Condition monitoring device and condition monitoring method

Abstract

This reduces the user's processing burden in setting at least one of the frequency range and time range used for monitoring the condition of bearings. [Solution] The condition monitoring device 100 displays a spectrogram image based on a spectrogram on a display device and accepts the user's specification of the frequency range and time range in the spectrogram image. Based on the spectrogram of the frequency range and time range specified by the user, the condition monitoring device 100 detects whether or not there is a bearing abnormality.

Description

Technical Field

[0006]

[0001] The present disclosure relates to a state monitoring device and a state monitoring method.

Background Art

[0002] For example, Japanese Unexamined Patent Application Publication No. 2020-056771 (Patent Document 1) discloses a state monitoring device that detects abnormalities in a bearing. This state monitoring device extracts data in a specific frequency band from the vibration data of the bearing using a band-pass filter. This state monitoring device calculates a feature amount based on the extracted data to detect an abnormality in the bearing.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0007] The bearing condition monitoring method of this disclosure comprises acquiring a detection signal relating to the vibration of the bearing. The condition monitoring method comprises generating a three-dimensional signal from a first target signal based on the detection signal. The condition monitoring method comprises displaying a three-dimensional image based on the three-dimensional signal on a display device. The condition monitoring method comprises receiving a specification from a user for at least one of a frequency range and a time range in the three-dimensional image displayed on the display device. The condition monitoring method comprises performing a predetermined process relating to the detection of an abnormality in the bearing based on the three-dimensional signal of the frequency range specified by the user. [Effects of the Invention]

[0008] According to this disclosure, the user's processing burden for setting at least one of the frequency range and time range used for monitoring the condition of bearings can be reduced. [Brief explanation of the drawing]

[0009] [Figure 1] This is a diagram showing an example configuration of a status monitoring system. [Figure 2] This is a functional block diagram of the state monitoring device according to the first embodiment. [Figure 3] This figure shows an example of a spectrogram image. [Figure 4] This figure shows an example of a frequency spectrum. [Figure 5]It is a flowchart showing the processing of the state monitoring device. [Figure 6] It is a diagram for explaining an impulse signal. [Figure 7] It is a functional block diagram of the state monitoring device of the second embodiment. [Figure 8] It is a functional block diagram of the pulse reduction processing unit. [Figure 9] It is a diagram for explaining pulse reduction processing. [Figure 10] It is a functional block diagram of the state monitoring device of the third embodiment. [Figure 11] It is a diagram for explaining HPSS (Harmonic / Percussive Sound Separation) processing. [Figure 12] It is a functional block diagram of the state monitoring device of the fourth embodiment. [Figure 13] It is a diagram for explaining the method of specifying a user in the first example. [Figure 14] It is a diagram for explaining the method of specifying a user in the second example. [Figure 15] It is a diagram for explaining the method of specifying a user in the third example. [Figure 16] It is a functional block diagram of the state monitoring device of the seventh embodiment. [Figure 17] It is a diagram showing a configuration example of the past analysis DB. [Figure 18] It is a diagram showing an example of the recommended range. [Figure 19] It is a diagram showing another display example of the recommended range. [Figure 20] It is a diagram showing another display example of the recommended range. [Figure 21] It is a diagram showing another display example of the recommended range. [Figure 22] It is a diagram showing an example of the monitoring system. [Figure 23] It is a functional block diagram of the state monitoring device of another embodiment. [Figure 24] It is a functional block diagram of the state monitoring device of another embodiment. [[ID=�8]]

Embodiments for Carrying Out the Invention

[0010] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. In the embodiments described below, when referring to the number, amount, etc., unless otherwise specified, the scope of the present disclosure is not necessarily limited to such number, amount, etc. The same parts and corresponding parts are assigned the same reference numerals, and repeated descriptions may not be repeated. It is initially planned to use the configurations in the embodiments in appropriate combinations.

[0011] <First Embodiment> [Condition Monitoring System] FIG. 1 is a diagram showing a configuration example of a monitoring system 300. The monitoring system 300 includes a bearing 10, a sensor 50, a condition monitoring device 100, a display device 150, and an input device 160.

[0012] In the example of FIG. 1, the bearing 10 includes a fixed ring 11, a rotating ring 12, and rolling elements 13. The bearing 10 is incorporated into a rotating machine. The rotating machine is, for example, a wind power generation device.

[0013] The sensor 50 detects parameters that enable the condition monitoring device 100 to detect abnormalities in the bearing 10. This parameter is, for example, a parameter that enables the condition monitoring device 100 to perform frequency analysis. This parameter is, for example, a vibration parameter related to the vibration that occurs when the bearing 10 to be monitored is operated. The vibration parameter is, for example, the acceleration of the vibration of the bearing 10.

[0014] The sensor 50 is any one of a vibration sensor, a pressure sensor, an AE (Acoustic Emission) sensor, a current sensor, and a microphone. The sensor 50 outputs a detection signal indicating the detected vibration parameter to the condition monitoring device 100.

[0015] The display device 150 displays an image based on image data from the condition monitoring device 100. The user of the condition monitoring device 100 performs various processes while viewing the image. For example, if the display device 150 displays an image indicating that a bearing 10 malfunction has occurred, the user of the condition monitoring device 100 notifies the wind power generation equipment operator that a bearing 10 malfunction has occurred.

[0016] The input device 160 is a user interface that receives input from the user. The input device 160 receives the user's specification of the frequency range in the spectrogram image, which will be described later.

[0017] The input device 160 may be a manually operated device such as a mouse or keyboard. Alternatively, the input device 160 may be a device that accepts user voice input (for example, a microphone). Furthermore, the display device 150 and the input device 160 may be integrated to form a touch panel.

[0018] The condition monitoring device 100 detects abnormalities in the bearing 10 based on the detection signal. The condition monitoring device 100 includes a CPU (Central Processing Unit) 202, a memory 204, and an interface 206. The CPU 202 corresponds to the “control device” in this disclosure. The control device may also be called a control circuit.

[0019] Memory 204 includes ROM (Read Only Memory) and RAM (Random Access Memory). ROM is non-rewritable, non-volatile memory, while RAM is volatile memory.

[0020] The ROM stores a program that describes the processing procedure for the CPU 202. The CPU 202 loads the program stored in the ROM into RAM or other memory and executes it. The interface 206 receives detection signals from the sensor 50, transmits image data of the image to be displayed on the display device 150, and accepts user input from the input device 160.

[0021] The processing performed by the status monitoring device 100 includes both actions performed by the status monitoring device 100 itself and actions performed jointly by the status monitoring device 100 and at least one information processing device (not shown).

[0022] [Functional Block Diagram] Figure 2 is a functional block diagram of the condition monitoring device 100 of the first embodiment. The condition monitoring device 100 performs predetermined processing related to the detection of abnormalities in the bearing 10. The predetermined processing in the first embodiment is to calculate a feature quantity for detecting abnormalities in the bearing 10 and to detect whether or not there is an abnormality in the bearing 10 based on the feature quantity. In the first embodiment, the feature quantity is a frequency corresponding to an abnormality in the bearing 10.

[0023] The condition monitoring device 100 includes an acquisition unit 102, an STFT (short-time Fourier transform) processing unit 104, a spectrogram display unit 106, a range processing unit 108, a BPF (band-pass filter) processing unit 110, an envelope processing unit 112, an FFT processing unit 114, and an analysis unit 116.

[0024] The acquisition unit 102 acquires a detection signal from the sensor 50. The detection signal is output to the STFT processing unit 104 and the BPF processing unit 110. The STFT processing unit 104 generates a spectrogram by performing a short-time Fourier transform on the first target signal based on the detection signal. The spectrogram corresponds to the "three-dimensional signal showing the time, frequency, and vibration intensity of the vibration of the bearing 10" as described in this disclosure. In the first embodiment, the first target signal is the detection signal. The spectrogram is input to the spectrogram display unit 106 and the BPF processing unit 110.

[0025] In the first embodiment, the operation performed on the detection signal is a short-time Fourier transform, but other operations may be performed. For example, a wavelet transform may be performed. When a wavelet transform is performed, a scalogram is generated as a three-dimensional signal.

[0026] The spectrogram display unit 106 displays a spectrogram image based on the spectrogram on the display device 150. The spectrogram image will be described later.

[0027] The range processing unit 108 acquires the time range and frequency range of the spectrogram image from the user specified by the input device 160. The frequency range is output to the BPF processing unit 110. The range processing unit 108 also extracts the time range component (signal) specified by the user from the detection signal from the acquisition unit 102 and outputs the extracted component to the BPF processing unit 110. The detection signal from which the time range component has been extracted by the range processing unit 108 corresponds to the "second target signal" of this disclosure.

[0028] The BPF processing unit 110 performs filtering on signals within a time range specified by the user, allowing them to pass through a frequency band within the frequency range from the range processing unit 108. The filtered signal is output to the envelope processing unit 112.

[0029] The envelope processing unit 112 performs envelope processing on the input signal. The envelope-processed signal is output to the FFT processing unit 114. The FFT processing unit 114 performs a Fourier transform on the envelope-processed signal. The frequency spectrum obtained by this Fourier transform is output to the analysis unit 116.

[0030] The analysis unit 116 analyzes the damaged areas of the bearing 10 based on the damage frequency and frequency spectrum corresponding to each damaged area of ​​the bearing 10. For example, the analysis unit 116 analyzes the damaged areas of the bearing 10 by comparing the damage frequency and the frequency spectrum.

[0031] The analysis unit 116 then displays the analysis results on the display device 150. For example, if the analysis unit 116 identifies damage to the rolling element 13, the analysis unit 116 displays an image on the display device 150 indicating that the rolling element 13 is damaged. Note that the processing of the envelope processing unit 112, the processing of the FFT processing unit 114, and the processing of the analysis unit 116 in Figure 2 correspond to the "predetermined processing" in this disclosure.

[0032] [Spectrogram image] Figure 3 shows an example of a three-dimensional spectrogram image 230 displayed on the display device 150 by the spectrogram display unit 106. In the spectrogram image 230 in Figure 3, the X axis (first axis) is the time axis, and the Y axis (first axis orthogonal to the second axis) is the frequency axis. In addition, the amplitude (vibration intensity) of the spectrogram is represented by color in the spectrogram image 230 in Figure 3.

[0033] The user can specify a region 250 on the spectrogram image 230 displayed on the display device 150 using the input device 160. The user can perform various operations on the spectrogram image 230, such as zooming in, zooming out, and scrolling, using the input device 160. The display device 150 displays an image representing the region 250 superimposed on the spectrogram image 230. The user can modify the specified region 250. In the first embodiment, the region 250 is rectangular.

[0034] The range processing unit 108 identifies a range as a frequency band, which is comprised of a lower limit corresponding to a pixel with the minimum frequency F1 and an upper limit corresponding to a pixel with the maximum frequency F2, as indicated by the specified region 250. The range processing unit 108 also identifies a range as a time range, which is comprised of a lower limit corresponding to a pixel with the minimum time T1 and an upper limit corresponding to a pixel with the maximum time T2, as indicated by the specified region 250.

[0035] Figure 4 is an example of a frequency spectrum obtained when the FFT processing unit 114 performs FFT processing on the signal specified by the user-specified region 250 in Figure 3. The status monitoring device 100 may display the image in Figure 4 together with the image in Figure 3, or it may choose not to display the image in Figure 4.

[0036] In summary, with conventional condition monitoring devices, users are required to manually input the upper and lower limits of the bandpass filter's frequency band, which can lead to the problem of cumbersome frequency band settings.

[0037] In contrast, the condition monitoring device 100 of the first embodiment accepts a specification of a frequency range from the user in the spectrogram image 230 (see Figure 3) displayed on the display device 150. The condition monitoring device 100 then detects an abnormality in the bearing 10 based on the spectrogram of this frequency range. Therefore, the user does not need to manually input numerical values ​​for the upper and lower limits of the bandpass filter frequency band, and can intuitively set the frequency band. Thus, the condition monitoring device 100 can reduce the processing burden on the user in setting the frequency band.

[0038] Furthermore, the condition monitoring device 100 also accepts the user's specification of a time range in the spectrogram image 230 displayed on the display device 150. The condition monitoring device 100 then detects abnormalities in the bearing 10 based on the spectrogram for the time range and frequency range specified by the user. Therefore, the condition monitoring device 100 can reduce the processing burden on the user in setting the time range used for monitoring the condition of the bearing 10.

[0039] [flowchart] Figure 5 is a flowchart showing the processing of the state monitoring device 100. First, in step S2, the state monitoring device 100 acquires a detection signal from the sensor 50. Next, in step S4, the state monitoring device 100 generates a spectrogram by performing a short-time Fourier transform on the detection signal.

[0040] Next, in step S6, the status monitoring device 100 displays a spectrogram image based on the spectrogram on the display device 150. Next, in step S8, the status monitoring device 100 determines whether or not the user has specified area 250. The status monitoring device 100 repeats the process in step S8 until the user has specified area 250.

[0041] When the user specifies area 250 (YES in step S8), in step S10, the status monitoring device 100 performs an abnormality detection process (a process to detect whether or not there is an abnormality) in response to the signal specified by the user.

[0042] <Second Embodiment> The detection signal acquired by the acquisition unit 102 in Figure 2 may include an impulse signal (pulse signal). Figure 6 is a diagram illustrating an impulse signal and the Fourier transform applied to said impulse signal. Figure 6(A) is a diagram illustrating an impulse signal. In the example in Figure 6(A), an impulse signal with infinite intensity is shown. The frequency bandwidth of this impulse signal spans a wide range. Therefore, when FFT processing is performed on this impulse signal, a wide-ranging frequency spectrum is generated, as shown in Figure 6(B).

[0043] Therefore, when the FFT processing unit 114 performs FFT processing on a detection signal that includes an impulse signal, the wide range of frequency components mentioned above may be added to the frequency peak indicating an abnormality in the bearing 10, potentially reducing the analytical accuracy of the analysis unit 116.

[0044] Figure 7 is a functional block diagram of the status monitoring device 100A of the second embodiment. The status monitoring device 100A has a pulse reduction processing unit 120 between the acquisition unit 102 and the STFT processing unit 104.

[0045] The pulse reduction processing unit 120 performs a reduction process to reduce the pulses contained in the detection signal. In the second embodiment, the first target signal is the signal to which the reduction process has been performed on the detection signal. The detection signal with reduced pulses is then output to the STFT processing unit 104 and the BPF processing unit 110. In this way, since the pulse reduction processing unit 120 performs a reduction process to reduce the pulses contained in the detection signal, the accuracy of the abnormality detection process of the bearing 10 can be improved.

[0046] Next, an example of the reduction process will be described. In the second embodiment, the pulse reduction processing unit 120 reduces the pulse by converting the impulse signal into white noise. Figure 8 is a functional block diagram of the pulse reduction processing unit 120. The pulse reduction processing unit 120 includes a calculation unit 1202, a setting unit 1204, and a replacement unit 1206.

[0047] The detection signal from the acquisition unit 102 is output to the calculation unit 1202. The calculation unit 1202 calculates the average amplitude A and variance σ of the detection signal for a time range specified by the user (see Figure 3).

[0048] Next, the setting unit 1204 determines the upper and lower limits of the first range based on the variance σ. The first range is the range that includes the average amplitude A. As shown in equation (1) below, the setting unit 1204 determines the upper limit of the first range by adding the value obtained by multiplying the first predetermined value by the variance σ to the average amplitude A. As shown in equation (2) below, the setting unit 1204 determines the lower limit of the first range by subtracting the value obtained by multiplying the first predetermined value by the variance σ from the average amplitude A.

[0049] Upper limit of the first range = Mean amplitude A + First predetermined value × Variance σ (1) Lower limit of the first range = Mean amplitude A - First predetermined value × Variance σ (2) The first predetermined value is, for example, "3".

[0050] Furthermore, the setting unit 1204 determines the upper and lower limits of the second range based on the variance. The second range is narrower than the first range and includes the amplitude mean value A. The setting unit 1204 determines the upper limit of the second range by adding the value obtained by multiplying the second predetermined value by the variance σ to the amplitude mean value A, as shown in equation (3) below. However, the second predetermined value is a smaller value than the first predetermined value, for example, "1". The setting unit 1204 determines the lower limit of the second range by subtracting the value obtained by multiplying the second predetermined value by the variance σ from the amplitude mean value A, as shown in equation (4) below.

[0051] Upper limit of the second range = Mean amplitude A + Second predetermined value × Variance σ (3) Lower limit of the second range = Mean amplitude A - Second predetermined value × Variance σ (4) The replacement unit 1206 then changes the data (amplitude) of the detection signal that does not belong to the first range to a random value so that it belongs to the second range. This change to a random value is performed, for example, using a random number generator. The replacement unit 1206 also converts extremely large amplitudes (amplitudes that do not belong to the first range) of the detection signal, such as impulse signals, so that they belong to the second range. Therefore, the pulse reduction processing unit 120 can reduce the pulses contained in the detection signal.

[0052] Figure 9 illustrates the effect of pulse reduction processing. Figure 9(A) is the detection signal containing pulse P. Figure 9(B) is the frequency spectrum after FFT is performed on the detection signal containing pulse P in Figure 9(A). Figure 9(C) is the detection signal with pulse P reduced. Figure 9(D) is the frequency spectrum after FFT is performed on the detection signal (signal with pulse P reduced).

[0053] In the frequency spectrum of example 9(D), peak frequency A is present. On the other hand, in the frequency spectrum of example 9(B), peak frequency A, which should be present, is not present.

[0054] Since the condition monitoring device 100A displays a spectrogram image with reduced pulse component P, it makes it easier for the user to specify region 250. Furthermore, since the condition monitoring device 100A performs abnormality detection processing of the bearing 10 based on the signal with reduced pulse component P, it can improve the accuracy of abnormality detection of the bearing 10.

[0055] <Third Embodiment> Figure 10 is a functional block diagram of the condition monitoring device 100B of the third embodiment. The condition monitoring device 100B has an HPSS (Harmonic Percussion Source Separation) processing unit 122 between the STFT processing unit 104 and the spectrogram display unit 106.

[0056] The HPSS processing unit 122 performs HPSS on the spectrogram to separate it into a percussion sound signal corresponding to the percussion sound and a harmonic sound signal corresponding to the harmonic sound. The abnormal signal component corresponding to the bearing 10 abnormality is highly correlated with the component of the percussion sound signal, and the related signal associated with the bearing 10 abnormality is the said percussion sound signal. On the other hand, the abnormal signal component is less correlated with the component of the harmonic sound signal.

[0057] Therefore, the HPSS processing unit 122 extracts the percussion sound signal. The percussion sound signal corresponds to the “relevant signal” in this disclosure. The percussion sound signal is a signal in the frequency domain.

[0058] The percussion sound signal is output to the spectrogram display unit 106 and the inverse STFT processing unit 180. The inverse STFT processing unit 180 generates a time-domain percussion sound signal by performing an inverse STFT on the frequency-domain percussion sound signal. The time-domain percussion sound signal corresponds to the "second target signal" of this disclosure.

[0059] The range processing unit 108 extracts signals within a time range specified by the user from the time-domain percussion sound signal and outputs the extracted signals to the BPF processing unit 110. The spectrogram display unit 106 displays an image based on the percussion sound signal as a spectrogram image. The BPF processing unit 110 also performs bandpass filtering on the percussion sound signal from which the time-range components have been extracted from the inverse STFT processing unit 180.

[0060] Figure 11 is a diagram illustrating the HPSS processing. Figure 11(A) is the spectrogram from the STFT processing unit 104. Figure 11(B) is the spectrogram of the harmonic sound signal. Figure 11(C) is the spectrogram of the percussion sound signal. The HPSS processing unit 122 extracts the spectrogram of the percussion sound signal in Figure 11(C) from the spectrogram in Figure 11(A), and outputs the extracted spectrogram of the percussion sound signal to the spectrogram display unit 106 and the BPF processing unit 110.

[0061] As described above, the condition monitoring device 100B displays a spectrogram image from which signals with little correlation to bearing abnormalities (harmonic signals) have been excluded, making it easier for the user to specify region 250. Furthermore, since the condition monitoring device 100B performs bearing abnormality detection processing using the signal from which signals with little correlation to bearing abnormalities (harmonic signals) have been excluded, the accuracy of bearing abnormality detection can be improved.

[0062] <Fourth Embodiment> Figure 12 is a functional block diagram of the status monitoring device 100C of the fourth embodiment. The status monitoring device 100C has a pulse reduction processing unit 120 and an HPSS processing unit 122, as described in the second embodiment. In the example of Figure 12, the pulse reduction processing unit 120 is placed before the STFT processing unit 104. A status monitoring device 100C having a pulse reduction processing unit 120 and an HPSS processing unit 122, as in the fourth embodiment, may also be provided.

[0063] <Fifth Embodiment> In the fifth embodiment, a method of specifying the user that differs from the first embodiment is described. Figure 13 is a diagram illustrating the user specification method in the first example. In the example of Figure 13, the user can set up multiple temporary areas. The user can then specify area 250 from among the multiple temporary areas. In the example of Figure 13, the specified area 250 and the multiple temporary areas 251, 252, and 253 are displayed. Note that the specified area 250 and the multiple temporary areas 251, 252, and 253 are displayed in different ways.

[0064] The condition monitoring device of the fifth embodiment detects whether or not there is an abnormality in the bearing 10 based on the spectrogram of the frequency range and time range of the designated region 250. Furthermore, each time the region designation is switched from among a plurality of hypothetical regions, the condition monitoring device detects whether or not there is an abnormality in the bearing 10 based on the spectrogram of the frequency range and time range of the designated region.

[0065] With this configuration, the user can specify areas suspected of being abnormal as multiple hypothetical areas and detect whether or not there is an abnormality in the bearing 10 by switching between the specified areas among these multiple hypothetical areas. Therefore, user convenience can be improved.

[0066] Figure 14 is a diagram illustrating the user specification method in the second example. In the example in Figure 3, a configuration was described in which the user specifies a rectangular area 250. In the second example, an example is shown in which a non-rectangular area 270 with a shape other than a rectangle is specified by the user. In the example in Figure 14, the non-rectangular area 270 is specified by a curve.

[0067] The range processing unit 108 identifies the range formed by the minimum time T1 and maximum time T2 indicated by the non-rectangular region 270 as the time range. The range processing unit 108 also identifies the range formed by the minimum frequency F1 and maximum frequency F2 indicated by the non-rectangular region 270 as the frequency range. In other words, the range processing unit 108 converts the non-rectangular region 270 into region 250 to identify the time range and frequency range specified by the user.

[0068] This diagram illustrates the user specification method for the third example. Figure 15 shows a point region 275 specified by multiple points 283. The range processing unit 108 identifies the range formed by the minimum time T1 and maximum time T2 indicated by the point region 275 as the time range. The range processing unit 108 also identifies the range formed by the minimum frequency F1 and maximum frequency F2 indicated by the point region 275 as the frequency range. In other words, the range processing unit 108 converts the point region 275 into a region 250 to identify the time range and frequency range specified by the user.

[0069] As shown in Figures 14 and 15, the state monitoring device of the fifth embodiment can receive non-rectangular regions 270 or point regions 275 from the user in the spectrogram image. Therefore, the state monitoring device of the fifth embodiment can improve user convenience and identify time ranges and frequency ranges specified by the user.

[0070] <Sixth Embodiment> Next, we will explain the "grid processing" described in parentheses in Figure 2 for the range processing unit 108. For example, suppose a user has previously specified a region 250 and wants to specify a region 250 again with the same frequency range and time range as that previous region 250. However, as explained in the first embodiment, since the user specifies the region 250 in the spectrogram image 230, there may be a difference in frequency range or time range between the previously specified region 250 and the newly specified region 250.

[0071] Therefore, the status monitoring device of the sixth embodiment performs the following grid processing (grid snapping) to suppress deviations in the frequency range or time range. The status monitoring device is configured by the user with a set frequency step size and a set time step size. The status monitoring device then changes the upper and lower limits of the frequency range specified by the user to the nearest multiple of the preset frequency step size and displays the spectrogram image on the display device 150. Similarly, the status monitoring device changes the upper and lower limits of the time range specified by the user to the nearest multiple of the preset time step size and displays the spectrogram image on the display device 150.

[0072] For example, let's describe the grid processing when the frequency step size and time step size are set to 25 Hz and 1 second, respectively. Suppose the user specifies the frequency range and time range as 405 to 1020 Hz and 1.2 to 39.9 seconds, respectively, in spectrogram image 230.

[0073] In such cases, the range processing unit 108 changes the lower limit (405Hz) and upper limit (1020Hz) of the frequency range specified by the user to 400Hz and 1025Hz, respectively, which are the closest multiples of the frequency step size (25Hz). Similarly, the range processing unit 108 changes the lower limit (1.2 seconds) and upper limit (39.9 seconds) of the time range specified by the user to 1 second and 40 seconds, respectively, which are the closest multiples of the time step size (1 second).

[0074] In this way, the range processing unit 108 changes the upper and lower limits of the frequency range specified by the user to the nearest multiple of the frequency step size as a grid process. Furthermore, the range processing unit 108 also changes the upper and lower limits of the time range specified by the user to the nearest multiple of the time step size as a grid process. Therefore, the status monitoring device 100 can suppress the above-mentioned deviations in the frequency range and time range.

[0075] As an alternative, the range processing unit 108 may perform grid processing on either the frequency range or the time range.

[0076] <Seventh Embodiment> As described above, the user specifies region 250 in the spectrogram image 230. However, if the user has little experience using the status monitoring device, it may be difficult to determine which region of the spectrogram image 230 to specify.

[0077] Therefore, the status monitoring device of the seventh embodiment displays a recommended time range and a recommended frequency range on the display device 150, which are recommended to be specified by the user.

[0078] Figure 16 is a functional block diagram of the condition monitoring device 100D of the seventh embodiment. In the example shown in Figure 16, the condition monitoring device 100D includes a recommended range display unit 132, a trained model 134, and a historical analysis DB (Database) 136.

[0079] First, the method for determining the recommended frequency range will be explained. Generally, the frequency range corresponding to abnormalities that may occur in the bearing 10 under inspection tends to be the same as the abnormal frequency range corresponding to abnormalities that have occurred in bearings of the same type as bearing 10 in the past. Therefore, the condition monitoring device 100D of the seventh embodiment identifies the abnormal frequency range corresponding to abnormalities that have occurred in bearings of the same type as bearing 10 in the past as the recommended frequency range.

[0080] Figure 17 shows an example of the configuration of the historical analysis DB136. As shown in Figure 17, the historical analysis DB136 associates a bearing type with the number of past abnormalities that have occurred for that bearing type and the frequency range corresponding to that abnormality. This frequency range is the frequency band of the bandpass filter set for that bearing.

[0081] In the example shown in Figure 11, for a bearing of type A1, one abnormality B corresponds to the corresponding abnormal frequency range, two abnormalities B corresponds to the corresponding abnormal frequency range, and three abnormalities B corresponds to the corresponding abnormal frequency range.

[0082] Furthermore, the recommended range display unit 132 stores the model number of the bearing 10 to be inspected. The recommended range display unit 132 refers to the past analysis DB 136 and identifies the abnormal frequency range corresponding to the stored bearing 10 model as the recommended frequency range.

[0083] Next, the method for determining the recommended frequency range will be explained. The recommended range display unit 132 uses a trained model 134 to determine the recommended time information. The trained model 134 is trained, for example, with training data that associates a noise-free spectrogram with a label indicating that the spectrogram is noise-free. Noise includes not only general signal noise but also disturbances caused by fluctuations in the rotational speed of the rotating wheel 12.

[0084] The recommended range display unit 132 identifies a time range free of noise as the recommended time range by applying the spectrogram from the STFT processing unit 104 to the trained model 134.

[0085] Figure 18 shows an example of the recommended range 280 displayed on the display device 150 by the recommended range display unit 132. The range in the X-axis direction and the range in the Y-axis direction of the recommended range 280 correspond to the recommended time range and recommended frequency range, respectively. The display mode of the image in the recommended range 280 differs from the display mode of the user-specified area 250.

[0086] Figure 19 shows another example of the recommended range 280. In the example in Figure 19, the recommended frequency range is the number of abnormal frequency ranges corresponding to the top number of occurrences (3 in the example in Figure 19) where abnormalities occurred most frequently.

[0087] Figures 20 and 21 show other examples of the recommended range 280. In Figure 20, the recommended range 280 shows the recommended frequency range but does not show the recommended time range. In Figure 21, the recommended range 280 shows the recommended time range but does not show the recommended frequency range.

[0088] As described above, the status monitoring device 100D displays at least one of the recommended frequency range and recommended time range on the display device 150, which it recommends the user specify. Therefore, even a user with little experience in specifying frequency ranges and time ranges can appropriately specify the time range and frequency range.

[0089] <Other Embodiments> (1) The predetermined process of the first embodiment has been described as a process for detecting a bearing abnormality. However, the predetermined process may also be a process for displaying an image showing a frequency spectrum on the display device 150. The image showing the frequency spectrum is the image shown in Figure 4. The user can identify the abnormality of the bearing 10 by visually viewing the image showing the frequency spectrum.

[0090] (2) In the embodiments described above, a configuration was described in which the user can specify both a time range and a frequency range. However, the condition monitoring device of the present disclosure may employ a configuration in which the user can specify either a time range or a frequency range.

[0091] (3) Figure 22 shows an example configuration of another monitoring system 300 to which the condition monitoring device 100 may be applied. The monitoring system 300 includes a bearing 10, a sensor 50, a sensor device 410, a server 420, and a client device 430. The sensor device 410 is a device attached to the sensor 50. The condition monitoring device 100 of this disclosure may be applied to any of the sensor device 410, the server 420, and the client device 430.

[0092] For example, the condition monitoring device 100 may be applied to a client device 430. In this case, the sensor device 410 acquires a detection signal, which is transmitted to the client device 430 via the server 420. The client device 430 performs predetermined processing related to the detection of abnormalities in the bearing 10 by performing the above processing on the detection signal. In this example, the client device 430 may also be omitted.

[0093] Furthermore, the condition monitoring device 100 may also be applied to the server 420. In this case, the sensor device 410 acquires a detection signal, which is then transmitted to the server 420. The server 420 performs predetermined processing related to the detection of abnormalities in the bearing 10 by executing the above-described processing on the detection signal. In addition, the display of the spectrogram image and user input are performed by at least one of the server 420 and the client device 430.

[0094] Furthermore, the condition monitoring device 100 may be applied to the sensor device 410. In this case, the sensor device 410 acquires a detection signal and performs the above-described processing on the detection signal to perform predetermined processing related to the detection of abnormalities in the bearing 10. In addition, the display of the spectrogram image and user input are performed by at least one of the sensor device 410, the server 420, and the client device 430.

[0095] (4) In the status monitoring device 100A shown in Figure 7, a configuration was described in which the pulse reduction processing unit 120 performs pulse reduction processing on the signal before the user extracts the time range and frequency range.

[0096] However, a configuration may be adopted in which the pulse reduction processing unit 120 performs pulse reduction processing on the signal after the time range specified by the user has been extracted. This configuration is also referred to as "Configuration X". Figure 23 is a functional block diagram of a state monitoring device 100D employing such Configuration X. The range processing unit 108 extracts components (signals) within the time range specified by the user from the detection signal from the acquisition unit 102 and outputs the extracted signal to the pulse reduction processing unit 120. The pulse reduction processing unit 120 performs pulse reduction processing on this signal. Then, the BPF processing unit 110 extracts components (signals) within the frequency range specified by the user from the signal after pulse reduction processing has been performed and outputs the extracted signal to the envelope processing unit 112. In the example of Figure 23, the detection signal corresponds to the "first target signal" of this disclosure, and the signal after pulse reduction processing by the pulse reduction processing unit 120 corresponds to the "second target signal" of this disclosure.

[0097] Furthermore, Figure 24 is a functional block diagram of a state monitoring device 100E in which the above configuration X is adopted for the state monitoring device 100C of Figure 12. The range processing unit 108 extracts components (signals) within a time range specified by the user from the signal from the inverse STFT processing unit 180 and outputs the extracted signal to the pulse reduction processing unit 120. The pulse reduction processing unit 120 performs pulse reduction processing on this signal. Then, the BPF processing unit 110 extracts components (signals) within a frequency range specified by the user from the signal after pulse reduction processing and outputs the extracted signal to the envelope processing unit 112. In the example of Figure 24, the detection signal corresponds to the "first target signal" of this disclosure, and the signal after pulse reduction processing by the pulse reduction processing unit 120 corresponds to the "second target signal" of this disclosure.

[0098] <Note> (Note 1) A bearing condition monitoring device, An interface for acquiring detection signals related to the vibration of the bearing, Equipped with a control device, The control device is From the first target signal based on the detection signal, a three-dimensional signal is generated indicating the time, frequency, and vibration intensity of the bearing vibration. A three-dimensional image based on the aforementioned three-dimensional signal is displayed on a display device. The user specifies at least one of the frequency range and time range in the three-dimensional image displayed on the display device. A condition monitoring device that performs a predetermined process for detecting abnormalities in the bearing using signals within a range specified by the user from among the second target signals based on the aforementioned detection signal.

[0099] (Note 2) The control device performs a reduction process to reduce the pulses included in the detection signal. The status monitoring device as described in Appendix 1, wherein the first target signal is a signal on which the reduction process has been performed on the detection signal.

[0100] (Note 3) The first target signal is the detection signal, The control device performs a reduction process to reduce the pulses included in the detection signals within the time range specified by the user. The second target signal has a post-processing signal after the reduction process has been performed. The control device is a state monitoring device as described in Appendix 1, which performs the predetermined processing using a signal within the frequency range specified by the user from among the post-execution signals.

[0101] (Note 4) The control device performs a process to extract from the first target signal a frequency-domain related signal that is associated with the bearing abnormality by executing HPSS (Harmonic Percussion Source Separation) on the first target signal. The control device converts the associated signal into a time-domain signal, The status monitoring device according to Appendix 1 or Appendix 2, wherein the second target signal is the related signal converted to the time domain.

[0102] (Note 5) The first target signal is the detection signal, The control device is By performing HPSS on the detection signal, a process is executed to extract related signals in the frequency domain that are associated with the bearing abnormality from the detection signal. The aforementioned related signals are converted into time-domain signals, A reduction process is performed to reduce the pulses contained in the signals within the time range specified by the user among the related signals that have been converted to the time domain. The second target signal has a post-processing signal after the reduction process has been performed. The control device is a state monitoring device as described in Appendix 1, which performs the predetermined processing using a signal within the frequency range specified by the user from among the post-execution signals.

[0103] (Note 6) The range specified by the user includes the aforementioned time range, The reduction process described above is A process for calculating the average amplitude and variance of the detection signal within the time range specified by the user, A process for determining the upper and lower limits of a first range including the amplitude mean based on the variance, A process for determining the upper and lower limits of a second range that is included in the first range but is narrower than the first range, based on the variance, A condition monitoring device according to any one of Appendix 2, Appendix 3, and Appendix 5, which includes a process of changing the amplitude of the detection signal that does not belong to the first range to a random value so that it belongs to the second range.

[0104] (Note 7) The control device changes the upper and lower limits of the frequency range specified by the user to the nearest multiple of the preset frequency step size, and displays the three-dimensional image on the display device, as described in any one of Appendix 1 to Appendix 6.

[0105] (Note 8) The control device changes the upper and lower limits of the time range specified by the user to the nearest multiple of the preset time step size, and displays a three-dimensional image on the display device, as described in any one of Appendix 1 to Appendix 7.

[0106] (Note 9) The control device is In the aforementioned three-dimensional image, a region or multiple points are received from the user, The range consisting of the minimum and maximum times indicated by the region or the plurality of points is identified as the time range. A condition monitoring device according to any one of the appendices 1 to 8, wherein the range comprising the minimum and maximum frequencies indicated by the region or the plurality of points is specified as the frequency range.

[0107] (Note 10) The control device is a status monitoring device according to any one of the appendices 1 to 9, wherein the control device displays on the display device at least one of a recommended frequency range and a recommended time range that the user is recommended to specify in the three-dimensional image displayed on the display device.

[0108] (Note 11) The control device is a condition monitoring device as described in Appendix 10, wherein the control device displays the abnormal frequency range when an abnormality occurs in another bearing of the same type as the bearing as the recommended frequency range on the display device.

[0109] (Note 12) The control device is a status monitoring device according to Appendix 10 or Appendix 11, wherein the control device displays the time range of the three-dimensional signal that does not contain noise as the recommended time range on the display device.

[0110] (Note 13) The condition monitoring device described in any one of the appendices 1 to 12, wherein the predetermined process is a process for detecting an abnormality in the bearing.

[0111] (Note 14) The predetermined processing of the axis is a process of displaying an image showing the frequency spectrum of the frequency range specified by the user on the display device, as described in any one of Appendix 1 to Appendix 13 of the status monitoring device.

[0112] (Note 15) A method for monitoring the condition of a bearing, To acquire a detection signal related to the vibration of the bearing, A three-dimensional signal is generated from the first target signal based on the aforementioned detection signal. Displaying a three-dimensional image based on the aforementioned three-dimensional signal on a display device, The display device accepts from the user the specification of at least one of the frequency range and time range in the three-dimensional image displayed on the display device, A condition monitoring method comprising performing a predetermined process for detecting an abnormality in the bearing based on a three-dimensional signal within the frequency range specified by the user.

[0113] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims rather than by the description of the embodiments above, and all modifications within the meaning and scope equivalent to the claims are intended to be included. [Explanation of Symbols]

[0114] 10 Bearing, 11 Fixed wheel, 12 Rotating wheel, 13 Rolling element, 50 Sensor, 100, 100A, 100B, 100C, 100D Condition monitoring device, 102 Acquisition unit, 104 STFT processing unit, 106 Spectrogram display unit, 108 Range processing unit, 110 BPF processing unit, 112 Envelope processing unit, 114 FFT processing unit, 116 Analysis unit, 120 Pulse reduction processing unit, 122 HPSS processing unit, 132 Recommended range display unit, 134 Learned model, 150 Display device, 160 Input device, 204 Memory, 206 Interface, 230 Spectrogram image, 250 Region, 270 Non-rectangular region, 275 Point region, 280 Recommended range, 300 Monitoring system, 1202 Calculation unit, 1204 Setting unit, 1206 Replacement unit.

Claims

1. A bearing condition monitoring device, An interface for acquiring detection signals related to the vibration of the bearing, Equipped with a control device, The control device is From the first target signal based on the detection signal, a three-dimensional signal is generated indicating the time, frequency, and vibration intensity of the bearing vibration. A three-dimensional image based on the aforementioned three-dimensional signal is displayed on a display device. The user specifies at least one of the frequency range and time range in the three-dimensional image displayed on the display device. A condition monitoring device that performs a predetermined process for detecting abnormalities in the bearing using signals within a range specified by the user from among the second target signals based on the aforementioned detection signal.

2. The control device performs a reduction process to reduce the pulses included in the detection signal. The status monitoring device according to claim 1, wherein the first target signal is a signal on which the reduction process has been performed on the detection signal.

3. The first target signal is the detection signal, The control device performs a reduction process to reduce the pulses included in the detection signals within the time range specified by the user. The second target signal has a post-processing signal after the reduction process has been performed. The status monitoring device according to claim 1, wherein the control device performs the predetermined processing using the signal within the frequency range specified by the user from among the post-execution signals.

4. The control device performs HPSS (Harmonic Percussion Source Separation) on the first target signal to extract related signals in the frequency domain that are associated with the bearing abnormality from the first target signal. The control device converts the associated signal into a time-domain signal, The status monitoring device according to claim 1 or claim 2, wherein the second target signal is the associated signal converted to the time domain.

5. The first target signal is the detection signal, The control device is By performing HPSS on the detection signal, a process is executed to extract frequency-domain related signals from the detection signal that are associated with the bearing abnormality. The aforementioned related signals are converted into time-domain signals, A reduction process is performed to reduce the pulses contained in the signals within the time range specified by the user among the related signals that have been converted to the time domain. The second target signal has a post-processing signal after the reduction process has been performed. The status monitoring device according to claim 1, wherein the control device performs the predetermined processing using the signal within the frequency range specified by the user from among the post-execution signals.

6. The range specified by the user includes the aforementioned time range, The reduction process described above is A process for calculating the average amplitude and variance of the detection signal within the time range specified by the user, A process for determining the upper and lower limits of a first range including the mean amplitude based on the variance, A process for determining the upper and lower limits of a second range that is included in the first range but is narrower than the first range, based on the variance, A state monitoring device according to any one of claims 2, 3, and 5, comprising the process of changing the amplitude of the detection signal that does not belong to the first range to a random value so that it belongs to the second range.

7. The status monitoring device according to any one of claims 1 to 3, wherein the control device changes the upper and lower limits of the frequency range specified by the user to the nearest multiple of the preset frequency step size, and displays the three-dimensional image on the display device.

8. The control device changes the upper and lower limits of the time range specified by the user to the nearest value among the multiples of the preset time step size, and displays the three-dimensional image on the display device, according to any one of claims 1 to 3.

9. The control device is In the aforementioned three-dimensional image, a region or multiple points are received from the user, The range consisting of the minimum and maximum times indicated by the region or the plurality of points is identified as the time range. A condition monitoring device according to any one of claims 1 to 3, wherein the range formed by the minimum frequency and maximum frequency indicated by the region or the plurality of points is specified as the frequency range.

10. The status monitoring device according to any one of claims 1 to 3, wherein the control device displays on the display device at least one of a recommended frequency range and a recommended time range that the user is recommended to specify in the three-dimensional image displayed on the display device.

11. The condition monitoring device according to claim 10, wherein the control device displays the abnormal frequency range when an abnormality occurs in another bearing of the same type as the bearing as the recommended frequency range on the display device.

12. The status monitoring device according to claim 10, wherein the control device displays the time range in which the three-dimensional signal does not contain noise as the recommended time range on the display device.

13. The condition monitoring device according to any one of claims 1 to 3, wherein the predetermined process is a process for detecting an abnormality in the bearing.

14. The status monitoring device according to any one of claims 1 to 3, wherein the predetermined processing of the axis is a process of displaying an image showing a frequency spectrum of the frequency range specified by the user on the display device.

15. A method for monitoring the condition of a bearing, To acquire a detection signal related to the vibration of the bearing, From the first target signal based on the detection signal, a three-dimensional signal is generated indicating the time, frequency, and vibration intensity of the vibration of the bearing. Displaying a three-dimensional image based on the aforementioned three-dimensional signal on a display device, The device accepts from the user the specification of at least one of the frequency range and time range in the three-dimensional image displayed on the display device, A condition monitoring method comprising: performing a predetermined process for detecting an abnormality in the bearing using a signal within a range specified by the user from among the second target signals based on the detection signal.

Images