WO2010091700A1

WO2010091700A1 - Method for the operation of a coriolis mass flowmeter, and coriolis mass flowmeter

Info

Publication number: WO2010091700A1
Application number: PCT/EP2009/001030
Authority: WO
Inventors: Thomas Bierweiler; Wolfgang Ens; Henning Lenz
Original assignee: Siemens Aktiengesellschaft
Priority date: 2009-02-13
Filing date: 2009-02-13
Publication date: 2010-08-19

Abstract

The invention relates to a method for operating a Coriolis mass flowmeter (1) as well as a Coriolis mass flowmeter according to the preamble of the independent claims. A first indicator is determined by determining a differential signal between the vibration signals (10) of the vibration pick-offs (9) of the Coriolis mass flowmeter and the frequency spectrum of the differential signal, said first indicator corresponding to a ratio between a weighted sum of harmonic frequencies and a fundamental frequency of the frequency spectrum. Furthermore, it is determined whether or not the first indicator exceeds a predefined threshold value. If the first indicator exceeds the predefined threshold value, at least one additional indicator is determined. The process fault identified by means of the first indicator is then classified on the basis of the at least two determined indicators.

Description

description

Method for operating a Coriolis mass flowmeter and Coriolis mass flowmeter

The invention relates to a method for operating a Coriolis mass flow meter and a Coriolis mass flow meter according to the preamble of the independent claims.

The measurement of the flow through process pipes of a plant is a necessary measure, for example, in process automation, which is why there are a variety of known measuring methods. Known measuring methods include Coriolis, magnetic-inductive, ultrasound, eddy current meters,

Pressure diaphragm, etc. With these measuring methods, a flow velocity and partly also the flow volume can be measured. Only a Coriolis mass flow meter delivers a direct mass flow measurement. However, when measuring the mass flow rate of liquids, the Coriolis mass flowmeter generally provides erroneous readings when gas components (gas void fraction) are carried.

Therefore, there are several proposals for detecting mitgeförderter gas shares in liquids. A Coriolis mass flow meter can determine not only the mass flow but also the density of the medium flowing through it. Therefore, a gas content in a liquid flowing through, for example, by comparing the specific density with a

Determine limit density. The problem with this approach, however, is that the density of the medium depends on different factors, e.g. Depending on the temperature, therefore, this method does not provide reliable results.

In the article "New parameters for analyzing the disturbances of a Coriolis mass meter", Technisches Messen, 74, 2007, Pp. 577-588, J. Goeke and M. Stock, various measures are presented for detecting gas fractions in a liquid whose flow is to be measured. In the amplitude characteristic, the discrete spectrum of the sensor signal of one of the two sensors is formed and a useful signal and a noise component are identified. The amplitude coefficient is then defined as the difference of the accumulated useful frequencies and the accumulated noise frequencies. However, the ratio of useful signal to noise component is subject to considerable temporal fluctuations. In addition, this method requires a high computing and storage costs, which must be considered in the design of an evaluation.

WO 2006/130415 A and WO 2006/127527 A describe methods for detecting and counting gas bubbles by means of additional excitation of other frequencies and comparison of the response with a reference spectrum. This makes it possible to determine both the mass flow rate and the density of the liquids and the gas. However, a problem with this approach is that it requires the use of a suitable excitation source, which is why the hardware of a conventional Coriolis mass flowmeter would be different. Furthermore, this approach is only applicable to certain gas bubble distributions in the liquid.

Furthermore, there are proposals for compensating the effects of entrained gas fractions on a measurement accuracy of the Coriolis mass flowmeter. For example, US2005284237 describes a method for correcting a mass flow measurement error. A neural network that has temperature, attenuation and density drop as input parameters corrects the mass flow error. However, only a partial compensation is achieved. Furthermore, this method requires some computing and memory overhead, which must be taken into account in the design of an evaluation. Thus, although a conventional Coriolis mass flowmeter is capable of detecting gas contents requested in liquids and at least partially compensating for an effect of the gas components on the measurement accuracy. However, a conventional Coriolis mass flowmeter is unable to detect cavitation and determine the location of occurrence. Cavitation is the formation of cavities in liquids by pressure fluctuations. As a result, a process tube, to which the Coriolis mass flowmeter is flanged, can be set in vibration. Cavitation directly in front of the Coriolis mass flowmeter leads to erroneous measured values, since the liquid in the measuring tubes of the measuring instrument then also carries cavities with it. Furthermore, pumps, valves, transducers, etc. can be damaged or even destroyed by cavitation. In addition, with low process pressure on the outflow side of the measuring section in the Coriolis mass flow measuring device, cavitation can occur, which can lead to damage of the measuring device. For this reason, in manuals to Coriolis mass flow meters in the

Headings Installation instructions or system pressure usually indicated that no cavitation may occur, as this may influence the measuring tube vibration and thus a measuring accuracy (see eg Endress + Hauser, Proline Promass 80/83 F, M, Technical Information, TI053E / 06 / en / ll .05, pages 15 and 20).

Therefore, it is an object of the present invention to enable a Coriolis mass flow meter to easily and reliably detect cavitation and determine the location of occurrence.

The objects underlying the invention are each achieved with the features of the independent claims. Preferred embodiments of the invention are specified in the dependent claims. The Coriolis mass flowmeter according to the preamble of the independent claims can detect in a liquid mitgeförderte gas components (two-phase flow) and thus a process disturbance by a first indicator is determined as described above. For this purpose, the frequency spectrum is calculated from the difference signal of the vibration pickup, for example by means of an FFT of 64 support points at a sampling rate of 20 kHz for a Coriolis mass flowmeter with a fundamental frequency of 650 Hz. The first indicator then corresponds to the quotient of the weighted sum of the Upper frequencies and the fundamental frequency. It is advantageous that such a phase information is clearly taken into account in the spectral analysis. This contains important information because the flow and the media density between the two vibration sensors change due to the multiphase flow, in particular due to the gas components entrained in the fluid.

According to the invention, at least one further indicator is now determined when a predetermined threshold value for the first indicator is exceeded. Preferably, the threshold for the first indicator is between 1.5 and 2.5, preferably about 2, wherein the determination of the threshold for the first indicator also depends on the size of the Coriolis mass flowmeter.

On the basis of these at least two specific indicators can then classify the detected process disorder. Preferably, all the specific indicators, including the first indicator, are used to discriminate the cause of the process disturbance so as to diagnose the exact process state. In particular, the classification allows the detection of cavitation. Furthermore, the location of the source of cavitation can be determined. In other words, the type and location of the process disturbance can be determined by the classification. This allows condition monitoring of the process. Based on this, it can also be determined whether erroneous measurements are available, so that additional Let a condition monitoring of the field device is possible. Basically, the more indicators that are included in the classification, the more reliable it is to determine cavitation and the location of the source of cavitation.

The classification according to the present invention preferably distinguishes between the following classes:

- no process fault - entrained gas without cavitation

- Cavitation far ahead of the device without gas

- Cavitation far ahead of the device with entrained gas

- Cavitation directly in front of the device without entrained gas

- Cavitation directly in front of the device with entrained gas - Cavitation in the device without entrained gas

- Cavitation in the device with entrained gas

- Cavitation directly after the device without entrained gas

- Cavitation directly after the device with entrained gas

According to one embodiment of the invention, the at least one further indicator is determined from the vibration signals, that is, the raw signals of the at least two vibration sensors. In this regard, as further indicators, for example, a phase angle, an amplitude value or amplitude estimated value, correlation coefficients of the two vibration sensors (eg 128 data points at a sampling rate of 20 kHz) and a difference of the vibration signal amplitudes of the at least two vibration sensors come into question. However, the at least one further indicator can also be obtained, for example, from an actuator current or an excitation frequency for the excitation arrangement or an amplification factor of the regulation. Furthermore, values derived from the here mentioned and other suitable indicators can be formed, such as their standard deviation, which are then included as further indicators in the classification. According to another embodiment of the invention, the at least one further indicator is calculated periodically and a predetermined number of values are stored for each further indicator. For example, the further indicators are calculated every 0.02 seconds and the last ten values of each indicator are stored in a memory of the evaluation device. Then the standard deviation for an indicator can be determined from its last ten values.

According to a further embodiment of the invention, the classification of the process disturbance takes place on the basis of a plurality of threshold values for each indicator. That is, each indicator that is determined is compared to multiple thresholds for that indicator. Experiments have shown that it is possible to make a good distinction between the respective process disturbances on the basis of several threshold values, in particular if either no gas is carried in the liquid or a gas volume fraction (gas void fraction) of more than 1.5% is present. In particular, if cavitation is present, cavitation can also be localized by the classification (before, in, or just after the Coriolis mass flow meter). The respective indicator threshold values can be derived from measured data or empirical values. It is also the case that the more threshold values are set for each indicator, the better it is possible to differentiate between process disturbances, so that a process disturbance can be reliably assigned to one of the above-listed classes.

According to an alternative embodiment of the invention, the classification of the process disturbance is based on minimum and maximum envelopes for each indicator. That is, a minimum and a maximum envelope are provided for each indicator instead of discrete thresholds. Although this variant requires a higher computational and memory overhead than the variant with the multiple thresholds per indicator, but so can the rigid Make the distinction between the discrete thresholds somewhat more flexible and better adapted to the actual circumstances. According to this alternative embodiment, each indicator lies within a certain range of the minimum and maximum envelopes provided for this purpose. Depending on this area, which is also referred to below as the residence area, it is possible to draw conclusions about the process status.

The envelopes do not necessarily have to be monotone or linear. Rather, the envelopes may be, for example, piecewise linear envelopes, nth order polynomials, spins, neural networks, exponential or logarithmic envelopes. In particular, the envelopes can be modeled in a manner known per se from a set of standard deviations for the respective indicator. For this purpose, standard deviations are preferably determined once for each indicator. From a subset of minimum values then the minimum curve and from a subset of maximum values the maximum curve by inverting a respective one

Envelope calculated. Furthermore, from the reversal of the envelope for each indicator, one or (in the case of non-monotonic envelopes) can specify several common areas. Based on the residence areas of the indicators, the process status can then be determined and thus a classification made. In order to determine the process disturbance even at low gas contents in the liquid, the residence areas of the indicators are preferably superimposed (superposition).

A particular advantage of the present invention is that the hardware of the Coriolis mass flowmeter does not have to be extended to determine further indicators. On the one hand, the other indicators are derived from the raw signals from the vibration sensors or from other signals from the measuring device, so that only existing (sensor) signals are used and evaluated. On the other hand, the determination of whether a threshold for the first Indicator is exceeded, the determination of other indicators and a classification of the process disturbance as software in an existing processor of the evaluation implement. Furthermore, since the process state can be identified and classified on the basis of the further features (eg entrained gas components, occurrence of cavitation, etc.), no additional sensors or actuators are required and the control circuit of the measuring device does not need to be changed.

Another advantage of the present invention is that a diagnostic statement about the process state and the field device is provided. Because from a certain gas content, the field device is out of specification. That is, the flow reading is no longer reliable and should no longer be considered valid process information. In addition, as described above, cavitation can set the measuring tubes and / or the process tube in vibration, which likewise leads to erroneous measured values. The Coriolis mass flowmeter according to the present invention can provide information on the occurrence of undesirable cavitation including the location of occurrence, thus providing the starting point for process optimization, such as, e.g. by increasing a backpressure.

Hereinafter, preferred embodiments of the invention will be explained in more detail with reference to the drawings.

Figure 1 shows a Coriolis mass flow meter according to an embodiment of the present invention;

FIG. 2 shows a pipeline of a plant to which the Coriolis mass flowmeter according to FIG. 1 is flanged;

Figure 3 shows an embodiment of a classifier according to the present invention; and FIG. 4 shows a flow chart for a method according to an embodiment of the present invention.

Figure 1 shows a Coriolis mass flowmeter according to an embodiment of the present invention. The Coriolis mass flowmeter 1 according to FIG. 1 operates on the Coriolis principle. A first measuring tube 2 and a second measuring tube 3 are arranged substantially parallel to one another. They are usually made from one piece by bending. The course of the measuring tubes is essentially U-shaped. A flowable medium flows according to an arrow 4 in the mass flow meter 1 and thus in the two located behind a not visible in the figure inlet splitter inlet sections of the measuring tubes 2 and 3 and corresponding to an arrow 5 from the outlet sections and the behind, also in the figure invisible discharge splitter off again. Flanges 6, which are fixedly connected to the inlet splitter and the outlet splitter, serve for fastening the mass flowmeter 1 in a pipeline, not shown in the figure. By a stiffening frame 7, the geometry of the measuring tubes 2 and 3 is kept substantially constant, so that even changes in the piping system in which the mass flowmeter 1 is installed, for example, due to temperature fluctuations, possibly lead to a low zero offset. An excitation arrangement 8, shown schematically in FIG. 1, which can consist, for example, of a magnet coil attached to the measuring tube 2 and a magnet immersed in the measuring tube 3, which is immersed in the magnetic coil, serves to generate mutually opposite oscillations of the two measuring tubes 2 and 3 whose frequency the natural frequency of the substantially U-shaped center portion of the measuring tubes 2 and 3 corresponds. Vibration sensors 9 are used to detect the Coriolis forces and / or based on the Coriolis forces oscillations of the measuring tubes 2 and 3, which arise due to the mass of the medium flowing through. The vibration signals 10, which are generated by the two vibration sensors 9, are emitted by an Value device 11 evaluated. For evaluation, the evaluation device 11 comprises a digital signal processor 12, in which the evaluation of signals according to the invention is implemented in the measuring device. Results of the evaluation, which will be described in detail below, are output on a display 13 and / or transmitted via an output, not shown in the figure, eg fieldbus, to a higher-level control station. Furthermore, a warning signal with information about a state of the process and / or the field device can be output via the display 13 or the output. In addition to the evaluation of the vibration signals 10 and possibly further signals in the measuring device, the evaluation device 11 also controls the exciter arrangement 8.

Notwithstanding the illustrated embodiment, the measuring tubes 2 and 3 may of course have other geometries, such as a V-shaped or a Ω-shaped center section, or it may be a different number and arrangement of exciter arrangements and vibration pickups are selected. The Coriolis mass flowmeter may alternatively have a different number of measuring tubes, for example a measuring tube or more than two measuring tubes.

According to the present invention, an evaluation performed in the evaluation device 11 includes determining a first indicator and determining whether the first indicator exceeds a predetermined threshold. If this is the case, the evaluation device 11 determines at least one further indicator in order to classify the process disturbance determined on the basis of the first indicator, taking into account the at least two specific indicators. That is, in the evaluation device 11, a classifier is implemented, which will be described with reference to Figure 3.

FIG. 2 shows a pipeline (process line) 14 of a plant to which the Coriolis mass flowmeter 1 is connected. measure of Figure 1 is flanged. Furthermore, FIG. 2 shows classes of process disturbances which can be determined on the basis of the Coriolis mass flowmeter according to the present invention. The classification primarily focuses on whether cavitation is present. In this regard, a distinction is made according to whether cavitation occurs far in front of the device, directly in front of the device, in the device or directly after the device. In addition, the classification may include whether gas bubbles are carried in the liquid or not.

FIG. 3 shows an exemplary embodiment of a classifier according to the present invention. This classifier classifies a process disorder based on multiple thresholds for each indicator used in the classification. In FIG. 3, an indicator with characteristic is designated, SD stands for standard deviation and stands for pickup for vibration pickup. The classifier of FIG. 3 can also be called a decision matrix, which is used for an allocation of process disturbances.

As shown in FIG. 3, the decision matrix comprises only one correlation coefficient of the two oscillators for each class of process disturbances, an interval with an upper and a lower limit. Exemplary values for the interval limits are: Sl = 0.5; S2 = 0.9; S3 = 0.97; S4 = 0.995; S5 = 0.9995; S6 = 0.99995; S7 = 1. With respect to all other features, the decision matrix of Figure 3 shows discrete thresholds, which for the sake of simplicity are not represented as absolute values but as "X". A "-" in the decision matrix means that all thresholds are undercut. To refine the classification, several thresholds are provided for each indicator. X = Threshold 1 exceeded, XX = Threshold 2 exceeded, XXX = Threshold 3 exceeded and XXXX = Threshold 4 exceeded. However, X does not mean that only Threshold 1 but not Threshold 2 is exceeded, but that Threshold 1 is definitely is exceeded. The same applies to the other thresholds XX, XXX and XXXX.

In order to be able to classify as precisely as possible, all the indicators provided in the decision matrix are determined each time, with the indicators provided in the decision matrix of FIG. 3 not being exhaustive, but also other indicators being included in the classification. In principle, the more indicators that are included in the classification of a process disturbance, the more precisely it is possible to differentiate between classes of process disturbances. Because not all indicators always help to make a precise statement about the nature of the process disorder. For example, it can be seen from FIG. 3 that the amplitude estimate does not allow a reliable statement about a process fault class. On the basis of the amplitude estimate, it can only be determined whether gas bubbles are conveyed in the liquid. However, it can not be determined if and where cavitation exists. By contrast, the standard deviation of the amplitude of the vibration signal of the first vibration sensor (SD (amplitude pickup I)) in the flow direction makes it very easy to see whether cavitation is present directly in front of the device and no gas is carried in the liquid. Because in this case the threshold 4 (XXXX) is crossed the only time.

FIG. 4 shows a flow chart for a method according to an embodiment of the present invention. In a first step S1, a first indicator is determined by determining a difference signal of the vibration signals of the vibration sensors (see FIG. 1) and its frequency spectrum. The first indicator corresponds to a quotient of a weighted sum of upper frequencies and a fundamental frequency of the frequency spectrum. In particular, the first indicator provides information about whether a process malfunction exists or not. In a second step S2, it is determined whether the first indicator has a predetermined threshold value. value exceeds or not. If the threshold is not exceeded, the procedure stops here. However, if the first indicator exceeds the predetermined threshold, then in a third step S3 at least one further indicator is determined. Based on the at least two specific indicators, the process fault detected by means of the first indicator is then classified in a fourth step S4.

Claims

claims

1. A method for operating a Coriolis mass flowmeter (1) with at least one measuring tube (2, 3) through which a medium flows, at least one excitation arrangement (8), which in the central region of the at least one measuring tube (2, 3 is arranged and this excites to vibrations, at least two vibration sensors (9), which in the longitudinal direction of the at least one measuring tube (2, 3) in front of and behind the at least one exciter assembly

(8) are arranged, and an evaluation device (11) which controls the at least one excitation arrangement (8), receives vibration signals (10) from the at least two vibration sensors (9) and determines a first indicator for a process fault, wherein a difference signal of

Oscillation signals (10) of the vibration sensor (9) and its frequency spectrum are determined and the indicator corresponds to a quotient of a weighted sum of upper frequencies and a fundamental frequency of the frequency spectrum, characterized by when a predetermined threshold value for the first indicator is exceeded,

Determining at least one further indicator and classifying the process disturbance using the at least two determined indicators.

2. The method of claim 1, wherein the further indicators from the vibration signals (10) are determined.

3. The method of claim 1, wherein the further indicators from an actuator current or an excitation frequency for the exciter assembly (8) are determined.

A method according to any one of the preceding claims, wherein the further indicators are periodically calculated and a predetermined number of values are stored for each further indicator.

5. The method of claim 1, wherein classifying the process disturbance based on multiple thresholds for each indicator.

6. The method of claim 1, wherein classifying the process disturbance based on minimum and maximum envelopes for each indicator.

7. Coriolis mass flowmeter (1), comprising: at least one measuring tube (2, 3) through which a medium flows, at least one excitation arrangement (8) which is arranged in the middle region of the at least one measuring tube (2, 3) and this vibrates, at least two vibration sensors (9), which in

Longitudinal direction of the at least one measuring tube (2, 3) are arranged in front of and behind the at least one exciter arrangement (8), and an evaluation device (11) which is set up to control the at least one excitation arrangement (8), vibration signals (10) of the at least receive two vibration sensors (9) and to determine a first indicator for a process disturbance, wherein a difference signal of the vibration signals (10) of the vibration sensor (9) and its frequency spectrum are determined and the indicator a quotient of a weighted sum of the upper frequencies and a Corresponds to the fundamental frequency of the frequency spectrum, characterized in that, when a predetermined threshold value for the first indicator is exceeded, the evaluation device (11) is further configured to determine at least one further indicator and the process disturbance by means of at least two specific indicator

8. System with at least one pipe (14), to which at least one Coriolis mass flowmeter (1) is flanged according to claim 7.