Method for verifying the valid signal content of a periodic sensor signal

Abstract

The present invention relates to a method for verifying the periodic effective signal component of a sensor signal of a flow meter, wherein the effective signal component of the sensor signal is used to determine a flow measurement value of a flowing medium, the method comprising: detecting a periodic measurement variable during a time interval t1 and outputting a periodic sensor signal; the sensor signal is subjected to a time / frequency analysis and a time-dependent spectrum of the sensor signal is determined; the time-dependent spectrum is checked with respect to features that are characteristic for the effective signal; and in the case where the expected characteristic feature is found in the spectrum, the component of the sensor signal having the characteristic feature is verified as an effective signal component, and the flow measurement value is determined based on the verified effective signal component; or in the case where the expected characteristic feature is not found in the spectrum, the component of the sensor signal without the characteristic feature is discarded as an interference signal component.

Description

Technical Field

[0001] The invention relates to a method for verifying the effective signal content of a periodic sensor signal of a flow meter, wherein the effective signal content of the sensor signal is used to determine a flow measurement value of a flowing medium. Background Art

[0002] A variety of flow meters are known from the prior art, which determine the value of the flow rate of a flowing medium based on different operating principles. All flow meters have in common that a measured variable is recorded using a sensor and a corresponding sensor signal is output by the sensor.

[0003] Eddy flow measurement is based on frequency measurement, for example. The frequency of the eddy currents interrupted by the introduction of an interfering body into the flowing medium is determined. For this purpose, a pressure sensor records the pressure fluctuations (measured variable) caused by the eddy currents interrupted at the interfering body. The sensor signal output by the pressure sensor is subjected to a frequency analysis. For example, in the case of magnetic induction flowmeters (in the case of correspondingly periodic excitation) and in the case of Coriolis mass flowmeters, there are also other periodic sensor signals.

[0004] In real systems, in addition to the really interesting valid signals, interference signals always appear, which may also be periodic. Therefore, the sensor signal output by the sensor has an effective signal component and an interference signal component. In the case of eddy current flow measurement, such interference signals may be caused, for example, by vibrations of the measuring meter, especially vibrations of the pump or valve. However, interference signals may also be caused, for example, by bubbles in the flowing liquid medium. These interference signals may cause: the flow measurement is affected and some flow measurement values ​​are distorted. For example, very strong interference signals may be mistakenly regarded as valid signals and the flow measurement value is determined based on them. Summary of the invention

[0005] The object of the present invention is therefore to provide a method with which useful signals and interfering signals can be reliably distinguished.

[0006] According to the invention, this object is achieved primarily and essentially by the method described below.

[0007] The method according to the present invention is characterized in that: first, in the detection step, the detection is performed at the time interval t 1 Preferably, the time interval t 1 should be selected so that a plurality of periods of the sensor signal are output. The more periods of the sensor signal are output, the more reliable the result of the method. In one embodiment, the time interval t 1can be selected by the user, wherein the user can in particular use empirical values ​​for the duration of a signal cycle to determine the time interval t 1 .

[0008] In an analysis step following the detection step, the sensor signal is subjected to a time / frequency analysis and a time-dependent frequency spectrum of the sensor signal is determined.

[0009] In order to obtain information not only about the frequency content of the signal at a single point in time, the method according to the invention determines the content of the sensor signal consisting of individual frequencies that change over time. Therefore, if we talk about the time-dependent spectrum in the context of this application, we mean determining the amplitude spectrum of the signal at multiple points in time.

[0010] In a preferred design, the time-dependent spectrum is output in the form of a spectrogram or a scalogram. The spectrogram or scalogram is used to present the image of the time-dependent spectrum of the signal. In order to present not only the frequency of the signal with respect to time but also the amplitude of the signal with respect to time, the frequency of the signal with respect to time is usually plotted in the spectrogram or scalogram, wherein the amplitude is presented by the color coding of the plotted signal.

[0011] According to the invention, it has been identified that periodic useful signals of a meter have characteristic features, based on which they can be clearly identified as useful signals and thus distinguished from interfering signals. Therefore, the method according to the invention further provides that in the checking step, the time-dependent frequency spectrum is checked with respect to the characteristic features for the useful signal.

[0012] In one embodiment, the time-dependent frequency spectrum is checked with respect to a plurality of mutually different characteristic features. This embodiment is based on the knowledge that a useful signal can have a plurality of characteristic features simultaneously. If the frequency spectrum is checked with respect to a plurality of these characteristic features, the reliability of the method according to the invention can be increased.

[0013] If the expected characteristic features are found, in a verification step, the components of the sensor signal having these characteristic features are verified as valid signal components. Subsequently, the flow measurement value is determined based on the components of the sensor signal verified as valid signal components. The method according to the invention makes it possible to identify the valid signal based on time-dependent effects, which can be used for analysis based on the time-dependent frequency spectrum.

[0014] The determination of the flow measurement value need not be based on the entire useful signal component, but can in particular be based on a portion of the useful signal. When determining the flow measurement value with the aid of a vortex flowmeter, in particular the frequency of the amplitude peak of the useful signal with the largest amplitude can be used as a basis.

[0015] If the expected characteristic feature is not found, components of the sensor signal which do not have this characteristic feature are discarded as interference signal components in a rejection step.

[0016] In an embodiment in which the spectrum is checked with respect to a plurality of characteristic features, a component of the sensor signal is verified as a valid signal component if it already has only one of the expected characteristic features. Thus, in this embodiment it is not provided that all expected characteristic features must also be found, but rather it is sufficient to find only one expected characteristic feature.

[0017] According to an alternative embodiment of the method of the invention, it is provided that before a component of the sensor signal is verified as a valid signal component, all expected characteristic features must be found. If only one of the expected characteristic features is not found, the corresponding component of the sensor signal is discarded as a disturbing signal component.

[0018] A further embodiment of the method according to the invention provides that the user can select the number of characteristic features that must be found in order for a component of the sensor signal to be validated as a valid signal component.

[0019] In a further embodiment of the method according to the invention, it is provided that the user can select the type of characteristic feature to be found, based on which the component of the sensor signal is verified as a valid signal.

[0020] In the method according to the invention, the analysis step can be implemented in different ways and methods. A particularly preferred embodiment provides that the time / frequency analysis is performed by means of a short-time Fourier transform. In a variant, the time / frequency analysis is performed by means of a Gabor transform. The Gabor transform is a special short-time Fourier transform in which a Gaussian function is used as a window function. If the short-time Fourier transform, i.e., in particular also the Gabor transform, is used as the basis, then in connection with the image presentation of the signal, we talk about a spectrogram.

[0021] In an alternative embodiment of the method according to the invention, the time / frequency analysis is performed by means of a wavelet transform. In the wavelet transform, a time-localized function is used as a basis, in particular a time-localized wave function. If the wavelet transform is used as a basis, then scalograms are mentioned in connection with the image presentation of the signal.

[0022] In another alternative embodiment of the method according to the invention, the time / frequency analysis is performed by means of a Hilbert-Huang transformation. In this embodiment, it is advantageous if the time / frequency ambiguity of the transformation is small.

[0023] In order to be able to reasonably identify a characteristic feature, it must first be carefully analyzed in an error-free state. In a preferred embodiment of the method according to the invention, it is provided that a predetermined value of the ratio of the amplitude of a signal component of a fundamental frequency to the amplitude of a higher-order signal component is selected in the frequency spectrum as a characteristic feature. In addition or as an alternative thereto, a value of the ratio of the amplitudes of two higher-order signal components can also be selected. Therefore, in order to be able to predetermine one or more values ​​for the amplitude ratio, these values ​​must first be determined in an error-free state. If the amplitude ratio then corresponds to a predetermined value during the flow measurement, the corresponding sensor signal component is verified as a valid signal component.

[0024] In a preferred embodiment of the method according to the invention, a predetermined amplitude modulation and / or a predetermined frequency modulation of the signal component is selected as a characteristic feature. Thus, in a first variant, the signal component is verified as a valid signal if it is amplitude modulated in a predetermined manner. In a second variant, the signal component is verified as a valid signal if it is frequency modulated in a predetermined manner. In a third variant, the two characteristic features are combined with one another and the signal component is verified as a valid signal if it is not only amplitude modulated in a predetermined manner but also frequency modulated in a predetermined manner.

[0025] If the time-dependent spectrum has higher-order signal components, amplitude modulation and / or frequency modulation can occur in addition to or exclusively in the higher-order signal components.

[0026] In another embodiment, the occurrence of jitter in a time-dependent spectrum, in particular in a spectrogram or a scalogram, is selected as a characteristic feature. Thus, if jitter can be found in the spectrum, the corresponding sensor signal component is verified as a valid signal component. Jitter is to be understood as a deviation in the periodicity of an essentially periodic signal. In the case of a vortex flowmeter, jitter is to be understood, for example, as the time difference between individual vortex interruptions. Thus, in the spectrum, jitter can be noticed due to the broadening of individual amplitude peaks.

[0027] In order to be able to reliably identify characteristic features in the spectrum, a further embodiment of the method according to the invention is characterized in that in a partition selection step, a partition of the spectrum over time is selected and in a partition analysis step, the partition is subjected to a further time / frequency analysis. Here, the partition can be selected by the user. Likewise, the partition can be selected according to a predetermined criterion, for example so that a predetermined amplitude value of a signal component is found in the spectrum and then the area around the corresponding signal component is selected as a partition in which the amplitude value drops to a predetermined percentage of the predetermined amplitude value.

[0028] In a preferred embodiment, the partition is subjected to a further Gabor transformation or a further wavelet transformation. The choice of the transformation can be made according to the transformation used in the analysis step. Therefore, in one embodiment, it is provided that the same transformation as in the analysis step is performed. In another embodiment, a transformation different from the transformation performed in the analysis step is performed.

[0029] A particularly preferred embodiment of the method according to the present invention is not only suitable for verifying the effective signal component of the periodic sensor signal. More precisely, this particularly preferred embodiment enables long-term diagnosis and thereby enables in particular the identification of changes in the effective signal component. The characteristic of this particularly preferred embodiment is that the method is executed multiple times in succession over time, wherein after each method cycle, information about the verified effective signal component is stored. The stored information can be, for example, a determined flow value, or can be information about one or more characteristic features found, i.e., for example, the value of the amplitude ratio of two peaks in the spectrum. After each method cycle, the currently verified effective signal component is compared with the effective signal component of the previous method cycle. In particular, the information stored separately can be compared. In the case where these effective signal components deviate from each other by exceeding a predetermined tolerance range, a notification is output. The output of the notification can be carried out in different ways and methods, i.e., for example, by outputting an optical signal or an acoustic signal. Similarly, the notification can be output by internal storage of the notification.

[0030] In one embodiment, the individual method cycles are executed continuously one after the other. In another embodiment, the individual method cycles are executed at intervals of a predetermined waiting time T. Here, the waiting time T can be constant between the individual method cycles and can be selected in particular by the user. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Now, there are various possibilities for designing and expanding the method according to the invention for verifying the effective signal component of a periodic sensor signal in detail. For this purpose, reference is made to the description of preferred embodiments in conjunction with the accompanying drawings. In the drawings:

[0032] Figure 1 A block diagram showing a first embodiment of the method according to the present invention;

[0033] Figure 2 A block diagram illustrating an embodiment for finding representative features in a spectrum;

[0034] Figure 3 A block diagram showing a second embodiment of the method according to the invention;

[0035] Figure 4a A schematic diagram of a purely valid signal is shown;

[0036] Figure 4b A schematic diagram of a sensor signal having a valid signal component and an interference signal component is shown;

[0037] Figure 5 shows a spectrum diagram of a valid signal that is stable in frequency and amplitude over time;

[0038] Figure 6 shows a spectrum diagram of an amplitude modulated effective signal;

[0039] Figure 7 shows a spectrum diagram of a frequency modulated effective signal;

[0040] Figure 8 The spectrum diagram of the effective signal which is not only amplitude modulated but also frequency modulated is shown;

[0041] Fig. 9 shows a spectrum diagram of a valid signal with harmonics;

[0042] Fig.10 shows a spectrum diagram of an effective signal with amplitude modulated harmonics;

[0043] Fig.11 shows a spectrum diagram of a useful signal with interference signal components occurring in a limited time; and

[0044] Fig.12The spectrum diagram of the effective signal in the presence of flow pulses is shown. DETAILED DESCRIPTION

[0045] exist Figure 1 1 shows a block diagram of a first embodiment of the method according to the invention for verifying the effective signal content of a periodic sensor signal of a flow meter. In a detection step 101, a measurement variable is detected by a sensor of the flow meter. The detection of the measurement variable is performed at a time interval t 1 During the period, the time interval t 1 For example, it can be predetermined by the user. Then, the sensor outputs a periodic sensor signal. The sensor signal has an effective signal component and an interference signal component. In the analysis step 102, the sensor signal is subjected to a time / frequency analysis. In the embodiment shown, the time / frequency analysis is performed by means of a Gabor transform. Not shown, but the present invention also includes: using any other transform, such as a wavelet transform, to perform the time / frequency analysis. After the time / frequency analysis, the time-related spectrum of the sensor signal is determined. In the inspection step 103, the spectrum is inspected for one or more characteristics that are characteristic of the effective signal. In the case of finding one or more expected characteristic features, in the verification step 104, the components of the sensor signal having these characteristic features are verified as effective signal components. Then, the flow measurement value can be determined based on the components of the sensor signal that are verified as effective signal components. If one or more expected characteristic features are not found in the spectrum, then in the discarding step 105, the corresponding sensor signal component is discarded as an interference signal component.

[0046] Figure 2 A block diagram of an embodiment for finding a characteristic signal in a spectrum, i.e. an embodiment of an inspection step 103 is shown. In order to find the expected characteristic features, in a partition selection step 103a, a partition of the time-dependent spectrum that should be examined more closely is selected. In this case, the selection of the partition can be performed by the user. In a partition analysis step 103b, the selected partition is subjected to a further time / frequency analysis. In the example shown, the further time / frequency analysis is performed by means of a wavelet transform.

[0047] Figure 3 A block diagram of another preferred embodiment of a method for verifying a valid signal component of a periodic sensor signal is shown. In this case, the first method step corresponds to Figure 1 The method steps shown in FIG. 1 are as follows: In a detection step 101, detection is performed at a time interval t 1A periodic measurement parameter is determined during the period, and a periodic sensor signal is output by the sensor. In an analysis step 102, the sensor signal is subjected to a time / frequency analysis. After determining the time-dependent spectrum of the sensor signal, in a check step 103, the determined spectrum is checked with respect to features that are characteristic for a valid signal. In a verification step 104, when the expected characteristic feature is found, the component of the sensor signal having the characteristic feature is verified as a valid signal component, and the flow measurement value is determined based on the verified valid signal component. Next, in an information storage step 106, information about the verified valid signal component is stored. This information can be, for example, a characteristic feature or can be a flow measurement value determined based on the verified valid signal component. Now, the method is as follows. Figure 3 The embodiment shown in FIG. 1 is characterized in that the method described so far is executed again, ie the method steps are executed multiple times in succession over time. Therefore, now in the second detection step 101 ', at another time interval t 2 During this time, the sensor detects the measured variable and outputs a periodic sensor signal. In the second analysis step 102', the sensor signal is subjected to a time / frequency analysis and the time-dependent spectrum of the sensor signal is determined. In the second check step 103', the spectrum is checked with respect to a characteristic feature. Here, the characteristic feature corresponds to the characteristic feature that has been sought in the first method cycle. In the case of finding the expected characteristic feature, in the second verification step 104', the sensor signal component with the characteristic feature is verified as a valid signal component, and another flow measurement value is determined based on the verified second valid signal component. In the second information storage step 106', the same information about the valid signal component as in the first method cycle is stored. In the comparison step 107 that now follows, the valid signal component of the second method cycle is compared with the valid signal component of the first method cycle. The comparison is performed based on the stored information of the corresponding valid signal component. In the case where these valid signal components deviate from each other by more than a predetermined tolerance range, in the notification step 108, a corresponding notification is output. Therefore, the embodiment of the method can achieve a good possibility of monitoring the flow measurement. In particular, this allows long-term monitoring, with which changes in the flow meter can be easily detected.

[0048] Figure 4a The spectrum of an undisturbed sensor signal, ie a sensor signal consisting only of the useful signal, is shown at time t. The useful signal has: 0 The amplitude A 0 The peak amplitude at frequency f 1 The amplitude A1 The first-order amplitude peak value; and at frequency f 2 The amplitude A 2 The second-order amplitude peak value. Characteristic for the useful signal shown is that the amplitude A of the first-order amplitude peak 1 is the fundamental frequency f 0 The amplitude of the peak amplitude under 0 half of the amplitude of the second-order amplitude peak 2 is the amplitude of the first-order peak amplitude A 1 is half as large and is therefore at the fundamental frequency f 0 The amplitude of the peak amplitude under 0 In the present case, the following feature is selected as a representative feature: the value of the ratio of the amplitude of the O-order amplitude peak to the amplitude of the (O+1)-order amplitude peak is one-half.

[0049] Figure 4b The spectrum of the following sensor signal at time t is shown, which sensor signal has a disturbance signal component in addition to the useful signal component. The useful signal component corresponds to Figure 4a The interference signal component consists of the effective signal at frequency f s0 The amplitude A s0 The additional peak amplitude and frequency f s1 The amplitude A s1 The interference signal component can be identified as an interference signal component due to the following circumstances: at frequency f s0 and f s1 The amplitude of the peak amplitude under A s1 / A s0 It is not worth half and thereby does not have that characteristic feature.

[0050] exist Figures 5 to 12 In FIG. 1 , time-dependent spectra are respectively shown in the form of spectrograms, which respectively show useful signals having different characteristic features.

[0051] exist Figures 5 to 12 In FIG. 1 , time t is plotted on the x-axis, frequency f is plotted on the y-axis and amplitude A is plotted in a color-coded manner. The corresponding color scale is shown on the right side of the spectrogram.

[0052] Figure 5 A useful signal which is stable over time in terms of frequency and amplitude is shown without any special characteristic features.

[0053] In order for a signal to be verified as a valid signal, it can be stipulated that the signal has exactly one characteristic feature. For example, the characteristic feature can be a certain amplitude modulation or a certain frequency modulation. Figure 6 The useful signal shown in has an amplitude modulation which corresponds to the amplitude modulation defined as the characteristic feature.

[0054] exist Figure 7 The valid signal shown in is frequency modulated in a representative manner. Therefore, if the signal has the same frequency modulation, the signal is verified as a valid signal.

[0055] In order to verify that a signal is a valid signal, it can also be provided that the signal has multiple characteristic features, for example, not only amplitude modulated in a certain way but also frequency modulated in a certain way. If the signal has only one of these two characteristic features, the signal is not verified as a valid signal, but is discarded as a interference signal. Figure 8 The useful signal shown in has two characteristic features: On the one hand, it has an amplitude modulation which is defined as a characteristic feature, and on the other hand, it also has a frequency modulation which is also defined as a characteristic feature.

[0056] like Figure 4a That way, Fig. 9 The useful signal is also shown, which has: 0 The amplitude A 0 The signal component at frequency f 1 The amplitude A 1 The first-order signal component at frequency f 2 The amplitude A 2 However, unlike Figure 4a , not just a time point t, but actually the entire spectrum over time. Characteristic for the effective signal shown is: the amplitude A of the first-order amplitude peak 1 With the fundamental frequency f 0 The amplitude of the peak amplitude A 0 The ratio is determined by the second-order amplitude peak amplitude A 2 The amplitude A of the first-order peak amplitude 1 In the present case, a certain value of the ratio of the amplitude of the 0th order amplitude peak to the amplitude of the (0+1)th order amplitude peak is selected as the representative feature.

[0057] exist Fig.10 In addition to the spectrum shown in Fig. 9 In addition to the characteristic features shown in , the characteristic features of the amplitude modulation of the harmonics are also shown. Therefore, if a signal has a characteristic amplitude modulation in addition to a predetermined amplitude ratio of signal components of different orders, then the signal is verified as a valid signal.

[0058] Fig.11 Shown in Figure 5 The effective signal with constant amplitude and constant frequency is shown in Fig.11 Temporally limited interference signal components also appear in the spectrogram shown in If, for example, the spectrogram is based on eddy current flow measurement, these temporally limited interference signals can be caused by gas bubbles that occur in the medium.

[0059] exist Fig.12 , which shows a useful signal in the presence of flow pulses. In the case of the useful signal shown, the frequency and amplitude of the signal are combined with each other in a characteristic manner.

[0060] Reference numerals

[0061] 101 Detection Step 1

[0062] 101' Detection Step 2

[0063] 102 Analysis Step 1

[0064] 102'Analysis Step 2

[0065] 103 Check Step 1

[0066] 103' Check step 2

[0067] 103a Partition Selection Steps

[0068] 103b Partition Analysis Steps

[0069] 104 Verification Step 1

[0070] 104' Verification Step 2

[0071] 105 discard step 1

[0072] 105' discard step 2

[0073] 106 Information Storage Step 1

[0074] 106' Information storage step 2

[0075] 107 Comparison Steps

[0076] 108 Notification Steps

Claims

1. A method for verifying the periodicity of a valid signal component of a sensor signal of a flow meter, wherein: The effective signal component of the sensor signal is used to determine a flow measurement value of the flowing medium. It is characterized in that In a detection step (101), a periodic measurement variable is detected during a time interval t1, and a periodic sensor signal is output; In an analysis step (102), the sensor signal is subjected to a time / frequency analysis and a time-dependent frequency spectrum of the sensor signal is determined; In a checking step (103), the time-dependent spectrum is checked for features that are characteristic of a valid signal; and In case the expected characteristic feature is found in the time-dependent spectrum, in a verification step (104), the component of the sensor signal having the characteristic feature is verified as a valid signal component, and the flow measurement value is determined based on the verified valid signal component; or If the expected characteristic feature is not found in the time-dependent spectrum, in a discarding step (105), the components of the sensor signal without the characteristic feature are discarded as interference signal components. Therein, in order to identify representative features, in a partition selection step (103a), partitions of the time-dependent spectrum are selected; and in a partition analysis step (103b), the partitions are subjected to further time / frequency analysis.

2. The method according to claim 1, characterized in that The time / frequency analysis is performed with the aid of a short-time Fourier transform.

3. The method according to claim 1 or 2, characterized in that: The time / frequency analysis is performed with the aid of a wavelet transform, a Gabor transform or a Hilbert-Huang transform.

4. The method according to claim 1 or 2, characterized in that: A predetermined amplitude modulation and / or a predetermined frequency modulation of the signal components is selected as a characteristic feature.

5. The method according to claim 1 or 2, characterized in that: A predetermined value of the ratio of the amplitude of a signal component of a fundamental frequency to the amplitude of a higher-order signal component and / or a value of the ratio of the amplitudes of two higher-order signal components is selected as a characteristic feature in the time-dependent spectrum.

6. The method according to claim 1 or 2, characterized in that: The occurrence of jitter in the time-dependent spectrum is selected as a characteristic feature.

7. The method according to claim 1 or 2, characterized in that: The method is performed at least twice in succession over time, wherein, after each method cycle, information about components of the sensor signal that are verified as valid signal components is stored; and after each method cycle, the currently verified valid signal components are compared with the valid signal components of the previous method cycle; and, if the valid signal components deviate from each other by more than a predetermined tolerance range, a notification is output.

Images