Method and system for monitoring a device state of a device

Description

[0001] The invention relates to a method for monitoring the state of a device, in particular a device with a rotatable component, wherein a structure-borne sound signal of the device is measured. Furthermore, the invention relates to a system for monitoring the state of a device, in particular a device with a rotatable component, wherein the system comprises a structure-borne sound meter configured to measure a structure-borne sound signal of the device.

[0002] Vibration-based monitoring of a component's fixture condition can be used to detect and minimize unplanned machine downtime caused by unexpected or unforeseen failures. Conventional fixture condition monitoring systems typically require extensive expert knowledge during installation and operation. The costs and effort for the equipment, installation, and operation are typically too high to allow widespread application.

[0003] Known condition monitoring systems are usually based either on a machine-specific configuration, which requires expert knowledge and often local adaptations to enable monitoring of the machine condition, or they only use the vibration signal power, which provides only very basic information for condition monitoring and with which reliable monitoring of a device's device condition is usually not possible.

[0004] DE 10 2017 124 281 A1 describes a method for monitoring an operating state of a device, in particular a device with a rotating component, wherein a structure-borne sound signal of the device is measured and a spectrum (X) or an envelope curve of the structure-borne sound signal is determined and, in order to determine an operating state of the device, at least one spectral audio feature (M1, M2) of the determined spectrum (X) is determined.

[0005] US 6 370 957 B1 describes a method for determining the operating condition of a rotating machine, comprising monitoring the machine under a basic operating condition and collecting basic vibration data.

[0006] DE 199 45 058 A1 describes a method for determining the remaining service life of the switching contacts in an electrical switching device, wherein structure-borne sound signals of the switching contact arrangement generated by the switching process are recorded, these are then subjected to a Fourier transformation in individual time windows to generate a sonogram, and individual function values are specifically determined from the sonogram and evaluated by means of an evaluation device.

[0007] US 2010 / 0 030 492 A1 describes a method comprising receiving an input signal containing information associated with a rolling element bearing and / or a piece of equipment containing the rolling element bearing. The method further comprises decomposing the input signal into a frequency-domain signal and determining at least one frequency family corresponding to at least one failure mode of the rolling element bearing. The method further comprises generating a reconstructed input signal using the at least one frequency family and the frequency-domain signal. Furthermore, the method comprises determining an indicator that identifies the overall condition of the rolling element bearing using the reconstructed input signal.The indicator may be determined using a base signal associated with either (i) normal operation of the rolling bearing and / or piece of equipment or (ii) faulty operation of the rolling bearing and / or piece of equipment (the severity of the faulty operation increasing over time).

[0008] US 2020 / 0 149 993 A1 describes a machine bearing arranged on a rotatable element, such as may be present on a vehicle, for example. A method for monitoring the functional status of a machine bearing comprises monitoring an acoustic signal via a microphone and simultaneously determining a rotational speed of the rotatable element connected to the machine bearing. The sound spectrum is correlated with the rotational speed of the rotating element, and a time-frequency analysis is performed to determine a sound spectrum. The sound spectrum is converted into a residual spectrum. A first feature associated with a first frequency band and a second feature associated with a second frequency band are extracted from the residual spectrum. The health status of the machine bearing is detected based on the first and second features and transmitted to a second controller.

[0009] US 2018 / 0 354 088 A1 describes a machine tool configured to set a first frequency band comprising the characteristic vibration frequency of the main shaft and a second frequency band comprising the characteristic vibration frequency of the mechanical structure; to store threshold values for each of the set frequency bands; to extract vibration components for each of the frequency bands from the output of a vibration sensor; and to trigger an alarm when the vibration amplitude exceeds a threshold value in one of the frequency bands.

[0010] US 2019 / 0 203 729 A1 describes a pumping device that can more accurately detect the occurrence of abnormal vibration. The pumping device comprises: a pump; an electric motor for driving the pump; an inverter as a speed-changing means for the electric motor; a vibration detector for detecting at least one vibration of the pump, one vibration of the electric motor, and one vibration of the inverter; and a control unit for controlling the pump. The controller includes a storage unit for storing a vibration value measured by the vibration detector. The storage unit has a memory table that stores the vibration measurement values measured at each step as the speed of the pump is gradually increased to a predetermined speed.

[0011] It is an object of the present invention to provide a reliable and / or cost-effective method for monitoring the device status of a device, which preferably enables uncomplicated installation and / or operation of the monitoring system. Furthermore, it is an object to provide a corresponding system.

[0012] The object is achieved according to the invention by a method having the features of patent claim 1.

[0013] According to the invention, this makes it possible to achieve cost-effective and reliable condition monitoring, which is also particularly advantageously automated, so that the requirements for the user's specialist knowledge of the monitoring system can be kept to a minimum. This creates the possibility of applying improved condition monitoring to a variety of devices and machines for which prior art systems and methods would be too complex and uneconomical. According to the invention, it is particularly advantageous that errors or defects in the device can be detected and assessed without the need for user configuration.

[0014] Mechanical defects in equipment with rotating components can be detected by monitoring machine vibrations. Initially, defects typically manifest themselves in either a lower or upper frequency band. As the defect severity increases, the resulting excitation increases in amplitude and can also spread across wider frequency ranges.

[0015] Preferably, the device state of the device is determined and / or checked at least as a function of the first time-domain feature state and the second time-domain feature state.

[0016] According to the invention, it is conceivable for the method to be a computer-implemented method. In particular, one, several, or all steps of the method are carried out by a computer, in particular in an automated manner.

[0017] The device state determined and / or checked according to the invention indicates in particular whether the device is functioning normally or without error or whether an error and / or an indication of an error is present.

[0018] It is preferably possible for an automatic alarm to be triggered depending on the specific device condition. Thus, a particularly advantageous vibration-based condition monitoring system with automatic alarm methods can be achieved.

[0019] The rotatable component is, in particular, a component that rotates during operation of the device. Such a rotating component typically results in a measurable structure-borne sound signal. With the aid of the present invention, fault conditions in rolling bearings or gears, for example, can be detected at an early stage. This advantageously enables the timely planning of maintenance measures and reduces the likelihood of unplanned downtime.

[0020] According to the invention, it is possible, for example, to use an audio feature-based anomaly detection system that reacts to a changed sound or a changed timbre or a change in the structure-borne sound signal of the device, and furthermore to use a frequency-dependent evaluation of the structure-borne sound signal, for example the energy of the structure-borne sound signal, to assess the severity of the error.

[0021] Advantageous further developments and refinements can be found in the dependent claims.

[0022] According to one embodiment of the present invention, it is conceivable that during operation of the device -- one or more further first time-domain features are determined for the first frequency band of the structure-borne sound signal and / or -- one or more further second time-domain features are determined for the second frequency band of the structure-borne sound signal, --- wherein the further first time-domain feature(s) determined during operation of the device are each compared with an upper and / or a lower further first threshold value, so that one or more further first time-domain feature states are determined, and / or --- wherein the further second time-domain feature(s) determined during operation of the device are each compared with an upper and / or a lower further second threshold value, so that one or more further second time-domain feature states are determined, wherein the device state of the device is additionally determined and / or checked as a function of the further first time-domain feature state(s) and / or the further second time-domain feature state(s). This makes it particularly advantageous to use a plurality of first time-domain features for the first frequency band of the structure-borne sound signal and / or a plurality of second time-domain features for the second frequency band of the structure-borne sound signal for state monitoring, thereby achieving a particularly high level of reliability.

[0023] According to one embodiment of the present invention, it is conceivable that in a training phase in which the device is in particular in a predetermined good state and / or normal operation, the first time domain feature of the first frequency band of the structure-borne sound signal of the device and the second time domain feature of the second frequency band of the structure-borne sound signal of the device are determined, wherein the upper and / or lower first threshold value relating to the first time-domain feature is determined and, in particular, stored using the first time-domain feature of the first frequency band determined in the training phase, wherein the upper and / or lower second threshold value relating to the second time-domain feature is determined and, in particular, stored using the second time-domain feature of the second frequency band determined in the training phase. It is particularly advantageous for the training phase to be carried out automatically, in particular before the condition monitoring during operation of the device. This allows the threshold values to be determined automatically and, thus, particularly conveniently, so that user-friendliness can be increased and susceptibility to errors can be reduced. The threshold values can be stored, for example, in a computer memory.

[0024] According to one embodiment of the present invention, it is conceivable that in the training phase, one or more further first time-domain features of the first frequency band of the structure-borne sound signal of the device and / or one or more further second time-domain features of the second frequency band of the structure-borne sound signal of the device are determined, wherein, with the aid of the further first time-domain feature(s) determined in the training phase, an upper and / or a lower further first threshold value, in particular for each of the first time-domain features, is determined and, in particular, stored, wherein, with the aid of the further second time-domain feature(s) determined in the training phase, an upper and / or a lower further second threshold value, in particular for each of the second time-domain features, is determined and, in particular, stored. Thus, threshold values can also be determined for further time-domain features in the training phase, which can then be used in the operation of the device while monitoring the device state of the device.

[0025] According to one embodiment of the present invention, it is conceivable that the first time domain feature is one of the following features of the structure-borne sound signal relating to the first frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy, wherein in particular the further first time domain feature(s) are each one of the following features of the structure-borne sound signal relating to the first frequency band: -- mean squared deviation and / or standard deviation, -- kurtosis, -- energy.

[0026] Other time-domain features specifically designed for monitoring device states are also conceivable. The first time-domain feature and the further first time-domain feature(s) are preferably different features from one another.

[0027] According to one embodiment of the present invention, it is conceivable that the second time domain feature is one of the following features of the structure-borne sound signal relating to the second frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy, wherein in particular the further second time domain feature(s) are each one of the following features of the structure-borne sound signal relating to the second frequency band: -- mean squared deviation and / or standard deviation, -- kurtosis, -- energy.

[0028] Other time-domain features specifically designed for monitoring device states are also conceivable. The second time-domain feature and the further second time-domain feature(s) are preferably different features from one another.

[0029] According to one embodiment of the present invention, it is conceivable that the structure-borne sound signal is a time-domain signal, wherein the structure-borne sound signal comprises in particular one or more of the following signals: -- Vibration acceleration signal, -- Vibration velocity signal, -- Vibration displacement signal. This makes it possible to use advantageous structure-borne sound signals for the process, which are determined using a structure-borne sound meter on the device.

[0030] According to one embodiment of the present invention, it is conceivable that the first frequency band is a lower frequency band of the structure-borne sound signal, and the second frequency band is an upper frequency band of the structure-borne sound signal. In particular, the lower frequency band lies below the upper frequency band. It is conceivable that the lower and upper frequency bands are spaced apart from one another. Alternatively, it is conceivable that the lower and upper frequency bands are adjacent to one another.

[0031] According to one embodiment of the present invention, it is possible for the lower frequency band to comprise a frequency range from 0 Hz to 750 Hz and / or for the upper frequency band to comprise a frequency range from 750 Hz to 3000 Hz, preferably to 5000 Hz. It is particularly conceivable for the lower frequency band to extend from 0 Hz to 750 Hz and / or for the upper frequency band to extend from 750 Hz to 3000 Hz, preferably from 750 Hz to 5000 Hz. Other values for the frequency bands are also conceivable. It is conceivable that, for the specific selection of the frequency bands, technical properties of the device and / or empirical values relating to the device and / or technical restrictions are taken into account. It is conceivable for the lower frequency band and / or the upper frequency band to be fixed. Alternatively, it would be possible for the lower frequency band and / or the upper frequency band to be variable.

[0032] According to the invention, the first time domain feature state can be assigned to one of at least three categories, in particular one of the following categories: -- "normal", -- "elevated", -- "high", wherein the determined first time-domain feature state is "normal" if the first time-domain feature of the first frequency band determined during operation of the device lies below the lower and below the upper first threshold value, wherein the determined first time-domain feature state is "elevated" if the first time-domain feature of the first frequency band determined during operation of the device lies above the lower first threshold value and below the upper first threshold value, wherein the determined first time-domain feature state is "high" if the first time-domain feature of the first frequency band determined during operation of the device lies above the lower first threshold value and above the upper first threshold value, and / or wherein the second time-domain feature state can be assigned to one of at least three categories, in particular one of the following categories: -- "normal", -- "elevated", -- "high",The determined second time-domain feature state is "normal" if the second time-domain feature of the second frequency band determined during operation of the device is below the lower and below the upper second threshold. The determined second time-domain feature state is "elevated" if the second time-domain feature of the second frequency band determined during operation of the device is above the lower second threshold and below the upper second threshold. The determined second time-domain feature state is "high" if the second time-domain feature of the second frequency band determined during operation of the device is above the lower second threshold and above the upper second threshold. The designation of the categories as "normal""Increased" and "high" are to be understood merely as an example and not as a limitation. Any other designations for the categories are also conceivable in accordance with the invention.

[0033] Accordingly, according to an embodiment of the present invention in which one or more further first time domain features are used, it is possible for the further first time domain feature state(s) to be assigned to one of at least three categories, in particular to one of the following categories: -- "normal", -- "elevated", -- "high", wherein a determined further first time-domain feature state is "normal" if the corresponding further first time-domain feature of the first frequency band determined during operation of the device is below the corresponding lower further first threshold value and below the corresponding upper further first threshold value, wherein a determined further first time-domain feature state is "elevated" if the corresponding further first time-domain feature of the first frequency band determined during operation of the device is above the corresponding lower further first threshold value and below the corresponding upper further first threshold value, wherein a determined further first time-domain feature state is "high",if the corresponding further first time domain feature of the first frequency band determined during operation of the device is above the corresponding lower further first threshold value and above the corresponding upper further first threshold value.

[0034] Accordingly, according to an embodiment of the present invention in which one or more further second time domain features are used, it is possible for the further second time domain feature state(s) to be assigned to one of at least three categories, in particular to one of the following categories: -- "normal", -- "elevated", -- "high", wherein a determined further second time-domain feature state is "normal" if the corresponding further second time-domain feature of the second frequency band determined during operation of the device is below the corresponding lower further second threshold value and below the corresponding upper further second threshold value, wherein a determined further second time-domain feature state is "elevated" if the corresponding further second time-domain feature of the second frequency band determined during operation of the device is above the corresponding lower further second threshold value and below the corresponding upper further second threshold value, wherein a determined further second time-domain feature state is "high",if the corresponding further second time domain feature of the second frequency band determined during operation of the device is above the corresponding lower further second threshold value and above the corresponding upper further second threshold value.

[0035] According to one embodiment of the present invention, it is conceivable that the determined device state of the device can be assigned to one of at least four categories, wherein the categories include in particular the following state categories: -- OK and / or normal, -- Suspected error and / or suspected error, -- Warning, -- Significant warning and / or danger.

[0036] The naming of the categories for the device status is intended merely as an example and not as a limitation. Any other naming of the categories is also conceivable according to the invention.

[0037] According to one embodiment of the present invention, it is conceivable that the determined device state of the device is determined from the individual time domain feature states considered in such a way that: -- (i) if at least one of the considered time-domain feature states is "elevated", the determined device state of the device is "suspected error" and / or "presumed error"; -- (ii) if at least one of the considered time-domain feature states is "high", the determined device state of the device is "warning"; -- (iii) if at least one of the considered first and / or further first time-domain feature states of the first frequency band is "elevated" and at least one of the considered second and / or further second time-domain feature states of the second frequency band is "elevated", the determined device state of the device is "warning";-- (iv) if at least one of the considered first and / or further first time-domain feature states of the first frequency band is "high" and at least one of the considered second and / or further second time-domain feature states of the second frequency band is "high", the determined device state of the device is "significant warning" and / or "danger"; -- (v) if all considered time-domain feature states are "normal" and / or none of the preceding conditions (i) to (iv) are met, the determined device state of the device is "ok" and / or "normal"; -- (vi) if several of the preceding conditions (i) to (v) are met simultaneously, the determined device state of the device is the device state with the highest warning level resulting from the individual conditions (i) to (v).

[0038] According to one embodiment of the present invention, it is conceivable that an embodiment of the present invention, in particular a previously described embodiment of the present invention, can be combined with an additional audio feature-based anomaly detection, for example, a detection according to DE 102017 124 281 A1. It is conceivable that if an anomaly is detected according to a method of DE 102017 124 281 A1, the device state of the device is assigned at least to the category "suspected error" and / or "presumed error" (or even to the category "warning" or "significant warning and / or danger").

[0039] According to the present invention, a temperature of the device is observed and / or evaluated. For this purpose, a temperature sensor is provided that is configured to measure a temperature of the device. If the measured temperature of the device is "elevated" (in particular, above a first temperature threshold), the device status of the device is assigned at least to the category "suspected error" and / or "presumed error." If the measured temperature of the device is "high" (in particular, above a second temperature threshold that is greater than the first temperature threshold), the device status of the device is assigned at least to the category "warning."

[0040] According to one embodiment of the present invention, it is conceivable that an embodiment of the present invention, in particular a previously described embodiment of the present invention, is combined with a standard vibration signal feature for which known threshold values exist. This can be, for example, a velocity-rms (root means square) value, as defined in ISO 20816. By applying such a standardized absolute threshold value, potential difficulties in determining the lower and upper (first and second) threshold values for the first time-domain feature / second time-domain feature / further first time-domain feature / further second time-domain feature, etc. (during the training phase, if the device is not in a normal (good) state) can be particularly advantageously reduced.

[0041] According to one embodiment of the present invention, it is conceivable that an embodiment of the present invention, in particular a previously described embodiment of the present invention, is combined with rules for determining the type or nature of a defect or error. For example, if the determined device state of the device has the category "warning" or "significant warning" and / or "danger," it is conceivable that the probable type of error is determined from the individual determined time-domain feature states (first time-domain feature state, second time-domain feature state, etc.). For this purpose, it is possible to use or consider further additional information, such as the system type, when determining the type of error.

[0042] According to one embodiment of the present invention, it is possible for a user to be automatically notified depending on the determined device state, particularly in the case of a suspected error, a warning, and / or a serious warning, so that the user can initiate countermeasures. Alternatively or additionally, it is conceivable for countermeasures to be automatically initiated, for example, shutting down the device, depending on the determined device state, particularly in the case of a suspected error, a warning, and / or a serious warning. Further countermeasures are also conceivable alternatively or additionally.

[0043] Another object of the present invention is a system for monitoring a device state of a device having the features of patent claim 11.

[0044] The system comprises, in particular, circuit means or a computer configured to execute the steps of a method according to an embodiment of the present invention, preferably in an automated manner. The features, embodiments, and advantages already described in connection with the method according to the invention or in connection with an embodiment of the method according to the invention can be applied to the system according to the invention.

[0045] Further details and advantages of the invention will be explained below with reference to the exemplary embodiments illustrated in the drawings. Herein: Fig. 1 is a schematic representation of a method according to an embodiment of the present invention. Fig. 2 is a schematic representation of a system according to an embodiment of the present invention.

[0046] In Fig. 1 a schematic representation of a method according to an embodiment of the present invention is shown.

[0047] With the aid of a structure-borne sound meter 2, data recording 100 takes place during operation of the device 1, in which a structure-borne sound signal 200 (or vibration signal) relating to a device 1 is determined.

[0048] In a feature extraction 110, at least one first time-domain feature F1L and preferably one or more further first time-domain features F2L, ..., FnL are determined for a first frequency band of the structure-borne sound signal 200. "F1L, ..., FnL" denote a total of "n" first time-domain features, where "n" is a natural number ("n" can be, for example, 2, 3, 4, 5, 6, 7, etc.). Furthermore, in the feature extraction 110, at least one second time-domain feature F1H and preferably one or more further second time-domain features F2H, ..., FmH are determined for a second frequency band of the structure-borne sound signal 200. "F1H, ..., FmH" denote a total of "m" second time-domain features, where "m" is a natural number ("m" can be, for example, 2, 3, 4, 5, 6, 7, etc.).

[0049] In a feature state evaluation 120, the determined first time-domain features F1L, ..., FnL of the first frequency band are each compared with an upper and a lower first threshold value. This comparison determines a corresponding separate first time-domain feature state SF1L, ..., SFnL for each of the considered first time-domain features F1L, ..., FnL.

[0050] Furthermore, in the feature state evaluation 120, the determined second time-domain features F1H, ..., FmH of the second frequency band are each compared with an upper and a lower second threshold value. This comparison determines a corresponding second time-domain feature state SF1H, ..., SFmH for each of the considered second time-domain features F1H, ..., FmH.

[0051] In a device state evaluation 130, the device state 140 of the device 1 is then determined from the determined time-domain feature states SF1L, ..., SFnL, SF1H, ..., SFmH.

[0052] According to one embodiment of the present invention, a method according to the invention may comprise the following steps.

[0053] Two sets of one or more time-domain features F1L, ..., FnL and F1H, ..., FmH are defined, each for a lower and an upper frequency band of a time-domain vibration signal (or structure-borne sound signal 200), which are used to determine a device condition 140 of a device 1. The time-domain features F1L, ..., FnL and F1H, ..., FmH can, for example, include a standard deviation, a kurtosis, or other features specifically developed for vibration-based condition monitoring systems. The time-domain vibration signal (or structure-borne sound signal 200) can, for example, be an acceleration signal or a velocity signal, which is determined, for example, using a structure-borne sound meter 2 on the device 1.

[0054] Baseline values for the considered time-domain features are preferably determined in an initial training phase, during which the device 1 is in an assumed good state and in operation. From these baseline values, an upper and a lower threshold value are determined for each of the considered time-domain features F1L, ..., FnL and F1H, ..., FmH (and in particular for each frequency band).

[0055] By monitoring the considered time domain features F1L, ..., FnL and F1H, ..., FmH during the further operation of the device 1, in particular after the training phase has been completed, a time domain feature state SF1L, ..., SFnL and SF1H, ..., SFmH can be determined for each of the considered time domain features F1L, ..., FnL and F1H, ..., FmH: "Normal" - the value of the measured time-domain characteristic F1L, ..., FnL and F1H, ..., FmH is below the lower and below the upper threshold; "Elevated" - the value of the measured time-domain characteristic F1L, ..., FnL and F1H, ..., FmH is between the lower and upper threshold; "High" - the value of the measured time-domain characteristic F1L, ..., FnL and F1H, ..., FmH is above both the lower and above the upper threshold. Statistical evaluation methods can be applied to minimize the influence of random outliers and / or noise when determining the time-domain characteristics F1L, ..., FnL and F1H, ..., FmH during operation of the device 1. A variety of evaluation methods are possible.

[0056] The device state 140 of the device 1 is obtained from the individual determined time-domain feature states SF1L, ..., SFnL and SF1H, ..., SFmH of the considered time-domain features F1L, ..., FnL and F1H, ..., FmH, in particular in accordance with the (previously described) evolutionary behavior with respect to the amplitude and frequency range of the excitations: Initially, defects typically appear in either a lower or an upper frequency band. With increasing defect severity, the excitation resulting from the defect increases in amplitude and can also spread across additional frequency ranges.

[0057] The determined device state 140 of the device 1 can be specified, for example, using the following categories or classes: -- "ok" and / or "normal", -- "suspected error" and / or "presumed error", -- "warning", -- "significant warning" and / or "danger".

[0058] According to one embodiment, the determined device state 140 of the device 1 is obtained from the individual considered time domain feature states SF1L, ..., SFnL, SF1H, ..., SFmH as follows: -- If at least one of the considered time-domain feature states is "elevated," the determined device state 140 of device 1 is "suspected error" and / or "presumed error." -- If at least one of the considered time-domain feature states is "high," the determined device state 140 of device 1 is "warning." -- If at least one of the considered first (or further first) time-domain feature states of the first (particularly lower) frequency band is "elevated," and at least one of the considered second (or further second) time-domain feature states of the second (particularly upper) frequency band is "elevated," the determined device state 140 of device 1 is "warning." -- If at least one of the considered first (or further first) time-domain feature states of the first (in particular lower) frequency band is "high" and at least one of the considered second (or further first) time-domain feature states of the first (in particular lower) frequency band is "high" andfurther second) time-domain feature states of the second (in particular, upper) frequency band is "high," the determined device state 140 of device 1 is "significant warning" and / or "danger." -- If all considered time-domain feature states are "normal" or none of the above conditions are met, the determined device state 140 of device 1 is "OK" and / or "normal." -- If several of the above conditions are met simultaneously, the determined device state 140 of device 1 is the one with the highest warning level resulting from the individual conditions.

[0059] It is possible that, depending on the determined device state 140, in particular in the case of a suspected error, a warning, and / or a serious warning, a notification is automatically sent to a user so that the user can initiate countermeasures. Alternatively or additionally, it is conceivable that, depending on the determined device state 140, in particular in the case of a suspected error, a warning, and / or a serious warning, countermeasures are automatically initiated, for example, switching off the device 1.

[0060] In Fig. 2A schematic representation of a system according to an embodiment of the present invention is shown. A structure-borne sound meter 2 is attached and / or arranged on a device 1 with a rotatable component 1'. The structure-borne sound meter 2 is designed to detect a structure-borne sound signal 200. The structure-borne sound signal 200 is provided to an evaluation device 3, in particular a computer, wherein the evaluation device 3 is configured to determine the time-domain features F1L, ..., FnL, F1H, ..., FmH, time-domain feature states SF1L, ..., SFnL, SF1H, ..., SFmH and the device state 140 of the device 1. In particular, the feature extraction 110, feature state evaluation 120 and / or device state evaluation 130 are performed by the evaluation device 3. It is conceivable that the evaluation device 3 is installed together with the structure-borne sound meter 2 or separately from the structure-borne sound meter 2. List of reference symbols

[0061] 1Device 1'Rotatable component 2Structural sound meter 3Evaluation device 100Data acquisition 110Feature extraction 120Feature state evaluation 130Device state evaluation 140Device state 200Structural sound signal F1LFirst time-domain feature F2L, ..., FnLFurther first time-domain features F1HzSecond time-domain feature F2H, ..., FmHFurther second time-domain features SF1LFirst time-domain feature state SF2L, ..., SFnLFurther first time-domain feature states SF1HzSecond time-domain feature state SF2H, ..., SFmHFurther second time-domain feature states

Claims

1. A method for monitoring the state of a device (1), in particular a device (1) with a rotatable component (1'), wherein a structure-borne sound signal (200) of the device (1) is measured, characterised in that during operation of the device (1) -- at least one first time-domain characteristic (F1L) is determined for a first frequency band of the structure-borne sound signal (200), and -- at least one second time-domain characteristic (F1H) is determined for a second frequency band of the structure-borne sound signal (200), wherein the first time-domain characteristic (F1L) of the first frequency band determined during operation of the device (1) is compared with an upper and a lower first threshold value, such that a first time-domain characteristic state (SF1L) is determined, wherein the second time-domain characteristic (F1H) of the second frequency band determined during operation of the device (1) is compared with an upper and a lower second threshold value, such that a second time-domain characteristic state (SF1H) is determined, wherein the state (140) of the device (1) is determined and / or checked at least as a function of the first time-domain characteristic state (SF1L) and / or the second time-domain characteristic state (SF1H), wherein the first time-domain characteristic state (SF1L) can be assigned to one of at least three categories, namely at least one of the categories: - "normal", - "increased", - "high", wherein the determined first time-domain characteristic state (SF1L) is "normal" if the first time-domain characteristic (F1L) of the first frequency band determined during operation of the device (1) is below the lower and below the upper first threshold value, - wherein the determined first time-domain characteristic state (SF1L) is "increased" if the first time-domain characteristic (F1L) of the first frequency band determined during operation of the device (1) is above the lower first threshold value and below the upper first threshold value, - wherein the determined first time-domain characteristic state (SF1L) is "high" if the first time-domain characteristic (F1L) of the first frequency band determined during operation of the device (1) is above the lower first threshold value and above the upper first threshold value, and / or wherein the second time-domain characteristic state (SF1H) can be assigned to one of at least three categories, namely at least one of the following categories: - "normal", - "increased", - "high", wherein the determined second time-domain characteristic state (SF1H) is "normal" if the second time-domain characteristic (S1H) of the second frequency band determined during operation of the device (1) is below the lower and below the upper second threshold value, - wherein the determined second time-domain characteristic state (SF1H) is "increased" if the second time-domain characteristic (S1H) of the second frequency band determined during operation of the device (1) is above the lower second threshold value and below the upper second threshold value, - wherein the determined second time-domain characteristic state (SF1H) is "high" if the second time-domain characteristic (S1H) of the second frequency band determined during operation of the device is above the lower second threshold value and above the upper second threshold value, characterised in that a temperature gauge is provided, which is configured to measure a temperature of the device (1) and the temperature of the device (1) is measured with the temperature gauge, wherein the state of the device (1) is assigned to at least one category of "suspected fault" and / or "presumed fault" if the measured temperature is above a first temperature threshold value.

2. The method according to claim 1, wherein the state of the device (1) is assigned to at least one category of "warning" if the measured temperature is above a second temperature threshold value greater than the first temperature threshold value.

3. The method according to one of the preceding claims, wherein during operation of the device (1) -- one or a plurality of further first time-domain characteristics (F2L, ..., FnL) are determined for the first frequency band of the structure-borne sound signal (200) and / or -- one or a plurality of further second time-domain characteristics (F2H, ..., FmH) are determined for the second frequency band of the structure-borne sound signal (200), --- wherein the further first time-domain characteristic(s) (F2L, ..., FnL) determined during operation of the device (1) are each compared with an upper and / or a lower further first threshold value, such that one or a plurality of further first time-domain characteristic states (SF2L, ..., SFnL) are determined, and / or --- wherein the further second time-domain characteristic(s) (F2H, ..., FmH) determined during operation of the device (1) are each compared with an upper and / or a lower further second threshold value, such that one or a plurality of further second time-domain characteristic states (SF2H, ..., SFmH) are determined, wherein the state (140) of the device (1) is additionally determined and / or checked as a function of the further first time-domain characteristic state(s) (SF2L, ..., SFnL) and / or the further second time-domain characteristic state(s) (F2H, ..., FmH).

4. The method according to one of the preceding claims, wherein, in a training phase in which the device (1) is in particular in a predetermined good state and / or normal operation, the first time-domain characteristic (F1L) of the first frequency band of the structure-borne sound signal (200) of the device and the second time-domain characteristic (F1H) of the second frequency band of the structure-borne sound signal (200) of the device (1) are determined, wherein the first time-domain characteristic (F1L) of the first frequency band determined in the training phase is used to determine and in particular store the upper and / or the lower first threshold value relating to the first time-domain characteristic (F1L), wherein the second time-domain characteristic (S1H) of the second frequency band determined in the training phase is used to determine and in particular store the upper and / or the lower second threshold value relating to the second time-domain characteristic (F1H).

5. The method according to claim 4, wherein in the training phase one or a plurality of further first time-domain characteristics (F2L, ..., FnL) of the first frequency band of the structure-borne sound signal (200) of the device (1) and / or one or a plurality of further second time-domain characteristics (F2H, ..., FmH) of the second frequency band of the structure-borne sound signal (200) of the device (1) are determined, wherein an upper and / or a lower further first threshold value, in particular for each of the first time-domain characteristics (F2L, ..., FnL), are determined and in particular stored with the aid of the further first time-domain characteristic(s) (F2L, ..., FnL) determined in the training phase, wherein an upper and / or a lower further second threshold value, in particular for each of the second time-domain characteristics (F2H, ..., FmH), are determined and in particular stored with the aid of the further second time-domain characteristic(s) (F2H, ..., FmH) determined in the training phase.

6. The method according to one of the preceding claims, wherein the first time-domain characteristic (F1L) is one of the following characteristics of the structure-borne sound signal (200) relating to the first frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy, wherein in particular the further first time-domain characteristic(s) (F2L, ..., FnL) are each one of the following features of the structure-borne sound signal (200) relating to the first frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy.

7. The method according to one of the preceding claims, wherein the second time-domain characteristic (F1H) is one of the following characteristics of the structure-borne sound signal (200) relating to the second frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy, wherein in particular the further second time-domain characteristic(s) (F2H, ..., FmH) are each one of the following features of the structure-borne sound signal (200) relating to the second frequency band: -- mean square deviation and / or standard deviation, -- kurtosis, -- energy.

8. The method according to one of the preceding claims, wherein the structure-borne sound signal (200) is a time-domain signal, wherein the structure-borne sound signal (200) comprises in particular one or a plurality of the following signals: -- vibration acceleration signal, -- vibration speed signal, -- vibration deflection signal.

9. The method according to one of the preceding claims, wherein the first frequency band is a lower frequency band of the structure-borne sound signal (200) and wherein the second frequency band is an upper frequency band of the structure-borne sound signal (200).

10. The method according to one of the preceding claims, wherein the determined state (140) of the device (1) is determined from the individual considered time-domain characteristic states (SF1L, ..., SFnL, SF1H, ..., SFmH) such that: -- (i) if at least one of the considered time-domain characteristic states (SF1L, ..., SFnL, SF1H, ..., SFmH) is "increased", the determined state (140) of the device (1) is "suspected fault" and / or "presumed fault"; -- (ii) if at least one of the considered time-domain characteristic states (SF1L, ..., SFnL, SF1H, ..., SFmH) is "high", the determined state (140) of the device (1) is "warning"; -- (iii) if at least one of the considered first and / or further first time-domain characteristic states (SF1L, ..., SFnL) of the first frequency band is "increased" and at least one of the considered second and / or further second time-domain characteristic states (SF1H, ..., SfmH) of the second frequency band is "increased" the determined state (140) of the device (1) is "warning"; -- (iv) if at least one of the considered first and / or further first time-domain characteristic states (SF1L, ..., SFnL) of the first frequency band is "high" and at least one of the considered second and / or further second time-domain characteristic states (SF1H, ..., SfmH) of the second frequency band is "high" the determined state (140) of the device (1) is "significant warning" and / or "danger"; -- (v) if all the considered time-domain characteristic states (SF1L, ..., SFnL, SF1H, ..., SFmH) are "normal" and / or none of the preceding conditions (i) to (iv) is fulfilled, the determined state (140) of the device (1) is "o.k." and / or "normal"; -- (vi) if a plurality of the preceding conditions (i) to (v) are simultaneously fulfilled, the determined state (140) of the device (1) is the state (140) with the highest warning level resulting from the individual conditions (i) to (v).

11. A system for monitoring the state of a device (1) with a rotatable component (1'), wherein the system is suitable for performing a method according to one of the preceding claims, and comprises a structure-borne sound sensor (2) configured to measure a structure-borne sound signal (200) of the device (1), wherein - the system is configured such that: during operation of the device (1) -- at least one first time-domain characteristic (F1L) is determined for a first frequency band of the structure-borne sound signal (200), and -- at least one second time-domain characteristic (F1H) is determined for a second frequency band of the structure-borne sound signal (200), - wherein the system is configured such that: -- the first time-domain characteristic (F1L) of the first frequency band determined during operation of the device (1) is compared with an upper and / or a lower first threshold value, such that a first time-domain characteristic state (SF1L) is determined, and -- the second time-domain characteristic (F1H) of the second frequency band determined during operation of the device is compared with an upper and / or a lower second threshold value, such that a second time-domain characteristic state (SF1H) is determined, - wherein the system is configured such that the state (140) of the device (1) is determined and / or checked at least as a function of the first time-domain characteristic state (SF1L) and / or the second time-domain characteristic state (SF1H), characterised in that the system comprises a temperature gauge, which is configured to measure a temperature of the device (1), wherein the system is configured such that the state of the device (1) is assigned to at least one category of "suspected fault" and / or "presumed fault" if the measured temperature is above a first temperature threshold value.