JP6639266B2 - Abnormality diagnosis device and abnormality diagnosis method - Google Patents

Description

  The present invention relates to a bearing device abnormality diagnosis device and a bearing device abnormality diagnosis method.

  Japanese Patent Laying-Open No. 2013-185507 (Patent Document 1) discloses a condition monitoring system (CMS: Condition Monitoring System) that can appropriately diagnose an abnormality of a device provided in a wind power generator. In this state monitoring system, it is diagnosed whether or not the main shaft bearing is damaged by using an effective value of vibration data measured by an acceleration sensor fixed to the main shaft bearing.

JP 2013-185507 A

  The effective value of the vibration data is defined as a root mean square (RMS) of the amplitude of the waveform of the vibration data. The effective value can be said to be an index value that is a measure of the magnitude of the amplitude included in the waveform of the vibration data.

  When the bearing device is damaged, vibrations due to the damage may be newly generated, or the vibrations that have been generated in a normal state may not be generated in an abnormal state. As described above, the amplitude of the waveform of the vibration data may be different between a normal state and an abnormal state. Therefore, an abnormality of the bearing device may be detected from the vibration data using the effective value of the vibration data.

  In calculating the effective value, generally, when calculating the root mean square of the amplitude in the acceleration vibration data, the amplitude of the vibration that is not caused by damage is also used. The amplitude of vibration not caused by damage hardly changes depending on the presence or absence of damage. Therefore, the change in the effective value due to the presence or absence of the damage almost depends on the change in the amplitude of the vibration caused by the damage.

  When damage occurs in a bearing device that rotates at a low speed (for example, about 100 rpm) such as a main shaft bearing of a wind power generator, a change in acceleration caused by the damage is often smaller than that in a bearing device that rotates at a high speed. It is difficult to distinguish from a change in acceleration that is not caused by the above. As a result, the effective value hardly changes depending on the presence or absence of damage. Therefore, if the abnormality diagnosis of the bearing device rotating at a low speed is performed using the effective value of the acceleration vibration data, an erroneous diagnosis may occur.

  A main object of the present invention is to provide an abnormality diagnosis device and an abnormality diagnosis method that can improve the accuracy of abnormality diagnosis of a bearing device by using acceleration vibration data.

  An abnormality diagnosis device according to the present invention detects damage to a bearing device based on vibration data of acceleration of the bearing device. The abnormality diagnosis device includes a filter and a diagnosis unit. The filter is configured to extract a vibration waveform belonging to a predetermined frequency band from the vibration data. The diagnosis unit determines the evaluation value calculated from the ratio between the first amplitude that is the amplitude of the vibration waveform, which exceeds the reference value, and the second amplitude of the vibration waveform after a lapse of a predetermined time from the time when the first amplitude occurs. Is configured to diagnose that the bearing device has been damaged when the pressure exceeds the threshold value.

  According to the present invention, in the abnormality diagnosis of the bearing device, a value calculated from a ratio between the first amplitude and the second amplitude of the vibration waveform after a lapse of a predetermined time from the time when the first amplitude has occurred is an evaluation value of the abnormality diagnosis. As a result, it is possible to perform an abnormality diagnosis focusing on the fact that the time waveform shapes of the vibrations are different between the normal state and the abnormal state. As a result, the accuracy of the abnormality diagnosis of the bearing device can be improved.

It is the figure which showed the structure of the wind power generator schematically. 5 is a flowchart illustrating a process of abnormality diagnosis performed by the data processing device. It is a figure showing an example of change of the effective value of vibration data. It is a figure showing the whole main spindle bearing composition at the time of normal. It is a figure showing the whole spindle bearing composition at the time of abnormalities. It is a figure which shows both the schematic diagram (a) of the waveform of the vibration data in normal time, and the schematic diagram (b) of the result of short-time Fourier transform. It is a figure which shows both the schematic diagram (a) of the waveform of the vibration data at the time of abnormality, and the schematic diagram (b) of the result of short-time Fourier transform. FIG. 7 is a diagram showing a waveform diagram (a) of vibration data measured in a normal state and a result (b) of performing a short-time Fourier transform on the vibration data. FIG. 7 is a diagram showing both a waveform diagram (a) of vibration data measured at the time of abnormality and a result (b) of performing a short-time Fourier transform on the vibration data. FIG. 3 is a functional block diagram for describing a functional configuration of a data processing device that performs abnormality diagnosis. It is a figure showing a vibration waveform (500-5000 Hz) at the time of normal. It is a figure which expands and shows a part of vibration waveform shown in FIG. It is a figure which shows the vibration waveform (500-5000Hz) at the time of abnormality. FIG. 14 is an enlarged view showing a part of the vibration waveform shown in FIG. 13. It is a figure showing change of the number of times that the ratio of the 1st amplitude and the 2nd amplitude exceeded the threshold. 5 is a flowchart for describing processing of a subroutine for calculating an evaluation value used in the first embodiment. It is a figure showing change of the frequency of occurrence when the ratio of the 1st amplitude and the 2nd amplitude exceeds a threshold. 15 is a flowchart for describing processing of a subroutine for calculating an evaluation value used in the second embodiment. It is a figure which shows the result of having performed the significant difference test by the t test about each evaluation value of a comparative example, Embodiment 1, and Embodiment 2.

  Hereinafter, embodiments will be described with reference to the drawings. In the following description, the same or corresponding elements have the same reference characters allotted, and detailed description thereof will not be repeated.

[Embodiment 1]
FIG. 1 is a diagram schematically illustrating a configuration of the wind turbine generator 1. As shown in FIG. 1, the wind turbine generator 1 includes a main shaft 10, a main shaft bearing 20, a blade 30, a speed increaser 40, a generator 50, an acceleration sensor 70, and a data processing device 80. . The main shaft bearing 20, the gearbox 40, the generator 50, the acceleration sensor 70, and the data processing device 80 are stored in the nacelle 90. Nacelle 90 is supported by tower 100.

  The main shaft 10 is connected to an input shaft of the gearbox 40 in the nacelle 90. The main shaft 10 is rotatably supported by a main shaft bearing 20. The main shaft 10 transmits the rotational torque generated by the blade 30 that has received the wind to the input shaft of the gearbox 40. The blade 30 is provided at the tip of the main shaft 10, converts wind power into rotational torque, and transmits the rotational torque to the main shaft 10.

  The main shaft bearing 20 is configured to include a rolling bearing, and includes, for example, a self-aligning roller bearing, a tapered roller bearing, a cylindrical roller bearing, a ball bearing, and the like. These bearings may be single-row or double-row.

  The acceleration sensor 70 is disposed on the main shaft bearing 20 of the main shaft, and measures vibration generated in the main shaft bearing 20.

  The speed increaser 40 is provided between the main shaft 10 and the generator 50, and increases the rotation speed of the main shaft 10 and outputs the rotation speed to the generator 50. As an example, the speed increasing device 40 is configured by a gear speed increasing mechanism including a planetary gear, an intermediate shaft, a high speed shaft, and the like. Although not particularly shown, a plurality of bearings for rotatably supporting a plurality of shafts are provided in the speed increasing device 40. The generator 50 is connected to the output shaft of the gearbox 40, and rotates to generate power by the rotation torque received from the gearbox 40. The generator 50 includes, for example, an induction generator. Note that a bearing for rotatably supporting the rotor is also provided in the generator 50.

  The data processing device 80 is provided inside the nacelle 90 and receives vibration data of the main shaft bearing 20 measured by the acceleration sensor 70. The data processing device 80 uses the vibration data received from the acceleration sensor 70 to perform an abnormality diagnosis as to whether or not the main shaft bearing 20 has been damaged. Although not shown, the acceleration sensor 70 and the data processing device 80 are connected by a wired cable, and are configured to be able to communicate data. Communication between the acceleration sensor 70 and the data processing device 80 may be performed by wireless communication. The data processing device 80 corresponds to the abnormality diagnosis device of the present invention.

The data processing device 80 diagnoses, based on the vibration data received from the acceleration sensor 70, whether the spindle bearing 20 has been damaged. FIG. 2 is a flowchart showing the abnormality diagnosis processing performed by the data processing device 80. As shown in FIG. 2, in the data processing device 80, in step (hereinafter, step is simply referred to as S) S <b> 1, the main shaft bearing 20 is damaged based on the vibration data received from the acceleration sensor 70. An evaluation value V for diagnosing whether or not is calculated, and the process proceeds to S2. In S2, the data processing device 80 determines whether the evaluation value V is an abnormal value. If the evaluation value V is equal to or _smaller than the determination value Vd (NO in S2), the data processing device 80 determines that the evaluation value V is a normal value and ends the processing. If the evaluation value V is greater than the determination value V _d (YES in S2), the data processing unit 80, abnormality in the spindle bearing 20 in S3 as an evaluation value V is abnormal values is to inform the user of the occurrence. Examples of the notification method include an audible or visual method such as a sound, lighting of a lamp, or transmission of a message. Determination value V _d may be appropriately determined by actual experiment or simulation.

  When an abnormality occurs in the main shaft bearing 20, the amplitude of the vibration caused by the abnormality may appear in the vibration data. It is known that an effective value is used as an evaluation value V for distinguishing between vibration data in a normal state and vibration data in an abnormal state by focusing on the difference in amplitude of the waveform of such vibration data. I have. The effective value is defined as the root mean square of the amplitude in the waveform of the vibration data. The effective value can be said to be an index value that is a measure of the magnitude of the amplitude included in the waveform of the vibration data.

FIG. 3 is a diagram illustrating an example of a change in the effective value of the vibration data. 3, until the time TR ₁ is in an abnormal state damage the main shaft bearing 20 has occurred. Spindle bearing 20 during the period from the time TR ₁ to time TR ₂ is replaced, at time TR ₂ after the spindle bearing 20 is a normal state. As shown in FIG. 3, the range in which the effective values are distributed at time TR ₁ before, and a range in which the effective value are distributed at time TR ₂ after almost identical. Therefore, it is difficult to clearly distinguish between the normal state and the abnormal state.

  In calculating the effective value, when calculating the root-mean-square of the amplitude in the acceleration vibration data, the amplitude of the vibration not caused by damage is also used. The amplitude of vibration not caused by damage hardly changes depending on the presence or absence of damage. Therefore, the change in the effective value due to the presence or absence of the damage almost depends on the change in the amplitude of the vibration caused by the damage.

  When damage occurs in a bearing device that rotates at a low speed (for example, about 100 rpm) such as the main shaft bearing 20 of the wind turbine generator 1, a change in acceleration caused by the damage is often smaller than that of a bearing device that rotates at a high speed. This makes it difficult to distinguish from a change in acceleration that is not caused by damage. As a result, as shown in FIG. 3, the effective value hardly changes depending on the presence or absence of damage. Therefore, if the abnormality diagnosis of the main shaft bearing 20 is performed using the effective value of the vibration data of the acceleration, an erroneous diagnosis may occur.

  Therefore, in the first embodiment, attention is paid to the fact that the frequency component included in the vibration generated by the collision between the rolling element and the retainer or the bearing ring (the inner ring or the outer ring) differs between the normal state and the abnormal state.

  The reason why the frequency component included in the vibration generated by the collision between the rolling element and the cage or the bearing ring differs between the normal state and the abnormal state will be described with reference to FIGS. FIG. 4 is a diagram showing an entire configuration of the main shaft bearing 20 in a normal state. As shown in FIG. 4A, the main shaft bearing 20 includes an inner ring 22, an outer ring 24, a retainer 26, and a plurality of rolling elements 28.

  In FIG. 4A, the main shaft 10 rotates in a direction indicated by an arrow D. As shown in FIG. 4A, a radial load is applied to the main shaft 10 in a direction indicated by an arrow N perpendicular to the rotation axis of the main shaft 10. Arrow G indicates the direction of gravity. A point PA on the inner peripheral surface of the outer ring 24 is located just above the rotation axis of the main shaft 10. The point PB on the inner peripheral surface of the outer ring 24 is located at a position rotated by 90 degrees in the rotation direction D from the position of the point PA. The point PC on the inner peripheral surface of the outer ring 24 is located vertically downward from the rotation axis of the main shaft 10.

  The inner ring 22 is fitted into and fixed to the main shaft 10, and rotates in the direction of arrow D integrally with the main shaft 10. The outer ring 24 is arranged on the outer peripheral side of the inner ring 22.

  The cage 26 is provided with a plurality of pockets Pkt for holding the rolling elements 28 at equal intervals. The retainer 26 is disposed between the outer peripheral surface of the inner race 22 and the inner peripheral surface of the outer race 24 while holding the rolling element 28 in the pocket Pkt. When the rolling element 28 rotates along the outer peripheral surface of the inner ring 22 with the rotation of the inner ring 22, the retainer 26 rotates between the outer peripheral surface of the inner ring 22 and the inner peripheral surface of the outer ring 24 together with the rolling element 28.

  The rolling element 28 rotates between the inner race 22 and the outer race 24 in the rotation direction D while being held in the pocket Pkt of the cage 26. As shown in FIGS. 4B and 4C, between the rolling element 28 and the pocket Pkt of the retainer 26 holding the rolling element 28, the rolling element 28 rotates in the pocket Pkt. A gap (pocket Pkt gap) is provided so as to be able to operate. Normally, grease is applied to the rolling elements 28 to reduce friction between the pockets Pkt of the retainer 26, the inner ring 22, and the outer ring 24.

  When the rolling element 28 moves from the point PC to the point PA, the rolling element 28 is deviated in the pocket Pkt in a direction opposite to the rotation direction D due to gravity. The rolling element 28 contacts the retainer 26 and is pushed up to the point PA by the retainer 26. When the rolling element 28 passes through the point PA, the rolling element 28 is deviated in the pocket Pkt in a direction opposite to the rotation direction D (see FIG. 4B).

  The rolling element 28 in the state shown in FIG. 4B receives gravity and moves in the pocket Pkt after passing through the point PA and before passing through the point PB, and the cage 26, the inner ring 22, or It collides directly with the outer ring 24 and the like (see FIG. 4C). Vibration occurs when the rolling element 28 collides with the cage 26, the inner ring 22, the outer ring 24, or the like.

  FIG. 5 is a diagram showing the entire configuration of the main shaft bearing 20 at the time of abnormality. In FIG. 5A, the vicinity of the point PC is cut by the rolling elements 28, and damage I occurs. At the point PC located directly below the rotation axis of the main shaft 10, the radial load when the rolling element 28 passes through the point PC is likely to be larger than other portions of the inner peripheral surface of the outer ring 24. Every time the rolling element 28 passes through the point PC, a large load is applied to the point PC, so that the point PC is easily damaged.

  When the damage I occurs, for example, a peeled piece scraped off from the inner peripheral surface near the point PC of the outer ring 24 by the rolling element 28 when the damage I occurs, or wear powder generated when the rolling element 28 passes through the damage I Such foreign matter S adheres to the inner race 22, the outer race 24, the retainer 26, and the rolling elements 28 (see FIG. 5B). When the rolling element 28 passes through the point PA and collides with the retainer 26, the inner ring 22 or the outer ring 24 in this state, the foreign matter S is present between the rolling element 28 and the retainer 26, the inner ring 22 or the outer ring 24. They often exist (see FIG. 5C). The impact when the rolling element 28 collides with the retainer 26, the inner ring 22, the outer ring 24, or the like is weakened by the foreign matter S. Therefore, the frequency component included in the vibration generated in the main shaft bearing 20 when the rolling element 28 collides with them becomes lower in the abnormal state than in the normal state.

  6 and 7, how the frequency component included in the vibration generated when the rolling element 28 collides with the cage 26, the inner ring 22, or the outer ring 24, etc., is different between the normal state and the abnormal state. I do. FIG. 6 is a diagram showing both the schematic diagram (a) of the waveform of the vibration data in the normal state and the schematic diagram (b) of the result of the short-time Fourier transform. FIG. 7 is a diagram showing both the schematic diagram (a) of the waveform of the vibration data at the time of abnormality and the schematic diagram (b) of the result of the short-time Fourier transform. The short-time Fourier transform performs a fast Fourier transform (FFT) by dividing vibration data at predetermined time intervals.

6 (a), the rolling elements 28 in each of the time _TN 1 to Tn _3, the retainer 26, are occurring collision with such an inner ring 22 or outer ring 24. In FIG. 7 (a), the rolling elements 28 in each of the time _TD 1 _{~TD 3,} the retainer 26, the collision between such inner 22 or outer ring 24, is happening.

As shown in FIG. 6 (b), the frequency components of the time _TN 1 to Tn ₃ is beyond the frequency _{F th.} On the other hand, as shown in FIG. 7 (b), the frequency components of the time _TD 1 _{~TD 3} is less than the frequency _{F th.} As described above, the frequency component included in the vibration generated when the rolling element 28 collides with the cage 26, the inner ring 22, the outer ring 24, or the like can be clearly distinguished between the normal state and the abnormal state.

  Hereinafter, how the frequency components included in the actually measured vibration data differ between the normal state and the abnormal state will be described with reference to FIGS. 8 and 9. FIG. 8 is a diagram showing a waveform diagram (a) of vibration data measured in a normal state and a result (b) of performing a short-time Fourier transform on the vibration data. FIG. 9 is a diagram showing a waveform diagram (a) of vibration data measured at the time of abnormality and a result (b) of performing a short-time Fourier transform on the vibration data. Each of FIG. 8B and FIG. 9B is a diagram showing a result of performing fast Fourier transform (FFT: Fast Fourier Transform) by dividing vibration data of 40 seconds at intervals of 0.1 second.

  8B and 9B, the horizontal axis represents time, and the vertical axis represents frequency components. In FIG. 8B and FIG. 9B, frequency components whose power spectral density is larger than a predetermined threshold value are plotted in order to extract a frequency band to be focused on in abnormality diagnosis. The power spectrum density corresponds to the intensity of a signal corresponding to each frequency component in the spectrum resulting from the FFT. Since the frequency component whose power spectrum density is equal to or less than the threshold value is hardly considered to be the vibration component caused by the collision of the rolling element 28 with the cage 26, the inner ring 22, the outer ring 24, etc., FIG. 9 (b) is not plotted.

  Comparing FIG. 8B and FIG. 9B, when the frequency component included in the vibration data exceeds 5000 Hz, the normal time (FIG. 8B) is abnormal (FIG. 9B). )), And almost nonexistent in abnormal times. On the other hand, the frequency component of 5000 Hz or less is included in the vibration data at most times in both the normal state and the abnormal state.

  The collision between the rolling element 28 and the retainer 26, the inner ring 22, the outer ring 24, or the like as shown in FIGS. 4C and 5C is repeated at predetermined time intervals as the main shaft 10 rotates. appear. The vibration caused by the collision occurs in a frequency band of 5000 Hz or less in an abnormal state. In a normal state, the vibration also occurs in a frequency band exceeding 5000 Hz. Therefore, the vibration generated by the collision between the rolling element 28 and the retainer 26, the inner ring 22, the outer ring 24, or the like is less likely to be attenuated in an abnormal state in a frequency band of 5000 Hz or less. That is, in the frequency band of 5000 Hz or less, the attenuation rate of the vibration caused by the collision between the rolling element 28 and the cage 26, the inner ring 22, the outer ring 24, or the like is smaller in the abnormal state than in the normal state.

  Therefore, in the first embodiment, a vibration waveform of 500 to 5000 Hz is extracted from the vibration data, and a first amplitude exceeding a reference value in the vibration waveform and a first amplitude exceeding a reference value after a lapse of a predetermined time from the time when the first amplitude occurs. The number of times that the ratio with the two amplitudes exceeds the threshold value is used as an evaluation value for abnormality diagnosis. By using, as an evaluation value, a value calculated from a ratio between the first amplitude and the second amplitude after a lapse of a predetermined time from the time when the first amplitude occurs, the time waveform shape of the vibration between the normal time and the abnormal time can be changed. An abnormality diagnosis focusing on the difference can be made. As a result, the accuracy of the abnormality diagnosis of the main shaft bearing 20 can be improved.

  FIG. 10 is a functional block diagram for describing a functional configuration of data processing device 80 that performs abnormality diagnosis. As shown in FIG. 10, the data processing device 80 includes a filter 81, a diagnosis unit 83, and a notification unit 84. The vibration data measured by the acceleration sensor 70 is input to the filter 81. The filter 81 may receive the vibration data from the acceleration sensor 70 in real time, or may read the vibration data stored in a memory (not shown) from the memory.

  The filter 81 extracts a vibration waveform Wv of 500 to 5000 Hz from the vibration data and outputs it to the diagnosis unit 83. Filter 81 includes, for example, a band-pass filter.

  The diagnosis unit 83 includes a computer such as a CPU (Central Processing Unit), and a volatile memory and a non-volatile memory for storing data necessary for abnormality diagnosis. Upon receiving the vibration waveform Wv, the diagnosis unit 83 performs an abnormality diagnosis as to whether or not the main shaft bearing 20 has been damaged. When determining that an abnormality has occurred in the main shaft bearing 20, the diagnosis unit 83 controls the notification unit 84 to notify the user that an abnormality has occurred in the main shaft bearing 20.

  The diagnosis unit 83 extracts the first amplitude exceeding the reference value E in the vibration waveform Wv (500 to 5000 Hz) in the abnormality diagnosis. The reason why the amplitude exceeding the reference value E is extracted from the vibration waveform Wv is to specify the time at which the rolling element 28 collides with the retainer 26, the inner wheel 22, the outer wheel 24, or the like. The change in acceleration that occurs when the rolling elements 28 collide with them appears relatively large in the waveform of the vibration data due to the impact due to the collision. Therefore, by extracting an amplitude exceeding the reference value E from the vibration waveform Wv, the time at which the rolling element 28 collides with the retainer 26, the inner wheel 22, the outer wheel 24, or the like can be specified. As a result, the amplitude almost irrelevant to the collision of the rolling elements 28 with them can be excluded from the calculation of the evaluation value.

  The reference value E at the time of extracting the first amplitude is desirably a value that hardly exceeds the amplitude generated regardless of the collision of the rolling element 28 with the retainer 26, the inner ring 22, the outer ring 24, or the like. . The reference value E can be appropriately determined by an actual machine experiment or simulation, and can be set to, for example, about five times the effective value of the vibration waveform Wv.

FIG. 11 is a diagram illustrating a vibration waveform Wv in a normal state. The vibration waveform Wv shown in FIG. 11 is a waveform extracted by the filter 81 of FIG. 10 from the vibration data shown in FIG. As shown in FIG. 11, the first amplitude PA _k exceeding the reference value E occurs at time TA _k (k = 1 to N).

Figure 12 is a diagram showing an enlarged vicinity of time TA _k of the vibration waveform Wv shown in Figure 11. As shown in FIG. 12, the second amplitude PB _k occurs at a time TB _k after a lapse of time ΔT from the time TA _k at which the first amplitude PA _k occurs. In the abnormality diagnosis according to the first embodiment, the number of times that the ratio of first amplitude PA _k to second amplitude PB _k exceeds threshold value _Rth is used as an evaluation value of whether an abnormality has occurred.

FIG. 13 is a diagram illustrating a vibration waveform Wv at the time of abnormality. The vibration waveform Wv shown in FIG. 13 is a waveform extracted by the filter 81 of FIG. 10 from the vibration data shown in FIG. As shown in FIG. 13, the first amplitude PA _k exceeding the reference value E occurs at a time TA _k (k = 1 to M).

Figure 14 is a diagram showing an enlarged vicinity of time TA _k of the vibration waveform Wv shown in Figure 13. As shown in FIG. 14, the second amplitude PB _k occurs at a time TB _k after a lapse of time ΔT from the time TA _k at which the first amplitude PA _k occurs. In FIG. 14, as in FIG. 12, the number of times that the ratio between the first amplitude PA _k and the second amplitude PB _k exceeds the threshold value _Rth is used as an evaluation value of whether an abnormality has occurred.

When the rolling element 28 collides with the cage 26, the inner ring 22, the outer ring 24, or the like (at the time TA _k ), the damping rate of the vibration is more normal in the abnormal state where most of the vibration occurs at 500 to 5000 Hz. Be smaller than time. Therefore, the ratio between the first amplitude PA _k and the second amplitude PB _k is larger in the abnormal state than in the normal state.

FIG. 15 is a diagram illustrating a change in the number of times the ratio between the first amplitude PA _k and the second amplitude PB _k exceeds the threshold value _Rth . 15, until the time TR ₁₁ is an abnormal state damage the main shaft bearing 20 has occurred. Spindle bearing 20 during the period from the time _{TR 11} to the time _{TR 12} are exchanged, at time _{TR 12} after the main shaft bearing 20 is a normal state. For example a value L _d as a determination value V _d, the case where the evaluation value V exceeds the determination value V _d is determined to be abnormal, that evaluation value V is determined as normal if less than or equal to the determination value V _d, and the normal An abnormal time can be distinguished. As a result, the accuracy of the abnormality diagnosis can be improved as compared with the comparative example.

FIG. 16 is a flowchart for explaining the processing (S1 in FIG. 2) of the subroutine for calculating the evaluation value V used in the first embodiment. As shown in FIG. 16, the data processing device 80 extracts the first amplitude PA _k (k = 1 to N) exceeding the reference value E from the vibration waveform Wv (500 to 5000 Hz) in S11, and proceeds to S12. Proceed. The data processing apparatus 80 advances a first amplitude PA _k is extracted a second amplitude PB _k at time TB _k after time ΔT has elapsed from the time TA _k generated processing to S13 in S12. The data processing device 80 calculates the ratio R _k between the first amplitude PA _k and the second amplitude PB _k in S13, and advances the processing to S14. The data processing unit 80 calculates the number L ratio R _k exceeds the threshold R _th as an evaluation value V in S14, and returns the process to the main routine to perform the abnormality diagnosis shown in FIG. The data processing unit 80, evaluation value V is larger than the determination value V _d informs the user that an abnormality has occurred. The threshold value _Rth can be appropriately obtained by simulation or actual machine experiment.

As described above, according to the first embodiment, in the abnormality diagnosis of the main shaft bearing 20, the first amplitude PA _k equal to or more than the reference value E in the vibration waveform Wv of 500 to 5000 Hz and the time TA _k at which the first amplitude PA _k occurs. By using, as an evaluation value, the number L of times that the ratio R _k of the vibration waveform to the second amplitude PB _k after the time ΔT has exceeded the threshold value _Rth , the rolling element 28 and the retainer 26 can be used in a normal state and an abnormal state. The abnormality diagnosis can be performed by focusing on the fact that the time waveform shapes of the vibrations generated by the collision with the inner ring 22 or the outer ring 24 are different. As a result, the accuracy of the abnormality diagnosis of the bearing device can be improved.

[Embodiment 2]
In the first embodiment, the ratio between the first amplitude and the second amplitude of the vibration waveform after a lapse of a predetermined time from the time when the first amplitude occurs is set as a threshold as an evaluation value of whether or not an abnormality has occurred. The case where the exceeded number is used as the evaluation value has been described. What is the evaluation value of whether or not an abnormality has occurred, as long as it is a value calculated from the ratio between the first amplitude and the second amplitude of the vibration waveform after a predetermined time has elapsed from the time when the first amplitude occurred. It may be something. In the second embodiment, a case will be described in which the frequency of occurrence when the ratio between the first amplitude and the second amplitude exceeds a threshold is used as an evaluation value of whether an abnormality has occurred.

  The second embodiment differs from the first embodiment in that the frequency of occurrence when the ratio between the first amplitude and the second amplitude exceeds a threshold value is used as the evaluation value of whether or not an abnormality has occurred. In the second embodiment, FIGS. 15 and 16 of the first embodiment are replaced with FIGS. 17 and 18, respectively. Other configurations are the same as those of the first embodiment, and thus description thereof will not be repeated.

FIG. 17 is a diagram illustrating a change in the occurrence frequency when the ratio between the first amplitude PA _k and the second amplitude PB _k exceeds a threshold value. 17, until the time _{TR 21} is a state abnormal spindle bearing 20 has occurred. Is the main shaft bearing 20 is exchanged between the time _{TR 21} and the time _{TR 22,} in after time _{TR 22} spindle bearing 20 is a normal state. As shown in FIG. 17, for example, a value F _d as determination value V _d, the case where the evaluation value V exceeds the determination value V _d is determined to be abnormal, and normal when the evaluation value V is less than or equal to the determination value V _d By making the determination, it is possible to distinguish between a normal time and an abnormal time. As a result, the accuracy of the abnormality diagnosis can be improved as compared with the comparative example.

FIG. 18 is a flowchart for explaining the processing (S1 in FIG. 2) of the subroutine for calculating the evaluation value V used in the second embodiment. As shown in FIG. 18, the data processing device 80 extracts a first amplitude PA _k (k = 1 to N) exceeding the reference value E from the vibration waveform Wv (500 to 5000 Hz) in S11, and proceeds to S12. Proceed. The data processing apparatus 80 advances a first amplitude PA _k is extracted a second amplitude PB _k at time TB _k after time ΔT has elapsed from the time TA _k generated processing to S13 in S12. The data processing device 80 calculates the ratio R _k between the first amplitude PA _k and the second amplitude PB _k in S13, and advances the processing to S24. The data processing unit 80, the ratio _{R k} is calculated a number of times L, which exceeds the threshold value _{R th} in S24, the process proceeds to S25. The data processing unit 80 calculates the occurrence frequency F = L / N where the ratio R _k exceeds a threshold value R _th in S25, and returns the process to the main routine to perform the abnormality diagnosis shown in FIG. The data processing unit 80, evaluation value V is larger than the determination value V _d informs the user that an abnormality has occurred.

As described above, according to the second embodiment, in the abnormality diagnosis of the main shaft bearing 20, the first amplitude PA _k equal to or larger than the reference value E in the vibration waveform Wv of 500 to 5000 Hz and the time TA _k at which the first amplitude PA _k occurs. By using, as an evaluation value, the frequency of occurrence F when the ratio R _k of the vibration waveform to the second amplitude PB _k after the lapse of the time ΔT exceeds the threshold value _Rth , the rolling element 28 can be held between normal and abnormal times. The abnormality diagnosis can be performed by focusing on the fact that the time waveform shapes of the vibrations generated by the collision with the device 26, the inner ring 22, or the outer ring 24 are different. As a result, the accuracy of the abnormality diagnosis of the bearing device can be improved.

  FIG. 19 is a diagram illustrating a result of performing a significant difference test by a t-test on each evaluation value of the comparative example, the first embodiment, and the second embodiment. The t-test is a test method for determining whether there is a significant difference between the average value of a certain sample group and the average value of another sample group. In FIG. 19, there is a significant difference between the average value of the evaluation values when the main shaft bearing 20 is damaged and the average value of the evaluation values when the main shaft bearing 20 is not damaged. Was determined. The number of samples is 21, respectively. The t value, which is a boundary value indicating whether a significant difference is recognized, is 2.021. In the t test, a significant difference is recognized when the test result exceeds the t value.

  As shown in FIG. 19, in the comparative example, since the test result is 0.274, which is smaller than the t value, no significant difference is recognized. On the other hand, in Embodiments 1 and 2, the test results are 2.09 and 6.01, respectively, and both exceed the t value, so that significant differences are recognized.

  From the results of the significant difference test, in the first and second embodiments, the evaluation value in the case where the spindle bearing 20 has no abnormality and the evaluation value in the case where the abnormality has occurred are compared with the comparative example using the effective value as the evaluation value. It can be said that a difference easily occurs with the evaluation value. Therefore, in the first and second embodiments, the occurrence of an abnormality is more likely to appear as a change in the evaluation value. Therefore, according to the first and second embodiments, the accuracy of abnormality diagnosis can be improved as compared with the comparative example.

  The embodiments disclosed this time are also expected to be implemented in appropriate combinations. The embodiments disclosed this time are to be considered in all respects as illustrative and not restrictive. The scope of the present invention is defined by the terms of the claims, rather than the description above, and is intended to include any modifications within the scope and meaning equivalent to the terms of the claims.

  Reference Signs List 1 wind power generator, 10 main shaft, 20 main shaft bearing, 22 inner ring, 24 outer ring, 26 cage, 28 rolling element, 30 blade, 40 gearbox, 50 generator, 70 acceleration sensor, 80 data processing device, 81 first Filters, 83 diagnostics, 90 nacelles, 100 towers.

Claims (4)

An abnormality diagnosis device that detects damage to the bearing device based on vibration data of acceleration of the bearing device,
A filter configured to extract a vibration waveform belonging to a predetermined frequency band from the vibration data,
The evaluation value calculated from the ratio between the first amplitude exceeding the reference value, which is the amplitude of the vibration waveform, and the second amplitude of the vibration waveform after a lapse of a predetermined time from the time when the first amplitude has occurred is a determination value. A diagnosis unit configured to diagnose that the damage has occurred in the bearing device when the value exceeds the threshold value.   2. The diagnostic unit according to claim 1, wherein the diagnostic unit is configured to calculate, as the evaluation value, the number of times that the ratio between the first amplitude and the second amplitude exceeds a threshold value during the measurement period of the vibration data. 3. Abnormal diagnostic device.   The diagnostic unit calculates, as the evaluation value, a ratio between the number of times the ratio between the first amplitude and the second amplitude exceeds a threshold value and the number of times the first amplitude occurs in the measurement period of the vibration data. The abnormality diagnosis device according to claim 1, wherein the abnormality diagnosis device calculates. An abnormality diagnosis method for detecting damage to the bearing device based on vibration data of acceleration of the bearing device,
Extracting a vibration waveform belonging to a predetermined frequency band from the vibration data,
The evaluation value calculated from the ratio between the first amplitude exceeding the reference value, which is the amplitude of the vibration waveform, and the second amplitude of the vibration waveform after a lapse of a predetermined time from the time when the first amplitude has occurred is a determination value. Diagnosing that the damage has occurred in the bearing device when the value exceeds the threshold value.

Images