JP2010266327A - Facility diagnosis device and facility diagnosis method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a facility diagnosis device and a facility diagnosis method capable of diagnosing an operation state highly accurately in a short period, with respect to a facility executing a plurality of different operations as well. <P>SOLUTION: This facility diagnosis device 1 includes: a sound collection part 10 for collecting operation sounds generated from the facility 2; a storage part 13 for storing a reference spectrum which is a time frequency spectrum of an operation sound of the facility 2 in a prescribed period when a plurality of operations of the facility 2 are executed, when the facility 2 is operated normally; a time frequency analysis part 21 for calculating the time frequency spectrum corresponding to operation sounds during a prescribed period, by performing time frequency conversion of operation sounds during the prescribed period collected by the sound collection part 10; a comparison part 22 for specifying a spectrum component corresponding to each operation sound in the plurality of operations of the facility 2, by comparing the time frequency spectrum with the reference spectrum; and a determination part 23 for determining whether an abnormality is generated in facility 2 or not, based on each spectrum component. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an equipment diagnosis apparatus and an equipment diagnosis method for diagnosing an operation state of equipment by analyzing sounds generated during operation of the equipment.

  2. Description of the Related Art Conventionally, techniques for diagnosing the operating state of a machine by analyzing sounds generated during the operation of the machine have been developed (see, for example, Patent Documents 1 and 2).

For example, the fault diagnosis apparatus disclosed in Patent Document 1 obtains the linear prediction coefficients of the zeroth order and the first and subsequent orders of the measured sound or vibration by frequency analysis means. Then, the failure diagnosis apparatus causes the display means to display the multiplication values obtained by multiplying the zeroth-order linear prediction coefficient and the linear prediction coefficient after the first order along the measurement time.
Further, the rotating device state diagnosis support apparatus disclosed in Patent Document 2 detects an abnormality based on a difference between an operation sound during normal operation and a collected operation sound.

JP 2003-57210 A JP 2005-77111 A

The diagnostic technique based on the analysis of the operation sound described above is effective for failure diagnosis of a device that performs a single operation, such as a rotating body that rotates at a constant speed. This is because there is basically only one type of operation sound that occurs when a device that operates in a single operation is normal, and the operation sound that is collected during the operation of that device is compared with the operation sound during normal operation. This is because it is easy.
However, for example, equipment that performs a plurality of different operations such as transportation, holding, assembling, or cleaning of parts, such as a production facility for automobile parts, generates an operation sound having different characteristics for each operation. Therefore, when trying to diagnose the operating state of the facility that performs a plurality of different operations by the diagnostic device and method using this diagnostic technology, the diagnostic device and method obtains an operation sound for each operation of the facility, It is necessary to analyze the operation sound. Therefore, the diagnostic apparatus and method using this diagnostic technique require a long time for diagnosing equipment that performs a plurality of different operations.

Also, when diagnosing equipment that performs multiple different operations using various measuring devices such as ammeters, voltmeters, and shape measuring devices, these measuring devices can be attached to the equipment to be diagnosed, Since the target equipment must be disassembled, the time required for diagnosing the equipment is enormous. In particular, when the device to be diagnosed is a production facility, the longer the time required for the diagnosis, the longer the time to stop the production facility, so the economic loss of the user of the production facility also increases. Therefore, it is desirable that the time required for diagnosis of the operating state is as short as possible.
On the other hand, if an operator tries to diagnose the operating state of equipment that performs multiple different operations by listening to the operation sound or by visual inspection, a malfunction that causes abnormal operation due to the sensory evaluation or intuition of the operator The location will be specified. For this reason, the detection accuracy of the fault location varies depending on the skill level and physical condition of the worker.

  Accordingly, an object of the present invention is to provide a facility diagnosis apparatus and facility diagnosis method capable of diagnosing the operation state in a short period of time and with high accuracy even for a facility that performs a plurality of different operations. is there.

According to the description of claim 1, an equipment diagnosis apparatus is provided as one aspect of the present invention. The facility diagnosis apparatus performs a plurality of operations of the facility (2) when the facility (2) is operating normally, and the sound collection unit (10) that collects the operation sound generated by the facility (2) A storage unit (13) for storing a reference spectrum, which is a time frequency spectrum of the operation sound of the equipment (2) for a predetermined period, and a time frequency conversion of the operation sound collected by the sound collection unit (10) for a predetermined period By comparing the time frequency spectrum with the reference spectrum, the time frequency analysis unit (21) that calculates the time frequency spectrum corresponding to the operation sound of the predetermined period, and each of the plurality of operations of the facility (2) The comparison part (22) which specifies the spectrum component corresponding to the operation sound of the above and the determination part (23) which determines whether or not an abnormality has occurred in the facility (2) based on each spectrum component.
By having such a configuration, this equipment diagnosis apparatus can diagnose the operation state of a plurality of different operations with high accuracy in a short period of time.

According to a second aspect of the present invention, the determination unit (23) is configured to determine whether the operation sound spectrum component corresponding to the first operation among the plurality of operations and the operation sound spectrum component corresponding to the second operation. When the time interval is different from the time interval between the first operation and the second operation during normal operation, it is preferable to determine that an abnormality has occurred in the facility (2).
Thereby, this equipment diagnostic device can detect that the operation timing of each part which equipment has has shifted.

According to the third aspect of the present invention, when the spectrum component corresponding to any one of the plurality of operations is not detected by the comparison unit (22), the determination unit (23) has an abnormality in the facility (2). It is preferable to determine that it has occurred.
Thereby, this equipment diagnostic device can detect that any operation which the equipment should have performed was not performed.

Alternatively, according to the description of claim 4, the determination unit (23) determines that an abnormality has occurred in the facility (2) when a spectral component that does not correspond to any of the plurality of operations is detected. preferable.
Thereby, since this equipment diagnostic apparatus can detect the abnormal sound which the equipment emitted, it can detect that some kind of abnormality has occurred in the equipment. Further, the frequency band of the spectral component corresponding to the abnormal sound is different from the frequency band of the spectral component of the operating sound due to the normal operation of the facility. Therefore, this equipment diagnosis apparatus can detect that an abnormality has occurred in the equipment even if an abnormal sound is generated while an operation sound is generated due to the normal operation of the equipment.

Furthermore, according to the description of claim 5, when the determination unit (23) determines that an abnormality has occurred in the facility (2), the spectrum component that is the basis for the determination is generated in the facility (2). When the abnormality cause condition corresponding to the specified abnormality is satisfied, it is preferable to estimate that the cause of the abnormality that has occurred in the facility (2) is a cause associated with the abnormality cause condition.
Note that the spectral component that is the basis for determining that an abnormality has occurred in the facility should be detected not only as a spectral component detected from the time-frequency spectrum but also as a spectral component that originally corresponds to one of the operations. Of the spectral components, it may be an undetected spectral component.

Furthermore, according to the description of claim 6, an equipment diagnosis method is provided as another embodiment of the present invention. The facility diagnosis method includes a step of collecting operation sound generated by the facility, and time-frequency conversion of the collected operation sound for a predetermined period in which a plurality of operations of the facility are performed, thereby performing an operation for a predetermined period. By calculating the time frequency spectrum corresponding to the sound, by comparing the time frequency spectrum with a reference spectrum that is the time frequency spectrum of the operation sound of the equipment for a predetermined period when the equipment is operating normally, The method includes a step of specifying a spectrum component corresponding to each operation sound of a plurality of operations of the facility, and a step of determining whether an abnormality has occurred in the facility based on each spectrum component.
By including such a procedure, this facility diagnosis method can diagnose the operation state of a facility that performs a plurality of different operations in a short period of time and with high accuracy.

  In addition, the code | symbol in the parenthesis attached | subjected to each said means is an example which shows a corresponding relationship with the specific means as described in embodiment mentioned later.

It is a schematic block diagram of the equipment diagnostic apparatus by one Embodiment of this invention. It is a functional block diagram of the control part of the equipment diagnostic apparatus shown in FIG. It is a figure which shows an example of a reference | standard spectrum. It is a figure which shows an example of the time frequency spectrum with respect to the operation sound of an installation. It is an operation | movement flowchart of the equipment diagnostic process performed by the equipment diagnostic apparatus which concerns on embodiment of this invention.

Hereinafter, a facility diagnosis apparatus according to an embodiment of the present invention will be described.
This facility diagnosis apparatus collects operation sounds generated during one operation cycle including a plurality of operations that the diagnosis target facility repeatedly executes at a constant cycle. And this equipment diagnostic apparatus calculates the time frequency spectrum of the operation sound of the 1 operation cycle by carrying out time frequency conversion of the collected sound. This facility diagnostic apparatus compares the calculated time-frequency spectrum with a reference time-frequency spectrum corresponding to the operation sound of one operation cycle when the diagnosis-target facility is operating normally, thereby providing a spectral component for each operation. By identifying the above, the presence or absence of equipment abnormality is determined.

FIG. 1 is a schematic configuration diagram showing an overall configuration of a facility diagnosis apparatus 1 according to one embodiment of the present invention.
In the present embodiment, the facility 2 which is an example of the facility to be diagnosed includes two rotating brushes 4-1 and 4-2 attached to the movable frame 5 for each workpiece 3 conveyed on the belt conveyor 7. The dust adhering to the surface of the work 3 is removed by sweeping in order. For this purpose, the equipment 2 has a control circuit 6, which controls the two rotating brushes 4-1 and 4-2 and the movable frame 5. That is, the movable frame 5 moves in parallel to the upper surface of the workpiece 3 or moves perpendicular to the upper surface of the workpiece 3 in accordance with a control signal received from the control circuit 6. For this reason, the control circuit 6 can bring either one of the rotating brushes 4-1 and 4-2 into contact with the workpiece 3. The rotating brushes 4-1 and 4-2 rotate at a constant rotational speed or stop according to the control signal received from the control circuit 6.
Further, the control circuit 6 outputs a signal indicating the operation state of the facility 2, for example, a signal indicating that the movable frame 5 starts a predetermined operation to the facility diagnosis apparatus 1.

The facility 2 operates the movable frame 5 and the rotating brushes 4-1 and 4-2 according to a predetermined operation sequence. And the installation 2 performs each operation | movement included in the operation cycle for every workpiece | work 3 by making operation | movement of the movable frame 5 and the rotation brushes 4-1 and 4-2 according to the predetermined order into 1 operation cycle.
Here, when the movable frame 5 of the equipment 2 moves, or when any of the rotating brushes sweeps the work 3, the operation by the movable frame 5 or the rotating brush 4-1 or 4-2 is mechanically operated. Sound is generated. Further, the operation sound generated when the movable frame 5 moves is different from the operation sound generated when the rotating brush 4-1 or 4-2 sweeps the workpiece 3. Furthermore, these operating sounds occur at different times.

  Therefore, the facility diagnosis apparatus 1 diagnoses the operation state of the facility 2 by analyzing the operation sound for each operation cycle of the facility 2. Therefore, as shown in FIG. 1, the equipment diagnosis apparatus 1 includes a microphone 10, an amplifier 11, an analog-digital converter 12, a storage unit 13, a communication unit 14, a display unit 15, and a control unit 16. And have.

  The microphone 10 functions as a sound collection unit, collects an operation sound generated by the operation of the equipment 2, and measures a measurement signal corresponding to the intensity of the collected sound as an analog electric signal in time series. Output as an audio signal. In this embodiment, a condenser microphone having a frequency band of 20 Hz to 20 kHz is used. However, another type of microphone such as a piezoelectric microphone may be used as the microphone 10. Further, the frequency band of the sound that can be collected is not limited to the above, and for example, a frequency band of 100 Hz to 18 kHz may be used. The optimum microphone system and measurable frequency band can be selected as appropriate according to the type, size, or structure of the equipment to be diagnosed.

The amplifier 11 amplifies the analog audio signal acquired from the microphone 10 and outputs it to the analog-digital converter 12. The analog-digital converter 12 converts the amplified analog audio signal into a digital audio signal. In this embodiment, the analog-to-digital converter 12 used is one that samples an input analog signal at a sampling frequency of 44.1 kHz and outputs it as a digital signal having a resolution of 16 bits. The analog-to-digital converter is not limited to the above-described one, and those having different sampling frequencies and resolutions may be used. For example, the analog-digital converter 12 having a sampling frequency of 48 kHz and a resolution of 12 bits may be used.
The digitized audio signal is sent to the control unit 16.

The storage unit 13 includes a magnetic recording medium such as a hard disk, a random access memory (RAM), a semiconductor memory such as a flash memory, a readable / writable optical recording medium such as a CD-RW and a DVD-R / W, and its access. Having at least one of the devices. The storage unit 13 stores programs used in the control unit 16 and various data. Furthermore, the memory | storage part 13 memorize | stores the reference | standard spectrum which is a time frequency spectrum of the operation sound for 1 operation cycle when the installation 2 is operating normally. Further, the storage unit 13 temporarily stores a digital audio signal corresponding to the operation sound measured during the diagnosis of the facility 2 together with the time when the digital audio signal is acquired until the control unit 16 analyzes it.
The storage unit 13 passes various programs, data, or reference spectrum to the control unit 16 in response to a request from the control unit 16.

The communication unit 14 is an interface for connecting the equipment diagnosis apparatus 1 and the equipment 2 or another apparatus (for example, a management apparatus that manages a process in which the equipment 2 is installed) via a line according to a predetermined communication standard. It has a circuit. Then, the communication unit 14 receives a signal indicating that the facility 2 starts a predetermined operation from the control circuit 6 of the facility 2 and passes the signal to the control unit 16. The communication unit 14 receives a control signal for controlling the facility 2, for example, a control signal for stopping the facility 2 or executing a specific operation from the control unit 16, and transmits the control signal to the control circuit 6 of the facility 2. .
Further, the communication unit 14 may output the diagnosis result information of the facility 2 received from the control unit 16 to another device.

The display unit 15 displays the diagnosis result of the equipment 2, other predetermined information, and the like, and includes, for example, a liquid crystal display. The display unit 15 displays the diagnosis result of the equipment 2 by the control unit 16. In addition, when the control unit 16 determines that an abnormality has occurred in the facility 2, the display unit 15 displays a message corresponding to the estimated cause of the abnormality.
The equipment diagnosis apparatus 1 may include a speaker that is separate from the display unit 15 or integrated with the display unit 15. When the facility diagnosis apparatus 1 determines that the facility 2 is not operating normally, the facility diagnosis apparatus 1 may cause the speaker to emit an alarm sound indicating that the facility 2 is not operating normally. Further, when the diagnosis result of the facility 2 by the control unit 16 is output to another device via the communication unit 14, the display unit 15 may be omitted.

The control unit 16 has one or a plurality of processors and their peripheral circuits. And the control part 16 diagnoses the operation state of the installation 2 by analyzing the digital audio | voice signal corresponding to the operation sound of the installation 2 received from the analog-digital converter 12. FIG.
FIG. 2 is a functional block diagram showing functions realized by the control unit 16. The control unit 16 includes a time frequency analysis unit 21, a comparison unit 22, and a determination unit 23. Each of these units included in the control unit 16 is a functional module implemented by a computer program executed on a processor included in the control unit 16. Alternatively, these units included in the control unit 16 may be dedicated arithmetic circuits that realize the functions of the units.

The time-frequency analysis unit 21 performs time-frequency conversion of the digital audio signal corresponding to the operation sound of the facility 2 for each operation cycle of the facility 2, thereby generating a time-frequency spectrum corresponding to the operation sound of one operation cycle of the facility 2. Ask for. Therefore, the time frequency analysis unit 21 starts the first operation of the next operation cycle from the time when the first operation of the operation cycle is started, for example, from the digital audio signal temporarily stored in the storage unit 13. A signal corresponding to the period until the set time is extracted as a signal corresponding to the operation sound of one operation cycle. The time when the first operation of the operation cycle is started can be the time when the control unit 16 receives a signal indicating that the first operation has started from the control circuit 6 of the facility 2.
In the present embodiment, the time-frequency analysis unit 21 calculates a time-frequency spectrum by performing wavelet transform on a digital audio signal corresponding to the operation sound of the facility 2. The wavelet transform is performed based on the following formula.
Here, h (t) is the intensity of the digital audio signal at the elapsed time t from the start of measurement, and the function Ψ (y) is the mother wavelet. Further, a is a parameter for determining the frequency, and b is a parameter for defining the parallel movement amount on the time axis.

In this embodiment, a Gabor function is used as the mother wavelet. As the mother wavelet, another function used as a mother wavelet in signal analysis such as speech analysis, such as a Mexican hat function, may be used instead of the Gabor function.
The time frequency analysis unit 21 passes the calculated time frequency spectrum to the comparison unit 22.

  The comparison unit 22 compares the time frequency spectrum received from the time frequency analysis unit 21 with the reference spectrum. And the comparison part 22 detects the spectrum component corresponded to each operation | movement contained in one operation cycle of the installation 2 from a time frequency spectrum.

  For example, the comparison unit 22 executes the pattern matching while shifting the spectral component corresponding to the specific operation included in the reference spectrum and the time-frequency spectrum in the time axis direction, thereby setting the time shift amount as a variable. To obtain the correlation value. Then, the comparison unit 22 converts the spectrum component when the spectrum component corresponding to the specific operation and the time-frequency spectrum are the best match, that is, the spectrum component of the time-frequency spectrum corresponding to the maximum correlation value to the specific operation. It shall correspond to. However, when the maximum correlation value is less than a predetermined threshold, the comparison unit 22 may determine that the spectrum component corresponding to the specific operation has not been detected. The predetermined threshold is set to an appropriate value according to the simulation or the experimental result. For example, the predetermined threshold value can be the minimum value of the correlation values of two spectral components obtained separately from the experimental results obtained from the experimental measurement results.

  Moreover, the comparison part 22 may limit the range of the time frequency spectrum made into the object of pattern matching to the time slot | zone when the specific operation | movement may be performed. For example, for the spectral component of the reference spectrum corresponding to the operation first performed in one operation cycle of the facility 2, the comparison unit 22 sets the range of the time frequency spectrum to be subjected to pattern matching from the beginning of the time frequency spectrum. You may limit to the range until 1 / 4-1 / 3 of the required time corresponded to 1 operation cycle passes.

  Furthermore, when the comparison unit 22 detects a spectral component in the time-frequency spectrum corresponding to one operation as a reference among the operations included in one operation cycle of the facility 2, the spectral component corresponding to the reference operation is detected. A time range in which spectral components corresponding to other operations are detected may be set on the basis of the time at which the signal exists. For example, if the equipment 2 is operating normally, the relative relationship between the time during which the movable frame 5 moves and the time during which the rotary brushes 4-1 and 4-2 are sweeping the workpiece 3 is It is decided in advance. Therefore, for example, it is assumed that the spectral component corresponding to the operation in which the movable frame 5 is lowered and the rotary brush 4-1 is brought into contact with the workpiece 3 is detected from the measured time-frequency spectrum. In this case, the comparison unit 22 has a time difference from the end of the operation of bringing the rotating brush 4-1 into contact with the work 3 to when the rotating brush 4-1 or the rotating brush 4-2 starts to sweep the work 3, respectively. Is determined based on the setting of the facility 2. And the comparison part 22 is the rotation brush 4-1 or the rotation brush 4- after the time which added the said time difference from the time when the spectral component corresponding to the operation | movement which makes the rotation brush 4-1 contact the workpiece | work 3 was detected. 2 detects a spectral component corresponding to the operation of sweeping the workpiece 3.

In addition, the comparison unit 22 has an abnormal sound having a predetermined intensity from spectrum components that do not match any spectrum component corresponding to the specific operation of the facility 2 represented in the reference spectrum in the measured time-frequency spectrum. Spectral components that are equal to or greater than the threshold are detected. This is because such a spectral component may correspond to an abnormal sound that does not occur when the facility 2 is operating normally. For example, when the average value of the spectrum components in a specific frequency band has an intensity equal to or higher than the abnormal sound threshold for a certain period, the comparison unit 22 converts the spectrum component in the specific frequency band during the certain period to an abnormal sound. It detects as a corresponding spectrum component.
The abnormal sound threshold is preferably set to a value that does not detect the spectral component corresponding to the environmental sound as the spectral component corresponding to the abnormal sound. For example, the abnormal sound threshold is larger than the maximum intensity of the spectrum component corresponding to the environmental sound when the movable part of the equipment 2 (that is, the rotating brushes 4-1, 4-2 and the movable frame 7) is not operating, And it is set to a value smaller than the intensity of the spectral component corresponding to the operation sound of any movable part of the facility 2. Moreover, it is preferable that the predetermined fixed period is set to a period in which a spectrum component corresponding to the sudden sound generated around the facility 2 is not detected as a spectrum component corresponding to the abnormal sound. For example, the fixed period is set to any value between 1 second and 10 seconds. In addition, when an abnormal sound related to an abnormality that is assumed to occur in the facility 2 is known in advance by an experiment or the like, the abnormal sound threshold value and the predetermined fixed period may be set according to a spectrum component corresponding to the abnormal sound. Good.

  The comparison unit 22 notifies the determination unit 23 of the time during which there is a spectrum component corresponding to each operation included in one operation cycle of the facility 2. In addition, when a spectrum component corresponding to any operation is not detected, the comparison unit 22 notifies the determination unit 23 of a signal indicating the operation that has not been detected. Further, when a spectral component corresponding to the abnormal sound is detected, the comparison unit 22 notifies the determination unit 23 of the frequency band and time in which the spectral component corresponding to the abnormal sound exists.

  The determination unit 23 receives the time from which the spectral component corresponding to each operation of the facility 2 exists, the frequency band and the time in which the signal indicating the operation not detected and the spectral component corresponding to the abnormal sound exist. Based on the above, the operating state of the facility 2 is diagnosed.

FIG. 3 is a diagram illustrating an example of a reference spectrum. In FIG. 3, the horizontal axis represents the elapsed time t, and the vertical axis represents the frequency f. The concentration distribution in the graph represented by the horizontal axis and the vertical axis represents the reference spectrum 300. In the reference spectrum 300, the higher the concentration, the stronger the spectral component.
Further, in the reference spectrum 300, the spectrum components 301 to 306 correspond to the operation sound of the facility 2, respectively. For example, the spectral component 301 corresponds to an operation sound when the movable frame 5 moves horizontally in parallel with the upper surface of the frame 3. The spectral component 302 corresponds to an operation sound when the movable frame 5 is lowered so that the rotary brush 4-1 is brought into contact with the frame 3. Further, the spectral component 303 corresponds to an operation sound when the rotating brush 4-1 sweeps the upper surface of the frame 3. The spectral component 304 corresponds to an operation sound when the movable frame 5 moves horizontally so that the rotary brush 4-2 is brought into contact with the frame 3. Further, the spectral component 305 corresponds to an operation sound when the rotating brush 4-2 sweeps the upper surface of the frame 3. The spectral component 306 corresponds to an operation sound when the movable frame 5 moves up so as to separate the rotary brush 4-2 from the frame 3.
As described above, in the reference spectrum, every time a component of the facility 2 operates, a spectrum component corresponding to a frequency corresponding to the time during which the operation is performed and the type of operation increases.

FIG. 4 is a diagram illustrating an example of a time-frequency spectrum obtained by performing time-frequency conversion on the digital sound signal of the operating sound measured for the facility 2. Also in FIG. 4, the horizontal axis represents the elapsed time t, and the vertical axis represents the frequency f. The concentration distribution in the graph represented by the horizontal axis and the vertical axis represents the measured time frequency spectrum 400. In the time-frequency spectrum 400, the higher the concentration, the stronger the spectral component. It should be noted that the spectra shown in FIGS. 3 and 4 are schematic for facilitating understanding of the invention.
Further, in the time-frequency spectrum 400, the spectrum components 401 to 406 correspond to the operation sound of the equipment 2, respectively. And the operation | movement of the installation 2 with which the spectral components 401-406 respond | correspond is equal to the operation | movement of the installation 2 with which the spectral components 301-306 in FIG. For example, the spectral components 401, 402, 404, and 406 correspond to operation sounds when the movable frame 5 moves, similarly to the spectral components 301, 302, 304, and 306. Similarly to the spectral components 303 and 305, the spectral components 403 and 405 correspond to operation sounds when the rotating brush 4-1 or the rotating brush 4-2 sweeps the upper surface of the frame 3.

Here, in FIG. 4, while the spectral component 403 corresponding to the operation sound when the rotating brush 4-1 sweeps the upper surface of the frame 3 is observed, a strong spectral component 407 having a lower frequency than the spectral component 403 is present. Observed. In the reference spectrum 300 shown in FIG. 3, there is no spectral component corresponding to this spectral component 407. Therefore, it can be seen that the equipment diagnosis apparatus 1 corresponds to the abnormal sound that the spectral component 407 is different from the operation sound due to the normal operation of the equipment 2.
Further, the time when the spectral component 406 occurs overlaps with the time zone when the spectral component 405 occurs. On the other hand, in the reference spectrum 300, the time when the spectral component 306 corresponding to the spectral component 406 occurs is after the time zone when the spectral component 305 corresponding to the spectral component 405 is generated. From this, the equipment diagnosis apparatus 1 can determine that the operation of the movable frame 5 corresponding to the spectrum component 406 is not normal.

  Therefore, the determination unit 23 determines whether the time interval between the spectral components corresponding to any two operation sounds of the plurality of operations of the facility 2 is different from the time interval between the two operations during normal operation. Find out. If the time interval between the spectral components corresponding to the two operation sounds is different from the time interval during normal operation, the determination unit 23 determines that an abnormality has occurred in the equipment 2. For example, the determination unit 23 obtains an absolute value of a difference between a time interval in which spectral components corresponding to two temporally continuous operations of the facility 2 exist and a corresponding two time interval in the reference spectrum. . If the absolute value of the difference is larger than a predetermined ratio with respect to the interval between two corresponding times in the reference spectrum, the determination unit 23 determines that an abnormality has occurred in the facility 2. Note that the predetermined ratio is a value obtained by multiplying the ratio between the maximum value of the fluctuation interval of the time interval and the time interval, which occurs even when the facility 2 is operating normally, by a safety factor of about 1.1, for example, 0.1 It can be.

  Moreover, the determination part 23 determines with abnormality having generate | occur | produced in the equipment 2 also when the spectrum component corresponding to one of the operation sounds is not detected, or when the spectrum component corresponding to the abnormal sound is detected. .

Furthermore, when the determination unit 23 determines that some abnormality has occurred in the facility 2, the determination unit 23 estimates the cause of the abnormality based on the measured time-frequency spectrum.
Accordingly, the determination unit 23 estimates the cause of the abnormality depending on whether or not the spectrum component that has caused the determination that an abnormality has occurred in the facility 2 satisfies an abnormality cause condition corresponding to various predetermined abnormalities.
For example, an abnormal spectrum that is a time-frequency spectrum at the time of occurrence of various abnormalities in the facility 2 assumed in advance is stored in the storage unit 13 in advance. And the determination part 23 calculates the correlation value of the spectrum component used as the cause which determines with the abnormality having generate | occur | produced in the installation 2, and each abnormal spectrum by pattern matching. When the determination unit 23 determines that the correlation value between the spectrum component that caused the determination that an abnormality has occurred and one of the abnormal spectra is higher than the abnormality identification threshold corresponding to the minimum correlation value that can be considered to be related to each other It is determined that the abnormal cause condition is satisfied. Then, the determination unit 23 estimates that an abnormality corresponding to the abnormal spectrum has occurred in the facility 2. For example, if the correlation value between the spectrum component 407 corresponding to the abnormal sound generated during the rotation of the rotating brush 4-1 shown in FIG. 4 and the abnormal spectrum corresponding to the wear of the rotating brush is higher than the abnormality specifying threshold value. The determination unit 23 estimates that the cause of abnormality of the equipment 2 is wear of the rotating brush 4-1.

Further, the abnormality cause condition may be defined by a spectrum intensity value in a specific frequency band and an occurrence time. For example, as one of the abnormal cause conditions, a spectrum corresponding to the operating sound of the movable frame 5 was observed while a spectral component corresponding to the sound of the rotating brush 4-2 sweeping the workpiece 3 was being observed. Is defined as an abnormality of the movable frame 5. In this case, the determination unit 23 is a time during which a spectral component corresponding to the sound of the rotary brush 4-2 sweeping the workpiece 3 is observed and a time during which a spectrum corresponding to the operation sound of the movable frame 5 is observed. If the relationship satisfies the above-described abnormality cause condition, it is estimated that an abnormality has occurred in the movable frame 5.
Further, as another example of the abnormality cause condition relating to the movable frame 5, no spectral component corresponding to the sound of any rotating brush sweeping the workpiece 3 is detected, and the movable frame 5 applies the rotating brush to the workpiece 3. It is specified that a spectral component corresponding to an operation sound moving in the direction of touching is not detected. In this case, the determination unit 23 estimates that an abnormality has occurred in the movable frame 5 when these two spectral components are not detected.

  The determination unit 23 reads a message corresponding to the estimated cause of abnormality from the storage unit 13. Then, the determination unit 23 causes the display unit 15 to display the message. Further, when determining that an abnormality has occurred in the facility 2, the determination unit 23 may transmit a control signal for stopping the facility 2 to the control circuit 6 of the facility 2 via the communication unit 14.

  Hereinafter, facility diagnosis processing by the facility diagnosis apparatus 1 according to one embodiment of the present invention will be described with reference to the flowchart shown in FIG. Note that the equipment diagnosis process shown in FIG. 5 is controlled by the control unit 16.

  As shown in FIG. 5, the facility diagnosis apparatus 1 collects the operation sound of the facility 2 via the microphone 10 (step S101). The collected operation sound is amplified by the amplifier 11 and then converted into a digital audio signal by the analog-digital converter 12. The operation sound that has become a digital audio signal is temporarily stored in the storage unit 13.

  Next, the time frequency analysis unit 21 of the control unit 16 calculates a time frequency spectrum of the operation sound corresponding to one operation cycle by performing time frequency conversion on the operation sound for one operation cycle of the equipment 2 (step S102). ). The time frequency spectrum is sent to the comparison unit 22 of the control unit 16. And the comparison part 22 is included in 1 operation cycle of the installation 2 by comparing a time frequency spectrum with the reference | standard spectrum which is the time frequency spectrum of the operation sound for 1 operation cycle when the installation 2 is operating normally. A spectral component corresponding to each operation is extracted (step S103). Then, the comparison unit 22 notifies the determination unit 23 of the processing unit 14 of the time during which the spectrum component corresponding to each operation included in one operation cycle of the facility 2 exists. In addition, when a spectrum component corresponding to any operation is not detected, the comparison unit 22 notifies the determination unit 23 of a signal indicating the operation that has not been detected. Further, when a spectral component corresponding to the abnormal sound is detected, the comparison unit 22 notifies the determination unit 23 of the frequency band and time in which the spectral component corresponding to the abnormal sound exists.

The determination unit 23 determines whether or not the interval between the spectral components corresponding to each operation of the facility 2 is different from the time interval of the operation during normal operation (step S104). If the interval between the spectral components corresponding to each operation is different from the time interval of the operation during normal operation, the determination unit 23 turns on an operation time abnormality flag that is one of the abnormality flags (step S105).
In step S104, when the interval between the spectral components corresponding to each operation is equal to the time interval of the operation during normal operation, or after step S105, the determination unit 23 is executed during one operation cycle of the equipment 2. It is determined whether or not spectral components corresponding to all operations are detected (step S106). If no spectral component corresponding to any operation is detected, the determination unit 23 turns on the non-operation abnormality flag, which is another one of the abnormality flags (step S107).
On the other hand, in step S106, when spectral components corresponding to all the operations executed during one operation cycle of the facility 2 are detected, or after step S107, the determination unit 23 determines which operation of the facility 2 is performed. It is determined whether or not a spectral component that does not correspond is detected (step S108). And when the spectrum component which does not respond | correspond to any operation | movement of the installation 2 is detected, the determination part 23 turns ON the abnormal sound flag which is another one of an abnormality flag (step S109).

In step S108, when a spectrum component that does not correspond to any operation of the facility 2 is not detected, or after step S109, the determination unit 23 determines whether any abnormality flag is ON ( Step S110). When any abnormality flag is OFF, the determination unit 23 determines that the facility 2 is operating normally. And the equipment diagnostic apparatus 1 complete | finishes equipment diagnostic processing.
On the other hand, if any abnormality flag is ON in step S110, the determination unit 23 estimates the cause of the abnormality based on the spectrum component that causes the determination that the abnormality has occurred (step S111). And the determination part 23 reads the message corresponding to the estimated abnormality cause from the memory | storage part 13, and displays the message on the display part 15 (step S112).
After step S112, the facility diagnosis apparatus 1 ends the facility diagnosis process.

  As described above, the facility diagnosis apparatus according to one embodiment of the present invention collects operation sounds generated during one operation cycle including a plurality of operations that the diagnosis target facility repeatedly executes at a constant cycle. . And this equipment diagnostic apparatus calculates the time frequency spectrum of the operation sound of the 1 operation cycle by carrying out time frequency conversion of the collected sound. The facility diagnostic apparatus compares the calculated time frequency spectrum with the reference time frequency spectrum corresponding to the operation sound of one operation cycle when the diagnosis target facility is operating normally, thereby operating sound of each operation. It is possible to accurately specify the spectral component corresponding to. Therefore, this equipment diagnosis apparatus can specify whether or not the spectrum component corresponding to each operation has been detected at a predetermined timing, or can identify the spectrum component corresponding to an abnormal operation sound that does not exist during normal operation. The presence or absence of abnormalities can be accurately determined. Thus, since this equipment diagnostic apparatus can determine the presence or absence of equipment abnormalities by simply collecting operation sounds for one operation cycle at a time, even for equipment that performs a plurality of different operations, The operation state can be diagnosed between and with high accuracy.

In addition, this invention is not limited to said embodiment. For example, the time frequency analysis unit divides a digital audio signal corresponding to an operation sound of one operation cycle of the equipment at a predetermined time interval, and performs fast Fourier transform or discrete cosine transform on each divided digital audio signal. Thus, a time frequency spectrum may be calculated. In this case, the predetermined time interval is preferably set to be shorter than, for example, a time during which individual operations included in one operation cycle of the facility continue. As a result, the spectral components corresponding to the operations performed at different times appear at different times, similar to the case where the digital sound signal corresponding to the operation sound of one operation cycle of the facility is wavelet transformed. Therefore, the time-frequency analysis unit can distinguish spectral components corresponding to two operations performed at different times.
The time-frequency analysis unit may acquire a digital audio signal corresponding to the operation sound for one operation cycle a plurality of times, and may perform time-frequency conversion on an average of the plurality of digital audio signals. Alternatively, the time frequency analysis unit calculates the time frequency spectrum calculated from the digital audio signal corresponding to the operation sound for one operation cycle for each of the plurality of operation cycles, and compares the averaged time frequency spectrum. You may output to a part. Thereby, the equipment diagnostic apparatus can avoid detecting the sound produced by the single event as the abnormal operation sound of the equipment.

Furthermore, according to another embodiment, the equipment diagnosis apparatus may include a video camera for sequentially photographing each part of the diagnosis target equipment. And the control part of an equipment diagnostic apparatus detects that the equipment was in the reference | standard state before starting the operation | movement included in each operation cycle based on the image of the equipment image | photographed one by one by the video camera. The equipment diagnostic apparatus collects the operation sound of the equipment for the time required to execute one operation cycle via the microphone from the time when it is determined that the equipment is in the reference state. Thereby, the facility diagnostic apparatus can acquire the time frequency spectrum corresponding to the operation sound for one operation cycle accurately, without modifying the existing facility. Therefore, this equipment diagnosis apparatus can diagnose the operating state of the equipment without any modification to the existing equipment.
Note that the control unit recognizes the characteristic shape of each movable part of the facility, for example, by pattern matching, for example, in order to detect that the facility is in the reference state based on the image of the facility. Then, the control unit can determine that the equipment is in the reference state by detecting that the positions of the characteristic shapes of the respective movable units exist at those positions in the reference state.
As described above, those skilled in the art can make various modifications within the scope of the present invention.

DESCRIPTION OF SYMBOLS 1 Equipment diagnostic apparatus 2 Equipment 3 Work 4-1, 4-2 Rotating brush 5 Movable frame 6 Control circuit 7 Belt conveyor 10 Microphone 11 Amplifier 12 Analog-digital converter 13 Memory | storage part 14 Communication part 15 Display part 16 Control part

Claims (6)

A sound collection unit (10) for collecting operation sounds generated by the facility (2);
A memory that stores a reference spectrum that is a time frequency spectrum of operation sound of the facility (2) during a predetermined period when a plurality of operations of the facility (2) are executed when the facility (2) is operating normally. Part (13);
A time-frequency analysis unit (21) that calculates a time-frequency spectrum corresponding to the operation sound of the predetermined period by performing time-frequency conversion on the operation sound of the predetermined period collected by the sound collection unit (10);
A comparison unit (22) for identifying a spectrum component corresponding to each operation sound of the plurality of operations of the facility (2) by comparing the time-frequency spectrum and the reference spectrum;
A determination unit (23) for determining whether an abnormality has occurred in the facility (2) based on the spectral components;
A facility diagnostic apparatus comprising:   The determination unit (23) is configured so that the time interval between the spectrum component of the operation sound corresponding to the first operation of the plurality of operations and the spectrum component of the operation sound corresponding to the second operation is a normal operation time. The equipment diagnosis apparatus according to claim 1, wherein when the time interval between the first operation and the second operation is different from each other, it is determined that an abnormality has occurred in the equipment (2).   The determination unit (23) determines that an abnormality has occurred in the facility (2) when a spectrum component corresponding to any of the plurality of operations is not detected by the comparison unit (22). The equipment diagnostic apparatus according to claim 1 or 2.   The said determination part (23) determines that the abnormality has generate | occur | produced in the said installation (2), when the spectrum component which does not respond | correspond to any of these some operation | movement is detected. The equipment diagnostic apparatus according to one item.   When the determination unit (23) determines that an abnormality has occurred in the facility (2), the spectrum component that is the basis for the determination corresponds to a specific abnormality that has occurred in the facility (2). The equipment diagnosis according to any one of claims 1 to 4, wherein when an abnormality cause condition is satisfied, the cause of the abnormality that has occurred in the equipment (2) is estimated to be a cause associated with the abnormality cause condition. apparatus. Collecting the operation sound generated by the equipment;
Calculating a time-frequency spectrum corresponding to the operation sound of the predetermined period by performing time-frequency conversion of the operation sound of the predetermined period in which a plurality of operations of the collected equipment are executed;
By comparing the time frequency spectrum with a reference spectrum that is a time frequency spectrum of the operation sound of the facility during the predetermined period when the facility is operating normally, each of the plurality of operations of the facility Identifying a spectral component corresponding to the operating sound;
Determining whether an abnormality has occurred in the facility based on each spectral component; and
A facility diagnosis method comprising: