Eugene
Miles, Garnett
John
Everett, Gray
Robert
James, Dragnea
Gabriel
| 1. | A system calibration process for a differential pressure flow measuring system for fluid flowing in a confined conduit, where such system comprises a primary flow sensor and a secondary electromechanical conversion apparatus, including electronic memory, for converting high and low fluid pressure signals derived from the primary flow sensor to an electric signal output representative of fluid flow, comprising the steps of; calibrating the secondary apparatus against known pressures over the operating range of the apparatus to obtain initial output characterization data and storing the characterization data in the electronic memory, sensing the fluid impact pressure and the fluid static low pressure for a plurality of known reference fluid flows to establish a plurality of high and low pressure signals over a range of flows, with the use of the initial output characterization data, transducing the said high and low pressure signals into measured electrical differential pressure signals representative of the difference between the sensed high and low pressure for each of the plurality of reference flow rates, determining a correction factor for the output of the secondary apparatus for each of the plurality of reference flow rates as a function of the deviation of each value of measured differential pressure from the nominal differential pressure corresponding to each of the said plurality of reference flow rates, modifying the data from the initially characterized secondary apparatus by overwriting! in the electronic memory, the initial characterization data with the calculated correction values. |
| 2. | The method of claim 1 wherein the step of determining correction factors for the output of the secondary apparatus for each of the plurality of reference flow rates includes the step of determining the nominal differential pressure output of the system by defining the nominal flow coefficient of the primary flow sensor. |
| 3. | The method of claim 2 and further including the step of determining the average actual flow coefficient value for the plurality of reference flows. |
| 4. | The method of claim 3 wherein the step of determining correction factors for the output of the secondary apparatus for each of the plurality of reference flow rates includes the step of correcting the measured differential pressure output values for the deviation of the actual flow coefficient from the nominal flow coefficient. |
| 5. | The method of claim 4 wherein the step of determining a correction factor for the output of the secondary apparatus for each of the plurality of 5 0 reference flow rates includes the step of subjecting the values of the measured differential pressure and the determined nominal differential pressure values to a statistical regression analysis to establish the correction factors. |
| 6. | For a differential pressure flow measuring system comprising a primary flow sensor and a secondary electromechanical conversion apparatus for converting high and low fluid pressure signals from the primary flow sensor to an electronic signal output representative of fluid flow, a system calibration process, comprising the steps of; applying a plurality of known reference fluid flow rates to the primary flow sensor over the system's operative range of flow conditions, sensing the electrical output of the secondary electronic conversion apparatus for each of the said plurality of reference fluid flow rates; establishing a correction factor for each of the plurality of reference flow rates which will tend to characterize a desired relationship between the fluid flow rate and the electrical output of the measuring system, and storing the correction factors in nonvolatile memory. |
| 7. | The method of claim 6 and further including the step of initially linearizing the secondary electromechanical apparatus against known pressures ex I over the operating range of the apparatus to obtain data for linearizing the output of the said electromechanical apparatus. |
| 8. | The method of claim 7 wherein the step of establishing a correction factor for each of the plurality of reference flow rates does so as a function of the deviation of each value of measured differential pressure from the nominal differential pressure corresponding to each of the plurality of reference flow rates. |
| 9. | The method of claim 8 and further including the step of modifying the initial linearization data for the secondary electromechanical apparatus with the determined correction factors. J l . |
| 10. | A system calibration process for differential pressure flow measurement in a system having fluid flowing in a confined conduit, and having a primary flow sensor and a secondary electromechanical conversion apparatus, including electronic memory, for converting high and low fluid pressure signals derived from the primary flow sensor to an electric signal representative of fluid flow, comprising the steps of; sensing the fluid impact pressure and the fluid static low pressure for a plurality of known reference fluid flows to establish a plurality of high and low pressure signals over a range of fluid flows, transducing the said high and low pressure signals into measured electrical differential pressure signals representative of the difference between the sensed high and low pressure for each of the plurality of reference flow rates, establishing a desired relationship between the reference flow rates and the electrical output of the secondary apparatus for each of the plurality of the reference flow rates, overwriting into the electronic memory the said established desired relationship of the secondary apparatus. |
| 11. | The method of claim 10 and further including the step of initially characterizing the secondary electromechanical apparatus against known pressures over the operating range of the apparatus to obtain data for characterizing the output of the said electromechanical apparatus. |
| 12. | The method of claim 1 1 wherein the step of establishing a desired relationship between the reference flow rates and the electrical output of the secondary apparatus for each of the plurality of the reference flow rates does so as a function of the deviation of each value of measured differential pressure from the nominal differential pressure corresponding to each of the plurality of reference flow rates. |
| 13. | The method of claim 12 and further including the step of modifying the data obtained by initially characterizing the secondary electromechanical apparatus with the determined correction factors. |
| 14. | The method of claim 13 and further including the step of storing the correction factors in nonvolatile memory. |
| 15. | The method of claim 10 where the step of establishing a desired relationship includes that of linearization. <?. |
The present invention relates to differential pressure based fluid flow
measurement, including rate, volume and mass, and more specifically to a
method for improving the accuracy of measurement by calibrating the system
as a whole, over its operating range, with known fluid flow references.
BACKGROUND OF THE INVENTION
Differential pressure based fluid flow meters, or measurement systems,
for fluid filling a confined channel, such as a pipe, typically include two basic
elements. A system comprises a fluid immersed sensor, as the primary element
of the system, and a secondary element comprising mechanical and electronic
means to convert the sensed fluid dynamics into a format usable for obtaining
needed information, such as rate of flow, volume of flow and mass of flowing
fluid.
Primary sensors which detect differential fluid pressure in the flow
include such specific types as averaging pitot tubes, orifice plates, flow nozzles
and venturi meters.
The secondary portion of the system may include a differential pressure
(DP) transducer, a static pressure transducer, a fluid temperature transducer and
a flow computer. The raw electrical output of these transducers is typically
processed into a signal which is transmitted for subsequent electronic
processing in additional apparatus, such as a flow computer. The combination
of the transducer and the subsequent electronic means, such as A/D converters,
amplifiers, and a central processing unit is often referred to in the trade as a DP
transmitter and that term will also be used in this specification. Often the
primary sensor and the secondary, or DP transmitter, are physically widely
separated and are commonly treated as separate instrumentation entities,
meaning that the two elements are calibrated separately and present their own
individual accuracy standards.
Preliminary to an understanding of the current state of the art of fluid
flow measurement, and the accuracy attainable with available systems, vis-a-vis
the present invention, it is useful to review the alternative methods of rating the
accuracy of measurement devices.
There are two methods of expressing accuracy. The first is usually
expressed as "percent of reading" or "percent of value." This method identifies
the error at a specific reading. In measurement devices the use of this method
implies that the device produces an error statement based on its readings over
the entire operating range. The second method is expressed as the "percent of
full scale" (percent of fs) accuracy standard. This method relates the error in a
device when it is measuring a quantity that represents 100% of the output. The
equivalent "percent of value" error figure for a device calibrated with a
"percent of full scale" accuracy statement is calculated using the following
equation:
VoFullScaleError % Reading Error = — — — - X 100 equation # 1
5 %oβcale
The differences between these two accuracy ranking methods are not
obvious until they are plotted. Figure 1 shows the results of each of the two
methods for expressing measurement error. Curve "A" shows a 1% of reading
error while curve "B" shows a 1% of full scale error statement. Both are
plotted over a 10: 1 turndown. In the example shown, a "±1% of full scale"
error is equivalent to a "±5% of value" error at a 20% of full range level (5: 1
turndown). Flow turndown is the ratio of the highest flow rate expected to be
measured by the system to the lowest flow rate expected to be measured, at
some stated accuracy. This quantity is typically expressed on one line with a
colon, such as 10: 1 for a turndown ratio of 10. Most flow measurement devices
have a minimum flow and a maximum flow that can be measured within stated
accuracy limits. A typical flow rate falls between these values.
In differential pressure measurement devices, flow rate Q is derived in
accordance with the formula Q = C yfOP , where C is a constant and DP is the
difference between the sensed high and low fluid pressures. Different types of
differential pressure sensors have their own benefits and drawbacks, but all of
them suffer from an inherent inability to provide fluid pressure outputs which
are absolutely related to a fluid flow rate in accordance with the stated
jnathematical formula over a wide operating range. The induced errors and
ultimate performance characteristics of these differential pressure flow sensors
are defined by and vary with the installation parameters of each device, such as
the shape and dimension of the sensor, location and number of its plenum
apertures and the sensor's position inside the fluid carrying pipe. Figure 3 is a
diagrammatic curve illustrating in solid line the mathematically ideal
relationship between fluid flow Q and differential pressure DP. The dotted line
curve illustrates the same relationship but with the inherent errors of a primary
sensing element taken into consideration. An averaging pitot tube type of
primary sensor or other DP sensor can be made to exhibit a close similarity to
the stated mathematical ideal over a wide flow range, but even those types of
sensors cannot offer much better than ±1% of reading over a 10: 1 turndown.
In addition to the error introduced by the primary, the components of the
secondary of a flow measurement system are each responsible for introducing
error into the system. Although error is contributed by all of the secondary
components, a large portion of it is induced by the non-linearity of the
transducers. Linearity, in this context, is the ability of the components to
produce, or closely approximate, a linear relationship between the actual
physical input and the DP transmitter output. Other components of the
secondary also contribute to non-linearity, resulting in secondary error. The
solid line curve of Figure 4 illustrates ideal linearity of the DP transmitter,
according to the formula: m.a. = B x DP, where m.a. is the secondary output in
milliamperes, B is a constant and DP is the differential fluid pressure. The
dotted line curve of Figure 4 diagrammatically illustrates the non-linear
performance of the secondary of the flow meter (without reference to the static
pressure or temperature of the fluid) according to the formula: m.a. = B x DP +
E S (DP) where E s is the % error in differential pressure due to the secondary.
While there may be several different representations of linearity, the term is
assumed to represent the independent linearity of the secondary of the flow
meter. A straight line is used in Figure 5 to minimize the maximum deviation
of the actual characteristic. The graph of this Figure has been normalized to
show linearity as a percent of reading error deviation from the ideal zero error
curve, which would be a straight horizontal line. Current flow measurement
practice recognizes the non-linearity of the secondary and attempts to
compensate for it by "characterizing" the DP transmitter. This is most often
done by employing dead weights, or some other form of calibrated DP source,
to simulate various pressures on the transducer diaphragm and then modifying
the electrical output to correct for the observed discrepancies. In this type of
process however, neither actual fluid flow nor the aberrant characteristics of the
primary sensor are given consideration in the linearizing attempt.
Accordingly, when independently calibrated primary and secondary
elements are combined, the resulting measurement system accuracy must be
determined by combining the error of the primary and secondary elements over
the selected range of operation in accordance with the "square root of the sum
of the squares" rule:
% System Error = JEpl + Es equation Wl
where: Ep = % Error in flow due to the primary; and
Eg = % Error in flow due to the secondary, i.e.
where :EDP = Error due to the differential pressure transducer, Ep = Error due to the static pressure transducer, and y = Error due to the temperature transducer.
The primary element's error contributes directly to the error in flow,
whereas the secondary device contributes to error in the differential pressure
(DP), which is the square of the error in flow. The percent error in DP will
depend upon the percent scale, as shown in Figure 1. For a DP transmitter or
meter, the contribution to the secondary's flow calculation error is:
% EDP - 1 X 100 equation #3
where: %Efs = the % of full scale error (accuracy) of the DP transmitter;
%DP = the % of scale at which the DP transmitter is operating.
Ep and E are similarly calculated.
To determine the error in flow at the minimum scale for the desired flow
turndown, equation 3 must be substituted into equation 2. To better illustrate
the point, reference is made to Figure 2 where the reading flow system accuracy
is plotted as a function of turndown for different types of flow meters. An
example of a flow meter with a % of reading accuracy, where there is no
secondary element contributing to the system error, is shown in solid lines. The
error of an orifice plate or other type of DP sensor working with a DP
transmitter is calculated using equations 2 and 3, above, and is plotted in the
graph of Figure 2 with dotted lines. In this exemplary curve the orifice plate
C
sensor itself contributes a ±1% of reading error (Ep) and the secondary element
has a ±0.1 % full scale (fs) error (EJJ).
After deriving the dotted line curve of Figure 2 , the problem of the prior
art custom of separately calibrating the primary and secondary flow meter
elements is apparent. The errors of the separate elements are additive within
the flow system, creating a combination where the total error can never be less
than the least accurate component of the system.
It is therefore the primary object of the invention to provide a synergistic
method of obtaining higher measurement system accuracy by calibrating the
assembled combination of primary and secondary system elements with actual
reference fluid flow.
A further object of the invention is to provide a method of measurement
system calibration that results in unexpectedly better accuracy than that
obtained by the square root of the sum of the squares of the individually
calibrated system elements.
Another object of the invention is to provide a method for measurement
system calibration which results in a wider flow turndown with higher
accuracy.
Another object of the invention is provide a process for calibrating a
flow measurement system where the flexibility of the secondary element's
linearization potential is advantageously employed to correct for errors in the
?
flow primary, as well as to correct for the nonlinearity in the secondary's own
components.
Other and still further objects, features and advantages of the invention
will become apparent upon a reading of the detailed description of a preferred
form of the invention.
The most pertinent known prior art is listed as follows:
Dieterich Standard Annubar® Flow Measurement Systems brochure,
where, on pages 2 and 3 (unnumbered), a flow measurement system of the type
described in this specification is shown, and on page 9, pressure transmitters
and their accuracy are discussed. This measurement system shown does not
contemplate the calibration and linearization method of the present invention.
Dieterich Standard Diamond II Annubar® Flow sensors catalog where,
on page 3 a diagram of a flow measuring system is shown which is similar to
the general system referred to in this specification, but without the linearization
method of the present invention, and on page 7 thereof accuracy of the sensor
primary is discussed.
U.S. Patent No. 5,111,827 to Borje Rantala, disclosing, in a respiratory
sampling device, a microprocessor to adjust a flow rate in accordance with the
composition of a gas, as determined by a gas analyzer.
U.S. Patent No 4,836,019 to Floyd W. Hagen, disclosing an air data
sensor having a computer programmed to take various compensation tables
from calibration curves for various sensors. The computer is programmed to
provide an appropriate correction factor to the pressure outputs.
U.S. Patent No. 2,869,367 to D.W. Moore, describing a system utilizing
pressure responsive diaphragms to change electrical resistances in order to
linearize the response characteristic of a system.
THE DRAWINGS
Figure 1 is a graph showing the relationship between "percent of
reading" and "percent of full scale" systems of defining measurement accuracy.
Figure 2 is a graph showing the percent of the reading flow system
accuracy as a function of turndown for two different types of fluid flow meters.
Figure 3 is a diagrammatic curve illustrating in solid line the
mathematically ideal relationship between fluid flow Q and differential
pressure DP. The dotted line illustrates the error of primary flow meter
measurements.
Figure 4 is a graph illustrative of a typical characteristic curve of the
output of a differential pressure transmitter (secondary) as related to differential
pressure. The solid line represents a linear relationship between differential
pressure and electrical system output. The dashed line is exemplary of an
°l
actual curve, taking into consideration the error E§ introduced by non-linearity
in the secondary.
Figure 5 is a normalized curve showing system output linearity as a
percent of reading error deviation from an ideal linear curve which would be
represented by the straight horizontal line with zero error.
Figure 6 is a block diagram of a typical DP fluid flow rate measurement
system adapted to use the calibration method of the present invention.
Figure 7 is a diagrammatic block flow diagram of the initial calibration
of the DP transmitter used in a system shown in Figure 6.
Figure 8 is an exemplary table showing percentage values for differential
pressure taken in ten percent increments over the operating range of a DP
transmitter where URL is the Upper Range Limit of the transmitter measured in
inches of water column. The table shows corresponding values of the initial
transmitter characterization, done as shown in Figure 7 (Calib. DP) and the
Corrected DP (DP corT ected)> shown in Figure 6.
Figure 9 is a block flow diagram of the calibration method of the present
invention, with exemplary tables of values taken from indicated points in the
process.
Figure 10 is a graph showing the relationship between Dp measure j ? as a
percentage of the upper range limit (URL) and the normalized DP, or DP', as a
percentage of URL
SUMMARY OF THE INVENTION
The method of the present invention calibrates and characterizes a
differential pressure flow measurement system as a unitary assembly, using an
actual calibrated flow on the primary sensor. By calibrating the system as a
whole, with actual reference fluid flow, instead of relying on the calibration of
the primary and secondary elements individually, an unexpected and synergistic
effect is produced whereby the system accuracy and turndown is better than
that obtained by the square root of the sum of the squares of the individually
calibrated system elements.
DETAILED DESCRIPTION
Before addressing the details of the calibration and characterization
method of the present invention, a brief description will be made of a flow
measurement system in which the calibration method would be used in order to
establish the background of the invention.
Figure 6 is a block diagram of a typical differential pressure dependent
low measurement system of the type to which the calibration method of the
present invention is particularly suited. In this measurement system a pitot tube
type of differential pressure flow sensor 4 functions as the primary element,
/ /
however traditional differential pressure devices such as an orifice plate,
venturi tube or flow nozzle can also act as the differential pressure sensing
device. The pitot tube shown is of the type described in U.S. Patent No.
4, 154,100 to James B. Harbaugh et al. and U.S. Patent No. 4,559,836 to Darrel
5 F. Coleman et ai, to which patents reference is made for a more complete
explanation of the DP flow sensor 4 of Figure 6.
The upstream facing side of the pitot tube 4 senses the average sensor
impact pressure of flowing fluid to establish a high pressure value. The
downstream facing side of the pitot tube senses low pressure. The high and low
lϋ fluid pressures are conducted from the plenums 6 and 8 of the flow sensor 4 to
a pressure transducer 10, the first stage of the differential pressure transmitter
1 1. The pressure transducer transforms the respective high and low fluid
pressures issuing out of the primary sensor into an electrical signal whose
character is a function of the differential pressure (DP), that is the difference
15 between the sensed high and low fluid pressures. A typical transducer would
be one equipped with a sensing element comprising a silicon diaphragm, into
the surfaces of which are diffused piezo resistors which comprise an electrical
bridge whose output is analogous to the differential pressure applied to the
transducer. Other types of DP transmitters may be used, such as capacitance or
20 variable reluctance. The electrical analog signal from the pressure transducer
10 is transformed in the A/D converter 12 to a digital signal for input into a
central processing unit (CPU) 16. The CPU 16 performs the square root
function as well as other calculations, including the system error corrections
required for producing an accurate final output signal of the flow measurement
system.
After calibration, in accordance with the method of the present
invention, and during operation in actual flow measuring conditions, the
electronic signal, representing the fluid differential pressure, addresses stored
calibration information in an EEPROM memory 18, which information may be
in the form of a look-up table or a polynomial equation 19, to provide collective
corrections to the errors in that signal which occur as a result of non-linearity in
the transmitter 1 1 and aberrations in the operation of the primary sensor 4.
Obtaining the information to be stored in the EEPROM to implement these
corrections is the result of the calibration and characterization process of the
present invention and will subsequently be described in detail. The corrected
DP signal is further processed by the CPU in the step labeled 27 to determine
the square root of the corrected DP. That value is used to determine the flow Q
which is then read out in an appropriate instrument 30 or applied to some
further process or instrumentation.
Having generally described the function of the flow meter in actual
operation, the method of arriving at the calibration coefficients which are stored
in the EEPROM will now be described. As mentioned earlier, both the primary
sensor and the secondary element contribute error and each adversely effects
the accuracy of the system. By providing correction for both of these sources
/J
of error at the same time the final flow system signal will have greater accuracy
than it would otherwise have by calibrating the primary and secondary elements
individually.
With respect to an understanding of the error contributed by the primary
sensor, refer again to Figure 3. As for the secondary, the DP transmitter's non¬
linear output is initially linearized against a calibrated DP source, such as a
dead weight tester. See Figure 7 for a diagrammatic illustration of an initial
linearization process of the DP transmitter by itself. At this point it should be
noted that linearization is a form of calibration that is used when the desired
output is a linear function of the input of a device. There are situations,
however when the desired output is exponential or a polynomial function or
some other function of the input. Therefore, to generically define the
calibration method of the present invention, which covers all such functions,
the terms "characterizing" or "characterization" may be used to broadly include
linearization, as well as the others mentioned.
The initial linearization process characterizes the transmitter's output
from the known differential forces supplied by a calibrated DP source. As a
result of this initial characterization, the transmitter output will be referred to as
DPcalibrated. As an example of the result of such initial calibration, reference
is made to Figure 8 where DP ca ijt>rated ΪS shown for increments of differential
pressure values from zero to 100% of the upper range limit (URL) of the
transmitter, as supplied by the calibrated DP source. As part of the initial
H
calibration process, these DP ca j values are stored in the non- volatile memory
of the transmitter's CPU. See Figure 6 for a showing of the central processing
unit.
However, in accordance with the objects of the present invention, it is
not enough to linearize the DP transmitter alone. The objects of the invention
are accomplished by characterizing the system DP output against the
transmitter's DP ca iJb ra ted output, plus the flow coefficient, K, for the primary
sensor, that is: DP ca ι + K — > DP corre cted.
In other words, the initial characterization of the DP transmitter is
further enhanced in the context of the total measurement system by taking into
consideration in the calibration, the K value variation of the primary sensor.
This is accomplished by storing one or more correction coefficients in the
EEPROM memory 18 and using those coefficients to characterize the
measurement system as an integral whole.
Reference to Figure 9 will facilitate an understanding of how the
correction coefficients 21 which are stored in the EEPROM are derived. The
numbers in the tables which are referenced to different portions of the diagram
are exemplary only, for the purpose of better understanding the process. First,
the K values for a chosen different number of flows are determined. The
αumber of fluid flow values chosen for the calibration process is a judgment
call. For purposes of this description, eleven will be chosen, covering the
operating range of the system in ten percent increments, starting with zero.
Although reference fluid flows may be obtained through a number of different
techniques, a preferred form involves the use of a highly accurate laboratory
weigh tank, which determines mass flow directly. A number of different
chosen reference flows are directed into a pipe having a primary sensor, such as
the averaging pitot tube 4 of Figure 6. The sensed high and low pressures are
directed to the already initially calibrated DP transmitter, from which a
differential pressure signal DP measure( j (DP^) is obtained. This signal, along
with an input signal Q, representing a calibrated reference fluid flow from the
weigh tank calculates the actual value K of the primary sensor, K ac t, according
to the formula: K ac t = — -^ — where A is a coefficient to include pipe
AVDPM diameter, pressure, temperature, etc. From the K ac t value for each of the
chosen flow rates an average K (K av g) is determined by known statistical
analysis methods. In the next step, a nominal DP or DP' is determined
according to the formula x DPVJ = DP . Following determination of
the nominal DP for each of the chosen flow rates, a statistical regression
analysis is used to arrive at one or more correction coefficients for the system
which are then stored in the EEPROM. Preferably, the regression uses the least
squares method to fit a line through a set of points, each point representing a
respective nominal DP value for one of the measured DP values. The
dependent variable in the analysis is the measured DP (DPj j), while the
independent variable is the nominal DP, DP'. In accordance with well known
Ik
techniques the regression analysis can be expanded for increased accuracy by
using additional independent variables, DP' 2 , DP' 3 , DP' 4 and DP' 5 . The
regression analysis yields the X Variables, shown in the table of Figure 9, with
their exemplary coefficient values. These are the coefficients which need to be
applied to the initial calibration values stored into the transmitter's EEPROM to
determine the new calibration values:
DP cor = Intercept + X V arl(DPcal) + Xvar2(DP C al) 2 + • • ■•Xvar5(DP C al) 5
where: X V arn i sme regression coefficient for the nth independent
variable;
intercept is the point where the straight line crosses the Y axis.
Depending on the desired accuracy, 2, 3 or 6 independent variables
could be used, instead of the five shown in this example.
When the memory is addressed in actual system operation the final
corrected DP value (DP cor ) is used to determine the correct flow output.
Calibration is carried out over the full flow range of the measurement
system, establishing an accuracy of at least ±0.5% over the full calibrated flow
range. The calculated correction values (corrected DP's) replace the
corrections in the EEPROM which were established during the initial
characterization of the transmitter, as shown in Figure 7. Characterizing the
flow meter with this method of calibration permits the electrical output 27 of
the meter to track the true reference flow, thereby compensating for all errors
due to non-linearity, and the flow coefficient of the primary system, as shown
in Figure 10.
/<?