The present application is related to and claims the priority benefit of German Patent Application No. 10 2017 115 251.2, filed on Jul. 7, 2017, and International Patent Application No. PCT/EP2018/066273, filed on Jun. 19, 2018, the entire contents of which are incorporated herein by reference.
The present invention relates to a measurement pickup for determining the mass flow rate of a liquid, comprising: at least one measurement tube for carrying the liquid, wherein the measurement tube has an inlet-side end section and an outlet-side end section; at least one support body on which the measurement tube is mounted in the region of the inlet-side end section and in the region of the outlet-side end section; at least one exciter for exciting vibrations of the measurement tube in at least one vibration mode; at least one vibration sensor for detecting the vibrations of the measurement tube and for outputting vibration-dependent signals; an operating and evaluation circuit for driving the at least one exciter, for receiving the vibration-dependent signals of the vibration sensor, and for ascertaining a measured value representing the mass flow rate as a function of at least the vibration-dependent signals of the sensor.
Published patent application DE 10 2010 003 948 A1 describes a generic measurement pickup and its operating procedure, for example.
The processing of the vibration-dependent signals, which, for example, takes place for evaluations, for control applications, for visual representation, for documentation purposes, etc., is increasingly carried out digitally. In the process, the varying vibration-dependent signals are initially recorded analogously. In a known manner, such analog signals can be sampled by sampling at a high sampling frequency and converted into a time-discrete, one-dimensional measurement signal having a sequence of temporally successive measured values. However, the measurement signal obtained in this way has a noise which can be pronounced to different degrees as a result of the varying properties of the liquid flowing in the measurement tube, for example as a result of the level of gas content. It is known to filter the measurement signal by means of a low-pass filter, which is however accompanied by a reduction of the data rate. The reduction of the data rate is problematic in particular in time-critical applications if a superordinate unit, for example a process control unit or process monitoring unit, has to be informed as promptly as possible about occurring changes in the mass flow rate.
It is known to suppress the noise with adaptive filters, but these filters are controlled as a function of the noise of the measured variable. However, this may result in a delayed adaptation of the filter to changed measurement conditions, in particular when media properties, such as a gas content of a liquid, change suddenly.
The object of the present invention is, therefore, to find a remedy.
The measurement pickup according to the invention for determining the mass flow rate of a liquid comprises:
In a development of the invention, the control variable is independent of the mass flow rate.
In a development of the invention, the control variable depends on a signal representing the level of the gas content.
In a development of the invention, the operating and evaluation circuit is designed to ascertain the control variable as a function of a gas-content-dependent fluctuation of a natural frequency of the measurement tube vibration.
In a development of the invention, the operating and evaluation circuit is designed to ascertain the control variable as a function of a gas-content-dependent damping of the measurement tube vibrations.
In a development of the invention, the operating and evaluation circuit is designed to ascertain the control variable as a function of a gas-content-dependent damping fluctuation of the measurement tube vibrations.
In a development of the invention, the low-pass filter comprises an adaptive filter, wherein the control variable has a weighting function with which a current, preliminary measured value contributes to a measured value to be output, wherein the weighting function depends on a current weighting time constant, which is ascertained as a function of the gas content, wherein the weighting function km(τ) is a function that decreases monotonously with an increasing weighting time constant τ, in particular a function of the reciprocal value of the time constant, for example of the type km(τ)=(1−exp(−(t/τ)), where t is the period between two preliminary measured values.
In a development of the invention, the operating and evaluation circuit is designed to first ascertain a time constant trend value τs as a function of current signals dependent on the gas content, to compare a last valid weighting time constant τi-1 with the time constant trend value τs, to increase the value in comparison to the last weighting time constant τi-1 in order to ascertain a current weighting time constant τi when the last valid weighting time constant τi-1 is less than the time constant trend value τs, and to reduce the value in comparison to the last weighting time constant τi-1 when the last valid weighting time constant τi-1 is greater than the time constant trend value τs.
In a development of the invention, the following applies to the case of a weighting time constant to be increased: (τi−τi-1)/t=a, where t is the time interval between two successive measured values, and a is a constant which is not more than 8, in particular not more than 4, and preferably not more than 2.
In a development of the invention, the following applies to the case of time constants to be reduced:
τi/τi-1=b, where 0.9<b<1, in particular 0.95<b<1, preferably 0.975<b<1.
In a development of the invention, the operating and evaluation circuit is designed to ascertain the time constant trend value τs as a function of the fluctuation of the natural frequency of the measurement tube vibration normalized with the square of the natural frequency.
In a development of the invention, the operating and evaluation circuit is designed to ascertain the time constant trend value τs as follows:
where X is a constant, fi is a current natural frequency of a measurement tube vibration, and kr, kFF, kF, are weighting factors, in particular constant weighting factors.
According to one embodiment of the invention, a current filtered measured value of the mass flow rate {dot over (m)}f,i={dot over (m)}f(ti) is ascertained by a weighted addition of the current unfiltered measured value of the mass flow rate {dot over (m)}i=={dot over (m)}(ti), to the previous filtered measured value of the mass flow rate {dot over (m)}f,i-1={dot over (m)}f(ti-1), that is to say:
{dot over (m)}f,i:=km(τi)·{dot over (m)}i+(1−km(τi))·{dot over (m)}f,i-1,
where km(τi) is a sequence of gas-content-dependent weighting factors, which were ascertained in the previously described manner.
The invention is now described in further detail on the basis of an exemplary embodiment shown in the drawings.
They show:
An exciter 8 is arranged between the two measurement tubes A and B. In the present embodiment, the exciter 8 is positioned at a reversal point of the arc formed by the two measurement tubes A and B in each case. The two measurement tubes A and B can be periodically deflected in relation to one another by the exciter 8 so that they execute bending vibrations. The two measurement tubes A and B are also mechanically coupled to one another on the inlet side and on the outlet side by corresponding coupling elements 10, 12.
Two vibration sensors 14, 16 respectively extend between the two measurement tubes A and B at an inlet-side and at an outlet-side section thereof. In the present embodiment, the two vibration sensors 14, 16 can each detect the change in distance between the two measurement tubes A, B, i.e., their combined amplitude. The vibration sensors 14, 16 each output a sensor voltage as a function of the vibrations of the measurement tubes A and B. This is initially an analog measurement signal in each case. Excitation of the exciter 8 by application of a corresponding excitation voltage and processing and evaluation of the analog measurement signals provided by the vibration sensors 14, 16 takes place by a correspondingly designed operating and evaluation circuit 18, which is shown in
A phase difference Δφ(t) of the vibrations of the two measurement tubes A, B between the two measuring points formed by the vibration sensors 14, 16 is to be determined in order to ascertain measured values of the mass flow rate {dot over (m)}(t). The mass flow rate of the fluid flowing in the pipeline can be determined in particular, as is known to the person skilled in the art, from this phase difference Δφ(t) as follows:
{dot over (m)}(t):=k·tan(Δφ(t)/2)/(2π·f),
where k is a constant specific to the respective measuring device, and f describes the current vibration frequency of the measurement tubes.
In order to determine a sequence Δφi=Δφ(ti) of discrete values of the phase difference Δφ(t), the operating and evaluation circuit 18 respectively determines, from the analog measurement signals provided by the vibration sensors 14, 16, the phase information φ1(ti), φ2(ti) of the vibration of the two measurement tubes A, B at the measuring points formed by the two vibration sensors 14, 16. For this purpose, in particular the analog measurement signal, which is respectively provided by the vibration sensors 14, 16 and is generally formed by a sensor voltage corresponding to the vibrations, can be sampled at a high sampling frequency, for example at 40 kHz. As a result, a time-discrete, one-dimensional measurement signal having a temporal sequence of measured values is obtained. As a rule, the individual measured values of the sequence are also quantized in order to enable digital processing. Furthermore, in the present embodiment, each measurement signal is converted into an analytical signal consisting of a real part R(ti) and an imaginary part I(ti). For this purpose, two filters having a phase difference of 90° may, for example, be used in parallel in a known manner. Furthermore, the present embodiment provides that the data rate of the analytical signal is reduced. This can be done, for example, via corresponding decimation stages. From the analytic signal, as is familiar to the person skilled in the art, (time-dependent) amplitude information A1(ti), A2(ti) and (time-dependent) phase information φ1(ti), φ2(ti) of the vibration of the two measurement tubes A, B can respectively be obtained at the respective measuring point. The phase difference Δφi=Δφ(ti) between the two measuring points formed by the vibration sensors 14, 16 can be obtained by forming the difference of the phase information φ1(ti), φ2(ti) of the vibrations detected by the two vibration sensors 14, 16. Since the signal processing in the present embodiment takes place largely digitally, the respective processed measurement signals are time-discrete so that specific instants ti are referred to in each case.
The operating and evaluation circuit as a rule evaluates the amplitude information A1(ti), A2(ti) and the phase information φ1(ti), φ2(ti). In particular, the excitation of the measurement tubes A, B by the exciter 8 is controlled in each case as a function of these variables.
The operating and evaluation circuit is further designed to determine, from the sequence of time-discrete phase difference values Δφi, a corresponding sequence of measured values of the mass flow rate through the pipeline {dot over (m)}i as follows:
{dot over (m)}i:=k·tan(Δφi/2)/(2π·fi),
where {dot over (m)}i:=m(ti), Δφi=Δφ(ti), and fi=f(ti).
Other signal processing or digital or analog signal processing alternative to the above-explained signal processing, such as amplification, zero point compensation, etc., can also be performed in order to provide the sequence of measured values of the mass flow rate {dot over (m)}i. This can take place both before and after the signal processing described below with an adaptive filter.
The sequence of measured values of the mass flow rate {dot over (m)}i, or measured values for short, is, in particular with fluctuating gas content of a measured medium flowing in the pipeline, subject to differently strong fluctuations, as shown in
The effective operating principle of the adaptive filter is made possible by controlling a filter parameter as a function of variables which are independent of the phase difference measurement signal, as explained in more detail below.
For operation, the adaptive filter requires a current time constant τi:=τ(ti), which in the present exemplary embodiment is ascertained as a function of the fluctuation of the vibration frequency of the measurement tube vibrations. Wherein the fluctuation, for its part, is also detected via a filter in order to avoid excessive changes in the time constant.
In this case, a current time constant trend value τs,i is first ascertained, to which a last valid time constant is compared. If the time constant trend value is greater than the last valid time constant τi-1 by more than a tolerance value, a currently valid time constant τi with a value greater than that of the last valid time constant is ascertained. If the time constant trend value is smaller than the last valid time constant τi-1, a currently valid time constant τi with a value smaller than that of the last valid time constant is ascertained.
In this exemplary embodiment, the current time constant trend value τs,i is ascertained on the basis of the fluctuation of the frequency of the measurement tube vibrations, which is normalized with the square of this frequency. Specifically, the filter uses the following equation:
The weighting factor kτs with which a current frequency fluctuation contributes to the current time constant trend value is a few % and can be ascertained as:
kτs=(1−exp(−ts/τk)),
where ts is the measurement interval, that is to say the time between two frequency measured values, for example 20 ms. τK is a defined time constant which is a multiple of ts, for example 0.5 s<τk<2 s, in particular τk=1 s.
The factor X is a constant having a value between 10 and 200, for example 50.
When the thus-ascertained time constant trend value τs,i is greater than the sum of the last valid weighting time constant τi-1 and a tolerance value, which, for example, corresponds to half a measurement interval ts, the weighting time constant is increased, for example by addition of a constant time increment, which corresponds to a measurement interval ts, for example 20 ms.
If, on the other hand, the time constant trend value τs,i is less than the last valid weighting time constant τi-1, the latter is multiplied by a factor b<1 in order to ascertain a current weighting time constant τi, i.e., τi=b·τi-1, where: 0.9<b<1, in particular 0.95<b<1, preferably 0.975<b<1.
The combination of increases in the weighting time constant by addition of a moderate increment with a reduction by multiplication by a factor <1 has the effect on the one hand that the weighting time constant increases relatively quickly with occurring gas contents starting from small values, but that it then, in contrast to a sequence of multiplications, does not increase exponentially without bounds, and that on the other hand the weighting time constant decreases exponentially as the gas content decreases so that the measured value ascertainment can be adapted quickly to the more favorable conditions. An example of a jump response of the weighting time constant to a jump of the time constant trend value is shown in
The weighting time constant τi contributes to the ascertainment of a weighting factor km(τi) for ascertaining a filtered measured value as follows:
km(τi)=(1−exp(−(ts/τi)), where ts is again the measurement interval. A current, filtered measured value {dot over (m)}f,i={dot over (m)}f (ti) is ascertained by a weighted addition of the currently unfiltered measurement signal {dot over (m)}i={dot over (m)}(ti) to the previous filtered measurement signal {dot over (m)}f,i-1={dot over (m)}f(ti-1), i.e.:
{dot over (m)}f,i=km(τi)·{dot over (m)}i+(1−km(τi)){dot over (m)}f,i-1.
As shown in
If the frequency fluctuations are used to ascertain the time constant trend value, the current frequency fluctuations and their normalizing frequency value contribute, as described above, with a weighting factor kτs to the time constant trend value, which is the function of a time constant.
Since the relative fluctuations of the fluctuations of the frequency are significantly greater than the relative fluctuations of the frequency, it may be expedient to use a larger time constant for the fluctuations than for the normalizing frequency. As shown in the following embodiment.
where X is a constant, fi is a current natural frequency of a measurement tube vibration, and kFF, kF, are weighting factors, in particular constant weighting factors, where:
kFF,<kF.
This ensures that a faster adaptation of the adaptive filter to an actually changing vibration frequency of the measurement tubes takes place.
Instead of the fluctuations of the vibration frequency of the measurement tubes, the damping of the vibration amplitudes A1(ti), A2(ti) at the current resonant frequency or the fluctuation of this damping can also be used. The damping Di=D(ti) results, for example, from a ratio of a current excitation signal Ei=E(ti) for exciting the vibrations and, for example, the sum of the amplitudes at resonance, i.e.:
Di=Ei/(A1(ti)+A2(ti))
A normalized current fluctuation of the FDi damping can be ascertained as follows:
FDi=(Di−Di−1)/Di−1
Based on these variables, a time constant trend value, with which the filter is to be controlled as described above, can be determined with appropriate normalizations in accordance with the methods described in connection with the frequency fluctuation.
Number | Date | Country | Kind |
---|---|---|---|
10 2017 115 251.2 | Jul 2017 | DE | national |
Filing Document | Filing Date | Country | Kind |
---|---|---|---|
PCT/EP2018/066273 | 6/19/2018 | WO |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2019/007679 | 1/10/2019 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
5555190 | Derby et al. | Sep 1996 | A |
6318156 | Dutton | Nov 2001 | B1 |
6389364 | Vyers | May 2002 | B1 |
20050274200 | Henry et al. | Dec 2005 | A1 |
20080243400 | Bell | Oct 2008 | A1 |
20170254688 | Patten | Sep 2017 | A1 |
Number | Date | Country |
---|---|---|
105865555 | Aug 2016 | CN |
102010003948 | Oct 2011 | DE |
1363108 | Dec 2002 | EP |
0104716 | Jan 2001 | WO |
2007022118 | Feb 2007 | WO |
2007035376 | Mar 2007 | WO |
Entry |
---|
Endress+Hauser Flowtec AG, E+H Flow Handbook, 3rd edition, 2006 (Year: 2006); provided by Applicant on Jun. 29, 2023. |
Number | Date | Country | |
---|---|---|---|
20200191632 A1 | Jun 2020 | US |