Diagnostic equipment, diagnostic methods, diagnostic programs

Abstract

To provide a diagnosis device, etc., with which it is possible to efficiently diagnose a device that includes a rotating member.SOLUTION: The diagnosis device comprises: a vibration waveform acquisition unit 171 that acquires the waveform of vibration that is generated between the rotary shaft member of a stirring unit that stirs a fluid by rotation and a shaft support member that pivotally supports the rotary shaft member while sealing off the fluid; a partial waveform extraction unit 172 that extracts a partial waveform for a prescribed duration from the waveform; a vibration strength acquisition unit 174 that acquires vibration strength that represents a prescribed frequency band in the partial waveform; a diagnosis unit that diagnoses at least one of the rotary shaft member and the shaft support member on the basis of the vibration strength; and a storage 18 that preserves the result of diagnosis by the diagnosis unit together with the vibration strength and does not preserve at least some of the waveform acquired by the vibration waveform acquisition unit 171, on which the foregoing are based, and the partial waveform extracted by the partial waveform extraction unit 172.SELECTED DRAWING: Figure 2

Description

Technical Field

[0006]

[0001] The present invention relates to a diagnostic device and the like.

Background Art

[0002] Patent Document 1 discloses a diagnostic device for a mechanical seal provided in a rotating machine such as a pump, which uses an acoustic emission (AE) signal. Through the analysis of the AE signal measured by an AE sensor in the time domain and frequency domain, the diagnostic device can detect or predict abnormal wear and friction in the mechanical seal.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The diagnostic device of Patent Document 1 diagnoses whether the wear and friction on the sliding surface of the mechanical seal are within the allowable range based on a spectrogram obtained by expanding the intensity of the AE signal in the time domain and frequency domain. However, since the spectrogram contains a huge amount of data, it leads to a decrease in the diagnostic speed and an increase in the storage capacity for storing the diagnostic results.

[0005] The present invention has been made in view of such a situation, and its object is to provide a diagnostic device and the like that can efficiently diagnose a device including a rotating member.

Means for Solving the Problems

[0006] To solve the above problems, a diagnostic device according to one aspect of the present invention comprises: a vibration waveform acquisition unit that acquires the waveform of vibrations generated between a rotating shaft member of an agitator that agitates a fluid by rotation and a support member that supports the rotating shaft member while sealing the fluid; a partial waveform extraction unit that extracts a partial waveform for a predetermined time from the waveform; a vibration intensity acquisition unit that acquires the vibration intensity that represents a predetermined frequency band in the partial waveform; and a diagnostic unit that diagnoses at least one of the rotating shaft member and the support member based on the vibration intensity.

[0007] In this embodiment, the vibration intensity representative of a predetermined frequency band is obtained from a partial waveform for a predetermined time period extracted by the partial waveform extraction unit. By using the vibration intensity representative of a predetermined time period and predetermined frequency band in this way, the amount of data used for diagnosing the rotating shaft member as a rotating member and the pivot member as a support member can be reduced, thereby preventing a decrease in diagnostic speed and the need for large storage capacity to save the diagnostic results.

[0008] Another aspect of the present invention is also a diagnostic device. This device comprises a vibration waveform acquisition unit that acquires a vibration waveform generated between a rotating member and a support member that supports the rotating member; a partial waveform extraction unit that extracts a partial waveform for a predetermined time from the waveform; a vibration intensity acquisition unit that acquires a vibration intensity representative of a predetermined frequency band in the partial waveform; and a diagnostic unit that diagnoses at least one of the rotating member and the support member based on the vibration intensity.

[0009] In this embodiment, the vibration intensity representative of a predetermined frequency band is obtained from a partial waveform for a predetermined time period extracted by the partial waveform extraction unit. By using vibration intensity representative of a predetermined time period and predetermined frequency band in this way, the amount of data used for diagnosing rotating members and support members can be reduced, thus preventing a decrease in diagnostic speed and the need for large-capacity storage to save diagnostic results.

[0010] A further aspect of the present invention is a diagnostic method. This method comprises: a vibration waveform acquisition step of acquiring a vibration waveform generated between a rotating shaft member of an agitator that stirs a fluid by rotation and a support member that supports the rotating shaft member while sealing the fluid; a partial waveform extraction step of extracting a partial waveform for a predetermined time from the waveform; a vibration intensity acquisition step of acquiring a vibration intensity that represents a predetermined frequency band in the partial waveform; and a diagnostic step of diagnosing at least one of the rotating shaft member and the support member based on the vibration intensity.

[0011] Furthermore, any combination of the above components, as well as conversions of the expression of the present invention between methods, apparatus, systems, recording media, computer programs, etc., are also valid embodiments of the present invention. [Effects of the Invention]

[0012] According to the present invention, devices including rotating members can be efficiently diagnosed. [Brief explanation of the drawing]

[0013] [Figure 1] The configuration of the diagnostic device is schematically shown. [Figure 2] This is a functional block diagram of the AE signal processing unit. [Figure 3] This shows an example of AE signal processing by the AE signal processing unit. [Figure 4] This shows an example of AE signal processing by the AE signal processing unit. [Figure 5] This map shows various AE phenomena that can occur on the sliding surface of a mechanical seal, according to the frequency and amplitude of the AE wave. [Figure 6] This example demonstrates a diagnostic method that uses waveforms in the time domain of AE intensity across multiple predetermined frequency bands as diagnostic data. [Figure 7] An example of the diagnostic department's prediction of when an abnormality will occur is shown. [Modes for carrying out the invention]

[0014] The embodiments for carrying out the present invention will be described in detail below with reference to the drawings. In the description and drawings, identical or equivalent components, members, and processes are denoted by the same reference numerals, and redundant descriptions are omitted. The scale and shape of the illustrated parts are set for convenience to facilitate the description and are not to be interpreted restrictively unless otherwise specified. The embodiments are illustrative and do not limit the scope of the present invention in any way. Not all features or combinations thereof described in the embodiments are necessarily essential to the invention.

[0015] Figure 1 schematically shows the configuration of a diagnostic device 100 according to an embodiment of the present invention. The diagnostic device 100 is a device for diagnosing a mechanical seal 1 attached for sealing the fluid in an agitator (only partially shown) which is equipped with an agitator blade 5 as an agitator that agitates the fluid in the agitator tank 2 by rotation. A cylindrical shaft 3, which serves as the rotation axis member of the agitator blade 5, is provided to penetrate the mechanical seal 1 vertically. Above the mechanical seal 1, the upper end of the shaft 3 is connected to a rotational power generation unit 4 including a motor and gearbox, and below the mechanical seal 1, the lower end of the shaft 3 is connected to the agitator blade 5. The rotational power generation unit 4 converts the rotational power generated by the motor into a desired speed or torque using a gearbox. The agitator blade 5 rotates due to the rotational power from the rotational power generation unit 4 transmitted via the shaft 3, agitating the fluid contained in the agitator tank 2.

[0016] A mounting flange 6 for attaching a mechanical seal 1 is provided on the top or upper surface of the stirring tank 2. The mechanical seal 1 comprises a casing 9 as its housing. A through hole larger than the circular cross-section or outer circumference of the shaft 3 is provided in the center of the inner circumference side of the casing 9 in a top view, so that the shaft 3 can pass through the mechanical seal 1 vertically. Such a mechanical seal 1 constitutes a support member that rotatably supports the shaft 3 while sealing the fluid in the stirring tank 2.

[0017] The casing 9 is placed on the mounting flange 6 of the stirring tank 2. The casing 9 and the mounting flange 6 are firmly fixed by a fixing device such as a clamp (not shown) so that fluid does not leak from their contact surfaces (the lower surface of the casing 9 and the upper surface of the mounting flange 6). An annular lower fixing ring 10 surrounding the outer periphery of the shaft 3 is provided at the central portion facing the through hole of the casing 9 (i.e., the shaft 3). The diameter of the inner peripheral surface of the lower fixing ring 10 facing the outer peripheral surface of the shaft 3 is larger than the diameter of the inner peripheral surface of the casing 9 facing the outer peripheral surface of the shaft 3 at the lower part of the casing 9. Therefore, even when the inner peripheral surface of the lower part of the casing 9 contacts the outer peripheral surface of the shaft 3, the inner peripheral surface of the lower fixing ring 10 and the outer peripheral surface of the shaft 3 do not contact or interfere with each other.

[0018] On the other hand, the upper surface of the lower fixing ring 10 faces or approaches the lower surface of an annular lower rotating ring 11 fixed surrounding the outer periphery of the shaft 3 with a minute gap in the order of microns interposed therebetween. A lower fluid film 12 formed by fluid (liquid or gas) is formed in this minute gap. The lower fluid film 12 constitutes a fluid sealing portion that prevents the outflow of the fluid in the stirring tank 2 to the outside of the stirring tank 2 and the inflow of the fluid (such as air) outside the stirring tank 2 into the stirring tank 2. Further, the lower fluid film 12 also functions as a "sliding surface" between the lower rotating ring 11 that rotates integrally with the shaft 3 and the lower fixing ring 10 fixed to the mechanical seal 1. Since the lower rotating ring 11 and the lower fixing ring 10 are designed so as not to contact directly in principle by the lower fluid film 12 therebetween (strictly designed in the region of boundary lubrication), the "sliding surface" mainly means the contact surface between the lower surface of the lower rotating ring 11 and the upper surface of the lower fluid film 12 and the contact surface between the upper surface of the lower fixing ring 1 and the lower surface of the lower fluid film 12.

[0019] When the shaft 3 rotates, sliding and friction occur in the lower fluid film 12 as a sliding surface (strictly speaking, interactions such as collisions and adhesions of surface protrusions and fine particles on the sliding surface at the micron scale), resulting in high-frequency vibrations, sound waves, and elastic waves (hereinafter collectively referred to as AE waves) called acoustic emissions (AE) due to tribological phenomena. One or more lower AE sensors 13 for mainly measuring AE waves from the lower fluid film 12 are provided on the outer periphery of the lower part of the casing 9. The lower AE sensor 13 may be provided on the inner peripheral side of the casing 9, for example, in the vicinity of the lower fluid film 12 as a sliding surface or the lower fixed ring 10. The lower AE sensor 13 measures AE waves of 10 kHz or more and 2 MHz or less generated between the shaft 3 (lower rotating ring 11) and the mechanical seal 1 (lower fixed ring 10).

[0020] The lower AE sensor 13 transmits an AE signal as a measurement result of the AE wave to the barrier & amplifier 14 by wire or wirelessly. The amplifier in the barrier & amplifier 14 supplies power to the lower AE sensor 13 and amplifies the AE signal received from the lower AE sensor 13 for the subsequent signal processing computer 15. The barrier in the barrier & amplifier 14 limits the current flowing through the lower AE sensor 13 to a specified value or less, which is necessary when the lower AE sensor 13 is installed in an explosion-proof area. In this case, it is preferable to use the lower AE sensor 13 with explosion-proof specifications and install the barrier & amplifier 14 outside the explosion-proof area or inside an explosion-proof box.

[0021] The casing 9 has a relatively large-diameter internal space that accommodates the lower rotating ring 11 and the upper rotating ring 11 (described later), which are integrally formed on the outer circumference of the shaft 3. This internal space is filled with fluid, which provides the fluid sealing function and rotational lubrication function of the mechanical seal 1. The upper part of the casing 9 has a relatively small diameter, and an annular upper fixing ring 10 surrounding the outer circumference of the shaft 3 is provided in the central part facing the through hole (i.e., the shaft 3). The diameter of the inner surface of the upper fixing ring 10 that faces the outer surface of the shaft 3 is larger than the diameter of the inner surface of the casing 9 that faces the outer surface of the shaft 3 at the bottom of the casing 9. Therefore, even if the inner surface of the bottom of the casing 9 and the outer surface of the shaft 3 come into contact, the inner surface of the upper fixing ring 10 and the outer surface of the shaft 3 do not come into contact or interfere with each other.

[0022] On the other hand, the lower surface of the upper fixing ring 10 faces or approaches the upper surface of the annular upper rotating ring 11, which is fixed around the outer circumference of the shaft 3, with a minute gap of microns in between. An upper fluid film 12 is formed in this minute gap by the fluid (liquid). The upper fluid film 12, like the lower fluid film 12, constitutes a fluid sealing part that prevents the fluid in the stirring tank 2 from flowing out of the stirring tank 2 and the fluid outside the stirring tank 2 from flowing into the stirring tank 2. Also, like the lower fluid film 12, the upper fluid film 12 functions as a sliding surface between the upper rotating ring 11, which rotates integrally with the shaft 3, and the upper fixing ring 10, which is fixed to the mechanical seal 1.

[0023] As the shaft 3 rotates, sliding and friction (more precisely, interactions such as collisions and adhesions of micron-scale surface protrusions and fine particles on the sliding surface) occur on the upper fluid film 12, which acts as a sliding surface, resulting in the generation of AE waves due to tribological phenomena. One or more upper AE sensors 13 are provided on the outer circumference of the upper part of the casing 9 to measure AE waves mainly from the upper fluid film 12. The upper AE sensors 13 may also be provided on the inner circumference of the casing 9, for example, near the upper fluid film 12 or the upper fixing ring 10, which act as a sliding surface. The upper AE sensors 13 measure AE waves between 10 kHz and 2 MHz that occur between the shaft 3 (upper rotating ring 11) and the mechanical seal 1 (upper fixing ring 10).

[0024] The upper AE sensor 13, like the lower AE sensor 13, transmits the AE signal, which is the result of measuring the AE wave, to the barrier & amplifier 14, which includes a barrier and an amplifier, via wired or wireless connection. The amplifier in the barrier & amplifier 14 supplies power to the upper AE sensor 13 and amplifies the AE signal received from the upper AE sensor 13 for the subsequent signal processing computer 15. The barrier in the barrier & amplifier 14 limits the current flowing through the upper AE sensor 13 to below a specified value and is necessary when the upper AE sensor 13 is installed in an explosion-proof area. A barrier & amplifier 14 is provided for each lower AE sensor 13 and upper AE sensor 13. Therefore, the number of barrier & amplifiers 14 is the same as the total number of lower AE sensors 13 and upper AE sensors 13. Hereafter, the total number of lower AE sensors 13 and upper AE sensors 13 or the number of barrier & amplifiers 14 will also be referred to as the number of channels.

[0025] The signal processing computer 15, which processes multi-channel AE signals measured by multiple AE sensors 13, comprises an AE signal input unit 16, an AE signal processing unit 17, a storage unit 18, and a control unit 19. The AE signal input unit 16 converts the multi-channel AE signals (analog signals) input from the multiple AE sensors 13 via multiple barriers and amplifiers 14 into digital signals and transfers them to the AE signal processing unit 17. The sampling frequency used by the AE signal input unit 16 when converting analog signals to digital signals is preferably sufficiently higher than the frequency of the AE signal (10 kHz or more and 2 MHz or less). For example, setting it to 10 MHz or higher allows for sufficiently detailed results in the frequency domain when performing the short-time Fourier transform described later. Note that the AE signal input unit 16 may also receive AE signals from AE sensors 13 and barriers and amplifiers 14 provided in different stirring devices.

[0026] The AE signal processing unit 17 is comprised of a GPU (Graphics Processing Unit) and processes the multi-channel AE signals transferred from the AE signal input unit 16 in real time at high speed in parallel. Thus, the GPU comprising the AE signal processing unit 17 is used as a so-called GPGPU (General-Purpose Graphics Processing Unit) for the routine (yet complex) pipeline processing of multi-channel AE signals, which are not image signals. Furthermore, all or part of the AE signal processing unit 17 (GPU) may be comprised of artificial intelligence capable of machine learning. Detailed processing of the AE signal processing unit 17 will be described later.

[0027] The AE signal processing unit 17 processes the multi-channel AE signal transferred from the AE signal input unit 16 to diagnose the information contained in the AE wave measured by the mechanical seal 1. The diagnostic data of the mechanical seal 1 generated by the AE signal processing unit 17 is transferred to and stored in the storage 18, either via the control unit 19 or without going through the control unit 19. Here, the large amount of multi-channel AE signal input to the AE signal input unit 16 that formed the basis of the diagnostic data generated by the AE signal processing unit 17 is not transferred to the storage 18, thus preventing the storage 18 from becoming excessively large.

[0028] Furthermore, the AE signal processing unit 17 processes multi-channel AE signals without going through the control unit 19. That is, the large number of multi-channel AE signals input to the AE signal input unit 16 are not transferred to the control unit 19, and the AE signal processing unit 17 processes these multi-channel AE signals exclusively using its own computing resources without substantially using the computing resources of the control unit 19. As a result, the control unit 19, which is composed of a CPU (Central Processing Unit) or the like, can concentrate on controlling the signal processing computer 15 and, consequently, the diagnostic device 100, and can maintain an appropriate control state even while a large number of multi-channel AE signals are being processed by the AE signal processing unit 17.

[0029] The network interface 20 is an interface between the signal processing computer 15 and the host computer 21. The network interface 20 can read and write data to the storage 18 of the signal processing computer 15, and can execute control commands and data transfers to the control unit 19 of the signal processing computer 15. The host computer 21 is a computer used by the user of the diagnostic device 100, and has diagnostic software 22 installed that detects or predicts abnormalities in the mechanical seal 1.

[0030] Users of the diagnostic device 100 can configure measurement and processing settings and check diagnostic data stored in the storage 18 through the GUI (Graphical User Interface) of the diagnostic software 22 connected to the network interface 20. Specifically, the measurement and processing settings configured by the GUI of the diagnostic software 22 are transferred to the control unit 19 via the network interface 20 and reflected in the AE signal input unit 16 and the AE signal processing unit 17 under the control of the control unit 19. In addition, administrators of the diagnostic device 100 can access the diagnostic software 22 to provide user support and access the network interface 20 to configure settings and check data related to the control and processing of each part of the signal processing computer 15.

[0031] Figure 2 is a functional block diagram of the AE signal processing unit 17. The AE signal processing unit 17 comprises a vibration waveform acquisition unit 171, a partial waveform extraction unit 172, a Fourier transform unit 173, a vibration intensity acquisition unit 174, and a diagnostic data generation unit 175. These functional blocks are realized through the cooperation of hardware resources such as the computer's central processing unit, memory, input devices, output devices, and peripheral devices connected to the computer, and software executed using them. Regardless of the type or location of the computer, each of the above functional blocks may be realized using the hardware resources of a single computer, or it may be realized by combining hardware resources distributed across multiple computers. In this embodiment, all or part of the functional blocks of the AE signal processing unit 17 are realized by the GPU that constitutes the AE signal processing unit 17.

[0032] The vibration waveform acquisition unit 171 acquires the waveform of an AE wave between 10 kHz and 2 MHz generated between the shaft 3 (rotating ring 11) and the mechanical seal 1 (fixed ring 10) as a multi-channel AE signal (digital signal) from the AE signal input unit 16. Figure 3(A) shows one channel of AE signal acquired by the vibration waveform acquisition unit 171. Hereafter, the AE wave waveform shown in Figure 3(A) will also be referred to as the overall waveform to distinguish it from the partial waveform described later.

[0033] The partial waveform extraction unit 172 extracts a partial waveform for a predetermined time from the entire waveform of the AE wave acquired by the vibration waveform acquisition unit 171. Specifically, the partial waveform extraction unit 172 sequentially cuts out a plurality of partial waveforms of a partial waveform extraction length L as a predetermined time from different extraction start times in the entire waveform of the AE wave. FIG. 3(B) shows a plurality of partial waveforms (in time-domain representation) sequentially extracted by the partial waveform extraction unit 172. The interval between the extraction start times of each partial waveform is set as the sampling period T. By making the sampling period T smaller than the partial waveform extraction length L (T < L), a plurality of adjacent partial waveforms overlap with each other temporally. That is, the end portion of the partial waveform earlier in time and the start portion of the partial waveform later in time coincide with each other. By making the sampling period T and the partial waveform extraction length L constant respectively, the processing of the AE signal processing unit 17 is standardized and diagnosis can be performed stably, but the sampling period T and / or the partial waveform extraction length L may be variable according to the situation.

[0034] The Fourier transform unit 173 performs a Fourier transform on each partial waveform extracted by the partial waveform extraction unit 172 to convert it from the time-domain representation to the frequency-domain representation (it can also be said that a short-time Fourier transform with a sampling period T is performed on the entire waveform of the AE wave acquired by the vibration waveform acquisition unit 171). FIG. 3(C) shows a plurality of partial waveforms in the frequency-domain representation after being Fourier-transformed by the Fourier transform unit 173. The vibration intensity acquisition unit 174 acquires a plurality of vibration intensities (hereinafter also referred to as AE intensities) representing a plurality of different predetermined frequency bands in the plurality of partial waveforms in the frequency-domain representation generated by the Fourier transform unit 173. In the example of FIG. 3(C), the first predetermined frequency band from about 0 MHz to about 0.5 MHz and the second predetermined frequency band from about 0.5 MHz to about 1 MHz are schematically shown in each partial waveform.

[0035] The vibration intensity acquisition unit 174 acquires the integral value or average value (the integral value divided by the width of the first predetermined frequency band) of each partial waveform over a first predetermined frequency band as the first AE intensity representing the first predetermined frequency band, and acquires the integral value or average value (the integral value divided by the width of the second predetermined frequency band) of each partial waveform over a second predetermined frequency band as the second AE intensity representing the second predetermined frequency band. As a result, as shown in Figure 3(D), data for two AE intensities representing the two predetermined frequency bands are obtained from each partial waveform of partial waveform extraction length L for each sampling period T. As will be described later, the host computer 21 (diagnostic unit) that performs diagnosis based on these time periods and frequency bands does not need to refer to the large number of multi-channel AE signals shown in Figure 3(A), thus enabling faster diagnostic processing and compression of the amount of diagnostic result data.

[0036] Figure 4(A) shows the overall waveform of the AE wave in Figure 3(A) over a longer period, and Figure 4(B) shows the waveforms of the two AE intensities in Figure 3(D) over the same period as Figure 4(A). As shown in Figure 4(B), periodic fluctuations that were not apparent in Figure 4(A) appear in the waveform of the AE intensity in the first predetermined frequency band (lower than the second predetermined frequency band). For example, if such periodic fluctuations are disturbed, it can be diagnosed that some kind of abnormality has occurred in the mechanical seal 1, etc. Figure 4(C) shows the overall waveform of the AE wave when a sudden fluctuation occurs in the first predetermined frequency band, and Figure 4(D) shows the waveforms of the two AE intensities in that case. As shown in Figure 4(D), the waveform of the AE intensity in the second predetermined frequency band hardly changes, but the waveform of the AE intensity in the first predetermined frequency band fluctuates greatly. In this way, by monitoring the waveform of the AE intensity in the first predetermined frequency band, sudden fluctuations and other issues in the first frequency band can be effectively detected.

[0037] Figure 3(C) illustrates a simple first predetermined frequency band and a second predetermined frequency band that do not overlap. However, it is preferable to set one or more predetermined frequency bands for which the vibration intensity acquisition unit 174 acquires a representative AE intensity, depending on the various AE phenomena that may occur in the rotating ring 11, which is the source of the AE wave. Figure 5 shows a mapping of various AE phenomena that may occur in the fluid film 12 as the sliding surface of the mechanical seal 1, according to the frequency and amplitude of the AE wave.

[0038] For example, in order to effectively detect "rolling friction," the vibration intensity acquisition unit 174 is set to acquire a typical AE intensity in a predetermined frequency range of approximately 0.1 MHz to approximately 0.4 MHz. In order to effectively detect "abrasive wear," the vibration intensity acquisition unit 174 is set to acquire a typical AE intensity in a predetermined frequency range of approximately 0.3 MHz to approximately 1.0 MHz. In order to effectively detect "severe wear," the vibration intensity acquisition unit 174 is set to acquire a typical AE intensity in a predetermined frequency range of approximately 1.0 MHz to approximately 1.5 MHz.

[0039] Thus, the multiple predetermined frequency bands to which the vibration intensity acquisition unit 174 applies may overlap. Conventionally, narrowband AE sensors with each frequency band as the measurement band were sometimes used to detect each AE phenomenon, but in this embodiment, a wideband AE sensor 13 that can measure various AE waves from 10 kHz to 2 MHz together is used, and the subsequent vibration intensity acquisition unit 174 applies one or more predetermined frequency bands corresponding to each AE phenomenon.

[0040] The diagnostic data generation unit 175 generates diagnostic data used by the host computer 21, which functions as a diagnostic unit, to diagnose the mechanical seal 1, based on the representative AE intensity for each time period and frequency band acquired by the vibration intensity acquisition unit 174. The diagnostic data may be the AE intensity itself acquired by the vibration intensity acquisition unit 174, or it may be obtained by further processing the waveform of the AE intensity acquired by the vibration intensity acquisition unit 174 using the partial waveform extraction unit 172, the Fourier transform unit 173, and the same processing as the vibration intensity acquisition unit 174, or it may be calculated by combining multiple AE intensities acquired by the vibration intensity acquisition unit 174.

[0041] In this way, the diagnostic data generation unit 175 can flexibly generate diagnostic data optimized for detecting the target event based on one or more AE intensities. For example, the diagnostic data generation unit 175 can generate various types of diagnostic data capable of detecting not only AE phenomena related to wear and friction on the fluid film 12 as the sliding surface of the mechanical seal 1, as shown in Figure 5, but also abnormalities caused by malfunctions of parts around the sliding surface of the mechanical seal 1.

[0042] Here, defects in parts around the sliding surface of the mechanical seal 1 include, for example, distortion of the sliding surface, errors in the installation angle of the sliding surface, loosening of the set screw, and defects in the coil spring. These effects manifest as mechanical vibrations of the sliding surface, causing fluctuations or vibrations in the actual contact area of ​​the sliding surface. While such mechanical vibrations can be measured by installing dedicated vibration sensors or acceleration sensors, this is not practical for the mechanical seal 1, where space constraints near the sliding surface (fluid film 12) are severe. Therefore, the inventor focused on an AE sensor 13 that can be easily installed on the outer circumference of the casing 9 of the mechanical seal 1, and attempted to detect mechanical vibrations of the sliding surface through the AE waves measured by it. As a result, while it is difficult to directly capture the mechanical vibrations of the sliding surface from the "AE waves" measured by the AE sensor 13, it was found that the "AE intensity" or "AE energy" after signal processing by the AE signal processing unit 17 as described above exhibits vibration components similar to those of the mechanical vibrations of the sliding surface. Therefore, according to the diagnostic data of this embodiment based on "AE intensity" or "AE energy," in addition to the AE phenomena shown in Figure 5 that could conventionally be diagnosed through the AE sensor 13, it is possible to detect abnormalities caused by malfunctions of parts around the sliding surface of the mechanical seal 1, which could not be diagnosed with the conventional AE sensor 13.

[0043] The host computer 21, which functions as a diagnostic unit, diagnoses the mechanical seal 1 based on the time-dependent changes in one or more diagnostic data generated by the diagnostic data generation unit 175 based on the AE intensity. Figures 6(A) to (E) show an example of diagnosis using the waveform in the time domain of the AE intensity acquired by the vibration intensity acquisition unit 174 for five different predetermined frequency bands (more precisely, plots for each sampling period T) as diagnostic data. Three threshold values ​​for diagnosis are pre-set for each predetermined frequency band and are held in the diagnostic index holding unit implemented by the host computer 21.

[0044] As shown in Figures 6(B) and (E), in a predetermined frequency band where the waveform of the diagnostic data is below the minimum first threshold, it is diagnosed as "Normal". As shown in Figure 6(A), in a predetermined frequency band where the waveform of the diagnostic data is above the minimum first threshold and below the second threshold (which is greater than the first threshold and less than the maximum third threshold), it is diagnosed as "Caution". As shown in Figure 6(D), in a predetermined frequency band where the waveform of the diagnostic data is above the second threshold and below the maximum third threshold, it is diagnosed as "Warning". As shown in Figure 6(C), in a predetermined frequency band where the waveform of the diagnostic data is above the maximum third threshold, it is diagnosed as "Need repair". These diagnostic results are stored in storage 18, but diagnostic results such as "Caution", "Warning", and "Need repair", which may require urgent action, are notified in real time to the user or administrator of the diagnostic device 100 via the GUI of the network interface 20 and / or the diagnostic software 22. As mentioned above, different predetermined frequency bands can correspond to different diagnostic items. Therefore, by performing diagnoses for each predetermined frequency band as shown in Figure 6, the host computer 21, which functions as a diagnostic unit, can individually diagnose different diagnostic items.

[0045] Furthermore, the diagnostic index holding unit in the host computer 21 may not be limited to simple thresholds for the waveform of each diagnostic data as described above, but may also hold and provide to the diagnostic unit various characteristic quantities in the waveform of each diagnostic data, such as maximum value, PP value, average value, RMS value, deviation, kurtosis, distortion, extreme value thresholds, and other diagnostic indices.

[0046] By configuring the diagnostic unit with machine learning-capable artificial intelligence, abnormalities in each part of the mechanical seal 1 can be effectively detected or predicted through comprehensive analysis of each predetermined frequency band and / or each diagnostic data. The training data or teacher data used for machine learning in the diagnostic unit includes the AE intensity and / or each diagnostic data for each predetermined frequency band when each part of the mechanical seal 1 is operating normally, and / or the AE intensity and / or each diagnostic data for each predetermined frequency band when each part of the mechanical seal 1 is not operating normally.

[0047] Since a typical mechanical seal 1 is a custom-made product, the impact on the AE intensity and / or diagnostic data in each predetermined frequency band when an abnormality occurs may vary significantly from one mechanical seal 1 to another. In such cases, it is preferable to train the diagnostic unit using only the AE intensity and / or diagnostic data in each predetermined frequency band (with outliers removed) when each part of the mechanical seal 1 is operating normally as training data. Based on this training data, the diagnostic unit can learn the "normal data space" in the AE intensity and / or diagnostic data for each predetermined frequency band, and can diagnose an abnormality when data that is significantly different from this "normal data space" (this distance can also be set empirically) is detected.

[0048] The host computer 21, which functions as a diagnostic unit, may predict when a malfunction will occur in the mechanical seal 1 based on one or more diagnostic data generated by the diagnostic data generation unit 175, and present this to the user of the diagnostic device 100 through the GUI of the diagnostic software 22. For example, as shown in Figure 7, for a diagnostic item for which a diagnostic score is presented on an evaluation scale from "0" to "5", if the current diagnostic score is within the acceptable range of "1" to "4", but it is predicted that in 500 hours (500h), the diagnostic score will approach the lower limit of the acceptable range of "1", and further, in 1000 hours (1000h), the diagnostic score will fall below the lower limit of the acceptable range of "1", increasing the probability of a malfunction, the user of the diagnostic device 100 can schedule maintenance on the mechanical seal 1 before the diagnostic score falls below the lower limit of the acceptable range of "1", for example, in 500 hours. The diagnostic score is calculated by a machine learning-trained diagnostic unit by comprehensively referring to the current AE intensity and / or diagnostic data for each predetermined frequency band, past AE intensity and / or diagnostic data for each predetermined frequency band during normal operation or when abnormalities occurred, and past evaluation data from disassembly and inspection of mechanical seal 1.

[0049] All or part of the processing in the AE signal processing unit 17 described above is performed substantially without the intervention of the control unit 19 (without substantially using the computational resources of the control unit 19). The involvement of the control unit 19 is limited to overall control of the AE signal processing unit 17 using minimal control resources. The storage 18 stores the diagnostic data generated by the diagnostic data generation unit 175, along with the AE intensity of a predetermined frequency band acquired by the vibration intensity acquisition unit 174 that formed the basis of the diagnostic data. On the other hand, the storage 18 does not store the overall waveform of the AE wave acquired by the vibration waveform acquisition unit 171 that formed the basis of these, the partial waveform in time-domain representation extracted by the partial waveform extraction unit 172, or the partial waveform in frequency-domain representation generated by the Fourier transform unit 173. Therefore, it is possible to prevent the storage 18 from becoming excessively large, and / or to operate the diagnostic device 100 continuously for a long period of time until the amount of data stored in the storage 18 reaches its capacity.

[0050] The present invention has been described above based on embodiments. The embodiments are illustrative, and it will be understood by those skilled in the art that various modifications are possible in combinations of their components and processing processes, and that such modifications also fall within the scope of the present invention.

[0051] In the embodiment, a diagnostic device 100 for diagnosing the mechanical seal 1 was illustrated, but according to the present invention, at least one of the rotating member and the support member can be diagnosed based on the waveform of vibrations occurring between any rotating member (including the shaft 3 in the embodiment) and any support member that supports the rotating member (including the stirring tank 2 and a bearing not shown that holds the shaft 3 in the embodiment). For example, the present invention can be applied to any rotating machine having a plurality of members (rotating members and support members) that are in contact with and / or sliding against each other, such as a pump, motor / gearbox, or gland packing.

[0052] The functional configurations of each device described in the embodiments can be realized using hardware resources, software resources, or through the collaboration of hardware and software resources. Hardware resources can include processors, ROMs, RAMs, and other LSIs. Software resources can include operating systems, applications, and other programs. [Explanation of symbols]

[0053] 1 Mechanical seal, 2 Agitation tank, 3 Shaft, 4 Rotation power generation unit, 5 Agitator blade, 6 Mounting flange, 9 Casing, 10 Fixing ring, 11 Rotating ring, 12 Fluid membrane, 13 AE sensor, 14 Barrier & amplifier, 15 Signal processing computer, 16 AE signal input unit, 17 AE signal processing unit, 18 Storage, 19 Control unit, 100 Diagnostic device, 171 Vibration waveform acquisition unit, 172 Partial waveform extraction unit, 173 Fourier transform unit, 174 Vibration intensity acquisition unit, 175 Diagnostic data generation unit.

Claims

1. A vibration waveform acquisition unit that acquires the waveform of AE waves, which are sound waves generated between a rotating shaft member of an agitator that stirs the fluid by rotation and a support member that supports the rotating shaft member while sealing the fluid, A partial waveform extraction unit extracts multiple partial waveforms for a predetermined time period from different extraction start times in the aforementioned waveform, A vibration intensity acquisition unit that acquires vibration intensity representative of a predetermined frequency band in each of the aforementioned partial waveforms, A diagnostic unit diagnoses the mechanical vibration of the sliding surfaces of the rotating shaft member and the support member based on diagnostic data obtained by further processing the vibration intensity waveform with the partial waveform extraction unit and the vibration intensity acquisition unit. A diagnostic device equipped with the following features.

2. The vibration waveform acquisition unit acquires the waveform of the AE wave, which is between 10 kHz and 2 MHz and generated between the rotating shaft member and the support member. The predetermined frequency bands are set to be at least 0.1MHz to 0.4MHz, 0.3MHz to 1.0MHz, and 1.0MHz to 1.5MHz. The diagnostic unit detects rolling friction in the predetermined frequency band of 0.1 MHz to 0.4 MHz, abrasive wear in the predetermined frequency band of 0.3 MHz to 1.0 MHz, and severe wear in the predetermined frequency band of 1.0 MHz to 1.5 MHz. The diagnostic device according to claim 1.

3. The diagnostic apparatus according to claim 1 or 2, wherein the vibration intensity acquisition unit acquires the integral value or average value of the vibration intensity over the predetermined frequency band in the partial waveform as the vibration intensity representative of the predetermined frequency band.

4. The partial waveform extraction unit extracts multiple partial waveforms for a predetermined time period from different extraction start times in the waveform. The vibration intensity acquisition unit acquires a plurality of vibration intensities that represent the predetermined frequency band in the plurality of partial waveforms. The diagnostic unit diagnoses at least one of the rotating shaft member and the support member based on the temporal changes in the plurality of vibration intensities. A diagnostic device according to any one of claims 1 to 3.

5. The diagnostic device according to claim 4, wherein the plurality of partial waveforms overlap each other in time.

6. The vibration intensity acquisition unit acquires multiple vibration intensities that represent different predetermined frequency bands in the partial waveform, The diagnostic unit diagnoses at least one of the rotating shaft member and the support member based on the plurality of vibration intensities. A diagnostic device according to any one of claims 1 to 5.

7. The diagnostic device according to claim 6, wherein the different predetermined frequency bands overlap with each other.

8. The diagnostic device according to any one of claims 1 to 7, wherein the vibration waveform acquisition unit acquires the waveform of the AE wave between the rotating shaft member and the pivot member, which is between 10 kHz and 2 MHz.

9. A control unit that controls the diagnostic device, A processing unit that performs at least a portion of the processing of the vibration waveform acquisition unit, the partial waveform extraction unit, the vibration intensity acquisition unit, and the diagnostic unit without going through the control unit, A diagnostic device according to any one of claims 1 to 8, further comprising:

10. The diagnostic apparatus according to any one of claims 1 to 9, further comprising storage that stores the diagnostic results from the diagnostic unit together with the vibration intensity, and does not store at least a portion of the waveforms acquired by the vibration waveform acquisition unit and the partial waveforms extracted by the partial waveform extraction unit that formed the basis for these results.

11. The diagnostic device according to any one of claims 1 to 10, wherein the diagnostic unit predicts the timing of an abnormality in at least one of the rotating shaft member and the support member based on the vibration intensity.

12. The diagnostic device according to any one of claims 1 to 11, wherein the diagnostic unit includes artificial intelligence capable of machine learning the normal data space for the vibration intensity in the predetermined frequency band and / or the diagnostic data, and diagnoses the mechanical vibration of the sliding surfaces of the rotating shaft member and the support member based on the distance of the vibration intensity in the predetermined frequency band and / or the diagnostic data from the normal data space.

13. A vibration waveform acquisition unit that acquires the waveform of an AE wave, which is a sound wave generated between a rotating member and a support member that supports the rotating member, A partial waveform extraction unit extracts multiple partial waveforms for a predetermined time period from different extraction start times in the aforementioned waveform, A vibration intensity acquisition unit that acquires vibration intensity representative of a predetermined frequency band in each of the aforementioned partial waveforms, A diagnostic unit diagnoses the mechanical vibration of the sliding surfaces of the rotating member and the support member based on diagnostic data obtained by further processing the waveform of the vibration intensity by the partial waveform extraction unit and the vibration intensity acquisition unit. A diagnostic device equipped with the following features.

14. A vibration waveform acquisition step that acquires the waveform of an AE wave, which is a sound wave generated between a rotating shaft member of a stirring section that agitates the fluid by rotation, and a support member that supports the rotating shaft member while sealing the fluid, A partial waveform extraction step of extracting multiple partial waveforms for a predetermined time period from different extraction start times in the aforementioned waveform, A vibration intensity acquisition step to acquire a vibration intensity that represents a predetermined frequency band in each of the aforementioned partial waveforms, A diagnostic step to diagnose the mechanical vibration of the sliding surfaces of the rotating shaft member and the support member based on diagnostic data obtained by further processing the waveform of the vibration intensity in the partial waveform extraction step and the vibration intensity acquisition step, A diagnostic method that includes the following features.

15. A vibration waveform acquisition step that acquires the waveform of an AE wave, which is a sound wave generated between a rotating shaft member of a stirring section that agitates the fluid by rotation, and a support member that supports the rotating shaft member while sealing the fluid, A partial waveform extraction step of extracting multiple partial waveforms for a predetermined time period from different extraction start times in the aforementioned waveform, A vibration intensity acquisition step to acquire a vibration intensity that represents a predetermined frequency band in each of the aforementioned partial waveforms, A diagnostic step to diagnose the mechanical vibration of the sliding surfaces of the rotating shaft member and the support member based on diagnostic data obtained by further processing the waveform of the vibration intensity in the partial waveform extraction step and the vibration intensity acquisition step, A diagnostic program that causes a computer to run.

Images