such that each of said
intervals has a lower and a higher endpoint value;
processor means, coupled to said output means and said memory means,
for linearizing said output frequency by using a predetermined correction
factor to produce an estimated frequency, said processor periodically
recalculating said correction factor by selecting an appropriate
interval containing said estimated frequency, then determining the
relative displacement of said estimated frequency within said interval,
the relative displacement being defined as the difference of the
upper endpoint and the estimated frequency which is divided by the
length of the interval, then calculating a new correction factor
(K) using the equation:
with KA being the calibration factor associated with the lower
endpoint value of said interval, X being said relative displacement,
and KB being the calibration factor associated with said higher
endpoint value.
2. The apparatus as described in claim 1 wherein said memory means
comprises a non-volatile memory.
3. The apparatus as described in claim 1 wherein said memory means
comprises random-access-memory circuits (RAM).
4. The apparatus as described in claim 3 wherein said RAM circuits
have a back-up power device.
5. The apparatus as described in claim 4 further comprising an
input means for receiving new calibration value pairs and storing
said pairs in said RAM, said calibration value pairs comprising
a calibration factor and an endpoint value.
6. The apparatus as described in claim 5 wherein said input means
also has the ability to receive a single calibration offset value
such that every value is shifted by said offset value.
7. In a fuel dispensing apparatus having a fuel line, a system
for accurately determining a flow rate of the fuel line, said system
comprising:
a turbine meter coupled to the fuel line having a pulse emitting
means for transmitting a pulse stream having a frequency proportional
to said flow rate;
a memory for storing a plurality of calibration values, each one
of said calibration values associated with a frequency which is
an endpoint of a predefined interval, such that each of said intervals
has a lower and a higher endpoint value;
a processor coupled to said turbine meter pulse emitting means
and said memory for observing said pulse stream such that for every
pulse from said meter a predetermined correction factor is summed
by said processor to determine said estimated frequency, said processor
periodically recalculating said predetermined correction factor
by selecting an appropriate interval containing said estimated frequency
within said interval, the relative displacement being defined as
the difference of the upper endpoint and the estimated frequency
which is divided by the length of the interval, then calculating
a new correction factor (K) using the equation:
with KA being the calibration factor associated with the lower
endpoint value of said interval, X being said relative displacement,
and KB being the calibration factor associated with said higher
endpoint value, then replacing said predetermined correction factor
with the value of said new correction factor.
8. The apparatus as described in claim 7 wherein said memory comprises
a non-volatile memory.
9. The apparatus as described in claim 7 wherein said memory comprises
random-access memory (RAM).
10. The apparatus as described in claim 9 wherein said RAM further
comprises a battery backed-up memory.
11. The apparatus as described in claim 7 further comprising an
input means for receiving new calibration value pairs and storing
said pairs in said memory, said calibration value pairs comprising
a calibration factor and an endpoint value.
12. The apparatus as described in claim 11 wherein said input means
also has the ability to receive a single calibration offset value
such that every calibration factor is shifted by said offset value.
13. In a flow meter apparatus, a method of producing a linearized
stream of output pulses, comprising the repetitive steps of:
checking for an input pulse wherein if the recent occurrence of
an input pulse is detected, a predetermined correction factor is
added to an input pulse counter and overflow is checked for, else
a cycle check is performed;
said overflow checking wherein if the presence of an overflow flag
is determined, an output pulse is sent and a frequency counter is
incremented by one and then performing said cycle check; else said
cycle check is performed;
said cycle check wherein if a predetermined period has elapsed,
a new correction factor is calculated by:
selecting two known frequencies having the current value of said
frequency counter within the interval which is determined by said
known frequencies, and having calibration factors associated with
said known frequencies;
determining the relative displacement of the value of said input
counter within said interval, the relative displacement being defined
as the difference of the upper endpoint and the input counter which
is divided by the length of the interval;
calculating a new correction factor (K) using the equation:
with KA being the calibration factor associated with said known
frequency having a lower value, X being said relative displacement,
and KB being the calibration factor associated with said known frequency
having a higher value;
resetting said frequency counter to zero;
performing said input pulse checking; else said input pulse checking
is performed.
14. The method described in claim 13 wherein said predetermined
period is approximately one second.
15. The method as described in claim 13 further comprising an initialization
step performed previous to said repetitive steps wherein said initialization
step comprises setting said input pulse counter and said frequency
counter to zero and setting said correction factor to a calibration
factor associated with the lowest of said known frequencies.
Description BACKGROUND OF THE INVENTION
The field of the invention is that of meters for measuring liquid
flow. More specifically, the field is that of turbine meters for
fuel dispensers.
Various devices are known in the art for measuring liquid flow.
One type of meter which is commonly used in fuel dispensers is a
piston type of positive displacement meter. Piston meters require
a chamber to accept and then exhaust a liquid. Often the chamber
size needed to accommodate the piston meter is a limiting factor
for liquid dispensing device, particularly in fuel dispensing installations.
A turbine or rotor meter is another type of liquid flow meter.
The turbine meter has, in effect, a small fan or rotor placed within
a liquid line. The liquid flow then drives the fan with the fluid
flow pressure turning the fan. The rotational velocity of the fan
is roughly proportional to the rate at which liquid flows through
the meter. The rotational velocity is converted into a train of
pulses by a transducer. The pulse rate if therefore roughly proportional
to the flow rate.
Turbine meters, due to their small size, are well suited for fuel
dispensing devices. However, problems exist with the use of turbine
meters. Turbine meters exhibit non-linear behavior, because the
observed rotational velocity of the fan is not exactly proportional
to the amount of liquid passing through the meter for all rates
of fluid flow. At slow and fast flow rates, a turbine meter is therefore
often inaccurate. Fuel dispensing devices have little tolerance
for erroneous measures; thus to utilize a smaller turbine meter,
its measuring error must be corrected.
An example of a prior art instrument for error compensation in
a flow meter is found in U.S. Pat. No. 3965341 (Honey). Honey
describes a flow rate computer which corrects the measured flow
rates for liquids having a wide range of viscosities. The Honey
flow rate computer uses pre-programmed values taken from a "universal
flow meter curve" to correct observed flow rates for various
viscosities. It uses a curve which plots correction values against
frequency divided by viscosity.
Another instrument is described in U.S. Pat. No. 4306457 (Fukui).
The Fukui device uses a predetermined table matching rotations per
second with an error offset value. The circuit described in the
Fukui patent measures the number of rotations which occur during
a predetermined interval. It does not estimate the error reduction
factor for any point between the number of rotations listed on its
table (e.g. 2 2.5 3 etc.).
A third instrument for error compensation for a flow meter is described
in U.S. Pat. No. 4581946 (Kanayama). The Kanayama device performs
a form of linear interpolation derived from a predefined flow rate
curve. The flow rate curve plots an error coefficient against flow
rate. A linear approximation of the flow rate curve is generated
by a sum of the least squares analysis. In this manner, the Kanayama
device estimates the actual flow rate on the basis of the measured
flow rate from the flow meter. However, the Kanayama device uses
only one of a predetermined set of predefined compensation constants
to correct the measured flow rate.
What is therefore needed is a simple, yet accurate, linearization
apparatus which can be used with a turbine flow meter. Also what
is needed is a turbine flow meter which reliably determines the
flow rate from a set of output pulses which defines the frequency.
An additional need is for a flow meter which is adaptable to changing
conditions and to a variety of dispensers.
The prior art lacks a small rotory meter which gives accurate results.
Also lacking is a method for correcting non-linear output which
both accurately measures flow rates and robustly adapts to changing
conditions.
BRIEF SUMMARY OF THE INVENTION
The turbine flow meter of the present invention provides several
advantages over the prior art. By linearizing its output pulses,
the meter has greater accuracy at low and high flow rates. By utilizing
a rotary flow meter rather than a positive displacement meter, the
present invention lessens the space required to house the meter.
In addition, the ability to dynamically change calibration values
provides a more robust meter which can adapt to changing conditions
and maintain its accuracy. These and other advantages will become
apparent from the following disclosure of the present invention.
One object of the present invention is to provide a linearized
turbine flow meter. Another object of the present invention is to
provide a linearizing device which operates according to output
pulses generated by the turbine flow meter. Still another object
of the present invention is a flow meter which can be used with
a variety of dispensers.
The present invention is an apparatus for correcting the measured
flow rate of a turbine flow meter in a fuel dispensing apparatus.
The present invention includes a turbine flow meter which is adapted
to measure the flow rate of a liquid by the rotational velocity
of its rotor. Additionally coupled to the flow meter is an output
means which generates pulses at a rate proportional to the rotational
velocity of the turbine meter.
The present invention also includes a memory for storing calibration
values from the flow meter calibration curve and which is coupled
to a processor. The processor determines the exact correction factor
needed to adjust the measured pulse signals to a true reading. The
processor first determines in which interval the measured frequency
occurs, then calculates the relative displacement of the measured
frequency within that interval. The processor calculates a new correction
factor on the basis of the calibration values at the endpoints of
the interval and the displacement of the measured frequency within
the interval.
Thus, the present invention describes a method of linearizing the
output signal of a non-linearly responsive flow meter.
DESCRIPTION OF THE DRAWINGS
The above mentioned and other features and objects of this invention,
and the manner of attaining them, will become more apparent and
the invention itself will be better understood by reference to the
following description of an embodiment of the invention taken in
conjunction with the accompanying drawings, wherein:
FIG. 1 shows the fuel dispensing apparatus of the present invention.
FIG. 2 is a schematic organizational diagram of the linearizing
components of the present invention.
FIG. 3 is a sectional view of the turbine flow meter of the present
invention.
FIG. 4 is a schematic diagram of the linearizer of the present
invention.
FIG. 5a is a graph depicting the deviation curve with no linearization.
FIG. 5b is a graph depicting the deviation curve with prior art
linearization.
FIG. 5c is a graph depicting the deviation curve with the end point
linearization of the present invention.
FIG. 6 is a flow chart diagram of the linearization method of the
present invention.
Corresponding reference characters indicate corresponding parts
throughout the several views of the drawings.
The exemplifications set out herein illustrate a preferred embodiment
of the invention, in one form thereof, and such exemplifications
are not to be construed as limiting the scope of the disclosure
or the scope of the invention in any manner.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
FIG. 1 shows a representation of a fuel dispensing apparatus compatible
with the linearized turbine flow meter according to the present
invention. Fuel dispenser 2 is shown having three fuel dispensing
lines 5 and having a display 4. Display 4 is used to show the amount
of fuel which has been provided by the pump nozzles 3 of fuel dispenser
2. To dispense fuel from fuel dispenser 2 the user takes a pump
nozzle 3 and places it in a fuel receptacle. By squeezing the nozzle
trigger, the user begins the fuel flow. A pump of fuel dispenser
2 creates a flow in the fuel dispensing line 5. The display 4 typically
shows the amount of fuel dispensed and the cost thereof. The amount
of fuel is determined as a function of output pulses of a flow meter
within fuel dispenser 2. The cost is the product of the amount and
the fuel price. Often, the user varies the flow rate by manipulating
the nozzle trigger, typically as fast as possible initially then
relatively slowly when the fuel receptacle is almost full.
Lawful operation of a fuel dispensing apparatus requires that the
apparatus measures the liquid dispensed with particular accuracy.
To most efficiently accomplish that goal, a turbine flow meter is
used to minimize the space required for the flow meter inside of
the fuel dispenser 2. The present invention provides for a small
turbine meter to give accurate flow measurements for both fast and
slow flow rates. It also allows for reprogramming the calibration
equipment to adjust to the changing conditions of the fuel dispenser.
These features as well as others will become apparent in the following
description of the present invention.
Although the preferred embodiment of the present invention describes
a flow meter in a fuel dispensing device, the present invention
need not be limited to that embodiment. The present invention could
be a portable device used for situations requiring a small but accurate
flow meter. Also, the flow meter could be permanently installed
as in a fuel dispensing device pictured in FIG. 1. Thus, the present
invention is adaptable to a variety of applications.
FIG. 2 shows a schematic diagram of the flow measurement system
of the present invention. A fuel line 16 is used to dispense fuel.
Coupled with fuel line 16 is flow meter 10. Flow meter 10 is equipped
such that it provides output signals in proportion to the rotational
velocity of the fan or rotor of the turbine meter. Those output
pulses are sent along input pulses line 20.
Linearizer 34 receives the input pulses from the flow meter 10.
Within the linearizer, the input pulses are corrected using correction
factors to produce an accurate stream of output pulses. These output
pulses reflect the true volume of liquid dispensed through fuel
line 16. In addition to receiving input pulses from input pulses
line 20 linearizer 34 also sends output pulses, on linearized output
pulses line 22 to counter 6.
Counters like counter 6 are well-known in the art. The counter
6 receives a number of input pulses which it converts to a number
signifying the amount of liquid volume which has passed through
fuel line 16. Its first step is to receive pulses from linearized
output pulses line 22. Then counter 6 totals the pulses received
and generates a representation of the amount of liquid which has
been dispensed through fuel line 16. The counter 6 is coupled to
display 4 and drives display 4 to show the amount of liquid measured.
Display 4 can be any one of a number of devices well known in the
art for projecting graphic information.
The present invention provides a linearizing flow meter which produces
an adjusted set of pulses, allowing counter 6 to accurately determine
the amount of flow through fuel line 16. The flow meter 10 is depicted
in greater detail in FIG. 3. The linearizer 34 is depicted in greater
detail in FIG. 4.
The flow meter 10 depicted in FIG. 3 comprises a flow channel
18 a rotor 12 and a pickup assembly 14. The flow meter 10 is a
standard type of turbine flow meter, which directs a liquid line
through a flow channel 18 having a rotor or fan. The rotor 12 covers
the flow area such that any liquid flow causes the rotor to move.
The pickup assembly 14 is placed close to rotor 12 so that assembly
14 can sense the movement of rotor 12 but is out of the rotational
path of rotor 12 so as not to inhibit the rotor movement. Thus,
pickup assembly need not be physically attached to flow channel
18.
Liquid flow through the flow meter 10 turns the rotor 12. The pickup
assembly 14 senses the rotary movement of rotor 12 and emits pulses.
The output pulse frequency is proportional to the rotational velocity
observed on rotor 12. Typically, the pickup assembly 14 is comprised
of a magnetic device which senses each rotation of the rotor 12
and sends one or more pulses for each such rotation or numbers of
rotations.
A variety of shapes and sizes of turbine flow meters can be used
with the present invention. The flow meter must have a means to
measure the rotor's velocity. Also, an ability to prove the flow
meter is necessary for the practice of the present invention. Without
proving ability, a calibration curve cannot be derived.
The flow meter pickup assembly 14 sends pulses over raw input pulses
line 20 to linearizer 34. A schematic diagram of linearizer 34 is
shown in FIG. 4. A processor 26 receives input pulses via raw input
pulses line 20. The processor 26 then determines a linearized output
which it sends on linearized output pulses line 22. To accomplish
this, processor 26 is coupled to thumbwheel switches 24 random
access memory (RAM) 28 and read only memory (ROM) 30. Additionally,
thumbwheel switches 24 and RAM 28 are coupled. The processor 26
in the preferred embodiment, has machine instructions stored in
read only memory (ROM) 30. The processor also has a number of correction
values stored in random access memory (RAM) 28. Thumbwheel switches
24 operate to input new correction values in RAM 28 when enabled
by processor 26.
The preferred embodiment of the present invention uses a battery
backed-up random access (RAM) memory. However, the use of many types
of electronic memory is practicable with the present invention.
Such memories include read only memories (ROM), electronically programmable
read only memories (EPROM), magnetic bubble memory, erasable electronically
programmable read only memories (EEPROM) or any other non-volatile
memory. Volatile memories (e.g. ordinary RAMs) can be used, but
the meter would lose the correction values when a power interruption
occurs. However, programmable non-volatile memories allow dynamic
change of the correction values.
To set the endpoints and their associated calibration values, a
number of methods can be used. The preferred embodiment of the present
invention has six endpoint values implicitly defined by the program
stored in the ROM 30. The correction values associated with these
endpoint values are stored in the RAM 28. The six correction values
can be changed by using thumb wheel switches 24.
The preferred embodiment of the present invention has seven thumb
wheel switches with an additional switch or button. Six of the thumb
wheel switches specify the correction value to be entered. The seventh
thumb wheel switch specifies which of the six endpoints to associate
the correction value with. An additional feature involves the ability
to offset all six correction values by the amount specified on the
first six thumb wheel switches. This feature is useful when changing
conditions uniformly shift the calibration curve a certain amount
from its predefined values. Thus, by offsetting all by a uniform
amount, the six correction values can be corrected instantly. Altering
memory according to the seven thumb wheel switches occurs by depressing
a button which is located near the thumb wheel switches.
The preferred embodiment has thumb wheel switches 24 coupled both
to processor 26 the switches' driver, and RAM 28 the switches'
depository. A human operator can enter correction values for the
various interval endpoints on the thumb wheel switches 24 the correction
values to be stored in RAM 28. Because the thumb wheel switches
24 do not contain any microprocessing circuitry, the processor 26
must store the correction values to RAM 28.
This embodiment having ROM, RAM, and thumbwheel switches is only
one configuration of the present invention. Any one of a number
of data entry devices can be used within the teachings of the present
invention. To practice the present invention, it is only necessary
to have a processor having access to memory for storing the correction
values. However, having a microprogram in the ROM greatly facilitates
performance of the correct linearization calculations. The preferred
embodiment uses a 80C31 chip made by Intel of Santa Clara, Calif.
The appendix to this specification is a copy of micro code for the
processor of the preferred embodiment, compiled by ASM 51.
The linearization of the present invention is graphically depicted
in FIG. 5. The graphs of FIG. 5 show deviation as a function of
flow rate (measured by the frequency of pulses by pickup assembly
14). Deviation expresses the over- or under-estimation of the actual
amount of liquid to have passed through the flow meter. A deviation
of zero shows an exactly accurate measurement of the liquid flow.
A value of 1 indicates the measured flow rate was twice the actual
flow rate. Similarly, a deviation of negative one-half shows the
measured flow rate was one-half the actual flow rate.
Graph A of FIG. 5 shows the actual error of a typical turbine flow
meter. A typical deviation curve is shown as measurement error curve
40a. The flow meter detects no liquid flow until the flow rate is
approximately 3 gallons per minute (gpm). At 3 gpm, the flow meter
under-estimates the actual liquid flow by a significant degree,
thus numerically the deviation is a negative fraction. As the flow
rate increases from approximately 3 gpm to 40 gpm, the deviation
goes from under-estimation to over-estimation. From 40 gpm and above
the deviation stays at approximately the same positive fractional
value, i.e. over-estimating the flow rate.
Graph B of FIG. 5 showns a prior art method of linearization of
the curve depicted in Graph A. By segmenting measurement error line
40a into differing intervals and attempting to linearizer each interval,
the average deviation of measurement error 40b decreases compared
to the no linearization case of 40a. However, looking to Graph C
of FIG. 5 it is apparent that the endpoint linearization shown
as measurement error 40c further decreases the total deviation.
Thus the present invention produces a more accurate measurement
of the turbine meter flow rate. The method used to construct FIG.
5 will become apparent upon consideration of the endpoint linearization
and the flow chart of FIG. 6.
To linearize a deviation curve, the curve is divided into intervals
and then each interval is interpolated to produce a set of lines
which estimate the deviation curve. The intervals are determined
by an empirical method which selects portions of the deviation curve
that are relatively linear within the interval. Alternately, intervals
could be evenly spaced. Even spacing works well for large numbers
of small intervals. However, a large number of intervals requires
increased storage and computation time. Preferably, a small number
of intervals is chosen each having a reasonably linear curve within
each interval. Any suitable method of interval selection can be
used.
Within each interval, the deviation curve is interpolated to produce
a straight line estimating the deviation curve within the interval.
The present invention uses an endpoint linearization method. This
assumes that the correction factor for any measured frequency lies
on the line between the immediately lower and higher endpoints.
Other methods of linear interpolation, such as the least sum of
the squares method or an empirical method, can be used. However,
any change in calibration of the flow meter would require extensive
recalculation to adapt those other methods. For the endpoint linearization
method, only recalibration at the interval's endpoints is needed.
Also, if the selected endpoints are located at the usual operating
speed of the flow meter, the endpoint linearizating is both more
accurate and more robust. Thus, endpoint linearization exhibits
the preferred behavior. The endpoint linearization algorithm is
expressed in correction factor change section 72 of FIG. 6.
Linearization flow chart of the present invention is depicted in
FIG. 6. The flow chart 50 shows the algorithm that processor 26
executes when the fuel dispenser 2 dispenses fuel. The algorithm
operates upon four variable quantities, a set of frequency endpoints,
and correction factors associated with those endpoints. In the algorithm,
the four variables are Q, K, S, and X. Q represents the rate of
flow by counting the number of pulses per second. K represents the
current correction factor which is used to determine when an output
pulse should be sent. S represents the sum of the correction factors
and it never exceeds one, although in the process of summing correction
factors overflow can occur which causes an output pulse. X represents
the displacement within the selected interval of the latest measured
frequency. Also, the Nth frequency is denoted by F.sub.N and the
correction factor associated with that frequency is denoted by COR.sub.N.
The first step of the flow chart is initialization 52. In step
52 both Q and S are assigned the value of 0 because they are used
as counters. K is assigned the correction value associated with
the endpoint having the smallest frequency because K must have an
initial value for the algorithm to operate predictably. The correction
value associated with the lowest endpoint is selected as K's initial
value because typically the initial flow is relatively slow.
After initialization 52 the algorithm begins its input loop 70.
Input loop 70 is a busy-wait type of loop that cycles approximately
once per second to recalculate the value of the correction factor.
The first step of input loop 70 is input pulse check 42 where the
processor checks for the recent occurrence of an output pulse from
flow meter 10. If no recent output pulse has occurred, the process
proceeds to cycle check 49 described below. If a recent output
pulse has occurred, the processor proceeds to correction summing
step 44.
In step 44 the value of S is incremented by the value of K. That
incrementing causes overflow if the sum would be greater than or
equal to one, although the value of S never exceeds one. After correction
summing 44 the processor checks for overflow in overflow check
step 45. If no overflow occurred in step 44 then the processor
proceeds to cycle check 49 described below. If overflow occurred,
then outputs pulse step 46 is executed.
Overflow occurs whenever the value of S+K equals or exceeds 1.
For example, assuming S=0.743 and K=0.674 the actual value of S+K
is 1.417. However in step 44 S is assigned the value of 0.417 and
an overflow flag is set. In step 45 the processor checks the overflow
flag to determine if sending an output pulse is appropriate. If
the overflow flag is set, then the processor clears the flag and
proceeds to execute step 46. However, upon finding the overflow
flag clear, the processor branches to cycle check step 49.
The processor sends an output pulse in output pulse step 46. Next,
is pulse counting step 48 where the variable Q is incremented by
one. The final step in input loop 70 is the previously mentioned
cycle check step 49 where the processor determines whether a second
has transpired since the start of the current iteration of input
loop 70. If a second has not transpired, then the input loop 70
restarts by execution of input pulse check 42. If a second has transpired
then the processor executes correction factor change section 72.
The correction factor change section 72 is executed approximately
once per second. Its purpose is to select the appropriate correction
factor to use in the next iteration of input loop 70. The procedure
shown in correction factor change section 72 is one embodiment of
the endpoint linearization method, with other equivalent methods
also existing. The first step of section 72 is interval selection
step 58.
In step 58 the estimated frequency is used to select the appropriate
interval for calculating the new correction factor. The interval
is selected such that the estimated frequency Q (for Q pulses per
second) lies directly between a lower endpoint and a higher endpoint,
both of which have an associated correction value. The lowest interval
is the fringe defined as any estimated frequency lower than the
lowest endpoint. Also, a second fringe interval is defined by the
highest endpoint and beyond. Typically, the estimated frequency
Q falls between two defined endpoints. In fact, intervals should
be selected such that normally occurring frequencies do not lie
within either fringe interval.
After selecting the appropriate interval, the Q value's displacement
within that interval is determined. In displacement calculation
step 60 the variable X is assigned a value which gives the estimated
frequency's percentage displacement within the selected interval.
The values of X range between zero and one. The displacement value
is calculated by subtracting the estimated frequency Q from the
higher endpoint then dividing the difference by the distance between
the two endpoints (that distance equaling the frequency of the higher
endpoint minus the frequency of the lower endpoint). The processor
uses the displacement value to calculate a new correction factor.
The processor computes the new correction factor in correction
calculation step 62. The new correction factor K is defined as the
sum of the correction factor of the lower endpoint value and the
displacement factor. Specifically, the displacement factor is derived
by subtracting the correction value of the lower endpoint from the
correction value associated with the higher endpoint and multiplying
that difference by the displacement value X. The new value of K
is derived by subtracting the displacement factor from the correction
value associated with the higher endpoint value. In equation form:
with COR.sub.A representing the correction value associated with
the lower endpoint and COR.sub.B representing the correction value
associated with the higher endpoint. In the preferred embodiment
of the present invention, the calculation of K's value is done differently
depending on which endpoint has a higher correction value associated
with it (see Appendix A, comments on bottom of page 6). The new
correction factor K's value indicates the error of the measured
frequency.
Lastly, in correction factor change section 72 pulse reset step
73 is executed. Step 73 assigns Q the value of 0 thus resetting
counter Q for another iteration of input loop 72. Then the processor
cycles to step 42 restarting input loop 70. The variable S is not
reset to 0 because it indicates a fractional pulse which should
be accounted for.
In the preferred embodiment, the processor performs the input loop
70 of the linearization flow chart once every second. However, it
would be obvious to one skilled in the art that a different sampling
rate could be used and still be within the teachings of the present
invention. By the foregoing method, the processor is able to correct
the stream of input pulses and send the counter linearized output
pulses.
In the preferred embodiment, the intervals are defined by the standard
behavior of the flow meter type which a particular implementation
is designed for. Thus, a particular flow meter is proven, and from
the tests a deviation curve is determined. The deviation curve is
then broken into intervals. The selection of interval endpoints
depends on the specific operating characteristics of the flow meter.
Endpoints are chosen such that the lines defined by those endpoint's
deviations most closely match the deviation curve between those
endpoints. The specific endpoint selection method chosen is influenced
by the number of endpoints which can be utilized by the processor
and the shape of the flow meter's deviation curve. Another possible
embodiment would have both the calibration values and the endpoint
values dynamically stored. However, for most standardized applications,
having predetermined endpoint values achieves the desired results.
While this invention has been described as having a preferred design
it will be understood that it is capable of further modification.
This application is therefore intended to cover any variations,
uses, or adaptations of the invention following the general principles
thereof and including such departures from the present disclosure
as come within known or customary practice in the art to which this
invention pertains and fall within the limits of the appended claims. |