Method and apparatus for detecting impact and vibration conditions in magnetostrictive transducers

Abstract

In the method described herein and in a device for operating an electromagnetic, in particular magnetostrictive, sensor or transducer (500), in particular a position measurement system (500, 505), in particular by evaluating signal variables (120, 125) in the region of a baseline or null line (105) of a measurement signal provided by the sensor / transducer (500), the signal variables (120, 125) are evaluated in the region of the baseline or null line (105) of the measurement signal provided by the sensor / transducer (500). A mechanical shock and / or a vibration state, for example occurring during operation of the sensor / transducer (500) or a vibration of the sensor / transducer (500) caused thereby, is detected.

Description

Technical Field

[0001] The present invention generally relates to the operation of magnetostrictive or magnetostrictively acting sensors or transducers referred to herein, and in particular to methods and apparatus for detecting shock and / or vibration conditions in magnetostrictive sensors or transducers, which methods and apparatus can be used, for example, in magnetostrictive position or speed measurement systems in the field of technical diagnostics. Background Art

[0002] Magnetostrictive sensors or transducers can be used in position measurement systems developed and distributed by the applicant in order to determine the position of an object along the sensitive axis of a corresponding sensor device. This position measurement system, which mainly operates inductively (e.g., product number BTL7-V50D-M0250-B-C003 of Balluff GmbH), is designed as a serial sensor device with a planar coil system manufactured using printed circuit board technology. The coil system has an excitation coil and a plurality of structured planar receiver coils. In order to determine the position of an object relative to the sensor device, the vertical component of the object's magnetic field acting on the coil plane is utilized and the magnetostrictive transducer is used to measure and record the vertical component. The amplitude and phase of the induced voltage of each of the receiver coils depend on the position of the object. The measuring range extends to approximately 130 millimeters (mm).

[0003] Furthermore, DE 10 2018 129 942 A1 (Balluff GmbH) provides a vibration sensor in such a sensor device to characterize the operating conditions of the sensor or transducer involved in this context. The control of the transducer can use the vibration information arising thereby in order to determine the reliability of the recorded sensor data.

[0004] Furthermore, DE 10 2017 116 828 A1 (Sick AG) proposes a photodetector for detecting objects in a monitoring area, the photodetector comprising a sensor module having a light transmitter for emitting a light signal into the monitoring area, a light receiver for receiving a light signal from the monitoring area and generating a corresponding received signal. Furthermore, a sensor evaluation unit for evaluating the received signal and generating process data (so-called "object determination signal") and for generating sensor module data is provided. A process data channel is used to output the process data. Furthermore, a condition monitoring module is provided, which has:

[0005] - a situation assessment unit for generating situation data,

[0006] - a first internal interface between the sensor module and the condition monitoring module for transmitting sensor module data to the condition monitoring module, and

[0007] - a ring buffer, which receives data consisting of condition data and / or sensor module data from the condition monitoring module via the second interface and temporarily stores the data.

[0008] The condition monitoring module is designed to output a first trigger signal through the second interface based on the condition data and / or the sensor module data, wherein the first trigger signal prevents further overwriting of data in the ring buffer for a defined period of time, and wherein the data can be output from the ring buffer through the condition data channel. The condition monitoring module is used to detect abnormal behavior of the sensor data to warn that the reliability of the sensor system is reduced. Summary of the invention

[0009] The object of the present invention is to specify a method and a device for accurately and reliably detecting shock and vibration conditions in a magnetostrictive or magnetostrictively acting transducer as described herein, without requiring expensive and technically complex modifications or changes to existing magnetostrictive transducers and / or to said sensor devices having such transducers.

[0010] The idea of ​​the invention is to detect the vibrations caused by mechanical impacts or the vibrations caused in the recorded, preferably amplified, measured values ​​or sensor signals of such a sensor device.

[0011] Since the well-known time measurement based on zero crossings is relatively inaccurate in such sensor devices and leads to measurement errors in the range of several millimeters in the position measurement, the proposed method in particular monitors the background noise or the corresponding baseline / zero line (i.e. "baseline") of the sensor signal in order to reliably detect the disturbance. The sensor signal is preferably digitized so that this monitoring can then be added to an existing sensor device in the form of a control program or control software.

[0012] The invention is based in particular on the finding that, in the case of mechanical undisturbance, a digitized signal curve or data points of a corresponding signal waveform calculated from a measurement signal supplied by a measuring sensor as described herein has a characteristic curve in the region of the baseline in the sensor device as described herein. As long as there are no such disturbances due to shocks and / or vibrations, the measurement data, i.e. the data points of the signal waveform, accumulate in the vicinity of the zero line. The data points present in a window centered around zero can therefore be statistically evaluated and correspondingly monitored.

[0013] It is worth mentioning here that in the methods and apparatus described herein, there may be one or more targets that cause the target signal (110) referred to herein. Therefore, the corresponding target signal (110) may appear at any location, so to speak, depending on the location of the underlying target (e.g., target magnet) along the waveguide.

[0014] According to the method of the invention, measurement data that deviate from the characteristic course of the previously described undisturbed signal waveform can be characterized as "unreliable", skipped during the evaluation of the measurement data or adopted with lower weight or even deleted during further processing. This exclusion, suppression or deletion can be performed based on whether the deviation of the indicator variable of the currently measured signal waveform from the corresponding indicator variable of the previously stored undisturbed signal waveform exceeds an empirically predeterminable measure, or in the case of a scalar indicator variable, exceeds the value of an empirically predeterminable threshold.

[0015] It should be noted that the evaluation of the measurement data is not necessarily skipped if the shock event is detected. This means that a shock / vibration situation can only be identified when the entire signal curve is already available and it is therefore too late to start evaluating the detected signal or signal curve to calculate the position of the target magnet. Therefore, the corresponding waveforms recorded in the case of a shock situation do not necessarily have to be deleted or completely excluded from the position evaluation.

[0016] It should also be noted that when generating position data by the sensor device or corresponding evaluation / control system referred to herein, unreliable measurement data may be skipped. However, the present invention is primarily concerned with the generation of said indicator value for identifying shock / vibration conditions and its use in the operation of the sensor device or corresponding position measurement system referred to herein.

[0017] In a method according to the invention for operating an electromagnetic, in particular magnetostrictive sensor or transducer, in particular a position measuring system, it is particularly provided that mechanical shocks and / or vibration conditions occurring during operation of the sensor / transducer or vibrations of the sensor / transducer caused thereby are detected by statistically evaluating signal fluctuations in the region of a baseline or zero line of a measurement signal provided by the sensor / transducer.

[0018] In the method according to the invention, when analyzing signal fluctuations in the region of the baseline / zero line, a comparison with a characteristic curve of the measurement signal of the undisturbed situation is preferably performed.

[0019] In the method according to the invention, it can further be provided that the measurement signal is digitized and the resulting data points are evaluated in a zero-centered, time-limited portion or excerpt of the measurement signal. However, it should be noted that preferably such portions or excerpts are not analyzed, but rather the complete waveform is analyzed.

[0020] In the method according to the invention, provision can further be made for the signal waveform of an undisturbed electromagnetic effect, in particular of a magnetostrictive material, to be used as the characteristic curve of the measurement signal in the undisturbed case.

[0021] In the method according to the invention, provision can further be made for measurement signals or measurement data which deviate from the characteristic curve in the undisturbed situation to be excluded from further processing or to be used with lower weighting or to be deleted.

[0022] It can further be provided that if a deviation of a currently measured measurement signal or corresponding measurement data exceeds an empirically predeterminable threshold value, the measurement signal or measurement data is excluded from further processing or is used with lower weighting or is deleted.

[0023] In the method according to the invention, it can further be provided that a deviation of a currently measured measurement signal or of corresponding measurement data can be quantified because a relevant increased value of at least one determined indicator variable is detected because the at least one indicator variable exceeds the value of an empirically predeterminable threshold value predetermined in the case of the undisturbed measurement.

[0024] In the method according to the invention, provision can further be made that the evaluation of signal fluctuations in the range of the baseline or zero line of the measurement signal is performed by statistical evaluation of data points generated from the measurement signal.

[0025] It may further be provided that, based on the waveform data generated from the measurement signal, a statistical evaluation is performed using one or more of the following mathematical evaluation methods:

[0026] - the “quartile method”, in which the difference between the first and last quartile of the waveform data (e.g., the quartile) is evaluated, said difference increasing under the influence of a shock;

[0027] - "Histogram method", which is based on the fact that the sides of the histogram of waveform data close to zero increase under the influence of an impact while the central part decreases, i.e. according to the ratio of the sides to the center, and is therefore evaluated according to the ratio of the sides to the center;

[0028] - "Standard Deviation Method", which is based on the fact that the standard deviation of waveform data close to zero increases under the influence of a shock, so waveform data close to the baseline / zero line is analyzed;

[0029] - A "data density method" in which the data density of the waveform data near the baseline / zero line is analyzed.

[0030] In the method according to the invention, provision can further be made that, in the case of the histogram method, the number of data points is evaluated as a function of the value deviation from the zero value.

[0031] In a method according to the invention, in particular a magnetostrictive sensor or transducer for operating a position measuring system with a magnetostrictive or magnetostrictive active or acting waveguide, wherein a mechanical rotational or torsional wave packet is generated in the waveguide by interrogation current pulses propagating in two directions along the waveguide, wherein a measuring sensor, i.e. a transducer or a sensor, arranged near one end of the waveguide is provided, the measuring sensor being for receiving the measurement signal for converting the rotational / torsional wave packet travelling in the waveguide into an electrical signal, wherein during a first time period of the measurement signal, the interrogation current pulse induces an oscillating interrogation noise in the received measurement signal in the region of the baseline / zero line and the interrogation noise caused by the interrogation current pulse is not generated. The invention relates to a method for measuring a collective mechanical response of the waveguide, wherein, in a second time interval of the measurement signal, an oscillating background noise and a target signal superimposed on the background noise are present on the interrogation noise in the region of the baseline / zero line, wherein the target signal is followed by another oscillating background noise. It can also be provided that, for at least one time range of the oscillating noise in the region of the baseline / zero line, the measured values ​​of the recorded measurement signal in the at least one time range are compared with corresponding measured values ​​of a sensor / transducer disturbed by a mechanical shock and / or vibration condition and an undisturbed sensor / transducer, and depending on the result of the comparison, a disturbed or undisturbed condition of the sensor / transducer is inferred.

[0032] In the method according to the invention, provision can further be made for the statistical evaluation to be limited to a time period of the measurement signal which is not influenced by the interrogation noise or the superimposed target signal.

[0033] In the method according to the invention, provision can finally also be made for the measured values ​​recorded in at least one time range of the sensor / transducer which is disturbed by mechanical shocks and / or vibration conditions to be disturbed by low-frequency noise.

[0034] It is also proposed that the device according to the invention for operating an electromagnetic, in particular magnetostrictive or magnetostrictive acting sensor or transducer, in particular a position measuring system, has in particular an evaluation unit, which is used to detect mechanical shocks and / or vibration conditions occurring during operation of the sensor / transducer or vibrations of the sensor / transducer caused thereby by evaluating signal fluctuations in the region of a baseline or zero line of a measurement signal provided by the sensor / transducer.

[0035] The device according to the present invention may also have: at least one amplifier, which is used to amplify the measurement signal supplied by the sensor / transducer; a comparator, which is used to compare at least one time interval of the amplified measurement signal in the area of ​​the baseline / zero line with the characteristic curve of the measurement signal of the signal waveform in the undisturbed situation; a flight time analysis unit, which supplies a position result; an analog-to-digital converter, which is used to convert the measurement signal into corresponding digital measurement data; and a statistical unit, which is used to perform a statistical evaluation according to the method according to the present invention in order to determine or supply an impact indicator variable.

[0036] The device according to the invention may also have: at least one amplifier, which is used to amplify the measurement signal supplied by the sensor / converter; at least one analog-to-digital converter, which is used to convert the measurement signal into corresponding digital measurement data; a digital signal processor, which is used to evaluate the measurement data; a time-of-flight analysis unit, which supplies position results; and a statistical unit, which is used to perform a statistical evaluation according to the method according to the invention in order to determine or supply an impact indicator variable.

[0037] The measurement data from the comparator preferably come from a magnetostrictive "time of flight" position sensor. Such sensors are subject to a high degree of measurement inaccuracy due to mechanical shocks and / or vibrations. In the case of relatively large disturbances, a measurement may even be completely impossible due to missing zero crossings of a data point or due to invalidation of the prerequisites for the aforementioned time of flight measurement. This is because applying the cross-correlation technique to the digitized waveform and determining the time of flight from the cross-correlated waveform results in reduced influences, i.e. smaller deviations in the determined position of the target object in the presence of vibrations or shocks.

[0038] Vibrations and / or shocks that affect the accuracy or reliability of position measurements in the sensor devices described herein can be reliably detected or determined by means of the method and device according to the invention. This means that such sensor devices can also be reliably used in machine-critical applications.

[0039] Thus, in the case of the position / speed measurements involved herein, the probability of erroneous control interventions or reactions of the control system based on the corresponding detected position information can be effectively reduced or even completely avoided. The method and the device also enable a rapid pulse-by-pulse display of the influence of shocks and vibrations on, for example, the underlying magnetostrictive position measurement system. Furthermore, it can be provided that, for example, the corresponding magnetostrictive sensor can accurately report the exact vibration conditions during its operation, for example in the form of a correspondingly provided condition monitoring function.

[0040] When the method or the device according to the invention is used in a conventional, i.e. comparator-based position measuring system with a magnetostrictive position sensor, only an analog-to-digital conversion system or A / D converter with a converter having a low or medium operating or clock speed is required. Such A / D converters are usually already available in common microcontrollers, wherein the implementation of the method according to the invention described in detail below results in a relatively low computational load. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Exemplary embodiments of the invention are depicted in the drawings and are explained in more detail in the following description.

[0042] In the attached picture:

[0043] Figure 1a to Figure 1d Not disturbed by mechanical shock pulses ( Figure 1a , Figure 1c ) and the case of mechanical shock pulse interference ( Figure 1b , Figure 1d ) of a magnetostrictive position measurement system;

[0044] Figure 2a to Figure 2d For evaluation using the "quartile" method according to the present invention Figure 1a and Figure 1b Exemplary waveform data obtained from the measurement curve shown in;

[0045] Figure 3a to Figure 3d Various impact indicator variables according to the evaluation according to the present invention depicted, for example, using a histogram;

[0046] Figure 4a , Figure 4b Examples of designating a data point as "unreliable" or not as "unreliable" due to detection or non-detection of a shock effect that may be implemented in accordance with the present invention; and

[0047] Figure 5a , Figure 5b Five exemplary embodiments of the device according to the invention for implementing the method according to the invention by means of a combined flow chart / block diagram, wherein Figure 5a and Figure 5b In the drawings, the same or similar parts have corresponding reference numerals. DETAILED DESCRIPTION

[0048] Magnetostrictive position sensor systems typically include:

[0049] - a waveguide made of a strong magnetostrictive material;

[0050] - an electronic circuit for applying, upon an initialization signal, an interrogation current pulse to the waveguide itself or to a conductive wire passing through a longitudinal hole of the tubular waveguide, in order to generate a temporary magnetization of the waveguide;

[0051] - a target magnet arranged in the vicinity of the waveguide and generating an initial magnetization of the waveguide, wherein the initial magnetization is overlaid with the magnetization induced by the interrogation current pulse, thereby generating a torsional wave packet propagating in two directions along the waveguide;

[0052] - a measuring sensor arranged near the proximal end of the waveguide, i.e. near the electronics of the waveguide, for converting the torsional waves of the waveguide into electrical signals;

[0053] - Most importantly, an attenuation zone / structure located at the other end of the waveguide to prevent (or attenuate) reflections of the torsional waves at this end of the waveguide; and

[0054] - An electronics block for evaluating the electronic signals from the measuring sensors.

[0055] It should be noted that in addition to applying the interrogation current pulses using a conductive or current-carrying wire guided through a longitudinal hole in the tubular waveguide, it is also possible to pass the excitation current through the (solid) magnetostrictive wire itself and use copper wire outside the magnetostrictive wire to conduct the current back to the excitation circuit.

[0056] Due to the technical design of this sensor system, the torsional wave packets and therefore the corresponding electronic signal pattern have a typical shape. The time interval measured between the triggering of the interrogation current pulse and the time of detecting the corresponding electronic pattern is used to characterize the position of the target magnet by means of the correlation between distance and time. The wave propagation speed is used as a proportionality factor between these two time points. The position detection task is thus transformed into a time measurement task.

[0057] In magnetostrictive and also ultrasonic-based position or distance sensors, the known measuring technique itself consists in starting a counter and an active edge of an "Init" signal, wherein a comparator is provided to detect a horizontal crossing of the electronic output signal of the sensor. When the comparator output is activated, the time counter is stopped. The comparator output can be preset or preprocessed by further comparators and sequential logic in order to prevent the counter from stopping due to non-target-related fluctuations, since the microelectronically integrable and highly accurate time counters mostly used here cannot react to a false stop without losing valid data.

[0058] Due to the coupling to the magnetostrictive waveguide, the electronic signal of the measuring sensor (or "pickup") or the sensor signal generated or supplied by the pickup is exposed to unwanted mechanical disturbances such as shocks and vibrations propagating along the waveguide to the pickup area.

[0059] Furthermore, position measurement systems based on the evaluation of the longitudinal magnetic effects of torsional waves have also become known, which are more sensitive to such disturbances than systems using only rotational waves. This is because in systems using only rotational waves, the aforementioned disturbances can be coupled into the waveguide from the mechanical environment with much lower efficiency.

[0060] Within the scope of the solution proposed herein, it was found that the cross-correlation technique is particularly effective for suppressing the influence of vibrations and shocks on the accuracy of magnetostrictive position measurements. Since the target-related signal components in the position measurement system involved herein represent relatively narrow-band wave packets in terms of frequency behavior, the cross-correlation generates an effective filtering of the corresponding measured waveform. This is because the frequency components around the center frequency of the waveform (approximately 100 kHz) can be retained when determining the position, and only the low-frequency components, i.e. a few kHz (e.g., up to 4 kHz), which are typical characteristics of vibrations and shocks, can be suppressed or excluded.

[0061] In experimental investigations, it was found that the undesired deviations of the position data due to external vibration influences in the case of digital determination of the zero crossing time of the cross-correlation waveform are three to ten times smaller than the deviations caused by the determination of the zero crossing time of the analog scanning signal based on an analog comparator, but cannot be completely eliminated. This is because, in addition to the coupling of mechanical interference, electronic interference also causes oscillations in the scanning signal, which also affect the reliability of the position estimation.

[0062] The starting point of the solution described herein is to only include or provide information relevant to the reliability of position measurement data generated by a magnetostrictive position sensor under external shock and vibration influences.

[0063] The basis for this information is the structure of the scan signal. Figure 1a Depicted is the structure of the measured signal without shock influences. In the figure, the recorded signal values ​​are plotted in [LSB] versus time in [μs]. In the bottom layer measurement, a radially arranged target magnet was used and the waveform was recorded with a scanning frequency of 3.125 MHz and a nominal resolution of 12 bits. Figure 1a The measurement signal shown in FIG. 1 starts with an oscillatory noise 100 caused by an interrogation pulse of approximately 4 μs long and the collective mechanical response of the waveguide that accompanies the interrogation pulse. This disturbance 100 is essentially constant, independent of other influences, and decays after approximately 20 μs. This interrogation noise is followed by a background noise 105 with a near zero point or baseline and a portion of the target signal 110 superimposed on the background noise 105, which portion appears as a short wavelet between approximately 105 μs and 135 μs. The target signal 110 is again followed by the background noise 115.

[0064] It should be noted that in addition to the collective mechanical response of the waveguide associated with the interrogation pulse, electronic crosstalk from the excitation to the measurement sensor is also possible. The exact structure of this interfering signal is not relevant in the present context.

[0065] It should also be noted that in the methods and apparatus described herein, there may be one or more targets that cause the target signal (110) referred to herein. Thus, the corresponding target signal (110) may appear anywhere, depending on the location of the target. Therefore, Figure 1a The shown course of the measurement signal is only one example and comprises every occurrence of such an undisturbed measurement signal which, in the case of a magnetostrictive position detection system corresponding to a target magnet, differs greatly from the actual position of the target.

[0066] By recording the signal progression of the undisturbed waveguide ( Figure 1a ) and the signal process in which the sensor is impacted ( Figure 1b ) can clearly see the impact detection concept described in this article.

[0067] Figure 1b Again, it shows Figure 1a The complete waveform shown in , i.e. including the interrogation noise and the target signal, however, the sensor is disturbed by the underlying low frequency noise 120, 125. In this case, a half-sine shock pulse with a duration of 2 ms and a maximum value of 150 g perpendicular to the waveguide is generated using an electrodynamic vibrator. In the case of the undisturbed waveform ( Figure 1a ), data points 105, 115 are arranged in the baseline region close to zero value, while the data points of waveforms 120, 125 affected by the shock are significantly deviated from zero value ( Figure 1b ).

[0068] This difference also Figure 1c and Figure 1d As shown in Figure 1a and Figure 1b The waveform data 130 and 135 are shown in corresponding histograms. In these histograms, the frequency of the number of data points (Y-axis) is plotted against the value of the corresponding signal deviation from the value zero (X-axis). In this example, the difference between the two distributions of the waveform data 130 and 135 within this limited data range around the zero value is used to distinguish whether the corresponding underlying measurement was made under the influence of a mechanical shock. Therefore, this distinction is based on the statistical characteristics of the shock effect.

[0069] Although it is possible through Figure 1c and Figure 1dThe distributions 130 and 135 shown in FIG easily distinguish between undisturbed data and data affected by the shock, but the evaluation is preferably limited to those parts of the signal form that are not affected by interrogation noise or superimposed target signals. Figure 1a and Figure 1b Limiting the horizontal axis to the amount of time required for the torsional wave to propagate through the magnetostrictive waveguide only once effectively prevents reflections (or echoes) of the torsional oscillations from the ends of the waveguide from causing further induced signals in the pickup, which would result in Figure 1c Thus, in the exemplary embodiments described herein, the aforementioned statistical evaluation of the complete waveform over the run time is applied to the entire length of the waveguide.

[0070] In the following, exemplary embodiments of the method or apparatus according to the present invention are described based on various impact detection methods using examples of “quartile method”, “histogram method”, and “standard deviation method”.

[0071] It should be noted that other corresponding or similar detection methods may also be used. However, the four methods mentioned for real-time characterization of the reliability of position measurement data are particularly suitable for effectively preventing erroneous control reactions of the control system of the measurement system (eg position detection system) involved in this article.

[0072] In order to characterize the actual impact of the mechanical disturbances involved in this paper, the following four evaluation strategies are proposed.

[0073] Example 1: "Quartile Method"

[0074] In this method, the waveform data elements are sorted in ascending order. Figure 1a and Figure 1b The data elements of the waveform depicted in Figure 2a and Figure 2b 2 , 215 or 220 , 235 of the underlying single data of the detected signal are depicted in LSB (LSB=“Least Significant Bit” or “Least Valued Byte”) units as a function of time. Figure 2c and Figure 2d A view of the same data magnified vertically 100 times (at least the data labeled 200, 215) is shown.

[0075] This evaluation method is based on the finding that the waveform in a disturbance-free measured signal contains a large number of data points near the value zero. Therefore, the statistical number of data points that lie in a narrow range or window of values ​​around zero must be related to the strength of the disturbance in question.

[0076] In this evaluation method, the data points, preferably only every nth data point (e.g. 100 data points), are sorted and the difference between the third quartile and the first quartile is calculated. For an undisturbed signal, this difference is small, but increases if the signal baseline is disturbed. The use of the quartiles is only preferred, but has the advantages described herein for waveguides of different lengths and different numbers of targets (i.e. magnetostrictive measurement systems for different numbers of target magnets).

[0077] It should also be noted that, for example Figure 1a and Figure 1b The data elements of the waveform shown in are sorted in ascending order, and the waveform values ​​sorted in this manner are simply plotted continuously so that the horizontal axis corresponds to the sequence number of the sorted waveform data points. The vertical axis is specified in the aforementioned LSB unit. The representation of the waveform data in LSB units also corresponds to the resolution of the corresponding analog-to-digital converter (ADC).

[0078] because Figure 1a Most of the data points in the time series of the undisturbed detection signal waveform shown in FIG are arranged at positions close to zero, so the gradient of the middle part of the sorted data series is very small, and the sorted data series ( Figure 2c ) are caused by the aforementioned interrogation noise and target signal. Figure 2d ) have increased deviations due to mechanical disturbances. In addition, the gradient in the middle of the sorted sequence is larger due to the wider distribution of the data.

[0079] The gradient of the middle portion is advantageously characterized by the difference between the first quartile 205 or 225 and the last quartile 210 or 230 of the sorted sequence. Thus, the undisturbed signal is characterized by a relatively small value difference, and the signal affected by the shock is characterized by a relatively large value difference. Figure 2a to Figure 2d The difference between the two quartiles marked by points 205, 210 and 225, 230 in Figure 2c The value in is 16 LSB and Figure 2d The value in is 600 LSB.

[0080] For the "quartile method", the waveform data points can only be sorted when they are available. However, using a field programmable gate array (FPGA), a real-time sorting algorithm can be implemented for a known total data length. The sorting algorithm corresponds to a structure of one LSB histogram per bin, which consumes memory area according to the ADC resolution (e.g., 4096 bins for a 12-bit ADC resolution).

[0081] The bin contents can be accumulated from the minimum corresponding ADC value, and if, for example, more than one quarter of the total number of data points is exceeded, the first "quartile" is found. In statistics, "quartiles" are a type of quartile that divides the current number of data points into four roughly equal-sized parts or quarters. The data is sorted from smallest to largest value to calculate such quartiles, i.e., quartiles are a form of ordinal statistics.

[0082] A similar method can be performed to find the third quartile. In the aforementioned FPGA, both the accumulation and determination of the quartiles by digital comparison can be run in parallel in real time. The implementation of the histogram-based quartile method can also be implemented using a microcontroller, which allows very short delays.

[0083] Example 2: "Histogram Method"

[0084] This method is based on Figure 1c and Figure 1d The knowledge obtained from the two histograms 130 and 135 shown in FIG. 1 is that as a feature of the undisturbed signal (hereinafter referred to as a “characteristic curve”), the portion of the distribution 105 and 115 near the zero line has relatively more data points than the peripheral portion ( Figure 1c ), while in the case of the shock-affected signal, the data points 120, 125 appearing near the zero line are more evenly distributed ( Figure 1d ). Therefore, in this example, the peripheral to average sum ratio ( Figure 1c ) is 0.250, while the case affected by the shock ( Figure 1d ) is 1.471.

[0085] Using this evaluation method, the data points in the lower and upper tertiles of a narrow zero-centered window are counted and summed, and the resulting sum is then divided by the number of data points in the middle tertile. Typically, a narrower distribution of data points from an undisturbed signal results in a smaller summation ratio, while a correspondingly wider distribution of data points from a disturbed signal results in a larger summation ratio.

[0086] Alternatively, the difference of the total sums within the two distributions may also be used. However, for relatively small mechanical disturbances, the total sum is less sensitive than the aforementioned ratio of the peripheral sum to the average sum.

[0087] The "histogram method" requires the use of a digital window comparator to construct a three-bin histogram and, after the required amount of data has been processed, the use of a logical addition component and a logical division component to generate the corresponding shock indicator parameters.

[0088] Example 3: "Standard Deviation Method" (or "Sigma" Method)

[0089] The distribution of data points around the zero line can also be characterized by its standard deviation. The standard deviation of the data points of the signal waveform involved in this article is calculated within a narrow window centered on zero. Generally, a narrower distribution of data points for an undisturbed signal results in a smaller sigma value, while a disturbed signal results in a wider distribution of data points, which corresponds to a larger sigma value.

[0090] Therefore, according to Figure 2c and Figure 2d , the standard deviation of the data points in the range of LSB values ​​between -36 and +36 for the undisturbed case ( Figure 2c ) is about 9.5 and for the disturbed case ( Figure 2d ) is approximately 19.9. It should be noted that without the above-mentioned restriction of the LSB value to between -36 and +36, the standard deviation of the data points of the entire waveform would be much less sensitive to the above-mentioned vibration and shock effects due to the larger values ​​of the interrogation noise and target signal portions.

[0091] The "Standard Deviation Method" requires some logical multiplications and accumulation of some values ​​until the required amount of data has been processed. Since the multiplication and accumulation operations can be performed within a single clock on most relevant microcontrollers, this method can be performed as a real-time method on microcontrollers as well as the aforementioned FPGA architectures.

[0092] Example 4: "Density Method"

[0093] The density of data points near zero is calculated by counting the data points of the signal waveform affected herein in an even narrower window near zero. The value is highest for a signal without interference and decreases as the data points move away from near zero due to ripples in the waveform.

[0094] Compared to the four methods described, other methods can also be considered, in which the spectrum of the measurement signal waveform is calculated and the low-frequency part of the spectrum is quantitatively characterized, because in this context the impact of shock and vibration on the measurement signal usually occurs below 5 kHz. However, the calculation of the spectrum limited to the low-frequency range (including by fast Fourier transform (FFT) or discrete Fourier transform (DFT)) is relatively computationally intensive and can only be started after the entire waveform has been detected. Therefore, such methods are not able to characterize the reliability of the actual (i.e. currently available) position measurement data in real time.

[0095] In the following, refer to Figure 3a to Figure 3d as well as Figure 4a and Figure 4b , describes the measurement results actually obtained from the above-mentioned position measurement system.

[0096] The shock pulses described above were repeatedly applied to a 76 cm long magnetostrictive sensor, while the interrogation pulses were output at a repetition rate of 500 Hz and the signal waveforms were recorded at a sampling frequency of 3.125 MHz. For each recorded waveform, the three methods described above were applied to generate different index amplitudes for the corresponding shock behavior. This method lasted for about 40 seconds, during which about 20,000 waveforms were recorded. Figure 3a to Figure 3c In the above, the resulting values ​​of the indicator variables "D_quantile", "D_histogram" and "D_stdev" described above are each plotted in arbitrary [au] units as a function of the time t in [s]. In addition, the deviation "D_posdev" is specified in Figure 3d is plotted against time in [s].

[0097] from Figure 3a and Figure 3d As can be seen in Figure 1, within the time range in which the shock pulse is present, the value of the described indicator variable has an increased value above the corresponding baseline, wherein in the corresponding undisturbed case the baseline has a typical course or typical, i.e. a predetermined value level. Furthermore, the shock pulse has a random effect on both the indicator value and the position deviation. While the indicator value for the disturbed case is always higher than the value for the undisturbed case, the position deviation is often even the opposite, i.e. negative, but in most cases positive.

[0098] It should also be noted that there does not appear to be a correlation between any two of the described index values, nor between the index values ​​and the positional deviations.

[0099] The position of the target magnet can be calculated based on the signal behavior recorded in this way. First, the cross-correlation between the reference waveform and the detection waveform is calculated, where the reference waveform is obtained by Figure 1a to Figure 1d as well as Figure 2a to Figure 2d The target position is defined (without reference to the inevitable constant offset value) as the position of the first negative zero crossing of the cross-correlation waveform and is calculated by linear regression of the data on this zero crossing. Figure 3d The corresponding deviations of the position values ​​calculated for the undisturbed measurement from the master position are plotted.

[0100] It should be noted here that the duration of the shock effect is much longer than the 2ms shock pulse itself. Furthermore, the slowdown in the deviation of the position data from its undisturbed value occurs later than the drop in the shock indicator variable, which is essentially attributable to the true target vibration despite the damping of the vibration of the sensor waveguide itself.

[0101] When at least one of the shock indicator variables exceeds a value of an empirically predeterminable threshold value previously determined during a measurement without vibration, the impact of shock-like vibrations on the position measurement can be quantified by identifying a corresponding increased value of the aforementioned indicator parameter of this measurement. Alternatively or additionally, several such variables obtained under unknown vibration conditions can be analyzed.

[0102] Figure 4a and Figure 4b Depicted are examples of marking or flagging or not marking or not marking data points as "unreliable" due to detected shock effects, which can be automatically performed using the described methods.

[0103] Figure 4a The shock indicator values ​​for a time span (0.8 s in this example) of the position measurement according to the "standard deviation method" are shown. In the diagram shown, the indicator variables "D_stdev" 405, 410 are likewise plotted in arbitrary [au] units as a function of time t in [s]. Figure 4a The circles 415, 420 shown in show those data points that were classified as unreliable using the following method and labeled accordingly:

[0104] When the shock indicator value exceeds the predefined threshold value of 400 (in Figure 4a At the point in time when the recorded impact index value drops below the predefined level 400 again, the corresponding position output data is marked or labeled as unreliable. These markings are maintained for an empirically predeterminable period of time. During this period, other measurement points or corresponding data points continue to be marked or labeled, namely in the exemplary embodiment ten (10) other measurement points. Figure 4a In the table, data not marked in the above way are shown as small dots.

[0105] exist Figure 4b The corresponding position data 425, 435 are depicted. Figure 4a The value of the threshold value of the mark / marker shown in 400 and the amount of output data for which the mark / marker is maintained can be determined empirically according to the desired dynamics of the entire system. Although relatively low threshold values ​​and relatively long durations of the mark / marker result in more reliable output data and therefore a more stable position measurement system, such values ​​also reduce the possible dynamics of the control system.

[0106] It should be noted that Figure 4a The shock-induced zero-point fluctuation shown in Figure 1 lasts longer than the underlying shock event itself. Figure 4b It can be seen that from Figure 4a The shock-induced fluctuations 430, 440 of the measured position data shown in FIG. Figure 4aThe recorded fluctuations of the indicator variable "D_stdev" shown in are of long duration. It should also be noted that for data points marked as unreliable, the data points of the corresponding signal waveform derived from the measurement data may also be adjusted, for example, they may be deleted completely.

[0107] It should be noted that the measurement data recorded by the sensor itself is not marked as "unreliable" or even deleted completely. Only the digital signal waveform derived from the measurement data or the position data obtained from the evaluation can be marked as "unreliable". Unreliable position data can also be suppressed by using adaptive filtering (e.g. using a Kalman filter) or by exclusion through appropriate weighting.

[0108] Figure 5a and Figure 5b An exemplary embodiment of an apparatus for implementing the method described above is depicted using a combined flow chart / block diagram.

[0109] The method and device for generating a shock indicator value are based on the digitization of the corresponding sensor signal.However, the device for detecting the target position and for generating the reliability information of the recorded position data described above in parallel, in particular for generating the shock indicator value described above, can be realized in different ways.

[0110] Figure 5a Two exemplary embodiments of the aforementioned device are shown, with and without additional dashed line components.

[0111] According to a first exemplary embodiment of the device, the signals to be analyzed are supplied by a measuring sensor ("pickup") 500, which is arranged near the proximal end of the magnetostrictive waveguide 505, i.e. near the electronics of the waveguide. Via a first line 510, these signals are fed to a first amplifier 515 in order to raise the signal level to a variable that can be further processed. The signals amplified in this way are fed via a second line 520 to a comparator (in this case with an integrated TDC) 525, in which the comparison described above is performed. The data resulting from this processing step are then fed to a time-of-flight analysis 530, which provides a position result 535.

[0112] It should be noted herein that the evaluation of the waveform data derived from the measurement data does not have to be terminated or interrupted if a shock event as referred to herein is detected. A shock / vibration condition can only be detected when the entire signal curve is already available, and it is therefore too late to start evaluating the signal / signal curve to calculate the position of the target magnet. Therefore, the corresponding waveform recorded under a shock condition does not have to be deleted or completely excluded from the position evaluation.

[0113] By the way, this also applies to Figure 5aThe comparator-based runtime measurement shown in , where the comparator and the “time-to-digital” converter already generate the position data. Figure 5a and Figure 5b In the lower evaluation path shown in , indicator values ​​are calculated from the signal waveform, and based on the indicator values ​​and their comparison with threshold values, the position data is marked as unreliable.

[0114] Via a third line 540 branching off from the second line 520, the amplified signal is also fed to an ADC 545, the digital output signal of which is fed to a statistical unit 550 for the aforementioned statistical evaluation of the waveform in order to determine or supply an impulse indicator variable 555 as a result.

[0115] In a second exemplary embodiment of the device depicted with dashed lines, the signal provided by the measuring sensor 500 is fed via a fourth line 560 to a second amplifier 565 and only then to the ADC 545. In the second exemplary embodiment, the third line 540 is therefore obsolete or not arranged. This enables the amplification factor or the corresponding level increase of the signal supplied by the measuring sensor 500 to be selected differently for further processing by the comparator 525 and the ADC 545 or the subsequent statistical evaluation 550 in order to be able to provide a suitable or optimal signal level for the two separate further processing steps 525-535 or 550, 555.

[0116] Figure 5b Three other exemplary embodiments of the aforementioned device with and without additional dashed components are shown. In these exemplary embodiments, the measurement signal supplied by the measurement sensor (“pickup”) 500 arranged at one end of the magnetostrictive waveguide 505 is also fed to the first amplifier 515 via the first line 510.

[0117] In the third exemplary embodiment, the amplified signal is fed to a first ADC 575 via a second line 570, where the amplified signal is first digitized. The digital signal or corresponding data is fed to a digital signal processor 585 via a third line 580 and further fed to a statistical unit 605 via a branched fourth line 600 for the aforementioned statistical evaluation of the aforementioned waveform.

[0118] The data supplied by the digital signal processor 585 are fed after evaluation to a time-of-flight analysis 590 which in turn provides a position result 595 .

[0119] The evaluation results supplied by the statistical unit 605 then in turn supply the impact indicator variable 610 described above.

[0120] exist Figure 5bIn the fourth exemplary embodiment of the device depicted with dashed lines in FIG. 5 , the amplified signal supplied by the first amplifier 515 is fed to the second ADC 635 via a branched additional fifth line 630. This enables, for example, the selection of different analog-to-digital conversion parameters for the respective subsequent components, namely the digital signal processor 585 and the statistical unit 605. For the purposes described, the digitized signal or the corresponding digital data is fed to the statistical unit 605 via a sixth line 640.

[0121] In a fifth exemplary embodiment of the apparatus, Figure 5b As is additionally depicted in dashed lines in FIG. 5 , the measurement signal supplied by the measuring sensor 500 via the first line 510 is additionally fed via a branched seventh line 615 to a second amplifier 620 and then only via an eighth line 625 to a second ADC 635 present according to the fourth exemplary embodiment. This in turn enables different selection of the amplification factor or the corresponding level increase of the signal supplied by the measuring sensor 500 for further processing by subsequent further processing in the digital signal processor 585 and in the statistical unit 605 connected via the sixth line 640, in order to be able to provide a suitable or optimal signal level for the two individual further processing steps 585-595 or 605, 610.

[0122] It should also be noted that in the fourth exemplary embodiment, the fourth line 600 is abandoned, while in the fifth exemplary embodiment, both lines 600, 630 are abandoned.

[0123] according to Figure 5a and Figure 5b In the exemplary embodiment shown in , the evaluation 530 of the run time of the target signal for determining the target position 535 is performed either by the high-resolution time-to-digital converter 525 or by generating a stop input signal by an analog zero comparator for the pre-amplified detection signal. This involves a digital evaluation of the corresponding detected waveform, which is generated by digitizing the detection signal.

[0124] from Figure 5a and Figure 5b It can also be seen that the requirements for the signal amplifier and the corresponding analog-to-digital converter (ADC) 545, 575, 635 and their parameters for generating the impact indicator value may be different from those required for evaluating the running time. Therefore, the evaluation can be performed by the comparator 525 (possibly in combination with the aforementioned "time to digital" converter (TDC 525)), or by digital signal processing using a digital signal processor.

[0125] As mentioned above, since only values ​​close to zero are relevant for shock detection, it is advantageous to apply a higher amplification - regardless of the saturation of the target-affected portion of the signal. On the other hand, unlike the case where the run time is determined from the digitized signal behavior by digital signal processing, the calculation of the shock index value does not require a high resolution or a precise analog-to-digital converter. Furthermore, since each method calculates the shock index value by statistical methods and the signal bandwidth of the shock- and vibration-related signal components is small, the scanning frequency of the associated ADC 545, 575, 635 is also not time-critical.

[0126] Finally, it should be noted that the method described above can be implemented in a conventional, comparator-based magnetostrictive position measuring system or in a purely software-based magnetostrictive position measuring system based on the evaluation of the signal waveform described above.

Claims

1. A method for operating an electromagnetic, in particular magnetostrictive sensor or transducer (500), in particular a position measuring system (500, 505), characterized in that mechanical shocks and / or vibration conditions occurring during operation of the sensor / transducer (500) or vibrations of the sensor / transducer (500) caused thereby are detected by statistically evaluating signal fluctuations (120, 125) in the region of a baseline or zero line (105) of a measurement signal provided by the sensor / transducer (500).

2. The method according to claim 1 is characterized in that, during the evaluation of the signal fluctuations (120, 125) in the region of the baseline / zero line (105), a comparison is made with a characteristic curve (130) of the measurement signal in an undisturbed situation.

3. The method according to claim 2 is characterized in that the measurement signal (545) is digitized and the resulting data points of the corresponding signal waveform (100-105) are evaluated in a zero-centered, time-limited part of the measurement signal (Figure 1a, Figure 1b).

4. The method according to claim 2 or 3, characterized in that the signal waveform of an undisturbed electromagnetic effect, in particular a magnetostrictive material, is used as the characteristic curve (105, 115) of the measurement signal in the undisturbed situation.

5. A method according to one of the preceding claims, characterized in that signal waveforms (100-115) that deviate from the characteristic curve (105, 115) in the undisturbed situation are marked as "unreliable" and / or are excluded from further processing or are given a lower weight or are deleted.

6. A method according to claim 5, characterized in that if the deviation of the currently recorded signal waveform (100-115) exceeds the value (400) of an empirically predeterminable threshold of at least one determined indicator variable (see Figure 4a), the signal waveform (100-115) is excluded from the further processing, or is given a lower weight, or is deleted.

7. A method according to claim 6, characterized in that the deviation of the currently measured signal waveform (100-115) is quantified because a relevant increased value of at least one determined indicator variable (see Figure 4a) is detected because the at least one indicator variable exceeds the value of an empirically predeterminable threshold value predetermined in the case of the undisturbed measurement (400).

8. A method according to one of the preceding claims, characterized in that the evaluation of signal fluctuations in the region of the baseline / zero line (105) of the signal waveform (100-115) is performed by statistically evaluating (550, 605) data points generated from the signal waveform (100-115).

9. The method according to claim 8, characterized in that the statistical evaluation is performed based on data points of the corresponding signal waveform generated from the measurement signal using one or more of the following mathematical evaluation methods: - "quartile method", in which the difference between the first quartile and the last quartile (e.g., the quartiles) of the data points of the signal waveform is evaluated; - "histogram method", in which both sides of the histogram of data points of the signal waveform close to the value zero are evaluated according to the ratio of the two sides to the center; - "Standard Deviation Method", in which the standard deviation of the data points of the signal waveform close to the baseline / zero line is analyzed; - "Data Density Method", in which the data density of data points near the baseline / zero line of the signal waveform is analyzed.

10. The method according to claim 9, characterized in that in the histogram method (130, 135) the number of data points is evaluated according to the value deviation from the value zero.

11. A method according to one of the preceding claims for operating a magnetostrictive sensor or transducer (500) of a position measurement system (500, 505) having a magnetostrictive waveguide (505), wherein a mechanical rotational or torsional wave packet is generated in the waveguide (505) by interrogation current pulses propagating in two directions along the waveguide (505), wherein a measurement sensor (500) arranged near one end of the waveguide (505) is provided, the measurement sensor being arranged for receiving the measurement signal for converting the rotational / torsional wave packet travelling in the waveguide (505) into an electrical signal, wherein during a first time period of the measurement signal the interrogation current pulse induces an oscillating interrogation noise (100) and a collective mechanical response of the waveguide (505) caused by the interrogation current pulse in the received measurement signal in the region of the baseline / zero line, and wherein, During a second time interval of the measurement signal, an oscillating background noise (105) and a target signal (110) superimposed on the background noise (105) are present on the interrogation noise (100) in the region of the baseline / zero line, wherein the target signal (110) is followed by a further oscillating background noise (115), the method being characterized in that for at least one time range of the oscillating noise (105, 115) in the region of the baseline / zero line (105), the measured values ​​of the recorded measurement signal in the at least one time range are compared with corresponding measured values ​​of a sensor / transducer (500) disturbed by mechanical shock and / or vibration conditions and an undisturbed sensor / transducer (500), and depending on the result of the comparison, a disturbed or undisturbed condition of the sensor / transducer (500) is inferred.

12. The method according to claim 11, characterized in that the statistical evaluation is limited to a time interval (see Figures 4a, b) of the data points of the signal waveform (100-115), which time interval is not affected by the interrogation noise (100) or the superimposed target signal (110).

13. Method according to claim 11 or 12, characterized in that the measured values ​​recorded in at least one time range of the sensor / transducer disturbed by mechanical shock and / or vibration conditions are disturbed by low-frequency noise (120, 125).

14. A device for operating an electromagnetic, in particular magnetostrictive sensor or transducer (500), in particular a position measurement system (500, 505), characterized in that it comprises an evaluation unit (515-555) for detecting any mechanical shock and / or vibration conditions occurring during operation of the sensor / transducer (500) or any vibrations of the sensor / transducer (500) caused thereby, by evaluating signal fluctuations in the region of a baseline or zero line (105) of a measurement signal provided by the sensor / transducer (500).

15. The device according to claim 14 is characterized by: at least one amplifier (515, 565), which is used to amplify the measurement signal supplied by the sensor / transducer (500); a comparator (525), which is used to compare at least one time interval of the amplified measurement signal in the area of ​​the baseline / zero line (105) with a characteristic curve (105, 115) of data points of the signal waveform (100-115) of the undisturbed situation; a flight time analysis unit (530), which supplies a position result (535); an analog-to-digital converter (545), which is used to convert the measurement signal into digital data points of the corresponding signal waveform; and a statistical unit (550), which is used to perform a statistical evaluation according to one or more of claims 1 to 13 in order to determine or supply an impact indicator variable (555).

16. An apparatus according to claim 14 or 15, characterized in that: at least one amplifier (515, 620), the at least one amplifier is used to amplify the measurement signal supplied by the sensor / converter (500); at least one analog-to-digital converter (575, 635), the at least one analog-to-digital converter is used to convert the measurement signal into data points of a corresponding signal waveform; a digital signal processor (585), the digital signal processor is used to evaluate the measurement data; a flight time analysis unit (590), the flight time analysis unit supplies a position result (595); and a statistical unit (605), the statistical unit is used to perform a statistical evaluation according to one or more of claims 1 to 13 in order to determine or supply an impact indicator variable (610).

Images