Determination of the speed of a component using a non-speed measuring sensor

Abstract

The invention relates to a method for determining a rotational speed of a component by means of a non-rotational speed-measuring sensor, comprising the steps of: providing (S1) sensor data from the non-rotational speed-measuring sensor in the time domain; transforming (S2) the sensor data from the time domain into a frequency spectrum and displaying the frequency spectrum with reference to frequency, time, and amplitude in a frequency-time diagram on a display unit; detecting (S3) a first input from a user as a specification of manually selected frequency peaks in the frequency-time diagram; determining (S4) median frequencies for all points in time in the frequency-time diagram by means of a median frequency algorithm and / or determining a ridge by means of a ridge-tracking algorithm; and determining (S5) a rotational speed curve (1) of the component over time based on the determined median frequencies or the ridge.

Description

[0001] The invention relates to a method for determining a rotational speed of a component by means of a non-rotational speed measuring sensor.

[0002] In vehicle development, the question of providing a measurement or estimate of the rotational speed of a rotating component often needs to be answered. A large number of components on the vehicle or in production lines rotate, and it is important to know these rotational speeds. Dedicated sensors are often available to determine the rotational speed (e.g. crankshaft speed, wheel speed sensors for an anti-lock braking system, etc.). In some cases, however, no sensors are provided, or the information from the sensors is not accessible or difficult to access, for example, the rotational speed of a fuel pump, the rotational speed of a water pump in the cooling circuit, the rotational speed of a fan motor, the wheel speeds of competition vehicles, or the transmission shaft speeds of competition vehicles.

[0003] Since the rotational speeds of components are of particular interest in NVH development ("NVH" is an abbreviation for "noise, vibration, and harshness"), for example, to create Campbell diagrams, determine critical speed ranges, or investigate noise sources, a simple method is required to determine the rotational speed of rotating components for these cases as well. Furthermore, since rotating components are not always easily accessible (such as a cooling water pump rotor or a fuel pump), optical or magnetic speed sensors are only of limited use.

[0004] Instead, accelerometers, microphones, or Hall sensors can be used, which are attached externally to or on the machine with the rotating component. Using a Fourier transformation, frequency spectra can then be calculated from the time data at specific time intervals. If certain characteristic excitation mechanisms of the rotating component are known (such as imbalance, electromagnetic forces in electric motors, or the number of blades in turbines), the speed of the rotating component can be determined from the spectrum recorded using the data from one of the aforementioned sensors with recording on the machine housing. To do this, an algorithm is used to try to select a characteristic line from a so-called color plot, which represents the frequency spectrum of the measured variable over the measurement time.The line represents frequencies of higher amplitudes that are caused by rotation and therefore depend on the speed.

[0005] In the prior art, it is known to calculate a color plot (showing an amplitude spectrum in relation to the measurement time) from time data from a sensor that does not directly measure the speed, as described above, and to generate a speed curve from this using additional user input. However, the generated speed curve is often unsatisfactory and / or the required (time-consuming) user input is very high: An algorithm typically requires information about a selected line - depending on the excitation mechanism - in order to then recognize this line from the generated spectra and generate a speed curve from it.

[0006] Such a procedure, known in the state of the art, is described, for example, in the documentation of the "ArtemiS SUITE" software from HEAD Acoustics GmbH: "RPM Generator. Deriving accurate revolution speed information from order curves." by Thilo Leitmann from NVH Sales, available at the time of writing at https: / / global.head-acoustics.com / downloads / eng / artemis / ArtemiS_SUITE_RPM_Generator_e.pdf

[0007] DE 10 2022 202 092 A1 also relates to a method for determining the rotational speed of a rotating component of a device, comprising: detecting a vibration signal emanating from the device over time for one or more evaluation times, wherein the vibration signal comprises a plurality of superimposed signal components; forming an amplitude spectrum in the frequency domain for at least one evaluation time from the detected vibration signal; determining a fundamental frequency and / or at least one harmonic of the amplitude spectrum by pitch detection; and determining the rotational speed of the rotating component at the evaluation time from the determined fundamental frequency and / or a determined harmonic.

[0008] DE 10 2022 201 317 A1 further relates to a method for detecting a rotational speed of a machine, comprising: detecting a vibration originating and / or caused by the machine; determining an approximate rotational speed of the machine; and determining a final rotational speed of the machine based on the approximate rotational speed and a characteristic value specific to the machine, wherein the characteristic value is a value that defines a distance between typically consecutive harmonics in a vibration caused by the machine.

[0009] DE 10 2017 104 207 A1 relates to a method for determining at least one rotational speed of a machine, wherein the complex spectrum of this variable is determined from the measurement of a vibration variable on the machine over time, wherein frequency interpolation is carried out; boundary conditions for the evaluation of the spectrum are defined, which include the permissible frequency range of an expected main rotational speed, a set of relative frequencies related to the main rotational speed in the form of frequency multipliers and a weighting factor for the respective relative frequency; a spectral probability density is calculated taking into account the boundary conditions, which results for each frequency of the permissible frequency range as the sum of the amplitude of the spectrum weighted by the respective weighting factor at the frequency multiplied by the respective frequency multiplier;and the main speed is determined from the frequency with the maximum probability density.;

[0010] DE 10 2021 204 884 A1 further relates to a method for measuring an indicative parameter of a rotational speed of a component, comprising the following steps: detecting a vibration signal obtained from a vibration sensor attached to the component, the method comprising: filtering the vibration signal, based on a predetermined frequency range, into a sinusoidal signal and generating a pulse wave signal from the sinusoidal signal to read the value of the parameter.

[0011] Finally, US 5744723 A relates to a method for determining the speed of a rotating shaft, comprising: detecting vibrations generated by the shaft to generate a test vibration signal at an unknown speed of the shaft, converting the test vibration signal into a test frequency spectrum, compiling a table of the peaks found in the test frequency spectrum, providing a reference frequency spectrum corresponding to a known speed of the shaft, compiling a table of the peaks found in the reference frequency spectrum and determining a stretch factor that provides an optimal correlation between the table of peaks of the test frequency spectrum and the table of peaks of the reference frequency spectrum, and calculating the unknown speed of the rotating shaft using the known speed of the reference frequency spectrum and the stretch factor.

[0012] In practice, these common methods, which also allow inferences about the speed of rotating components such as actuators, have the disadvantage of being inaccurate or inaccurate during very dynamic speed changes or cyclical operation (rapid switching between on and off). If the speed of a component such as an actuator is not available via the CAN bus, this complicates system understanding. It also complicates NVH analysis and the search for the causes of noise phenomena. Perceived noises cannot be assigned to a specific component or operating state.

[0013] The object of the invention is to estimate a rotational speed of the rotating component in an improved manner from a time course of sensor data of a sensor that does not directly measure the rotational speed in the area of ​​influence of a rotating component.

[0014] The invention is based on the features of the independent claims. Advantageous developments and refinements are the subject of the dependent claims.

[0015] A first aspect of the invention relates to a method for determining a rotational speed of a component by means of a non-rotational speed measuring sensor, comprising the steps: - Providing sensor data from the non-speed measuring sensor in the time domain; - Transforming the sensor data from the time domain into a frequency spectrum and displaying the frequency spectrum with reference to frequency, time and amplitude in a frequency-time diagram on a display unit; - Capturing an initial user input as a specification of manually selected frequency peaks in the frequency-time diagram; - Determining median frequencies for all time points of the frequency-time diagram using a median frequency algorithm and / or determining a ridge using a ridge tracking algorithm; and - Determination of a rotational speed curve of the component over time based on the determined median frequencies or the burr.

[0016] Depending on the properties of a rotating component, the characteristic influences such as forces at the measurement location that cause the peaks in the frequency spectra vary. Without knowledge of the relationship between the rotational speed and the measured spectrum, it is difficult to draw conclusions about the rotational speed. However, the invention takes advantage of the fact that in cases where more is known about the rotating component, such as information about an imbalance, the number of teeth on a gear wheel on a transmission shaft, the number of blades on a turbine, the number of pole pairs on an electric motor, the number of cylinders on an internal combustion engine, or other information, the order of the peak in the frequency spectrum can be determined. The term "order" refers to the ratio of the frequency of the peak to the rotational speed.

[0017] By evaluating measured time data from a sensor on / next to a machine housing, the speed of a rotating machine part inside can be determined. The recorded time data can include, in particular, accelerations on the machine housing, sound pressure levels near the machine, magnetic field strengths on the machine housing if an electric motor is present, or other suitable measurement variables.

[0018] The frequency-time diagram is calculated and can be displayed as a color plot. The user can draw a rough line to approximate the desired line corresponding to the temporal speed curve. Next, the user can specify the distance from this rough line at which the exact curve should be searched. This corresponds to a search range for a processing unit that performs the subsequent computational steps. In an alternative embodiment, the search range can also be specified automatically.

[0019] The median frequency of the selected frequency range is then calculated for all time points using computer implementation, or the ridge tracking algorithm is used.

[0020] Both median frequency algorithms and ridge tracking algorithms are known in the art. Median frequency algorithms typically utilize power spectral densities and a cumulative sum vector of these. Further information on a ridge tracking algorithm can be found, for example, in the publication "On the extraction of instantaneous frequencies from ridges in time-frequency representations of signals" by the authors D. Latsenko, P.V.E. McClintock, and A. Stefanovska, available at https: / / arxiv.org / pdf / 1310.7276.pdf.

[0021] Both methods produce a line from which, with the information about the order, the speed can be deduced. If the result is unsatisfactory, the time range with the error can be re-drawn with a rough line. To achieve greater accuracy, either a more precise line can be drawn or a smaller distance can be set around the line to search for the exact curve.

[0022] Advantageously, a quick and simple method is obtained that reliably achieves good results. Methods known in the prior art are more time-consuming or often produce poorer results. When using the method according to the invention, which can be carried out completely automatically on a computing unit except for the manual input, only a user input in the form of a rough line is required. This only takes seconds, and after that the search area for the algorithm is narrowed down to such an extent that very reliably good results are achieved. A disadvantage of the prior art can be avoided in this way, namely that a stop of the rotating component or a sudden jump in speed is missed. In order to correct these errors, a great deal of effort and significantly more complex user input are often necessary in the prior art.The combination of a coarse user input and an algorithm for line detection (median frequency or ridge tracking) achieves a very good result with only moderate effort for the user.

[0023] According to an advantageous embodiment, a second input from the user is recorded as a specification of a search area around the first input.

[0024] According to a further advantageous embodiment, the detection of the first input and the determination of median frequencies or a ridge are carried out iteratively in order to determine the speed curve.

[0025] According to a further advantageous embodiment, the detection of the first input, the detection of the second input, and the determination of median frequencies or a ridge are carried out iteratively in order to determine the speed curve.

[0026] According to a further advantageous embodiment, the non-speed-measuring sensor is one of the following: acceleration sensor, whose sensor data are measured accelerations; sound pressure sensor, whose sensor data are measured sound pressures; Hall sensor, whose sensor data are magnetic field strengths.

[0027] According to a further advantageous embodiment, the sensor data is transformed by a Fourier transformation.

[0028] According to a further advantageous embodiment, the user's first input comprises a line. The line may have corners, edges, jumps, kinks, and the like, but in all cases has only a 1-dimensional shape.

[0029] According to a further advantageous embodiment, the frequency-time diagram is output as a color plot, with a time and a frequency on each coordinate axes and a specific color for specific amplitudes in the sensor data.

[0030] According to a further advantageous embodiment, the sensor data is output at frequencies in the audible range to support the user's initial input. This allows the user to listen to the time signal (e.g., an acceleration). This is helpful, for example, to determine whether the component was stationary during a certain time period and therefore had a zero rotational speed.

[0031] Further advantages, features, and details will become apparent from the following description, in which at least one embodiment is described in detail—possibly with reference to the drawings. Identical, similar, and / or functionally equivalent parts are provided with the same reference numerals.

[0032] They show: Fig. 1: A frequency-time diagram of a Fourier-transformed measured quantity. Fig. 2: Example specifications in the frequency-time diagram of the Fig. 1 by a user. Fig. 3: An example of a machine-determined speed curve of a rotating component, determined using the frequency-time diagram of the Fig. 1 and the requirements of the Fig. 2.

[0033] Fig. Figure 1 shows a frequency-time diagram with amplitudes in colors as a color plot. Time is indicated on the horizontal axis, and frequency on the vertical axis. Such a frequency-time diagram is obtained by first providing sensor data from a non-speed-measuring sensor, such as an acceleration sensor, in the time domain (S1) and transforming it using Fourier transformation (S2) to obtain information about a frequency spectrum.

[0034] In the next step, outlined in Fig. 2, a user input is recorded S3 at an input interface of a computer on which the process runs and the calculation steps are carried out. This input comprises a relevant line 5, which is roughly traced, with color-contrasting contours showing a time and frequency curve. In addition to the traced line 5, the user also specifies a search area 3 in the form of an upper and lower boundary around the line 5. This defines a search area for the computing unit, and a reference point for the desired speed curve is predefined by the rough line 5. Finally, the specified line 5 is refined by an algorithm on the computing unit by applying a median frequency algorithm or a ridge tracking algorithm S4. The curve thus determined is significantly refined compared to the specified line 5.

[0035] If the user now specifies and presets a second order, a time course of a speed can be determined S5, as in Fig. 3. The horizontal axis describes time and the vertical axis a rotational frequency of the component.

[0036] Although the invention has been illustrated and explained in detail by preferred embodiments, the invention is not limited by the disclosed examples, and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention. It is therefore clear that numerous variations exist. It is also clear that exemplary embodiments are truly only examples and should not be construed as limiting the scope, possible applications, or configuration of the invention in any way.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or arrangement of individual elements mentioned in an exemplary embodiment, without departing from the scope of protection defined by the claims and their legal equivalents, such as further explanations in the description. List of reference symbols 1 Speed ​​curve 3 Search area 5 Line S1 Provision S2 Transform S3 Capture S4 Determine S5 Determine QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2022 202 092 A1

[0007] DE 10 2022 201 317 A1

[0008] DE 10 2017 104 207 A1

[0009] DE 10 2021 204 884 A1

[0010] US 5744723 A

[0011] Cited non-patent literature

[0000] Thilo Leitmann from NVH Sales, available at the time of writing at https: / / global.head-acoustics.com / downloads / eng / artemis / ArtemiS_SUITE_RPM_Generator_e.pdf

[0006] On the extraction of instantaneous frequencies from ridges in time-frequency representations of signals” by the authors D. Latsenko, PVE McClintock, A. Stefanovska, available at https: / / arxiv.org / pdf / 1310.7276.pdf

[0020]

Claims

[1] Computer-implemented method for determining a rotational speed of a component by means of a non-rotational speed measuring sensor, comprising the steps: - Providing (S1) sensor data of the non-speed measuring sensor in the time domain; - Transforming (S2) the sensor data from the time domain into a frequency spectrum and displaying the frequency spectrum with reference to frequency, time and amplitude in a frequency-time diagram on a display unit; - capturing (S3) a first input from a user as a specification of manually selected frequency peaks in the frequency-time diagram; - determining (S4) median frequencies for all time points of the frequency-time diagram using a median frequency algorithm and / or determining a ridge using a ridge tracking algorithm; and - Determining (S5) a speed curve (1) of the component over time on the basis of the determined median frequencies or the burr. [2] Method according to claim 1, wherein a second input of the user is detected as a specification of a search area (3) around the first input. [3] Method according to one of the preceding claims, wherein the detection of the first input and the determination of median frequencies or a ridge are carried out iteratively in order to determine the speed curve (1). [4] Method according to claims 2 and 3, wherein the detection of the first input, the detection of the second input, and the determination of median frequencies or a ridge are carried out iteratively in order to determine the speed curve (1). [5] Method according to one of the preceding claims, wherein the non-speed-measuring sensor is one of the following: acceleration sensor, whose sensor data are measured accelerations; sound pressure sensor, whose sensor data are measured sound pressures; Hall sensor, whose sensor data are magnetic field strengths. [6] Method according to one of the preceding claims, wherein the sensor data is transformed by a Fourier transformation. [7] Method according to one of the preceding claims, wherein the first input of a user comprises a line (5). [8] Method according to one of the preceding claims, wherein the frequency-time diagram is output as a color plot, with a time and a frequency respectively on coordinate axes and a specific color for specific amplitudes in the sensor data. [9] A method according to any one of the preceding claims, wherein the sensor data is output in frequencies in the audible range to support the user's first input.

Images