MXPA00007469A - Generalized modal space drive control system for a vibrating tube process parameter sensor - Google Patents

Abstract

A drive system is taught for controlling the modal content of any number of drive signals used to excite any number of drives on a vibrating conduit such as is found in a Coriolis mass flowmeter or a vibrating tube densimeter. One or more motion signals are obtained from one or more spatially distinct feedback sensors. The motion signals are preferably filtered using a multi-channel modal filter to decompose the motion signals, each of which contain modal content at a plurality of vibration modes, into n single degree of freedom modal response signals. Each modal response signal corresponds to one of the vibration modes at which the vibrating conduit is excited. The n modal response signals are input to a drive channel having a separate processing channel for each of the n modal response signals. Within each drive controller channel, the respective modal response signal is compared to a desired modal response setpoint and the resulting mode error signal is amplified by a modal gain to produce a modal excitation signal for each mode. The modal excitation signal represents the modal excitation necessarily applied to the vibrating conduit to cause the modal response to match the modal setpoint for the given mode. The modal excitation signals are transformed from the modal domain back to the physical domain and mapped to the physical locations of the drives. The resulting drive signals are applied to the drives to excite the conduit.

Description

GENERALIZED MODAL SPACE CONTROL SYSTEM FOR A VIBRATORY TUBE PROCESS PARAMETER SENSOR FIELD OF THE INVENTION This invention pertains to the field of drive systems for causing a duct of a vibration tube process parameter sensor to oscillate. In particular, the present invention pertains to a system operating in the modal domain to drive several impellers with any modal force pattern.

ESTABLISHMENT OF THE PROBLEM It is known to use coriolis effect mass flowmeters to measure mass flow and other information of materials flowing through the conduit. Exemplary Coriolis flowmeters are described in U.S. Pat. Nos. 4,109,524 of August 29, 1978, 4,491,025 of January 1, 1985, and Ref. 31,450 of February 11, 1982, all for J. E. Smith et al. These flow meters have one or more conduits of straight or curved configuration. Each conduit configuration in a coriolis mass flow meter has a set of natural vibration modes, which can be single, torsional, radial or coupled type bends. Each REF: 121342 conduit is driven to oscillate to a resonance in one of these natural modes. The material flows into the flow meter from a connected duct on its inlet side of the flow meter, is directed through the duct or ducts, and exits the flowmeter through the outlet side. The natural vibration modes of the system filled with vibrating material are defined in part by the combined mass and stiffness characteristics of the conduits and the material flowing into the conduits. When there is no flow through the flow meter, all points along the pipe oscillate, because there is no driving force applied, with identical phase or zero flow phase, depending on the driven vibration mode. As the material begins to flow, the coriolis forces cause a change in the phase difference while either of two points along the conduit. The phase on the inlet side of the conduit delays the impeller, while the phase on the outlet side drives or advances to the impeller. The pick-up sensors are placed in the duct to produce sinusoidal signals representative of the movement of the duct. The signals emitted from the feedback sensors are presented to determine the change in the phase difference between the feedback sensors. The change to the phase difference between two pickup driving signals is proportional to the mass flow rate of the material through the duct.

A typical component and of all the coroliz flowmeters, and of all vibration tube densitometers, is the drive or excitation system. The drive system works to apply a periodic physical force to the conduit which causes the conduits to oscillate. The drive system includes at least one impeller mounted to the conduit or conduits of the flow meter. The drive typically contains one or many well-known arrangements, such as, but not limited to, a voice coil in which a magnet is mounted to the conduit if a coil of wire is mounted in the other conduit in an opposite relationship to the magnet. A driving circuit continuously applies a periodic, typically sinusoidal or square-shaped driving signal to the driving coil. Through the interaction of a continuous alternating magnetic field produced by the coil in response to the periodic driving signal and the constant magnetic field produced by the magnet, both flux conduits are initially forced to vibrate in an opposite sinusoidal pattern which is maintained from that moment. Those skilled in the art will recognize that any device capable of converting an electrical signal into a mechanical pit is suitable for application as an impeller (see U.S. Patent 4,777,833 issued to Carpenter and assigned in its presentation to Micro Motion., Inc.). Furthermore, one does not need to use a sinusoidal signal, but rather any periodic signal may be appropriate as the driving signal (see U.S. Patent 5,009,109 issued by Kalotay et al., And assigned in its inception to Micro Motion, Inc.). A typical mode, although it is not the only way in which coriolis flowmeters are driven to vibrate in the first bending mode out of phase. The first bending mode out of phase is the fundamental bending mode in which the two tubes of a double tube coriolis flow meter vibrate one in opposition to the other. However, this is not the only mode of vibration present in the vibratory structure of a coriolis flowmeter driven in the first bending mode out of phase. Of course, there are higher modes of vibration which are usually excited. In addition, there may also be a result of fluid flow through the vibratory flow tube and the consequent coriolis forces, a first out-of-phase torsion mode which is excited in the same way as the other modes. There are also modes of phase and lateral vibration. Finally, there are hundreds of vibration modes actually excited in a coriolis flowmeter that are driven to oscillate in the first bending mode out of phase. Even within a relatively narrow range of frequency near the first bending out of phase, there are at least several additional vibration modes. In addition to the multiple modes that are excited by the driving excitation of the flow tubes, the modes can be excited due to external vibrations in the flow meter. For example, a pump located elsewhere in a process line can generate a vibration along the pipe that excites a vibration mode in a coriolis flow meter. Another reason that additional and undesirable modes are sometimes excited in a coriolis flowmeter is when the manufacturing tolerances are such that the driving elements are not located symmetrically on the flow tubes. This results in the impeller establishing eccentric forces within the flow tubes and therefore excites multiple modes of vibration. Therefore, a coriolis flowmeter driven to oscillate or resonate in the first bending mode out of phase actually has a conduit or conduits that oscillate in many other modes in addition to the first bending mode out of phase. Meters driven to oscillate in a different mode to the first bending mode out of phase experience the same phenomenon of multiple excited modes in addition to the proposed drive mode. Existing drive systems process a feedback signal, typically one of the feedback sensor signals, to produce a driving signal. Unfortunately, the driving feedback signal contains responses in other ways in addition to the desired mode of excitation. Therefore, the boosted feedback signal is filtered through a frequency domain filter to remove unwanted components and the filtered signal is then amplified and filtered to the impeller. However, the frequency domain filter used to filter the driving feedback signal is not effective in isolating the desired single driving mode from other mode responses present in the driving feedback signal. There are responses outside resonance of other modes which are close to the resonance frequency of the desired mode. There may also be resonant responses at frequencies that approximate the desired resonance frequency. In any case, the filtered feedback signal, ie, the driving signal, typically contains a modal content at different frequencies than just the desired excitation mode of the flow tube. A driving signal constituted by a resonant response of inputs of multiple modes, through the impeller, energy to the flow tube which excites each mode for which the driving signal contains a modal content. Such a multiple-mode driving signal causes operational problems in coriolis flow meters. further, the frequency domain filters introduce a phase delay in the filtered driving signal. This may result in a requirement for greater driving energy in order to drive the flow tube to the desired amplitude. This is described in DE 19634663A granted to Fuji Electronic Co. Ltd.

A problem caused by a multi-mode driving signal is that external vibrations such as pipe vibrations are reinforced by the driving signal. If the pipe vibrations external to the coriolis flowmeter cause the flowmeter to vibrate, the driving feedback signal contains the response to pipe vibration. The frequency domain filter does not remove the unwanted response if the pipe vibration is at least partly within the frequency band of the filter. The filtered feedback signal, which includes the unwanted response to pipe vibration, is amplified and applied to the impeller. The impeller then operates to reinforce the excitation mode of the pipe vibration. An additional problem of a signal having a modal content at multiple frequencies occurs with respect to the density measurement performed by a coriolis mass flow meter. The density measurement in a coriolis flowmeter of a vibration tube densitometer is based on the measurement of the resonant frequency of the vibratory flow tube. A problem arises when the flow tube is driven in response to a driving signal that contains a modal content in multiple modes. The superposition of the multiple modes in the driving signal may result in a flow tube that is driven out of resonance from a real resonant frequency of a desired driving mode. Density measurement error may result. The problems indicated above describe various conditions under which a driving signal designed to excite a single mode is degraded so as to excite multiple modes. It is known that a modal filter can be used to produce a driving signal from at least two feedback signals wherein the modally filtered driving signal has a modal content only in the desired vibration mode. A modal filter is used to increase the desired drive mode and suppress one or more undesired modes. There are situations where one wishes to excite multiple modes and therefore a driving signal that has a modal content in multiple modes is required. The simultaneous excitation of two modes requires a driving signal that has a modal content in these two modes. There may be other reactions to drive a vibratory conduit so that multiple modes are excited and multiple modes are suppressed. For example, a superimposition in the time domain of a driving signal having a modal content in a first mode, and a second driving signal having a modal content in a second mode produces a double-mode driving signal. The impellers can be attached to a flow tube so that undesired modes are not excited and the advantages of multiple impellers are particularly described. There are no drive control systems for producing drive signals that have a desired modal content in multiple modes or for producing drive signals filtered to multiple drivers. There is a need for a drive control system that is easily adapted to a process of a plurality of drive feedback signals to produce a plurality of drive signals, each of which has a modal content that influences multiple modes.

ESTABLISHMENT OF THE SOLUTION The problems identified above, and others, are solved, and a technical advance is obtained in the field of the generalized modal space driving control system of the present invention. The present invention provides a generalized drive control system operating in the modal domain to produce a driving signal or signals for one or more drivers in a vibratory structure. The impeller system receives multiple impulse feedback signals, decomposes the modal content of the vibratory duct into modal responses of a single degree of freedom (SDOF), the process of modal SDOF response signals to select the desired quantities of each mode and transforms the results to the physical domain for application to the impellers. Therefore, using the generalized driving control system of the present invention, one or more driving signals are produced to excite one or more impellers and thus excite or suppress certain modes of the vibratory structure. In addition, the drive control system of the present invention operates on, and easily switches between, multiple operating configurations. For example, a set of driving signals is generated by the present invention during a flow measurement operation configuration and an alternative set of driving signals that are generated by the present invention during an axial tension measurement operation configuration. The drive control system of the present invention is used to control the vibration modes of a vibratory structure which includes at least one vibratory duct. The vibratory structure may also include additional vibratory ducts or one or more balance bars which are not wetted by the fluid to be measured by the flow meter. In addition, the vibratory structure may include ridges or covers, for example. The feedback sensors and the impellers are placed in any portion or portions of the vibratory structure according to the present invention, to carry out the excitation of the desired modes and the suppression of the undesired modes.

In one embodiment of the present invention, at least one feedback sensor provides a signal of the movement of the vibratory duct. The movement signal has a modal content in multiple modes, each of which is associated with a certain frequency. The multiple frequency bandpass filter produces multiple modal response signals. Each modal response signal is associated with one of the vibration modes present in the vibratory conduit. In another embodiment of the present invention, at least two feedback sensors provide movement signals indicative of the movement of the vibratory passage. The movement signals have a modal content in a plurality of vibration modes. The motion signals are input to a multi-channel modal filter having a channel for each mode to be altered, eg, excited or suppressed, by the drive control system. Each channel of the modal filter produces a modal response signal corresponding to one of the plurality of vibration modes in which the conduits vibrate. An impeller controller has a channel for each modal response signal that operates to produce a modal excitation signal corresponding to each mode of vibration. The modal excitation signal corresponds to a given vibration mode indicating the degree to which the given mode is present in the final drive signal or signals. A modal to physical force projector receives the modal excitation signals, transforms them into a physical domain and transmits one or more driving signals for application to one or more impellers. An additional embodiment of the present invention utilizes a frequency bandpass or other temporary filters in addition to the modal filters in order to generate modal response signals for further processing. In a further embodiment of the present invention, a combination of frequency bandpass filters and modal filters is used. The only output of each modal filter channel is passed through a frequency domain filtre. The modal filter of the present invention is constituted of a channel for each mode of vibration affected by the driving system. For each vibration a modal response signal is produced which is substantially an SDOF signal which is affected by the driving system of the present invention. The configuration of the modal filter channels, once established, does not need to be reconfigured even when the drive system switches from a first operation configuration to a second operation configuration. Likewise, the physical modal force projector is configured in accordance with the various methods described herein and does not need to be reconfigured even when the drive system switches from a first operating configuration to a second operating configuration.

Each channel of the driver controller receives an input from one of the modal response signals SODF. The modal response signal is compared to an established modal response point to produce an error mode signal. In the error mode signal is amplified by a gain stage to produce a modal driving signal for each mode altered by the driving system of the present invention. The relative amplitude of the modal excitation signals corresponding to each affected vibration mode indicates the care with which a given mode contributes to the driving signal or signals produced by the present driving system. For example, if a given vibration mode is to be suppressed, an "unwanted" mode, then the set modal response point for the corresponding modal response signal is zero. The resulting modal excitation signal is transformed to the physical domain, that is, an applied force, by operation of the projector from modal to physical force. The force is applied to the vibrating tube to suppress the unwanted mode. The drive system of the present invention provides multiple operation configurations. Thus, a coriolis flow meter, for example, uses the drive system of the present invention can alternate between various operating configurations, each of which requires a different driving methodology. Consider the following example. In a first operating configuration, a certain mode of vibration may be undesirable and is therefore suppressed by the drive system of the present invention. However, in a second operating configuration, the same vibration mode may be desirable and therefore is increased by the drive system of the present invention. The only changes to the drive system of the present invention from one operation configuration to another operation configuration are the modal response reference point for each channel of the driving controller. A search table contains the appropriate reference points for the various operating configurations of a coriolized flow meter or vibratory densimeter. The drive system of the present invention transforms multiple motion signals from the time domain to the modal domain. In the modal domain, the motion signals are decomposed into a SODF modal response signal for each vibration mode that is affected by the drive system. Each modal response signal is processed to produce a corresponding modal excitation signal. The modal excitation signal represents the magnitude of the corresponding vibration mode which is applied as a component of the driving signal or signals to carry out the desired modal response reference point. The modal excitation signals are transformed from the modal domain to the temporal domain to produce driving signals that result in a force applied to the impellers.

BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows a coriolis flowmeter and the associated electronic meter equipment; Figure 2 shows a block diagram of a generalized modal space driving control system according to the present invention; Figure 3 shows a blog diagram of a driving circuit according to the present invention; Figure 4 shows a frequency response function of a representative flow tube feedback signal and the additional frequency response functions which represent the contribution of the constitutive vibration modes of the feedback signal; Figure 5 shows a blog diagram of a multiple channel modal filter according to the present invention; Figure 6 shows a processor of modal response signals according to the present invention; Figure 7 is a flow diagram illustrating the process steps for transforming a scalar modal excitation signal for a physical force projection vector; Figure 8 is a flow chart illustrating the process steps for selecting modal filter weighting coefficients by trial and error; Figure 9 is a flow chart illustrating the process for selecting modal filter weighting coefficients when calculating the inverse or pseudoinverse of the eigenvector matrix; Figure 10 shows an alternative driver controller of the present invention, - Figure 11 is a flow chart summarizing the operation of the present invention. Figure 12 shows a modal response signal generator according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION Coriolis flow meter in creneral - FIGURE 1 Figure 1 shows a coriolis 5 flowmeter comprising a Coriolis meter assembly 10 and an electronic meter equipment 20. The electronic meter equipment 20 is connected to the meter assembly 10 by means of conductors 100 to provide density, mass flow rate, volume flow rate, total mass flow and other information on the path 26. A flow meter structure is described. coriolis, although it is evident to those skilled in the art that the present invention can be practiced in conjunction with a vibratory tube densitometer without the additional measurement capability provided by a coriolis mass flow meter. In addition, the present invention can be applied to a coriolis or densitometer flowmeter using more than one impeller or more than two acquisition sensors. The meter assembly 10 includes a pair of flanges 101 and 101 ', a manifold 102 and flow tubes 103A, 103B. Connected to the flow tubes 103A and 103B is the impeller 104 and the pickup sensors 105 and 105 '. The connecting bars 106 and 106 'serve to define the axes W and W around which each flow tube oscillates. When the flowmeter 10 is inserted into a pipe system (not shown) which conveys the process material to be measured, the material enters the meter assembly 10 through the flange 101, passes through the manifold 102 when the material is directed to enter the flow tubes 103A, 103B, flows through the flow tubes 103A and 103B and back into the manifold 102 from where the meter 10 exits through the rebord.e 101 '. The flow tubes 103A and 103B are appropriately selected and mounted to the manifold 102 so that they have substantially the same distribution of mass, moments of inertia and elastic modules around the folding axes W-W and W'-W 'respectively. The flow tubes extend outwardly from the manifold in an essentially parallel manner. The flow tubes 103A and 103B are driven by the impeller 104 in opposite directions around their respective folding axes W and W 'and what is called the first bending mode out of phase of the flow meter. The impeller 104 may comprise any of many well-known arrangements, such as a magnet mounted to the flow tube 103A and an opposite coil mounted to the flow tube 103B and through which an alternating current is passed to vibrate both flow tubes. An appropriate drive signal is applied by the electronic meter equipment 20, via the conductor 110, to the impeller 104. As indicated above and described herein in greater detail, the present invention is suitable for use with any number of impellers. The description of Figure 1 is provided solely as an example of the operation of a coriolis flow meter and is not intended to limit the teachings of the present invention. The electronic meter equipment 20 receives the left and right speed signals that appear on conductors 111 and 111 ', respectively. The electronic meter equipment 20 produces a driving signal that appears in the conductor 110 and causes the driver 104 to vibrate the tubes 103A and 103B. The present invention, as described herein, can produce multiple driving signals for multiple impellers. The electronic meter equipment 20 processes the left and right speed signals to calculate the mass flow rate and the density of the material passing through the meter assembly 10. This information is applied by the electronic equipment 20 meter on the path 26 to a means of use (not shown). It is known to those skilled in the art that the coriolis 5 flowmeter is very similar in structure to a vibratory tube densitometer. The densitometers of vibrating tube also use a vibratory tube through which the fluid flows or, in the case of a sample type densitometer, within which the fluid is retained. Vibrating tube densitometers also use a drive system to excite the flow tube to vibrate. Vibration tube densitometers typically use only a feedback signal since a density measurement requires only the measurement of the frequency and a phase measurement is not necessary. The descriptions of the present invention herein apply equally to vibratory tube densitometers. Those skilled in the art will recognize that there is a coriolis flowmeter in advance that has two feed signals available to introduce a modal filter, and an existing vibratory tube densitometer has only one typically available feedback signal. Therefore, one only needs to provide additional feedback signals in a tube densitometer in order to apply the present invention to a vibratory tube densitometer.

Electronic equipment in general - FIGURE 2 Figure 2 shows a blog diagram of an electronic device 20 meter for a general case where there are L feedback sensors 202A-202L and M impellers 206A-206M. L and M are any integer greater than one. The electronic electrode 20 meter includes a mass flow rate circuit 30 and a driving circuit 40. The mass flow rate circuit 30 is one of many known circuits for calculating the mass flow rate of a fluid through a vibratory tube based on the phase difference between two points in the vibratory tube. Typically, two feedback sensors 202A-202L are also used by the mass flow rate circuit 30 to produce a mass flow rate information, although the feedback sensors used by the driving circuit 40 can be separated from the sensors of feedback used by the mass flow rate circuit 30. The mass flow circuit 30 produces the output of a means of use (not shown) on the line 26. The means of use can be, for example, a screen. The details of the mass flow rate circuit 30 are well known to those skilled in the art and do not form part of the present invention. See U.S. Pat. Re 31,450 issued to Smith on November 29, 1983 and assigned in his presentation to Micro Motion, Inc. or the U.S. patent. 4,879,911 issued to Zolock on November 14, 1989 and assigned on its cover to Micro Motion Inc., or the U.S. patent. 5,231,884 issued to Zolock on August 3, 1993, and assigned on its cover to Micro Motion, Inc., for exemplary information regarding the mass flow rate circuit 30. The driver circuit 40 receives motion signals on path 204A-204L from the feedback sensors 202A-202L, respectively. The driving circuit 40 generates M driving signals on the paths 208A-208M to the drives 206A-206M, respectively. The operation of the present invention utilizes at least two movement signals from at least two feedback sensors that detect the movement of two different points in a vibratory duct. The operation of the present invention requires at least one impeller. The remainder of the discussion herein focuses on the operation of the driving circuit 40.

Modal space driver driver in creneral - FIGURE 3-4 Figure 3 shows the driving circuit 40 including a multi-channel modal filter 310 and a processor 320 of modal response signals. The vibratory conduit or conduits 302 schematically represent the conduits 103A-103B or any other single or multiple conduit geometry. As indicated above, the vibratory conduits 302 can be part of a vibratory structure that includes other vibratory elements. Although this embodiment is described with respect to the vibratory conduits 302 as the overall vibratory structure, it is understood, as indicated above, that the vibratory conduits may comprise only a portion of a vibratory structure. Force 304 from impellers 206A-206M causes vibratory lines 302 to vibrate. As the vibratory lines 302 vibrate, movement 306 is detected by the feedback sensors 202A-202L. The feedback sensors 202A-202L are located at different points of the vibratory structure, ie, at different points on the vibratory conduits 302 and produce motion signals on the paths 204A-204L which are input to the driving circuit 40. The motion signals are input to a multi-channel modal filter 310. The multi-channel modal filter 310 decomposes the modal content of the motion signals to produce N modal response signals on the 312A-312N paths. Figure 4 illustrates the operation of a multi-channel modal filter 310. Graphs 401-404 provide vertical axes that represent the logarithmic relationship of the response amplitude of the conduit over the input force amplitude. The graph 401 illustrates the frequency response function 405 (FRF) for a vibrating conduit such as the vibrating conductors 302. An FRF characterizes the dynamics between the force applied to a structure in one position and the resulting movement of the structure in another position. Thus, the FRF 405 is, for example, the FRF at the position of one of the feedback sensors 202 on the vibratory lines 302. The FRF 405 has a modal content in modes A, B and C within a certain frequency range. Each of the movement signals generated by the feedback sensors 202A-202L has a modal content similar to that of FRF 405. The multi-channel modal filter 310 operates to decompose the modal content of the motion signals to produce modal response signals of a single degree of freedom (SDOF) substantially over trajectories 312A-312N and which has the FRF as illustrated in graphs 402-404. The graph 402 illustrates the FRF 406 which corresponds to a modal response signal on the path 312A for the A mode of the first out-of-phase folding component of the FRF 405. The graph 403 illustrates the FRF 407 which corresponds to a modal response signal on the path 312B, for the B mode of the first component of the out-of-phase torsion mode of the FRF 405. The graph 404 illustrates the FRF 408 which corresponds to a modal response signal on the 312N path for C mode or second phase folding component of the FRF 405. Therefore, the multiple channel modal filter 310 is movement signals L on the paths 204A-204L. The output for the multi-channel modal filter 310 is N SDOF modal response signals on the 312A-312N paths, wherein each modal response signal represents the modal content of the vibration lines 302 in a single mode. The multi-channel modal filter 310 may be configured to produce modal response signals for any mode present in the frequency range of interest. The modal response signals are input to the processor 320 of modal response signals. As described in more detail with respect to Figure 6, a processor 320 of modal response signals processes the modal response signals to produce M drive signals on paths 208A-208M. In the simplest case there is only one impeller and therefore only one driving signal.

Multiple channel modal filter - FIGURE 5 Figure 5 is a more detailed description of a multi channel modal filter 310. The multiple channel modal filter 310 is constituted of modal filter channels 500A-500N. The channels 500A-500N, of modal filter are identical but for the gains of their respective amplifiers 504A-504N, 505A-505N and 506A-506N. As described herein, the gains of the amplifiers in the modal filter channels 500A-500N are established such that each modal filter channel 500A-500N transmits a modal response signal corresponding to a vibration mode present on the Tubes 302 vibrators. Only one of the modal filter channels 500A-500N is described in detail, since there is a complete common condition between the configuration of the multiple channels. With reference to the modal filter channel 500A, the motion signals from the feedback sensors 202A-202L are input to the amplifiers 504A-506A, respectively. The amplifier 506A represents a variety of additional amplifiers for receiving motion signals from a variety of additional feedback sensors 202A-202L. The amplifier 504A has a gain of RG1, the amplifier 505A has a gain of AG2, and the amplifier 506A has a gain of AGL. The AGX-AGL gains are referred to as the weighting factors applied by the modal filter channel 500A to the motion signals on the paths 204A-204L. The outputs of the amplifiers 504A-506A on the paths 507A-509A are referred to as the weighted feedback signals. The weighted feedback signals are summed by the additive 510A to produce a modal SDOF response signal on the path 511A. The modal response speed signal on path 511A is input to integrator 512A. The integrator 512A integrates the modal response speed signal on the path 511A to produce a modal response displacement signal on the path 312A.

Selection of the modal filter weighting factors -FIGURES 8-9 The selection of the weighting factors for each channel of the multi-channel modal filter 310 is discussed in more detail below with respect to Figures 8-9. There are several methods, one of which can be used to select the weighting factors for the modal filters applied to the driving circuit of a coriolis mass flow meter. The means by which the weighting factors are determined is not critical and any method or combination of methods is appropriate and equivalent. A method to select the weighting factors for the modal filters of a coriolis drive circuit is by trial and error. As indicated with respect to Figures 3-5, the desired result of the multi-channel modal filter 310 is to produce modal SDOF response signals corresponding to the vibration modes present in the vibratory conduits 302. Figure 8 is a flow diagram illustrating the steps that are used to select the modal filter weighting coefficients for a single channel of a multi-channel modal filter 310 using the trial-and-error approach. Steps 801-804 are repeated until the appropriate SDOF modal response signal is obtained as an output of, for example, channel 500A. Steps 801-804 are carried out using a real coriolis flowmeter, suitably instrumented to provide the necessary feedback signals together with a driving circuit that allows the change of the gains of the modal filter amplifiers. Alternatively, the feedback signals can be recorded, for example in a digital audio tape format, and can be reapplied to a modal filter drive circuit with each pass through steps 801-804. Alternatively, stages 801-804 are executed using a numerical model of a Coriolis flowmeter and an associated driving circuit. The process begins with step 800 and continues with step 801 where a first set of weighting coefficients is selected. During step 801, one may select a whole new set of weighting coefficients (GA gains -AGx, in the case of channel 500A) each time step 801 is executed or one may select a new weighting coefficient for only one signal of feedback each time step 801 is executed. During step 802, the feedback signals are applied to channel 500A, where each modal filter amplifier has the gain set determined by step 801. During step 803, measures and records the output signal of the filter. The process proceeds from step 803 to decision block 804. Decision block 804 operates to determine the filter channel output signal if it is an SDOF modal response signal for the appropriate vibration mode. If it is determined, by operation of decision block 804, that the signal output through the filter channel is substantially a modal SDOF response signal for the desired vibration mode, then processing continues to decision block 805, where determines if there are more channels in the multiple channel filter for which weighting coefficients are required. If it is determined that the signal output through the filter channel is not substantially a SODF modal response signal for the desired vibration mode, then the process returns to step 801. A new set of weighting coefficients are selected during the step 801, and steps 802-804 are processed again to locate a set of weighting coefficients that produce a SDOF modal response signal. This process is repeated for each of the channels 500A-500N, by operation of the decision blog 805, until each of the channels 500A-500N produces a SDOF modal response signal corresponding to a vibration mode. Another method for selecting the weighting coefficients for each channel of a multi-channel modal filter 310 is to calculate the inverse or pseudo-reverse of an eigenvector matrix. As indicated above, a vibration flow tube of a coriolis flow meter has a combination of vibration modes present. When analyzing the flow tube movement in the physical coordinates, for example, the singular response in points and divisions of the flow tube, requires the analysis of coupled equations which do not easily provide useful information about the movement of the flow tube. However, one can use a modal transformation to transform a vector of physical responses to modal responses or modal coordinates of the system. The standard modal transformation is given by: (1) = F? where: x is the vector of the physical response coordinates f is the matrix of the eigenvector, the columns of which are the eigenvectors of the flux tube (also referred to as modal vectors) of interest, and? is the vector of the modal response coordinates.

The eigenvector matrix can be developed, as described below, by any coriolis flowmeter flow tube. The physical vectors can be considered as the input, that is, the feedback signals, to the modal filter. Therefore, equation (1) is solved for?, And the response or responses of modal coordinates, as follows: (2) ? = ffx By placing equation (1) in the form of equation (2), it requires taking the pseudoinverse of the matrix f of the eigenvector. If the matrix of the eigenvector is square and not singular, then the inverse of the matrix of the eigenvector (f-1) is used in equation (2), instead of the pseudoinverse. The matrix of the eigenvector is square and not singular when the number of feedback signals from the flow tube is equal to the number of modes considered and the modal vectors are linearly independent. The following example is used to illustrate the process by which one calculates the pseudoinverse of a modal matrix to determine the weighting coefficients-for a channel of a modal filter of multi-channel modal filter. One can use a physical or numerical model of the flowmeter to construct the own vector matrix. In the following example a numerical model of the flowmeter is used. A finite element model is constructed from the tubes of a coriolis mass flow meter model CMF100 (manufactured by Micro Motion, Inc. of Boulder, Colorado). The model is fixed to ground at the ends of the flow tubes so that, in a physical flowmeter, it is connected to a multiple of flowmeter. Finite element modeling techniques are well known to those skilled in the art and do not form part of the present invention. The exemplary finite element model is constructed using SDRC-Ideas and analyzed by MSC / NASTRAN, a finite element code available from MacNeal-Schwendler. Those skilled in the art of finite element modeling recognize that alternately any finite element code can be used. The positions of the feedback sensors are modeled to produce an output representative of the relative movement between the positions in the flow tube of a magnet and the corresponding coil of the right feedback, the impeller and the left feedback. "Scalar points" are a standard technique in advanced dynamic analysis. See "A Finite Element for the Vibration Analysis of a Fluid-Conveying Timeshenko Bean." (AIAA document 93-1552), for more information regarding the finite element modeling of coriolis flowmeters. The eigenvalue coefficients of the CMF 100 model are extracted from the finite element model to construct the following 3-row matrix by 10 columns of the eigenvector for the CMF 100 sensor: 0 25 08 0 0 0 -40 .3 0 0 0 36. 78 '3) Fcomplßto 0 35. 39 0 0 0 0 0 0 0 - 36. 55 0 25. 08 0 0 0 40. 3 0 0 0 36. 78 Each complete row of the full eigenvector of equation (3) corresponds to a physical position in the flow tube. The first row corresponds to the left pick-up position, the second row corresponds to a driving position and the third row corresponds to the right pick-up position. Each column in the complete vector matrix fcomplete corresponds to a vibration mode. This matrix is used from one known for the finite element model to model the signals generated by the pickup sensors. The matrix is used as described below, to develop the weighting coefficients for each channel of the impulse circuit modal filter. The "modes" columns with zeros in the complete eigenvector matrix are in "in-phase modes". This means that there is no relative movement between the tubes because both tubes move with the same speed and direction. Thus, the sensors used to provide feedback signals, speed sensors in this example, in themselves act as a kind of modal filter to filter out all in-phase modes. The foomp? Eto matrix of the complete eigenvector is reduced by removing all the columns in phase. . 1 -40.3 36.8 (4) freducido 35.4 0 -36.6 25.1 40.3 36.8 Equation (4) is the reduced fructific vector of eigenvector. Equation (1), the standard modal transformation, is rewritten using the freduclted matrix of eigenvector a reduced, as follows: (5) where? b is the first modal coordinate response of out-of-phase bending mode,? t is the first modal coordinate response of out-of-phase torque mode and? 2b is the second modal coordinate response of bending mode off phase and FSA is the physical response of sensor A feedback, FSB is the physical response of sensor B feedback and FSL is the physical response of sensor L feedback. If the pickup response and the reduced eigenvector matrix are known, the modal coordinate responses can be determined as in (6) by premultiplying equation (5) by the inverse or inverse pseud? Of the reduced eigenvector matrix.

The reduced eigenvector matrix is inverted by importing the matrix into a standard commercial mathematical calculation package such as Mathcad and using one of the standard investment or pseudoinvestment functions available in these calculation packages. The resulting equation is shown as equation (7): (7) The numerical coefficients in equation (7) are the weighting factors for the modal filter amplifiers in a coriolis flowmeter driven circuit. For example, if one wishes to extract the first bending mode out of phase of the feedback signals, as in the case for channel 500A, then the first row of the previous modal filter vector matrix is used, as follows: (8)? B = 8.2389 (FSA) + 16.5795 (FSB) + 8.2389 (FSL) The first modal vector coefficients of out-of-phase folding mode are multiplied by 103 to simplify equation (8). With reference to channel 500A of Figure 5, the GX gain of amplifier 504 is set to 8.2389 '(the modal filter vector coefficient corresponding to the sensor A feedback, AG2 gain is set at 16.5795 (the modal filter vector coefficient corresponding to the feedback sensor B) and the AGL gain is set to 8.2389 (the modal filter vector coefficient corresponding to the feedback sensor L). Likewise, the coefficients of the second and third rows of equation 7 are used as the weighting coefficients for channels 500B-N, respectively. Therefore, each of the channels 500A-N produces an SDOF modal response signal corresponding to a vibration mode present in the vibratory conduits 302. The weighting factors are scaled linearly as a group to provide a modal response signal on the path 312A having the appropriate amplitude to input it into the processor 320 of modal response signals. Figure 9 is a flow diagram illustrating the process steps for determining the modal filter coefficients of the driving circuit when calculating the inverse or pseudo-inverted of the eigenvector matrix. The calculation of the inverse or pseudo-inverse of the eigenvector matrix described above and with respect to Figure 9 is known to those skilled in the advanced dynamic analysis art and is a useful tool for determining the modal coefficients of the driving circuit. The flow chart of Figure 9 begins with element 900 and proceeds to step 901. During step 901, an eigenvector array is constructed. As indicated above, a method for determining the eigenvectors for the eigenvector matrix is to construct a finite element model of the flowmeter from which the eigenvectors are extracted. Another solution is to use an experimental modal analysis to determine the eigenvectors directly from a physical sample of the flowmeter. Experimental modal analysis is well known to those skilled in the art and their methods and use are not part of the present invention. Once the eigenvectors are obtained by any appropriate method, the own vector matrix is compiled. Equation (3) is an example of a complete eigenvector matrix for ten vibration modes at three points in the flow tubes. Each column of the eigenvector matrix represents a different mode while the number of rows of the eigenvector matrix represents the degrees of freedom. The proper vector matrix is then reduced to the modes for it to be filtered. For the current example, this is done by eliminating the columns with zeros as coefficients. For the exemplary structure and the sensors described herein, columns (modes) with coefficients such as zero are in in-phase modes. Processing proceeds from step 901 to step 902. During step 902, the inverse or pseudo-reverse of the eigenvector matrix is computed. Each row of the inverse or pseudo-inverted matrix of the eigenvector contains modal filter coefficients associated with a particular mode. This is expressed in general by equation (2) and is shown by the previous example by equation (7). The processing then proceeds to step 903. During step 903, the appropriate modal filter weighting coefficients are selected, as discussed above, to produce a different SDOF modal response signal from each channel of the multi-channel modal filter. .

Generator of modal response signals - FIGURE 12 Figure 12 shows a generator 1200 of modal response signals. The generator of modal response signals 1200 is an alternative to the multi-channel modal filter 310 to produce modal response signals on the 312A-312N trajectories. The generator of modal response signals 1200 uses frequency bandpass filters 1202A-1202N to decompose a single motion signal, from the feedback sensor A in this example, into modal response signals where each modal response signal corresponds to a vibration mode present in vibrating conduits 302. With reference to Figures 12 and 4, the bandpass filter 1202A is configured to pass the frequency A, shown in Figure 4. Similarly, the bandpass filter 1202B is configured to pass the frequency B and the bandpass filter 1202N is configured to pass frequency C. Integrators 1206A-1206N integrate the bandpass filter signals on paths 1204A-1204N to produce displacement signals from the speed signal output by the sensor A feedback. Those skilled in the art of time domain filters recognize that any of the different different filtering techniques can be used in bandpass filters 1202A-1202N and include, but are not limited to, digital signal processing techniques. In a further embodiment of the present invention, the bandpass filters of Figure 12 are used in conjunction with the multiple channel modal filter of Figure 5. For example, bandpass filter 1202A is applied to the output of the amplifier 504A, the bandpass filter 1202B is applied to the output of the amplifier 505A and the bandpass filter 1202N is applied to the output of the amplifier 506A.

Processor of signals of modal answers - In general -FIGURA 6 The multi-channel modal filter 310, or alternatively the generator 1200 of modal response signals, produces N modal response signals, as just described, on the 312A-312N paths. The circuit paths 312A-312N are input to the processor 320 of modal response signals. The processor 320 of modal response signals processes the N modal response signals to produce N driving signals on the paths 208A-M. Figure 6 shows a blog diagram of the processor 320 of modal response signals. The modal response signal processor includes a driver 602 driver, a projector 604 of modal force to physics and an add-on stage 606. In general, the driving controller 602 receives N modal response signals on the 312A-312N trajectories and for each modal response signal, determines the deviation of the modal response signal from a desired reference point for the modal response signal . This deviation or "error mode signal" is amplified by the 610A-610N amplifiers with a modal gain (GA-GN) to produce a modal excitation signal that is applied through the 206A-206M impellers to vibrate the conduits 302, modifies the modal content of the vibratory conduits 302, so that the resulting modal response signals approximate their corresponding desired modal reference points. Modal excitation signals are communicated from the driving controller 602 to a 604 modal-to-physical force projector on the 618A-618N trajectories.

The modal excitation signals on the trajectories 618A-618N represent, in the modal domain, excitations which are applied to the vibratory conduits 302. However, of course, the vibratory conduits 302 are in the physical domain and therefore the modal excitation signals must be converted to regulated physical excitations or forces. This transformation is carried out by a 604 projector of modal to physical force. The 604 modal-to-physical force projector is constituted of a separate 620A-620N channel for each modal response signal. There is an output from channel 620A-620N for each impeller 206A-206M. The outputs of force 604, modal to physical projector 604 are fed to step 606 of the adder. The outputs of the 604 projector from modal to physical force are summed, as described in more detail in the following; to produce M driving signals on paths 208A-208M.

Impeller controller - FIGURE 6 The driver controller 602 is constituted of a channel 601A-601N for each of the N modal response signals. The function and operation of each channel 601A-601N is similar and therefore only one of channels 601A-601N will be described in detail. The operation of the remaining channels is clear from the following description of channel 601A. The description of the driving controller 602 is organized in a first discussion of the modal reference points and a second discussion of the modal gains.

Modal reference point The channel 601A of the driver 602 receives a modal response signal on the path 312A for a multi-channel modal filter 310. For example, suppose that the modal response signal processed by the channel 601A corresponds to the first bending mode out of phase for the vibratory lines 302. The modal response signal is introduced to subtraction stage 608A as the subtrahend entry. The minuend introduced to the subtraction step 608A is the modal reference point on the path 612A. The modal reference point is the desired level for the corresponding modal response signal. The error mode output signal on path 614A indicates the extent to which the modal response signal deviates from the modal reference point. The modal reference point on the path 612A can be, but is not limited to, a fixed voltage that is provided by a voltage reference (not shown) or a value retrieved from a memory (not shown).

The modal reference point itself is determined as follows. As indicated in the above with respect to equation 1, the physical response (x) of a system is related to the modal response (?) Of a system by a set of modal vectors (f). The physical response at a given position j due to a given mode i is represented by equation 9: (9) x = Fjiií'llii where: x ± j is the scalar physical response at position j due to mode i; -JÍ is the element of the proper vector matrix for the mode duct and the position row j; and is the scalar modal response of mode i. Equation 9 can be rewritten in the form of equation 10 and solved for a modal reference point? 1 (s) for node i.

Suppose that one wishes the reference point (x1.) Of the desired physical displacement for the first out-of-phase folding mode which is 0.015 inches in the feedback sensor A. Referring again to equation 4, the matrix element (fj: ¡.) Corresponding to the feedback sensor A (first row) and the first folding mode out of phase (first column) is 25.1. Therefore,? 1 (the modal reference point for point i (the first folding mode out of phase in this example) is equal to 0.0006.

Modal gain In subtraction step 608A subtracts the modal response signal on path 312A from the modal reference point on path 612A to produce an error mode signal on path 614A to gain stage 610A. As indicated above, each channel of the driving controller 602 converts a modal response signal to a modal excitation signal which, when applied to the vibratory conduits 302, causes the vibratory conduits to oscillate such that a modal response signal Approach the modal reference point for the given mode. The error mode signal on the path 614A is amplified by the modal gain stage 610A. In a linear time invariant system, the modal response is related to the modal excitation by the system parameters, that is, the mass stiffness and the damping. In the modal space, these parameters are the modal mass, the modal stiffness and the modal damping. For a system with normalized mass eigenvectors, the modal mass is the unit, and the modal stiffness is the square of the natural frequency. In a system excited in resonance, as in the case with a coriolis flowmeter, the nominal gain to be obtained from the modal response to modal excitation is expressed by equation 12: (??) -? N < : ID where; ^ is damping; ? n is mode frequency, - J7 is the modal response for the i mode; and tsr ± is the modal excitation for mode i.

Using the current example, where the modal reference point, ie, the desired modal response is given by equation 11, the gain, G, applied by gain step 610A is set as shown in equation 12 as follows : 2? V = 2- (0.0005) • (106.2-2p) 2 = 445.254 sec "(12) I The gain stage 610A is therefore configured to have a gain of 445.254. The drive excitation signal on the path 618A therefore is the modal force that needs to be applied to the vibrating conduits 302 in order that the modal response signal for the corresponding mode coincides with the corresponding modal reference point.The remaining stage, as described below, is for transforming the modal excitation into physical excitations that are applied in the form of driving signals on the paths 208A-208M The process described above with respect to the channel 601A of the driving controller 602 is repeated for each of the remaining channels 601B-601N for each of the remaining modal response signals Note that the present invention provides a direct means for influencing one or more modes in a coriolis flowmeter, for example, if one wishes to suppress a certain mode, then the modal reference point for that mode is zero. Therefore, any signal in general in the corresponding modal response signal produces an error mode signal. If one wishes to suppress multiple modes, then the modal reference point for each of the appropriate modes is set to zero. Similarly, one can choose to excite multiple modes by selecting the appropriate modal reference points for each mode that is to be excited.

Projector of modal force to physics - FIGURE 7 Each channel 601A-601N of the driver 602 transmits a scalar modal excitation signal which must be expanded in a vector signal to drive each of the impellers m. A projection vector of modal force to physical is determined, each element of which is the gain applied to the outputs 618A-618N from the driving controller 602 to scale the scalar modal excitation signal for each impeller, at an appropriate amplitude. The Nr scalar mo- tor signal is the modal force needed to drive the r mode to the desired response amplitude. The current modal force applied to the sensor in the r mode is Nractual, as shown in equation 13. where: Fr is the vector of physical force necessary to excite the desired mode; Nr is the scalar mode drive signal; and f is the transposition of a single column of the modal eigenvalue matrix.

In order to ensure that Nractual - Nr, as desired, the physical force vector, Fr, must satisfy the following relation, F? FJ. = 1. There are many different physical force vectors which satisfy this relationship. In fact, any arbitrary vector, A,., Which is not totally orthogonal to f ^, that is, frtFr = 0, can be scaled to a desired physical force vector, Fr, as follows: Note that the physical force vector (Fr) is a single column per m row vector, where each row corresponds to a specific driving position. Also note that frt needs to be reduced to the physical positions of impeller. This makes the frt matrix different from the matrix used for the modal filter and the driver driver configuration which is reduced to the feedback detector positions. Equation 14 needs to be solved for the physical force vector (Fr), that is, equation 14 needs to be inverted. This is done by finding an A vector, such as: A / Fr = 1 (15) Multiplying both sides of Equation 15 with the vector Ar, then the transformation is obtained from the modal excitation to the physical force, as shown in equation 16: A ^ (16) Equation 16 is a point product of two vectors which means that one is working with vectors and not with matrices. By applying certain properties of the vectors, it is determined that the vector A ,, is expressed as shown in equation 17: where (| the symbol is the vector length operator.

A calculation of the exemplary force projection is shown below for the case where the coriolis flowmeter has two impellers (m = 2). Referring again to Equation 6, one of the two impellers is located coincident with the feedback sensor A, and the other of the two impellers is located coincident with the feedback sensor L. The first out-of-phase folding mode, A,., Is then calculated as shown in the equation 18: A "= 0. 0398 IJbf. sec2 (18) 0.0398 inches It has been shown with respect to equations 9-10 that the modal reference point for the first out-of-phase folding mode for this example is 0.0006. It has been shown with respect to equations 11-12 that the modal gain for the first out-of-phase folding mode for this example is 4.45.254 sec "2. Therefore, the modal reference point multiplied by the gain modal for the first off-phase folding mode in this example, ie, the scalar modal excitation signal on path 618A, is equal to 0.2672. These quantities are substituted into equation 15 to determine the physical force vector Fr for the first folding mode out of phase, as shown in equation 19: A * N í lbf (19) This result tells us that in order to push the sensors to the first bending mode out of the desired phase, a force of 48 grams (0.0106 pounds) must be applied to the two impellers. The force projection vectors are developed in a similar manner for each mode of interest altered by a drive system of the present invention. In order to optimize the drive system of the present invention for a particular drive scheme, consideration must be given to selecting the appropriate force projection vectors. There are many issues which may be considerations when choosing the shape of the vector. First, the peak force which any impeller must generate in order to obtain the vibration amplitude of the desired flow tube must be minimized. An additional consideration is to minimize the total energy which dissipates the drive system to obtain the desired flow tube vibration amplitude. Another consideration is to minimize the residual response or other modes. Figure 11 is a flowchart discussed below as a summary of the present invention. The step 1114 of FIG. 11, the force projection step, is observed in greater detail in FIG. 7. FIG. 7 begins with step 700 from step 1112 of FIG. 11. As indicated in the above , there is a channel 620A-620N separate force projector for each modal response signal. During step 702, the force projector array is determined by reducing the total eigenvector array to a single column in relation to the appropriate mode. The processing then continues in step 704.

During step 704, the inversion matrix is determined (as described above, by using standard vector properties.) The processing now proceeds to step 706, during step 706, equation 15 is solved for the projector vector of force (Fr). Steps 704 and 706 have the effect of scaling the force projector array to the correct amplitude, as shown in equation 14. Processing then proceeds to step 708. During step 708, Corresponding forces of each channel of the force projector are added to provide m impeller signals The process then proceeds to step 710 and returns to step 1112 of figure 11. A modal force-to-physics propector 604 is implemented using analog components discrete, as described with respect to the rest of the drive circuit 40. Alternatively, a projector 604 of modal force to physical and the rest of the driver circuit 40 is implemented using analog to digital converters (ADC) to convert analog motion signals into digital signals. The digital signals are then processed using digital signal processing (DSP) techniques and the resulting driving signals are converted back to analog signals using digital to analog converters (DACs). Those skilled in the art of electronic signal processing will recognize that many different analog or digital solutions (ADC, DSP, DAC) can be used to implement the teachings of the present invention and that all of them are equivalent and within the scope of the equivalents for the present invention.

Addition stage - FIGURE 6 The signal output from the 604 mode to physical force projector is summed in the add-on stage 604 to produce driving signals 208A-208M, as follows. As indicated above, each channel 620A-620N of a 604 modal-to-physical force projector produces as many outputs (M) as there are impellers 206. The totality of the outputs of the 604 modal-to-physical force projector correspond to an A-impeller. , for example, are added by the adder 622A to produce a driving signal 208A. Therefore, the output 620AnA, the output 6206 ^ and the output 620NDA are input to the adder 622A. The output of the adder 622A is the impeller signal 208A. The impeller signals 208B-208N are formed in a similar manner by the operation of the adders 622B-622M. The number of impellers M may be different from the number of modal response signals N which may be different from the number of movement signals N.

Multiple operation configurations - FIGURE 10 The unique design of the drive system of the present invention allows for significant and total changes of the drive schemes by only selecting a different set of modal reference points. The generalized modal space impeller control system of the present invention operates in different operating configurations by establishing different modal reference points for each operation configuration. For example, a first mode of operation for a flowmeter 5 is a mass flow configuration. In the mass flow configuration, for example of the flow meter 5 of Figure 1, the first bending mode out of phase is excited and the other modes are suppressed. A second configuration of a flowmeter 5 is a pressure measurement configuration. In the pressure measurement configuration, the first bending mode out of phase and the first bending mode out of phase are both excited, and the other modes are suppressed. The flowmeter 5 switches between the first and second operating configurations and this impeller is optimized for each mode. Figure 10 is similar to the description of Figure 6 of the processor 320 of modal response signals but with the addition of the memory 1002 and which operates with the configuration selector 1004. The memory 1002 is a read-only memory (ROM) for example. Loaded in memory 1002 is a modal reference point table (not shown). The modal reference point table includes a set of N modal reference points for each operating configuration of the flow meter. Each set of N modal reference points includes a modal reference point corresponding to N modal response signals. The operation of the configuration selector 1004 provides a control signal on the path 1006 to the memory 1002, which determines the set of modal reference points transmitted by the memory 1002 over the paths 612A-N. By operating configuration selector 1004 and memory 1002, it is preferably part of electronic meter circuit 20. The operating operation selector 1004 may, for example, be a microprocessor (not shown) in electronic meter equipment 20. The switch from a first operating configuration to one or more alternative operating configurations can be carried out in various ways and is not important for the present invention. An example is that the operation selector switch 1004 can be configured to switch every 30 seconds between the first operation configuration and the second operation configuration. Another example is that the operation configuration selector 1004 can be configured so that the first operation configuration is a default operation configuration, and that the second operation configuration is used only when, for example, a user presses a button (not shown) which requests a pressure measurement of the flow meter 5. In addition, a first operation configuration may always be running and the second operation configuration is periodically used in addition to the first operation configuration.

Summary - FIGURE 11 Figure 11 is a summary of a flow chart of the drive system of the present invention. Steps 1102-1106 relate to the transformation of the physical movement of a vibrating conduit to modal SDOF response signals. Steps 1108-1112 operate in the modal domain to generate modal excitation signals from the modal response signals. It is in the modal domain where the signals are manipulated to alter the modes of vibration. Steps 1114-1118 operate to transform the modal excitation signals to the physical domain, map the excitation signals to the impellers and apply the impeller signals to the impellers. The process begins with step 1102, when a coriolis or densimeter flowmeter begins operation. In step 1104, the movement signals L are received from the feedback sensors L indicating the movement of various points along the vibrating conduit. During step 1106, N modal response signals are generated from L motion signals. Each of the modal response signals, typically an SDOF signal, corresponds to a vibration mode present in the vibrator conduit. The N modal response signals are input to an N-channel driver controller during step 1108. During step 1110, an error mode signal is generated for each of the modal response signals by applying each modal response signal. to its corresponding modal reference point. The error mode signal is amplified by a mode gain, during step 1112, to produce a modal excitation signal for each mode. During step 1114, each modal excitation signal is transformed to a physical force vector where each element of the physical force vector corresponds to one of the impellers m. Step 1114 is described in greater detail in FIG. 7. The processing then continues with step 1116 wherein the m drive signals are applied to the impellers which causes the vibrating duct to vibrate in the desired modes. Processing concludes with step 1118. Although specific embodiments have been described herein, it is expected that a person skilled in the art can and design alternative driving control systems 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 by the applicant to carry out the aforementioned invention, is the conventional one for the manufacture of the objects or products to which it refers.

Claims (22)

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

1. A driving circuit that generates M driving signals and applies the M driving signals to a driving means fixed to a conduit to cause the driving means to vibrate to the conduit, wherein the driving signals are generated from L movement signals received by the driving circuit from a motion sensor means fixed to the conduit generating L movement signals in response to the oscillations of the conduit, where L and M are integers greater than 1, the driving circuit is characterized in that it includes: a modal response signal generating means arranged to receive L movement signals and to produce N modal response signals from the L motion signals, in where each of the N modal response signals corresponds to a different one of a plurality of vibration modes of The conduit and wherein N is greater than 1; The modal response signal generating means comprises N modal filter channels, wherein each of the N modal filter channels are arranged to receive at least two motion signals L as input, analyze the received motion signals to determine the amount of vibration of the conduit in a particular mode, and produce one of the N modal response signals as an exit; a means responsive to the production of N modal response signals arranged to produce N modality excitation signals from N modal response signals by subtracting each of N modal signals from a modal reference point for a particular mode of a signal of modal responses and applying a mode gain to a result of the subtraction, wherein each of the N modality excitation signals represents the modal excitation necessary to obtain a desired modal response level for the respective one of the plurality of modes of vibration; and a means responsive to a production of N modality excitation signals arranged to produce M drive signals from N modal excitation signals, wherein M drive signals cause the driving means to vibrate the duct in the desired modes.

2. The driving circuit according to claim 1, characterized in that the driving means includes M impellers fixed to the conduit, and each of the M impellers responds to one of the M drive signals generated by the driving circuit.

3. The driving circuit according to claim 2, characterized by the means for producing M driving signals includes: M driving signal channels which respond to the production of N modal driving signals to produce M driving signals.

4. The driving circuit according to claim 1. characterized in that the modal signal response modes include: N frequency bandpass filters, wherein each of the N frequency bandpass filters receives a different L signals of movement generated by the motion sensor means as an input and transmitting a respective N of modal response signals.

5. The driving circuit according to claim 4, characterized in that the modal response signal modality further includes: N integration means for integrating each of the modal response signals.

6. The driving circuit according to claim 1, characterized in that the motion sensor means includes: L movement sensors to produce L movement signals indicative of the movement of the conduit at a point of attachment of L motion sensors to the conduit wherein each of the L motion signals has a modal content in each of the plurality of modes of movement. vibration.

7. The driving circuit according to claim 6, characterized in that each N modal filter channels include: a first weighting means for applying a first weight factor to the first of at least two L motion signals to develop a first weighted signal , - a second weighting means for applying a second weighting factor to the second of one of at least two of L motion signals to develop a second weighted signal; and a modal filter addition means for combining the first weighted signal and the second weighted signal to produce one of the N modal response signals.

8. The driving circuit according to claim 6, characterized by each N modal response signals is substantially a single degree of freedom of modal response signal corresponding to one of the plurality of vibration modes.

9. The driving circuit according to claim 7, characterized in that the first and second weighting factors are determined through trial and error.

10. The driving circuit according to claim 7, characterized by the first and second weighting factors are determined through experimental analysis.

11. The driving circuit according to claim 7, characterized by the first and second weighting factors are determined through numerical analysis.

12. The driving circuit according to claim 1, characterized by the means to produce N modal excitation signals includes: N impeller driver channels, each having as input one of the N modal response signals and each produces one of the N Modal excitation signals as an exit.

13. The driving circuit according to claim 12, characterized by each of the N driver driver channels includes: a modal response reference mode for defining the modal response level for a given one of a plurality of vibration modes; a comparison means for comparing the modal response level with a corresponding one of N modal response signals to produce an error mode signal; and a gain means that responds to a production of an error mode signal to produce one of N modal excitation signals.

14. The driving circuit according to claim 13, characterized in that the modal response reference point for at least one of the N driver driver channels defines a zero modal response level for one of the N modal response signals that corresponds to at least one of the N driver driver channels whereby a corresponding modal drive signal having a zero level is produced from at least one of the N driver driver channels.

15. The driving circuit according to claim 13, characterized in that the modal response reference point for at least one of the N driver controller channels defines a non-zero modal response level for one of the N signaling signals. response modalities corresponding to at least one of the N driver impeller controllers, whereby a corresponding modal excitation signal having a non-zero level is produced from at least one of the N impeller driver channels.

16. The driving circuit according to claim 1, characterized in that it further includes: a means of "selecting to select one of a plurality of operating configurations for the driving circuit; an adjustment means responsive to a selection of one of a plurality of operation settings to adjust the modal response reference points.

17. The driving circuit according to claim 16, characterized in that the selection means includes: a memory containing a set of modal response reference points for each of the plurality of operating configurations; and an operation configuration selector to choose one of the sets of modal response reference points in the memory.

18. The driving circuit according to claim 17, characterized in that the adjustment means includes: a means that responds to the operation configuration selector to replace a first set of modal response reference points in the N driver controllers with a second set of modal response reference points.

19. The driving circuit according to claim 1, characterized by the means of the modal force projector to physical includes: N means, each has an input of one of the N modal excitation signals and each responds to a respective N of modal excitation signals to produce a drive component signal; the drive component signal represents a force in the impeller necessary to influence the vibration mode corresponding to a desired modal response level; and an adding means for adding the N impeller component signals transmitted from the N means of transformation from modal to physical signal to produce a driving signal.

20. A method for generating drive signals and applying the drive signals to a drive system fixed to a conduit in order to cause the drive system to vibrate the conduit, the method is characterized by including the steps of: receiving L signals of movement from of L motion sensors where each of the L motion signals is generated by one of the motion sensors and is indicative of the movement of the duct at a point of attachment of a motion sensor to the duct, the L motion signals have a modal content to a plurality of vibration modes and wherein L is an integer greater than 1; decompose the modal content of the L motion signals to produce N modal response signals, wherein each of the N modal response signals represents the amount of vibration of the conduit in a different one of the plurality of vibration modes and each of the N modal response signals are generated from at least two L motion signals; N generate modal excitation signals which respond to N modal response signals in which each of the N modality excitation signals represents the modal excitation necessary to obtain the desired modal response level for the respective vibration mode and are produced by subtracting the modal response signal from the modal establishment point and applying a modal gain to the result of the subtraction; transform the N modal excitation signals into M driving signals; and apply the M impulse signals to M impellers which causes the duct to vibrate where M is at least 1.

21. The method according to claim 21, characterized by the decomposition step includes: filtering the L motion signals through N modal filter channels wherein each of the N modal filter channels receives the entire L motion signals as an input, and transmits one of the N modal response signals.

22. The method according to claim 21, characterized in that the generation step includes: receiving N modal response signals on N respective driver driver channels and within each of the respective N driver driver channels; and amplifying the respective error mode signal by a mode gain to generate a respective modal excitation signal. SUMMARY OF THE INVENTION An impeller system for controlling the modal content of various driving signals used to drive several impellers in a vibratory duct such as is found in a coriolis mass flowmeter or a vibratory tube densimeter is disclosed. One or more movement signals are obtained from one or more spatially different feedback sensors. The motion signals are preferably filtered using a multiple channel modal filter to decompose the motion signals, each of which contains a modal content in a plurality of vibration modes, in n degrees of vibration, in n unique freedoms of Modal response signals. Each of the modal response signals corresponds to one of the vibration modes in which the vibrator conduit is excited. The n modal response signals are input to a drive channel which has a separate processing channel for each of the n signals. response manners. Within each impulse controller channel, the respective modal response signal is compared to a desired modal response reference point and the resultant mode error signal is amplified by a modal gain to produce a modal excitation signal for each mode. The modal excitation signal represents the modal excitation necessarily applied to the vibratory conduit to cause the modal response to coincide with the modal reference point for the given mode. Modal excitation signals are transformed from a modal domain back to a physical domain and mapped to physical positions of the impellers. The resulting impulse signals are applied to the impellers to excite the conduit.