JP2002181038A - Abnormality diagnosis device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an abnormality diagnosis device capable of easily installing a vibration detection means additionally without taking a measure for reducing signal attenuation and remodeling an existing device. SOLUTION: An AE sensor 19 on which oil prevention treatment is applied is charged into an oil bath 3 adjacent to a bearing 2, and the AE sensor 19 detects vibration for propagating lubricating oil in the oil bath 3, inputs a detection signal into an abnormality diagnosis means 7, and diagnoses an abnormality based on a situation of vibration. The AE sensor 19 can be installed without applying silicone grease to it and remodeling the existing device.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality diagnostic device for diagnosing abnormalities such as wear and cracks in a bearing portion supporting a rotary shaft.

[0002]

2. Description of the Related Art For example, in a power plant equipped with a gas turbine or a steam turbine, a large number of devices having a sliding member and a rotating shaft supported by bearings are provided. During operation of the equipment, fine wear powder and the like are generated from the sliding member and the support (contact part) of the rotating shaft, and the wear powder and the like become foreign matters and are mixed into the lubricating oil and seize on the support. And so on. In addition, bearings (particularly ball bearings) have a lifetime, and if the operation is continued beyond the lifetime, damage such as cracks may occur.

[0003] As a means for quickly diagnosing the occurrence of serious damage that hinders the normal operation of such equipment, abnormality diagnosis for measuring vibration, which is an indicator of the operating state of a plant, has been conventionally performed, and is supported by the abnormality diagnosis. Department management is underway. In the conventional abnormality diagnosis, vibration detection means (such as an ultrasonic vibration sensor) is provided on the bearing to detect vibrations at the bearing part, and the detected vibration is analyzed. The state and the place of occurrence are estimated. Then, when any abnormality is estimated, an alarm or the like is issued to cause a worker on site to take necessary measures such as replacing parts or stop the operation of the plant.

[0004]

In the conventional abnormality diagnosis, an ultrasonic vibration sensor or the like is attached to a bearing casing or the like in order to measure a vibration which is an index of a plant operating state. In this case, silicon grease or the like is applied to reduce signal attenuation due to the gap between the contact surfaces.However, if the application is not sufficient or the grease leaks, the obtained signal becomes extremely weak. When a plurality of ultrasonic vibration sensors are installed in the same apparatus, if there is an ultrasonic vibration sensor with an insufficient amount of application, the obtained signal level varies, and accurate information cannot be obtained. In addition, when an ultrasonic vibration sensor is attached to a bearing casing or the like using an existing device, a significant modification is required, and it is difficult to additionally install the ultrasonic vibration sensor.

[0005] The present invention has been made in view of the above circumstances, it is not necessary to take measures to reduce the signal attenuation,
Moreover, an object of the present invention is to provide an abnormality diagnosis device that can easily install a vibration detection unit without modifying an existing device.

[0006]

An abnormality diagnosis apparatus according to the present invention for achieving the above object comprises at least one or more vibration detecting means for detecting a state of vibration of a bearing portion supporting a rotating shaft, and a vibration detecting means. The bearing that supports the rotating shaft is immersed in the lubricating oil sump, and the vibration detecting means is disposed in the lubricating oil in the lubricating oil sump. Diagnostic means include:
It is characterized by having a function of detecting vibration propagating the lubricating oil and diagnosing abnormality.

[0007] The abnormality diagnosis means is provided with a vibration generation detecting function for detecting sudden vibration generation, intermittent vibration generation and continuous vibration generation based on the information of the vibration detection means. . Further, the abnormality diagnosis means has a storage function of storing a diagnosis result as time-series data, a function of deriving a variation pattern based on the time-series data stored in the storage function, and diagnosing an abnormality according to the variation pattern. Is provided. Further, a plurality of vibration detecting means are provided in the circumferential direction of the rotating shaft, and the abnormality diagnosing means divides the rotating shaft into a plurality of regions in the circumferential direction, synthesizes the output states of the plurality of vibration detecting means, and It is characterized in that a function of specifying a generation area is provided. In addition, the vibration generation detection function is to detect sudden vibration generation, intermittent vibration generation, and continuous vibration generation by sampling the state of vibration from the vibration detection means at low frequency and sampling at high frequency. It is characterized by.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a schematic configuration of a rotary shaft supporting portion provided with an abnormality diagnosis device according to an embodiment of the present invention, FIG. 2 is a block configuration of abnormality diagnosis means, and FIG. A graph showing the result of confirming the medium propagation signal is shown. The rotating shaft support shown in FIG. 1 is applied to, for example, a bearing of a rotating shaft of an air heater that collects exhaust heat of a generator in a power plant equipped with a gas turbine or a steam turbine. still,
The location to which the present invention is applied is not limited to the bearing portion of the rotating shaft of the air heater, and can be applied to the bearing portion of the rotating shaft of various devices.

As shown in FIG. 1, a rotating shaft 1 is rotatably supported by bearings 2 (ball bearings) (bearing portions). The lower part of the bearing 2 is arranged in an oil bath 3, and lubricating oil is supplied to and discharged from the oil bath 3 from a lubrication line 4.
The lubricating oil in the oil bath 3 is supplied to the bearing 2 with the rotation of the rotating shaft 1. An ultrasonic vibration sensor (AE sensor) 19 as an oil-proof vibration detecting means is provided in the oil bath 3 adjacent to the sliding portion. And AE
The sensor 19 detects the vibration propagating the lubricating oil in the oil bath 3. The detection signal of the AE sensor 19 is input to the abnormality diagnosis means 7 (control device). The abnormality diagnosing means 7 diagnoses an abnormality based on the state of vibration (vibration occurrence detection function). That is, the AE sensor 19 detects the crack propagation of the material and the elastic wave at the time of direct contact, and can obtain detailed information such as the current state of the sliding portion (contact portion).

Normally, when the AE sensor is mounted on the casing or the like of the bearing 2, silicon grease or the like is applied to reduce signal attenuation due to a gap between the contact surfaces. Also,
When attaching to the casing etc. of the bearing 2 with existing equipment,
It requires significant remodeling and makes it difficult to install additional AE sensors. In contrast, in the present embodiment, the AE sensor 19
Is put in the oil bath 3 so that the AE sensor 19
The space between the bearing and the bearing 2 is filled with oil, and there is no need to apply silicon grease or the like. In addition, since the oil bath 3 is provided with a refueling port and an opening for sampling, the existing device can be remodeled by introducing the AE sensor 19 subjected to the oil-proofing process from the refueling port and the opening. The AE sensor 19 can be installed without any
9 becomes very easy to install.

In general, liquids are known to transmit sound waves relatively well. FIG. 3 shows an ultrasonic signal propagating in a solid (circle in the figure) and an ultrasonic signal propagating in oil (● in the figure).
3 shows a graph in which the signal levels are compared. The horizontal axis represents the load N. Formally, an increase in the load corresponds to a severer sliding state of the bearing. For both the ultrasonic signal propagating in the solid and the ultrasonic signal propagating in the oil, the signal level that increases with increasing load is compared as a ratio to the initial voltage value. When the two ultrasonic signals are compared, when the load applied to the bearing increases, an equivalent increase in the signal level can be confirmed in each case. For this reason, there is no problem when applied to vibration measurement using an ultrasonic signal propagating in oil instead of an ultrasonic signal propagating through a solid.

As shown in FIG. 2, a detection signal of the AE sensor 19 is subjected to processing such as amplification and filter processing by an amplifier 8 and then passed through an AD converter 9 to calculate a vibration parameter. Is input to The vibration processing unit 10 quantifies the sliding state (contact state) by parameterizing the input data into the number of ultrasonic events, the amplitude, the characteristic frequency, and the like. The information of the sliding state (contact state) based on the vibration quantified by the vibration processing unit 10 is sent to the first diagnosis unit 14,
The first diagnostic unit 14 uses a neural network that self-organizes diagnostic criteria by learning, an expert system that has diagnostic criteria created based on past knowledge and experience of experiments, and the like, and performs short-term evaluation of sliding conditions and diagnosis of damaged parts. Is diagnosed.

That is, according to the detection signal of the AE sensor 19,
First, a slight level of sliding surface direct contact where damage is not surfaced is detected. The severity of the sliding state can be evaluated based on the absolute value of the amplitude or the rate of change from the initial value. Further, by frequency analysis, it is possible to detect synchronization with the rotation frequency and to estimate a damaged element.

The details of the short-term diagnosis will be described later with reference to FIGS.

Further, as shown in FIG.
Is also sent to the time-series data storage unit 15, and A
It is stored in the time-series data storage unit 15 as an instantaneous value of a diagnostic parameter based on the output of the E sensor 19 or a diagnostic history.
The data stored in the time-series data storage unit 15 is sent to a second diagnosis unit 16, which creates a self-organized diagnosis criterion by learning from a neural network, learning from past knowledge, experience from experiments, and the like. An expert system or the like having the diagnostic criteria described above evaluates a sliding state and diagnoses a tendency in consideration of a variation pattern of a damaged portion or the like.

The details of the diagnosis of the tendency are shown in FIGS.
Will be described later.

The details of the abnormality diagnosis (vibration detection function) based on the vibration based on the signal processing status of the AE sensor 19 will be described with reference to FIGS. FIG. 4 shows a block diagram of a vibration generation detecting function, FIG. 5 shows a processing flowchart of a vibration signal, and FIG. 6 shows a temporal change of a vibration generation state.

As shown in FIG. 4, the signal of the AE sensor 19 is sampled by the low frequency analyzer 21 and the high frequency analyzer 22. With respect to the vibration signal in the low frequency band sampled by the low frequency analysis device 21, parameters are extracted by the frequency analysis processing unit 23 using the original waveform and the envelope waveform (one waveform obtained by averaging a plurality of original waveforms). I do. With respect to the vibration signal in the high frequency band sampled by the high frequency analysis device 22, parameters are extracted by the signal shape evaluation signal processing unit 24 using the original waveform. In addition, the amplitude evaluation signal processing unit 25 of the vibration signal waveform extracts parameters for both the vibration signal in the low frequency band and the vibration signal in the high frequency band.
These parameters are used together for diagnosis. The high frequency band is set to a frequency of about megahertz, and the low frequency band is set to an appropriate frequency according to the scale of the device, the number of rotations of the rotating shaft 1 to be operated, and the like.

Since the vibration signal has different phenomena suitable for measurement in the low-frequency band and the high-frequency band, detailed diagnosis of abnormalities can be made by combining the two diagnoses and using them for analysis.

In a relatively low frequency band, a signal corresponding to the rotation frequency is mainly detected. For example, it has been known that damage corresponding to the inner and outer rings and the rolling elements of the bearing 2 generates a signal having a correlation with the rotation speed of the bearing 2. For this reason, it is possible to diagnose a damaged portion from the occurrence cycle.

The waveform of the vibration signal in the high-frequency band is analyzed, and the signal is generated continuously, independently, or the signal generated independently is generated intermittently. Or identify. That is, as shown in FIG. 5, when a vibration signal is generated, it is determined in step S1 whether another vibration signal is generated within a predetermined time, and when it is determined that another vibration signal is not generated. Single vibration (state of Fig. 6 (a))
It is assumed. If it is determined in step S1 that another vibration signal is generated within a predetermined time, it is determined in step S2 whether or not the signal is a continuous wave. If it is determined in step S2 that the wave is not a continuous wave, the vibration is determined to be intermittent vibration (the state shown in FIG. 6C).
If it is determined in step S2 that the wave is a continuous wave, continuous vibration (the state shown in FIG. 6B) is set.

The solitary vibration corresponds to the propagation of cracks inside the material or on the surface, and the intermittent vibration corresponds to the phenomenon of stick-slip or fretting.
It is considered that continuous vibration corresponds to direct contact between two surfaces due to poor lubrication, skidding, or damage to a sliding surface. For this reason, the sliding state (contact state) can be diagnosed by analyzing the waveform of the vibration signal.

The AE sensor 19 will be described with reference to FIGS.
The function of specifying the specific attachment status and the abnormality occurrence position (the function of identifying the abnormality occurrence area) will be described.
FIG. 7 shows the concept of the arrangement state of the AE sensor 19, and FIGS. 8 to 10 show the output state of the AE sensor 19.

As shown in FIG. 7, two AE sensors 1 are arranged in the oil bath 3 in the circumferential direction of the rotary shaft 1 (the circumferential direction of the bearing 2).
Assuming that the bearings 9a and 19b are arranged concentrically 180 degrees symmetrically, the bearing 2 is divided into two regions A and B in the circumferential direction with the boundary S interposed therebetween.
Suppose that it is divided into Two AE sensors 19a, 19
The approximate position of the output source and the presence or absence of movement of the output source are estimated from the time difference between the output signals b. Since the inside of the bearing 2 has a relatively complicated structure, accurate position locating may not be desired even when a normal position locating algorithm is used. For this reason, the two AE sensors 19a and 19b are mainly used for detecting the movement of the vibration source.
Are arranged concentrically around the center. If the two AE sensors 19a and 19b are arranged around the bearing 2,
It is not necessarily limited to concentric shapes.

In the case of two AE sensors 19a and 19b,
Based on the time difference at which the vibration wave is detected, it is possible to know whether a vibration signal has occurred in the region A, the region B, or on the boundary S line. Generally, since the bearing 2 has a concentric three-dimensional structure, it is possible to project three-dimensional information onto a two-dimensional plane for evaluation.

For example, as shown in FIG.
If there is no time difference between the signal output of the AE sensor 9b and the signal output of the AE sensor 9b, it is estimated that an abnormality has occurred on the boundary S line. As shown in FIG. 9, when the signal output of the AE sensor 19a occurs first and then the signal output of the AE sensor 19b occurs with a time lag, it is estimated that an abnormality has occurred in the region A. As shown in FIG. 10, first, the signal output of the AE sensor 19a is generated, then the signal output of the AE sensor 19b is generated with a time difference T1, then the signal output of the AE sensor 19a is generated, and then the time difference T2 ( When the signal output of the AE sensor 19b is generated with T1 <T2), it is estimated that an abnormality has occurred in the area A and the place of occurrence has moved to the rear in the rotation direction.

In this way, by estimating the location where the abnormality has occurred to some extent, it is possible to provide useful information for a work plan at the time of maintenance and the like.

It should be noted that the number of AE sensors 19 and the number of areas into which the bearing 2 is divided can be three or more.
The spatial resolution can be improved by increasing the number of 9s and the number of divided areas. Further, by arranging the AE sensor 19 also in the height direction, the same evaluation can be extended to three dimensions.

A situation in which the sliding state is diagnosed in consideration of the history of the output signal in a time series will be described with reference to FIGS. FIG. 11 shows a block configuration of the time-series diagnosis, and FIG. 12 shows the situation of the output signal.

The instantaneous value 31 and the instantaneous value diagnostic result 32 (diagnosis history) of the diagnostic parameter based on the output of the AE sensor 19 are stored as time-series data 33 (time-series data storage unit 15 in FIG. 2). A time series diagnosis unit 34 (the second diagnosis unit 16 in FIG. 2) that performs a diagnosis based on the series data 33 is provided. Thus, by accumulating the time-series data 33 for a predetermined time and using the variation pattern for diagnosis, it is possible to predict the value of the diagnostic parameter and determine the cause of the variation even if the instantaneous value 31 varies. And make reliable diagnosis.

That is, as shown in FIG. 12 (a), when the level of the signal of interest gradually rises (gradually increases), the signal level rises in the future by considering its history and the threshold value increases. Is expected to be exceeded, and it is possible to give a notice of the occurrence of an abnormality at an early stage before the threshold is exceeded. Further, as shown in FIG. 12 (b), when a state in which an abnormal signal level is detected only for a short time continues for a long period of time (unsteady state), sudden abnormalities continue to occur. It can be diagnosed that the moving state is in an unstable state.
Further, as shown in FIG. 12 (c), if a signal level exceeding the threshold value is temporarily detected and normalizes immediately (sudden), the signal level abnormality is determined as a temporary phenomenon, and the abnormality is determined. Diagnosis can be suppressed.

In the abnormality diagnosis based on the instantaneous value, an abnormality may frequently occur due to various factors, although there is no problem in the actual operation of the machine, which is a practical problem. For this reason, as described above, deriving a fluctuation pattern based on time-series data and diagnosing an abnormality according to the fluctuation pattern are effective for machine state diagnosis and remaining life evaluation. In particular, when the bearing 2 is a rolling bearing, a design based on a finite life is usually made, and it is effective to diagnose an abnormality according to a fluctuation pattern.

[0033]

The abnormality diagnosis apparatus of the present invention diagnoses an abnormality based on at least one or more vibration detecting means for detecting a state of vibration of a bearing portion supporting a rotating shaft, and detection information of the vibration detecting means. The bearing that supports the rotating shaft is immersed in the lubricating oil sump, and the vibration detecting means is disposed in the lubricating oil in the lubricating oil sump. Function to detect abnormalities and detect abnormalities, eliminating the need to interpose a fluid that reduces signal attenuation in the installation of vibration detection means, and easily opening through openings etc. without modifying existing equipment. The vibration detecting means can be installed in the lubricating oil reservoir, and the additional installation of the vibration detecting means becomes very easy.

The abnormality diagnosis means is provided with a vibration generation detection function for detecting sudden vibration generation, intermittent vibration generation and continuous vibration generation based on information from the vibration detection means. Diagnosis can be made based on the status of damage or damage.

The abnormality diagnosing means includes a storage function for storing the diagnosis results as time-series data, and a variation pattern derived based on the time-series data stored in the storage function to diagnose an abnormality according to the variation pattern. Function to predict the value of a diagnostic parameter,
Even if there is a variation in the instantaneous value, a reliable diagnosis can be performed by determining the cause of the variation.

Further, a plurality of vibration detecting means are provided in the circumferential direction of the rotating shaft, and the abnormality diagnosing means divides the rotating shaft into a plurality of areas in the circumferential direction, and synthesizes output states of the plurality of vibration detecting means. Since the function for specifying the abnormality occurrence area is provided, it is possible to estimate the abnormality occurrence part to some extent, and it is possible to provide useful information to a work plan at the time of maintenance.

Further, the vibration generation detecting function samples the situation of the vibration from the vibration detecting means at a low frequency and at a high frequency, so that sudden vibration generation, intermittent vibration generation and continuous vibration generation can be obtained. Since the detection is performed, it is possible to diagnose abnormalities corresponding to various phenomena.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram of a rotating shaft support provided with an abnormality diagnosis device according to an embodiment of the present invention.

FIG. 2 is a block diagram of an abnormality diagnosis unit.

FIG. 3 is a graph showing a result of confirming a propagation signal in oil.

FIG. 4 is a block diagram of a vibration generation detection function.

FIG. 5 is a processing flowchart of a vibration signal.

FIG. 6 is a graph showing a change over time in a state of occurrence of vibration.

FIG. 7 is an explanatory diagram of the concept of the arrangement of the AE sensor 19;

FIG. 8 is an explanatory diagram of an output state of the AE sensor 19;

FIG. 9 is an explanatory diagram of an output state of the AE sensor 19;

FIG. 10 is an explanatory diagram of an output state of the AE sensor 5;

FIG. 11 is a block diagram of a time series diagnosis.

FIG. 12 is a diagram illustrating the situation of an output signal.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Rotation shaft 2 Bearing 3 Oil bath 4 Lubrication line 7 Abnormality diagnosis means 8 Amplifier 9 A / D converter 10 Vibration processing unit 14 First diagnosis unit 15 Time series data storage unit 16 Second diagnosis unit 19 Ultrasonic vibration sensor (AE sensor) 21 low frequency analyzer 22 high frequency analyzer 23 frequency analysis processor 24 signal shape evaluation signal processor

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G01M 19/00 G01M 19/00 A G01N 29/14 G01N 29/14 (72) Inventor Masaya Kawano Nagasaki 5-717-1, Fukahori-cho, Nagasaki-shi, Nagasaki Research Laboratory, Mitsui Heavy Industries, Ltd. (72) Makito Katayama 6-16-1, Hikoshima-Enoura-cho, Shimonoseki-shi, Yamaguchi F-term in Shimonoseki Shipyard, Mitsubishi Heavy Industries, Ltd. (Reference) 2G024 AC02 BA21 CA13 FA02 FA04 FA06 2G047 AC08 BA05 EA11 GG01 GG09 GG15 GG19 GG33 GG37 3J101 AA02 AA52 AA62 BA77 CA03 FA24 GA24 GA26

Claims (5)

[Claims] The apparatus comprises at least one or more vibration detecting means for detecting a state of vibration of a bearing portion supporting a rotating shaft, and abnormality diagnosing means for diagnosing an abnormality based on detection information of the vibration detecting means. The bearing that supports the shaft is immersed in the lubricating oil sump, and the vibration detecting means is disposed in the lubricating oil in the lubricating oil sump. The abnormality diagnosing means detects the vibration propagating through the lubricating oil to diagnose the abnormality. An abnormality diagnosis device having a function. 2. The abnormality diagnosis means according to claim 1, further comprising: a vibration generation detecting function for detecting sudden vibration generation, intermittent vibration generation, and continuous vibration generation based on information of the vibration detection means. An abnormality diagnosis device, characterized in that: 3. The abnormality diagnosing means according to claim 1, further comprising: a storage function for storing diagnosis results as time-series data; and a derivation pattern based on the time-series data stored in the storage function. And a function of diagnosing abnormalities in response to the abnormalities. 4. The vibration detecting means according to claim 1, wherein a plurality of vibration detecting means are provided in a circumferential direction of the rotating shaft; An abnormality diagnosis device provided with a function of synthesizing the output states of the above and specifying an abnormality occurrence area. 5. The vibration generation detecting function according to claim 2, wherein the vibration state from the vibration detection means is sampled at a low frequency and at a high frequency, so that sudden vibration generation, intermittent vibration generation and An abnormality diagnostic device that detects continuous vibration generation.