Abstrict A method of obtaining a measure of flow from an electromagnetic
flow meter is disclosed. A shaped excitation waveform is applied
to the flow meter and an output is received from the meter. The
sampled output data from the meter is then resolved into a non-flow
waveform component and a flow waveform component to derive a measure
of flow. Resolving the data into expected waveform components rather
than correlating each frequency component of the sampled output
data with a sine wave may allow readings to be obtained without
processing each frequency component separately. Also, the method
may allow more accurate removal of the non-flow component from the
output signal, and may decrease the time delay before the initial
reading and allow the use of a greater reading repetition rate.
Claims What is claimed is:
1. A method of obtaining a measure of flow from an electromagnetic
flow meter, comprising: applying a shaped excitation waveform to
a coil of an electromagnetic flow meter; receiving and sampling
an output signal from the meter in response to the application of
the shaped excitation waveform; and resolving sampled output data
from the meter into a non-flow waveform component and a flow waveform
component so as to derive a measure of flow.
2. A method according to claim 1 wherein the shaped excitation
waveform comprises a pulsed waveform.
3. A method according to claim 1 wherein the shaped excitation
waveform includes a first pulse rise section of a first polarity.
4. A method according to claim 3 wherein the shaped excitation
waveform includes a second pulse rise section.
5. A method according to claim 4 wherein the second pulse rise,
substantially constant, and pulse decay sections are of the opposite
polarity to the first pulse rise section, substantially constant,
and pulse decay sections.
6. A method according to claim 3 wherein the shaped excitation
waveform includes a second substantially constant section.
7. A method according to claim 3 wherein the shaped excitation
waveform includes a second pulse decay section.
8. A method according to claim 1 wherein the shaped excitation
waveform includes a substantially constant section of a first polarity.
9. A method according to claim 1 wherein the shaped excitation
waveform includes a pulse decay section.
10. A method according to claim 1 wherein the shaped excitation
waveform comprises at least one substantially constant section.
11. A method according to claim 10 wherein the constant section
of the shaped excitation waveform is used to determine background
interference.
12. A method according to claim 11 wherein the background interference
is subtracted from the sampled output data.
13. A method according to claim 1 wherein the shaped excitation
waveform comprises a plurality of pulse rise sections, substantially
constant sections, and pulse decay sections.
14. A method according to claim 13 wherein alternate substantially
constant sections are of alternate opposing polarities.
15. A method according to claim 1 wherein the shaped excitation
waveform is a substantially square wave waveform.
16. A method according to claim 15 wherein alternate pulses of
the square wave are of opposing polarities.
17. A method according to claim 1 wherein resolving comprises
performing a weighted least squares fit between the sampled output
data and a model of the expected waveform.
18. A method according to claim 1 wherein the non-flow waveform
component is represented in the form z.Vz(t).
19. A method according to claim 18 wherein Vz(t) is based on the
differential of the current input to produce the excitation waveform
(dI/dt).
20. A method according to claim 18 wherein Vz(t) is based on a
zero signal captured empirically.
21. A method according to claim 20 wherein Vz(t) is based on an
average of a plurality of captured zero signals.
22. A method according to claim 18 wherein z is a scalar multiplier
term, and wherein resolving includes determining the value of z.
23. A method according to claim 1 wherein the flow waveform component
is represented in the form f.Vf(t).
24. A method according to claim 23 wherein Vf(t) is determined
empirically.
25. A method according to claim 24 wherein Vf(t) is determined
empirically and is based on an average of a plurality of captured
flow signals.
26. A method according to claim 23 wherein Vf(t) is modelled by
a mathematical function.
27. A method according to claim 23 wherein f is a scalar multiplier
term, and wherein resolving comprises determining the value of the
scalar multiplier term f.
28. A method according to claim 1 further comprising determining
a measure of a trend within the sampled output data.
29. A method according to claim 28 wherein the measure of a trend
is subtracted from the sampled output data.
30. A method according to claim 1 wherein correlating is performed
over a window that does not contain an integer number of periods
of all frequency components.
31. A method according to claim 30 wherein the window is shorter
than the period of the lowest frequency component.
32. A method of determining a measure of a non-flow signal from
an electromagnetic flow meter comprising: applying a shaped excitation
waveform to an electromagnetic flow meter; receiving an output from
the meter in response to application of the shaped excitation waveform;
resolving sampled output data from the meter into a non-flow waveform
component and a flow waveform component; and determining a measure
of the non-flow waveform component based at least in part on the
resolved sampled output data.
33. A method according to claim 32 further comprising determining
a calibration measurement for the meter based on the measure of
the non-flow component.
34. A method according to claim 33 further comprising applying
a correction to flow measurements determined by the flow meter based
on the calibration measurement determined.
35. A method according to claim 32 further comprising detecting
a fault condition for the meter based on the measure of the non-flow
component.
36. A method according to claim 32 wherein the shaped excitation
waveform is a pulsed waveform.
37. A method according to claim 32 wherein the non-flow waveform
component is represented in the form z.Vz(t), and wherein Vz(t)
is a function of the current input to form the excitation waveform
and z is a scalar multiplier term.
38. A method according to claim 32 wherein the flow waveform component
is represented in the form f.Vf(t), wherein Vf(t) is a function
of the flow rate of a fluid through the flow meter, and wherein
f is a scalar multiplier term.
39. An apparatus for obtaining a measure of flow from an electromagnetic
flow meter comprising: means for applying a shaped excitation waveform
to the coil of an electromagnetic flow meter; means for receiving
an output from the meter; and means for processing the output from
the meter to resolve sampled output data from the meter into a non-flow
waveform component and a flow waveform component so as to derive
a measure of flow.
40. An apparatus for determining a measure of a non-flow signal
from an electromagnetic flow meter comprising: means for applying
a shaped excitation waveform to an electromagnetic flow meter; means
for receiving an output from the meter; means for resolving sampled
output data from the meter into a non-flow waveform component and
a flow waveform component; and means for determining a measure of
the non-flow waveform component.
41. A computer readable medium including a program for executing
a method of obtaining a measure of flow from an electromagnetic
flow meter, comprising: applying a shaped excitation waveform to
an electromagnetic flow meter; receiving an output from the meter;
and resolving sampled output data from the meter into a non-flow
waveform component and a flow waveform component to derive a measure
of flow.
Description CROSS-REFERENCE TO RELATED APPLICATIONS
This application is related to U.S. patent application Ser. No.
10/817765 filed on the same day as the present application, having
the same inventor as the present application, and entitled "ELECTROMAGNETIC
FLOW METER". The disclosure of the above-described filed application
is hereby incorporated by reference.
BACKGROUND OF THE INVENTION
The present invention relates to electromagnetic flow meters. However,
aspects of signal processing techniques disclosed herein may be
more broadly applied. The operating principles of Electromagnetic
Flow Meters are well known, discussed for example in GB-A-2380798.
Where the sensing electrodes are in contact with the fluid, due
to electrochemical or other effects, a DC potential is usually present
across the electrodes even when there is no coil excitation, i.e.
no field. That component of the signal is independent of the flow
and is generally not static. For example, it may drift randomly
with time, flow and temperature. This inhibits the ability to determine
the flow in a static fashion. To overcome this some form of dynamic
excitation to the coils is typically provided in order to generate
a dynamic component at the electrodes that can be differentiated
from the background DC (or slowly varying) bias signal. This dynamic
signal is normally pulsed DC or an AC signal.
While using some kind of alternating signal to excite the meter
is usually necessary, it does introduce its own particular problem,
namely that electromagnetic coupling between the coil current and
the electrode wires typically creates a signal at the electrodes
when there is no fluid flow through the conduit. This signal is
completely independent of any flow-generated signal at the electrodes
and so the total signal received is the sum of the unwanted `zero`
term (the non-flow component) and the flow generated signal.
In a sinusoidally excited meter, this unwanted signal is typically
at nominally 90 degrees to the wanted signal and hence is often
termed the `quadrature` signal. The unwanted signal is related to
the rate of change of current in the coils and consequently is often
described as being due to the `transformer effect` where the coil
winding is the primary and the electrode wiring is the secondary.
In a perfectly symmetrical sensor, the signal created in the electrode
wiring in a typical sensor should be zero but manufacturing tolerances
mean that there is always some residual area in the electrode wiring
`loop` that picks up some of the primary current.
In a sinusoidally excited meter it is known to include some kind
of phase adjustment in the phase-sensitive detector to null out
this unwanted signal (this process can be performed in hardware,
manually or automatically or in software). Performance of the system
is often limited by the success of this zero removal system. In
many systems where the adjustment is only made at one operating
condition (factory calibration) then the system may exhibit errors
at other operating conditions, typically temperature or installation
related.
AC meters excited with a sinusoidal waveform (or a plurality of
frequencies as we have previously shown) can be beneficial but require
substantially continuous excitation. To reduce power consumption,
for example in a battery-powered meter, intermittent excitation
is preferred, and pulsed DC meters are typically used.
In `pulse` or `square wave` driven meters, the approach is usually
to provide adequate settling time after the coil current is changed
for the transient quadrature signal to die away (since it is related
to the rate of change of coil current). These meters work well because
there is substantially no quadrature signal but the need for the
signal to die away has implications on the maximum speed at which
the meter can be operated.
Conversely, if a measurement is not taken sufficiently quickly,
there is a potential problem of zero drift. The "square wave"
will not normally be true square wave as the slew rate will be limited
by the apparatus and will often be deliberately limited to reduce
effects of the high rate of change of current.
SUMMARY OF CERTAIN INVENTIVE EMBODIMENTS
The system and method described herein may provide a novel solution
to at least some of the problems described above whereby a pulsed
or square wave excitation can be used but the meter can be operated
significantly faster than would normally be possible.
According to a first aspect the invention provides a method of
obtaining a measure of flow from an electromagnetic flow meter comprising
applying a shaped excitation waveform, receiving an output from
the meter and resolving sampled output data from the meter into
a non-flow waveform component and a flow waveform component to derive
a measure of flow.
By shaped, we mean a waveform having a non-sinusoidal shape. The
waveform can be considered as a shaped waveform comprising a plurality
of frequency components, although the components may not be explicitly
added as the signal may be directly generated in the time domain,
for example using a Digital Signal Processor (DSP).
In this method, by taking the novel step of resolving into expected
waveform components, rather than the conventional method of correlating
each frequency component of the sampled output data with a sine
wave, a number of advantages can be obtained. Although the shaped
excitation waveform will effectively contain multiple frequency
components, it has been found that it is not necessary to process
each component separately to obtain a measure of flow.
Preferably, the shaped excitation waveform comprises a pulsed waveform.
For a pulsed waveform, using the present method it may not be necessary
to wait for transient signals to decay before a reading can be taken.
This may provide the advantage over prior art systems that the delay
time before the initial reading is taken can be reduced and the
rate of repetition of reading of the meter can be increased.
Preferably, the shaped excitation waveform includes a first pulse
rise section of a first polarity.
Preferably, the shaped excitation waveform includes a substantially
constant section of a first polarity.
Preferably, the shaped excitation waveform includes a pulse decay
section.
Preferably, the shaped excitation waveform includes a second pulse
rise section.
Preferably, the shaped excitation waveform includes a second substantially
constant section of the opposite polarity to the first substantially
constant section.
Preferably, the shaped excitation waveform includes a second pulse
decay section. Preferably, the second pulse rise, substantially
constant and pulse decay sections are of the opposite polarity to
the first pulse rise section, substantially constant and pulse decay
sections.
Preferably, the shaped excitation waveform includes at least one
substantially constant section. This may be, for example, at the
beginning and/or end of a group of pulses in a pulsed DC system
or it may be an inter-pulse constant section.
The constant section of the excitation waveform may advantageously
be used to determine background interference using correlation.
For example, the constant section may be used to determine 50 Hz
interference.
Preferably, the background interference is subtracted from the
sampled output data; this is preferable to simple filtering.
Preferably, the shaped excitation waveform includes a plurality
of pulse rise sections, substantially constant sections and pulse
decay sections.
Preferably, alternate substantially constant sections are of alternate
opposing polarities.
The features provided may allow a pulsed shaped excitation waveform
to be used. Preferably, a substantially square wave waveform may
be used (although the slew rate will normally be limited and so
it will not be a perfect square wave). Preferably wherein alternate
pulses of the square wave are of opposing polarities, although more
complex patterns having opposing polarities.
This technique may provide a number of advantages, for example,
resolving the output data into components may allow the method to
be used with a wide variety of input waveforms, since the method
is not dependent on the input waveform used.
The method of resolving into a non-flow component and a flow component
may be more accurate and more flexible than subtracting a constant
predetermined zero signal from the sampled output data, since the
zero signal subtracted may not accurately represent the zero signal
at the time the reading was taken.
Preferably, the resolving step comprises performing a weighted
least squares fit between the sampled output data and a model of
the expected waveform, the expected waveform comprising a flow component
and a non-flow component. Other methods of resolving the output
data into components to fit the data to the model of the expected
waveform may also be used.
According to one embodiment, the non-flow waveform component may
be represented in the form z.Vz(t).
Vz(t) may be based on the differential of the current input to
produce the excitation waveform, that is dI/dt. This has been found
to provide a good model for the non-flow component term.
According to an alternative embodiment, Vz(t) may be based on a
zero signal captured empirically. In this embodiment, the non-flow
component is preferably based on an average of a plurality of captured
zero signals. This may provide a more accurate representation of
the non-flow component. The zero signals may be captured at a plurality
of different operating conditions, for example at a plurality of
different temperatures, and the non-flow component used may be varied
depending on the operating conditions when the sampled output data
is taken.
z is preferably a scalar multiplier term and may be designated
the non-flow co-efficient. The value of the co-efficient is preferably
determined during the resolving step. The co-efficient preferably
represents the relative amount of the non-flow component present
in the sampled output data.
According to one embodiment, the flow waveform component may be
represented in the form f.Vf(t).
In one embodiment, Vf(t) may be determined empirically. For example,
the component may be determined experimentally using one or more
flow rates at one or more sets of operating conditions. Preferably,
the component may be based on an average of a plurality of captured
flow signals. This may allow experimental errors in the expected
flow component to be reduced and may provide a more accurate expected
flow component.
According to an alternative embodiment, Vf(t) may be modelled by
a mathematical function. For example, Vf(t) may be expected to be
generally proportional to the magnetic field.
f is preferably a scalar multiplier term and may be designated
the flow co-efficient. The value of the co-efficient is preferably
determined during the resolving step. The co-efficient preferably
represents the relative amount of the flow component present in
the sampled output data.
According to a preferred embodiment, the method preferably further
comprises determining a measure of a trend within the sampled output
data. A trend may result, for example, from experimental effects
such as heating of the components of the flow meter. The trend may
be, for example, a linear trend or an exponential trend.
Preferably, the measure of a trend may be subtracted from the sampled
output data. This may allow a more accurate measurement of the flow
rate to be determined.
According to one embodiment, the correlating step is performed
over a window that does not contain an integer number of periods
of all frequency components. Hence the system described herein may
be more flexible and determination of the flow rate may not be dependent
on the input waveform used.
Preferably, the window may be shorter than the period of the lowest
frequency component. This may allow the measurement of the flow
rate to be obtained with a higher bandwidth than the lowest frequency
component.
According to a further aspect there is provided a method of determining
a measure of a non-flow signal from an electromagnetic flow meter
comprising: applying a shaped excitation waveform to an electromagnetic
flow meter; receiving an output from the meter; resolving sampled
output data from the meter into a non-flow waveform component and
a flow waveform component; determining a measure of the non-flow
component.
In prior art flow meters, a predetermined non-flow component is
subtracted from the sampled output data to remove the non-flow effects
from the output signal. For example, the part of the sampled output
data that is due to transformer effects within the meter itself
can be removed in this way. Determining the magnitude of this non-flow
component itself has not seemed particularly useful, since it provides
no indication of the flow rate of the fluid. However, it has been
appreciated by the inventors that, surprisingly, determining the
magnitude of this non-flow component itself can provide a useful
tool to improve the accuracy of flow readings. For example, determination
of the non-flow component can be used to check that the flow meter
is within calibration tolerance and is free from faults. This may
prevent systematic errors being introduced to flow rates that are
subsequently determined. If the flow meter is in an environment
with particularly high or low temperatures, or with highly fluctuating
temperatures, it may be particularly useful to ensure that the meter
remains within calibration tolerance and the present method may
allow this to be done while the flow meter is in situ and over an
extended period of time.
The ability for the meter to provide a meaningful self-diagnosis
may be very beneficial--a dynamic estimate of the meter accuracy
may be provided. For example, it might be that under unfavourable
conditions, the meter is only 1% accurate but under good signal
conditions may be 0.05% accurate or better. The meter may determine
a measure of accuracy, preferably based on the non-flow signal,
and may report this with the measurement value. Downstream equipment
may use the dynamically reported figure, rather than having to assume
a worst case accuracy of 1%.
Preferably, the method further comprises determining a calibration
measurement for the meter based on the measure of the non-flow component.
Preferably, the method further comprises applying a correction
to flow measurements determined by the flow meter based on the calibration
measurement determined. Hence a non-flow component determined in
this way may also be input back into the system to be incorporated
into later readings that are taken in the presence of flowing fluid
to determine more accurately the flow rate.
According to a further embodiment, the method may further comprise
detecting a fault condition for the meter based on the measure of
the non-flow component. A fault with the meter may be brought to
the attention of the operator so that readings may be adjusted accordingly
or so that the fault may be corrected.
Preferably, the shaped excitation waveform is a pulsed waveform.
The waveform preferably has at least some of the features of the
waveform described for the previous aspect. Preferably, the waveform
is a generally square wave pulse waveform.
Preferably, the non-flow waveform component is represented in the
form z.Vz(t), wherein Vz(t) is a function of the current input to
form the excitation waveform and z is a scalar multiplier term.
The value of z is preferably determined as part of the present method.
Preferably, the flow waveform component is represented in the form
f.Vf(t), wherein Vf(t) is a function of the flow rate of a fluid
through the flow meter and f is a scalar multiplier term.
Features of the first method may be applied to the second method
and corresponding advantages may be provided.
According to a further aspect, there is provided apparatus for
obtaining a measure of flow from an electromagnetic flow meter comprising:
means for applying a shaped excitation waveform; means for receiving
an output from the meter; and means for processing the output from
the meter to resolve sampled output data from the meter into a non-flow
waveform component and a flow waveform component to derive a measure
of flow.
Features of the method aspect described above may be applied to
the apparatus aspect and may provide corresponding advantages.
According to a further aspect there is provided a computer program
or computer program product for carrying out a method according
to the method aspects described above or any of their preferred
features.
BRIEF DESCRIPTION OF THE DRAWINGS
Embodiments of the invention will now be described by way of example
with reference to the accompanying drawings in which:
FIG. 1 illustrates one possible coil excitation signal (current)
for a meter;
FIG. 2 shows an electrode signal produced by a sensor when driven
by the coil excitation signal illustrated in FIG. 1;
FIG. 3 shows the total signal, that is the zero signal and the
flow-induced signal for a particular flow rate;
FIG. 4 shows the electrode signal of FIG. 3 with the zero signal
of FIG. 2 subtracted;
FIG. 5 is a schematic diagram of one embodiment of the invention.
DETAILED DESCRIPTION OF CERTAIN INVENTIVE EMBODIMENTS
FIG. 1 illustrates one possible coil excitation signal for a meter.
The waveform illustrated may be created by a controlled drive voltage,
for example by a Central Processing Unit (CPU) and a Digital-to-Analogue
(D/A) converter. As described in more detail below, in the present
embodiment, the precise waveform is not important, however features
of the pulse may include that it is zero outside the active pulse
region 110 and that the pulse has some AC content, i.e. it goes
positive 112 and negative 114 at least once. In most embodiments,
drive voltages may be slew-rate limited, since this may create signals
which are less problematic for the equipment as rapid changes make
the relationship between field and applied current less predictable.
FIG. 2 illustrates an electrode signal produced by a real sensor
in the absence of fluid flowing past the sensor when driven by the
coil current shown in FIG. 1. This signal may be known as the zero
signal.
This electrode signal has high frequency noise and a small amount
of mains interference and other corruption. The signals captured
here are from an unshielded sensor. Signals captured from production
sensors, which may have shields incorporated with them may not show
as much noise. However, the techniques described herein can preferably
operate in the presence of such distortions. In the present embodiment,
the electrode signal captured can be sufficiently accurately described
as a scaled rate of change of the current (I) with respect to the
time (t), that is a dI/dt of the coil excitation signal shown in
FIG. 1.
When there is flow present, there will be an additional component
which is generally proportional to the flow rate and to the magnetic
field, the magnetic field being generally proportional to the coil
current (but not exactly because of factors such as hysteresis and
magnetic circuit eddy currents).
FIG. 3 shows the total signal for a particular flow rate, that
is the non-flow signal and the flow-induced signal. Points 310 and
312 (at about x=4400 and about x=9500) are the points where, after
a longer settling time, most prior art systems would choose to derive
the flow signal as, in principle, the flow rate can be determined
based on the result of calculating (Va-Vb)/2. Note further that
in this case, the signal at A and B is not settled before the coil
current is changed.
It will be appreciated that the non-flow signal could be subtracted
from this waveform and, if the non-flow signal can be expected to
remain completely constant, then this would be a useful thing to
do. FIG. 4 shows the waveform after FIG. 2 the non-flow signal,
has been subtracted.
It can be seen that the waveform is now closer to an ideal signal.
However, the technique described herein has more flexibility than
this, particularly when the non-flow signal might be expected to
change, for example for wide temperature applications.
According to the present embodiment, the technique is to find coefficients
z and f such that:
matches the measured shaped waveform as closely as possible.
Where Vz(t) is a captured non-flow signal and Vf(t) is the expected
flow component in the absence of a non-flow/quadrature term.
In the present embodiment, z and f are scalar multipliers that
represent the amount of the non-flow and flow components present
in the sampled output data.
This approach may be more flexible and accurate than simply subtracting
a predefined non-flow signal, particularly if the non-flow signal
being subtracted does not accurately reflect the current non-flow
signal.
As will be appreciated, the technique is generic and does not depend
in any way upon the exact waveforms being used, though it is useful
for the waveforms to have certain characteristics (or to not have
certain characteristics), some of which are discussed in more detail
below.
An example of a fitting technique that may be used to determine
the co-efficients z and f is weighted least squares. The technique
has been found to fit the real data over all measured flow rates
with no systematic residual error, just random noise.
The flat, zero section outside the pulse 110 may be used to determine
50 Hz (or other mains or other significant source e.g. 60 Hz) interference
using correlation. The magnitude and phase of the interference as
derived from this section may then be subtracted from the whole
window. This is found to be more effective than trying to filter
out the interference as the waveform is less distorted.
The flat zero sections may also be used to identify and remove
any linear trend from the measured electrode signal prior to further
processing.
Multiple captures of the non-flow signal can be taken (offline)
and may then be averaged to create a very clean non-flow signal
prior to being used in a weighted least squares regression.
Similarly, multiple flow signals can be taken and may be averaged
to get the Vf(t) term prior to a weighted least squares regression.
The technique may also be used to identify that zero is drifting
if z shows a systematic shift from the factory calibrated or site
calibrated value.
In one embodiment, the current waveform may be used as the Vf(t)
and/or dI/dt may be used as the Vz(t).
One embodiment of a practical system will now be described in more
detail with reference to FIG. 5. A microprocessor 10 may be provided
which may maintain a `time` variable to calculate drive signals.
The drive signals, for example a square wave signal, may be applied
to the drive coil in the sensor 3 via power amplifier 2. The electrode
signals may then be amplified 4 and may be converted into discrete
samples by ADC 5. Software modules running on the microprocessor
10 may be used to implement the methods described herein. The amplified
output of the flow meter may be resolved into a flow component and
a non-flow component 6 and a measure of the non-flow component may
be determined 8. This may allow a correction term to be calculated
9 and fed back into the system 6 to allow a more accurate flow rate
to be determined 6 and output 7.
The length of the samples may be set, to a fixed period (e.g. 25
ms is convenient) or may be adjustable, for example by setting a
parameter at the microprocessor 10.
In one embodiment, a confidence value may also be determined for
each flow reading output. This may be used, for example, to "weight"
the data for filtering purposes. A number of different filtering
and weighting algorithms may be used and the choice may depend on
the application.
Modifications of detail may be made to the method and system described
and features disclosed herein may be provided independently or in
other combinations. |