`
`[19]
`
`[11] Patent Number:
`
`4,823,614
`
`
`
`[45] Date of Patent: Apr. 25, 1989
`Dahlin, Erik B.
`
`[54]
`
`[76]
`
`CORIOLIS-TYPE MASS FLOWMETER
`
`Inventor: Dahlin, Erik B., 1936 Arroyo Seco
`Dr., San Jose, Calif. 95125
`
`[21]
`
`App]. No.: 873,201
`
`[22]
`
`Filed:
`
`Jun. 11, 1986
`
`[63]
`
`[51]
`[52]
`[58]
`
`[56]
`
`Related U.S. Application Data
`
`Continuation-impart of Ser. No. 856,939, Apr. 28,
`1986, abandoned.
`
`Int. Cl.4 ................................................ G01F 1/84
`
`U.S. Cl. ..............
`73/861.38; 73/198
`Field of Search ......................... 73/861.37, 861.38
`
`References Cited
`
`U.S. PATENT DOCUMENTS
`
`Re. 31,450 11/1983 Smith ............................... 73/861.38
`3,276,257 10/1966 Roth ................................. 73/861.38
`3,329,019
`7/1967 Sipin.
`3,355,944 12/1967 Sipin .
`3,485,098 12/1969 Sipin .
`4,011,757
`3/1977 Baatz.
`4,109,524
`8/1978 Smith.
`4,127,028 11/1978 Cox .
`4,187,721
`2/1980 Smith.
`.
`4,263,812 4/1981 Zeigner et a1.
`4,393,720 7/ 1983 Takahashi et a1.
`.
`4,393,724 7/ 1983 Werkrnann et a1.
`4,422,338 12/1983 Smith ............................... 73/861.38
`4,491,025
`1/1985
`73/861.38
`
`4,559,833 12/1985 Sipin ................................. 73/861.38
`4,622,858 11/1986 Mizerak .
`4,628,744 12/1986 Lew .
`4,658,657 4/1987 Kuppers ........................... 73/861.38
`
`4,680,970 7/ 1987 Simonsen et a1.
`.
`....... 73/861.38
`4,703,660 11/ 1987 Brenneman ...................... 73/861 .38
`
`.
`
`FOREIGN PATENT DOCUMENTS
`
`.
`1/ 1984 European Pat. Off.
`0119638A1
`3329544A1 9/ 1983 Fed. Rep. of Germany .
`3503841A1
`8/1986 Fed. Rep. of Germany .
`
`OTHER PUBLICATIONS
`
`R. M. Langdon of GEC Research Labs, Marconi Re-
`search Ctr., Sensors & Instruments Div., West Hanning-
`feld Rd., Great Baddow, Essex, UK Review Article
`titled: Resonantor Sensors—A Review in J. Phys. E.
`Sci., Instrum., vol. 18, 1985, Printed in Great Britain, 3
`pages.
`Bopp & Reuther. Titled: Massedurchflussrnesser Sys-
`tem Rheonik Serie RHM Cover page plus 2 pages.
`W. A. Wildhack: “Review of Some Methods of Flow
`Measurement” in Science, 8/1954, One page, showing
`FIG. 12 Schematic Drawing of “Vibro—Gyro”.
`Endress & Hauser, m-Point Massflowmeter.
`
`Primary Examiner—Herbert Goldstein
`Attorney, Agent, or Firm—Ciotti & Murashige, Irell &
`Manella
`
`[57]
`
`ABSTRACT
`
`A Coriolis-type mass flowmeter in which the flow tube
`is vibrated at a resonance frequency approximately
`equal to the frequency for forced or natural vibration in
`a higher anti-symmetric mode, such as the second
`mode. In the preferred embodiment the flowmeter is
`symmetrical and has sections of oval cross section that
`provide low bending resistance to the vibration at the
`points where the amplitude of vibration is the largest.
`The preferred embodiment uses electronic signal detec-
`tion/processing means that generates two signals pro-
`portional to flow tube velocity in the direction of vibra-
`tion at equal distance but on opposite sides of the plane
`of symmetry of the tube, generates a sum and a differ-
`ence of the two signals, integrates the sum, demodulates
`the integrated signal and the difference of the two sig-
`nals to produce peak amplitude signals, and divides the
`peak amplitude signals to produce an output that is
`proportional to mass flow rate. The preferred embodi-
`ment is further equipped with a novel acoustic wave
`suppressor.
`'
`
`21 Claims, 8 Drawing Sheets
`
`
`
`Micro Motion 1021
`
`1
`
`Micro Motion 1021
`
`
`
`. US. Patent
`
`Apr. 25, 1989
`
`Sheet 1 of8
`
`4,823,614
`
`_.9“.
`
`2
`
`
`
`. US. Patent
`
`Apr. 25, 1989
`
`Sheet 2 of 8
`
`4,823,614
`
`A‘ m
`FIG 2A I‘
`
`
`
` |Y
`
`Z
`
`FIG. 20
`
`3
`
`
`
`US. Patent
`
`Apr. 25, 1989
`
`Sheet 3 of 8
`
`‘
`
`4,823,614
`
`FLUID FLOW
`
`
`
`4
`
`
`
`US. Patent
`
`Apr. 25, 1989
`
`Sheet 4 of 8
`
`4,823,614
`
`
`
`3~
`_ u" n' ".4" -.t- ..'- u.".1‘ -.' .'.
`'.-.-‘-'-..-
`.1
`..'. ..‘-‘- -
`-
`- .' h‘.
`- -.' '
`' ”5L
`, -_‘~ .’-,- '.'.
`~ - ; -. ~‘,','. :2- -.‘
`'- :-. '.'.--,‘.'.~ . -‘ - ...‘:..'k:.-" .I-I'1“.‘..‘..‘ .- ‘: -.“|
`r-.....-......-u-.-.....-
`
`
`
`
`5
`
`
`
`. US. Patent
`
`Apr. 25, 1989
`
`Sheet 5 of8
`
`4,823,614
`
`
`
`
`DRIVE
`
`FIG. TB
`
`6
`
`
`
`US. Patent
`
`Apr. 25, 1989
`
`Sheet 6 of 8
`
`4,823,614
`
`mmcE
`
`
`
`-oozmo
`
`m0k<43
`
`NO_
`
`004m>
`
`momzmm
`
`
`
`<0..9...
`
`
`
`_O_
`
`UOJm>
`
`momzmm
`
`7
`
`
`
`
`
`US. Patent
`
`Apr. 25, 1989
`
`Sheet 7 of8
`
`4,823,614
`
`
`
`‘
`
`VELOCITY
`SENSOR
`
`
`
`
`FIG IOC
`
`VELOCITY
`SENSOR
`
`DRIVE LOCATION
`FOR SYMMETRIC MOTION
`
`AMPLIFIER
`
`MASS
`FLOW RATE
`
`W
`PROFILE
`
`'
`
`FIG.
`
`||
`
`SENSOR
`
`+%+
`
`FLOW TUBE
`
`|24
`
`I25
`
`
`
`
`w-
`‘u
`
`
`I.
`
`
`
`i-
`I
`
`‘ .
`
`E
`
`
`
`
`FIG. I2
`
`
`
`8
`
`
`
`US. Patent
`
`Apr. 25, 1989
`
`Sheet 8 of 8
`
`4,823,614
`
`NV.
`
`
`
`
`9
`
`
`
`4,823,614
`
`1
`
`CORIOLIS-TYPE MASS FLOWMETER
`
`CROSS-REFERENCE TO RELATED
`APPLICATION
`
`2
`a single flow measurement the flow tube’s motion is
`periodic and any one of these physical variables for
`almost any point on the flow tube together with known
`vibrating frequency and amplitude permits determina-
`tion of the flow rate. The dependency of flow rate de-
`termination on drive frequency and amplitude has been
`of fundamental importance in the design of prior Corio-
`lis mass flowmeters.
`When mass flow rate changes the flow tube motion
`changes which is the principle of the flow measurement.
`However, if the drive amplitude changes, the Coriolis
`portion of the motion changes also. If one did not know
`the new amplitude (for example by absence or inaccu-
`racy of amplitude measurement), the flowmeter may
`not distinguish a flow rate change from the amplitude
`change. Similarly, if the fluid temperature changes, the
`flow tube wall temperature would also change. The
`elasticity coefficient (Young’s modulus) for the flow
`tube material changes with temperature impacting the
`Coriolis force induced motion. A change of 10° C.
`would potentially bring the flowmeter outside specified
`calibration accuracy (for example 0.2% of reading) if
`the flow tube was made of stainless steel.
`Fluid pressure change modifies the cross section di-
`mension of the flow tube and, thereby, its bending prop-
`erties. Large pressure changes which may occur in
`practice can jeopardize calibration accuracy unless the
`flowmeter design eliminates this hazard.
`Major considerations for Coriolis flowmeters are
`calibration sensitivity and immunity to density change.
`Process fluids seen by the flowmeter may undergo ex-
`tensive density change. The reason may be change of
`fluid temperature and composition. The density change
`will modify the natural frequency of vibration for the
`flow tube. Since the flow tubes are usually driven in the
`immediate vicinity of a natural frequency,
`the drive
`frequency will change with density. The flowmeter
`design determines the extent or complete absence of
`calibration error due to density shift.
`Another problem for Coriolis flowmeters (as well as
`other types) is entrainment of gases in the fluid. The gas
`may be in the form of visible or microscopic size bub-
`bles. Gas entrainment causes both density change and
`change in the coupling between the fluid and the wall of
`the measurement tube essential for the Coriolis flowme-
`ter. Generally Coriolis flowmeters today exhibit signifi-
`cant to intolerable errors in calibration when the gas
`entrainment reaches a magnitude of 1% to 3% by vol-
`ume of gas to volume of fluid.
`There is little distinction in principle between Corio-
`lis flowmeters of one or two tube design especially
`when in the two tube design the tubes are symmetrical
`and the measurement reference for drive and motion
`
`sensing of one flow tube is the other tube. A single flow
`tube device must use a reference which is not a tube
`with process fluid. It can be a tube without process
`fluid, a blade Spring or the reference can be the housing
`itself. A major consideration is mounting requirements
`to eliminate influence from floor vibrations or pressure
`pulsation in the process fluid. Another major consider-
`ation is that the calibration of the flowmeter does not
`degenerate excessively when the fluid density changes.
`In the past all single flow tube Coriolis mass flowme-
`ters which have had flow tubes with an inside diameter
`larger than k inch and have employed a single flow tube
`design have required extremely complicated mounting.
`Even after being bolted to a ton of concrete such meters
`
`This application is a continuation-in—part of copend-
`ing U.S. patent application Ser. No. 856,939, filed Apr.
`28, 1986, now abandoned.
`DESCRIPTION
`
`Technical Field
`
`This invention is in the field of direct mass flowme-
`ters. More particularly, it concerns a Coriolis-type mass
`flowmeter.
`v
`
`BACKGROUND
`
`A. Mass Flowmeters
`Mass flowmeters (or direct mass flowmeters) have
`sensing means which respond uniquely to mass flow
`rate. Other flowmeters employ, for example, sensing
`means which respond to differential pressure or fluid
`velocity. If one needs to measure mass flow rate with
`such devices one must perform separate measurement
`of density and infer some flow distribution pattern in the
`cross section of the meter and also infer fluid flow pat-
`tern, such as turbulence. They also require Newtonian
`fluid behavior, which is often not met.
`Thus for reason of measurement simplicity alone, the
`direct mass flowmeters are very desirable. Additionally,
`other flowmeters generally lend themselves much bet-
`ter to volume flow rate measurement (gallons per min-
`ute or liter per second) than to mass flow measurement
`(tons/hour or kilograms/second). In practice the mass
`flow measurement is much more useful because chemi-
`cal reactions require blending of proportional mass (not
`volume) of ingredients and product
`specifications
`mostly refer to mass percentage of ingredients not vol-
`» ume percentage. Thus this represents another major
`advantage of direct mass flow measurement over other
`techniques.
`Coriolis flowmeters are direct mass flowmeters. They
`employ the principle of the Coriolis force and use the
`influence of a pattern of such forces upon a flow tube
`carrying the fluid within the meter. Devices disclosed
`to date employ one or two flow tubes which may split
`the fluid stream and carry a fraction each or may carry
`the fluid stream serially through both tubes. The flow
`tubes are typically vibrated by magnetic force coupling
`between a drive coil and permanent magnet, one or
`both of which are attached to a flow tube. To permit
`attachment to outside pipes the end of the flow tubes do
`not participate in the vibration.
`For each part of a flow tube which is momentarily
`not parallel with the axis of rotation for the element, a
`Coriolis force is produced. The force acts through the
`body of the fluid, which will produce pressure on the
`flow tube wall. The magnitude of the Coriolis force is
`proportional to the mass flow rate, the angular velocity
`of rotation and the sine of the angle between flow direc-
`tion within the element and the direction of the rotation
`vector.
`
`Under the aggregate of Coriolis forces upon the dif-
`ferent parts of the flow tube the flow tube will have
`motion in addition to the motion caused by the drive
`(vibrating) motion. “Motion” in this application is used
`to describe position, velocity, acceleration of a point or
`aggregate of points on the flow tube or any time-deriva-
`tive or time-integral of these variables. Over the time of
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`55
`
`60
`
`65
`
`10
`
`10
`
`
`
`4,823,614
`
`3
`have been reported to be unduly influenced by floor
`vibration in an ordinary industrial plant. Double flow
`tube design of similar capacity functions properly in this
`environment. There is thus an advantage in double tube
`design.
`the double tube design has significant
`However,
`drawbacks. It is more costly and may require a flow-
`splitting section and a flow-combining section made out
`of cast bodies and having the flow tubes attached by
`welding. This involves extra cost compared with single
`flow tube design. It also introduces a physical hazard
`due to welding attachments which are far more prone
`to stress corrosion than the flow tube material itself.
`
`Pressure drop is a major factor in many Coriolis flow-
`meter applications. These meters have become widely
`used for highly viscous fluids and thick slurries, for
`example, asphalt, latex paint and peanut butter. In order
`to keep the pressure drop compatible with the pumping
`capacity in the line, it is necessary in such applications
`to work with low mass flow rates. It is also necessary
`that the flowmeter does not introduce a pressure drop in
`excess of available pump capacity. In other words, it is
`of interest to employ a flowmeter with a large diameter
`and short flow tube but still seeing a low mass flow rate.
`This may introduce a sensitivity requirement beyond
`the capacity of all current Coriolis flowmeter designs.
`Acoustic waves generated by pumps and other pro-
`cess equipment can cause considerable deterioration of
`Coriolis flowmeter measurements especially if these
`waves are periodic and have frequencies in the domain
`of the drive frequency or the natural frequency of the
`mode shape closest to the bending caused by the Corio-
`lis forces. Frequent transient random acoustic disturb-
`ances may cause similar problems. The flowmeter may
`lose the ability to distinguish between motion caused by
`such disturbance from a flow rate change.
`The many restrictions stated above are major factors
`in Coriolis flowmeter design. The present invention
`addresses all of these restrictions.
`Descriptions of Coriolis flowmeters often state that
`the vibration maintained is at a “natural frequency” of a
`mode of free (unforced) vibration. In contrast, forced
`vibrations exhibit the phenomenon of resonance imply-
`ing maximum response to the driving force. The reso-
`nance frequency fora structure of very low damping
`(typical for most Coriolis flowmeters) is almost the
`same but not exactly the same as its natural frequency at
`the proper mode. The magnitude of existing damping
`and the method of forcing determine the difference.
`Since all Coriolis flowmeters are exposed to some force
`to maintain the vibrations, there must necessarily be a
`difference between the exact natural frequency and the
`one which is actually achieved. In order to simplify the
`description this small distinction will be ignored and
`where disclosures have been made stating that the struc-
`tures are vibrated at a “natural frequency” this will be
`considered equivalent with operation at resonance with
`some selected mode.
`B. Prior Patents
`
`The advantages of the Coriolis flowmeter principle
`has stimulated the development of many patents.
`Among them are the following:
`US. Pat. No. 3,329,019 (Sipin) discloses two straight
`single tube Coriolis mass flowmeter embodiments. The
`preferred embodiment (FIGS. 2, 3, 4 of the patent)
`employs a driving force in the center of a uniform flow
`tube using a fixed frequency, mechanical drive. The
`bending produced by the drive motion is of single polar-
`
`5
`
`10
`
`15
`
`20
`
`25
`
`3O
`
`35
`
`45
`
`50
`
`55
`
`60
`
`65
`
`4
`ity along the whole flow tube at every instant of opera-
`tion (ignoring the effect of Coriolis force bending). The
`flow tube is semi-pinned at the ends with blade-springs.
`Strain gauges on these springs sense the tube motion
`and the difference between the strain gauge signals is
`computed by a bridge circuit. This difference represents
`‘ flow and is displayed on a meter.
`One major drawback of the Sipin meter is that the
`mechanical drive forcing vibration of a low-damped
`mechanical system will introduce complex signal wave-
`forms with resulting poor measurement sensitivity. The
`sensitivity of measurement using that drive bending
`pattern is low due to relatively low angular velocities
`produced along the beam with small Coriolis force
`magnitudes. If sensitivity is enhanced by driving near
`the first mode natural frequency, the mechanical system
`would become excessively irregular causing huge dis-
`turbance levels on the signals. A further drawback is
`that the meter is highly sensitive to amplitude and fre-
`quency of the drive motion making overall performance
`prohibitively low. Another drawback is the seals which
`are employed for attaching the flow tubes to the meter
`body in order to give the springs freedom of action.
`Under process conditions such seals would be prone to
`leak.
`
`The major differences between this Sipin device and
`the present invention are:
`1. The present invention drives the beam in resonance
`with a higher anti-symmetric frequency mode. This
`requires much less driving force and power consump-
`tion and also furnishes smooth operation.
`2. The present invention drives the beam with oppos-
`ing polarity on each half of the beam at every instant in
`time. The preferred embodiment of the invention uses a
`nonuniform beam cross section profile and mass distri-
`bution. Combined with feature (1) this results in an
`order of magnitude higher sensitivity. This design is
`insensitive to drive amplitude and frequency changes.
`The second embodiment in the Sipin patent (FIGS. 7,
`8, 9 of the patent) employs a single, straight, uniform
`flow tube which is attached with bellows to inlet and
`outlet pipe sections. It is vibrated in the center with an
`electromagnetic drive at constant frequency. The drive
`motion represents a single polarity waveform along the
`tube at any instant in time. Two coil/magnet sensors
`mounted on the flow tube near the bellows produce
`velocity signals. These signals are fed to two amplifiers,
`one of which forms the sum, the other the difference
`between the velocity signals. The difference is em-
`ployed as a measure of the makss flow rate. The sum is
`a measure of drive amplitude and is used in a feedback
`control for drive amplitude control.
`The second Sipin embodiment has reduced calibra-
`tion sensitivity to drive amplitude due to feedback con-
`trol. However, the flowmeter calibration accuracy will
`depend on the stability of the control loop to hold the
`amplitude within the performance specifications typi-
`cally required for Coriolis flowmeters. Furthermore,
`the meter is sensitive to the constancy of the drive fre-
`quency regulation and unavoidable inaccuracy will
`directly impact flowmeter calibration accuracy. The
`meter depends on a soft bellows for obtaining sensitiv-
`ity. Bellows tend to introduce very large calibration
`sensitivity to fluid static pressure change and this would
`create a major practical problem. With harder bellows
`the sensitivity of this design pattern is much less than
`the present invention.
`
`11
`
`11
`
`
`
`4,823,614
`
`6
`deflection pattern is predominantly in a mode which
`does not have a widely different natural frequency.
`Major differences between the disclosure and a basic
`aspect of the present invention are that the invention
`uses substantially straight flow tubes rather than U
`tubes and that the invention forces vibration in the sec-
`ond (or higher) mode not the lowest mode with respect
`to rotation around the mounts. Finally the present in-
`vention is independent of changes in damping.
`U.S. No. Re 31450 (Smith) covers a commercial
`product, now withdrawn, produced by Micromotion
`Inc. The primary embodiment (FIGS. 1-8 of the patent)
`employs a single U-shaped flow tube with cantilever
`attachment to a fixed mount at the ends of the flow tube.
`An electromagnetic drive using peak detector feedback
`vibrates the flow tube with parallel drive motion of
`each leg of the U except for the influence of Coriolis
`forces. The drive coil is mounted on a blade spring and
`the drive magnet on the flow tube. The blade spring is
`manually adjusted to match the natural frequency of the
`flow tube in its first vibrating mode when the flow tube
`is filled with a fluid of a particular density. If the fluid
`had a different density, a different adjustment
`(by
`weight modification) must be done. The flow tube is
`vibrated at the resonance frequency of its first natural
`mode of vibration. Optical sensors measure the twist of
`the U tubes at symmetrical points on the legs similar to
`Cox & Gonzales. The sensor signals are an on/off-type
`using the effect of a shadow of blades mounted on the
`legs. When the shadow of the first leg arrives at the
`relaxed position (midplane) it triggers a photodetector
`to start a counter. When the other leg triggers its photo-
`detector, the count is taken. The time differential for the
`midplane arrival is used for proportional determination
`of the mass flow rate. The tiwsting motion of the U
`coincides with the motion occurring during free natural
`vibration in mode with higher frequency than the drive
`frequency.
`The main drawback of the method described by U.S.
`No. Re 31450 is that it depends on the housing as posi-
`tion reference. Any gradual warping of this housing due
`to stress release or other reason would bring in a sus-
`tained calibration shift. The instrument is extremely
`sensitive to floor and outside pipe vibrations and re-
`quires rigid attachment to huge weights such as con-
`crete blocks. The assumption of linear relationship be-
`tween flow rate and midplane time differential applies
`only for small
`time differentials which restricts the
`range of measurement. The design is also sensitive to
`fluid density change and natural frequency change due
`to changes in fluid temperature and pressure. Drift of
`the feedback-controlled drive amplitude will also cause
`calibration errors.
`‘
`U.S. Pat. No. 4,422,338 (Smith) covers a meter sold
`by Micromotion Inc. under the name C-Model. This
`patent discloses a design which is, in principle, identical
`to the preferred embodiment of U.S. No. Re 31450. The
`difference is that the optical sensors have been replaced
`by magnet/coil velocity sensors and that
`the blade
`spring for drive counterbalance has been replaced by an
`empty flow tube. The signals from the velocity sensors
`are integrated and amplified into square waveform. This
`provides waveform and phase identity with the signals
`produced by the optical “shadow sensors” employed in
`the No. Re 31450 design. The signals are used in the
`same manner to determine the difference in arrival time
`at the midplane. The drive system in this disclosure is
`the same as disclosed in No. Re 31450. In order to elimi-
`
`5
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`55
`
`60
`
`65
`
`5
`The differences between the second Sipin embodi-
`ment and the present invention are:
`l. The present invention uses opposing polarity of
`drive position and velocity for each half of the beam.
`This invention measures the response in areas of the
`beam where the wave shape of the Coriolis force in-
`duced motion is similar to the bending mode for the
`lowest natural frequency giving far higher sensitivity.
`This invention is fundamentally independent of both
`drive amplitude and frequency regardless of the accu-
`racy or absence of control of either variable. The pre-
`ferred embodiment of this invention uses a nonuniform
`cross section profile and mass distribution. This inven-
`tion does not employ a bellows.
`U.S. Pat. Nos. 3,355,944 and 3,485,098 (Sipin) show
`single tube meters with some curvature as well as full
`shaped U-tubes in different embodiments. All employ a
`central drive, implying single polarity deviation of the
`flow tube from equilibrium at all times. Again this is
`exactly opposite to the philosophy of the present inven-
`tion. An embodiment without bellows is shown. These
`embodiments depend on flow tube relaxed position
`curvature.
`
`U.S. Pat. No. 4,109,524 (Smith) discloses a three sec-
`tion flow tube with a center section connected with
`bellows to the outer sections. Each section is straight
`and uniform and all three have coinciding central axes.
`A fixed frequency, mechanical drive in the center
`moves the sections so that at any instant of time the
`deflection from the center position is of the same polar-
`ity for all points on all three sections. This is opposite to
`the present invention. The Smith design employs force
`balance repositioning of the center beam. The torque is
`measured each time the central beam passes through its
`central position. The magnitude of that torque is em-
`ployed as a measure of the mass flow rate. U.S. Pat. No.
`4,109,524 has the same differences from the present
`invention as the meters of the Sipin patents. Addition-
`ally, the preferred embodiment of the present invention
`does not employ torque measurement and it does not
`perform “snapshot observation” at a central position.
`U.S. Pat. No. 4,127,028 (Cox & Gonzales) discloses a
`double flow tube meter design. The flow tubes are of
`identical shape and construction. Both are U-shaped but
`with the legs of the U drawn together. They are
`mounted in parallel, cantilevered fashion on a fixed
`mount. The fluid flows through in the same direction
`through both U tubes. The drive is electromagnetic
`with a drive coil on one U (at the bottom point of the
`bight end) and a magnet on the other U in the equireso-
`nance frequency associated with the lowest frequency
`mode with vibration around the mounting point. The
`Coriolis forces twist the U tubes so that they are no
`longer plane. This motion element corresponds to an-
`other vibrating mode. The U tubes are shaped so that
`the natural frequency of the response mode is nearly the
`same as the drive resonance frequency. The objective is
`to enhance the response magnitude. A drawback not
`discussed by the patentees is that the response will be-
`come dependent on the natural damping which is intro-
`duced by fluid/flow tube interaction and internal crys-
`tal motion in the tube walls. Coil/magnet sensors mea-
`sure the velocities of the sidelegs of the U tubes with
`respect to each other at chosen symmetrical locations
`on each leg.
`The Cox & Gonzales meter is similar to the present
`invention in the use of forced vibration at a natural
`frequency and a structural design where the response
`
`12
`
`12
`
`
`
`4,823,614
`
`7
`nate drift which could (and does) occur in the integra-
`tors, these circuits are limited to the frequency domain
`of the drive frequency and no integration is performed
`at a low or zero frequency. Except for the freedom from
`gradual shift in housing position this design has all the
`other drawbacks pointed out for the design of No. Re
`31450.
`Major differences between the two Smith patents just
`described and the present
`invention are as follows:
`First, an aspect of this invention uses substantially
`straight flow tubes, not U tubes. This invention drives
`the flow tube at a resonance frequency corresponding
`to a high anti-symmetric mode not to its first mode. This
`invention does not employ any time measurement in its
`preferred embodiment. In an alternative embodiment,
`this invention uses time measurement for phase detec-
`tion without reference to a midplane location. In this
`embodiment this invention does not assume linear rela-
`tionships. This invention in its preferred embodiment
`does not use a blade spring or empty flow tubeas coun-
`terbalance in the drive force application. This invention
`uses a detection method which eliminates all sensitivity
`to fluid density, inaccuracy in drive amplitude regula-
`tion, and drive frequency change.
`U.S. Pat. No. 4,491,025 (Smith & Cage) describes
`meters that are sold by Micromotion Inc. under the
`designation “D-Model”. The meters use two U tubes
`similar to Cox & Gonzales but the legs are straight in
`the vicinity of the mount and over most of the sides of
`the tubes. It uses an electromagnetic drive acting be-
`tween the tubes creating relative motion. Magnet/coil
`velocity sensors determine the relative motion between
`the sensors as in the Cox & Gonzales design. Some
`elements of the meters of U.S. Pat. No. 4,422,038 are
`incorporated in the preferred embodiment of U.S. Pat.
`No. 4,491,025, namely the integrators and amplifiers
`which convert the resulting position signals to square
`waves. These square waves control a counter for mea-
`surement of time differential at arrival at “the respective
`midplanes” of the tubes. The devices sold by Micromo-
`tion which are stated to be covered by this patent do not
`have any feature for “midplane” or “relaxed position”
`determination but simply trigger the counter for timing
`when the integrated velocity signals pass at some preset
`and constant deviation from zero. The patent also dis-
`closes “plenums” or two small chambers which are
`provided at the inlet and outlet of the flow tubes. The
`fluid is split into two streams and recombined with a
`plenum handling each of those functions. No physical
`mechanism explaining how the plena improve the oper-
`ation is furnished. They do not allow wave bypass or
`attenuation by a tuned hydraulic circuit as disclosed in
`the present application.
`The meter of U.S. Pat. No. 4,491,025 has the same
`limitations as that of U.S. Pat. No. 4,422,338 except that
`the anchoring of the flowmeter to a huge mass of mate-
`rial is not necessary.
`U.S. Pat. No. 4,559,833 (Sipin) describes a commer-
`cial Coriolis flowmeter sold by Smith Meter Company.
`It employs single or double, parallel flow tubes in differ-
`ent embodiments. The flow tubes are S-shaped. The
`drive force is applied at the center of the S. Sensing
`devices are mounted near the top and bottom of the S.
`In one embodiment
`the sensors are optical on/off
`switches and difference in arrival time at a fixed posi-
`tion is used as a measure of flow rate. Another embodi-
`ment employs analog deflection sensors and the differ-
`ence of the two transducer signals is used to determine
`
`10
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`8
`the flow rate. A counterbalancing spring for the drive is
`also presented as an additional embodiment of the drive
`system.
`In contrast to U.S. Pat. No. 4,559,833, the present
`invention does not use an S-shaped conduit as a flow
`tube. In the preferred embodiment, this invention uses
`the difference in separately located motion sensor sig-
`nals. However, the present invention creates a unique
`previously undetected advantage by a particular combi-
`nation of such differential signals with other mathemati-
`cal operation. This gives major advantages over the
`design disclosed in U.S. Pat. No. 4,559,833 in terms of
`independency to density, drive frequency and ampli-
`tude shifts.
`
`U.S. patent application Ser. No. 775,739 (in the name
`of the present applicant) describes commercial products
`sold by Exac Corporation. That application describes
`single and multi flow tube design where each flow tube
`has a helical design (cross-over loop). The drive vibra-
`tion is at the lowest resonance frequency of the struc-
`ture. The Coriolis forces twist the loop and produce a
`response predominantly in the third mode of natural
`vibration. In one embodiment sensors on each side of
`the loop are used for differential phase measurement
`using nonlinear relationship including tangent function.
`This embodiment requires determination of drive fre-
`quency to be used in the measurement algorithm. It
`employs a temperature sensor attached to the flow tube
`to furnish compensation for fluid temperature change.
`Another disclosed sensing embodiment uses a position
`or velocity measurement between two tube sections at
`the crossover point. Electromagnetic dampeners are
`presented for restricting loop vibrations. A disclosure is
`made of a velocity feedback control loop for continuous
`feedback regulation of loop velocity in the direction of
`the drive.
`U.S. Pat. No. 4,660,421, issued to Dahlin et al., also
`describes commercial products sold by Exac Corpora-
`tion. The application expands on a special version of the
`helical loop design described in Ser. No. 775,739. The
`latter discloses a general helical loop which might have
`a circle as projection on a certain plane or have a pro-
`jection of any other shape. The former designates a
`projected shape with elongation in the direction of the
`opposed situated inlet and outlet flanges. It also shows
`the usage of a horseshoe magnet and coil as an embodi-
`ment of local velocity sensing.
`The present invention is different from the meters
`described in the referenced U.S. patent applications due
`to the absence of crossover loops. The present invention
`operates with a drive frequency which is the resonance
`frequency for a higher mode than the mode in which
`the response to the Coriolis forces occurs, which is
`opposite to the teachings of Ser. No. 775,739 and U.S.
`Pat. No. 4,660,421. The instant invention accomplishes
`in its preferred embodiment frequency, independency
`without the complexity of explicit measurement of that
`frequency. This invention is also fundamentally inde-
`pendent of frequency instead of relying on a'particular
`approximate formula for adjustment
`to frequency
`change. The present invention has an additional advan-
`tage over the devices of Ser. No. 775,739 and U.S. Pat.
`No. 4,660,421, namely that it has fundamental indepen-
`dency to drive amplitude shift in contrast to an indepen-
`dency which is valid only as long as constancy in pulse
`wave form is maintained.
`C. Related Product Literature
`
`13
`
`13
`
`
`
`9
`literature dis-
`The following commercial product
`closes technology related to the present invention.
`Hewlett-Packard