Method for diagnosing a fault in a measuring device and a measuring device

Abstract

The invention relates to a method for diagnosing a fault in a measuring device having a measuring head that can be moved along a measurement path, comprising the following steps: - (10) capturing a reference signal by repeatedly travelling along the measurement path at a target speed of the measuring head and detecting at least one parameter in a fault-free state of the measuring device; - (12) capturing a measurement signal by travelling along the measurement path at the target speed of the measuring head and detecting the parameter; - (14) transferring the measurement signal and the reference signal from the time domain to the frequency domain with a first predefined time window; - (16) determining a plurality of values of the energy content of the measurement signal and of the reference signal in the first time window; - (18) probabilistically evaluating the determined values of the energy content of the measurement signal in relation to the determined energy contents of the reference signal in the first time window by means of extreme value theory, and determining whether the determined energy content of the measurement signal represents a fault of the measuring device.

Description

[0001] Title: Method for diagnosing a fault in a measuring device and a measuring device

[0002] Description

[0003] The invention relates to a method for diagnosing a fault in a measuring device having a measuring head movable along a measuring path, with features of claim 1, and to a measuring device having features of the independent claim.

[0004] In machine tools, measuring systems for length or speed that are not completely encapsulated, especially external ones (not mounted in the motor), are usually exposed to contamination and mechanical stress. This can lead to phenomena such as unwanted noise or incorrect measurements. Causes can be contamination, alignment and / or damage to the

[0005] The problem could be with the measuring tape or the measuring head. This is often not properly recognized, and a replacement of the material is attempted. This measure is generally unsuccessful because the incorrectly recognized error (such as contamination, incorrect alignment, and / or damage) is not corrected.

[0006] It is therefore an object of the present invention to provide a method for diagnosing a fault in a measuring device and a measuring device, wherein the above disadvantages are eliminated.

[0007] The above object is achieved by a method for diagnosing a fault in a measuring device having the features of claim 1. The measuring device comprises a measuring head movable along a measuring path. The method comprises the steps:

[0008] Acquiring a reference signal by traversing the measuring section once, in particular multiple times, at a target speed of the measuring head and detecting at least one parameter in a fault-free state of the measuring device. In other words, a reference signal is determined in a state in which there is no fault in the measuring device. The parameter is detected in particular as a function of time. The parameter is therefore recorded or measured in particular on a time scale. The target speed also assigns the parameter to a location on the measuring section. The measuring device can thus be measured (multiple times) in a (fault-free) basic state. Acquiring a measurement signal by traversing the measuring section once, in particular multiple times, at the target speed of the measuring head and detecting the parameter.In other words, a measurement signal is determined in a state in which an error could be present in the measuring device. The parameter is recorded in particular as a function of time. The parameter is therefore detected or measured in particular on a time scale. The measuring device can thus be measured (multiple times) in a state in which an error may be present. The target speed is in particular a speed specified (e.g. by a user or a program) at which the measuring head is moved along the measuring section during the acquisition of the reference signal and during the acquisition of the measurement signal. The target speed is in particular constant.

[0009] Transferring the measurement signal and the reference signal from the time domain to the frequency domain using a first window. The first window has, in particular, a start time, an end time, and a window length.

[0010] Determine the energy content of the measurement signal and the reference signal in the first window.

[0011] Probabilistic evaluation of the determined energy signal of the measurement signal in relation to the energy content of the reference signal in the first window using extreme value theory and determining whether the determined energy content of the measurement signal represents an error in the measuring device.

[0012] This makes it possible to calculate the energy content of the measurement signal around a central position of the first window and to use it to approximate the measurement signal. The probabilistic evaluation makes the method robust, particularly against the influence of measurement noise. Using extreme value theory, it is possible to make a statement about how likely it is that the determined energy content of the measurement signal is still an extreme value of the measuring device in the error-free state (ground state) or already represents a fault in the measuring device. In other words, from the detected parameter, a probability can be determined as to whether or not there is an error in the measuring device. If this probability is small, it can be assumed that the detected parameter represents an error in the measuring device or that the measuring device is not in the error-free state (ground state).This makes it possible to provide a standardized, rapid, and / or low-effort procedure for diagnosing a measuring device error. The procedure can be automated, eliminating the user as a possible source of error. The procedure can therefore be reproduced with consistent quality.

[0013] According to a further development of the method, the method may comprise the following steps:

[0014] Determining at least one second window by shifting the first window along a time axis of the measurement signal. The second window has in particular a start time, an end time and a window length. For example, the start time of the first window can be shifted along the time axis of the measurement signal. The first window and the second window can overlap. In other words, the start time of the second window can be before the end time of the first window. The shift of the first window along the time axis can represent a shift of a viewing point along the measurement section. Due to the overlap of the windows, the location of the fault can be better localized. For example, if...If an error is detected in the first window and no error is detected in the second window, then it can be concluded that the error lies in the first window and outside an overlap area between the first and second windows. The error location can thus be further narrowed down within a single window (in this example, within the first window).

[0015] Transfer the measurement signal and the reference signal from the time domain to the frequency domain using the second window.

[0016] Determine the energy content of the measurement signal and the reference signal in the second window.

[0017] Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in the second window using extreme value theory and determining whether the determined energy content of the measurement signal represents an error in the measuring device.

[0018] This makes it possible to easily determine whether an error is present in the first window and / or the second window. This allows the error (if an error is detected) to be localized and / or limited in time. A more precise localization of the error can be achieved. According to a further development of the method, the method can include the following steps:

[0019] Determining a plurality of second windows by shifting the first window along the time axis of the measurement signal. The entire measurement path can be covered by the first window and the plurality of second windows. Two adjacent windows can overlap. Covering the entire measurement path ensures that no error is overlooked.

[0020] Transfer the measurement signal and the reference signal from the time domain to the frequency domain with each of the second windows.

[0021] Determine the energy content of the measurement signal and the reference signal in each of the second windows.

[0022] Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in each of the second windows using extreme value theory and determining whether the respectively determined energy content of the measurement signal represents an error in the measuring device.

[0023] This allows the entire measuring section to be easily examined for any faults. If a fault is detected, it can be precisely located.

[0024] According to a further development of the method, the first window, the second window, and / or the second windows can each have the same start time, the same end time, and / or the same window length for the reference signal and the measurement signal. The first window, the second window, and / or the second windows can each be configured identically.

[0025] By adjusting the respective windows, the location of the error can be adjusted as desired. For example, an overlap area between two adjacent windows and / or the window length can be adjusted.

[0026] According to a further development of the method, the transfer of the measurement signal from the time domain to the frequency domain can be carried out using a (windowed) short-time Fourier transformation. Alternatively or additionally, the transfer of the reference signal from the time domain to the frequency domain can be carried out using a (windowed) short-time Fourier transformation. It is also conceivable that other (energy-preserving) signal transformations can be used to transfer the measurement signal and / or the reference signal from the time domain to the frequency domain.

[0027] In particular, due to the good frequency resolution of the short-time Fourier transformation, the transfer of the measurement signal and / or the reference signal from the time domain to the frequency domain can be optimized.

[0028] According to a further development of the method, the parameter can be an actual speed, an actual position and / or an actual current of the measuring head. In this case, a distinction must be made between the actual speed and the target speed. The target speed means a predetermined (in particular constant) speed with which the measuring head is moved along the measuring section. The actual speed means a measured speed which is detected during a movement of the measuring head (detected parameter). If there is an error in the

[0029] If an error occurs in the measuring device, e.g., in the form of contamination along the measuring path, the actual speed is influenced by the error. The corresponding measuring signal can thus exhibit fluctuations corresponding to the error. In other words, an error (e.g., contamination) along the measuring path leads to a change in the measuring signal.

[0030] By determining various parameters, the diagnosis of the error can be further optimized and / or made more flexible.

[0031] According to a further development of the method, the energy content of the measurement signal can be determined using Parseval's theorem. Alternatively or additionally, the energy content of the reference signal can be determined using Parseval's theorem.

[0032] Parseval's theorem can be represented by the following formula:

[0033] Parseval's theorem states that the total energy of a signal (measurement signal or reference signal) can be determined by summing the power per sample over time or the

[0034] Spectral power can be calculated versus frequency. In other words, the energy of a signal in the time domain is equal to its energy in the frequency domain.

[0035] According to a further development of the method, the reference signal can be defined as normally distributed. The distribution of the reference signal can be determined empirically. It is also conceivable to base the reference signal on a distribution other than the normal distribution.

[0036] For example, individual differences between individual measuring devices, which may arise due to manufacturing tolerances, can be taken into account in a model of "normality". The measurement noise (in particular that which cannot be influenced) can be assumed to be normally distributed. This allows, in particular, the determined energy content of the measurement signal to be viewed as a normally distributed random variable centered around the determined energy content of the reference signal.

[0037] According to a further development of the method, the method may comprise the step:

[0038] Determining a probability with which the determined energy content of the measurement signal is an extreme value of the reference signal.

[0039] This makes it possible to determine the probability of an error occurring using simple means.

[0040] According to a further development of the method, the method may comprise the steps of: determining at least one limit value for the probability.

[0041] Triggering an alarm signal when the limit value is reached and / or exceeded. The alarm signal can be issued visually and / or acoustically, for example.

[0042] This allows limit values ​​to be introduced in a statistically interpretable manner. For example, a probability of 1% can be set as the limit value. Thus, if the limit value of 1% is reached or exceeded, it can be assumed with a 99% probability that the determined energy content of the measurement signal is not an extreme value of the reference signal, and thus that an error is present.

[0043] According to a further development of the method, the measurement signal can be acquired during a lubrication cycle of the measuring device. Alternatively or additionally, the measurement signal can be acquired during a machine startup of the measuring device.

[0044] By implementing ("hiding") a measuring run in a lubrication cycle or machine start-up of the measuring device, downtimes of the measuring device can be avoided or at least reduced. The method can thus be implemented in necessary work steps of the measuring device. In other words, the method is essentially carried out parallel to main time and does not require any additional machine time. According to a further development of the method, the acquisition of the measuring signal can be carried out within a maximum time of 10 seconds.

[0045] This allows the method to be used as time-efficiently as possible, so that downtime of the measuring device can be avoided or at least reduced.

[0046] The method also allows for a statement, particularly regarding the type of detected error. This allows a distinction to be made between a local error (e.g., contamination) along the measuring section and a tilt of the measuring section or the measuring head. The presence of a tilt can be detected, for example, in a characteristic peak in a linear spectrum at the second harmonic multiple of the measuring signal.

[0047] The above object is achieved by a measuring device with a measuring head movable along a measuring path, having the features of the independent claim. The measuring device is configured to carry out the method according to the above explanations.

[0048] With regard to the advantages that can be achieved, reference is made to the relevant explanations of the method. The measures described in connection with the method and / or those explained below can be used to further refine the measuring device.

[0049] Further features, details and advantages of the invention emerge from the wording of the claims and from the following description of an embodiment with reference to the

[0050] Drawing. It shows:

[0051] Fig. 1 is a flowchart of a method for diagnosing a fault in a measuring device with a measuring head movable along a measuring path.

[0052] Figure 1 shows a flowchart of a method for diagnosing a fault in a measuring device with a measuring head movable along a measuring path. The method comprises the following steps:

[0053] 10: Acquisition of a reference signal by one, in particular multiple, travel of the measuring section at a desired speed of the measuring head and detection of at least one parameter in a fault-free state of the measuring device.

[0054] The parameter can be an actual speed, an actual position and / or an actual current of the measuring head.

[0055] 12 : Recording a measuring signal by traveling along the measuring section once, in particular multiple times, at the desired speed of the measuring head and detecting the parameter .

[0056] The measurement signal can be acquired during a lubrication cycle and / or during machine startup of the measuring device. The measurement signal can be acquired within a maximum time of 10 seconds. 14: Transferring the measurement signal and the reference signal from the time domain to the frequency domain using a first window.

[0057] 16 : Determining the energy content of the measurement signal and the reference signal in the first window .

[0058] 18 : Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in the first window using extreme value theory and determining whether the determined energy content of the measurement signal represents an error in the measuring device.

[0059] The procedure comprises the following steps:

[0060] 20: Determining at least one second window by shifting the first window along a time axis of the measurement signal. The first window and the second window may overlap.

[0061] 22 : Transferring the measurement signal and the reference signal from the time domain to the frequency domain using the second window.

[0062] 24 : Determining the energy content of the measurement signal and the reference signal in the second window .

[0063] 26 : Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in the second window using extreme value theory and determining whether the determined energy content of the measurement signal represents an error in the measuring device.

[0064] Alternatively or additionally, the method may comprise the steps of:

[0065] 21: Determining a plurality of second windows by shifting the first window along the time axis of the measurement signal. The entire measurement path can be covered by the first window and the plurality of second windows. Two adjacent windows can overlap.

[0066] 23 : Transferring the measurement signal and the reference signal from the time domain to the frequency domain with each of the second windows.

[0067] 25 : Determining the energy content of the measurement signal and the reference signal in each of the second windows .

[0068] 27 : Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in each of the second windows by means of the extreme value theory and determination of whether the respectively determined energy content of the measurement signal represents an error of the measuring device.

[0069] The first window, the second window, and / or the second windows can each have the same start time, the same end time, and / or the same window length for the reference signal and the measurement signal, in particular, they can be identical. The transfer of the measurement signal and / or the reference signal from the time domain to the frequency domain can be carried out by means of a (windowed) short-time Fourier transformation.

[0070] The energy content of the measurement signal and / or the reference signal can be determined using Parseval's theorem.

[0071] The reference signal can be defined as normally distributed.

[0072] The method may include the step:

[0073] 28: Determining a probability with which the determined energy content of the measurement signal is an extreme value of the reference signal.

[0074] The procedure may include the steps:

[0075] 30: Specify at least one limit for the (determined) probability.

[0076] 32: Triggering an alarm signal when the (specified) limit value is reached and / or exceeded.

Claims

Patent claims 1 . A method for diagnosing a fault in a measuring device with a measuring head movable along a measuring path, comprising the steps of: ( 10 ) detecting a reference signal by traveling along the measuring section once, in particular multiple times, at a desired speed of the measuring head and detecting at least one parameter in a fault-free state of the measuring device; ( 12 ) Acquiring a measuring signal by traveling along the measuring section once, in particular multiple times, at the desired speed of the measuring head and detecting the parameter; ( 14 ) transferring the measurement signal and the reference signal from the time period into the frequency domain with a first window; ( 16 ) Determining the energy content of the measurement signal and the reference signal in the first window; ( 18 ) Probabilistic evaluation of the determined energy content of the measurement signal in relation to the energy content of the reference signal in the first window by means of extreme value theory and determination of whether the determined energy content of the measurement signal represents an error in the measuring device.

2. Method according to claim 1, characterized by the steps: ( 20 ) Determining at least one second window by shifting the first window along a time axis of the measurement signal , in particular wherein the first window and the second window overlap; (22) Transferring the measurement signal and the reference signal from the period to the Frequency space with the second window; (24) Determining the energy content of the measurement signal and the reference signal in the second window; (26) Probabilistically evaluating the determined energy content of the measurement signal with respect to the energy content of the reference signal in the second window using extreme value theory and determining whether the determined energy content of the measurement signal represents an error in the measuring device.

3. Method according to claim 2, characterized by the steps: (21) determining a plurality of second windows by shifting the first window along the time axis of the measurement signal, wherein the entire measurement path is covered by the first window and the plurality of second windows, in particular wherein two adjacent windows overlap; (23) transferring the measurement signal and the reference signal from the time domain to the frequency domain with each of the second windows; (25) determining the energy content of the measurement signal and the reference signal in each of the second windows; (27) Probabilistically evaluating the determined energy content of the measurement signal with respect to the energy content of the reference signal in each of the second windows by means of extreme value theory and determining whether the respective determined energy content of the measurement signal represents an error in the measuring device.

4. Method according to claim 2 or 3, characterized in that the first window, the second window and / or the second windows each have the same start time, the same end time and / or the same window length for the reference signal and the measurement signal, in particular are identical.

5. Method according to one of the preceding claims, characterized in that the transfer of the measurement signal and / or the reference signal from the time domain to the frequency domain is carried out by means of a short-time Fourier transformation.

6. Method according to one of the preceding claims, characterized in that the parameter is an actual speed, an actual position and / or an actual current of the measuring head.

7. Method according to one of the preceding claims, characterized in that the energy content of the measurement signal and / or the reference signal is determined by means of Parseval's theorem.

8. Method according to one of the preceding claims, characterized in that the reference signal is defined as normally distributed.

9. Method according to one of the preceding claims, characterized by the step: (28) Determining a probability with which the determined energy content of the measurement signal is an extreme value of the reference signal.

10. Method according to the preceding claim, characterized by the steps: ( 30 ) setting at least one threshold value for the probability ; (32) Triggering an alarm signal when the limit value is reached and / or exceeded.

11. Method according to one of the preceding claims, characterized in that the detection of the measurement signal is carried out during a lubrication cycle and / or during a machine start-up of the measuring device. 12 . Method according to one of the preceding claims, characterized in that the detection of the measuring signal is carried out within a maximum time of 10 seconds. 13 . Measuring device with a measuring head movable along a measuring path, characterized in that the measuring device is set up to carry out the method according to one of the preceding claims.

Images