MXPA01000510A

MXPA01000510A - Parameter sensors for vibrating conduit utilizing normal modal decomposition

Info

Publication number: MXPA01000510A
Application number: MXPA/A/2001/000510A
Authority: MX
Inventors: Timothy J Cunningham; Stuart J Shelley; David F Normen; Gary E Pawlas
Original assignee: Micro Motion Inc
Priority date: 1998-07-16
Filing date: 2001-01-15
Publication date: 2001-12-04

Abstract

A plurality of motion signals is received representing motion at a plurality of locations of a vibrating conduit containing material. The received plurality of motion signals is processed to resolve the motion into a plurality of real normal modal components. A process parameter is estimated from a real normal modal component of the plurality of real normal modal components. The motion signals may be processed by applying a mode pass filter to produce an output that preferentially represents a component of the motion associated with a real normal mode of the vibrating conduit. A process parameter may be estimated from the filtered output using conventional phase difference techniques. Real normal modal motion is estimated from the received plurality of motion signals, and a process parameter is estimated from the estimated real normal modal motion. For example, motion may be estimated in respective first and second real normal modes, the second real normal mode being preferentially correlated with a Coriolis force.

Description

VIBRATORY DUCT PARAMETER DETECTORS, OPERATING METHODS AND COMPUTER PROGRAM PRODUCTS THAT USE REAL NORMAL MODAL DECOMPOSITION BACKGROUND OF THE INVENTION FIELD OF THE INVENTION The present invention relates to process parameter detectors, methods of operation and computer program products, and more particularly to vibratory duct parameter detectors, method of operation and computer program products.

ESTABLISHMENT OF THE PROBLEM Coriolis effect mass flowmeters are commonly used to measure mass flow and other information for materials flowing through a conduit. Exemplary Coriolis type flowmeters are described in U.S. Patent Nos. 4,109,524 of August 29, 1978, 4,491,025 of January 1, 1985 and Re. 31,450 of February 11, 1982, all to J. E. Smith et al. These flow meters typically include one more conduit having a straight or curved configuration. Each conduit can be seen as a set of vibration modes including, for example, Ref: 126195, single, torsional, radial and coupled bending modes. In a typical mass flow application, each conduit is driven to oscillate at a resonance in one of its natural modes as material flows through the conduit. The vibration modes of the system filled with vibrating material are altered by the combined mass and stiffness of the ducts as well as the characteristics of the material flowing within the ducts. A typical component of a Coriolis flux meter is the activation or excitation system. The activation system works to apply a periodic physical force to the duct that causes the duct to oscillate. The drive system typically includes at least one actuator mounted in the conduit or conduits of the flow meter. The actuator may comprise one of many well-known electromechanical devices, such as a moving coil device having a magnet mounted on a first conduit and a fixed coil mounted on a second conduit, in an opposite relationship to the magnet. An impeller typically applies a periodic wave, for example a sine or square wave, which drives the signal to the drive coil. The periodic driven signal causes the actuator to drive the two conduits in an opposite periodic pattern. When there is indeed a "zero" flow through the conduit of a driven flow meter, the points along the conduit tend to oscillate approximately with the same phase or a "zero flow" phase with respect to the impeller, depending on the mode of flow. the driving vibration. As the material begins to flow from the flowmeter inlet, through the conduit and out of the flux meter outlet, Coriolis-type forces arising from the material flow tend to induce phase shifts between spatially separated points along the conduit , with the phase on the inlet side of the duct generally delaying the actuator and the phase on the outlet side of the duct generally ahead of the actuator. The induced phase shift between the two positions in the conduit is approximately proportional to the mass flow rate of the material through the conduit. Unfortunately, the precision of the measurements obtained using phase shift or time delay methods can be compromised by lack of linearity and asymmetries in the structure of the flow meter, as well as by vibration introduced into the structure of the flowmeter by external sources such as bombs These effects can be reduced, for example, by the use of balanced mechanical designs that reduce the effects of external vibration and by utilizing frequency domain filtering to remove the frequency components associated with undesirable vibrations. However, mechanical design approaches can be restricted by geometric considerations and frequency domain filtering may not be effective to remove the unwanted vibrational energy that occurs at or near the resonant frequencies of interest, for example, the frequency booster used to excite the conduit.

BRIEF DESCRIPTION OF THE INVENTION In light of the above, an object of the present invention is to provide vibratory duct parameter detectors, operating methods and computer program products which can provide accurate estimates of process parameters in detector ducts having structural non-linearities and asymmetries, and in the presence of external vibration. This and other objectives, features and advantages are provided by the vibration duct process parameter detectors, operating methods and computer program products in which the motion signals representing motion of a vibratory duct are processed to separate the movement of the conduit in a plurality of real normal modal components from which a process parameter such as mass flow can be estimated. In one embodiment of the present invention, a mode pass filter is applied to the mode and movement signals to produce an output that preferentially represents a duct movement component associated with one or more actual normal modes, for example, one or more real normal modes that are preferably correlated with the Coriolis force associated with a material in the conduit. An estimate of a process parameter such as mass flow can then be generated from the output using, for example, conventional phase difference techniques. In another embodiment according to the present invention, the actual normal modal movement is estimated from the plurality of movement signals and a process parameter is estimated from a subset of the estimated normal normal modal movement, for example, from of motion in real normal modes that correlate preferentially with Coriolis force. Because the movement of the conduit is separated into real normal modal components, more accurate estimates of the process parameters can be obtained. For example, a mode pass filter can pass duct movement components that correspond to the actual normal modes that correlate closely with Coriolis forces, while attenuating duct movement components associated with external noise sources. The output filtered in this way can be less corrupted by noise and vibration and therefore can be used advantageously to generate an accurate estimate of a process parameter such as mass flow. Similarly, the estimated normal normal modal movement of selected modes correlated with the Coriolis force can be used to generate an accurate estimate of a process parameter while eliminating modal movement that is attributable to other sources. In particular, according to the present invention, a process parameter detector for determining a process parameter includes a conduit configured to contain material, and a plurality of operating transducers operating to produce a plurality of movement signals that represent the movement. in a plurality of duct positions. A real normal modal splitter responds to the plurality of motion and operating transducers to process the plurality of motion signals to separate the movement represented by the plurality of movement signals into a plurality of real normal modal components. A process parameter estimator responds to the actual normal modal splitter and is operative to estimate a process parameter from the actual normal modal component of the plurality of real normal modal components. In one embodiment of the present invention, the actual normal modal splitter comprises a step filter operable to produce an output from the plurality of motion signals that preferentially represent a component of the movement associated with the actual normal mode of the duct, for example, a real normal mode that correlates preferentially with the Coriolis force. The process parameter estimator responds to the mode pass filter and is operative to estimate a process parameter from the output. In another embodiment of the present invention, the actual normal modal splitter comprises a real normal modal movement estimator operative to estimate the actual normal modal movement of the received plurality of motion signals. The process parameter estimator is operative to estimate a process parameter from the estimated normal normal modal movement, for example, from the movement estimated for a set of real normal modes correlated strictly with Coriolis force. The actual normal modal movement estimator may comprise a means for estimating the movement in a first real normal mode, and a means for estimating the movement in a second normal normal mode that correlates preferentially with a Coriolis force. The process parameter estimator may include a means to normalize the estimated movement in the second real normal mode with respect to the movement estimated in the first real normal mode to produce a normalized estimate of movement in the second normal real. A means can be provided for estimating a process parameter from the normalized movement estimate in the second real normal mode. According to the method aspects of the present invention, a plurality of movement signals are received which represent the movement in a plurality of positions of a vibratory duct containing material. The received plurality of movement signals is processed to separate the movement into a plurality of real normal modal components. A process parameter is estimated from a real normal modal component of the plurality of real normal modal components. According to a method aspect of the present invention, the motion signals can be processed to produce an output that preferentially represents a component of the movement associated with the actual normal mode of the vibratory duct. A mode pass filter can be applied to the plurality of motion signals and a process parameter can be estimated from the filtered output. For example, first and second centered signals representing movement in the first and second respective positions of the conduit can be produced. A process parameter can be estimated by determining a phase difference between the first filtered signal and a second filtered signal and a mass flow estimate from the determined phase difference. According to another aspect of the method of the present invention, the actual normal modal movement, that is, the movement in a plurality of single release degree systems (SDOF), is estimated from the received plurality of motion signals. A process parameter is estimated from the estimated normal normal modal movement. For example, a movement can be estimated in the respective first and second normal modes, the second normal mode is correlated preferentially with a Coriolis force. A process parameter can be estimated by normalizing the movement estimated in the second real normal mode with respect to the movement estimated in the first real normal mode to produce a normalized motion estimate in the second real normal mode. A process parameter can be established from the normalized movement estimate in the second real normal mode. In accordance with the other aspects of the present invention, a plurality of real normal modes are excited in a conduit, a plurality of motion signals are received which represent the response movement to the excitation and a plurality of motion signals are processed to separate the movement of the conduit in a plurality of real normal modal components from which a process parameter can be estimated. The excitation applied to the duct can be a wide band excitation, such as a series of substantially coherent excitations of varying frequencies or the excitation produced when transferring energy from a material in the conduit through, for example, fluid-structure interaction (FSI). A computer program product for estimating a process parameter according to the present invention includes a first computer readable program code means for processing a plurality of motion signals representing the movement of a conduit containing material to separate movement in a plurality of real normal modal components. A second means of computer readable program code estimates a process parameter from a real normal modal component of the plurality of real normal modal components. In a first embodiment, the first computer readable program code means comprises a computer readable program code means for processing the plurality of motion signals to produce an output that preferentially represents a movement component associated with the actual normal mode of the vibratory conduit. The second computer readable program code means comprises a computer readable program code means for estimating a process parameter from the output. In another embodiment, the computer readable program code means comprises a computer readable program code means for estimating actual normal modal movement from a plurality of moving signals. The second computer readable program code means comprises a computer readable program code means for estimating a process parameter associated with the material from the estimated actual normal modal movement. In this way, improved process parameter estimates can be provided.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 illustrates an exemplary process parameter detecting conduit structure. Figure 2 illustrates one embodiment of a process parameter detector, according to the present invention. Figure 3 illustrates another embodiment of a process parameter detector according to the present invention. Figure 4 illustrates an exemplary embodiment of a mode phase filter, in accordance with the present invention. Figure 5 illustrates an exemplary embodiment of a process parameter estimator, according to the present invention.

Figures 6 and 7 illustrate illustrative exemplary operations according to mode pass filtering aspects of the present invention. Figures 8-12 illustrate modal magnitude effects for a process parameter detector conduit structure. Figure 13 illustrates another embodiment of a process parameter detector, according to the present invention. Figure 14 illustrates an exemplary embodiment of a real normal modal motion estimator in accordance with the present invention. Figure 15 illustrates an exemplary embodiment of a process parameter estimator, in accordance with the present invention. Figures 16 and 17 illustrate exemplary operations for estimating a process parameter from estimated actual normal modal movement, in accordance with aspects of the present invention. Figures 18A and 18B illustrate real normal modes excited by broadband excitation arising from a fluid structure interaction (FSI).

DETAILED DESCRIPTION OF THE MODALITIES The present invention will now be described in greater detail in the following, with reference to the accompanying drawings, in which embodiments of the invention are shown. Those skilled in the art will appreciate that the invention may be embodied in many different forms and should not be considered as limited by the embodiments set forth herein; rather, these embodiments are provided so that this description will be thorough and complete and will encompass the entire distance of the invention to those skilled in the art. In the drawings, reference numerals and the like refer to similar elements therethrough. The following discussion largely refers to Coriolis type flowmeters in which a process parameter of a material processing system is estimated, for example the mass flow velocity for a material such as a fluid flowing through a vibratory conduit. Those skilled in the art will appreciate, however, that the present invention is also applicable to vibration duct process parameter detectors other than on-line detectors. For example, the present invention is applicable to vibratory tube densitometers of the sampling type which include a conduit configured to contain a sample of a material extracted from a material processing system. In the embodiments described above, the motion signals represent the movement of a detector conduit and are processed to separate the movement of the conduit in a plurality of real normal modal components. The actual normal modal decomposition can be implemented in numerous ways. For example, a mode pass filter can be used to pass the motion components of the detector conduit which are associated with a set of desired actual normal modes while attenuating conduit movement components associated with other real undesirable normal modes. Although the modal responses corresponding to the duct movement do not need to be explicitly determined, nevertheless the model pitch filtering "separates" the duct movement into respective components associated with respective actual normal modes. Alternatively, the actual normal modal movement, that is, the movement in coordinate systems of a plurality of systems of a single degree of freedom (SDOF), can be estimated explicitly from the movement signals and can be used to accelerate process parameter estimates. For example, mode step filtering can be obtained using a two-stage process that involves the application of a modal transformation to transform the conduit movement to the corresponding real normal modal movement and the application of a selective inverse modal transformation to the modal movement actual normal to produce a filtered output that preferentially represents components of the duct movement associated with one or more desired actual normal modes. This filtered output can be processed to estimate a process parameter using, for example, conventional phase difference techniques. According to another aspect of the present invention, a process parameter is determined directly from an estimated actual normal modal movement. For example, a real normal modal movement for a real normal mode that correlates closely with the Coriolis force is normalized with respect to a normal normal modal movement otherwise to generate a scaling factor. The scaling factor is used to estimate the mass flow.

Modal Comfort of a Vibratory Conduit The behavior of a vibratory structure such as a detector conduit can be described in terms of one or more natural modes having associated natural vibration frequencies. The modes and associated natural frequencies can be described mathematically by eigenvectors and associated eigenvalues, the eigenvectors are unique in relative magnitude but not in absolute magnitude and octagonal in relation to the mass and stiffness of the structure. The linear independent set of vectors can be used as a transformation to decouple equations that describe the movement of the structure. In particular, the response of the structure to an excitation can be represented as a superposition of scaled modes, the scaling represents the contribution of each mode to the movement of the structure. Depending on the excitement, some modes may contribute more than others. Some modes may be undesirable because they can contribute energy to the resonant frequency of the desired modes and therefore may corrupt the measurements taken at the resonant frequency in a desired way, such as phase difference measurements taken at the driven frequency. . Conventional flowmeters typically use structural and temporal filtering to reduce the effects of undesirable modes. Conventional structural filtering techniques include the use of mechanical features such as clamp bars designed to decouple in-phase and out-of-phase bending modes, actuators placed so that undesirable modes and transducers are less likely to be placed so that they are less sensitive to undesirable ways. Structural filtering techniques can be very effective in reducing energy in undesired ways, but can be limited to geometric and manufacturing constraints. Temporary filtering techniques typically modify the transducer signals based on the time domain or frequency domain parameters. For example, a typical Coriolis flowmeter may include frequency domain filters designed to remove frequency components that correlate significantly with undesired modes. However, the energy out of resonance of the unwanted modes can contribute considerably to the energy in the resonant frequency in a desired way. Because frequency domain filters are generally not effective in differentiating the contribution of multiple modes to a given frequency, the contribution of unwanted modes to a measurement frequency can be a major source of error in process parameter measurements . A detector conduit structure with negligible damping and zero flux can be assumed to have a purely normal natural or normal vibration mode, that is, in each mode, each point of the structure simultaneously reaches a maximum displacement, however, a real conduit That it has a non-negligible damping and with material flowing through it generally has a complex excitation response, that is, the points of the structure do not generally reach a maximum amplitude simultaneously. The movement of the conduit structure can be described as a complex mode that has real and imaginary components or, alternatively, components of magnitude and phase Coriolis forces imparted by the material that flows introduce complexity into the movement of the detector conduit. Even if it is complex, the movement of a conduit structure can be described as a superposition of scaled or "normal" natural modes, such as the real and imaginary parts of a complex mode that are linearly independent by definition. To represent complex motion, complex scaling coefficients are used to combine the actual constitutive normal modes. The particular real normal modes can be closely correlated with the imaginary component of the complex mode as long as they correlate less significantly with the real component of the complex mode. In consecuense, these particular real normal modes can correlate more closely with the Coriolis forces associated with the material in the detector conduit and can therefore provide information to generate an accurate estimate of a parameter associated with the material.As an illustrative example, a double curved tube of a 7.6 cm Coriolis flux meter is analyzed. (3 inches). Figure 1 illustrates a conceptual model of the conduit structure of this meter. The conventional speed transducers 105A, 105B, 105C oriented to measure the velocity in a z direction are placed in the respective right, drive and right positions in the conduit assembly 10. The respective accelerometers 105D, 105E are placed in the respective places of the conduits 103A, 103B near the position of the right transducer, and they are oriented to measure the lateral acceleration along a direction x. The outputs of accelerometers 105D, 105E are integrated to produce lateral absolute velocity information. A response vector can be constructed. { x •, respueßstm > from the outputs of the motion transducers 105A-E: right answer, z response of the impulse, z ix r- reply X response response, z response response, xz responselateral, x (1) where the skewed lateral response is a response along a 45 degree direction with respect to the x and z axes. A real normal modal "filter" matrix [f], that is, a real normal modal transformation matrix in relation to the physical movement vector. { x "* pue-.t"? to a vector. { 12.}. of normal normal modal movement can be identified so that: The actual normal modal transformation matrix [f] can be identified using numerous techniques. For example, trial and error or inverse techniques, such as those described in U.S. Patent Application Serial No. 08 / 890,785 filed July 11, 1997, assigned to the assignee of the present application may be used. incorporated herein by reference in its entirety and, if the text is physically present, and in the United States Patent Application entitled "Generalized Modal Space Drive Control for a Vibrating Tube Process Parameter Censure," filed on February 25, 1998, assigned to the assignee of the present application and incorporated as a reference in its entirety if the text is physically present. For the exemplary conduit structure 10 of Figure 1, an actual normal modal transformation matrix [f] was determined experimentally: [f] = (3) from left to right, the columns of the real normal modal transformation matrix [f] represent a first mode of bending out of phase, a lateral mode in phase, a lateral mode out of phase, a mode of torsion out of phase and a second bending mode out of phase, respectively. The matrix [f] of modal transformation can be used to separate the physical movement represented by the motion vector. { xreapacat?) in real normal modal components. For example, equation (2) can be solved explicitly for the modal movement vector. { n} , by premultiplying both sides of equation (2) by the inverse of the matrix [f] of modal transformation: . { »7 } = [f] -l. { x response). (4 ) where, for the exemplary structure of Figure 1, 0.002 0.0165 0.0138 0.0057 0.003 '0.034 0.00008 -0.0846 0.272 0.105 [f] "i = -0.001 0.00008 -0.0387 0.132 -0.067 kg. sec cm -0.176 0.00 0.0167 0.002 0.0008 0. 017 -0 .0103 .0.009 .004 .001 (5) As described in detail aguí, the real normal modal movement. { 17.}. it can be used directly to estimate a process parameter associated with one or more actual normal modes of the conduit structure, for example modes associated with Coriolis force. Alternatively, the modal transformation matrix [f] can be used to identify a pitch filter which can be applied to the physical movement. { xrß.pUest «} to produce a focused physical domain response that preferentially includes components of physical movement. { xrß pußata} associated with one or more conduit modes. This filtered response can be used to estimate a process parameter. Figure 2 illustrates an essential embodiment of a vibratory duct parameter detector 5 according to the present invention. The detector 5 includes a conduit assembly 10. The conduit assembly 10 includes an inlet rim 101, an outlet rim 101 ', a manifold 102 and a first and second conduit 103A, 103B. The clamp bars 106, 106 'connect the conduits 103A, 103B.

Connected to the conduits 103A, 103B is an actuator 104 which is operative to vibrate the conduits 103A, 103B which respond to an impeller 20. A plurality of motion transducers 105A-E are operative to produce a plurality of motion signals which represent the movement in a plurality of positions of the conduits 103A, 103B, for example the signals which represent displacement, velocity or acceleration of the conduits 103A, 103B. The motion transducers 105A-E may include a variety of devices such as elicoidal velocity transducers, optical or ultrasonic motion detectors, accelerometers, initial velocity detectors and the like. The conductors 100 are connected to the actuator 104 and the motion transducers 105A-E. When the conduit assembly 10 is inserted into a material processing system, the material flowing in the material processing system enters the conduit assembly 10 through the entrance flange 101. The material then follows through the multiple 102, where it is directed to the interior of the conduits 103A, 103B. After leaving the conduits 103A, 103B, the material flows back to the manifold 102 and exits the meter assembly 10 through the outlet rim 101 '. As the material flows through conduits 103A, 103B, it results in Coriolis sources altering conduits 103A, 103B.

The conduits 103A, 103B may be driven by the actuator 104 in opposite directions around their respective bending axes - and W'-W, directing what is commonly referred to as the first bending mode out of phase in the conduit assembly 10. The actuator 104 may comprise any of the many well-known devices such as a linear actuator which includes a magnet mounted to the first conduit 103A and an opposite spool mounted to the second conduit 103B. An alternating current induced by a driving signal provided by an impeller 20 via the driver 110 passes through the coil, generating mechanical force that vibrates the conduits 103A, 103B. The excitation provided by the actuator 104 may be substantially coherent, for example, confined to a narrow frequency range or may be wideband, as described in greater detail below. While the parameter detector 5 illustrated in Figure 2 is shown with an integral actuator 104, those skilled in the art will appreciate that the vibration of the conduits 103A, 103B according to the present invention can be obtained by other techniques. For example, wide-band excitation external to the conduit assembly 10 can be generated by sources such as pumps or compressors and can be transported to the conduit assembly 10, for example, via one of the flanges 101, 101 '. Similarly, broadband excitation can be generated by energy transfer from a material in conduits 103A, 103B through a fluid-structure interaction mechanism (FSI), as described in greater detail in the following. A real normal modal splitter 30 responds to the movement transducers 105A-E and is operative to separate the movement of the conduits 103A, 103B represented by the signals on the conductors 111 in a plurality of actual normal modal components. A process parameter estimator 40 responds to the actual normal modal splitter 30 and is operative to generate an estimate 45 of a process parameter from the plurality of actual normal modal components. As described herein, the actual normal modal separator 30 can separate the conduit movement in numerous ways, for example by way filtering the conduit movement mode or by estimating the actual normal modal movement corresponding to the conduit movement.

FILTERED PHASE OF MODE In accordance with one aspect of the present invention, a "mode pass filter" is applied to motion signals that produce an output representing a filtered form of the duct movement in which duct movement components associated with duct are attenuated. the unwanted modes.

The mode pass filter represents a product of the actual normal modal transformation that establishes a map of the movement of the conduit within the movement in a plurality of systems of a single degree of freedom (SDOF), that is, a real normal modal movement, and a normal selective inverse real modal transformation and which establishes the map of the selected portions of the real normal modal movement, that is, the movement in a set of desired real normal modes back to the physical domain. You can use an inverse real normal modal transformation matrix [f] to translate a vector. { ? } of normal normal modal movement to a vector. { Xfíltr? Do} of filtered motion in which the components associated with the actual unwanted normal modes are attenuated: . { ^ filr. * ..}. = [F * 1 i) (6) For the exemplary structure of Figure 1, a normal inverted normal modal transformation [f] matrix is constructed from the real normal modal transformation matrix [f] by substituting the elements of the normal modal transformation matrix [f] real associated with the normal unwanted normal modes with zeros: As shown in equations (6) and (7), the components of the vector. { - «« «^« t ,} of movement of conduit that correspond to the real unwanted normal modes can be attenuated by using a matrix [f] of normal modal transformation, real inverse selective that corresponds to the matrix [f] of real normal modal transformation with zeros substituting those elements of the matrix [f] of actual normal modal transformation associated with the real unwanted normal modes. However, those skilled in the art will appreciate that the attenuation of these components can be obtained by using nonzero values for those elements of the matrix [f *] of normal modal transformation, real inverse selective. By combining equations (4) and (6): x .f il trado. { ? '; .respues a} , (8) where the matrix [?] Of step filter mode is given by [?]) = [f *] tf] "1 (9) Matrix [ ? Step filter mode processes the vector. { xreBpueBta} of duct movement so that the vector. { xfiltr? ds} of filtered output movement preferentially represents components of the vector. { Xraspue t? e duct movement associated with one or more of the desired modes. The matrix [?] Of filter step mode can also be generated by: [?] = [f] A] [f] - \ where [A] is a matrix whose elements outside the diagonal are zeros, with selected diagonal elements that correspond to the desired modes associated with one, for example, The filtered output. { x £ IlCrado} it can be processed to generate accurate estimates of process parameters such as mass flow. For example, the exit. { xfiltr? do} it can be processed according to conventional time difference phase Coriolis measurement techniques. For the exemplary system illustrated in Figure 1, this can be carried out by determining a phase difference between the components of the filtered output corresponding to the right and left transducers 105A, 105C, for example using zero crossing or techniques of similar phase difference is such as those described in U.S. Pat. RE31,450 for Smith, U.S. Patent 4,879,911 for Zolock, and U.S. Patent 5,231,884 for Zolock, or similar techniques of phase difference or of time, using a digital signal processor (DSP) or a similar digital computing device. The information provided by the additional transducers 105B, 105D, 105E can therefore be used to filter out components of the pipeline movement which are associated, for example, with undesired side modes. The number of positions in a detector conduit represented by the movement signals can be deliberately chosen to exceed the number of actual normal modal components within which the conduit movement is separated. In that case, the real normal modal transformation matrix and the normal inverse selective normal transformation matrix have more rows than columns. Consequently, a generalized inverse of the actual normal modal transformation matrix can be used to calculate the mode step filter matrix in equation (9). in this way, the movement signals provide a source on certain information to separate the movement of the conduit in the given number of real normal modes. The estimated process parameters for such determined information are therefore spatially integrated, which provides potentially more accurate estimates. Spatial integration is described in a United States patent application entitled "Improved Vibrating Conduit Parameter Sensors and Methods of Operation Therefore Utilizing Spatial Integration", assigned to the assignee of the present application and filed concurrently with this document. Figure 3 illustrates an exemplary parameter detector 5 which implements the mode pass filter, according to the present invention. The actual normal modal splitter 30 includes a pass filter 330 so as to respond to the motion transducers 105A-E. The pass-through filter 30 separates the movement of the conduits 103A, 103B represented by the movement signals generated by the transducers 105A-E in a plurality of actual normal modal components by producing a filtered output 35 that preferentially represents one or more components of the movement of the conduits 103A, 103B associated with one or more of a plurality of real normal modes, for example actual normal modes associated with the Coriolis force imparted by a material contained in the conduits 103A, 103B. Figure 4 illustrates an exemplary mode of the filter 330 of step mode and the estimator 40 of process parameter. A sampler 432, for example a sampling and holding circuit or the like, provides a means for receiving motion signals 431 of movement transducers, sampling movement signals 431 and producing samples 433 therefrom for subsequent conversion to values 435 of digital signal by an analog to digital (A / D) convert 434. The detailed operations of the sampler 432 and A / D 434 can be performed by many circuits known to those skilled in the art, and need not be discussed in more detail here. Those skilled in the art will appreciate that movement signals 431 can be processed in numerous ways. For example, parasitic wave filtering, post-sampling filtering and similar signal processing can be applied. It will also be understood that, in general, the receiver and conversion means illustrated in Figure 3 can be implemented using special purpose hardware, firmware or software that operates under general purpose or special data processing devices or combinations thereof. . For example, analog-to-digital sampling and conversion functions can be integrated with the 105A-E transducers. The portions of the mode pass filter 330 can be constituted in a computer 50, for example a microprocessor, microcontroller, digital signal processor (DSP) or the like. The computer 50 may comprise, for example, a DSP with tubing especially suited for linear algebraic calculations, such as a DSP of the TMS320C4X family of the DSPs sold by Texas Instruments, Inc. Configured with an appropriate program code, for example software or firmware and data stored, for example, in a storage medium 60 such as a random access memory (RAM), an electrically erasable programmable read-only memory (EEPROM), a magnetic disk or the like, the computer 50 provides a means 436 to generate a motion vector 437 that represents the movement of a detector conduit from digital 435 values. The pass-through filter 330 includes means 438 for multiplying the motion vector 437 by a mode pass filter matrix to produce a filtered motion vector 35 preferentially representing components of the behavioral motion vector associated with one or more of the desired modes. The process parameter estimator 40 can also be increased in the computer 50. Constituted, for example as software or firmware running on the computer 50, the process parameter estimator 40 calculates an estimate 45 of a process parameter, for example , calculates an estimated mass flow rate, for the filtered motion vector 35. As illustrated in Figure 5, for example, the process parameter estimator 40 may include a means 542 for determining a phase difference between the components of a filtered output 35 and a means 544 for estimating a mass flow from the determined phase difference. Figure 6 illustrates the operations 600 for estimating a process parameter according to the filtering aspects of the mode step of the present invention. A plurality of movement signals is received by representing the movement in a plurality of positions of a vibratory conduit containing material from a material processing system (block 610). The received signals are processed to separate the movement of moving conduit, in a plurality of real normal modes by producing an output that preferentially represents a component of the movement of the conduit associated with the actual normal mode, for example, a mode that preferentially correlates with the Coriolis force imparted by a material passing through the conduit (block 620). A process parameter is estimated from the output (block 630).

Figure 7 illustrates operations 700 for estimating a process parameter according to another aspect of step filtering of the present invention. Motion signals are received (blog 710) and processed to generate a motion vector that represents the movement of the detector conduit (blog 720). The motion vector is multiplied by a selective inverse normal normal transformation matrix to generate a filtered motion vector that preferentially represents components of the motion vector associated with one or more real normal modes (block 730). A process parameter is estimated from the filtered motion vector (block 740). It will be understood that the blocks or combinations of blocks in the flow chart illustrations of Figures 6 and 7 can be implemented using a computer readable program code, for example, program or data instructions, or both, operated on a computer. or data processor such as a computer 50 illustrated in Figure 4. As used herein, a computer-readable program code may include, but is not limited to, things such as operating system instructions. (for example, object code), high-level language instructions and the like, as well as data which may be read, accessed or used in some other way along with such program instructions.

The program code may be loaded into a computer or similar data processing apparatus including, but not limited to, a microprocessor, a microcontroller, a digital signal processor (DSP) or the like. The combination of the program code and the computer can provide a device that is operative to implement a function or functions specified in a blog or blog of the flowchart illustrations. Similarly, the program code may be loaded into a computer or data processing device such that the program code and the computer provide a means to carry out the function or functions specified in a block or blocks of a hardware diagram. flow. The program code may also be stored in a computer readable storage medium such as a magnetic disk or tape, a bubble memory, a programmable memory device such as a programmable, electrically erasable read-only memory (EEPROM), or Similary. The stored program code can direct a computer to access the storage medium to operate so that the program code stored in the storage medium forms a manufacturing article that includes a means of program code to implement the program code. Function or functions specified in blog or flowchart blog. The program code can also be loaded into a computer to cause a series of operational steps to be carried out, so a process is implemented so that the program code, together with the computer, provides steps to implement the functions specified in a block or blocks of a flowchart. Accordingly, the blocks of the flow chart illustrations support an operating apparatus for performing the specified functions, combinations of means for carrying out the specified functions, combinations of steps performing the specified functions and a program code means readable in computer constituted in a computer readable storage medium to carry out the specified functions. It will also be understood that, in general, each block of the flowchart illustrations, and combinations of blogs in the flow chart illustrations may be implemented by special purpose hardware, software or firmware executed on a general-purpose computer, or combinations thereof. For example, the blog functions of the flow chart illustrations can be implemented by an application-specific integrated circuit.

(ASIC), a programmable gate arrangement or a similar special purpose device, or program instructions and data loaded and executed by a microprocessor, microcontroller, DSP or other general purpose calculation device.

Estimation of a process parameter from the estimated modal movement According to another aspect of the present invention, a process parameter such as mass flow is estimated directly from the estimated actual normal modal movement, i.e., from an estimate of motion in a model of the detector conduit that It comprises a plurality of systems of a single degree of freedom (SDOF). As discussed in the above, a complex mode can be represented as an overlap of real normal modes scaled by complex scaling coefficients: . { Foo ple3o} = - * -. { «} , (10) where . { Fcoppi.jo } is a vector of complex mode or, [F] is a matrix of the real mode vectors constituent for the vector of complex mode or,. { F00 »p? .j0} Y . { to} is a vector of generally complex scaling coefficients. In a stream of Q-rioli.s. the? rc * p? tr rrt-? The ccr x or detector becomes complex or by the Coriolis acceleration associated with the fluid flowing in the detector conduit. Accordingly, the scaling coefficient for the imaginary part of the complex flow mode is proportional to the flow velocity and can provide information for determining the flow velocity. However, the magnitude of. { or} it is susceptible to an absolute magnitude of the real normal modal movement which, as mentioned before, can be arbitary. Figures 8 and 10 illustrate the actual components (Re [ax]) ml, (Re [ax]) m2 of a scaling coefficient ax corresponding to a first bending mode out of phase in the first and second respective absolute modal magnitudes ml fía, and figures 9 and 11 illustrate imaginary components (Im [a2]) al, (Im [a2]) a2 of a scaling coefficient a2 which corresponds to a first mode of output phase torsion over a range of mass flow velocities in modal magnitudes m? / M2 , for a 7.6 cm (3 inch) double tube Coriolis flowmeter, as illustrated in Figure 1. As can be seen in Figures 8 and 10, the actual components (ReEa) ml, (ReCo) a2 for the first Scale coefficient ax are generally independent of the flow velocity, while Figures 9 and 11 illustrate that the imaginary components of the second scaling coefficient a2 show a substantially linear dependence on the flow velocity. However, both the actual components (ReAa) ml, (Re [a) n2 of the first scaling coefficient or.! as the imaginary components (Im [a2]) ml, (Im [a2]) m2 of the second a2 scaling coefficient depend on the absolute modal magnitude. One aspect of the present invention arises from the realization that because the relative magnitude of the real normal modes representing a complex motion are invariant (assuming there are no structural changes), a flow measurement can be made based on the coefficient a2 of torsion mode scaling that is insensitive to the absolute modal magnitude, by normalizing the imaginary component Im [2] of a2 with respect to the real component Re.a of a ... As illustrated in figure 12, the normalization of the imaginary component curve (Im [a2]) ^ of Figure 9 with respect to the real component provides a curve for the normalized imaginary component (Ip? a2]) m2? normalI2Jldo which is substantially identical to the curve for the component imaginary (ImEa3]) "-., which illustrates that the normalized imaginary component of a2 is generally insensitive to the absolute magnitude of mode. For the exemplary 7.6 cm (3 inch) Coriolis detector described above, a flow rate calibration factor k C τ can be determined at a known flow rate by determining the normalized imaginary component of the first torsion mode scaling coefficient out of phase and dividing it by the known flow rate: Jm [a2] i? E [ax] known ^ cal speed. known To determine a speed velocity of unknown mass velocidadaonoald ?, the normalized imaginary component of the first coefficient of scaling of torsion mode out of phase corresponding to the unknown flow velocity is determined and multiplied by the calibration factor ka? T: unknown velocity = _ - ^^ unknown Rearranging equation (10): . { to } = [F] "1 { F complex J (13) The shape of equation (13) is similar to that of equation (4). By analogy, the duct movement vector corresponds to the unimodal complex motion vector, and the scaling coefficient vector corresponds to the actual normal modal motion vector. Consequently, the mass flow for the exemplary 7.6 cm (3 inch) Coriolis detector can be estimated by determining a ratio of the estimated modal responses for the torsion and bending modes of the conduit. The technique described above can be generalized. For example, higher order estimates can be obtained by using a combination of scaling coefficient for modes associated with normalized Coriolis force with respect to one or more scaling coefficients associated with modes that correlate, for example, with an excitation applied to the detector conduit. For the 7.6 cm (3 inch) Coriolis meter of an exemplary curved tube, this may involve using coefficients associated with higher order torsion and bending modes, respectively. The technique described above is also applicable to different conduit configurations. The bending modes and torsion mode described for the exemplary dual tube Coriolis detector of 7.6 cm (3 inches) can generally be classified as symmetric modes and antisymmetric modes, respectively. Generally speaking, when defining a plane of symmetry orthogonal to the flow axis of a detector conduit, the symmetric modes represent modes in which the movement on the first side of a plane of symmetry is reflected by the movement on the second side of the plane. symmetry plane. For example, the folding modes of the U-shaped Coriolis detector conduit of Figure 1 are symmetrical with respect to the illustrated y-z plane. The antisymmetric modes represent those modes in which the movement of the first side of a plane of symmetry represents a reflection and a phase rotation of the movement on the second side of the plane of symmetry. For example, the torsion modes of the U-shaped detector conduit of Figure 1 are antisymmetric with respect to the y-z plane of Figure 1. The straight tube detectors show similar symmetrical and antisymmetric modes. Accordingly, the present invention also extends to parameter detectors that utilize one or more straight conduits and, in general, to detectors that utilize a variety of conduit configurations. As mentioned above with respect to mode pass filtering, the number of positions in a detector conduit can be chosen to exceed the number of real normal modal movement estimation aspects, this means that the real normal modal transformation matrix EF ] has more rows than columns. To estimate the actual normal modal movement [?], We use a generalized inverse EF] f of the real normal modal transformation matrix [F], that is: . { ? } = [f] '. { ? answer J (14) The resulting estimate of the real normal modal movement is therefore specially integrated, the motion signals provide an overdetermined source of information to separate the movement of the conduit. { ? ra, «puß« t ,} within the real normal modal movement. { ? } . Figure 13 illustrates an exemplary parameter detector 5 for carrying out process parameter estimation directly from an estimate of the actual normal modal movement. A normal modal splitter 30 comprises an actual normal modal motion estimator 1330 responsive to the motion signals produced by a plurality of motion transducers 105A-E. The normal normal modal movement estimator 1330 separates the movement of the conduits 103A, 103B represented by the movement by generating an estimate of the actual normal modal movement. An estimator 40 of the process parameter generates a process parameter 45 estimated from the estimated actual normal modal movement. Figure 14 illustrates an exemplary embodiment of the actual normal modal motion estimator 1330. The actual normal modal estimator 1330 may be constituted in a computer 50, for example, a microprocessor, microcontroller, digital signal processor (DSP) or the like. For example, the computer 50 may comprise an online DSP suitable especially for linear algebraic calculations, such as one of the DSPs of the TMS320C4X family of the DSPs sold by Texas Instruments, Inc. Configured with an appropriate program code, eg software or firmware, or both and data stored, for example, in a storage means 60 such as a random access memory (RAM), an electrically erasable programmable read-only memory (EEPROM) with a magnetic disk or the like, the computer 50 provides a means 1438 for multiplying a motion vector 437 by a modal transformation matrix to produce an estimated actual normal modal motion vector 35. The process parameter estimator 40 can also be implemented in the computer 50. Constituted, for example, as a software or firmware running on the computer 50, the process parameter estimator 40 calculates an estimate 45 of a process parameter, for example. example calculates the mass flow rate from the estimated actual normal modal motion vector 35. As illustrated in FIG. 15, the actual normal modal movement estimator 1330 may comprise a means 1532 for generating an estimate 1533 of motion in a first real normal mode, and a means 1534 for generating an estimate 1535 of motion in a second mode normal real. For example, the first and second modes may correspond to the first bending mode out of phase and the first twisting mode out of phase, respectively, for a double U-shaped tube flowmeter such as that illustrated in Figure 1. process parameter estimator 40 may include a means 1542 to normalize the second normal normal 1535 motion with respect to the first real normal 1533 motion to produce an estimated 1543 normalized motion in the second normal normal mode. The means 1544 for estimating a process parameter responds to normalized means 1542 to produce a process parameter 45 estimated from the 1543 normalized estimate. As illustrated, the responses 1533, 1535 estimated manners can be fed back to an impeller 20 to selectively excite one or more of the selected actual normal modes, as described in the above-mentioned patent application "Generalized Modal Space Drive Control for a Vibrating Tube Process Parameter Sensor ". Figure 16 illustrates the steps 1600 for estimating a process parameter according to an aspect of the present invention. A plurality of motion signals are received which represent the movement of a detector conduit (block 1610). The movement signals are processed to separate the conduit movement in a plurality of real normal modes when estimating the movement in a plurality of real normal modes (block 1620). A process parameter is estimated from the estimated actual normal modal movement (block 1630). Figure 17 illustrates operations 1700 for estimating a process parameter according to another aspect of the present invention. A plurality of movement signals representing movement of a detector conduit are received (block 1710) and a movement vector is generated from them (block 1720). The motion vector is multiplied by a modal transformation matrix to generate a real normal modal motion vector that represents the actual normal modal movement that corresponds to the duct movement (block 1730). A component of the real normal modal movement vector that corresponds to a first mode, for example a mode correlated with a Coriolis force imparted by a material in the detector conduit, is normalized with respect to a component of the actual normal modal movement vector corresponding to a second mode, for example, a substantially uncorrelated mode with the Coriolis force, such as a driving mode (block 1740). From the normalized component of the estimated modal response, a process parameter such as mass flow is determined (block 1750). It will be understood that the blocks or combinations of blocks in the flowchart illustrations of Figures 16 and 17 can be implemented using a computer readable program code, for example program instructions or data, or both operated on a computer or a computer. data processor such as computer 50 illustrated in Figure 14. As used herein, the computer readable program code may include, but is not limited to such things as operating system instructions (e.g., object code) , high level language instructions and the like, as well as data which may be read, accessed or used in some other way along with such program instructions. The program code may be loaded into a computer or similar data processing apparatus that includes, but is not limited to, a microprocessor, a microcontroller, a digital signal processor (DSP) or the like. The combination of the program code and the computer can provide an apparatus that is operative to implement a function or functions specified in a block or blocks of the flowchart illustrations. Similarly, the program code may be loaded into a computer or data processing device so that the program code and the computer provide a means to carry out the function or functions specified in a block or flowchart blog. . The program code may also be stored in a computer-readable storage medium such as a magnetic disk or tape, a bubble memory, a programmable memory device such as an electrically erasable programmable read-only memory (EEPROM), or the like. . The stored program code can direct a computer to access the storage medium for functions so that the program code stored in the storage medium forms an article of manufacture that includes a means of program code to implement the function or functions specified in a blog or blocks of a flowchart. The program code can also be loaded into a computer to cause a series of operational steps to be carried out, so a process is implemented so that the program code, together with the computer, provides steps to implement the functions specified in a blog or blog of a flowchart. Accordingly, the blocks of the flowchart illustrations support an operating apparatus for carrying out the specified functions, combinations of means for carrying out the specified functions, combinations of steps performing the specified functions and a program code means. readable in computer constituted in a computer readable storage medium to perform the specified functions. It will also be understood that, in general, each block of the flowchart illustrations and combinations of blocks in the flow chart illustrations may be implemented by special purpose hardware, software or firmware executed on a general purpose computer, or combinations of them. For example, the functions of the blocks of flowchart illustrations can be implemented by an application-specific integrated circuit (ASIC), a programmable gate array or similar special purpose device, or by program instructions and data loaded on and executed by a microprocessor, microcontroller, DSP or other general purpose computing device.

Wide band excitation The actual normal modal decomposition according to the present invention allows the use of wideband excitation. In fact, broadband excitation may be desirable because it excites both the desired and unwanted actual normal modes which can provide more complete information to identify the actual desired modal responses and generate precision parameter estimates from the same. Wideband excitation may include, but is not limited to, random or frequency sweep excitation. The frequency sweep excitation may be supplied by one or more actuators operatively associated with the detector conduit, for example by applying a series of substantially coherent excitations to a detector conduit at varying frequencies. Random excitation can be applied to a conduit by one or more actuators using, for example, a broadband actuator that drives a signal or pulse of a predetermined duration. Wide-band excitation can also be provided by random environmental forces applied to a detector conduit, for example by energy transfer from a material contained in the detector conduit, for example fluid-structure interaction (FSI) or transported vibrations of some other way to the conduit structure for pumps, compressors and the like. The efficiency of broadband excitation using ambient excitation such as that provided by FSI has been investigated using a CFM300 double-tube Coriolis flowmeter manufactured by Micro Motion, Inc. which operates in a passive manner, i.e. without excitation from of an actuator. Figures 18A and 18B illustrate amplitudes at frequencies that correspond to the actual normal modes of the detector at mass flow rates of 90.7 kg per minute and 272 kg per minute (200-600 pounds per minute), respectively. As can be seen from these figures, the amplitudes at frequencies corresponding to the first out-of-phase torsion mode 1801A, 1801B and the second out-of-phase bending mode 1802A, 1802B vary with the mass flow, which illustrates that Wide ambient band excitation may provide information for the actual normal modal decomposition of the duct movement to determine a process parameter such as mass flow according to the actual normal modal decomposition techniques such as those indicated in detail in the foregoing. The ability to use wideband excitation provided by ambient vibration sources such as FSI allows the construction of "passive" vibratory duct detectors, i.e., detectors that do not incorporate an actuator. Such a passive detector can be particularly advantageous in applications of limited power or in applications in which the driving power may present a safety problem, for example applications in explosive or flammable environments. conclusion As described here, the process parameter estimates are generated by using a real modal decomposition of a response from a detector to excitation conduit. The conduit response is separated into real normal modal components, for example movement in a plurality of real normal modes or components of the movement of behavior associated with such real normal modes. The actual normal modal components can then be used to generate estimates of process parameters such as mass flow, density, viscosity and the like.

Because the conduit response is separated into the actual normal components, contributions from sources such as pump vibration and fluid turbulence can be removed by filtering by ignoring, eliminating, attenuating or otherwise eliminating by filtration. Real normal modal components not associated with Coriolis forces. In this way, you can obtain parameter estimates that are potentially more accurate than those obtained by conventional techniques. In addition, because a detector according to the present invention can distinguish between modes, unimodal or almost unimodal excitation is not required. In fact, a detector conduit can be excited using only environmental excitation. Accordingly, a "passive" parameter detector can be constructed, i.e., a detector that includes only motion transducers and no actuator, a configuration which can be particularly advantageous in explosive environments or in applications in which the power consumption It is an important consideration. The actual normal modal decomposition according to the present invention can be used to improve the operation of conventional parameter detectors. For example, a conventional curved tube Coriolis flowmeter can be improved by the proportion of additional motion transducers that produce motion signals that can be processed by a mode pass filter to provide modally filtered signals for use in Coriolis measurements. conventional phase difference. A conventional flow meter can also be upgraded, for example by replacing conventional Coriolis electronic measurement circuits with a packet of DSP-based electronic detector circuits that implement a real normal modal motion estimator and a process parameter stevedor that directly calculates the mass flow or other process parameters from an estimated actual normal response. In addition to conventional straight and curved tube detectors, the modal decomposition used in accordance with the present invention may also be advantageous in detectors having non-conventional conduit geometries, for example asymmetric or unbalanced. Those skilled in the art will appreciate that the techniques of the present invention can be combined with other techniques to provide improved process parameter techniques. For example, the spatial filtering provided by the real normal modal decomposition technique described here can be combined with the frequency domain filtering to obtain a temporal filtering. Because the present invention is amenable to linear algebraic calculations, spatial integration can be provided in conjunction with the present invention by producing motion signals that provide information for various positions that exceed the number of actual normal modal components within which the notion of conduit. Therefore, an overdetermined source of information is provided for duct movement resolution in the actual normal modal components, as described in the above-mentioned U.S. patent application, "Improved Vibrating Conduit Parameter Sensors and Methods of Operation Therefor Utilizing Spatial Integration ". The drawings and the specification of the present application describe embodiments of the invention. Although specific terms are used, they are used in a generic and descriptive sense only and not for purposes of limitation. It is expected that persons skilled in the art can and use or sell the alternative modalities that are within the scope of the following claims either literally or under the doctrine of equivalents. It is noted that in relation to this date, the best method known to the applicant to carry out the aforementioned invention, is that which is clear from the present description of the invention.

Claims

CLAIMS Having described the invention as above, the content of the following claims is claimed as property:

1. An apparatus for estimating a process parameter of a material flowing through a conduit that is oscillated by an impeller, in which the process parameter is estimated from a plurality of motion signals that represent the movement of the conduit Since they are generated by a plurality of motion transceivers fixed to the conduit, the apparatus is characterized in that it comprises: a real normal modal separator which receives the plurality of movement signals from the plurality of motion transducers associated with the conduit and which processes the plurality of movement signals to separate the movement represented by the plurality of movement signals in a plurality of real normal modal components; and a process parameter estimator that responds to a generation of the plurality of real normal modal components and that estimates a process parameter from a real normal modal component of the plurality of real normal modal components.

2. The apparatus in accordance with the claim 1, characterized in that the actual normal modal splitter comprises a step filter operating in order to produce an output of the plurality of motion signals which preferentially represents a component of the movement associated with a real normal mode of the conduit, - and in which the process parameter estimator that responds to the generation of the output by the mode pass filter and that estimates a process parameter from the output.

3. The apparatus in accordance with the claim 2, characterized in that the mode pass filter is operative to preferentially pass a component of the duct movement associated with a real normal mode.

4. The apparatus according to claim 2, characterized in that the mode pass filter represents a product of a real normal modal transformation that establishes the movement map in a physical domain for movement in a plurality of system of a single degree of freedom and a normal reverse selective normal modal transformation that establishes the motion map in a selected set of plurality of systems of a single degree of freedom for movement in a physical domain.

5. The apparatus according to claim 2, characterized in that the mode pass filter is operative to produce an output that preferentially represents a component of the duct movement associated with a real normal mode that is preferentially correlated with a Coriolis force associated with the material in the conduit.

6. The apparatus according to claim 2, characterized in that the process parameter estimator comprises a mass flow estimator.

7. The apparatus according to claim 2, characterized by further comprising a means for processing received motion signals to produce a motion vector and wherein the mode pass filter comprises a means for multiplying the motion vector by a matrix Step filter mode to produce a filtered motion vector.

8 ^ The apparatus in accordance with the claim 2, characterized in that the mode pass filter is operative to produce a first filtered signal representing the movement in the first position of the conduit and a second filtered signal representing the movement in the second position of the conduit; and wherein the process parameter estimator comprises: means for determining a phase difference between the first filtered signal and the second filtered signal; and a means that responds to a determination of the phase difference, to estimate the mass flow from the determined phase difference.

9. The apparatus according to claim 1, characterized by: the actual normal modal splitter comprises an operative normal modal motion estimator operative to estimate the actual normal modal movement from the received plurality of motion signals; and in which the process parameter estimator is operative to estimate the process parameter from the estimated normal normal modal movement.

10. The apparatus according to claim 9, characterized in that the actual normal modal movement estimator is operative to apply a real normal modal transformation to the received plurality of motion signals to generate an estimate of the actual normal modal movement.

11. The apparatus according to claim 9, characterized in that: the actual normal modal movement estimator comprises: means for generating a motion vector; and a means for multiplying the motion vector by a real normal modal transformation matrix to produce a real normal modal motion vector; and wherein the process parameter estimator comprises a means for estimating the process parameter from the actual normal modal movement vector.

12. The apparatus according to claim 9, characterized in that: the actual normal modal movement estimator is operative to estimate the movement in a plurality of real normal modes, - and in which the process parameter estimator is operative to estimate a parameter of the process from the estimated movement for a subset of the plurality of real normal modes.

13. The apparatus according to claim 12, characterized in that: - 6d - the actual normal modal movement estimator comprises: a means for estimating movement in a first real normal mode; and a means for estimating movement in a second normal normal mode which correlates preferentially with a Coriolis force; and in which the process parameter estimator comprises: a means for normalizing the movement estimated in the second real normal mode with respect to the movement estimated in the first real normal mode to that produced from a normalized motion estimate in the second normal normal mode; and a means for estimating a process parameter from the normalized movement estimate in the second real normal mode.

14. The apparatus according to claim 13, characterized in that the means for normalization comprises a means for normalizing an imaginary component of the movement estimated in the second real normal mode with respect to a real component of the movement estimated in the first real normal mode.

15. The apparatus according to claim 14, characterized in that the means for estimating the movement in a first real normal mode comprises a means for estimating the movement in a real normal mode which correlates preferentially with an excitation applied to the conduit.

16. The apparatus according to claim 13, characterized in that the process parameter estimator comprises a means for estimating a process parameter from the normalized estimate of movement in the second normal real mode, a known mass flow and a predetermined normalized estimate. of movement in the second normal normal mode to a known mass flow.

17. The apparatus in accordance with the claim 16, characterized by the process parameter estimator comprising a means for estimating the mass flow.

18. The apparatus according to claim 13, characterized by: the means for estimating movement in a first real normal mode comprises a means for estimating movement in a symmetric mode; the means for estimating the movement in a second normal real mode comprises a means for estimating the movement in an antisymmetric mode; the means for normalization comprises a means for determining a relation of an imaginary component of the movement estimated in the antisymmetric mode with respect to a real component of the movement estimated in the symmetric mode; and in which the means for estimating a process parameter from the normalized movement estimate in the second normal real mode comprises a means for estimating a process parameter from the determined ratio.

19. The apparatus in accordance with the claim 18, characterized in that the means for determining the process parameter from the determined ratio comprises a means for estimating a process parameter from the determined ratio, a known mass flow and a predetermined ratio corresponding to the known mass flow .

20. The apparatus in accordance with the claim 19, characterized in that the means for estimating a process parameter from the determined ratio comprises a means for estimating the mass flow.

21. The apparatus according to claim 1, characterized in that the plurality of movement signals represents an overdetermined information source for separating the movement of the conduit in a plurality of real normal modal components.

22. A method for estimating a process parameter of a material flowing through a conduit that is vibrated by an impeller fixed to the conduit from a plurality of motion signals generated by a plurality of motion transducers fixed to the conduit, the The method is characterized in that it comprises the steps of: receiving the plurality of movement signals representing a movement in a plurality of vibratory duct positions from the plurality of motion transducers associated with the duct containing material; processing the received plurality of movement signals to separate the movement into a plurality of actual normal modal components; and estimating a process parameter from the actual normal modal component of the plurality of real normal modal components.

23. The method according to claim 22, characterized in that: the processing step comprises the step of processing the received plurality of motion signals to produce an output that preferentially represents a component of the movement associated with the actual normal mode of vibratory duct; and wherein the estimation stage comprises the step of estimating a process parameter from the output.

24. The method according to claim 23, characterized in that the processing step comprises the step of applying a mode pass filter to the plurality of movement signals.

25. The method according to claim 24, characterized by the mode pass filter representing a product of a real normal modal transformation which establishes the movement map in a physical domain with respect to the movement in a plurality of systems of a single degree of freedom and a normal reverse selective normal modal transformation which establishes a motion map in a selected set of the plurality of systems of a single degree of freedom with respect to movement in a physical domain.

26. The method according to claim 24, characterized in that the processing step comprises the step of producing an output which preferentially represents a component of the duct movement associated with a real normal mode which correlates preferentially with a Coriolis force.

27. The method according to claim 24, characterized in that the estimation step comprises the step of estimating the mass flow.

28. The method in accordance with the claim 24, characterized by the processing step comprising the steps of: processing the received plurality of motion signals to produce a motion vector; and multiplying the motion vector by a mode pass filter matrix to produce a filtered motion vector.

29. The method in accordance with the claim 24: characterized by the processing step comprises the step of producing a first filtered signal that represents the movement in a first position of the conduit and a second filtered signal represents the movement in a second position of the conduit, - and where the estimation stage it comprises the steps of: determining a phase difference between the first filtered signal and the second filtered signal; and estimate the mass flow from the first determined phase difference.

30. The method according to claim 23, characterized by: the processing step comprises the step of estimating the actual normal modal movement from the received plurality of motion signals; and wherein the step of estimating a process parameter comprises the step of estimating a process parameter associated with the estimated actual normal modal movement material.

31. The method according to claim 30, characterized in that the actual normal modal movement estimation step comprises the step of applying a real normal modal transformation to the received plurality of motion signals to generate an estimate of the actual normal modal movement.

32. The method according to claim 31, characterized by: the actual normal modal movement estimation step comprises the steps of: generating a motion vector from the plurality of motion signals; and multiplying the motion vector by a real normal modal transformation matrix to produce a real normal modal movement vector, - and in which the stage of estimating a process parameter comprises the step of estimating a process parameter from the Real normal modal movement vector.

33. The method according to claim 30, characterized in that: the step of estimating real normal modal movement comprises the step of estimating the actual normal modal movement for a plurality of real normal modes, - and where the stage of estimating a parameter The process comprises the estimation of a process parameter from the actual normal modal movement of a subset of the plurality of real normal modes.

34. The method according to claim 33, characterized in that: the step of estimating real normal modal movement comprises the steps of: estimating the movement in a first real normal mode; and estimating the movement in a second normal normal mode that correlates preferentially with a Coriolis force; and wherein the step of estimating a process parameter comprises the steps of: normalizing the movement estimated in the second real normal mode with respect to the movement estimated in the first real normal mode to produce a normalized estimate of movement in the second normal mode real, - and estimate a process parameter from the normalized movement estimate in the second real normal mode.

35. The method according to claim 34, characterized in that the normalization step comprises the step of normalizing an imaginary component of the movement estimated in a second real normal mode with respect to a real component of the movement estimated in the first real normal mode.

36. The method in accordance with the claim 34, characterized in that the step of receiving a plurality of movement signals comprises the step of receiving the plurality of movement signals while applying an excitation to the conduit; and wherein the step of estimating movement in a first real normal mode comprises the step of estimating the movement in a real normal mode that is preferentially correlated with the excitation applied to the conduit.

37. The method in accordance with the claim 34, characterized in that the step of estimating a process parameter comprises the step of estimating a process parameter from a normalized estimate of movement in the second normal real mode, a known mass flow and a predetermined normalized estimate of movement in the second normal mode real to the known mass flow.

38. The method according to claim 37, characterized in that the step of estimating a process parameter comprises the step of estimating the mass flow.

39. The method according to claim 34, characterized in that: - 7u - the step of estimating movement in a first real normal mode comprises the step of estimating the movement in a symmetric mode; wherein the step of estimating movement in a second normal real mode comprises the step of estimating movement in an asymmetric mode; wherein the normalization step comprises the step of determining an imaginary component relationship of the movement estimated in the asymmetric mode with respect to a real component of the movement estimated in the symmetric mode; and wherein the step of estimating a process parameter from a normalized movement estimate in the second normal real mode comprises the step of estimating a process parameter from the determined ratio.

40. The method according to claim 39, characterized in that the step of estimating a process parameter comprises the step of estimating a process parameter from the determined ratio, a known mass flow and a predetermined ratio corresponding to the mass flow known.

41. The method according to claim 40, characterized in that the step of estimating a process parameter from the determined ratio comprises the step of estimating a mass flow.

42. The method according to claim 22, characterized in that the number of positions exceeds the number of real normal modal components separated such that the plurality of movement signals represents an overdetermined source of information to separate the movement of the conduit in a plurality of Real normal modal components.

43. The method according to claim 22, characterized in that the reception stage is preceded by the step of driving a plurality of real normal modes in the conduit, and wherein the receiving step comprises the step of receiving a plurality of movement that represent movement in its response to excitement.

44. The method according to claim 43, characterized in that the excitation step comprises the step of applying a broadband excitation to the duct.

45. The method according to claim 43, characterized in that the excitation step comprises the step of applying a series of substantially coherent excitations of variable frequencies.

46. The method according to claim 43, characterized in that the excitation step comprises the step of transferring energy from a material in the conduit to thereby excite a plurality of real normal modes. SUMMARY OF THE INVENTION A plurality of movement signals are received which represent the movement in a plurality of positions of a vibratory duct containing material. The received plurality of movement signals is processed to separate the movement into a plurality of real normal modal components. A process parameter is estimated from the actual normal modal component of the plurality of real normal modal components. According to one aspect, the movement signals can be processed by applying a mode pass filter to produce an output that preferentially represents a movement component associated with a real normal mode of the vibratory passage. A process parameter can be estimated from the filtered output using, for example, conventional phase difference techniques. According to another aspect, the actual normal modal movement is estimated from the received plurality of motion signals, and a process parameter is estimated from the estimated actual normal modal movement. For example, the movement in the first and second respective normal modes can be estimated, the second normal mode actually correlates preferentially with a Coriolis force. A process parameter can be estimated by normalizing the movement estimated in the second real normal mode with respect to the movement estimated in the first real normal mode to produce a normalized estimate of movement in the second normal real mode and by estimation of a process parameter from the normalized estimate of movement in the second real normal mode.