HK1056392B - Vibratory level sensor - Google Patents

Description

Oscillation limit state sensor

Technical Field

The invention relates to a method for determining and/or monitoring a predetermined filling state of a filling material in a container having a vibrating rod probe directed into the interior of the container, which vibrating rod probe is part of an electromechanical oscillation system.

Background

DE 3348119C 2 discloses a limit state sensor for determining a predetermined filling state, which has a vibrating rod probe that projects into a container and is excited to oscillate by an electrical excitation device. The frequency of the oscillation depends on the filling state in the container. The amplification and feedback of the electrically detected mechanical oscillation signal to the excitation input produces a system capable of oscillation, commonly referred to as an oscillator.

Similar devices and methods are described, for example, in DE 19720519C 2, EP 0985916 a1 and EP 0985917 a 1. They have in common that the described oscillation limit state sensor comprises in principle an electromechanical system which excites a continuous self-oscillation. Different possibilities are disclosed for building such oscillators, which are distinguished from each other in terms of the implementation of the mechanical oscillating element, the electromechanical converter system and the electrical scheme of the oscillator.

In particular, at least two oscillating bars, which are either coaxial or arranged parallel to one another, are used as mechanical oscillating elements.

This mechanical system capable of oscillation is usually driven electromechanically via a piezoelectric element excited by alternating current, and a piezoelectric element is likewise used to detect the mechanical oscillation and convert it back into electrical oscillation. For this purpose, the drive and detection means may be divided into two separate piezoelectric elements, into two separate areas of a single piezoelectric element, or may unify both functions in a single piezoelectric element.

The entire electromechanical oscillation system can be equivalently regarded as an electrical quadripole network in separate drive and detection elements, which transmits a voltage with a frequency-dependent amplitude and phase shift, which is applied to the two input poles, to the two output poles. The transmission performance of this four-terminal network is comparable to that of an electrical tank circuit.

A single piezoelectric element, used for both actuation and sensing, typically has two terminals, which can thus be equivalently characterized as an electrical two-terminal network. Electrical two-terminal networks are generally characterized by their electrical impedance, which may be frequency-dependent.

It is also to be noted here that a single piezoelectric element, which can be understood as an electrical two-terminal network, can be modeled on the performance of an electrical four-terminal network by time-division multiplexing, i.e. by periodic switching between the drive and detection functions on the terminals.

With the described oscillating system, which can be equivalently regarded as an electrical two-terminal network or four-terminal network, an electrical oscillator can be built up in a known manner. The components of the oscillator are always amplification and feedback devices, at least one of which should be frequency selective. A preferred embodiment of the oscillator contains, for example, a frequency selection in the feedback loop of the feedback amplifier circuit, which can be regarded as a four-terminal network of the electric resonant circuit.

In other embodiments, the oscillator is characterized by a feedback amplification device, wherein the amplification factor is dependent on the oscillation device, which can be considered as an electrical two-terminal network. The frequency-selective amplifier thus formed is dependent on the electrical impedance of the two-terminal network characterizing the oscillating structure in terms of its amplification.

A problem with all these oscillation limit state sensors is the design of the oscillator in such a way that reliable starting or continuous oscillation of the oscillator is ensured under all operating conditions. By immersing the sensor in the filling, for example, the damping of the oscillating system is increased, the natural frequency and the phase shift of the oscillating system are changed, and the feedback or amplification characteristic of the oscillator is thus subject to more or less strong variations. In order to reliably oscillate an oscillator under all the occurring operating characteristics of the electromechanical oscillation element, complex and expensive circuit and/or design measures are often required. Furthermore, the oscillating elements, which may be considered as two-terminal network or four-terminal network transmission elements, may have similar characteristics at different, partially distant frequency points. If the conditions for oscillation of the oscillator for different frequencies can thus be met, the system oscillates more randomly at one or the other of the frequencies and therefore the container filling status cannot be judged. Expensive measures must therefore be taken to try to always oscillate the system at the prescribed, desired frequency and to achieve a reliable report of the fill state of a container above or below a certain level.

Disclosure of Invention

The object of the invention is therefore to specify a method which permits the limit state detection with reliable oscillation sensors which project into the container and with little effort.

This object is achieved by the following scheme.

The method according to the invention for determining and/or monitoring a predetermined filling state of a filling material in a container having a vibrating rod probe directed into the interior space of the container, which vibrating rod probe is a component of an electromechanical oscillation system, is characterized in that during the step of determining and/or monitoring the predetermined filling state of the filling material, the oscillation system is excited at different frequencies and the response of the system is measured at all frequencies; after the excitation of the oscillating system, one or more characteristic parameters are measured and this or these parameters are taken into account when evaluating the measurements.

In the simplest case, this means that the determination and evaluation of the electrical transmission behavior of the oscillating system can be regarded as a four-terminal network, or that the impedance of the oscillating element, which operates as a two-terminal network, is determined and evaluated. This is achieved in particular via a frequency band in which the transmission performance or the impedance has a characteristic feature by which the filling state can be unambiguously inferred.

In a first preferred embodiment the oscillating system is excited with different frequencies and the response of the system is measured at all frequencies. In order to excite with different frequencies, a certain number of discrete frequency points can be set in succession in a targeted manner and the frequency process can be repeated periodically. Alternatively, it is also possible to drive continuously through a frequency band. This way of proceeding is known with the keyword "frequency wobble".

For an oscillating system, which can be regarded as a four-terminal network, the frequency-dependent ratio of the output amplitude to the excitation amplitude and/or of the output phase to the input phase of the alternating voltage can be measured and evaluated. For an oscillating system operating as a two-terminal network, a series circuit is formed from a known impedance and the impedance of the two-terminal network circuit to be evaluated, via which a variable-frequency input voltage or excitation voltage is fed. By measuring the ratio of the voltages across the entire series circuit to the voltages across the two-terminal network and, if necessary, by measuring the phase shift of these voltages, frequency-dependent characteristic values of the two-terminal network impedance can be determined and evaluated.

In a second preferred embodiment, the system capable of oscillating is excited in the form of periodic pulses or in the form of jumps. The response of the electromechanical system is measured and evaluated after the end of a time-limited, usually short, excitation. This reaction is usually a damped mechanical oscillation or an electrical oscillation measurable on an electromechanical converter. On which the parameters starting amplitude, frequency and decay time constant can be measured, from which a decision can be derived as to whether the limit state sensor is immersed in the filling or not.

Finally, in a third preferred embodiment, a noise voltage of a generally limited frequency band, the so-called rosy noise, is applied to the input side, preferably here in the form of a four-terminal network, of the oscillating system. In this case, an electrical signal is measured which contains only certain frequency components as a response to the excitation at the output. Most of the frequency components of the broadband excitation signal are filtered out by the electromechanical oscillation system. The analysis of the output signal is preferably performed in the frequency domain, e.g. via a filter bank, or more simply by a fourier transformation of the signal from the time domain to the frequency domain. This requires digitization of a time-continuous voltage signal and implementation of a Fast Fourier Transform (FFT). The output amplitude, marked on the frequency axis, enables a decision to be made as to the instantaneous transmission performance of the oscillating system and, therefore, as to the fill coverage of the oscillating element of the sensor.

Drawings

The invention is described in detail below with the aid of examples and with reference to the accompanying drawings. Shown is that:

figure 1 is a first embodiment of a circuit arrangement for implementing the method of the invention previously described in connection with frequency swing,

figure 2 is a signal diagram of the circuit arrangement of figure 1,

figure 3 shows a second embodiment of the circuit arrangement according to the invention,

figure 4 shows a third embodiment of the circuit arrangement according to the invention,

figure 5 is a graph of the signal of figure 4,

FIG. 6 shows a fourth embodiment of a circuit arrangement according to the invention, and

fig. 7 is a graph of the signals of fig. 6.

Detailed Description

Voltage U from AC source 4 at input pole_INAn electromechanical oscillatory system 1, which can be regarded as an electrical four-terminal network, is excited, the oscillatory system 1 having separate excitation and detection oscillatory elements 2, 3. The mechanical oscillation element, which projects into the container, for example as an oscillation fork, is not explicitly indicated in this figure, but its role as a coupling between the electromechanical oscillation elements 2, 3 is merely indicated in the form of arrows. The voltage U present at the output of the four-terminal network is amplified in an amplifier 5_OUT。

The frequency of the ac power supply 4, which is also referred to in this case as a voltage-controlled oscillator VCO, is continuously varied over a certain range by the control, evaluation and output unit 6. In order to be able to make a decision about the operating state of the quadriterminal 1, i.e. about the fill coverage of the limit state sensor, the amplified output voltage U must be evaluated_OUT'. Respectively to input voltage U_INThe relative relationship of (a) can be analytically processed to not only amplitude a, but also phase * of this voltage. The amplitude and/or phase difference versus wobble time (which corresponds to the frequency) curve depends in a characteristic manner on the sensor coverage. By applying an output voltage U_OUT' feed into an amplitude detector 7, e.g. rectifying and smoothing the AC voltage, and the input voltage U_INAre fed together into a phase detector 8 to derive a time-or frequency-dependent voltage U_A(t) or U_A(f) And U is_*(t) or U_*(f) In that respect By means of the comparators 9, 10, switching pulse edges can be generated at points in time/frequency which exceed or fall below certain voltage values which are characteristic of the determined amplitude and/or phase difference, which switching pulse edges can be evaluated by means of the control, evaluation and output unit 6. The result of this evaluation is frequency values, at which the amplitude curve and/or the phase curve of the quadriterminal network have extreme values or strongly vary. From these frequency values, the coverage state of the sensor can be unambiguously deduced.

Fig. 2 exemplarily shows a typical signal curve in a circuit point within fig. 1. The control voltage versus time of the VCO 4 is marked in the graph 2aVoltage curve of (2). In the case of an assumed linear relationship between control voltage and frequency, a linear sawtooth rise likewise represents the excitation voltage U_INThe frequency rises with time.

In fig. 2b the output voltage U is plotted_OUTOr U_OUT' amplitude over time t or frequency f. Since time and frequency are linearly combined according to graph a, the time axis and the frequency axis of the frequency process are proportional to each other. At the output of the detector 7 called U_A(t) or U_A(f) Has a characteristic extreme value that shifts upon immersion of the sensor. The curve drawn with a solid line here indicates the curve on a sensor which oscillates freely in air, and the curve drawn with a dashed line indicates the curve on an oscillating element which is immersed, for example, in water. Graph 2c shows the output voltage U in a similar manner_OUT' to U_INIs called U_*(t) or U_*(f) Generated by the phase detector 8 according to fig. 1. In addition, the comparator thresholds of the comparators 9, 10 are plotted in the two graphs 2b and 2c, respectively, and these comparators 9, 10 provide a switching signal instead of a strong signal change. The time position or frequency position of these switching signals results, for example, in the resonance frequencies f1 and f2 which are characteristic for oscillations in air and in the immersed state.

In a variant of the arrangement of fig. 1, the comparators 9, 10 are omitted, and instead the voltage U is applied as a dashed line_A(f) Or U_*(f) Directly to the control, analysis processing and output unit 6. This control, evaluation and output unit 6 may, for example, comprise a microcontroller with an integrated analog/digital or digital/analog converter as a main part. It can thus simultaneously generate a time-discrete control voltage value of the VCO 4 via the D/a converter and the amplitude U_a(f) Sum phase difference U_*(f) The system response in the form is converted to a digital value and stored. The thus stored data record of a single-use frequency sequence can be evaluated for extreme values and strong variations by known mathematical methods. Determining the sensor-related characteristic from the characteristic signal features foundThe end result of the immersion state of the element and outputs this information in a known manner as a emptying or filling report.

A further possibility, which is not shown in fig. 1, is obtained by the additional omission of the two detectors 7 and 8. Instead, the input and output voltage U can be digitized directly_INOr U_OUT' and also the amplitude or phase analysis process is performed digitally within the microcontroller.

A further advantageous development, which is likewise not shown in fig. 1, consists in the narrow-band filtering of the output voltage U_OUTOr U_OUT'. The bandpass filter used for this purpose tracks the frequency set by the VCO 4 and therefore passes only a narrow frequency band corresponding to the excitation frequency. Thus effectively suppressing all interfering frequencies. Either by discrete parts or in U_INAnd U_OUT ^(′)The variable-frequency bandpass filter necessary for this can be implemented digitally in the form of a program process of a micro-air controller in the case of direct digitization.

The response of the system to the excitation is determined by the comparatively high quality of the electromechanical oscillation system 1 and is comparatively delayed. This means that for an accurate determination of the transmission behavior of the frequency in the quadripole network 1, the frequency course must be carried out correspondingly slowly. However, if a fast response time of the sensor to filling or emptying is required, it is not absolutely necessary to accurately determine the four-terminal network transmission performance. Since there is generally a large frequency difference between the resonance in the freely oscillating sensor and the immersed sensor, a rapid frequency process is sufficient, in which the system cannot oscillate to the full output amplitude when excited, and the position of the extreme values can be identified without problems despite the reduced amplitude. In spite of the rapid frequency swing, the resonance frequency can be determined particularly well from the phase curve, since the phase difference changes by 180 ° rather abruptly at this point. This strong variation is also noticeable when traversing the band frequency swing more quickly than is allowed for an exact measurement. When the resonance position is found quickly but also relatively inaccurately in this way, the frequency band of the frequency swing can be reduced considerably to the surroundings of the resonance position, so that the frequency swing speed can thus be reduced again and the resolution of the transfer function becomes greater. If the resonance location suddenly disappears from the monitored frequency window, the sensor output can be immediately switched from "empty" to "filled", or vice versa, and the subsequent rapid frequency process confirms that either the resonance location in the frequency has shifted or the lack of a resonance location suggests a sensor failure that can be immediately made visible.

Fig. 3 shows a device which differs from the device described above from fig. 1 in that a single electromechanical transducer 2' is used for excitation and detection, and the oscillating system can now be regarded as a two-terminal network. In this case the determination of the impedance of the two-terminal network 1' replaces the determination of the transmission performance of the four-terminal network 1. For this purpose, the alternating voltage U generated by the VCO 4_INTo excite a series circuit of known impedances 11 with a two-terminal network 1' for which an impedance should be determined. By finding U_INTo U_OUTBy taking the amplitude ratio of the two voltages and/or the phase difference between the two voltages, the impedance behavior of the two-terminal network 1' with respect to frequency can be deduced and the operating state of the sensor can be deduced therefrom. The electronics 12' of this device can then correspond to the electronics 12 from fig. 1. In the sense of application, all the statements made for the device from fig. 1 can also be applied to this circuit variant.

Fig. 4 differs in principle from fig. 1 only by the way in which the system can be excited in oscillation. Instead of a continuous or discrete frequency swing, an oscillating system is excited by the electrical shock. This impact is achieved by a voltage jump or a short voltage pulse. For this purpose, fig. 4 shows an implementation via a dc power supply 13, a resistor 14 and a switch 15. In the case of an open switch 15, the power supply 13 charges the electromechanical converter 2 to a certain voltage value via the resistor 14 (see fig. 5 a). At time t1, the switch is closed in a controlled manner by control, evaluation and output unit 6, as a result of which the electromechanical oscillation system illustrated by quadripole network 1 is excited to natural oscillation. A ringing like that shown in fig. 5b is generated at the output of the four-terminal network. By amplification, amplitude detection (see fig. 5c) and comparison with a comparator response threshold, for example, a decay duration τ can be determined which is characteristic for the immersion or not of the sensor element (see fig. 5 d). Additionally or alternatively, the frequency and maximum amplitude of the damped natural oscillation can be evaluated. The methods used herein are well known to the skilled person and are not described in detail herein. Technically, the electrical "tapping" of a mechanically oscillating system means a broadband frequency excitation to which the system reacts by filtering.

The method of which an embodiment is shown in fig. 6 is also based on this principle. Unlike fig. 4, a broadband frequency excitation is achieved by noise source 16. As shown in fig. 7a, this noise source 16 uniformly generates noise across the frequency band of interest to which the electromechanical oscillatory system may respond. Only a small frequency component of the original excitation spectrum is still present at the output of the four-terminal network 1. By determining the frequency of this component, which is shown in fig. 7b as the voltage plotted against the frequency axis, the limit state detection is effected by means of this sensor, as already explained above. The frequency analysis of the output signal can be carried out in the control, analysis and output unit 6', for example by a narrow-band adjustable bandpass filter or, preferably, by transforming the time signal into the frequency domain by means of a fast fourier transform FFT.

Claims (7)

1. Method for determining and/or monitoring a predetermined filling state of a filling material in a container having a vibrating rod probe directed into the inner space of the container, which vibrating rod probe is part of an electromechanical oscillation system,

exciting the oscillatory system at different frequencies during the step of determining and/or monitoring a predetermined filling state of the filling, and measuring the reaction of the system at all frequencies,

after the oscillating system has been excited, one or more characteristic parameters are measured and this or these parameters are taken into account when evaluating the measurements.

2. Method according to claim 1, characterized in that the electrical transmission properties of the oscillating system, which can be regarded as a four-terminal network, or the impedance of the oscillating element, which operates as a two-terminal network, are determined and evaluated as characteristic parameters.

3. A method according to claim 2, characterized in that said transmission properties or impedances have characteristic features which permit a clear conclusion of said filling state by determining and evaluating said features.

4. A method according to claim 1, characterized in that the excitation is carried out with different frequencies, wherein a certain number of discrete frequency points are adjusted one after the other and the frequency process is repeated periodically.

5. A method according to claim 1, characterized by continuously passing a given frequency band in a frequency-oscillating manner to excite said oscillating system.

6. A method according to claim 1, characterized in that the oscillating system is regarded as a four-terminal network and the frequency-dependent ratio of the output amplitude to the excitation amplitude and/or of the output phase to the input phase of the alternating voltage is measured and evaluated.

7. Method according to claim 1, characterized in that the oscillating system is operated as a two-terminal network and a series circuit is formed from a known impedance and the impedance of the two-terminal network circuit to be evaluated, through which a frequency-variable input voltage or excitation voltage is fed, wherein the frequency-dependent characteristic values of the two-terminal network impedance are determined and evaluated by measuring the ratio of the voltage across the entire series circuit to the voltage across the two-terminal network and, if necessary, by measuring the phase shift of these voltages.