 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates generally to electromagnetic flowmeters wherein
excitation current for the electromagnetic coil is a low-frequency wave,
and more particularly to a flowmeter converter connected to the flowmeter
electrodes which includes a delayed automatic gain or smoothing network
for suppressing excessive noise.
In an electromagnetic flowmeter, the liquid whose flow rate is to be
measured is conducted through a flow tube provided with a pair of
diametrically-opposed electrodes, a magnetic field perpendicular to the
longitudinal axis of the tube being established by an electromagnet. When
the flowing liquid intersects this field, a voltage is induced therein
which is transferred to the electrodes. This voltage, which is
proportional to the average velocity of the liquid and hence to its
average volumetric rate, is then amplified and processed in a converter to
actuate a recorder or indicator.
The magnetic field may be either direct or alternating in nature, for in
either event the amplitude of voltage induced in the liquid passing
through the field will be a function of its flow rate. However, when
operating with direct magnetic flux, the D-C signal current flowing
through the liquid acts to polarize the electrodes, the magnitude of
polarization being proportional to the time integral of the polarization
current. With alternating magnetic flux operation, polarization is
rendered negligible, for the resultant signal current is alternating and
therefore its integral does not build up with time.
Though A-C operation as disclosed, for example, in the Cushing U.S. Pat.
No. 3,693,439 is clearly advantageous in that polarization is obviated and
the A-C flow induced signal may be more easily amplified, it has distinct
drawbacks. The use of an alternating flux introduces spurious voltages
that are unrelated to flow rate and, if untreated, give rise to inaccurate
indications. The two spurious voltages that normally are most troublesome
are:
1. stray capacitance-coupled voltages from the coil of the electromagnet to
the electrodes, and
2. induced loop voltages in the input leads. The electrodes and leads in
combination with the liquid extending therebetween constitute a loop in
which is induced a voltage from the changing flux of the magnetic coil.
The spurious voltages from the first source may be minimized by
electrostatic shielding and by low-frequency excitation of the magnet to
cause the impedance of the stray coupling capacitance to be large. But the
spurious voltage from the second source is much more difficult to
suppress.
The spurious voltage resulting from the flux coupling into the signal leads
is usually referred to as the quadrature voltage, for it is assumed to be
90.degree. out of phase with the A-C flow-induced voltage. Actual tests
have indicated that this is not true in that a component exists in-phase
with the flow-induced voltage. Hence, that portion of the "quadrature
voltage" that is in-phase with the flow-induced voltage signal
constitutes an undesirable signal that cannot readily be distinguished
from the flow-induced signal, thereby producing a changing zero shift
effect.
Pure "quadrature" voltage has heretofore been minimized by an electronic
arrangement adapted to buck out this component, but elimination of its
in-phase component has not been successful. Existing A-C operated
electromagnet flowmeters are also known to vary their calibration as a
function of temperature, fluid conductivity, pressure and other effects
which can alter the spurious voltage both with respect to phase and
magnitude. Hence it becomes necessary periodically to manually re-zero the
meter to correct for the effect on zero by the above-described phenomena.
All of the adverse effects encountered in A-C operation of electromagnetic
flowmeters can be attributed to the rate of change of the flux field
(d.phi.)/dt, serving to induce unwanted signals in the pick-up loop. If,
therefore, the rate of change of the flux field could be reduced to zero
value, then the magnitude of quadrature and of its in-phase component
would become non-existent. Zero drift effects would disappear.
When the magnetic flux field is a steady state field, as, for example, with
continuous d-c operation, the ideal condition (d.phi.)/dt=0 is satisfied.
But, as previously noted, d-c operation to create a steady state field is
not acceptable, for galvanic potentials are produced and polarization is
encountered.
In the patent to Mannherz et al., U.S. Pat. No. 3,783,687, whose entire
disclosure is incorporated herein by reference, there is disclosed an
electromagnetic flowmeter in which the excitation current for the
electromagnetic coil is a low-frequency wave serving to produce a
periodically-reversed steady state flux field, whereby unwanted in-phase
and quadrature components are minimized without giving rise to
polarization and galvanic effects. This low frequency wave is derived by
means of a presettable scaler coupled to the standard a-c power line (60
Hz) and is at a frequency in the order of 17/8, 33/4, 71/2 or 15 Hz.
When the fluid being metered takes the form of a coarse slurry containing
solid particles such as sand, fly ash or salt which impinge on the surface
of the electrodes as the slurry passes through the meter tube, it has been
found that a substantial noise component is generated. This makes signal
detection more difficult and in some instances impossible. The meter
electrodes in combination with the fluid acting as an electrolyte define a
galvanic cell, and when the solids in the slurry strike the electrodes and
alter their interface to the fluid, this action brings about a rapid
change in galvanic voltage, thereby generating low frequency noise.
A second source of noise arises when the flowmeter is run partially full.
In the case of electrodes which make direct contact with the fluid, the
resultant sloshing of the fluid on the surface of the electrodes produces
excessive galvanic noise. In the case of electrodes which are capacitively
coupled to the fluid, the sloshing action results in changes in dielectric
constant (1 to 80) to produce capacitive noise.
Noise is any voltage that does not convey measurement information. Under
the most favorable circumstances where noise has been minimized by
filtering or other expedients, there are still certain sources of noise
present resulting from the granular nature of matter and energy. While
noise fluctuations may be small compared with the total energy transfer
involved in most measurements, the existence of a noise background limits
the ultimate sensitivity to which a measurement can be carried.
With electromagnetic flowmeters having a coil excited by an alternating
current or a periodically interrupted direct current, when the fluid being
metered is a slurry containing solid particles, a spectrum of noise
voltages is generated. But the frequency components of this spectrum do
not include the frequency of the standard a-c power line (i.e., 50 or 60
Hz). Hence in a magnetic flowmeter in which the excitation current is at
the line frequency, no difficulty is experienced in discriminating between
the flow-induced signal and noise components to provide a favorable
signal-to-noise ratio.
But with an electromagnetic flowmeter of the above-described Mannherz type
in which the excitation frequency is well below 60 Hz, it has been found
that the frequency components of the noise spectrum actually lie within
the excitation frequency range; hence discrimination between noise and
signal cannot be effected by bandpass filtering, and the signal-to-noise
ratio is unfavorable. Indeed, in some instances, the magnitude of the
noise is of sufficient strength to cause a 100% output change in the
converter, in addition to causing the average reading to be incorrect.
SUMMARY OF INVENTION
In view of the foregoing, the main object of this invention is to provide
an electromagnet flowmeter having low frequency excitation for metering a
slurry having solid particles therein which impinge on the flowmeter
electrodes, the flowmeter including means to increase the response time of
the converter and thereby decrease the influence of noise thereon.
More particularly, it is an object of this invention to provide a flowmeter
whose excitation current is constituted by a periodic wave whose frequency
is low relative to the frequency of the standard a-c power line from which
the wave is derived, the flowmeter converter including a delayed automatic
smoothing or gain network for suppressing excessive electrode noise.
Briefly stated, these objects are attained in an electromagnetic flowmeter
in accordance with the invention, wherein a liquid to be metered is
conducted through the flow tube of a primary to intersect a transverse
magnetic field produced by an electromagnet, the resultant signal induced
in the liquid being transferred to a pair of electrodes mounted at
diametrically-opposed points on the tube.
To avoid spurious signal components arising from stray couplings between
the electromagnet and the loop constituted by the electrodes and the
liquid bridging the electrodes, the coil of the electromagnet is energized
by a low-frequency wave whose frequency is well below the standard a-c
power line frequency.
To avoid noise components produced when the liquid being metered is a
slurry containing solid particles which impinge on the surface of the
electrodes, the flowmeter secondary includes a delayed automatic gain or
smoothing network which is activated when a signal having an excessive
noise component is detected which persists for a predetermined period,
activation serving to extend the response time of the secondary and
thereby reduce the influence of the noise component therein whereby the
output of the secondary has a favorable signal-to-noise ratio.
OUTLINE OF DRAWINGS
For a better understanding of the invention as well as other objects and
further features thereof, reference is made to the following detailed
description to be read in conjunction with the accompanying drawings,
wherein:
FIG. 1 is a block diagram of an electromagnetic flowmeter which includes
noise suppression delayed smoothing circuit in accordance with the
invention; and
FIG. 2 shows a delayed automatic gain circuit in accordance with the
invention.
DESCRIPTION OF INVENTION
The Basic Flowmeter
Referring now to FIG. 1, there is shown a flowmeter system in accordance
with the invention constituted by a primary in the form of an
electromagnetic flowmeter adapted to produce a low-frequency signal output
whose amplitude is proportional to the flow rate of the liquid being
measured, and a secondary which converts this low level a-c signal to a
proportional d-c current output signal.
In addition to an analog d-c current output, the secondary also yields a
pulse output whose frequency is proportional to flow rate, the pulse
output being useful in operating digital registers, counters or bath
control systems. By means of proper scale factoring, registry of total
accumulated flow may be integrated from this rate signal.
The flowmeter primary includes a flow tube 10 through which the liquid 11
to be measured is conducted. An electromagnet having a coil 12 is provided
to establish a magnetic field transverse to the direction of flow which is
parallel to the longitudinal axis of the tube. Electrodes 13 and 14 in
contact with the flowing fluid are disposed at diametrically-opposed
points on the flow tube on an axis which is perpendicular both to the
longitudinal direction of flow and the transverse magnetic field. The
invention is also applicable to electrode arrangements in which the
electrodes are capacitively coupled to the fluid, in which event the
resultant noise is of capacitive origin, as explained in the introduction.
As is well known, a voltage is induced in the liquid whose flow intersects
the magnetic field, this voltage being transferred to the electrodes to
produce a signal at flowmeter output terminals that reflects the flow
rate. This signal is referred to as the flow-induced signal to distinguish
it from spurious signal components that are independent of flow rate and
noise.
Magnet coil 12 is energized by a relatively low-frequency square wave
derived from a full-wave rectifier power supply constituted by a
transformer 16 whose primary is connected to an a-c power line through a
regulator 17, the line supplying the usual 50 or 60 Hz current. The
secondary of transformer 16 is connected to the input junctions of a
full-wave rectifier bridge 18, whose output junctions are connected to the
respective movable contacts of two single-pole single-throw switches 19
and 20 whose fixed contacts are both connected to one end of magnet coil
12. The other end of the coil is connected to the center tap of the
secondary of transformer 16.
When switch 19 is closed and switch 20 is simultaneously open, the
rectified output is applied to the magnet coil in one polarity, and when
switch 19 is open and switch 20 is simultaneously closed, the polarity is
reversed. While for purpose of explanation, switches 19 and 20 are shown
as mechanical devices, in practice these switches are in electronic form
and may be constituted by thyratrons, triacs or any other type of
electronic switching device in vacuum tube or solid-state form.
Switches 19 and 20 are activated at a rate which is low relative to the
frequency of the a-c line. This is accomplished by means of a presettable
scaler or frequency divider 21 to which the 60 Hz line voltage is applied
as a clock signal, the scaler yielding low frequency pulses in the order
of 17/8, 33/4, 71/2 or 15 Hz.
The low frequency pulses from the scaler are applied to the firing
electrodes of the two triac switches (or whatever other electronic
switching devices are used) to alternately turn on the triacs and thereby
connect either the positive or the negative side of the full-wave
rectified 60 Hz voltage to the magnet coil. Thus when switch 19 is closed,
current flows through the magnet coil in one direction; and when switch 20
is closed, the current flows in the reverse direction.
Because the output of the full-wave rectifier is a raw unfiltered direct
voltage, it is composed of a continuous train of half-cycle pulses, all of
the same polarity. But with the low-frequency switching action, the
voltage applied to magnet coil 12 is periodically reversed in polarity, as
a result of which the current passing through the coil has a 120 Hz ripple
component.
Because the electromagnet has a relatively high inductance, it functions as
a filter choke and, in practice, it takes out as much as 75 percent of the
ripple component. The remaining portion of the 120 Hz ripple component
that appears in the flow-induced signal is smoothed out at the summing
junction of the secondary via the filter action of the associated error
amplifier, to be later described. This obviates the need for filter
capacitors associated with the choke, as in conventional filters. Thus,
the system functions as if it were excited by a "square wave equivalent"
having a substantially constant amplitude.
The flow-induced signal appearing at output terminal 15 of the flowmeter
primary is fed to a secondary constituted by a converter. This converter
is essentially an all solid-state a-c feedback system producing a
frequency output (and optional current) whose rate is proportional to
flow.
In the converter, the flow-induced signal appearing at output terminals 15
is applied to the first stage of the converter which is an a-c
pre-amplifier 22. This signal has a generally square-wave formation but
for the spikes appearing at the points of polarity reversal. These spikes
are the result of switching transients or surges and have a duration
depending on the inductance-resistance time constant of the
electromagnetic circuit.
The constant level portion of the square wave reflects the steady state
condition of the magnetic field and has an amplitude that is directly
proportional to the velocity of liquid passing through the flow tube.
Hence it is only this portion of the signal which is of interest for
accurate measurement purposes.
The output of pre-amplifier 22 is applied through a blocking capacitor 23
to one input of a summing junction 24 to whose other input is fed the
output of a range attenuation circuit 25 from an error signal type of
feedback loop. The error signal produced by a comparison of the flow
signal and the feedback signal in the summing junction is amplified in a-c
error amplifier 26, which is provided with a sub-negative feedback circuit
27 adapted to attenuate all frequencies lower and higher than that of the
error signal.
The a-c output of error amplifier 26 is applied as a 0.degree. signal with
respect to the flow signal to a full-wave demodulator 29 and is also
applied to an inverting (1:1) amplifier 28 whose output is applied as a
180.degree. signal to full-wave demodulator 29. The operation of the
demodulator is synchronized with the low-frequency switching rate of the
magnet coil and is so gated as to block the applied error signal at those
points corresponding to the point of polarity reversal, the blockage being
maintained for the duration of the inductance-resistance time constant of
the electromagnetic circuit. In this way, the d-c output of the
demodulator reflects only the steady state magnetic flux condition, the
spike portions of the flow-induced signal being suppressed.
In order to so synchronize the demodulator, the frequency divider 21 which
responds to the 50 or 60 cycle signal to produce low-frequency control
pulses for governing the electromagnetic switching action is provided with
suitable logic to produce gating pulses at the same low-frequency rate.
These gating pulses are coincident with the steady state portion of the
flow-induced signal. Thus, the demodulator is activated only during the
steady state intervals and is otherwise blocked. As a consequence, the
secondary only looks at the flow-induced signal during the point that
d.phi./dt is equal to zero.
The d-c output pulses produced by demodulator 29 are applied to a Miller
integrator including an input resistor 30A, a d-c amplifier 31 and a
feedback capacitor 30B to produce a direct-voltage error signal whose
magnitude is a function of flow rate. This d-c amplified error signal is
used as a controlled bias for a d-c to frequency converter 32 that, in
practice, may take the form of a blocking oscillator.
The blocking oscillator translates the d-c error signal level to a variable
frequency pulse train which exhibits a duty cycle that is proportional to
the error signal. This variable duty cycle error signal VD is used to
drive the output circuits of the system as well as serving as the take-off
point for the error signal feedback circuit.
For purposes of feedback, the variable duty cycle error signal must first
be restored to a proportional low-frequency signal (i.e., 17/8 Hz or
whatever low frequency is in effect). This is accomplished by means of a
sampling circuit 33 coupled to the dc-to-duty cycle converter 32 and
acting to sample in-phase reference voltage RV derived from the
electromagnet circuit. To generate this reference voltage, a fractional
ohm resistor 34 is interposed between magnet coil 12 and the center tap of
the secondary of transformer 16, the voltage drop thereacross depending on
current flow through the magnet coil. This voltage is applied to an
operational amplifier 35 to produce reference voltage RV at its proper
level. The output of sampler 33 is constituted by the duty-cycle pulses
derived from the d-c to frequency converter, enveloped by the low
frequency square wave reference voltage RV. This output is fed to summing
junction 24 through the range attenuator 25. Inasmuch as this feedback
signal depends on the amplitude of reference voltage RV, any variation in
the voltage as a result of line fluctuations will proportionately change
the feedback signal. Since it is the ratio of the flow-induced signal to
the feedback signal that constitutes the measurement criterion, no loss of
accuracy will be experienced with variations in line voltage (within
reasonable limits).
The signal from the dc-to-duty cycle converter is applied to a
duty-cycle-to-dc converter F/I, which converts the pulses of the former
into an analog d-c output that is proportional to fluid flow rate. The
signal from the dc-to-duty cycle converter is also applied to a pulse
scaler F/F which converts the applied pulses into engineering units which
are available to drive an external counter.
Instead of reversing polarity, the electronic switches may be arranged to
merely interrupt the flow of unidirectional half-wave pulses to the magnet
coil so that instead of a square-wave current in the coil in which
successive square wave pulses alternate in polarity, all of the
square-wave pulses are of the same polarity.
As pointed out previously, when the liquid 11 being metered is a slurry
having solid particles therein which impinge on the electrode surface in
the course of passage through flow tube 10, noise voltages are generated
whose magnitude is relatively high. With low-frequency excitation
currents, the excitation frequency range lies within the spectrum of noise
components; hence the noise components tend to mask the flow-induced
signal and it is difficult to obtain accurate flow rate readings.
Delayed Smoothing Network
In those situations where a coarse slurry rather than a liquid free of
particles appears intermittently or where the flow tube is intermittently
only partially full, it is then desirable to provide a response time of 4
seconds which is operative whenever the meter is quiet; that is, when
there is no slurry and a full pipeline. When, however, an excessive
noise-producing condition is encountered, it is then desirable to
automatically extend the response time, but only after excessive noise has
persisted for a predetermined period, as, for example, more than 5, 10 or
20 seconds. This is necessary in order to avoid an automatic extension of
response time as a reaction to momentary noise transients.
To this end, a delayed smoothing network is provided that includes a
capacitor 36 which, when the network is activated, is shunted across the
feedback capacitor 30B of the Miller integrator to extend the response
time of the converter. The value of this capacitor may be 5 to 50
microfarads, depending on the desired delay. The capacitor switching
action is controlled by a relay 37 whose pivoted armature 37' normally
engages a contact A connecting capacitor 36 to ground. When relay 37 is
energized, armature 37' is then caused to engage a contact B which serves
to connect capacitor 36 across capacitor 30B.
Relay 37 is connected through the emitter-collector path of a switching
transistor 38 across a d-c supply (12 V). The base of this transistor is
connected to the output of a control amplifier 39 whose input is connected
through an R-C timing circuit 40 and a diode 41 to the output of A-C error
amplifier 26.
Thus the output of the error signal channel is sensed; and if the error
signal exceeds normal balance conditions, this being indicative of the
presence of excessive noise, the timing circuit begins to time out. If the
error signal returns to normal in 4 seconds or less with a change in flow,
the bias applied to amplifier 39 will not be sufficient to effect
energization of relay 37.
If, however, the excessive error signal is due to noise and it persists for
a prolonged period (i.e., 5 to 20 seconds), relay 37 will then be
energized to switch in smoothing capacitor 36 to extend the response time
to the converter from a 4 second to a 100 second interval, or whatever
other extension is provided by the parameters selected for the delay
network. In the Miller integrator, the switching in of an additional
feedback capacitance acts to increase the loop time constant and thereby
greatly extends the response time.
Delayed Automatic Gain
Referring now to FIG. 2, an arrangement is shown providing delayed
automatic gain to extend the response time of the converter and thereby
reduce the influence of noise components thereon.
The circuit shown in FIG. 2 represents an insert to be made in FIG. 1
between inverting amplifier 28 and full-wave demodulator 29. This insert
includes a pair of differential amplifiers 26A and 26B, the 180.degree.
output from inverting amplifier 28 in FIG. 1 going to the non-inverting
input of amplifier 26A whose output is applied to the one input of
demodulator 29, the 0.degree. output from amplifier 26 in FIG. 1 which
goes to the other input of the demodulator also going to the non-inverting
input of amplifier 26B. The output of amplifier 26B is applied to the
inverting input of amplifier 26A.
Error amplifier 26B is provided with a first feedback path between the
output and the inverting input, the path being defined by a feedback
resistor 42 shunted by series-connected diodes 43 and 44. To effect
delayed gain, a second feedback path is provided which is connected
between the output of amplifier 26B and its inverting input, this path
being constituted by a resistor 45 coupled through an amplifier 46 to a
resistor 47.
Amplifier 46 is normally maintained in an inactive state by a control
transistor 48 (FET) which is coupled to the output of a control amplifier
49. The output of error signal amplifier 26B is connected through a diode
50 and a timing circuit 51 to the non-inverting input of control amplifier
49 which functions to apply a control signal to transistor 48 to activate
amplifier 46 and thereby render the second feedback path effective.
Because of the additional feedback, the feedback loop is reduced, thereby
extending the response time.
As with the first embodiment, the timing circuit 51 serves to render the
automatic gain system operative only when an excessive error signal is
sensed indicative of noise, and this noise persists for a predetermined
period.
In a delayed action converter in accordance with the invention, the reason
why an extended responsive time reduces the influence of the noise
component in the signal yielded by the meter electrodes is that the
converter is a closed loop system which behaves like a low-pass filter
that blocks all frequencies above a break point. It is this break point or
corner that is set by the response time of the converter. Thus if the
response time were one second, the break point would occur at about 1 Hz;
but if the system was slowed to a ten-second response time, the corner
would be moved down to one-tenth Hz.
While there has been shown and described a preferred embodiment of an
electromagnetic flowmeter having noise suppression network in accordance
with the invention, it will be appreciated that many changes and
modifications may be made therein without, however, departing from the
essential spirit thereof. Thus while the networks to extend response time
have been disclosed in the context of converters included in magnetic
flowmeters, these networks are also usable in other applications wherever
the need exists to suppress noise components accompanying a signal.
* * * * *
|
|
 |
|
 |
Description |
|
