WO2009096551A1 - Diagnostic system for bearing - Google Patents

Abstract

This object aims to provide a diagnostic system for a bearing, configured to allow a manager to intuitively understand a failure condition of a rotation bearing or a slide bearing set at a slow rotation machine installation and to reliably diagnose the failure. A diagnostic system (10) for a bearing in a rotation machine installation is provided with a damage occurrence detection sensor (20) set on a fixed member of the bearing, a monitoring diagnostic device (30) connected with the sensor, and a diagnosis notification means (40) that displays an abnormality occurrence condition by a percentage. The monitoring diagnostic device (30) is comprised of a memory (37) that stores measurement data detected by the sensor, a standard level calculation unit (50) that calculates a level of the abnormality judgment standard, and a judgment unit (59) that equally divides one rotation time of a continuous rotation of a spindle supported by the bearing or one rotation time by an intermittent operation of the spindle into a plurality of sections, compares the extraordinary judgment standard level with the measurement data of each section, and judges whether or not there is an abnormality situation for every section.

Description

Bearing diagnosis system

The present invention relates to a bearing diagnosis system, and in particular, allows an administrator to intuitively grasp the abnormal state of a bearing used in a rotating machine facility. Specifically, in the signal of the sensor that monitors the bearing state, the signal component indicating the abnormality or damage of the bearing is small, so even if the signal cannot be distinguished from the signal due to disturbance, the bearing state can be evaluated accurately. It is what you are doing.

In general, a large number of rotating machinery equipment is used in industrial plants, and maintenance management of the rotating machinery equipment is performed in order to make maximum use of equipment efficiency. In the maintenance management, monitoring of the state of rolling bearings and plain bearings, which have a high frequency of failure and a large influence, is regarded as the most important.
Usually, in a rotating machine facility operating at a rotational speed of approximately 100 rpm or more, quantification of the facility state represented by bearing state monitoring by measurement of vibration signals and sound signals of the operation state is performed.

For example, in order to diagnose the state of a rolling bearing, a technique is used in which a vibration sensor is attached to a housing of the bearing and damage to the rolling bearing is detected from the vibration waveform.
FIG. 22 shows a vibration waveform of the rolling bearing in which the outer ring is damaged. As shown in FIG. 22, it can be clearly observed that the vibration amplitude is modulated every time the rolling element passes through the damaged part of the outer ring.
However, in a low-speed rotating machine with a rotation speed of 100 rpm or less, the vibration generated by damage is very small. Therefore, such vibration is caused by vibration generated by rotation of the bearing or other factors generated around the bearing. The S / N ratio becomes very poor due to being buried in vibration. For this reason, the state of the bearing cannot be diagnosed by simply detecting the amplitude modulation.
For this reason, conventionally, as a technique for improving the SN ratio and monitoring the state of the bearing reliably, a method of improving the SN ratio by performing special signal processing on the vibration signal, or measuring an AE (acoustic emission) signal. A diagnostic method based on the above has been proposed.

For example, in Japanese Patent Application Laid-Open No. 6-323899 (Patent Document 1), vibration of a low-speed rotating machine is detected by a bearing portion, and the vibration detection signal is subjected to band-pass filter processing to obtain a spectrum domain power at the time of bearing damage. Extract the eigenband component representing the increasing abnormal state, calculate the crest factor of this eigenband component, calculate the crest factor of the calculated crest factor, and compare it with the preset threshold value to diagnose the abnormality of the low-speed rotating machine A method of performing is proposed.
Paying attention to the fact that the peak value due to noise is not so varied, and the period in which the sudden waveform due to bearing damage appears is longer than that due to noise, calculating the crest factor, leveling the noise and lowering the level, due to damage The peak value of the signal component is highlighted.

In addition, as another diagnostic method using a vibration signal, a vibration time waveform that is raised to a power and exceeds a threshold is counted as an event (abnormality). For example, the count number at a certain time interval such as one hour or one day is calculated. There is a method of quantifying the abnormality of the bearing by monitoring the increase / decrease and the increase / decrease of the count number per one rotation.

Further, as a diagnosis by measuring the AE signal, one occurrence of the AE signal is counted as one event, for example, increase / decrease of the count number at regular time intervals such as one hour or one day, or the count number per one rotation. There is a method of quantifying the bearing abnormality by monitoring the increase / decrease of the bearing. Also, the event duration is integrated with one occurrence of the AE signal as one event, for example, the increase or decrease of the event integration time at regular time intervals such as 1 hour or 1 day, or the event integration time per rotation There is also a method of quantifying the bearing abnormality by monitoring the increase and decrease.

However, when monitoring a bearing abnormality from the AE signal or vibration signal count or accumulated time, the administrator of the rotating machine equipment looks at the count number displayed on the warning means such as a monitor or the accumulated time result, and the numerical value itself. Has a problem that the state of failure of the bearing cannot be grasped intuitively.

Furthermore, in the method of monitoring the bearing abnormality by comparing the AE signal with the threshold value, the AE signal can detect a very small change in the equipment state, but reacts sensitively, so it cannot be said to be abnormal. Even if it exists, there are many overdetections that determine that it is abnormal, and there is a problem that it is difficult to handle in measurement and diagnosis.
In addition, the reference value for determining the occurrence of an event (occurrence of an abnormality) is usually a fixed value calculated from the data based on the normal signal collected during the initial adjustment period of the rotating machinery equipment. Many. As described above, when the reference value is a fixed value and a failure diagnosis of a bearing or the like is performed by comparing the reference value and the signal, the signal level changes when the rotational speed of the rotating machine equipment changes and a load fluctuation occurs. There is a problem that an event cannot be correctly judged and a malfunction is induced.

Furthermore, it is not an equipment that rotates continuously more than a plurality of revolutions, but an intermittent operation equipment that performs one operation less than one rotation and repeats the operation, such as a ladle turret that is one of steel facilities. In this case, there is a problem that it is difficult to quantify an abnormal state because data for one rotation cannot be obtained by one measurement.

Furthermore, when time waveform data such as an AE signal or vibration signal is subjected to an FFT operation and a cycle in which an abnormality occurs in the bearing is detected from the frequency analysis result, the amount of data handled in a low-speed rotating facility such as 100 rpm or less is enormous. Therefore, analysis is difficult in a general analysis device.

For example, in a medium / high-speed rotating facility of 1000 rpm, the abnormal cycle time of a bearing that occurs once per rotation is about 60 msec, and the period can be detected by measuring about 1 second even if data acquisition for multiple rotations is considered. Become.
However, in the case of low-speed rotation, the cycle in which the abnormality occurs becomes longer as the rotational speed becomes lower. Therefore, the abnormal cycle time of the bearing that occurs once in one rotation in the rotating facility of 1 rpm is about 60 sec, Considering the data acquisition, it is necessary to measure about 1000 seconds.
Furthermore, in the case of a rotating facility of 0.1 rpm, the abnormal cycle time of the bearing that occurs once per rotation is about 600 seconds, and taking data acquisition for a plurality of rotations requires measurement of about 10,000 seconds.

Normally, in the case of time waveform data measurement for the purpose of bearing abnormality detection, since the measurement is performed using the waveform after the detection process, the sampling frequency is 1 kHz to 3 kHz.
When the sampling frequency is 1 kHz, the number of data required for each rotation speed may be 1000 data for 1000 rpm, but 1000000 data for 1 rpm and 10000000 data for 0.1 rpm. Estimating the data capacity from the number of data, the data capacity is 2 kbytes at 1000 rpm, 2 Mbytes at 1 rpm, and 20 Mbytes at 0.1 rpm. There is a problem that it takes time.

Further, in Japanese Patent No. 3920715 (Patent Document 2), the collected vibration signal is divided at a preset time interval, and the frequency of the vibration signal is analyzed for each divided section to obtain a power spectrum. A signal processing method for improving the S / N ratio by discriminating a signal at a normal location and a signal at an abnormal location from a power value is disclosed.
In the above signal processing method, paying attention to the fact that the total power value is larger in the vibration of the abnormal portion than in the vibration of the normal portion, the average power spectrum in the remaining divided sections obtained by removing the divided section having a large total power value is obtained. It is regarded as a power spectrum at a normal location, and an abnormal vibration is detected by obtaining a ratio between the average power spectrum and the power spectrum for each divided section.

However, in the abnormality diagnosis detection method of Patent Document 2, when disturbance vibrations that occur irregularly occupy most of the vibration waveform in a certain divided section, the total power value of the power spectrum in that section also increases. Therefore, a section with a large total power value does not necessarily correspond to vibration at an abnormal portion of the bearing. As a result, disturbance vibration may be erroneously recognized as vibration at an abnormal portion of the bearing.

JP-A-6-323899 Japanese Patent No. 3920715

The present invention has been made in view of the above problems, and allows an administrator to intuitively grasp the state of failure of a rolling bearing or a sliding bearing attached to a rotating machine facility, without storing a large amount of measurement data. However, it is an issue to reliably perform failure diagnosis.
In addition, it is an object to make it possible to accurately diagnose the state of the bearing even when the disturbance signal is large and the S / N ratio of the signal component indicating abnormality or damage of the bearing is very low.

In order to solve the above-mentioned problem, as a first invention, there is provided a diagnostic system for a bearing in a rotary machine facility,
A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
A monitoring and diagnosis device connected to the sensor;
A diagnostic notification means for connecting with the monitoring diagnostic device and displaying an abnormal state in percentage;
With
The monitoring and diagnosis apparatus includes:
A storage unit for storing measurement data detected by the sensor;
A reference level calculation unit that calculates an abnormality determination reference level based on the measurement data stored in the storage unit;
One rotation time or intermittent operation time of the rotating shaft supported by the bearing is equally divided into a plurality of sections, and the abnormality determination reference level is compared with the measurement data of each section, for each section. And a determination unit for determining whether or not there is an abnormality is provided.
One rotation time or one intermittent operation time of the rotating shaft supported by the bearing is divided into 100 sections for display as a percentage.

According to the bearing diagnosis system of the present invention, the time required for one rotation of the rotating shaft supported by the bearing of the rotating equipment or one intermittent operation of the intermittent operation equipment is divided into 100 sections, and an event is generated for each section. (Abnormality occurrence) is diagnosed and the number of sections where an abnormality has occurred in the rolling bearing is calculated. For this reason, the number of sections in which the abnormality occurs is a percentage of the time (width) of the bearing abnormal state with respect to one rotation time or one intermittent operation time, and the facility administrator looks at the number of sections in which the abnormality has occurred. The time occupied by the abnormal state of the bearing per one rotation time or one intermittent operation time can be intuitively recognized.

In addition, when the abnormality determination is performed using the abnormality occurrence count of the signal such as AE (Acoustic Emission) or the integrated value of the abnormality occurrence time as in the prior art, it cannot be said that the abnormality is caused by the continuous AE occurrence phenomenon. However, in the present invention, one rotation time or one intermittent operation time is divided into a plurality of sections, and the abnormality of the rolling bearing is determined based on the number of sections in which the abnormality has occurred. Therefore, overdetection can be suppressed.

The number of rotations of the rotating shaft of the rotating machine equipment supported by the bearing is preferably 0.1 rpm or more and 300 rpm or less, more preferably 0.1 rpm or more and 150 rpm or less, and further preferably 0.1 rpm or more and 100 rpm or less.
The bearing is preferably a rolling bearing or a sliding bearing.
Furthermore, it is preferable that the sensor is any one of a vibration acceleration pickup, an acoustic emission (AE), an ultrasonic sensor, and a sound detection sensor fixed to the bearing housing.

In the reference level calculation unit of the monitoring and diagnosing device, the abnormality determination reference level is set to a constant multiple of the average value of the measurement data stored in the storage unit for each rotation or one intermittent operation of the rotating shaft supported by the bearing. is doing.
As described above, the abnormality determination reference level is set for each rotation or one intermittent operation of the rotating shaft supported by the bearing, so that the rotation speed and load of the rotating machine equipment to which the bearing is attached are changed. In addition, it is possible to determine the failure of the bearing in accordance with the rotational speed and load.

In the determination unit of the monitoring diagnostic device,
A diagnostic parameter calculator that calculates a diagnostic determination parameter that is the number of sections in which an abnormality has occurred;
It is preferable that a simple diagnosis determination unit that compares the diagnosis determination parameter with a predetermined abnormality determination criterion to determine whether or not there is an abnormality is provided.

The diagnosis parameter is calculated by the diagnosis parameter calculation unit, and the diagnosis determination parameter is compared with the abnormality determination reference value by the diagnosis determination unit. When the time occupied by the abnormal state of the bearing per one rotation time or one intermittent operation time exceeds the abnormality determination reference value, it can be diagnosed that the bearing has failed.

The determination unit of the monitoring / diagnostic apparatus sets an abnormal section only when the section in which abnormality is determined is a set number of adjacent sections within 2 to 10, and abnormal determination in a section less than the set number is an abnormality that is removed as noise. It is preferable that a generation section continuation determination unit is provided.

¡When a failure occurs in a rolling bearing, it is considered that an abnormality does not occur only in one section but an abnormality occurs in a plurality of sections continuously. For this reason, the continuity of adjacent sections is evaluated, and sections where the number of adjacent sections determined to be abnormal are less than the set number are excluded from the total number of sections in which an abnormality has occurred. Noise can be removed, and erroneous determination of failure due to the noise can be prevented.

Furthermore, it is preferable that the determination unit of the monitoring diagnostic apparatus includes an averaging processing unit that averages the diagnosis determination parameters for a plurality of rotations or a plurality of intermittent operations.
By using the diagnosis determination parameters on average, noise included in the measurement data can be removed, and erroneous determination of failure diagnosis due to the noise can be prevented.

It is preferable that the abnormality determination criterion to be compared with the diagnosis determination parameter is set at a plurality of levels such as a caution level and a danger level.
A plurality of abnormality determination criteria are provided, and the occurrence of abnormality is diagnosed by comparing each level of the caution level and the danger level with the diagnosis determination parameter, so that the bearing failure can be diagnosed in more detail.

In addition, the determination unit of the monitoring and diagnosis apparatus calculates the degree of coincidence of the determination results of occurrence of abnormalities for each distinction by shifting the abnormal cycle reference position of the abnormality determination result table for a plurality of rotations or a plurality of intermittent operations. It is preferable to include a synchronous search processing unit that searches for an abnormal cycle reference position where the frequency becomes high and automatically calculates an abnormal cycle from the abnormal cycle reference position.

The synchronous search processing unit gradually shifts the abnormal cycle reference position and creates a correction table based on the abnormality determination result table. The degree of coincidence of the determination result of occurrence of abnormality for each section of the correction table is calculated for each abnormal period reference position, and the abnormal period reference position in the correction table having the highest degree of coincidence is detected. An abnormal period in which an abnormality occurs in the bearing is calculated from the thus determined abnormal period reference position.

Furthermore, the present invention enables accurate fault diagnosis of the bearing state even when the disturbance signal is large and the S / N ratio of the signal component indicating abnormality or damage of the bearing is very low. The second invention and the third invention are provided.

A second invention is a diagnostic system for a bearing in a rotary machine facility,
A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
A monitoring and diagnosis device connected to the sensor;
A diagnostic notification means for connecting to the monitoring diagnostic device and displaying a diagnostic result;
With
The monitoring and diagnosis apparatus includes:
A storage unit for storing a signal waveform detected by the sensor;
A signal waveform calculation unit stored in the storage unit;
The computing unit is
Means for dividing the signal waveform into j = 1, 2,..., N according to the rotational speed of the bearing;
Means for obtaining a frequency spectrum Fjk (j = 1, 2,..., N, k = 1, 2,..., M) for each of the divided sections;
Means for obtaining a spectrum sum Sj (j = 1, 2,..., N) in a predetermined frequency range of the frequency spectrum;
Creating a waveform signal Sx in which the spectrum sums Sj are arranged in time series, and obtaining a frequency spectrum Ss of the waveform signal Sx;
Means for obtaining a kurtosis Ks of the spectral waveform Ss;
The spectrum waveform obtained from the kurtosis Ks is displayed by the diagnosis notification means.

The third invention is
A bearing diagnosis system for rotating machinery equipment,
A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
A monitoring and diagnosis device connected to the sensor;
A diagnostic notification means for connecting to the monitoring diagnostic device and displaying a diagnostic result;
With
The monitoring and diagnosis apparatus includes:
A storage unit for storing a signal waveform detected by the sensor;
A signal waveform calculation unit stored in the storage unit;
The computing unit is
Means for applying a bandpass filter to the signal waveform;
Means for dividing the signal waveform after passing through the band-pass filter into j = 1, 2,..., N according to the rotational speed of the bearing;
Means for determining the rms value RMSj (j = 1, 2,..., N) of the signal or equivalent peak value Pj for each of the divided sections;
Means for generating a waveform signal Sy in which the rms value RMSj (j = 1, 2,..., N) or the equivalent peak value Pj is arranged in time series, and obtaining the frequency spectrum Si of the waveform signal Sy;
Means for obtaining a kurtosis degree Ki of the spectral waveform Si;
The spectrum waveform for which the kurtosis degree Ki is obtained is displayed by the diagnosis notification means.

In the second and third inventions, the degree of kurtosis is obtained from a waveform signal detected by a vibration sensor or the like, an abnormality occurrence state is indicated by the degree of kurtosis, and whether the abnormality is sudden disturbance vibration or It is possible to judge whether it is due to damage.
That is, kurtosis is used as an index representing the degree to which the frequency spectrum of the waveform signal detected by the vibration sensor has a peak at a specific frequency. In the normal distribution, the value of the kurtosis is small, and the distribution having a sharper shape than the normal distribution has a larger value of the kurtosis. Therefore, when the sharpness is high, it indicates that the vibration having periodicity is occurring. Therefore, when the sharpness is high, it indicates that the vibration generated by the bearing damage is included. On the other hand, when the kurtosis is low, the vibration is sudden disturbance vibration, so it can be determined that the vibration is caused by a factor other than the bearing damage.

In any one of the first to third inventions,
Calculated from the bearing device information and / or the rotation speed information, and calculated by the calculation abnormality occurrence period or an integer multiple of the calculation abnormality occurrence period, which is different for each abnormality cause, by the synchronous search processing unit or the calculation unit. An abnormality period is compared, and when the degree of coincidence between the calculation abnormality occurrence period or an integer multiple of the calculation abnormality occurrence period and the abnormality period is high, an abnormality caused by the cause corresponding to the calculation abnormality occurrence period is present in the bearing. It is preferable to include a determination unit that diagnoses the occurrence.
In the case of a rolling bearing, the cause of the abnormality is at least one of inner ring damage, outer ring damage, and rolling element damage,
In the case of a sliding bearing, it is an abnormal metal contact of the rotating shaft that occurs once or a plurality of times in one rotation or one intermittent operation. Specifically, the abnormal metal contact is a metal contact of the rotating shaft due to abnormal vibration of the rotating shaft, a metal contact of the rotating shaft due to an oil film abnormality of the slide bearing, or the like.

The abnormality occurrence cycle differs for each cause of abnormality. In the case of a rolling bearing, the abnormality occurrence cycle corresponding to each abnormality cause is calculated from the bearing device information and the rotation speed information. Further, in the case of a slide bearing, an abnormality occurrence cycle corresponding to the rotation speed is caused by shaft contact. When the calculated abnormal occurrence period or an integer multiple of the calculated abnormal occurrence period is compared with the abnormal period, the cause of the abnormality corresponding to the calculated abnormal occurrence period has occurred in the bearing. Diagnose.
In this way, the abnormality occurrence period corresponding to each abnormality cause is calculated, and the abnormality cause occurring in the bearing is calculated by comparing the abnormality abnormality period with the calculation abnormality occurrence period or an integer multiple of the calculation abnormality occurrence period. Can be identified.

The movable type combining the sensor, the monitoring diagnostic device, and the diagnostic notification unit, or the diagnostic notification unit may be wirelessly connected to the monitoring diagnostic device to be portable.
By making the rolling bearing diagnostic system movable, the diagnostic system can be carried, and a sensor can be attached to the rolling bearing to be diagnosed to diagnose a failure of the rolling bearing.
Further, by wirelessly connecting the monitoring diagnostic apparatus and the diagnostic communication means, the administrator can carry the diagnostic communication means and know the state of the low-speed rotating machine equipment.

As described above, according to the bearing diagnosis system of the first invention, the time required for one rotation of the rotating shaft supported by the bearing of the rotating equipment or one operation of the intermittent operation equipment is divided into approximately 100 sections, The occurrence of abnormality is diagnosed for each section, and the number of sections in which the abnormality has occurred is calculated. For this reason, the number of sections in which the abnormality occurs is a percentage of the time (width) of the bearing abnormal state with respect to one rotation time or one intermittent operation time, and the facility administrator looks at the number of sections in which the abnormality has occurred. The time occupied by the abnormal state of the bearing per rotation time can be intuitively recognized.

In addition, the determination unit creates a correction table by gradually shifting the abnormal cycle reference position from the abnormality determination result table that is the determination result of occurrence of abnormality in each section for a plurality of rotations or a plurality of intermittent operations, and the degree of coincidence becomes highest. An abnormal cycle reference position is detected, and an abnormal cycle in which an abnormality occurs in the bearing is calculated from the abnormal cycle reference position. Anomaly occurrence period is different for each cause of anomaly, and it is generated in the bearing by comparing the anomaly period with the calculated anomaly occurrence period calculated for each anomaly cause or an integer multiple of the calculated anomaly occurrence period. It is possible to identify the cause of abnormality.

According to the second and third aspects of the present invention, the abnormality occurrence state can be visually displayed by displaying the signal waveform detected by the sensor as a spectrum waveform having a sharpness.
This enables quantitative evaluation of the periodicity of signals generated from rotating machinery, so that the state of the bearings can be evaluated even if a signal indicating abnormalities or damage of the bearings is buried in a disturbance signal that occurs suddenly. Can do. As a result, even if the signal S / N ratio is very low as in a low-speed rotating machine, it is possible to detect a bearing abnormality.
Further, since the signal can be divided into small sections and the state of the bearing can be determined from the rms value and the equivalent peak value in the divided section, the present invention is realized using an rms value calculation circuit and an equivalent peak calculation circuit. be able to. As a result, the data processing device does not require a large capacity data storage means or a high speed arithmetic unit. Furthermore, it is possible to diagnose and determine the cause of occurrence of an abnormality in the rolling bearing.

It is a block diagram which shows 1st Embodiment of the rolling bearing diagnostic system which is this invention. It is a block diagram which shows the structure of a monitoring diagnostic apparatus. It is a block diagram which shows operation | movement of CPU. It is explanatory drawing of the operation | movement which divides | segments 1 rotation time into 100 areas and compares the measurement data with an abnormality determination reference | standard level for every area, and diagnoses abnormality generation. It is an example of an abnormality determination result table. It is explanatory drawing of the continuity determination of an abnormality generation area. It is explanatory drawing of the correction table before the determination at the time of the continuity determination of an abnormality generation area, and after determination. It is explanatory drawing which calculates | requires a diagnostic determination parameter determination value. It is explanatory drawing of preparation of the correction table in case rotation time and abnormality occurrence period or m times of abnormality occurrence period differ. It is the figure which showed from which section of the original table the correction table is comprised. It is explanatory drawing of preparation of the correction table in case rotation time and abnormality occurrence period or m times of abnormality occurrence period are the same. It is explanatory drawing which calculates | requires a total matching degree. It is explanatory drawing of the structure of a rolling bearing, (A) is sectional drawing, (B) is a fragmentary sectional view of the perpendicular direction of (A). It is a block diagram of the monitoring diagnostic apparatus which shows 2nd Embodiment. It is a block diagram of the bearing diagnostic system which shows 3rd Embodiment. It is a block diagram of the bearing diagnostic system which shows the modification of 3rd Embodiment. It is a block diagram of the monitoring diagnostic apparatus which shows 4th Embodiment. It is a block diagram of the monitoring diagnostic apparatus which shows 5th Embodiment. (A) (B) is a graph which shows the vibration waveform of the bearing obtained with the conventional diagnostic apparatus. (A), (B), and (C) are graphs showing the kurtosis degrees Ks and Ki of the bearing obtained in the fourth embodiment. It is a figure which shows the frequency spectrum Ss obtained in 4th Embodiment. It is a graph which shows a prior art example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Rolling bearing diagnostic system 11a Rolling bearing 20 Damage occurrence detection sensor 21 Tachometer 30 Monitoring diagnostic device 40 Monitor 50 Time data storage processing unit 51 Diagnostic parameter calculation unit 54 Averaging processing unit 55 Simple diagnosis determination unit 56 Synchronous search processing unit 57 Cause diagnosis processing unit A, Ax Diagnosis determination parameter E Abnormality determination reference level T1 Abnormality determination result table T2 created by the diagnostic parameter calculation unit Abnormality determination result table Ta created by the abnormality occurrence section continuous determination unit Ta correction table Tx Calculation abnormality occurrence Period Ba Abnormal period 61 Vibration sensor 62 Rolling bearing unit 63 Amplifier 64 A / D converter 65, 67 Arithmetic processing unit 66 Output unit 71 Storage unit 72 Waveform dividing unit 73 Frequency spectrum calculating unit 74 Frequency spectrum sum calculating unit 75 Frequency spectrum sum Waveform signal creation Stage 76 waveform signal of the frequency spectrum computing unit 77 the frequency spectrum of the kurtosis calculation means 78 cause of the abnormality judgment means 81 rms value calculating means 82 rms value of the waveform signal generating means

Embodiments of the present invention will be described with reference to the drawings.
1 to 13 show a first embodiment of the first invention.
The bearing diagnosis system 10 for a rotating machine facility according to the present invention diagnoses the state of a rolling bearing 11a of a low-speed rotating motor, which is one of rotating machine facilities 11 installed in a steel facility factory.
As shown in FIG. 1, the diagnostic system 10 includes a damage occurrence detection sensor 20 mounted on a rolling bearing 11 a of a low-speed rotating machinery facility 11 that continuously rotates a plurality of rotations. The monitor / diagnosis apparatus 30 connected, the monitor 40 which comprises the diagnostic notification means connected with the monitor / diagnosis apparatus 30, and the tachometer 21 which measures the rotation speed of the motor are provided.

The damage occurrence detection sensor 20 is an AE sensor that detects acoustic emission, and is fixed to a bearing housing 11d of a low-speed rotation motor 11c provided on the foundation frame 11b with screws. The tachometer 21 is attached to the load side bearing mount 11e. The attachment position of the damage occurrence detection sensor 20 is not limited to the above position.
Further, the rotational speed information of the motor may be stored in advance in the microcomputer 34 of the monitoring / diagnosis apparatus 30 without attaching the tachometer 21. Furthermore, instead of the damage occurrence detection sensor 20, a vibration acceleration pickup that detects vibration, an ultrasonic sensor, or a sound detection sensor may be used.

As shown in FIG. 2, the monitoring / diagnosis device 30 includes a signal amplification circuit 31, a filter circuit 32, a detection circuit 33, and a microcomputer 34.
The signal amplifier circuit 31 is an amplifier that amplifies the signal measured by the damage occurrence detection sensor 20.
The filter circuit 32 is a band-pass filter that removes noise components from the signal amplified by the signal amplifier circuit 31. In the case of the vibration acceleration pickup, the filter circuit 32 is a filter that passes a band of 1 kHz to 20 kHz. When an AE sensor is used, a filter that passes a band of 50 kHz to 500 kHz is used.
The detection circuit 33 detects (envelope processing) the signal from which noise has been removed by the filter circuit 32. Depending on the signal, the detection circuit 33 may not be provided, and in the case of a vibration signal, a power process may be performed instead of detection.

The microcomputer 34 includes a CPU 35, a ROM 36, a RAM 37, and a port 38.
The port 38 of the microcomputer 34 A / D converts the signal from the detection circuit 33 at a predetermined sampling period.
The ROM 36 and the RAM 37 constitute a storage unit. The RAM 37 stores an abnormality determination result table calculated by the CPU 35 and temporarily stores data calculated by the CPU 35. The ROM 36 stores the operation of the CPU as software.
As shown in FIG. 3, the CPU 35 includes a reference level calculation unit 50 and a determination unit 59. The determination unit 59 includes a diagnostic parameter calculation unit 51, an abnormality occurrence section continuation determination unit 52, an averaging processing unit 54, Each unit includes a simple diagnosis determination unit 55, a synchronous search processing unit 56, and a cause diagnosis processing unit 57. The CPU 35 reads out the software from the ROM 36 and executes it to operate each unit.

The reference level calculation unit 50 receives the rotation speed information of the motor from the tachometer 21 and calculates a time required for one rotation of the motor (one rotation time). In addition, the time data output from the port 38 is stored in the RAM 37 for a time length of one or more rotations of the motor.
In addition, without providing the tachometer 21, the rotation number information of the motor 11c may be stored in advance in the reference level calculation unit 50, and one rotation time may be calculated using the rotation number information.

The reference level calculation unit 50 calculates the abnormality determination reference level E. First, an average value of one rotation time length data, which is time data stored in the RAM 37, is obtained from the equation (1).
Next, the abnormality determination reference level E is obtained from equation (2). In this embodiment, m is 4, but may be 2.0 or more and 10.0 or less according to the measurement target.
Average value per rotation = Total time length data for one rotation ÷ Number of time length data for one rotation Formula (1)
Abnormality judgment reference level E = 1 rotation average value × m Formula (2)

Further, as shown in FIG. 4, the diagnostic parameter calculation unit 51 divides one rotation time into equal intervals of 1/100, and sets each section as section No. 1 to Section No. 100. Section No. 1 in the order of the measured data for one rotation time length. 1 to Section No. Sort into 100 sections.
Next, section no. 1 to Section No. It is determined whether there is data exceeding the abnormality determination reference level E in the data allocated to each of the 100 sections. For example, in FIG. 6, 7 and 8 are abnormal occurrence sections because the data exceeds the abnormality determination reference level E.

A table T1 of the abnormality determination result is created for each rotation, with the section having data exceeding the abnormality determination reference level E as the abnormality occurrence section Y. FIG. 5 is an example of the abnormality determination result table T1, and shows whether or not an abnormality has occurred for each section with respect to the rotation of the motor from the latest to nine rotations.
Further, the total value of the number of sections in which an abnormality has occurred within one rotation time is calculated and used as the diagnosis determination parameter A. That is, the diagnosis determination parameter A is the number of abnormality occurrence sections (Y) within one rotation time length.

The abnormality occurrence section continuation determination unit 52 detects a section in which an abnormality occurs continuously from the table T1 of the abnormality determination result obtained by the diagnosis parameter calculation unit 51. For example, as shown in FIG. 6-No. In FIG. 8, the abnormality occurs continuously for three sections, and the number of consecutive sections is three. In addition, section No. 11-No. 14 is the number of consecutive occurrence sections is four. On the other hand, the section No. 98 is a section where an abnormality has occurred. In 97 and 99, no abnormality has occurred, and the number of consecutive occurrence sections is one.

After detecting the abnormal continuous occurrence section in this manner, the abnormal occurrence section continuous determination unit 52, as shown in FIG. 7, no abnormality has occurred in the section where the number of consecutive occurrence sections is less than the number k of continuity determination sections ( N) and excluded from the total number of abnormality occurrence sections (diagnosis determination parameter A). In FIG. 7, when k = 4 and the number of continuously generated sections is 3 or less, it is assumed that no abnormality has occurred. The continuity determination interval number k may be selected from 2 to 10.
The abnormality occurrence section continuation determination unit 52 creates an abnormality determination result table T2 that is corrected by determining continuity for each rotation, and stores it in the RAM 37.

The averaging processing unit 54 uses the correction abnormality determination result table T2 for the latest n rotations, as shown in FIG. 8, to obtain a diagnosis determination parameter An that is the number of sections in which an abnormality has occurred for each rotation.
Further, the average of the diagnosis determination parameters An for the latest n rotations is obtained from the equation (3), and is set as the diagnosis determination parameter determination value Ax. In FIG. 8, the average of the diagnostic determination parameters An for the latest eight rotations is obtained with n = 8, but n may be any value from 1 to 16.

The diagnosis determination parameter determination value Ax is averaged every rotation and is a moving average value. That is, as shown in FIG. 8, when the averaging processing unit 54 receives a table of abnormality determination results before two rotations, a diagnosis determination parameter determination value Ax1 is obtained from the diagnosis determination parameters An before nine rotations and two rotations before. When the table before one rotation is received, the diagnosis determination parameter determination value Ax2 is obtained from the diagnosis determination parameter An from one rotation before eight rotations. When the most recent table is received, the most recent diagnosis determination parameter from seven rotations before The diagnosis determination parameter determination value Ax3 is obtained from An.

The simple diagnosis determination unit 55 stores the attention level determination reference value as 5, for example, and the danger level determination reference value as 15, for example. The diagnosis determination parameter determination value Ax is compared with the determination reference value every rotation, and if the diagnosis determination parameter determination value Ax is larger than each determination reference value, a failure of a caution level and a danger level occurs in the rolling bearing 11a. It is determined that there is.
The simple diagnosis determination unit 55 displays the determination result on the monitor 40.

Next, identification of the cause of abnormality of the rolling bearing 11a will be described.
When the simple diagnosis determination unit 55 determines that a failure of the attention level or the danger level has occurred, the synchronization search processing unit 56 and the cause diagnosis processing unit 57 identify the cause of the abnormality of the rolling bearing 11a. .
The synchronous search processing unit 56 and the cause diagnosis processing unit 57 identify inner ring damage, outer ring damage, and rolling element damage as an abnormality of the rolling bearing 11a. When the rolling bearing 11a is damaged, although details will be described later, an abnormality occurs at a different cycle (abnormality generation cycle) depending on the damaged portion. The synchronous search processing unit 56 uses the abnormality determination result table T2 to obtain the abnormal cycle Bm that is closest to the rotation cycle of the rotating machine, and the cause diagnosis processing unit 57 determines the cause of the failure from the relationship between the abnormal cycle Bm and the abnormality occurrence cycle. I have identified.

Details of the operations of the synchronous search processing unit 56 and the cause diagnosis processing unit 57 will be described.
The synchronous search processing unit 56 automatically detects the abnormal cycle reference position and determines the abnormal cycle Bm. The abnormal period Bm is a period obtained by multiplying the abnormality generation period in which an abnormality occurs due to damage to the rolling bearing 11a by m, and the value of m is a value that makes the abnormal period Bm closest to the rotation period of the rotating machine. The abnormal cycle Bm is not necessarily synchronized with the rotation cycle of the rotating machine, and the abnormal cycle Bm may or may not be the same cycle as the rotation cycle of the rotating machine.
A method for obtaining the abnormal period Bm will be described below.
First, the correction table Ta is created using the abnormality determination result table T2 obtained by the averaging processing unit 54.
In the diagnosis parameter calculation unit 51, the abnormality determination result table T2 is created so as to be 100 sections based on the rotation cycle reference position that divides the rotation cycle. First, the abnormality cycle Bm is based on the abnormality determination result table T2 in the past direction. When it is assumed that the period is 101 sections long by one section, a correction table Ta having 100 sections is created.

As shown in FIG. 9, the abnormality determination result table T2-0 to the table T2-4 are arranged, and the section No. of the latest table T2-4 is arranged. With reference to 100, the abnormal period Bm is divided in the past direction. Since the abnormal period Bm is assumed to be 101, the section No. of the table T2-4 is past in the past direction. 100 to the section No. in table T2-3. 101 section up to 100 is abnormal cycle BmA, section No. of table T2-3. 99 to section T.2 of table T2-2. 101 section up to 99 is abnormal cycle BmB, section No. of table T2-2. 98 to section T. 2 of table T2-1. 101 section up to 98 is abnormal cycle BmC, section No. of table T2-1. 97 to section T.0 of table T2-0. 101 section up to 97 is the abnormal period BmD.

Here, the abnormal cycle reference position is set as the section No. in table T2-4. 100, section No. of table T2-3. 99, section No. of table T2-2. 98, section No. of table T2-1. 97, and the abnormal cycle reference position is defined as the 100th section of the correction table Ta, and 100 sections are extracted from the abnormal cycle reference position in the past direction, and the correction table Ta-4 to the abnormal cycle BmA, BmB, BmC, BmD Ta-1. At this time, for example, the section number of table T2-3. 100 etc. do not enter the correction table Ta.

FIG. 10 is a diagram showing from which section of the original table T2 the correction tables Ta-4 to Ta-1 are configured. FIG. 9 shows a case where the search interval difference in FIG. Similarly, when the rotation time is assumed to be 102 or 103 in the past direction by 2 or 3 sections longer in the table T2, that is, when the search section difference in FIG. Ask for. In FIG. 10, only the search section difference is described up to 3, but the search section difference is set to 0 to 10 to obtain the correction table Ta. The search interval difference may be set to 0-20.
The search section difference 0 is the case where the original table T2 and the correction tables Ta-4 to Ta-1 are the same, and shows the case where the rotation period and the abnormal period Bm coincide as shown in FIG.

Next, the total matching degree h is calculated for each of the correction tables Ta having the search section differences of 0 to 10 thus obtained.
The total matching degree h is obtained as follows.
First, in the correction tables Ta-4 to Ta-1 for the latest four rotations, the same section No. The number of occurrences of abnormality that is the number of occurrences of abnormality is obtained. For example, the section no. In 3, since an abnormality has occurred in the correction tables 1 and 4, the number of abnormality occurrence determinations is 2.
Next, for each section, the section number. The separate matching degree f is obtained from equation (7).
j is the number of correction tables. In FIG. 10, since the correction tables for the latest four rotations are used, j is four.
Section No. Another degree of coincidence f = number of abnormality occurrence determination (Y) ÷ j × 100% Equation (7)

Further, an abnormality occurrence (Y) determination interval number k which is the number of intervals in which an abnormality has occurred is obtained, and an interval No. Section Nos. 1 to 100 Σf which is the sum of the different matching degrees f is obtained, and the total matching degree h is obtained from the equation (8).
Total matching degree h = Σf ÷ k equation (8)

The total matching degree h is obtained for each of the correction tables Ta when the search section difference is 0 to 10, and the correction table Ta having the highest total matching degree h is selected.
The abnormal cycle Bm is determined from the abnormal cycle reference position to the next abnormal cycle reference position in the selected correction table Ta.

That is, the abnormal period Bm is the number of sections obtained by the equation (9). Since the time Ta per section is obtained by one rotation time R ÷ 100, the abnormal cycle Bm is expressed by the equation (10). All of these operations are automatically performed in the synchronous search processing unit 56.
Number of sections Ea = search section difference of selected correction table Ta + 100 (9)
Abnormal cycle Bm = time Ta × number of sections Ea per section (10)

The cause diagnosis processing unit 57 specifies the cause of the failure of the rolling bearing 11a using the abnormal cycle Bm obtained by the synchronous search processing unit 56.
FIG. 13 shows the structure of the rolling bearing 11a, an outer ring 60 fixed to the rotating machine case, an inner ring 61 fixed to the shaft of the rotating machine and arranged concentrically with the outer ring, and between the outer ring 60 and the inner ring 61. It consists of a plurality of spherical rolling elements 62 that are arranged to roll freely.

Possible causes of failure of the rolling bearing 11a include inner ring damage, outer ring damage, and rolling element damage. When these damages occur, the abnormality occurrence period for each damage is expressed by the equation (11) according to the geometric dimension of the rolling bearing 11a. It is calculated | required by-Formula (13). That is, the number of rolling elements is Z, the rolling element diameter is d, the pitch circle diameter is D, the contact angle between the rolling elements 62 and the rolling surface is α, and the rotational frequency of the rotating machine to which the rolling bearing 11a is attached is fr. Then, the abnormality occurrence period Tin when the inner ring is damaged is expressed by Expression (11), the abnormality generation period Tout when the outer ring is damaged is expressed by Expression (12), and the abnormality generation period Tball when the rolling element is damaged is expressed by Expression (13). The rotation frequency fr is the number of rotations (rpm) from the tachometer divided by 60.

These abnormality occurrence cycles Tin, Tout, and Tball are multiplied by s as a calculation abnormality occurrence cycle Tx, and calculation is performed by changing s from 1 to 10. In other words, the calculation abnormality occurrence period Tx is the calculation result time Tin1 to Tin10 of Tin × s, the calculation result time Tout1 to Tout10 of Tout × s, and the calculation result time Tball1 to Tball10 of Tball × s.
The abnormal cycle Bm obtained by the synchronous search processing unit 56 is compared with the calculated calculation abnormality occurrence cycle Tx, that is, Tin1 to Tin10, Tout1 to Tout10, and Tball1 to Tball10. When the value of the calculation abnormality occurrence period Tx substantially coincides with the abnormality period Bm, the damage corresponding to the calculation abnormality occurrence period Tx is diagnosed as an abnormality cause, and these diagnosis results are displayed on the monitor.
Specifically, a predetermined time width is determined around the abnormal period Bm, and when the value of the calculated abnormality occurrence period Tx is within the time width, the damage corresponding to the calculated abnormality occurrence period Tx is diagnosed as the cause of the abnormality. . For example, when Tin 10 is within a predetermined time width centered on the abnormal period Bm, it is diagnosed that the inner ring is damaged.
There may be a plurality of abnormal causes. For example, when Tin 10 and Tout 8 are within a predetermined time width centered on the abnormal period Bm, it is diagnosed that the inner ring and the outer ring are damaged.

In the present embodiment, a sliding bearing may be attached to the rotary machine equipment 11. In this case, if an abnormality occurs due to an abnormal metal contact of the rotating shaft and the abnormality occurs once per rotation, the calculation abnormality occurrence period Tx is obtained by 1 / fr, and the calculation abnormality generation period Tx The cause of the abnormality is diagnosed by comparing the abnormal period Bm.
Further, even when an abnormality occurs multiple times (2 to 10 times) in one rotation in the slide bearing, the abnormality occurrence period Tn is obtained by one rotation period / number of abnormality occurrences per rotation (p), and the abnormality occurrence period Tn The cause of the abnormality is diagnosed by comparing p times the number of times with the abnormal period Bm.

According to the present invention, the time required for one rotation of the rolling bearing 11a of the equipment rotating at a low speed or one operation of the intermittent operation equipment is divided into about 100 sections, and the occurrence of abnormality (occurrence of abnormality) is diagnosed for each section. The number of sections in which an abnormality has occurred is calculated. For this reason, the number of sections in which the abnormality occurs is a percentage of the time (width) of the bearing abnormal state with respect to one rotation time or one operation time, and the facility administrator looks at the number of sections where the abnormality has occurred, The time occupied by the abnormal state of the bearing per rotation time can be intuitively recognized.

Further, the synchronization search processing unit 56 and the cause diagnosis processing unit 57 gradually shift the abnormal cycle reference position from the abnormality determination result table that is the determination result of abnormality occurrence in each section for a plurality of rotations or a plurality of intermittent operations. And the abnormal cycle reference position with the highest degree of coincidence is detected. An abnormal period in which an abnormality occurs in the bearing is calculated from the thus determined abnormal period reference position. The abnormality occurrence period is different for each abnormality cause, and the degree of coincidence can be determined by comparing the abnormality abnormality period with the calculation abnormality occurrence period calculated for each abnormality cause or an integer multiple of the calculation abnormality occurrence period. When there is a high calculation abnormality occurrence cycle, it can be diagnosed that an abnormality cause corresponding to the calculation abnormality occurrence cycle is occurring in the bearing.

FIG. 14 shows a second embodiment. The second embodiment is an embodiment of the first invention.
In the second embodiment, one operation operation of the rotating facility is less than one rotation, and the rolling bearing 11a of the intermittent operation facility that repeatedly performs the operation is set as a diagnosis target.
An example of the intermittent operation facility is a ladle turret that is one of steel facilities. The ladle turret repeats the operation of 1/2 rotation at a low speed of 1 rpm per operation.
As shown in FIG. 14, a limit switch 22 is connected to a microcomputer 34 instead of the tachometer 21. The limit switch 22 causes the CPU 35 to detect operation start and operation stop in the intermittent operation facility.

In this case, the reference level calculation unit 50 of the CPU 35 stores the signal from the damage occurrence detection sensor 20 in the RAM 37 for a length equal to or longer than one intermittent operation time.
In addition, the diagnostic parameter calculation unit 51 obtains a one-time intermittent operation average value by dividing the total of one-time intermittent operation time length data stored in the RAM 37 by the number of data. The abnormality determination reference level E is obtained from the one-time intermittent operation average value × m.
Further, the one-time intermittent operation time is divided into 100 sections, and an abnormality determination result table is created by comparing with the abnormality determination reference level E for each section.

The averaging processing unit 54 obtains the diagnosis determination parameter A and the determination value of the diagnosis determination parameter A using the abnormality determination result table for the latest j times of intermittent operation and the table obtained in the same equipment operation state. Yes.
The same equipment operation state means that in the case of a motor that rotates 1/2 turn in one operation and repeats 1/2 turn operation as A → B → A → B → A → B, The operation state of B is said. As described above, the predetermined processing is performed using the tables obtained in the A operation state or the tables obtained in the B operation state.
Also in the synchronous search processing unit 56, the correction table Ta is obtained using tables obtained in the same equipment operation state.

According to the present invention, even in the case of the rolling bearing 11a attached to the intermittent operation facility, the number of sections in which the abnormality has occurred is expressed as a percentage of the time (width) of the bearing abnormal state with respect to one operation time. The manager can intuitively recognize the time occupied by the abnormal state of the bearing per operation time by looking at the number of sections in which the abnormality has occurred.

In addition, since another structure and an effect are the same as that of 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

FIG. 15 shows a third embodiment. The third embodiment is an embodiment of the first invention.
The diagnosis system 10 according to the third embodiment is a movable type in which the monitoring diagnosis device 30 and the monitor 40 are combined. The monitoring / diagnosis device 30 and the monitor 40 are, for example, laptop computers.
The monitoring / diagnosis device 30 is connected to the damage occurrence detection sensor 20, and the damage occurrence detection sensor 20 is attached to the bearing housing 11d with a magnet. Alternatively, a dedicated jig may be attached to the damage occurrence detection sensor 20, and the sensor may be fixed by manually pressing the sensor against the bearing housing 11d. Further, in a place where a person cannot approach the bearing housing 11d, only the damage detection sensor 20 is fixed to the bearing housing 11d in advance with a screw or an adhesive, and a portable measuring instrument is connected to the sensor for diagnosis. Also good.
By making the diagnostic system 10 portable as a movable type, it is possible to carry out the diagnostic system 10 to the rolling bearing 11a to be diagnosed and perform an abnormality diagnosis.
In addition, since another structure and effect are the same as that of 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

FIG. 16 shows a modification of the third embodiment, in which a diagnostic device 50 in which a damage detection sensor 20, a monitoring diagnostic device 30, and a tachometer 21 are integrated is fixed to a rolling bearing 11a. Further, the monitoring / diagnosis apparatus 30 is provided with a communication unit, and is wirelessly connected to a mobile phone 41 constituting the diagnosis notification unit. The diagnosis device 50 transmits the diagnosis result obtained by the simple diagnosis determination unit 55 to the mobile phone 41.
When the administrator carries the mobile phone 41, the diagnosis result of the rotating machine equipment 11 can be received even at a location away from the rotating machine equipment 11.
In addition, since another structure and effect are the same as that of 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

FIG. 17 shows a fourth embodiment. The fourth embodiment is an embodiment of the second invention.
In the fourth embodiment, a vibration sensor 61 is attached to a rolling bearing unit 62 on one side of a roll rotating at 46 rpm of a rotating machine facility, and the state of the bearing is monitored.
The output signal of the vibration sensor 61 is amplified to a predetermined level by the amplifier 63, and then transmitted to the A / D converter 64. The output signal of the A / D converter 64 is transmitted to the arithmetic processing unit 65, and the calculation is performed. The processing device A is connected to an output device 66 comprising a diagnostic notification means such as a monitor.

The arithmetic processing unit 65 includes a storage device 71 connected to an A / D converter 64, a waveform dividing unit 72, a frequency spectrum calculating unit 73, a frequency spectrum sum calculating unit 74, a frequency spectrum sum waveform signal generating unit 75, and a waveform signal. Frequency spectrum calculation means 76, frequency spectrum kurtosis calculation means 77, and abnormality cause determination means 78.
The A / D converter 64 discretizes the vibration waveform at a sampling frequency of 50 kHz, the arithmetic processing unit 65 calculates the kurtosis degree Ks from the discrete waveform, and the calculated kurtosis degree Ks by the output unit 66 in the past. At the same time as displaying a trend graph together with the calculation result, an alarm is displayed when the kurtosis Ks reaches a predetermined level.

The rotational speed of the bearing unit 62 whose state is monitored in advance is input to the arithmetic processing unit 65, and a vibration waveform of one rotation or more of the bearing can be stored in the storage device 71.
The waveform dividing means 72 divides the discretized vibration waveform into 128 sections (j = 1, 2,..., 128) every 512 points. The vibration waveform discretized at 50 kHz is divided into one divided section every 512 points, and one divided section has a time length of 10.24 msec. Therefore, when the 128 divided sections are combined, the vibration waveform between 128 × 10.24 msec = 1.31 s is analyzed, which is equal to the vibration waveform for one rotation of the bearing.
The frequency spectrum calculation means 73 performs a fast Fourier transform on the vibration waveform subjected to the anti-aliasing filter processing to perform frequency spectrum Fjk (j = 1, 2,..., 128, k = 1, 2,. 512).
The frequency spectrum sum calculation means 74 calculates the sum Sj (j = 1, 2,..., 205) of the spectrum Fjk (k = 11, 12,..., 205) corresponding to the frequency spectrum in the 1 kHz to 20 kHz band among the calculated Fjk. 128).
A waveform signal Sx is created by arranging Sj by the waveform signal creation means 75 of the frequency spectrum sum.
The frequency spectrum calculating means 76 of the waveform signal calculates the frequency spectrum Ss of the waveform signal Sx, and the sharpness calculating means of the frequency spectrum calculates Equation (14) to obtain the kurtosis Ks.

In Expression (14), xi represents each spectrum of the frequency spectrum Ss, and xa is an average value thereof. N ₀ is the number of xi.

The abnormality cause determination means 78 is inputted in advance with the number of rolling elements Z of the rolling bearing of the bearing unit 12, the diameter d of the rolling elements, the pitch circle diameter D, and the contact angle α between the rolling elements 62 and the transfer surface. Together with the speed, the period of occurrence of anomaly when damage occurs is calculated. Further, from the peak frequencies fs1, fs2, fs3,... Of the frequency spectrum Ss obtained by the frequency spectrum calculating means 76 of the waveform signal, the corresponding periods Ts1, Ts2, Ts3,. If they match, it is determined that the abnormality has occurred in the rolling bearing of the bearing unit 12.

Hereinafter, an abnormality diagnosis function using the vibration sensor 61 as a sensor for monitoring the bearing state will be described.
A vibration waveform indicating the state of the bearing is detected by the vibration sensor 61, the detected vibration waveform is collected and stored, and transmitted to the arithmetic processing unit 65 via the amplifier 63 and the A / D converter 64.
In the arithmetic processing unit 65, the storage unit 71 stores the vibration waveform, and the waveform dividing means 72 divides the stored waveform into n sections. Here, for example, if the time length of the divided section is tmsec, it is desirable to determine the divided section function n according to the rotational speed of the bearing. That is, if the division function n is set in accordance with the number of rotations of the bearing, a vibration waveform having a length of time corresponding to one rotation, two rotations, three rotations,...

When the inner ring, outer ring, rolling element, and cage of a rolling bearing are damaged, abnormal vibration occurs when the rolling element passes through the damaged part, and almost the same vibration as normal state without damage occurs when it does not pass through the damaged part. To do. Further, when vibration due to disturbance is added to this, the divided sections are (1) a section where vibration close to a normal state occurs, (2) a section where disturbance vibration is added to vibration close to a normal state, and (3) abnormalities. It can be roughly divided into a section in which vibration is generated and (4) a section in which disturbance vibration is added to abnormal vibration.

The frequency waveform Fjk (j = 1, 2,..., N, k = 1, 2,..., M) of the vibration waveform divided by the waveform dividing means 72 is divided by the frequency spectrum calculating means 73 for each divided section. )
Next, the sum of the frequency spectrum Fjk is obtained by the frequency spectrum sum calculating means 74.
Next, a waveform signal is created by the waveform signal signal means 7 having a frequency spectrum sum.

The frequency spectrum is the lowest in the section where the vibration close to the normal state (1) occurs, and the frequency spectrum is the highest in the section where the disturbance vibration is added to the abnormal vibration (4). Therefore, when the sum of the frequency spectra is obtained, it can be determined whether the divided section includes abnormal vibrations, disturbance vibrations, or both depending on the magnitude of the sum of the frequency spectra Fjk.
Here, when obtaining the sum of the frequency spectrum Fjk, it is desirable to obtain the sum by limiting to a spectrum in a specific frequency range.
Usually, the vibration generated by damage to the rolling bearing is a relatively high frequency vibration of 1 kHz to 40 kHz. For example, it is clear that a vibration of several tens of Hz is a disturbance vibration of a low frequency. The state can be determined more accurately by obtaining the spectral sum Sj (j = 1, 2,..., N) from which the spectral components have been deleted.

The spectrum sum Sj (j = 1, 2,..., N) is a representative value of n divided sections, and the waveform signal Sx in which these are arranged in a time series has a high value in a portion including abnormal vibration or disturbance vibration. As shown, a portion including vibrations close to normal is a time waveform signal indicating a low value.
The vibration of a rotating machine such as a bearing is generated with rotation and thus has periodicity. On the other hand, since the disturbance signal is generated suddenly and randomly, there is no periodicity. For this reason, when the frequency of the waveform signal Sx is analyzed, a spectrum having a peak at a specific frequency is obtained in the vibration generated by the bearing abnormality, whereas the disturbance vibration generated at random has a shape in which the frequency spectrum is dispersed.

In the present embodiment, the kurtosis degree Ks is used as an index representing the degree to which the frequency spectrum Ss of the waveform signal Sx has a peak at a specific frequency.
In the normal distribution, the kurtosis is 3, and as the distribution has a sharper shape than the normal distribution, the value of the kurtosis increases. Therefore, when the kurtosis Ks is high, it indicates that vibration having periodicity is generated, and therefore vibration generated by damage to the bearing is included. Conversely, when the kurtosis Ks is low, the vibration is a sudden disturbance vibration, so it can be determined that the vibration is caused by a factor other than the bearing damage.
Therefore, information for determining the state of the bearing can be obtained by displaying the value of the kurtosis Ks and its change with time as a trend graph.

FIG. 18 shows a fifth embodiment. The fifth embodiment is an embodiment of the third invention.
Similar to the fourth embodiment, a vibration sensor 61 is attached to the rolling bearing unit 62 on one side of a roll rotating at 46 rpm to monitor the state of the bearing.
The output signal of the vibration sensor 61 is amplified to a predetermined level by the amplifier 63, passes through a bandpass filter 80 having a passband of 1 kHz to 20 kHz, and is then transmitted to the A / D converter 64.
In the fifth embodiment, an analog filter is used as the bandpass filter 80, but after the A / D converter 64 stores the vibration waveform after A / D conversion, the stored vibration waveform passes through the digital filter. It is possible to extract a component of 1 kHz to 20 kHz.
An output signal from the A / D converter 64 is output to the arithmetic processing unit 67.

The arithmetic processing unit 67 includes a storage device 71 and a waveform dividing unit 72, an rms value calculating unit 81 connected to the waveform dividing unit 72, and an rms waveform signal generating unit 82. The rms waveform signal creation means 82 is connected to the frequency spectrum calculation means 76 of the waveform signal, and the frequency spectrum calculation means 76 of the waveform signal is connected to the frequency spectrum kurtosis calculation means 77 and the abnormality cause determination means 78. To the output means 66.
The kurtosis is calculated by the frequency spectrum kurtosis calculation means 77 from the discrete waveform of the frequency spectrum calculated by the frequency spectrum calculation means 76 of the waveform signal, and the calculated kurtosis is displayed by the output device 66. The same as in the fourth embodiment.

In the arithmetic processing unit 67, the waveform dividing means 72 divides the vibration waveform into a total of 128 sections (j = 1, 2,..., 128), with 512 discrete vibration waveforms as one divided section. ing.
The rms value calculation means 81 obtains RMSj (j = 1, 2,..., 128) for each divided section.
The process for obtaining the waveform signal Sy, the frequency spectrum Si of the waveform signal Sy, and the kurtosis Ki of the frequency spectrum Si from the obtained RMSj is the same as in the fourth embodiment.
Further, if the rms value calculation means 81 is replaced with equivalent peak calculation means, the kurtosis degree due to the equivalent peak Pj can be obtained in exactly the same manner. The cause of damage is also diagnosed and determined in the same manner as in the fourth embodiment.

The fifth embodiment employs the following method as a method of abnormality diagnosis.
Conventionally, Fast Fourier Transform has been widely used as a method for calculating a frequency spectrum from a time waveform signal.
A discrete time waveform signal f (n) (n = 0, 1,..., N−1) and a discrete Fourier spectrum between F (k) (k = 0, 1,..., N−1) The Parsval equation shown in equation (15) holds.

Therefore, although there is a mathematical difference between the simple absolute sum and the square sum, if the sum of the frequency spectrum in the divided section is large, the square sum of the vibration waveform and the simple absolute sum are also considered to be large. . In addition, in the vibration waveform, vibration is measured with reference to the balance position, so the average value is almost zero. In this way, it is recognized that the rms value or equivalent peak value of the vibration waveform in the divided section and the sum of the frequency spectrum have a correlation.

In the fourth embodiment, the above relationship is used, and the sum of the frequency spectra in the divided sections is replaced with an rms value or an equivalent peak value. However, as described above, in order to associate the rms value or equivalent peak value with the spectrum sum obtained by limiting to a specific frequency range, the rms value or equivalent peak is previously applied to the vibration waveform after passing through the bandpass filter. You must find the value.
A waveform signal in which the rms value obtained for each divided section j = 1, 2,..., N is RMSj and the equivalent peak value is Pj with respect to the vibration waveform after passing through the band-pass filter. If Sy is created and the kurtosis degree Ki of the frequency spectrum Si is obtained, the state of the bearing can be determined in the same manner as the method using the sum of the frequency spectra.

Further, the occurrence of abnormalities in the rolling bearing can be considered to be inner ring damage, outer ring damage, and rolling element damage. When these damages occur, the abnormality generation period for each damage depends on the geometric dimensions of the rolling bearing. It is obtained by the equations (11) to (13) described in the embodiment.
That is, if the number of rolling elements is Z, the rolling element diameter is d, the pitch circle diameter is D, each contact between the rolling elements 62 and the transfer surface is α, and the rotational frequency of the rotating machine to which the rolling bearing 11a is attached is fr. The abnormality occurrence period Tin when the inner ring is damaged is expressed by Expression (11), the abnormality generation period Tout when the outer ring is damaged is expressed by Expression (12), and the abnormality generation period Tball when the rolling element is damaged is expressed by Expression (13). The rotation frequency fr is the number of rotations (rpm) from the tachometer divided by 60.

The calculation is performed by changing s from 1 to 10 with s times the abnormality occurrence period Tin, Tout, and Tball as the calculation abnormality occurrence period Tx. In other words, the calculation abnormality occurrence period Tx is the calculation result time Tin1 to Tin10 of Tin × s, the calculation result time Tout1 to Tout10 of Tout × s, and the calculation result time Tball1 to Tball10 of Tball × s.

Assuming that specific frequencies at which the frequency spectrum Ss of the waveform signal Sx shows a peak are fs1, fs2, fs3,... (A plurality of peaks may occur), the corresponding periods are Ts1 = 1 / fs1, Ts2 = 1 / fs2, Ts3 = 1 / fs3,... When the value of the calculation abnormality occurrence period Tx substantially coincides with the periods Ts1, Ts2, Ts3,..., The damage corresponding to the calculation abnormality occurrence period Tx is diagnosed as an abnormality cause, and these diagnosis results are used as the kurtosis Ks. Display with Specifically, when a predetermined time width is determined around the period Ts1, Ts2, Ts3,... And the value of the calculation abnormality occurrence period Tx is within the time width, damage corresponding to the calculation abnormality occurrence period Tx is determined. Diagnose the cause of the abnormality.
There may be a plurality of abnormal causes. For example, when Tin 10 and Tout 8 are within a predetermined time width centered on the abnormal periods Ts1 and Ts2, it is diagnosed that the inner ring and the outer ring are damaged.
With respect to the frequency spectrum Si of the waveform signal Sy, the cause of the abnormality can be diagnosed in the same manner as described above.

"Experimental example"
FIGS. 19A and 19B show vibration waveforms obtained by the conventional method, and FIGS. 20A, 20B, and 20C show vibration waveforms obtained by the diagnostic system of the fifth embodiment.

FIG. 19A shows a vibration waveform before replacement of the rolling bearing, and FIG. 19B shows a vibration waveform after replacement. In the bearing before replacement, flaking occurred in the outer ring and the inner ring, and flaking also occurred in some of the rolling elements.
The vibration waveform after replacement is a replacement of a new rolling bearing with no damage.
In the vibration waveforms of FIGS. 19A and 19B, there is no significant difference before and after the replacement, and the conventional method cannot determine the quality of the bearing.

FIG. 20 shows the degree of kurtosis when the fifth embodiment is adopted for the vibration waveform of FIG. The vibration waveform is measured four times before and after the replacement, and the kurtosis is obtained from each vibration waveform. Before the exchange,
(1) Sharpness by spectral sum is 26 to 43
(2) Sharpness by rms value is 20 to 39
(3) Sharpness by equivalent peak value is 20-40
Met.
In contrast, after replacement,
(1) Sharpness by spectral sum is 3-8
(2) Sharpness by rms value is 3-11
(3) The kurtosis by the equivalent peak value is 4-12
It was confirmed that the bearing condition can be accurately determined.

FIG. 21 shows a vibration waveform in which the diagnosis system of the fourth embodiment is applied to a rolling bearing rotating at 100 rpm.
FIG. 21 is an example of the frequency spectrum Ss obtained by the frequency spectrum calculating means 76 of the waveform signal, and the peak frequencies appear at fs1 = 6.5 Hz and fs2 = 13.0 Hz. That is, it is suggested that an abnormality appears at intervals of the cycle Ts1 = 0.154 s and Ts2 = 0.077 s. On the other hand, in the rolling bearing, since the number of rolling elements Z = 10, the rolling element diameter d = 6.35 mm, the pitch circle diameter D = 27.60 mm, and the contact angle α = 9 degrees between the rolling elements and the transfer surface, Tout = 0.155 s is obtained as an abnormality occurrence period at the time of damage. Since this is almost equal to twice Ts1 and Ts2, it can be determined that the outer ring of the rolling bearing is damaged.

Claims (15)

1. A bearing diagnosis system for rotating machinery equipment,
   A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
   A monitoring and diagnosis device connected to the sensor;
   A diagnostic notification means for connecting with the monitoring diagnostic device and displaying an abnormal state in percentage;
   With
   The monitoring and diagnosis apparatus includes:
   A storage unit for storing measurement data detected by the sensor;
   A reference level calculation unit that calculates an abnormality determination reference level based on the measurement data stored in the storage unit;
   One rotation time or intermittent operation time of the rotating shaft supported by the bearing is equally divided into a plurality of sections, and the abnormality determination reference level is compared with the measurement data of each section, for each section. And a determination unit for determining whether there is an abnormality.
2. The bearing diagnosis system according to claim 1 or 2, wherein one rotation time or one intermittent operation time of the rotating shaft supported by the bearing is divided into 100 sections in order to display the percentage.
3. In the reference level calculation unit of the monitoring and diagnosing device, the abnormality determination reference level is set to a constant multiple of the average value of the measurement data stored in the storage unit for each rotation or one intermittent operation of the rotating shaft supported by the bearing. The bearing diagnostic system according to claim 1 or claim 2.
4. The determination unit of the monitoring diagnostic device compares a diagnosis parameter calculation unit that calculates a diagnosis determination parameter, which is the number of sections in which an abnormality has occurred, with the diagnosis determination parameter and a predetermined abnormality determination criterion to determine whether there is an abnormality. The bearing diagnosis system according to any one of claims 1 to 3, further comprising a simple diagnosis determination unit for determining.
5. The determination unit of the monitoring / diagnostic apparatus sets an abnormal section only when a section in which abnormality is determined is a set number of consecutive adjacent sections within 2 to 10, and abnormality determination for sections less than the set number is an abnormality that is removed as noise. The bearing diagnosis system according to any one of claims 1 to 4, further comprising a generation section continuation determination unit.
6. The bearing diagnosis system according to any one of claims 1 to 5, further comprising an averaging processing unit that averages the diagnosis determination parameters for a plurality of rotations or a plurality of intermittent operations.
7. The bearing diagnosis system according to any one of claims 1 to 6, wherein the abnormality determination criterion to be compared with the diagnosis determination parameter is set at a plurality of levels such as a caution level and a danger level.
8. The determination unit of the monitoring / diagnosis device calculates the degree of coincidence of the determination results of occurrence of abnormalities for each distinction by shifting the abnormal cycle reference position of the abnormality determination result table for a plurality of rotations or a plurality of intermittent operations, and the highest degree of coincidence A synchronous search processing unit that searches for an abnormal cycle reference position and automatically calculates an abnormal cycle from the abnormal cycle reference position;
   Comparing a calculation abnormality occurrence cycle or an integer multiple of the calculation abnormality occurrence cycle, which is calculated from bearing device information and / or rotation speed information and has a different value for each cause of abnormality, and the abnormality cycle calculated by the synchronous search processing unit If the degree of coincidence between the calculation abnormality occurrence period or an integer multiple of the calculation abnormality occurrence period and the abnormality period is high, an abnormality caused by the cause corresponding to the calculation abnormality occurrence period occurs in the bearing. The bearing diagnosis system according to claim 1, further comprising a cause diagnosis determination unit for diagnosis.
9. A bearing diagnosis system for rotating machinery equipment,
   A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
   A monitoring and diagnosis device connected to the sensor;
   A diagnostic notification means for connecting to the monitoring diagnostic device and displaying a diagnostic result;
   With
   The monitoring and diagnosis apparatus includes:
   A storage unit for storing a signal waveform detected by the sensor;
   An arithmetic processing unit for the signal waveform stored in the storage unit;
   The arithmetic processing unit includes:
   Means for dividing the signal waveform into j according to the rotational speed of the bearing;
   Means for determining a frequency spectrum Fjk for each of the divided sections;
   Means for obtaining a spectrum sum Sj in a predetermined frequency range of the frequency spectrum;
   Creating a waveform signal Sx in which the spectrum sums Sj are arranged in time series, and obtaining a frequency spectrum Ss of the waveform signal Sx;
   Means for obtaining a kurtosis Ks of the spectral waveform Ss;
   A diagnostic system for a bearing, characterized in that the kurtosis Ks and a spectrum waveform obtained from the kurtosis Ks are displayed by the diagnostic notification means.
10. A bearing diagnosis system for rotating machinery equipment,
    A sensor for detecting the occurrence of damage attached to a fixed member of the bearing;
    A monitoring and diagnosis device connected to the sensor;
    A diagnostic notification means for connecting to the monitoring diagnostic device and displaying a diagnostic result;
    With
    The monitoring and diagnosis apparatus includes:
    A storage unit for storing a signal waveform detected by the sensor;
    An arithmetic processing unit for the signal waveform stored in the storage unit;
    The arithmetic processing unit includes:
    Means for applying a bandpass filter to the signal waveform;
    Means for dividing the signal waveform after passing through the band-pass filter into j according to the rotational speed of the bearing;
    Means for determining the rms value RMSj or equivalent peak value Pj of the signal for each of the divided sections;
    Means for generating a waveform signal Sy in which the rms value RMSj or equivalent peak value Pj is arranged in time series, and obtaining a frequency spectrum Si of the waveform signal Sy;
    Means for obtaining a kurtosis degree Ki of the spectral waveform Si;
    The bearing diagnosis system, wherein the kurtosis degree Ki and the spectrum waveform obtained from the kurtosis degree Ki are displayed by the diagnosis notification means.
11. The monitoring and diagnosis apparatus includes a determination unit,
    The determination unit is a unit that calculates a calculation abnormality occurrence period or an integer multiple of the calculation abnormality occurrence period that is calculated from the bearing device information and the rotation speed information and has different values for each abnormality cause, or the frequency spectrum Ss. Compares the peak period calculated by means for obtaining Si, and corresponds to the calculation abnormality occurrence period when the degree of coincidence between the peak calculation period and an integer multiple of the calculation abnormality generation period and the peak period is high The bearing diagnosis system according to claim 9 or 10, wherein an abnormality due to the cause of occurrence is diagnosed in the bearing.
12. The rotational speed of the rotating shaft of the rotating machine supported by the bearing is 300 rpm or less,
    The bearing diagnosis system according to claim 1, wherein the bearing is a rolling bearing or a sliding bearing.
13. The bearing diagnostic system according to any one of claims 1 to 12, wherein the sensor is any one of a vibration sensor, a displacement sensor, an acoustic emission sensor, an ultrasonic sensor, and a sound detection sensor.
14. In the case of a rolling bearing, the cause of the abnormality is at least one of inner ring damage, outer ring damage, and rolling element damage,
    The bearing diagnostic system according to claim 8 or 11, wherein in the case of a slide bearing, the abnormal metal contact of the rotating shaft that occurs once or a plurality of times in one rotation or one intermittent operation.
15. 15. The movable type in which the sensor, the monitoring diagnostic device, and the diagnostic notification unit are combined, or the diagnostic notification unit is wirelessly connected to the monitoring diagnostic device and is portable. The described bearing diagnostic system.

Images