JP2003021555A - Abnormality monitoring equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To monitor whether abnormality exists in abnormality monitoring equipment itself, and decide accurately whether abnormality exists, regarding abnormality monitoring equipment monitoring whether abnormality exists in machine equipment as an object of diagnosis. SOLUTION: A signal for generating simulated abnormal sound is sent to a speaker 240 from an abnormality monitoring equipment main body 200, simulated abnormal sound is generated from the speaker 240, and sensing is performed by using a sound collecting microphone. The main body 200 makes an inverse filter act upon a received acoustic waveform signal, and obtains a residual signal, thereby checking whether correct detection is performed. The inverse filter corresponds to the same operating state as the machine equipment 21 as an object of diagnosis.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality monitoring device for monitoring the presence or absence of abnormality in a diagnosis target.

[0002]

2. Description of the Related Art Conventionally, equipment diagnosis has been executed or proposed by various equipment diagnosis methods for determining the presence or absence of abnormality of equipment or equipment. In this equipment diagnosis, the equipment is broken or it is necessary to stop immediately, and not only the serious failure is detected, but rather, the bearing in the rotating machine before the serious failure is damaged. However, it is still possible to continue operating sufficiently for some time, such as when some moving parts have worn out, but it is necessary to detect anomalies that may lead to serious failure in the future if left as it is. There is.

As a typical example of such equipment diagnosis method,
Obtain an acoustic vibration waveform when the device or equipment is in a normal state, perform spectrum analysis of the acoustic vibration waveform to investigate its characteristics, and detect the acoustic vibration waveform of the device or equipment when detecting the presence or absence of abnormality. Obtain and analyze the spectrum, and detect anomalies depending on whether there is a peak of a specific frequency component that is not seen in normal times, or whether the combination of peaks is the same as that in normal times. Is known to do.

Further, Japanese Patent Laid-Open No. 7-43259 discloses that
Obtain the acoustic vibration waveform when the device or equipment is in a normal state, create an inverse filter based on the acoustic vibration waveform, and obtain the acoustic vibration waveform of the device or equipment when detecting the presence or absence of abnormality. It has been proposed to detect an abnormality of a device or equipment by applying a previously obtained inverse filter to the acoustic vibration waveform to obtain a residual signal and analyzing the residual signal.

Further, in Japanese Unexamined Patent Publication No. 8-304124, a plurality of acoustic vibration waveforms when the equipment or equipment is in a normal state are obtained and, for example, one acoustic vibration waveform is selected from the plurality of acoustic vibration waveforms. An inverse filter is created based on this, and the inverse filter is applied to, for example, the remaining acoustic vibration waveforms to obtain a plurality of residual signals, and a plurality of statistical variables are obtained based on each of the plurality of residual signals. Even when detecting the presence or absence of abnormality, a plurality of acoustic vibration waveforms of the device or equipment are obtained, and the above-described inverse filter obtained in advance is applied to the plurality of acoustic vibration waveforms to generate a plurality of residual errors. Signals are obtained, multiple statistical variables are obtained based on the multiple residual signals, multiple statistical variables obtained in the normal state, and multiple statistics obtained when detecting the presence / absence of abnormality In between the variables, for example the F-test and t
It has been proposed to detect the presence or absence of abnormality in the device or equipment by performing verification or estimation by a method such as verification.

The method of detecting an abnormality in equipment or facilities by performing the above-mentioned spectrum analysis is also a very effective method depending on the nature of the equipment or facilities to be diagnosed,
The method of creating the inverse filter and the method of performing the statistical test are more effective methods.

[0007]

However, no matter which of the various equipment diagnosis methods described above is adopted, the abnormality monitoring device itself that incorporates the equipment diagnosis method and monitors whether or not there is an abnormality in the diagnosis target is If it does not operate normally, not only cannot the abnormality of the diagnosis target be accurately detected, but it is believed that the abnormality monitoring device is operating normally, rather increasing the possibility of a serious accident. It can also result in being lost.

Therefore, it has been conventionally known to perform facility diagnosis while confirming that the abnormality monitoring device and the sensor itself are normal. For example, Japanese Patent Laid-Open No. 5-44917 proposes to give a simulated signal to a sensor for diagnosis, and Japanese Patent Laid-Open No. 5-12.
In Japanese Patent No. 581, it is proposed that the input of the sensor is cut off and the signal is interrupted only when the input is cut off to determine whether or not the input is normal.

However, although the conventionally proposed technique can partially confirm that it is normal, it is possible to comprehensively confirm whether or not the abnormality monitoring device correctly maintains the desired performance. There is a problem that it is difficult to do.

In view of the above circumstances, it is an object of the present invention to provide an abnormality monitoring device capable of monitoring the presence or absence of an abnormality in a diagnosis target and correctly determining the presence or absence of the abnormality.

[0011]

An abnormality monitoring apparatus of the present invention that achieves the above object is an abnormality monitoring apparatus for monitoring the presence or absence of an abnormality in a diagnosis target, by capturing a predetermined physical quantity that reflects the state of the diagnosis target. Based on a sensor for obtaining a signal representing a physical quantity, reference data derived from a reference signal obtained when the diagnosis target is in a normal state by the sensor, and a diagnosis signal obtained by the sensor during abnormality monitoring, Abnormality determination unit that determines the presence or absence of abnormality, a simulated abnormal state generation unit that gives the sensor a physical quantity that simulates when the diagnostic target is in an abnormal state, the reference data, and the sensor simulated abnormal state Based on the simulated abnormal component mixing signal obtained by the sensor at the timing of capturing the physical quantity given by the generator, it is possible to determine whether or not there is an abnormality in the abnormality monitoring device itself. Characterized by comprising a self-abnormality determining unit for.

The abnormality monitoring apparatus of the present invention, by giving a physical quantity (for example, sound or vibration) simulating when the object to be diagnosed is in an abnormal state and checking whether or not it is judged to be abnormal, so to speak, When the abnormality monitoring device is determined to be abnormal, it is determined that the abnormality monitoring device is normal, and it is possible to comprehensively guarantee that the abnormality monitoring device is normal.

Here, in the above-mentioned abnormality monitoring apparatus of the present invention, there is provided a reference arithmetic operation section for obtaining reference data by an operation including an operation for obtaining an inverse filter based on the reference signal, and the diagnosis object abnormality judging section is provided with a diagnostic signal. An operation including an operation for obtaining a residual signal by applying an inverse filter to the is performed, and the presence or absence of abnormality of the diagnosis target is determined based on the result of the operation. It is preferable that an operation including an operation for obtaining a residual signal by performing an inverse filter on a signal is performed, and whether or not there is an abnormality in the abnormality monitoring device itself is determined based on the result of the operation.

Here, the above-mentioned "operation including an operation for obtaining an inverse filter" is a concept including a case in which only an operation for obtaining an inverse filter is included. In that case, the inverse filter is used as the reference data. can do. The "operation including an operation for obtaining an inverse filter" may include an operation for obtaining an inverse filter.
-304124, a plurality of reference signals are obtained, an inverse filter is created based on, for example, one reference signal of the plurality of reference signals, and the inverse filter is used as a plurality of other reference signals. May be used to calculate a plurality of statistical variables. In that case, the plurality of statistical variables thus obtained can serve as reference data.

The above-mentioned "calculation including a calculation for obtaining a residual signal by applying an inverse filter" is also the same as the above, and may be composed of only a calculation for obtaining a residual signal, or the above-mentioned. As described in Japanese Patent Application Laid-Open No. 7-43259, data convenient for determining the presence / absence of abnormality by calculating the residual signal such as obtaining the moving average value of the power of the residual signal is obtained. A calculation is performed, or an inverse filter is applied to a plurality of diagnostic signals as described in Japanese Patent Laid-Open No. 8-304124 to obtain a plurality of residual signals, and based on the plurality of residual signals. It may be an operation for obtaining a plurality of statistical variables.

If an inverse filter is used, steady noise can be eliminated from the signal, and the presence / absence of abnormality in the diagnosis object can be determined with higher accuracy.

Further, in the above-mentioned abnormality monitoring apparatus of the present invention, the simulated abnormal state generation unit determines to the sensor that the diagnosis target is abnormal by the abnormality monitoring apparatus when the abnormality monitoring apparatus is in a normal state. It is characterized by providing a physical quantity simulating an abnormal state of a level corresponding to the lower limit.

By doing so, the detection limit of this abnormality monitoring device can be guaranteed.

Further, in the abnormality monitoring device of the present invention,
From the simulated abnormal component mixed signal, the sensor includes a simulated abnormal component removal unit that removes the simulated abnormal component caused by the physical quantity given from the simulated abnormal state generation unit to generate a simulated abnormal component removal signal. The determination unit, with respect to the timing when the sensor is given a physical quantity that simulates that the diagnosis target is in an abnormal state by the simulated abnormal state generation unit, based on the reference data and the simulated abnormal component removal signal, It is preferable to determine whether or not there is an abnormality in the diagnosis target.

Depending on the object to be diagnosed, there is a case in which it is not allowed to continuously operate, and a blank time which is not monitored occurs even for a part of the time. In such a case, as described above, the simulated abnormal component removal signal is generated, and the presence / absence of the diagnosis target is determined based on the reference data and the generated simulated abnormal component removal signal, thereby generating the simulated abnormal state. The presence / absence of the abnormality of the diagnosis target can be determined also for the time when the sensor gives the sensor a physical quantity simulating that the diagnosis target is in the abnormal state.

[0021]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below.

FIG. 1 is a block diagram showing a basic embodiment of the abnormality monitoring device of the present invention.

The abnormality monitoring device 10 is a device for monitoring the presence or absence of abnormality in the diagnosis target 20, and includes a sensor 11, a reference calculation unit 12, a diagnosis target abnormality determination unit 13, a simulated abnormal state generation unit 14, and a self abnormality determination unit. 15 and a simulated abnormal component removing unit 16.

The sensor 11 captures a predetermined physical quantity that reflects the state of the diagnosis target 20, such as sound or vibration, and obtains a signal representing the physical quantity.

The reference calculation unit 12 calculates the reference data derived from the reference signal by a calculation including a calculation for obtaining an inverse filter based on the reference signal obtained when the diagnosis target 20 is in a normal state. Is. The reference data may be the inverse filter itself or the statistical variable as described above, but the inverse filter itself is adopted as the reference data in this embodiment.

Further, the diagnosis target abnormality determination unit 13 determines the presence or absence of abnormality of the diagnosis target based on the above-mentioned reference data (inverse filter in this embodiment) and the diagnostic signal obtained by the sensor 11 at the time of abnormality monitoring. It is a thing.

Here, this diagnosis target abnormality determination unit 13
Performs an operation including an operation for obtaining a residual signal by applying an inverse filter to the diagnostic signal, and determines whether or not there is an abnormality in the diagnosis target 20 based on the result of the operation.

This "calculation for obtaining the residual signal by applying the inverse filter" may be composed of only the calculation for obtaining the residual signal, as described above, or the above-mentioned Japanese Patent Laid-Open No. 7-43259. As described in the publication, the moving average value of the power of the residual signal is calculated,
The residual signal is calculated to obtain data convenient for determining the presence or absence of abnormality, or an inverse filter is applied to a plurality of diagnostic signals as described in Japanese Patent Laid-Open No. 8-304124. May be applied to obtain a plurality of residual signals, and a plurality of statistical variables may be obtained based on the plurality of residual signals.

In this embodiment, the moving average value of the power of the residual signal is obtained, and the presence or absence of abnormality of the diagnosis target 20 is determined based on the moving average value of the power.

Further, the simulated abnormal state generating section 14 gives the sensor 11 a physical quantity (for example, sound or vibration) simulating when the diagnosis target 20 is in an abnormal state. The physical quantity provided by the simulated abnormal state generation unit 14 does not need to completely simulate the physical quantity when the diagnosis target 20 is in an abnormal state as it is, and causes the abnormality monitoring device 10 to determine that it is abnormal. Anything will do. For example, when the sensor 11 is a sensor that captures a sound emitted from the diagnosis target 20 to obtain a signal, the simulated abnormal state generation unit 14 is emitted from the diagnosis target 20 when the diagnosis target 20 is in an abnormal state. It is not necessary to provide the sensor 11 with a sound having a waveform simulating the sound waveform as it is, and a completely different sound on the waveform is generated by the sensor 11.
May be given to

Here, a signal obtained by the sensor 11 when a physical quantity simulating when the diagnosis target 20 is in an abnormal state is given from the simulated abnormal state generation section 14 is referred to as a "simulated abnormal signal mixed signal". The simulated abnormal signal mixed signal is a signal including both components of the physical quantity given from the simulated abnormal state generation unit 14 and the physical quantity reflecting the operating state of the diagnosis target 20.

Further, the self-abnormality determination unit 15 is obtained by the sensor 11 at the timing when the sensor 11 captures the reference data (inverse filter in this embodiment) and the physical quantity given from the simulated abnormal state generation unit 14. The presence / absence of abnormality of the abnormality monitoring device itself is determined based on the simulated abnormal component mixed signal. This self-abnormality determination unit 15
Then, in the present embodiment, an operation including an operation for obtaining a residual signal by applying an inverse filter to the simulated abnormal component mixed signal is performed, and based on the result of the operation,
Whether or not there is an abnormality in the abnormality monitoring device itself is determined.

From the simulated abnormal state generating section 14,
The sensor 11 is provided with a physical quantity simulating an abnormal state at a level corresponding to a lower limit of determination for determining that the diagnosis target 20 is abnormal by the abnormality monitoring device when the abnormality monitoring device 10 is in a normal state. So here
It is possible to determine the performance of this abnormality monitoring device, including whether or not it is possible to correctly determine the presence or absence of abnormality of the diagnosis target up to the detection limit.

Further, the simulated abnormal component removing section 16 removes the simulated abnormal component resulting from the physical quantity given from the simulated abnormal state generating section 14 from the simulated abnormal component mixed signal to generate the simulated abnormal component removing signal. Is. In the present embodiment, the simulated abnormal state generation unit 14 gives a physical quantity such as sound or vibration to the sensor based on a reproducible signal such as a pseudo-random signal such as an M-sequence.
When 0 is stationary, a physical quantity (such as sound or vibration) is given to the sensor 11 from the simulated abnormal state generation unit 14 in advance based on the reproducible signal. The signal derived from only the physical quantity given to 11 is obtained and stored, and the simulated abnormal component removing unit 16 stores it from the simulated abnormal component mixed signal obtained by the sensor 11 while the diagnosis target 20 is operating. By subtracting the stored signal, a simulated pseudo abnormal component removal signal is generated. This simulated pseudo abnormal component removal signal corresponds to a diagnostic signal when the simulated abnormal state generation unit 14 is not operating, and the diagnosis target abnormality determination unit 13 diagnoses the sensor 11 by the simulated abnormal state generation unit 13. Target 20
The timing at which the physical quantity simulating that the abnormal state is given is diagnosed based on the reference data (inverse filter here) and the simulated abnormal component removing signal obtained by the simulated abnormal component removing unit 16. The presence or absence of abnormality of the target 20 is determined.

A more specific embodiment will be described below.

FIG. 2 is a system conceptual diagram showing an embodiment of the abnormality monitoring apparatus of the present invention.

The diagnosis target machine equipment 21 corresponds to an example of the diagnosis target according to the present invention, and the machine control device 30.
It operates according to the control signal from.

The mechanical equipment 21 to be diagnosed has a sound collecting microphone 2 as a sensor for obtaining a signal for diagnosing whether or not there is an abnormality.
10 are deployed in the vicinity. The sound collecting microphone 210 senses the mechanical operation sound generated with the operation of the diagnostic target mechanical equipment 21, and converts it into an acoustic waveform signal. The acoustic waveform signal obtained by the sound collection microphone 210 is taken into the abnormality monitoring device main body 200. Further, control status information indicating the control status of the diagnostic target machine equipment 21 is input to the abnormality monitoring apparatus main body 200 from the machine control apparatus 300.
The abnormality monitoring device main body 200 uses the control state information to
It is possible to know the current operating state of the diagnostic target machine equipment 21.

In the abnormality monitoring device 200, the sound collecting microphone 210
In response to the input of the acoustic waveform signal from the machine controller 300 and the input of the control state information from the machine controller 300, first, when it is known that the diagnosis target machine equipment 21 is operating normally, For each operating state of 21, the inverse filter is created based on each acoustic waveform signal obtained when the diagnostic target mechanical equipment 21 is in each operating state. In addition, at the time of abnormality diagnosis, for each operating state of the diagnostic target mechanical equipment 21, the inverse filter created in the same operating state as the operating state is applied to the acoustic waveform signal to obtain the residual signal, Further, the power of the residual signal is obtained, and when the power is larger than a predetermined threshold value, it is determined that the diagnostic target mechanical equipment 21 has an abnormality.

Further, a speaker 240 as an abnormal sound generator is provided near the sound collecting microphone 210, and at the time of self-diagnosis for determining whether or not the abnormality monitoring apparatus body itself is normal, the abnormality monitoring apparatus is provided. Speaker 240 from main body 200
A signal for generating a simulated abnormal sound is sent to the speaker, a simulated abnormal sound is emitted from the speaker 240, and the simulated abnormal sound is sensed by the sound collecting microphone 210 together with the mechanical operation sound. From 210, the acoustic waveform signal including the component of the simulated abnormal noise is transmitted to the abnormality monitoring device main body 200.
Be transmitted to. In the abnormality monitoring device main body 200, a residual signal is applied to the received acoustic waveform signal by applying an inverse filter corresponding to the same operating state as the operating state of the diagnostic target mechanical equipment 21 when the acoustic waveform signal is received. The power is calculated, and the power is further calculated. When the power is larger than a predetermined threshold value, it is determined that the abnormality monitoring device is operating normally. Speaker 24
The volume of the simulated abnormal noise generated from 0 is adjusted so that the power thus obtained is at the volume of the abnormality detection limit at which the power is slightly higher than the threshold.

FIG. 3 is an external perspective view of an abnormality monitoring computer that operates as an embodiment of the main body of the abnormality monitoring device of the present invention (the main body of the abnormality monitoring device shown in FIG. 2).
An abnormality monitoring apparatus as an embodiment of the present invention is a combination of an abnormality monitoring apparatus main body including hardware of the abnormality monitoring computer 100 and software executed therein, and a sensor, a speaker, and the like (not shown). It is realized by.

The abnormality monitoring computer 100 has a C
A main body 101 including a PU, a RAM memory, a magnetic disk, a communication board, etc., and a CRT display 1 for displaying a screen on its display screen 102a according to an instruction from the main body.
02, a keyboard 103 for inputting operator's instructions and character information in the abnormality monitoring computer, and by designating an arbitrary position on the display screen, an instruction corresponding to an icon or the like displayed at that position is given. A mouse 104 for inputting is provided.

A CD-ROM 105 (see FIG. 4) is removably loaded into the main body 101, and the loaded CD-
A CD-ROM drive for driving the ROM 105 is also incorporated.

Here, the abnormality monitoring program is stored in the CD-ROM 105, and the CD-ROM 105 is stored.
Is loaded in the main body 101, and the abnormality monitoring program stored in the CD-ROM 105 is installed in the magnetic disk of the abnormality monitoring computer 100 by the CD-ROM drive. Abnormality monitoring computer 1
When the abnormality monitoring program installed in the magnetic disk No. 00 is activated, the abnormality monitoring computer 100 operates as an abnormality monitoring apparatus main body.

FIG. 4 is a hardware configuration diagram of the abnormality monitoring computer 100 shown in FIG.

In this hardware configuration diagram, a central processing unit (CPU) 111, RAM 112, magnetic disk controller 113, CD-ROM drive 115, mouse controller 116, keyboard controller 11 are shown.
7, display controller 118, communication board 1
19, a D / A conversion board 120 and an A / D conversion board 121 are shown, which are interconnected by a bus 110.

As described with reference to FIG. 3, the CD-ROM drive 115 is loaded with the CD-ROM 105, and accesses the loaded CD-ROM 105.

The communication board 119 is connected to the machine controller 300 (see FIG. 2) for controlling the diagnosis target, and the communication board 119 has a control status indicating the control status of the diagnosis target from the machine controller. Information is entered. The control state information includes a timing signal representing the operating state of the diagnosis target.

The D / A conversion board 120 is connected to the speaker 240 shown in FIG.
At 20, the simulated abnormal sound data for generating the simulated abnormal sound is converted from the speaker 240 into an analog signal.

A sensor such as the sound collecting microphone 210 shown in FIG. 2 is connected to the A / D conversion board 121. The A / D conversion board 121 has a role of inputting an analog signal picked up by the sensor, converting the analog signal into a digital signal, and taking the signal into the inside.

In FIG. 4, the magnetic disk 114 accessed by the magnetic disk controller 113 and the mouse 10 controlled by the mouse controller 116 are shown.
4, the keyboard 103 controlled by the keyboard controller 117, and the display controller 11
A CRT display 102 controlled by 8 is also shown.

Next, the abnormality monitoring program executed in the abnormality monitoring computer shown in FIGS. 3 and 4 will be described. The abnormality monitoring program described below is an abnormality monitoring program executed in the abnormality monitoring computer when the abnormality monitoring computer shown in FIGS. 3 and 4 is used as the abnormality monitoring apparatus main body 200 of the system shown in FIG. Is.

FIG. 5 is a flowchart of the inverse filter creation program executed in the preparation stage for abnormality monitoring.

This inverse filter creation program corresponds to the reference calculation unit 12 in the block diagram shown in FIG. When it is known that the diagnostic target mechanical equipment 21 shown in FIG. 2 operates normally, this inverse filter creation program performs normal operation by, for example, newly installing the diagnostic target mechanical equipment 21 or performing maintenance and performing a trial run or the like. After it is found that the operation is performed, the diagnosis target mechanical equipment 21 is executed while operating in the same manner as the main operation.

In the inverse filter creation program of FIG. 5, first, the control state information from the machine control device 300 of FIG. 2 is acquired (step a1), and the diagnostic target machine equipment 21 has newly created an inverse filter. It is determined whether or not it is in the proper operating state (step a2).

When the diagnostic target machine equipment 21 is in a new operating state, the acoustic waveform signal in that operating state obtained by the sound collecting microphone 210 is received (step a).
3) Based on the received acoustic waveform signal, an inverse filter corresponding to the operating state is created (step a)
4). If it is determined in step a2 that the new filter is not in the new operating state, that is, the inverse filter has already been created for the current operating state, the process directly proceeds to step a5.

In step a5, it is judged whether or not the creation of the inverse filter has been completed for all the operating states. If there is an operating state in which the inverse filter has not yet been created, the process returns to step a1 and the above processing is repeated. When the inverse filters are created for all the operating states, the execution of the inverse filter creation program ends.

FIG. 6 is a flowchart of the abnormality monitoring program. The abnormality monitoring program is executed when the abnormality is monitored after the inverse filters are created for all the operating states of the diagnostic target machine equipment 21.

In this abnormality monitoring program, control status information is first acquired (step b1), and the acquired control status information is selected for determining whether or not the abnormality monitoring device itself is operating normally. It is determined whether or not the predetermined operating state is displayed (step b2). As the predetermined operating state, in the present embodiment, the state in which the diagnostic target mechanical equipment 21 is stopped is selected, but other operating states other than the stationary state, which do not require monitoring for abnormality. You may choose.

The control state information acquired in step b1 is
When the operation state other than the predetermined operation state (here, the stopped state) is displayed, the process proceeds to step b3, and the acoustic waveform signal from the sound collecting microphone 210 is received.

In step b4, the diagnostic target machine equipment 21
Is diagnosed. That is, in this embodiment,
The received acoustic waveform signal is subjected to an inverse filter created for the same operating state as the current operating state to obtain the residual signal, and the power of the residual signal is obtained, and the power and threshold And are compared.

In this abnormality diagnosis, if it is determined that the power of the residual signal is equal to or more than the threshold value and an abnormality is detected (step b5), the process proceeds to step b6, and the abnormality occurs in the diagnostic target machine equipment 21. To that effect, an alarm is issued.

Alternatively, when it is determined in this abnormality diagnosis that the power of the residual signal is less than the threshold value, that is, when it is determined that the diagnostic target mechanical equipment 21 is normal (step b5), For reference, normal operation is recorded together with the current time and the current operation state (step b7), the process returns to step b1, and the above abnormality monitoring is repeated.

When it is determined in step b2 that the diagnostic target mechanical equipment 21 is in a predetermined state operation (here, the stationary state), the process proceeds to step b8, and last time, it is confirmed that the abnormality monitoring device itself is normal. It is determined whether or not a predetermined time or more has elapsed since the start. If the time is less than the predetermined time, it is not necessary to check the presence / absence of abnormality of the apparatus itself so frequently, and the process returns from step b8 to step b1.

When it is determined in step b8 that the abnormality monitoring device itself has an abnormality for the predetermined time or more, it proceeds to step b9 and the speaker 24 is activated.
A simulated abnormality was generated from 0, and at that time the sound collection microphone 2
The acoustic waveform signal obtained in step 10 is received (step b1
0), based on the received acoustic waveform signal, abnormality diagnosis is performed in the same manner as in step b4 (step b11).

Here, in step b9, the speaker 2
The volume of the simulated abnormal noise generated from 40 is determined in step b12.
When the power of the residual signal is compared with the threshold value in, the volume of the detection limit at which the power is slightly above the threshold value when the abnormality monitoring device is operating stably and normally. Is adjusted to.

In step b12, it is judged whether or not there is an abnormality in the abnormality monitoring device itself. Here, it is normal when the power of the residual signal is equal to or more than the threshold value, and the power is less than the threshold value. Sometimes it is determined to be abnormal.

When it is determined that there is an abnormality in step b12, the process proceeds to step b13, and an alarm is issued to the effect that there is some abnormality in the abnormality monitoring device itself.

On the other hand, if it is determined to be normal in step b12, the fact that this abnormality monitoring device is operating normally is recorded together with the current time for later reference (step b
14), the process returns to step b1 and the abnormality monitoring is continued.

FIG. 7 is a flowchart showing another example of the abnormality monitoring program. The abnormality monitoring program shown in FIG. 7 is executed in the abnormality monitoring computer shown in FIGS. 3 and 4 instead of the abnormality monitoring program shown in FIG. 6, and the diagnostic target machine equipment 21 is continuously operated. This is particularly effective when there is a stationary state or a state where it is not necessary to monitor the abnormality and there is a need to constantly monitor the abnormality.

Compared with the abnormality monitoring program shown in FIG. 6, steps c1 to c12 of the abnormality monitoring program shown in FIG. 7 are different from step b of the abnormality monitoring program shown in FIG.
The only difference is that the determination of No. 2 does not exist. Excluding this point, step b1 of the abnormality monitoring program shown in FIG.
~ Step b13 is the same as that of step b13, and only the different points will be described here for the overlapping portions.

That is, in the abnormality monitoring program of FIG. 6, the process proceeds to step b8 only when it is in a predetermined operating state (for example, a stationary state) in step b2, and it is checked whether or not there is an abnormality in the abnormality monitoring device itself. Since the abnormality monitoring program has no determination step corresponding to step b12 in FIG. 6, no matter what the current operating state is,
The abnormality of the abnormality monitoring device itself is also detected.

When it is determined in step c11 that the abnormality monitoring device itself is normal, the process proceeds to step c13.
From the acoustic waveform signal obtained in step c9, the signal component resulting from the simulated abnormal sound emitted from the speaker 24 is removed. To explain further, the acoustic waveform signal received in step c9 is the signal component of the simulated abnormal sound from the speaker 240 and the operation of the diagnostic target mechanical equipment 21 at the same timing as the simulated abnormal sound is emitted from the speaker 240. The signal component of the resulting machine operation sound is mixed, and in step c13, only the signal component resulting from the simulated abnormal noise from the speaker 240 is removed from both of these signal components to perform the machine operation. Only the signal component due to the sound is extracted.

The signal component resulting from the simulated abnormal sound emitted from the speaker 240 is preliminarily detected by the mechanical equipment 2 to be diagnosed.
When 1 is stopped, a simulated noise is emitted from the speaker 240, and the acoustic waveform signal obtained by the sound collecting microphone at that time is stored. From the speaker 240,
The same simulated noise as that at that time is always emitted, and therefore, by subtracting the stored signal, it is possible to remove the signal component caused by the simulated noise.

When the simulated abnormal sound component removal signal is generated by removing the signal component caused by the simulated abnormal sound as described above in step c13, the simulated abnormal sound component removed signal is used as the operation at that time. The abnormality diagnosis is performed by applying the inverse filter that has been obtained for the same operating state as the state (step c14). Abnormality is determined by comparing the power of the residual signal with the threshold value (step c15). When the power of the residual signal is larger than the threshold value, the process proceeds to step c16, and the diagnostic target mechanical equipment 2
A warning is issued that an abnormality has occurred in 1. On the other hand, when the power of the residual signal is less than the threshold value, it is recorded that both the diagnosis target mechanical equipment 21 and this abnormality monitoring device are operating normally together with the current time and the current operating state (step c17). ), The process returns to step c1 and the abnormality monitoring is continued.

According to this abnormality monitoring program, it is possible to monitor the presence or absence of abnormality of the diagnostic target machine equipment at all timings including the timing at which a simulated abnormal noise is emitted from the speaker 240 to diagnose the abnormality of the abnormality monitoring device itself. This is extremely effective when there is no room for monitoring abnormalities.

FIG. 8 is a system diagram showing an example in which the present invention is applied to monitor gas leakage from a pipe.

Since a gas leak sound in the ultrasonic range is emitted when gas leaks, a sound collecting microphone having sensitivity to sound in the ultrasonic range is used here as a sound collecting microphone, and a supersonic noise generator An ultrasonic speaker that emits a simulated gas leak sound in the sound wave region is adopted. Other configurations are the same as those of the embodiment described with reference to FIGS.

FIG. 9 is a system diagram showing another example to which the present invention is applied.

In the system of FIG. 9, compared with the system of FIG. 8, instead of the ultrasonic speaker in the system of FIG. 8, an example is shown in which a simulated leaking jet nozzle for jetting a safe gas is provided. . A small shutoff valve, etc. is installed between the tank containing the safe gas and the simulated leaking jet nozzle. A safe gas is ejected, and a simulated noise generated at that time is picked up by a sound collection microphone.

This system is effective in constructing a storm type system.

In each of the above embodiments, the sound is captured by the sound collecting microphone, but a vibration sensor for picking up the vibration of the diagnosis object is provided, and the abnormality is monitored based on the signal picked up by the vibration sensor. May be performed, or the abnormality monitoring may be performed by capturing a different physical quantity other than sound and vibration.

Next, an inverse filter and a method of detecting the presence / absence of abnormality using the inverse filter will be described.

An arbitrary time series signal can be regarded as an output when white noise is input to an appropriate linear system. Determining the corresponding linear system from a given time series signal is called linear predictive analysis, and there is an established method.
An autoregressive model (AR model) is usually obtained in this way. When the sampled and discretized time series signals are X (n), n = 1, 2, ...
The signal X (n) at the time point is determined from the data of M time points before that as follows.

[0085]

[Equation 1]

Here, e (n) is a virtual input signal to the linear system, which is white noise. Given a time series signal,
By obtaining the set of coefficients {A _k } from the data,
An autoregressive model for the time series signal is determined.

When the coefficient set {A _k } is obtained, Y (n) is defined as follows using the time-series signal data {X (n)}. At this time, Y (n) is called a linear predicted value of X (n).

[0088]

[Equation 2]

When the following quantity is calculated,
From the expressions (1) and (2), X (n) -Y (n) = e (n) (3), and the residual becomes white noise. That is, when the prediction value Y (n) obtained from the previous M data is subtracted from the time-series signal data X (n) at the n-th time point, the input white noise is obtained. Here, obtaining the residual e (n) by subtracting the predicted value Y (n) from X (n) is referred to as applying an inverse filter. If a certain time series signal can be represented by an appropriate autoregressive model as described above, white noise is obtained by causing an inverse filter configured using the same to act on the original time series signal. That is, the input signal is whitened by the inverse filter. In this case, the input time-series signal does not have to be the signal itself used when designing the inverse filter, and if the autoregressive models are the same, that is, signals with the same characteristics, then a whitened signal is obtained as the output. You can However, if the characteristics of the time-series signal are different from those used in the design, whitening is not performed and white noise is not obtained even if the inverse filter is operated.

Therefore, an inverse filter is configured in advance using the first time-series signal carrying normal operating sound, vibration, etc. (operating sound, etc.), and carries the operating sound, etc. at any time. To detect a time-series signal (residual signal) different from the normal time by obtaining a new second time-series signal to be applied and applying an inverse filter to the second time-series signal to monitor the output. Can be done.

In this embodiment, specifically, the following signal processing method can be adopted.

First, in the reference calculation unit 12 of FIG. 1, 1024 points of sound signal data or vibration signal data obtained when the diagnosis target is in a normal state are used to perform FFT.
(Fast Fourier Transform) is performed, and then the power spectrum is obtained. Then it is IFFT (Inverse Fast Fourier Transform)
To obtain the autocorrelation function, and use it to Levinso
n algorithm (for example, see "Introduction to Digital Signal Processing" by Mikami, published by CQ Publishing) to obtain the coefficient {Ak} of the inverse filter.

After that, in the abnormality presence / absence determining unit 340, the inverse filter is operated to obtain the residual signal, and the calculation for obtaining the residual signal is performed by the coefficient {a [k]} in this embodiment. Is performed by moving average calculation.

Here, in order to obtain the moving average of the power of the residual signal, first, 128 data are extracted from the time series of the residual signal, and FFT, power spectrum calculation, IFFT are performed.
After that, the autocorrelation function is obtained, and the power is obtained from the peak value at the origin. After that, the power is sequentially obtained while shifting the starting point of the data by 50 points.

As the inverse filter, as an example, the order M
= 27, the coefficient {a [k]} shown in Table 1 is adopted.

[0096]

[Table 1] a [0] = 1.000000 a [1] = − 2.887330 a [2] = 3.947344 a [3] = − 3.535249 a [4] = 2.447053 a [5] = -1.620133 a [6] = 1.315352 a [7] =-1.268161 a [8] = 0.937471 a [9] =-0.380573 a [10] =-0.040919 a [ 11] = 0.284076 a [12] = − 0.353665 a [13] = 0.397849 a [14] = − 0.533185 a [15] = 0.501902 a [16] = − 0.238178 a [17] = − 0.003048 a [18] = 0.192220 a [19] = − 0.166854 a [20] = − 0.010498 a [21] = 0.061383 a [22] = 0.017323 a [23] =-0.014146 a [24] =-0.131247 a [25] = 0.239157 a [26] =-0.242444 a [27] = 0.115678 13 shows an example of a waveform obtained from equipment in a normal state, FIG. 10 shows a signal waveform of a sound signal sampled from equipment in a normal state, and FIG. 11 shows an inverse filter applied to this sound signal. FIG. 12 shows the signal waveform of the residual signal after the error, FIG. 12 shows the power spectrum of this residual signal, and FIG. 13 shows the moving average of the power of the residual signal.

FIGS. 14 to 17 show an example of waveforms obtained when the equipment is in an abnormal state, and each drawing shows the signal waveform of the same format as FIGS. 10 to 13, respectively. Even if the waveforms of the sound signals obtained from the equipments in the normal state and the abnormal state are respectively directly compared with each other, it is difficult to immediately judge the normality / abnormality from these. However, it is possible to judge normality / abnormality by analyzing FIG. 11 and FIG. 15 showing residual signals obtained by applying an inverse filter to these.

As can be easily understood by comparing FIGS. 11 and 15, the amplitude of the residual signal is extremely small when the equipment is in a normal state. Will be extremely large. Therefore, it is possible to judge normality / abnormality based on the maximum value of the power in the residual signal. For example, by determining that the amplitude of the residual signal that is 10 dB or more larger than the maximum power of the signal obtained when the equipment is in a normal state is abnormal, and the amplitude of the residual signal that is less than this is normal, , It is possible to judge normality / abnormality.

Further, as shown in FIGS. 12 and 16, in the spectrum obtained by Fourier transforming the residual signal, if an abnormality occurs, the power spectrum increases. For example, the power peak in FIG. 12 is 100 dB or less, but in FIG. 16, the power peak is almost 120 dB.
It has reached dB.

Further, by comparing FIG. 13 and FIG. 17 showing the moving average of the power of the residual signal, it is found that this moving average increases due to the abnormality. For example, when the moving average data is 20 dB or more larger than the maximum value of the moving average when the equipment is in the normal state, it can be determined as abnormal, and when the data is less than this, it can be determined as normal. When this method is adopted, it is particularly easy to determine whether or not there is an abnormality, and it is possible to detect an abnormality in a short time. Therefore, it is possible to perform real-time detection in the field, which is particularly preferable. Depending on the type of anomaly, the presence of the defect can be detected more accurately by detecting the power spectrum than by analyzing the moving average of the power.

In the embodiment described above, an inverse filter is created as reference data according to the present invention, a residual signal is obtained, and the presence or absence of abnormality is detected based on the residual signal. The present invention does not necessarily have to adopt this detection method, and may employ, for example, the above-mentioned spectral analysis method or statistical estimation or test method.

[0102]

As described above, according to the present invention, it is possible to monitor the presence / absence of an abnormality in a diagnosis target and correctly determine the presence / absence of the abnormality.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic embodiment of an abnormality monitoring device of the present invention.

FIG. 2 is a system conceptual diagram showing an embodiment of an abnormality monitoring device of the present invention.

FIG. 3 is an external perspective view of an abnormality monitoring computer that operates as an embodiment of the main body of the abnormality monitoring apparatus of the present invention (the abnormality monitoring apparatus main body shown in FIG. 2).

FIG. 4 is a hardware configuration diagram of the abnormality monitoring computer shown in FIG.

FIG. 5 is a flowchart of an inverse filter creation program.

FIG. 6 is a flowchart of an abnormality monitoring program.

FIG. 7 is a flowchart showing another example of an abnormality monitoring program.

FIG. 8 is a system diagram showing an example in which the present invention is applied to monitor gas leakage from a pipe.

FIG. 9 is a system diagram showing another example to which the present invention is applied.

FIG. 10 is a waveform diagram of a sound signal obtained from equipment in a normal state.

11 is a signal waveform diagram obtained by applying an inverse filter to the signal of FIG.

FIG. 12 is a power spectrum diagram obtained from the signal of FIG.

FIG. 13 is a diagram showing a moving average of electric power obtained from the signals of FIG. 11.

FIG. 14 is a waveform diagram of a sound signal obtained from equipment in an abnormal state.

15 is a signal waveform diagram obtained by applying an inverse filter to the signal of FIG.

16 is a power spectrum diagram obtained from the signal of FIG.

FIG. 17 is a diagram showing a moving average of electric power obtained from the signals of FIG. 15;

[Explanation of symbols]

10 Abnormality monitoring device 11 sensors 12 Reference calculator 13 Diagnosis target abnormality determination unit 14 Simulated abnormal state generation part 15 Self-abnormality determination unit 16 Simulated abnormal component removal unit 20 diagnosis target 21 Mechanical equipment to be diagnosed 200 Abnormality monitoring device body 210 Sound collecting microphone 240 speakers 300 Machine control device

Claims (4)

[Claims] 1. An abnormality monitoring device for monitoring the presence or absence of an abnormality in a diagnosis target, comprising: a sensor for capturing a predetermined physical quantity that reflects the state of the diagnosis target to obtain a signal representing the physical quantity; Based on a reference signal derived from a reference signal obtained when in a state, and a diagnostic signal obtained by the sensor during abnormality monitoring, a diagnosis target abnormality determination unit that determines the presence or absence of abnormality of the diagnosis target, and A simulated abnormal state generation unit that gives a sensor a physical amount that simulates when the diagnosis target is in an abnormal state, the reference data, and the sensor at a timing at which the sensor captures the physical amount given from the simulated abnormal state generation unit. And a self-abnormality determining unit that determines whether or not there is an abnormality in the abnormality monitoring device itself based on the simulated abnormal component mixing signal obtained by That abnormality monitoring apparatus. 2. A reference calculation unit that calculates the reference data by a calculation including a calculation that calculates an inverse filter based on the reference signal, wherein the diagnosis target abnormality determination unit applies the inverse filter to the diagnosis signal. By performing an operation including an operation to obtain a residual signal by determining the presence or absence of an abnormality of the diagnosis target based on the result of the operation, the self-abnormality determination unit, the simulated abnormal component mixed signal to the 2. An operation including an operation for obtaining a residual signal by operating an inverse filter is performed, and whether or not there is an abnormality in the abnormality monitoring device itself is determined based on the result of the operation. The abnormality monitoring device described. 3. The simulated abnormal condition generation unit causes the sensor to have a level corresponding to a lower limit at which the abnormality monitoring device determines that the diagnosis target is abnormal when the abnormality monitoring device is in a normal state. The abnormality monitoring device according to claim 1, wherein a physical quantity simulating an abnormal state is given. 4. A simulated abnormal component removing unit for generating a simulated abnormal component removing signal by removing from the simulated abnormal component mixed signal the simulated abnormal component caused by the physical quantity given by the simulated abnormal state generation unit by the sensor. The diagnosis target abnormality determination unit, with respect to the timing when the sensor is given a physical quantity simulating that the diagnosis target is in an abnormal state by the simulated abnormal state generation unit,
The abnormality monitoring device according to claim 1, wherein the presence or absence of abnormality of the diagnosis target is determined based on the reference data and the simulated abnormal component removal signal.