Abnormal noise detection device

Abstract

To provide an abnormal sound detector capable of detecting even slight abnormal sound.SOLUTION: An abnormal sound detector detects abnormal sound contained in operation sound of object equipment being an inspection object. The abnormal sound detector comprises: a sound acquisition part (10) for acquiring operation sound of object equipment, as a sound pressure time signal that is a time signal indicating aged deterioration of the intensity of a sound pressure signal; a signal conversion part (22) for converting the sound pressure time signal to a frequency characteristic indicating the intensity of the sound pressure signal per frequency; an abnormal sound accentuation part (23) for obtaining an abnormal sound accentuation signal which accentuates the intensity of a sound pressure signal of a frequency related to abnormal sound of the object equipment among sound pressure signals per frequency for the frequency characteristic, on the basis of the frequency characteristic; and an abnormal sound determination part (25) for determining whether or not abnormal sound is contained in operation sound of the object equipment on the basis of the abnormal sound accentuation signal.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present disclosure relates to an abnormal sound detection device that detects abnormal sounds from a target device that is an inspection target. [Background technology]

[0002] Conventionally, there is known an abnormality occurrence factor identification system that analyzes a sound emitted from a target device and identifies the cause of an abnormal sound contained in the sound (for example, see Patent Document 1). This abnormality occurrence factor identification system creates divided data by dividing sound pressure data including the abnormal sound at predetermined time intervals, and identifies the cause of the abnormal sound based on the degree of abnormality obtained from the FFT data of each divided data. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-81364 Summary of the Invention [Problem to be solved by the invention]

[0004] However, when the abnormal sound generated by an abnormality in the target device is contained in a small amount of abnormal sound, such as immediately after the abnormality begins to occur, the sound pressure signal of the abnormal sound is small when analyzing the sound pressure data. When the abnormal sound contained in the sound emitted from the target device is small in this way, it is difficult to detect the occurrence of the abnormal sound in an abnormal sound detection device that detects abnormal sounds by analyzing sounds emitted from the target device, such as the abnormality occurrence cause identification system described in Patent Document 1.

[0005] In view of the above, an object of the present disclosure is to provide an abnormal sound detection device capable of detecting even a slight abnormal sound. [Means for solving the problem]

[0006] The invention described in claim 1 is An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires an operating sound of a target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic indicating the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; A storage unit (21) for storing allophone feature amounts; a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal; Equipped with The target device has multiple component devices (M1, M2, C), The abnormal sound emphasis unit obtains the abnormal sound emphasis signal by performing at least an increasing adjustment out of an increasing adjustment for increasing the intensity of a sound pressure signal of a frequency estimated to be included in the abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics and a decreasing adjustment for decreasing the intensity of a sound pressure signal of a frequency different from the frequency estimated to be included in the abnormal sound of the target device. At the same time, an abnormal sound emphasis signal is obtained for each of the plurality of component devices by emphasizing the intensity of a sound pressure signal of a frequency included in an operating sound when each of the plurality of component devices is operating in a normal state, among the sound pressure signals for each frequency in the frequency characteristics; the sound analysis unit determines an allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound determination unit determines whether or not an abnormal sound is included in the operating sound of the target device based on the abnormal sound features for each of the multiple component devices in the past stored in the memory unit and the abnormal sound features for each of the multiple component devices currently determined by the sound analysis unit. The invention described in claim 5 is as follows: An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires an operating sound of a target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic indicating the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; Abnormal noise information, which is information about abnormal noises contained in the operating sounds of the target device and allophone features A storage unit (21) that stores the above in advance, a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal; Equipped with The target device has multiple component devices (M1, M2, C), The abnormal sound information is set based on an operation sound of each of the plurality of component devices when the plurality of component devices are operating, The abnormal sound emphasis unit performs at least one of an increase adjustment for increasing the intensity of a sound pressure signal of a frequency estimated to be included in an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics, and a decrease adjustment for decreasing the intensity of a sound pressure signal of a frequency different from the frequency estimated to be included in the abnormal sound of the target device, and obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency included in the operating sounds of the multiple component devices stored in the storage unit among the sound pressure signals for each frequency in the frequency characteristics. and determining an abnormal sound emphasis signal for each of the plurality of component devices by emphasizing the intensity of a sound pressure signal of a frequency included in an operating sound when each of the plurality of component devices is operating in a normal state, among the sound pressure signals for each frequency in the frequency characteristics; the sound analysis unit determines an allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound detection unit is The abnormal sound feature amounts for each of the plurality of component devices in the past stored in the storage unit and the abnormal sound feature amounts for each of the plurality of component devices currently obtained by the sound analysis unit Based on this, it is determined whether or not the operating sound of the target device includes an abnormal sound. In addition, claims 7 The invention described in An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires an operating sound of a target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic indicating the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; A storage unit (21) for storing allophone feature amounts; a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal; Equipped with The target device has multiple component devices (M1, M2, C), the abnormal sound emphasis unit obtains an abnormal sound emphasis signal for each of the plurality of component devices by emphasizing the intensity of a sound pressure signal of a frequency included in an operating sound when each of the plurality of component devices operates in a normal state, among the sound pressure signals for each frequency in the frequency characteristics; the sound analysis unit determines an allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound determination unit determines whether or not an abnormal sound is included in the operating sound of the target device based on the abnormal sound features for each of the multiple component devices in the past stored in the memory unit and the abnormal sound features for each of the multiple component devices currently determined by the sound analysis unit.

[0007] In this way, by determining whether or not an abnormal sound is included in the operating sound of the target device based on an abnormal sound emphasis signal that emphasizes the intensity of the sound pressure signal of the frequency related to the abnormal sound of the target device, it is possible to detect the abnormal sound even if the abnormal sound is slight.

[0008] Note that the reference symbols in parentheses attached to each component indicate an example of the correspondence between the component and specific components described in the embodiments described below. [Brief description of the drawings]

[0009] [Figure 1] 1 is a schematic diagram illustrating the configuration of an abnormal sound detection device according to a first embodiment. [Diagram 2] FIG. 4 is a diagram illustrating an example of a sound pressure time signal. [Diagram 3] FIG. 2 is a schematic configuration diagram of an EQ section according to the first embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of frequency characteristics. [Diagram 5] FIG. 13 is a diagram showing an example of an abnormal sound template. [Figure 6] FIG. 4 is a diagram showing an example of an abnormal sound emphasis characteristic. [Figure 7] FIG. 4 is a diagram illustrating an example of an emphasis sound signal. [Figure 8] 4A to 4C are diagrams for explaining a control process executed by an EQ unit according to the first embodiment. [Figure 9] FIG. 11 is a schematic diagram illustrating the configuration of an abnormal sound detection device according to a second embodiment. [Figure 10] FIG. 13 is a diagram illustrating an example of an abnormal cutting template. [Figure 11] 13 is a diagram for explaining a control process executed by an EQ unit according to the second embodiment. FIG. [Figure 12] FIG. 4 is a diagram illustrating an example of frequency characteristics. [Figure 13] FIG. 4 is a diagram showing an example of a first motor emphasis characteristic. [Figure 14] FIG. 11 is a diagram showing an example of a second motor emphasis characteristic. [Figure 15] FIG. 13 is a diagram showing an example of a cutting emphasis characteristic. [Figure 16] FIG. 11 is a schematic diagram illustrating the configuration of an abnormal sound detection device according to a third embodiment. [Figure 17] FIG. 11 is a schematic configuration diagram of an EQ section according to a third embodiment. [Figure 18] FIG. 13 is a diagram illustrating a setting unit of an input device according to a third embodiment. [Figure 19] 13 is a diagram for explaining a control process executed by an EQ unit according to the third embodiment. FIG. [Figure 20] FIG. 4 is a diagram showing an example of a first motor emphasis characteristic. [Figure 21] FIG. 11 is a diagram showing an example of a second motor emphasis characteristic. [Figure 22] FIG. 13 is a diagram showing an example of a cutting emphasis characteristic. [Diagram 23] FIG. 13 is a diagram illustrating a setting unit of an input device according to a modified example of the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, the embodiments of the present disclosure will be described with reference to the drawings. In the following embodiments, the same reference numerals are used for parts that are the same as or equivalent to those described in the preceding embodiments, and the description thereof may be omitted. In addition, in the embodiments, when only a part of the components is described, the components described in the preceding embodiments can be applied to the other parts of the components. In the following embodiments, each embodiment can be partially combined with each other, even if not specifically stated, as long as it does not cause any trouble in combination.

[0011] (First embodiment) This embodiment will be described with reference to Figs. 1 to 8. The abnormal sound detection device 1 of this embodiment is a device that determines whether or not an abnormal sound is included in an operation sound of a target device P that is an inspection target, based on the operation sound of the target device P. The target device P is, for example, equipment installed in a factory. An abnormal sound is a sound that is generated when the target device P operates in an abnormal state that is not a normal state. As shown in Fig. 1, the abnormal sound detection device 1 includes a sound acquisition unit 10, an EQ unit 20, a sound output unit 30, and a speaker 50. The abnormal sound detection device 1 acquires the operation sound of the target device P from a microphone MC provided outside the abnormal sound detection device 1.

[0012] The microphone MC is a sound detection unit that detects the operating sound of the target device P that is the target for abnormal sound detection, and outputs a signal corresponding to the detected operating sound to the outside as a sound pressure signal. The microphone MC is installed in the target device P, and transmits the sound pressure signal corresponding to the operating sound of the target device P as an analog signal to the sound acquisition unit 10, for example, at predetermined intervals.

[0013] The sound acquisition unit 10 is an interface unit that receives a sound pressure signal corresponding to an operation sound transmitted from the microphone MC. The sound acquisition unit 10 samples the analog signal transmitted from the microphone MC at a predetermined sampling interval and converts the analog signal into a digital signal. The sound acquisition unit 10 acquires the operation sound generated when the target device P operates as a sound pressure time signal Ts1, which is a time signal indicating a change in the intensity of the sound pressure signal over time. As shown in FIG. 2, the sound pressure time signal Ts1 includes information on the change in the intensity of the sound pressure signal over time. The sound acquisition unit 10 outputs the acquired sound pressure time signal Ts1 to the EQ unit 20.

[0014] The EQ unit 20 is configured as a computer equipped with a CPU and the like. As shown in FIG. 3, the EQ unit 20 includes a storage unit 21, a signal conversion unit 22, an abnormal sound emphasis unit 23, a sound analysis unit 24, an abnormal sound determination unit 25, and a signal inverse conversion unit 26. The storage unit 21 includes a RAM, a ROM, a writable nonvolatile storage medium, and the like. The RAM, the ROM, and the writable nonvolatile storage medium are all non-transient substantial storage media. The storage unit 21 stores in advance programs for the signal conversion unit 22, the abnormal sound emphasis unit 23, the sound analysis unit 24, the abnormal sound determination unit 25, and the signal inverse conversion unit 26 to execute processes, and information on the operation sound of the target device P, which will be described later.

[0015] The signal conversion unit 22, the abnormal sound emphasis unit 23, the sound analysis unit 24, the abnormal sound determination unit 25, and the signal inverse conversion unit 26 are devices that execute processes according to programs recorded in the ROM of the storage unit 21 or in a writable non-volatile storage medium. When executing the programs, the signal conversion unit 22, the abnormal sound emphasis unit 23, the sound analysis unit 24, the abnormal sound determination unit 25, and the signal inverse conversion unit 26 use the RAM of the storage unit 21 as a working area.

[0016] That is, the EQ unit 20 executes the programs stored in the storage unit 21 to function as a signal conversion unit 22, an abnormal sound emphasis unit 23, a sound analysis unit 24, an abnormal sound determination unit 25, and a signal inverse conversion unit 26. The EQ unit 20 may include a plurality of circuit modules in one-to-one correspondence with the storage unit 21, the signal conversion unit 22, the abnormal sound emphasis unit 23, the sound analysis unit 24, the abnormal sound determination unit 25, and the signal inverse conversion unit 26.

[0017] The signal conversion unit 22 converts the sound pressure time signal Ts1 transmitted from the sound acquisition unit 10 into a signal indicating the intensity of the sound pressure signal for each frequency. Specifically, the signal conversion unit 22 extracts only a predetermined length of time from the input sound pressure time signal Ts1 and performs a short-time Fourier transform to convert it into a frequency characteristic Fs1 indicating the intensity of the sound pressure signal for each frequency, as shown in FIG.

[0018] The signal conversion unit 22 of this embodiment converts the sound pressure time signal Ts1 into the intensity of the sound pressure signal for each frequency in the frequency band from 0 Hz to 20,000 Hz, which is generally considered to be the approximate maximum frequency in the audible range. The frequency characteristic Fs1 shown in Fig. 4 shows an example including information on the magnitude of the sound pressure signal of the operation sound of the target device P and the magnitude of the sound pressure signal of an abnormal sound included in the operation sound.

[0019] The abnormal sound emphasis unit 23 emphasizes the intensity of the sound pressure signal of a frequency related to an abnormal sound from the target device P among the sound pressure signals for each frequency in the frequency characteristic Fs1 converted by the signal conversion unit 22. The abnormal sound emphasis unit 23 emphasizes the intensity of the sound pressure signal of a frequency related to an abnormal sound from the target device P that is determined in advance in the storage unit 21 among frequencies from 0 to 20000 Hz in the frequency characteristic Fs1.

[0020] 5, which is related to the actual operation sound when the operation sound of the target device P includes an abnormal sound. Specifically, the storage unit 21 stores in advance information on an abnormal sound template Fs2 in which the operation sound including the actual abnormal sound of the target device P is converted into a sound pressure signal for each frequency of 0 to 20000 Hz in the same manner as the frequency characteristic Fs1, and the sound pressure signal of the frequency included in the abnormal sound is emphasized. In other words, the abnormal sound template Fs2 is based on the operation sound when the target device P is in an abnormal state.

[0021] Furthermore, feature quantities related to the intensity of the sound pressure signal of the abnormal sound template Fs2 are stored in advance in the storage unit 21. Specifically, the storage unit 21 stores in advance information about a template vector Vt obtained by converting the intensity of the sound pressure signal for each frequency indicated by the abnormal sound template Fs2 into a feature vector.

[0022] The storage unit 21 also stores in advance information on frequencies in the 0 to 20000 Hz frequency band in the abnormal sound template Fs2 that have relatively large sound pressure signals compared to other frequencies. In the abnormal sound template Fs2, the frequencies that have relatively large sound pressure signals compared to other frequencies are frequencies that are included in abnormal sounds among the frequencies included in the operating sounds of the target device P.

[0023] Incidentally, the abnormal sound included in the operation sound of the target device P occurs when the target device P operates in an abnormal state, not in a normal state. The abnormal sound generated when the target device P operates in an abnormal state is likely to contain relatively the same frequency each time the target device P changes from a normal state to an abnormal state. For this reason, in the abnormal sound template Fs2 set based on the operation sound when the target device P is in an abnormal state, a frequency with a relatively large sound pressure signal compared to other frequencies is a frequency estimated to be included in the abnormal sound when the target device P changes from a normal state to an abnormal state. The abnormal sound template Fs2 corresponds to abnormal sound information. Also, the template vector Vt corresponds to an estimated feature amount.

[0024] Hereinafter, among the frequency bands in the abnormal sound template Fs2, a frequency whose sound pressure signal is relatively large compared to other frequencies will be referred to as an abnormal sound frequency Fa, and a frequency different from the abnormal sound frequency Fa will be referred to as a normal frequency Fn.

[0025] In this embodiment, for example, the abnormal sound frequency Fa is set to a frequency band having a predetermined bandwidth centered on the frequency at which the sound pressure signal in the abnormal sound template Fs2 is maximized. The predetermined bandwidth is determined in advance and is set to, for example, 100 Hz.

[0026] 5, when the frequency at which the sound pressure signal is maximized is 2400 Hz, the abnormal sound frequency Fa is set to a frequency band having a bandwidth of 2350 Hz to 2450 Hz within the band of 0 to 20000 Hz, and the normal frequency Fn is set to a frequency band of 0 to 20000 Hz that is smaller than 2350 Hz and larger than 2450 Hz.

[0027] The predetermined bandwidth may be set to a bandwidth smaller than 100 Hz or may be set to a bandwidth larger than 100 Hz. The predetermined bandwidth is set appropriately in the abnormal sound template Fs2.

[0028] The information on the abnormal sound template Fs2, template vector Vt, abnormal sound frequency Fa, and normal frequency Fn is obtained from experimental results or the like in which operating sounds, including actual abnormal sounds, are collected when the target device P is operated in an abnormal state and sound analyzed. In the abnormal sound detection device 1, the information on the abnormal sound template Fs2, template vector Vt, abnormal sound frequency Fa, and normal frequency Fn, which are obtained from experimental results or the like performed in advance before abnormal sound detection, is stored in the storage unit 21.

[0029] In the abnormal sound template Fs2 shown in Fig. 5, in order to emphasize the sound pressure signal of the frequency estimated to be included in the abnormal sound, when the sound pressure signal of the actual operation sound containing the abnormal sound is displayed as a sound pressure signal for each frequency, the sound pressure signal of the abnormal sound frequency Fa is increased by a predetermined increase amount. Furthermore, in the abnormal sound template Fs2, in order to emphasize the sound pressure signal of the frequency included in the abnormal sound, when the sound pressure signal of the actual operation sound containing the abnormal sound is displayed as a sound pressure signal for each frequency, the sound pressure signal of the normal frequency Fn is decreased by a predetermined decrease amount. As a result, in the abnormal sound template Fs2 shown in Fig. 5, the sound pressure signal of the frequency included in the abnormal sound is emphasized in the sound pressure signal of the actual operation sound containing the abnormal sound.

[0030] The predetermined increase amount and the predetermined decrease amount for emphasizing the sound pressure signal of the frequency estimated to be included in the abnormal sound are set in advance as values ​​capable of emphasizing the intensity of the sound pressure signal of the abnormal sound frequency Fa among the frequencies in the frequency characteristic Fs1. The absolute values ​​of the predetermined increase amount and the predetermined decrease amount may be set to the same value or may be set to different values.

[0031] The abnormal sound emphasis unit 23 performs an equalizing process to emphasize the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device P when the abnormal sound detection device 1 determines whether or not an abnormal sound is included in the operating sound of the target device P. Specifically, the abnormal sound emphasis unit 23 emphasizes the intensity of a sound pressure signal of an abnormal sound frequency Fa predefined in the storage unit 21 among the sound pressure signals for each frequency of 0 to 20000 Hz in the frequency characteristic Fs1 determined by the signal conversion unit 22. The abnormal sound emphasis unit 23 of this embodiment performs an increase adjustment to increase the sound pressure signal of the abnormal sound frequency Fa among the sound pressure signals in the frequency characteristic Fs1 by a predetermined increase amount, and performs a decrease adjustment to decrease the sound pressure signal of the normal frequency Fn by a predetermined decrease amount.

[0032] That is, the abnormal sound emphasizing unit 23 increases the sound pressure signal of the abnormal sound frequency Fa in the frequency characteristic Fs1 by a predetermined increase amount that is the same as the sound pressure signal increased in the abnormal sound frequency Fa of the abnormal sound template Fs2. Also, the abnormal sound emphasizing unit 23 decreases the sound pressure signal of the normal frequency Fn in the frequency characteristic Fs1 by a predetermined decrease amount that is the same as the sound pressure signal decreased in the abnormal sound template Fs2. By performing equalization in this manner by the abnormal sound emphasizing unit 23, the frequency characteristic Fs1 converted by the signal conversion unit 22 has the intensity of the sound pressure signal of the frequency related to the abnormal sound of the target device P emphasized, as shown in FIG. 6. Also, the waveform of the sound pressure signal obtained by equalizing the frequency characteristic Fs1 converted by the signal conversion unit 22 can be made to approach the waveform of the abnormal sound template Fs2.

[0033] Hereinafter, the signal in which the intensity of the sound pressure signal of the frequency related to the abnormal sound of the target device P is emphasized by the abnormal sound emphasis unit 23 will be referred to as the abnormal sound emphasis characteristic Fs3. The abnormal sound emphasis characteristic Fs3 corresponds to the abnormal sound emphasis signal.

[0034] The allophone emphasis unit 23 outputs information on the allophone emphasis characteristic Fs3 to the sound analysis unit 24 and the signal inverse conversion unit 26.

[0035] The sound analysis unit 24 calculates an allophone feature amount related to the intensity of a sound pressure signal in the allophone emphasis characteristic Fs3 based on the allophone emphasis characteristic Fs3 input from the allophone emphasis unit 23. In this embodiment, the sound analysis unit 24 obtains the allophone feature amount as an emphasis vector Vp obtained by converting the intensity of a sound pressure signal for each frequency indicated by the allophone emphasis characteristic Fs3 into a feature vector. The sound analysis unit 24 outputs information on the calculated emphasis vector Vp to the allophone determination unit 25.

[0036] The abnormal sound determination unit 25 determines whether or not an abnormal sound is included in the operation sound of the target device P based on information on the emphasis vector Vp input from the sound analysis unit 24 and the template vector Vt pre-stored in the storage unit 21. Specifically, the abnormal sound determination unit 25 finds the similarity between the abnormal sound emphasis characteristic Fs3 and the abnormal sound template Fs2 by comparing the emphasis vector Vp, which is an abnormal sound feature amount of the abnormal sound emphasis characteristic Fs3, with the template vector Vt, which is a feature amount of the abnormal sound template Fs2. Then, the abnormal sound determination unit 25 determines whether or not an abnormal sound is included in the operation sound of the target device P based on information on the similarity between the abnormal sound emphasis characteristic Fs3 and the abnormal sound template Fs2. The abnormal sound determination unit 25 will be described in detail in the description of the operation below.

[0037] The signal inverse conversion unit 26 converts the abnormal sound emphasis characteristic Fs3 input from the abnormal sound emphasis unit 23 into a signal indicating a change in the intensity of the sound pressure signal over time and outputs the signal. Specifically, the signal inverse conversion unit 26 performs an inverse short-time Fourier transform on the abnormal sound emphasis characteristic Fs3 based on the frequency characteristic Fs1 obtained by the Fourier transform, and then converts the abnormal sound emphasis characteristic Fs3 into an emphasis sound signal Ts2, which is a time signal indicating a change in the intensity of the sound pressure signal over time. As shown in FIG. 7, the emphasis sound signal Ts2 contains information on the change in the intensity of the sound pressure signal over time. In addition, the emphasis sound signal Ts2 emphasizes the sound pressure signal of the frequency included in the abnormal sound with respect to the operation sound of the target device P. The signal inverse conversion unit 26 outputs the converted emphasis sound signal Ts2 to the sound output unit 30.

[0038] The sound output unit 30 is a signal output unit that converts the emphasized sound signal Ts2 input from the signal inverse conversion unit 26 into, for example, an analog signal and outputs it. The sound output unit 30 outputs to the speaker 50 a signal corresponding to a sound pressure signal in which a sound pressure signal of a frequency included in an abnormal sound is emphasized with respect to the operating sound of the target device P.

[0039] The speaker 50 is a sound generating unit that outputs a sound that emphasizes an abnormal sound relative to the operating sound of the target device P, based on a signal input from the sound output unit 30. The speaker 50 is installed, for example, in the vicinity of an operator who operates the abnormal sound detection device 1, and issues a sound that emphasizes an abnormal sound relative to the operating sound of the target device P to alert the operator.

[0040] The operation of the abnormal sound detection device 1 configured as above will be described below with reference to Fig. 8. When the sound acquisition unit 10 receives a signal corresponding to the operating sound of the target device P from the microphone MC, the sound acquisition unit 10 outputs the acquired sound pressure time signal Ts1 to the signal conversion unit 22.

[0041] When the signal conversion unit 22 receives the sound pressure time signal Ts1 from the sound acquisition unit 10, the signal conversion unit 22 converts the sound pressure time signal Ts1 into the frequency characteristic Fs1 shown in Fig. 4 by performing a short-time Fourier transform on the sound pressure time signal Ts1. The signal conversion unit 22 outputs the converted frequency characteristic Fs1 to the allophone emphasis unit 23.

[0042] When the abnormal sound emphasis unit 23 receives the frequency characteristic Fs1 converted by the signal conversion unit 22, it reads information on the abnormal sound frequency Fa and the normal frequency Fn in the abnormal sound template Fs2 stored in the storage unit 21. Then, the abnormal sound emphasis unit 23 determines the abnormal sound emphasis characteristic Fs3 based on the frequency characteristic Fs1, the abnormal sound frequency Fa, and the normal frequency Fn.

[0043] Specifically, the abnormal sound emphasis unit 23 performs an increase adjustment to increase the intensity of the sound pressure signal of a frequency of 2350 Hz to 2450 Hz by a predetermined increase amount in the frequency characteristic Fs1 converted by the signal conversion unit 22. Furthermore, the abnormal sound emphasis unit 23 performs a decrease adjustment to decrease the intensity of the sound pressure signal of a frequency lower than 2350 Hz and the sound pressure signal of a frequency higher than 2450 Hz by a predetermined decrease amount. As a result, the abnormal sound emphasis unit 23 obtains the abnormal sound emphasis characteristic Fs3 shown in FIG. 5 in which the intensity of the sound pressure signal in the frequency band related to the abnormal sound of the target device P is emphasized with respect to the frequency characteristic Fs1 converted by the signal conversion unit 22. The abnormal sound emphasis unit 23 outputs information on the abnormal sound emphasis characteristic Fs3 to the sound analysis unit 24 and the signal inverse conversion unit 26.

[0044] When the information on the abnormal sound emphasis characteristic Fs3 is input from the abnormal sound emphasis unit 23, the sound analysis unit 24 calculates an emphasis vector Vp based on the abnormal sound emphasis characteristic Fs3. The sound analysis unit 24 outputs information on the calculated emphasis vector Vp to the abnormal sound determination unit 25.

[0045] When information on the emphasis vector Vp is input from the sound analysis unit 24, the abnormal sound determination unit 25 reads information on the template vector Vt stored in the storage unit 21. Then, the abnormal sound determination unit 25 finds the similarity between the abnormal sound emphasis characteristic Fs3 and the abnormal sound template Fs2 by comparing the emphasis vector Vp with the template vector Vt. In this embodiment, the abnormal sound determination unit 25 calculates the similarity between the abnormal sound emphasis characteristic Fs3 and the abnormal sound template Fs2 as the cosine similarity of the emphasis vector Vp with respect to the template vector Vt. Then, the abnormal sound determination unit 25 determines whether or not the operating sound of the target device P includes an abnormal sound based on whether or not the calculated cosine similarity is equal to or greater than a predetermined determination threshold value.

[0046] The cosine similarity is a value between 0 and 1, and the higher the similarity of the emphasis vector Vp to the template vector Vt, the higher the value. In other words, the higher the similarity between the abnormal sound template Fs2 and the abnormal sound emphasis characteristic Fs3, the higher the cosine similarity. The predetermined determination threshold is a reference value determined in advance by experiments or the like in order to determine whether or not the operating sound of the target device P includes an abnormal sound, and is set to, for example, 0.2.

[0047] When the cosine similarity value is 0.2 or greater, the abnormal sound determination unit 25 determines that an abnormal sound is included in the operation sound of the target device P. On the other hand, when the cosine similarity value is not determined to be 0.2 or greater, the abnormal sound determination unit 25 determines that an abnormal sound is not included in the operation sound of the target device P. The abnormal sound determination unit 25 outputs the determination result as to whether or not an abnormal sound is included in the operation sound of the target device P to the signal inverse conversion unit 26.

[0048] The predetermined determination threshold may be a value smaller than 0.2 or may be a value larger than 0.2, so long as the value is capable of determining whether or not the operating sound of the target device P includes an abnormal sound.

[0049] Incidentally, the cosine similarity between the template vector Vt calculated based on the abnormal sound template Fs2 shown in Fig. 5 and the emphasis vector Vp calculated based on the abnormal sound emphasis characteristic Fs3 shown in Fig. 6 was approximately 0.229. In other words, the similarity between the abnormal sound template Fs2 and the abnormal sound emphasis characteristic Fs3 was approximately 0.229.

[0050] In contrast, the cosine similarity of the vector amount calculated in the same manner as the emphasis vector Vp as the feature amount of the frequency characteristic Fs1 shown in Fig. 4 for the template vector Vt calculated based on the abnormal sound template Fs2 shown in Fig. 5 was approximately 0.147. In other words, the similarity between the abnormal sound template Fs2 and the frequency characteristic Fs1 was approximately 0.147.

[0051] In this way, when analyzing a sound pressure signal when an abnormal sound is included in the operating sound of the target device P, the degree of similarity to each sound pressure signal is significantly different between when the intensity of the sound pressure signal of the frequency related to the abnormal sound is emphasized and when it is not emphasized. The degree of similarity can be increased by emphasizing the intensity of the sound pressure signal of the frequency related to the abnormal sound compared to when the intensity of the sound pressure signal of the frequency related to the abnormal sound is not emphasized. Therefore, it is possible to easily determine whether or not an abnormal sound is included in the operating sound of the target device P by using a predetermined determination threshold value.

[0052] When the signal inverse conversion unit 26 receives from the abnormal sound determination unit 25 a determination result that the operation sound of the target device P contains an abnormal sound, the signal inverse conversion unit 26 performs an inverse short-time Fourier transform on the abnormal sound emphasis characteristic Fs3 input from the abnormal sound emphasis unit 23 to convert it into an emphasis sound signal Ts2 shown in Fig. 7. The signal inverse conversion unit 26 then outputs the converted emphasis sound signal Ts2 to the sound output unit 30. The sound output unit 30 outputs a signal corresponding to the emphasis sound signal Ts2 input from the signal inverse conversion unit 26 to the speaker 50. The speaker 50 then issues a sound that emphasizes the abnormal sound in relation to the operation sound of the target device P.

[0053] When the signal inverse conversion unit 26 receives from the abnormal sound determination unit 25 a determination result that the operation sound of the target device P does not contain any abnormal sound, it does not perform an inverse short-time Fourier transform on the abnormal sound emphasis characteristic Fs3 input from the abnormal sound emphasis unit 23, and does not output the emphasis sound signal Ts2 to the sound output unit 30. For this reason, the speaker 50 does not emit a sound that emphasizes an abnormal sound in relation to the operation sound of the target device P.

[0054] As described above, the abnormal sound detection device 1 obtains an abnormal sound emphasis characteristic Fs3 that emphasizes the intensity of the sound pressure signal of the frequency related to the abnormal sound of the target device P for the operating sound of the target device P, and determines whether or not the operating sound of the target device P contains an abnormal sound based on the abnormal sound emphasis characteristic Fs3.

[0055] In this way, by determining whether or not the operating sound of the target device P contains abnormal noise based on the abnormal noise emphasis characteristic Fs3, which emphasizes the intensity of the sound pressure signal of the frequency related to the abnormal noise of the target device P, it is possible to easily detect the abnormal noise even if the abnormal noise contained in the operating sound of the target device P is slight.

[0056] Furthermore, according to the above embodiment, the following effects can be obtained.

[0057] (1) In the above embodiment, the abnormal sound emphasis unit 23 performs an increasing adjustment on the sound pressure signal for each frequency in the frequency characteristic Fs1 to increase the sound pressure signal of the abnormal sound frequency Fa, and performs a decreasing adjustment on the sound pressure signal of the normal frequency Fn.

[0058] This makes it possible to easily emphasize the intensity of the sound pressure signal of the frequency related to abnormal sounds with respect to the operation sounds of the target device P. Also, compared to a case where either an increase adjustment or a decrease adjustment is made to the sound pressure signal in the frequency characteristic Fs1, the intensity of the sound pressure signal of the frequency related to abnormal sounds with respect to the operation sounds of the target device P can be more emphasized.

[0059] (2) In the above embodiment, the abnormal sound template Fs2, which is information about abnormal sounds contained in the operation sound of the target device P, is determined in advance in the storage unit 21.

[0060] According to this, by comparing the abnormal sound emphasis characteristic Fs3 with the abnormal sound template Fs2, which is information regarding abnormal sounds contained in the operating sound of the target device P, it is possible to easily determine whether or not the operating sound of the target device P contains abnormal sounds.

[0061] (3) In the above embodiment, the abnormal sound template Fs2 is set based on the operating sound of the target device P operating in an abnormal state.

[0062] According to this, it is possible to determine whether or not the operating sound of the target device P contains an abnormal sound based on the abnormal sound template Fs2 that is set based on the operating sound of the target device P actually operating in an abnormal state, thereby improving the accuracy of the determination.

[0063] (4) In the above embodiment, the sound analysis unit 24 calculates an emphasis vector Vp, which is an abnormal sound feature amount related to the intensity of a sound pressure signal in the abnormal sound emphasis characteristic Fs3. The storage unit 21 also stores a template vector Vt, which is a feature amount related to the intensity of a sound pressure signal in the abnormal sound template Fs2. The abnormal sound determination unit 25 then determines whether or not the operating sound of the target device P includes an abnormal sound, based on the emphasis vector Vp calculated by the sound analysis unit 24 and the template vector Vt stored in the storage unit 21.

[0064] According to this, it is possible to determine whether or not the operating sound of the target device P contains an abnormal sound based on the respective feature amounts of the abnormal sound emphasis characteristic Fs3 and the abnormal sound template Fs2, making it possible to make this determination easily.

[0065] (5) In the above embodiment, the abnormal sound detection device 1 includes a signal inverse conversion unit 26 that converts the abnormal sound emphasis characteristic Fs3 determined by the abnormal sound emphasis unit 23 into an emphasis sound signal Ts2 that indicates a change over time in the intensity of the sound pressure signal, and outputs the emphasis sound signal Ts2. The abnormal sound detection device 1 also includes a speaker 50 that generates a sound that emphasizes an abnormal sound relative to the operating sound of the target device P, based on the emphasis sound signal Ts2.

[0066] According to this, a sound in which the abnormal sound is emphasized with respect to the operating sound of the target device P can be issued to, for example, an operator who operates the abnormal sound detection device 1.

[0067] (First modified example of the first embodiment) In the above-described first embodiment, an example has been described in which the abnormal sound emphasis unit 23 performs an increasing adjustment and a decreasing adjustment when emphasizing the intensity of the sound pressure signal in the frequency band related to the abnormal sound of the target device P with respect to the frequency characteristic Fs1, but this is not limiting. For example, the abnormal sound emphasis unit 23 may perform only one of the increasing adjustment and the decreasing adjustment to emphasize the intensity of the sound pressure signal in the frequency band related to the abnormal sound of the target device P with respect to the frequency characteristic Fs1.

[0068] (Second Modification of the First Embodiment) In the above-described first embodiment, an example has been described in which, when the abnormal sound determination unit 25 determines that the operation sound of the target device P includes an abnormal sound, the abnormal sound detection device 1 causes the speaker 50 to emit a sound that emphasizes the abnormal sound from the operation sound of the target device P, but this is not limiting. For example, the abnormal sound detection device 1 may be configured to cause the speaker 50 to emit the operation sound of the target device P regardless of the process performed by the abnormal sound determination unit 25 to determine whether or not the operation sound of the target device P includes an abnormal sound.

[0069] Second embodiment Next, the second embodiment will be described with reference to Figs. 9 to 15. In this embodiment, the information on abnormal sounds contained in the operation sound of the target device P that is stored in advance in the storage unit 21 is different from that in the first embodiment. In this embodiment, the processes performed by the abnormal sound emphasis unit 23, the sound analysis unit 24, and the abnormal sound determination unit 25 are also different from those in the first embodiment. In this embodiment, the differences from the first embodiment will be mainly described, and descriptions of the same parts as in the first embodiment may be omitted.

[0070] The target equipment P in this embodiment is, for example, a cutting machine that performs cutting processing on a workpiece (not shown). As shown in Fig. 9, the target equipment P has a tool T that cuts the workpiece as a component for performing cutting processing. The target equipment P also has a first motor M1 for operating the tool T and a second motor M2 for moving the workpiece as components for performing cutting processing. In this way, the target equipment P has a plurality of component devices.

[0071] In the present embodiment, the target equipment P simultaneously generates a first motor sound, which is the operating sound of the first motor M1, a second motor sound, which is the operating sound of the second motor M2, and a cutting sound of the tool T cutting the workpiece when performing cutting processing. The abnormal sound detection device 1 determines whether or not an abnormal sound is included in the operating sounds including the first motor sound, the second motor sound, and the cutting sound, which are simultaneously generated when the cutting machine, which is the target equipment P, performs cutting processing. Hereinafter, the operating sounds including the first motor sound, the second motor sound, and the cutting sound generated when the target equipment P operates may also be simply referred to as the operating sound of the target equipment P.

[0072] The storage unit 21 of this embodiment stores information on the operation sounds of the first motor M1 and the second motor M2 and information on the cutting sound of the tool T as information on abnormal sounds included in the operation sounds of the target device P. The storage unit 21 stores information on the frequencies included in the first motor sound when the first motor M1 operates in a normal state and the second motor sound when the second motor M2 operates in a normal state, independently of each other. The storage unit 21 also stores information on the frequencies included in the cutting sound when the target device P normally performs cutting processing, independently of information on the frequencies included in the first motor sound in a normal state and the second motor sound in a normal state. In other words, the storage unit 21 stores information on the frequencies included in the operation sounds when the first motor M1, the second motor M2, and the tool T operate normally for each of the constituent devices.

[0073] Specifically, the storage unit 21 stores information on frequencies at which sound pressure signals are relatively large compared to other frequencies for each of the operation sounds of the first motor M1, the second motor M2, and the normal cutting sound of the tool T. The storage unit 21 of this embodiment indicates the magnitude of the sound pressure signal for each frequency of 0 to 20,000 Hz for each of these component devices, and stores information on whether the sound pressure signal is relatively large compared to other frequency bands for each band when the 0 to 20,000 Hz range is divided into a plurality of frequency bands.

[0074] More specifically, the storage unit 21 stores information on bands in which the sound pressure signals are relatively large compared to other bands when the frequency band of 0 to 20000 Hz is divided into a first frequency band R1 to an eighth frequency band R8 having equal bandwidths of 2500 Hz for each of these constituent devices. The information on the sound pressure signals in the first frequency band R1 to the eighth frequency band R8 for each of these constituent devices is used to obtain a cutting abnormality template Fs4 shown in Fig. 10 described later and when the abnormal sound emphasis unit 23 executes a control process described later.

[0075] In this embodiment, the normal sound of the first motor M1 is such that, among the first frequency band R1 to the eighth frequency band R8, the sound pressure signals in the first frequency band R1 and the second frequency band R2 are relatively large compared to the sound pressure signals in the other bands. Also, the normal sound of the second motor M2 is such that, among the first frequency band R1 to the eighth frequency band R8, the sound pressure signals in the fifth frequency band R5, the sixth frequency band R6, the seventh frequency band R7, and the eighth frequency band R8 are relatively large compared to the sound pressure signals in the other bands. And, the normal cutting sound of the tool T is such that, among the first frequency band R1 to the eighth frequency band R8, the sound pressure signals in the fifth frequency band R5, the sixth frequency band R6, the seventh frequency band R7, and the eighth frequency band R8 are relatively large compared to the sound pressure signals in the other bands.

[0076] Hereinafter, among the frequencies included in the normal sound of the first motor M1 in the first frequency band R1 to the eighth frequency band R8, the band with a relatively large sound pressure signal is also referred to as the first motor band RM1. Also, among the frequencies included in the normal sound of the second motor M2 in the first frequency band R1 to the eighth frequency band R8, the band with a relatively large sound pressure signal is also referred to as the second motor band RM2. And, among the frequencies included in the normal cutting sound of the tool T in the first frequency band R1 to the eighth frequency band R8, the band with a relatively large sound pressure signal is also referred to as the cutting band RC. Information on the first motor band RM1, the second motor band RM2, and the cutting band RC is stored in advance in the storage unit 21.

[0077] Incidentally, when an abnormal sound included in the operation sound of the target device P is caused by a constituent device, the abnormal sound of the constituent device occurs when the constituent device operates in an abnormal state. The abnormal sound generated when the constituent device operates in an abnormal state is highly likely to include a frequency included in the operation sound generated when the constituent device operates in a normal state. For this reason, the frequency with a relatively large sound pressure signal in the operation sound when the first motor M1, the second motor M2, and the tool T, which are the constituent devices of the target device P, are each in a normal state, is a frequency estimated to be included in the abnormal sound when the target device P goes from a normal state to an abnormal state. For this reason, the first motor band RM1, the second motor band RM2, and the cutting band RC are information on the frequency related to the abnormal sound included in the operation sound of the target device P in this embodiment.

[0078] The first motor band RM1 is obtained from the results of an experiment in which the actual first motor sound is collected and analyzed when the first motor M1 is operated independently in a normal state. The second motor band RM2 is obtained from the results of an experiment in which the actual second motor sound is collected and analyzed when the second motor M2 is operated independently in a normal state. The cutting band RC is obtained from the results of an experiment in which the actual cutting sound is collected and analyzed when cutting is normally performed by the tool T.

[0079] In addition, the memory unit 21 of this embodiment stores a cutting abnormality template Fs4 shown in Figure 10 which relates to operating sounds that include all of the abnormal sounds of the first motor sound, the second motor sound, and the cutting sound.

[0080] In the cutting abnormality template Fs4, the sound pressure signal of the operation sound of the target device P, which actually includes the first motor sound, the second motor sound, and the cutting sound, is converted into a sound pressure signal for each frequency, and the sound pressure signals of the first motor band RM1, the second motor band RM2, and the cutting band RC are emphasized. Specifically, in the cutting abnormality template Fs4 shown in Fig. 10, the sound pressure signals of the first motor band RM1, the second motor band RM2, and the cutting band RC are increased by a predetermined increase amount, similar to the case of obtaining the abnormal sound template Fs2 in the first embodiment.

[0081] In this embodiment, all of the first frequency band R1 to the eighth frequency band R8 are included by spanning the first motor band RM1, the second motor band RM2, and the cutting band RC. Therefore, in the abnormal cutting template Fs4 shown in Fig. 10, the sound pressure signals for all frequencies from 0 to 20000 Hz are increased by a predetermined increment.

[0082] If it is not possible to span all of the first frequency band R1 to the eighth frequency band R8 using either the first motor band RM1, the second motor band RM2, or the cutting band RC, there is no need to increase the sound pressure signal of the band that cannot be spanned.

[0083] Furthermore, the storage unit 21 prestores a cutting template vector Vtc obtained by converting the sound pressure signal intensity for each frequency indicated by the cutting abnormality template Fs4 into a feature vector as a feature amount related to the sound pressure signal intensity of the cutting abnormality template Fs4. The cutting abnormality template Fs4 is information related to abnormal sounds contained in the operating sound of the target device P, and corresponds to abnormal sound information. Moreover, the cutting template vector Vtc corresponds to an estimated feature amount.

[0084] The cutting abnormality template Fs4 and the cutting template vector Vtc are obtained from experimental results, etc., in which actual operation sounds are collected and analyzed when the first motor M1 and the second motor M2 are in an abnormal state and cutting is not performed normally by the tool T. In addition, the cutting abnormality template Fs4 is based on the first motor sound, the second motor sound, and the cutting sound, which are operation sounds generated when the first motor M1, the second motor M2, and the tool T, which are components of the target device P, respectively, operate in an abnormal state.

[0085] The operation of the abnormal sound detection device 1 of this embodiment configured as described above will be described below with reference to Fig. 11. Note that in the description of the operation of this embodiment using Fig. 11, the description of the same control processes as those in the operation of the first embodiment described using Fig. 8 may be omitted.

[0086] When the signal conversion unit 22 receives a time signal indicating a change over time in the intensity of the sound pressure signal from the sound acquisition unit 10, the signal conversion unit 22 converts the signal into a frequency characteristic Fs5 shown in Fig. 12 by performing a short-time Fourier transform. The signal conversion unit 22 outputs the converted frequency characteristic Fs5 to the allophone emphasis unit 23.

[0087] When the abnormal sound emphasis unit 23 receives the frequency characteristic Fs5 converted by the signal conversion unit 22, it reads information on the first motor band RM1, the second motor band RM2, and the cutting band RC stored in the storage unit 21. Then, based on the frequency characteristic Fs5, the first motor band RM1, the second motor band RM2, and the cutting band RC, the abnormal sound emphasis unit 23 emphasizes the intensity of the sound pressure signal of frequencies related to abnormal sounds among the frequencies in the frequency characteristic Fs5.

[0088] Specifically, the abnormal sound emphasis unit 23 performs an increase adjustment on the frequency characteristic Fs5 converted by the signal conversion unit 22, by increasing the intensity of the sound pressure signal in each of the first motor band RM1, the second motor band RM2, and the cutting band RC by a predetermined increase amount individually.

[0089] That is, the abnormal sound emphasis unit 23 obtains a first motor emphasis characteristic Fm1 shown in Fig. 13, which emphasizes the abnormal sound of the first motor M1 by increasing the intensity of the sound pressure signal in the first motor band RM1 with respect to the frequency characteristic Fs5. The first motor emphasis characteristic Fm1 is obtained by increasing the intensity of the sound pressure signal in the first motor band RM1 with respect to the frequency characteristic Fs5 shown in Fig. 12.

[0090] The abnormal sound emphasis unit 23 obtains a second motor emphasis characteristic Fm2 shown in Fig. 14, which emphasizes the abnormal sound of the second motor M2 by increasing the intensity of the sound pressure signal in the second motor band RM2 with respect to the frequency characteristic Fs5. The second motor emphasis characteristic Fm2 is obtained by increasing the intensity of the sound pressure signal in the second motor band RM2 with respect to the frequency characteristic Fs5 shown in Fig. 12.

[0091] Then, the abnormal sound emphasizing unit 23 obtains a cutting emphasis characteristic Fc shown in Fig. 15 in which the intensity of the sound pressure signal in the cutting band RC is increased with respect to the frequency characteristic Fs5 to emphasize the abnormal cutting sound of the tool T. The cutting emphasis characteristic Fc is obtained by increasing the intensity of the sound pressure signal in the cutting band RC with respect to the frequency characteristic Fs5 shown in Fig. 12.

[0092] The abnormal sound emphasis unit 23 outputs information on the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc to the sound analysis unit 24 and the signal inverse conversion unit 26. Note that the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc correspond to the abnormal sound emphasis signal.

[0093] When information on the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc is input from the abnormal sound emphasis unit 23, the sound analysis unit 24 determines the abnormal sound feature quantities of each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc.

[0094] Specifically, the sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the first motor emphasis characteristic Fm1 as a first motor vector Vm1 obtained by converting the intensity of the sound pressure signal for each frequency indicated by the first motor emphasis characteristic Fm1 into a feature vector. The sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the second motor emphasis characteristic Fm2 as a second motor vector Vm2 obtained by converting the intensity of the sound pressure signal for each frequency indicated by the second motor emphasis characteristic Fm2 into a feature vector. The sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the cutting emphasis characteristic Fc as a cutting vector Vc obtained by converting the intensity of the sound pressure signal for each frequency indicated by the cutting emphasis characteristic Fc into a feature vector.

[0095] The sound analysis unit 24 outputs to the abnormal sound determination unit 25 information on the calculated first motor vector Vm1, second motor vector Vm2, and cutting vector Vc.

[0096] When information on the first motor vector Vm1, the second motor vector Vm2, and the cutting vector Vc is input from the sound analysis unit 24, the abnormal sound judgment unit 25 reads information on the cutting template vector Vtc stored in the storage unit 21. Then, the abnormal sound judgment unit 25 compares each of the first motor vector Vm1, the second motor vector Vm2, and the cutting vector Vc with the template vector Vt. As a result, the abnormal sound judgment unit 25 determines the similarity between each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc and the abnormal cutting template Fs4.

[0097] Specifically, the abnormal sound determination unit 25 calculates the similarity between the first motor emphasis characteristic Fm1 and the cutting abnormality template Fs4 as the cosine similarity between the first motor vector Vm1 and the cutting template vector Vtc. The abnormal sound determination unit 25 also calculates the similarity between the second motor emphasis characteristic Fm2 and the cutting abnormality template Fs4 as the cosine similarity between the second motor vector Vm2 and the cutting template vector Vtc. Furthermore, the abnormal sound determination unit 25 calculates the similarity between the cutting emphasis characteristic Fc and the cutting abnormality template Fs4 as the cosine similarity between the cutting vector Vc and the cutting template vector Vtc.

[0098] Then, the abnormal sound determination unit 25 determines whether or not at least one of the calculated cosine similarities corresponding to the first motor vector Vm1, the second motor vector Vm2, and the cutting vector Vc has a value of 0.2 or more.

[0099] When at least one of the cosine similarities corresponding to the first motor vector Vm1, the second motor vector Vm2, and the cutting vector Vc is equal to or greater than 0.2, the abnormal sound determination unit 25 determines that an abnormal sound is included in the operating sound of the target device P. Then, the abnormal sound determination unit 25 determines that the component device corresponding to the vector quantity having a calculated cosine similarity value of 0.2 or greater corresponding to the first motor vector Vm1, the second motor vector Vm2, and the cutting vector Vc is the cause of the abnormal sound.

[0100] In contrast, if the abnormal sound detection unit 25 does not determine that all of the cosine similarities corresponding to the calculated first motor vector Vm1, second motor vector Vm2, and cutting vector Vc are 0.2 or greater, it determines that the operating sound of the target equipment P does not include abnormal sounds.

[0101] For example, the cosine similarity of the first motor vector Vm1 calculated based on the first motor emphasis characteristic Fm1 shown in Fig. 13 to the cutting template vector Vtc calculated based on the cutting abnormality template Fs4 shown in Fig. 10 was about 0.10. Also, the cosine similarity of the second motor vector Vm2 calculated based on the second motor emphasis characteristic Fm2 shown in Fig. 14 to the cutting template vector Vtc calculated based on the cutting abnormality template Fs4 shown in Fig. 10 was about 0.25. And, the cosine similarity of the cutting vector Vc calculated based on the cutting emphasis characteristic Fc shown in Fig. 15 to the cutting template vector Vtc calculated based on the cutting abnormality template Fs4 shown in Fig. 10 was about 0.15.

[0102] In this case, the abnormal sound determination unit 25 determines that an abnormal sound is included in the operating sound of the target device P. Furthermore, the abnormal sound determination unit 25 determines that the cause of the abnormal sound is the second motor M2 among the first motor M1, the second motor M2, and the tool T, which are the components of the target device P. When the abnormal sound determination unit 25 determines that an abnormal sound is included in the operating sound of the target device P, it outputs information on the component components that are the cause of the abnormal sound to the signal inverse conversion unit 26.

[0103] When the signal inverse conversion unit 26 receives information on the component that is the cause of the abnormal sound from the abnormal sound determination unit 25, the signal inverse conversion unit 26 performs an inverse short-time Fourier transform on the characteristics corresponding to the component determined to be the cause of the abnormal sound from among the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc. As a result, the characteristics corresponding to the component determined to be the cause of the abnormal sound are converted into a time signal indicating the change over time in the intensity of the sound pressure signal. Then, the signal inverse conversion unit 26 outputs the time signal obtained by the conversion to the sound output unit 30. The sound output unit 30 outputs a signal corresponding to the time signal input from the signal inverse conversion unit 26 to the speaker 50. Then, the speaker 50 issues an operating sound of the target device P in which the operating sound of the component determined to be the cause of the abnormal sound is emphasized among the components of the target device P.

[0104] As described above, in the target device P having the first motor M1, the second motor M2, and the tool T, the cutting abnormality template Fs4 is set based on the operation sounds of the first motor M1, the second motor M2, and the tool T. In addition, the abnormal sound emphasis unit 23 individually emphasizes the intensities of the sound pressure signals of the first motor band RM1, the second motor band RM2, and the cutting band RC stored in the storage unit 21, among the sound pressure signals in the frequency characteristic Fs5 converted by the signal conversion unit 22. In this way, the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc are obtained by the abnormal sound emphasis unit 23 for each component device.

[0105] According to this, the abnormal sound determination unit 25 can compare the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc with the cutting abnormality template Fs4 in which the sound pressure signals of the first motor band RM1, the second motor band RM2, and the cutting band RC are emphasized. As a result, when the operating sound of the target equipment P contains an abnormal sound, it is possible to estimate the component equipment that causes the abnormal sound among the component equipments, the first motor M1, the second motor M2, and the tool T.

[0106] (First modified example of the second embodiment) In the above-described second embodiment, an example has been described in which the speaker 50 issues a sound that emphasizes the operation sounds of the component devices that cause the abnormal sound when the abnormal sound determination unit 25 determines that the operation sounds of the target device P do not include abnormal sounds, but the present invention is not limited to this. For example, the abnormal sound detection device 1 may be configured to issue operation sounds in which the operation sounds of the first motor M1, the second motor M2, and the tool T are all individually emphasized by performing an inverse short-time Fourier transform on all of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc.

[0107] (First modified example of the second embodiment) In the above-described second embodiment, an example has been described in which the first motor band RM1, the second motor band RM2, and the cutting band RC are based on the operation sounds when the first motor M1, the second motor M2, and the tool T, which are the components of the target device P, are in a normal state, but the present invention is not limited to this. The first motor band RM1, the second motor band RM2, and the cutting band RC may be based on the operation sounds when the first motor M1, the second motor M2, and the tool T are in an abnormal state.

[0108] Third embodiment Next, the third embodiment will be described with reference to Figs. 16 to 22. In this embodiment, the information on the operating sound of the target device P stored in advance in the storage unit 21 is different from that in the first and second embodiments. In this embodiment, in addition to a part of the configuration of the EQ unit 20 being different, the processes performed by the abnormal sound emphasis unit 23, the sound analysis unit 24, and the abnormal sound determination unit 25 are different from those in the first and second embodiments. In this embodiment, the parts different from the first and second embodiments will be mainly described, and the description of the parts similar to the first and second embodiments may be omitted.

[0109] The target equipment P in this embodiment is a cutting machine similar to that in the second embodiment. Similarly to the second embodiment, the abnormal sound detection device 1 in this embodiment determines whether or not an abnormal sound is included in the operation sounds, including a first motor sound, a second motor sound, and a cutting sound, generated when the cutting machine, which is the target equipment P, performs cutting processing.

[0110] However, unlike the first and second embodiments, the storage unit 21 of this embodiment does not store information about abnormal sounds contained in the operation sounds of the target device P detected before abnormal sound detection is performed. Specifically, the storage unit 21 does not store information about the first motor band RM1, the second motor band RM2, the cutting band RC, the cutting abnormality template Fs4, and the cutting template vector Vtc. In other words, the storage unit 21 does not store information about the operation sounds of the target device P, including normal sounds and abnormal sounds of the component devices, the first motor M1, the second motor M2, and the tool T.

[0111] In contrast to this, the storage unit 21 of this embodiment is configured to be capable of storing the abnormal sound feature amounts of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc determined by the abnormal sound emphasis unit 23, as will be described later. The storage unit 21 is also configured to be capable of storing setting information to be set in a change amount setting unit 40, which will be described later.

[0112] 16, the abnormal sound detection device 1 of this embodiment is different from the first and second embodiments in that it additionally includes a change amount setting section 40. The change amount setting section 40 is connected to an input device D provided externally to the abnormal sound detection device 1.

[0113] In addition, the sound analysis unit 24 of this embodiment is connected to the memory unit 21, as shown in FIG. 17, and is configured to be able to output the abnormal sound features of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc calculated by the sound analysis unit 24 to the memory unit 21.

[0114] In addition, the abnormal sound detection unit 25 of this embodiment is connected to the memory unit 21 and is configured to be able to read from the memory unit 21 the abnormal sound feature quantities of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc that the sound analysis unit 24 previously output to the memory unit 21.

[0115] The change amount setting unit 40 is a device that executes processing according to a program recorded in the ROM of the storage unit 21 or a writable non-volatile storage medium. The change amount setting unit 40 sets the amount of change in the sound pressure signal when the abnormal sound emphasis unit 23 emphasizes the intensity of the sound pressure signal of a frequency included in the operation sound of a constituent device of the target device P. Specifically, the change amount setting unit 40 is configured to be capable of setting in advance an increase amount when increasing the intensity of the sound pressure signal of a frequency included in the operation sound of the constituent device, and to be capable of setting an decrease amount when decreasing the intensity of the sound pressure signal of a frequency different from the frequency included in the operation sound of the constituent device.

[0116] The change amount setting unit 40 is configured to be able to set these increase / decrease amounts and decrease amounts for each frequency of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc determined by the abnormal noise emphasis unit 23.

[0117] Specifically, the amount of increase and decrease of the sound pressure signal set by the change amount setting unit 40 can be set within a predetermined increase / decrease range based on the predetermined increase or decrease amount shown in the first embodiment. The predetermined increase / decrease range is set in advance. For example, the predetermined increase / decrease range in this embodiment is set to -3 dB to +3 dB. Note that the predetermined increase / decrease range is not limited to the range of -3 dB to +3 dB, and the minimum value may be set to a value smaller than -3 dB or larger than -3 dB. Note that the maximum value of the predetermined increase / decrease range may be set to a value smaller than +3 dB or larger than +3 dB.

[0118] In this embodiment, the amount of increase and decrease of the sound pressure signal set by the change amount setting unit 40 can be set for each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc, for each of the predetermined frequency bands. Specifically, the amount of increase and decrease of the sound pressure signal set by the change amount setting unit 40 can be set for each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc, for each of the bands in the first frequency band R1 to the eighth frequency band R8.

[0119] More specifically, the change amount setting unit 40 can set the amount of increase when the abnormal sound emphasis unit 23 increases the sound pressure signal of the first motor band RM1 and the amount of decrease when the abnormal sound emphasis unit 23 decreases the sound pressure signal of a frequency band different from the first motor band RM1.

[0120] In addition, the change amount setting unit 40 can set the amount of increase when the abnormal sound emphasis unit 23 increases the sound pressure signal of the second motor band RM2 and the amount of decrease when the abnormal sound emphasis unit 23 decreases the sound pressure signal of a frequency band different from the second motor band RM2.

[0121] The change amount setting unit 40 can set the amount of increase when the allophone emphasis unit 23 increases the sound pressure signal in the cutting band RC and the amount of decrease when the allophone emphasis unit 23 decreases the sound pressure signal in a frequency band different from the cutting band RC. The input side of the change amount setting unit 40 is connected to the input device D, and these increase and decrease amounts are controlled by signals input from the input device D. The output side of the change amount setting unit 40 is connected to the storage unit 21 of the EQ unit 20, and the change amount setting unit 40 outputs setting information set by signals input from the input device D to the storage unit 21.

[0122] The input device D is an input unit for performing operations for setting the amounts of increase and decrease in the sound pressure signal set by the change amount setting unit 40. The input device D may be attached to the abnormal sound detection device 1, or may be provided in a location remote from the abnormal sound detection device 1. Alternatively, the input device D may be a terminal carried by an operator who operates the abnormal sound detection device 1. In this case, the amounts of increase and decrease in the sound pressure signal set by the change amount setting unit 40 are controlled by wireless communication between the input device D and the change amount setting unit 40.

[0123] As shown in Fig. 18, the input device D includes a setting unit Ds having an operation button Db. The input device D is configured to be able to control the amount of increase and decrease in the sound pressure signal set by the change amount setting unit 40 for each of the component devices, the first motor M1, the second motor M2, and the tool T, by an operator operating the operation button Db. The input device D controls the amount of increase and decrease in the sound pressure signal for each of the first frequency band R1 to the eighth frequency band R8 set by the change amount setting unit 40.

[0124] 18 shows settings for increasing the sound pressure signals of the first frequency band R1 and the second frequency band R2, which are the first motor band RM1, by an additional 3 dB based on a predetermined increase amount when the abnormal sound emphasis unit 23 calculates the first motor emphasis characteristic Fm1. The operator sets the increase / decrease amount and the decrease amount when the abnormal sound emphasis unit 23 calculates the first motor emphasis characteristic Fm1 based on the first motor band RM1 obtained from the results of an experiment conducted in advance before abnormal sound detection.

[0125] 18 also shows settings for increasing the sound pressure signals of the fifth frequency band R5 to the eighth frequency band R8, which are the second motor band RM2, by an additional 3 dB based on a predetermined increase amount when the abnormal sound emphasis unit 23 calculates the second motor emphasis characteristic Fm2. The operator sets the increase / decrease amount and the decrease amount when the abnormal sound emphasis unit 23 calculates the second motor emphasis characteristic Fm2, based on the second motor band RM2 obtained from the results of an experiment conducted in advance before abnormal sound detection.

[0126] 18 shows a setting for increasing the sound pressure signals of the third frequency band R3 to the sixth frequency band R6, which are the cutting band RC, by an additional 3 dB based on a predetermined increase amount when the abnormal sound emphasis unit 23 calculates the cutting emphasis characteristic Fc. Furthermore, FIG. 18 shows a setting for decreasing the sound pressure signals of the first frequency band R1, the second frequency band R2, the seventh frequency band R7, and the eighth frequency band R8, which are bands different from the cutting band RC, by an additional -3 dB based on a predetermined decrease amount.

[0127] The operator sets the amount of increase or decrease when the abnormal sound emphasis unit 23 calculates the cutting emphasis characteristic Fc, based on the cutting band RC obtained from the results of an experiment performed in advance before abnormal sound detection.

[0128] The operation of the abnormal sound detection device 1 of this embodiment configured as described above will be described below with reference to Figure 19. Note that in the description of the operation of this embodiment using Figure 19, descriptions of control processes that are the same as the operation in the first embodiment described using Figure 8 or the operation in the second embodiment described using Figure 11 may be omitted.

[0129] When the abnormal sound emphasis unit 23 receives the frequency characteristic Fs5 converted by the signal conversion unit 22, it reads information on the increase / decrease and decrease amount of each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc stored in the storage unit 21. Then, the abnormal sound emphasis unit 23 emphasizes the intensity of the sound pressure signal of frequencies related to abnormal sound among the frequencies in the frequency characteristic Fs5, based on the increase / decrease amount and decrease amount of each of the frequency characteristic Fs5, the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc.

[0130] Specifically, the abnormal sound emphasis unit 23 obtains a first motor emphasis characteristic Fm1 by emphasizing the intensity of the sound pressure signal in the first motor band RM1 relative to the frequency characteristic Fs5 converted by the signal conversion unit 22. The abnormal sound emphasis unit 23 also obtains a second motor emphasis characteristic Fm2 by emphasizing the intensity of the sound pressure signal in the second motor band RM2 relative to the frequency characteristic Fs5 converted by the signal conversion unit 22. The abnormal sound emphasis unit 23 then obtains a cutting emphasis characteristic Fc by emphasizing the intensity of the sound pressure signal in the cutting band RC relative to the frequency characteristic Fs5 converted by the signal conversion unit 22. In this way, the abnormal sound emphasis unit 23 obtains the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc for each component device.

[0131] Here, in this embodiment, the increase and decrease amounts of the sound pressure signal shown in Figure 18 are set for each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc by the input signal input from the input device D.

[0132] For this reason, as shown in Fig. 20, the first motor emphasis characteristic Fm1 determined by the abnormal sound emphasis unit 23 has the sound pressure signal in the first motor band RM1 increased by the change amount setting unit 40 compared to the first motor emphasis characteristic Fm1 shown in Fig. 13 of the second embodiment. For this reason, the first motor emphasis characteristic Fm1 further emphasizes the sound pressure signal of the frequency contained in the operation sound of the first motor M1 compared to a case where the sound pressure signal in the first motor band RM1 is not increased by the change amount setting unit 40. Note that the dashed line shown in Fig. 20 indicates the sound pressure signal when the sound pressure signal in the first motor band RM1 in the first motor emphasis characteristic Fm1 is not emphasized by the change amount setting unit 40.

[0133] 21, the second motor emphasis characteristic Fm2 determined by the abnormal sound emphasis unit 23 has the sound pressure signal in the second motor band RM2 increased by the change amount setting unit 40 compared to the second motor emphasis characteristic Fm2 shown in FIG 14 of the second embodiment. Therefore, the second motor emphasis characteristic Fm2 further emphasizes the sound pressure signal of the frequency included in the operating sound of the second motor M2 compared to a case where the sound pressure signal in the second motor band RM2 is not increased by the change amount setting unit 40. The dashed line shown in FIG 21 indicates the sound pressure signal when the sound pressure signal in the second motor band RM2 in the second motor emphasis characteristic Fm2 is not emphasized by the change amount setting unit 40.

[0134] As shown in Fig. 22, the cutting emphasis characteristic Fc calculated by the abnormal sound emphasis unit 23 is such that the sound pressure signal in the cutting band RC is increased by the change amount setting unit 40 compared to the cutting emphasis characteristic Fc shown in Fig. 15 of the second embodiment. Furthermore, in the cutting emphasis characteristic Fc shown in Fig. 22, the sound pressure signal in a band different from the cutting band RC is decreased by the change amount setting unit 40 compared to the cutting emphasis characteristic Fc shown in Fig. 15 of the second embodiment.

[0135] Therefore, in the cutting emphasis characteristic Fc, the sound pressure signal of the frequency contained in the operation sound when the tool T cuts the workpiece is further emphasized, compared with the case where the sound pressure signal of the cutting band RC is not increased and the sound pressure signal of the band different from the cutting band RC is not decreased by the change amount setting unit 40. Note that the dashed line shown in Fig. 22 indicates the sound pressure signal in the case where the sound pressure signal of the cutting band RC in the cutting emphasis characteristic Fc is not emphasized by the change amount setting unit 40.

[0136] The abnormal sound emphasis unit 23 outputs information on the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc to the sound analysis unit 24 and the signal inverse conversion unit 26.

[0137] When information on the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc is input from the abnormal sound emphasis unit 23, the sound analysis unit 24 determines the abnormal sound feature quantities of each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc.

[0138] Specifically, the sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the first motor emphasis characteristic Fm1 as a first motor feature amount Pm1, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal in the first motor emphasis characteristic Fm1. The sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the second motor emphasis characteristic Fm2 as a second motor feature amount Pm2, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal in the second motor emphasis characteristic Fm2. The sound analysis unit 24 obtains an abnormal sound feature amount related to the intensity of a sound pressure signal in the cutting emphasis characteristic Fc as a cutting feature amount Pc, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal in the cutting emphasis characteristic Fc.

[0139] The sound analysis unit 24 outputs information on the calculated first motor characteristic amount Pm1, second motor characteristic amount Pm2, and cutting characteristic amount Pc to the storage unit 21 and the abnormal sound determination unit 25.

[0140] The storage unit 21 stores information on the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc input from the sound analysis unit 24.

[0141] When information on the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc is input, the abnormal sound detection unit 25 determines whether or not the operating sound of the target equipment P includes an abnormal sound based on the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc.

[0142] Specifically, the abnormal sound judgment unit 25 compares the first motor feature amount Pm1 determined by the sound analysis unit 24 in the current control cycle with the first motor feature amount Pm1 determined by the sound analysis unit 24 in the past and stored in the storage unit 21. The abnormal sound judgment unit 25 also compares the second motor feature amount Pm2 determined by the sound analysis unit 24 in the current control cycle with the second motor feature amount Pm2 determined by the sound analysis unit 24 in the past and stored in the storage unit 21. Then, the abnormal sound judgment unit 25 compares the cutting feature amount Pc determined by the sound analysis unit 24 in the current control cycle with the cutting feature amount Pc determined by the sound analysis unit 24 in the past and stored in the storage unit 21.

[0143] The first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc previously determined by the sound analysis unit 24 are, for example, the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc determined by the sound analysis unit 24 in the immediately previous control cycle.

[0144] Then, the abnormal sound determination unit 25 determines the changes over time of each of the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc by determining the differences between the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc based on the comparison result. The abnormal sound determination unit 25 determines whether or not the operating sound of the target device P includes an abnormal sound based on the changes over time of each of the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc.

[0145] Specifically, the abnormal sound judgment unit 25 judges whether or not the operating sounds of the component equipment include abnormal sounds based on whether or not the amount of change in the first motor feature amount Pm1, the amount of change in the second motor feature amount Pm2, and the amount of change in the cutting feature amount Pc are each greater than or equal to a predetermined judgment reference value.

[0146] In this embodiment, the predetermined judgment reference value is a reference value determined in advance by experiments or the like in order to determine whether or not abnormal sounds are included in the operation sounds of the first motor M1, the second motor M2, and the tool T. The predetermined judgment reference values ​​for determining whether or not abnormal sounds are included in the operation sounds of the first motor M1, the second motor M2, and the tool T may be set to different values ​​corresponding to the first motor M1, the second motor M2, and the tool T, respectively, or may be set to the same value.

[0147] The abnormal sound detector 25 determines whether or not each of the amount of change in the first motor characteristic amount Pm1, the amount of change in the second motor characteristic amount Pm2, and the amount of change in the cutting characteristic amount Pc is equal to or greater than a predetermined determination reference value.

[0148] When the value of at least one of the change amount of the first motor characteristic amount Pm1, the change amount of the second motor characteristic amount Pm2, and the change amount of the cutting characteristic amount Pc is equal to or greater than a predetermined judgment reference value, the abnormal sound judgment unit 25 judges that an abnormal sound is included in the operating sound of the target device P. Then, the abnormal sound judgment unit 25 judges that the component device of the target device P, that is, the first motor M1, the second motor M2, and the tool T, whose change amount of the abnormal sound characteristic amount is equal to or greater than the predetermined judgment reference value, is the cause of the occurrence of the abnormal sound.

[0149] In contrast, if the abnormal sound detection unit 25 does not determine that all of the calculated changes in the first motor feature Pm1, the second motor feature Pm2, and the cutting feature Pc are greater than or equal to a predetermined judgment reference value, it determines that the operating sound of the target equipment P does not include abnormal sound.

[0150] As described above, the abnormal sound detection device 1 that detects abnormal sounds from a target device P having a first motor M1, a second motor M2, and a tool T as its constituent devices includes a sound analysis unit 24 that determines the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc, which are the abnormal sound feature amounts. The abnormal sound detection device 1 also includes a storage unit 21 that stores the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc determined for each constituent device. The abnormal sound determination unit 25 then determines whether or not the operating sound of the target device P includes an abnormal sound based on the past first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc, and the current first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc determined by the sound analysis unit 24.

[0151] This makes it possible to determine whether or not the operating sound of the target device P includes an abnormal sound, without having to analyze in advance by experiment or the like the operating sound of the target device P when it operates in an abnormal state.

[0152] Furthermore, according to the above embodiment, the following effects can be obtained.

[0153] (1) In the above embodiment, the abnormal sound determination unit 25 determines whether or not the operating sound of the target equipment P includes an abnormal sound based on the changes over time of the first motor feature Pm1, the second motor feature Pm2, and the cutting feature Pc obtained by the sound analysis unit 24.

[0154] According to this, since it is determined whether or not the operating sound of the target device P includes an abnormal sound based on changes in the operating sounds of the constituent devices of the target device P, it is easy to detect the occurrence timing when an abnormality occurs in the target device P.

[0155] (2) In the above embodiment, the change amount setting unit 40 is provided, which sets the amount of change when changing the intensity of the sound pressure signal of the frequencies in the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc.

[0156] This makes it easier to emphasize the sound pressure signals of the frequencies related to abnormal sounds in the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc at a desired level, which makes it easier for the abnormal sound determination unit 25 to determine whether or not an abnormal sound is included in the operation sound of the target device P. Furthermore, for example, when an operation sound including an abnormal sound is to be notified to an operator so that the operator can perform an abnormal sound detection task, the sound pressure signals of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc can be set to a desired level according to the operator's own desire.

[0157] (First modified example of the third embodiment) In the above-mentioned third embodiment, an example has been described in which the change amount setting unit 40 is configured to be able to set the amount of increase and decrease of the sound pressure signal for one or more of the eight bands indicated by the first frequency band R1 to the eighth frequency band R8, but the present invention is not limited to this. For example, the number of bands for which the change amount setting unit 40 can set the amount of increase and decrease of the sound pressure signal may be less than eight or more than eight.

[0158] If the number of bands for which the change amount setting unit 40 can set the increase and decrease amounts of the sound pressure signal is too small, it may be difficult to emphasize the intensity of the sound pressure signal at frequencies related to abnormal sounds from the target device P, and if the number is too large, the operation for setting the increase and decrease amounts may become cumbersome. For this reason, the number of bands for which the change amount setting unit 40 can set the increase and decrease amounts of the sound pressure signal may be set to 8 to 12, which makes it easy to emphasize the intensity of the sound pressure signal at frequencies related to abnormal sounds from the target device P and does not make setting the increase and decrease amounts cumbersome.

[0159] The change amount setting section 40 may be configured to set the amount of increase and decrease of the sound pressure signal of the selected frequency in one band selected from the eight bands for each band set in more detail. In this case, for example, as shown in Fig. 23, the change amount setting section 40 may be configured to set the amount of increase and decrease of the sound pressure signal for one or more bands out of ten bands set in advance in the one selected band.

[0160] (Second modified example of the third embodiment) In the above-described third embodiment, the abnormal sound feature of the first motor emphasis characteristic Fm1 is obtained as the first motor feature Pm1, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the first motor emphasis characteristic Fm1. The abnormal sound feature of the second motor emphasis characteristic Fm2 is obtained as the second motor feature Pm2, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the second motor emphasis characteristic Fm2. Furthermore, the abnormal sound feature of the cutting emphasis characteristic Fc is obtained as the cutting feature Pc, which is the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the cutting emphasis characteristic Fc. However, the method of obtaining the abnormal sound feature of each of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc is not limited to this.

[0161] For example, the maximum value of the sound pressure signal of the first motor emphasis characteristic Fm1 may be used as the abnormal sound feature amount of the first motor emphasis characteristic Fm1. Also, the maximum value of the sound pressure signal of the second motor emphasis characteristic Fm2 may be used as the abnormal sound feature amount of the second motor emphasis characteristic Fm2. And, the maximum value of the sound pressure signal of the cutting emphasis characteristic Fc may be used as the abnormal sound feature amount of the cutting emphasis characteristic Fc.

[0162] Alternatively, the abnormal sound feature of the first motor emphasis characteristic Fm1 may be the average value of the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the first motor emphasis characteristic Fm1 at a predetermined sampling time. Also, the abnormal sound feature of the second motor emphasis characteristic Fm2 may be the average value of the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the second motor emphasis characteristic Fm2 at a predetermined sampling time. And, the abnormal sound feature of the cutting emphasis characteristic Fc may be the average value of the absolute value of the difference between the maximum and minimum values ​​of the sound pressure signal of the cutting emphasis characteristic Fc at a predetermined sampling time.

[0163] (Third modified example of the third embodiment) In the above-mentioned third embodiment, an example was described in which the determination as to whether or not the operation sound of the target device P includes an abnormal sound is performed based on the time-dependent changes of the first motor feature amount Pm1, the second motor feature amount Pm2, and the cutting feature amount Pc, but the present invention is not limited to this. For example, the determination as to whether or not the operation sound of the target device P includes an abnormal sound may be performed based on the results of statistical analysis of the abnormal sound feature amounts of the first motor emphasis characteristic Fm1, the second motor emphasis characteristic Fm2, and the cutting emphasis characteristic Fc for each predetermined sampling time. For the statistical analysis, for example, a mixed Gaussian distribution, a Poisson distribution, a normal distribution, or the like may be used.

[0164] (Fourth Modification of the Third Embodiment) In the above-mentioned third embodiment, an example was described in which the past first motor feature amount Pm1, second motor feature amount Pm2, and cutting feature amount Pc are adopted as the first motor feature amount Pm1, second motor feature amount Pm2, and cutting feature amount Pc in the immediately preceding control cycle, but the present invention is not limited to this. For example, the first motor feature amount Pm1, second motor feature amount Pm2, and cutting feature amount Pc previously calculated by the sound analysis unit 24 may be average values ​​of the first motor feature amount Pm1, second motor feature amount Pm2, and cutting feature amount Pc calculated at a predetermined sampling time.

[0165] (Other embodiments) Representative embodiments of the present disclosure have been described above, but the present disclosure is not limited to the above-described embodiments and can be modified in various ways, for example, as described below.

[0166] In the above-described first embodiment, an example has been described in which the abnormal sound detection device 1 includes the signal inverse conversion unit 26 that converts the abnormal sound emphasis characteristic Fs3 into a time signal, and the speaker 50 that generates a sound that emphasizes an abnormal sound relative to the operation sound of the target device P, but the present invention is not limited to this. For example, the abnormal sound detection device 1 may be configured not to include the signal inverse conversion unit 26 that converts the abnormal sound emphasis characteristic Fs3 into a time signal, and the speaker 50 that generates a sound that emphasizes an abnormal sound relative to the operation sound of the target device P.

[0167] In the above embodiment, an example has been described in which the intensity of the sound pressure signal is increased by a predetermined increase amount when the abnormal sound emphasis unit 23 performs an upward adjustment, and the intensity of the sound pressure signal is decreased by a predetermined decrease amount when the abnormal sound emphasis unit 23 performs a downward adjustment, but the present invention is not limited to this. When performing an upward adjustment, the abnormal sound emphasis unit 23 may multiply the sound pressure signal of the abnormal sound frequency Fa by a predetermined coefficient to make the sound pressure signal of the abnormal sound frequency Fa larger than the sound pressure signal before multiplication. Also, when performing a downward adjustment, the abnormal sound emphasis unit 23 may multiply the sound pressure signal of the normal frequency Fn by a predetermined coefficient to make the sound pressure signal of the normal frequency Fn smaller than the sound pressure signal before multiplication.

[0168] In the above-mentioned second and third embodiments, the abnormal sound emphasis unit 23 has been described as increasing and decreasing the sound pressure signal of one or more bands obtained by dividing 0 to 20000 Hz into eight equal frequency bands of 2500 Hz each, but is not limited thereto. For example, the abnormal sound emphasis unit 23 may increase and decrease the sound pressure signal of one or more bands obtained by dividing 0 to 20000 Hz into bands of different sizes from 2500 Hz. In addition, the abnormal sound emphasis unit 23 may increase and decrease the sound pressure signal of one or more bands obtained by dividing 0 to 20000 Hz into octave bands or 1 / 3 octave bands with a center frequency of 10000 Hz.

[0169] In the above-mentioned second and third embodiments, an example has been described in which the abnormal sound detection device 1 determines whether or not an abnormal sound is included in the operating sounds, which include the first motor sound, the second motor sound, and the cutting sound that are generated simultaneously when the target equipment P performs cutting processing; however, the present invention is not limited to this.

[0170] For example, the abnormal sound detection device 1 may determine whether or not an abnormal sound is included in each of the operating sounds, which include a first motor sound, a second motor sound, and a cutting sound, which are generated at different times when the target equipment P performs cutting processing.

[0171] In the above-described embodiments, it goes without saying that the elements constituting the embodiments are not necessarily essential, except in cases where it is particularly expressly stated that they are essential or where they are clearly considered to be essential in principle.

[0172] In the above-described embodiments, when numerical values ​​such as the number, values, amounts, ranges, etc. of components of the embodiments are mentioned, they are not limited to the specific numbers unless expressly stated as required or clearly limited to a specific number in principle.

[0173] In the above-described embodiments, when referring to the shapes, positional relationships, etc. of components, etc., there is no limitation to those shapes, positional relationships, etc., unless specifically stated or in principle limited to a specific shape, positional relationship, etc. [Explanation of symbols]

[0174] 10 Sound acquisition section 22 Signal conversion section 23 Allophone Emphasis Section 25 Abnormal sound detection section

Claims

1. An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires the operation sound of the target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic that indicates the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; A storage unit (21) for storing the allophone feature amount; a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal, The target device has a plurality of constituent devices (M1, M2, C), the abnormal sound emphasis unit obtains the abnormal sound emphasis signal by performing at least an increasing adjustment of increasing an intensity of a sound pressure signal of a frequency estimated to be included in an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics and a decreasing adjustment of decreasing an intensity of a sound pressure signal of a frequency different from the frequency estimated to be included in the abnormal sound of the target device, and obtains the abnormal sound emphasis signal for each of the plurality of constituent devices by emphasizing an intensity of a sound pressure signal of a frequency included in an operating sound when each of the plurality of constituent devices operates in a normal state among the sound pressure signals for each frequency in the frequency characteristics; the sound analysis unit determines the allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound detection device, wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound feature amounts for each of the plurality of component devices in the past stored in the memory unit and the abnormal sound feature amounts for each of the plurality of component devices currently determined by the sound analysis unit.

2. The storage unit pre-stores abnormal sound information which is information regarding abnormal sounds included in the operation sound of the target device, The abnormal sound detection device according to claim 1 , wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal and the abnormal sound information.

3. The abnormal sound detection device according to claim 2 , wherein the abnormal sound information is set based on an operating sound of the target device operating in an abnormal state.

4. The abnormal sound information is set based on an operating sound of each of the plurality of component devices when the plurality of component devices are operating, 3. The abnormal sound detection device according to claim 2, wherein the abnormal sound emphasis unit determines the abnormal sound emphasis signal for each of the plurality of constituent devices by emphasizing intensity of a sound pressure signal of a frequency that is included in the operation sounds of the plurality of constituent devices stored in the storage unit, among the sound pressure signals for each frequency in the frequency characteristics.

5. An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires the operation sound of the target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic that indicates the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; a storage unit (21) for storing in advance abnormal sound information, which is information regarding abnormal sounds contained in the operation sound of the target device, and the abnormal sound feature quantity; a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal, The target device has a plurality of constituent devices (M1, M2, C), the abnormal sound information is set based on operation sounds of the plurality of constituent devices when the plurality of constituent devices are operating, the abnormal sound emphasis unit performs at least one of an increasing adjustment for increasing an intensity of a sound pressure signal of a frequency estimated to be included in an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics, and a decreasing adjustment for decreasing an intensity of a sound pressure signal of a frequency different from the frequency estimated to be included in the abnormal sound of the target device, and obtains the abnormal sound emphasis signal by emphasizing an intensity of a sound pressure signal of a frequency included in an operating sound of the plurality of constituent devices stored in the storage unit among the sound pressure signals for each frequency in the frequency characteristics, and obtains the abnormal sound emphasis signal for each of the plurality of constituent devices by emphasizing an intensity of a sound pressure signal of a frequency included in an operating sound when each of the plurality of constituent devices operates in a normal state among the sound pressure signals for each frequency in the frequency characteristics, the sound analysis unit determines the allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound detection device, wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound feature amounts for each of the plurality of component devices in the past stored in the memory unit and the abnormal sound feature amounts for each of the plurality of component devices currently determined by the sound analysis unit.

6. The abnormal sound information indicates the operating sound including the abnormal sound of the target device by the intensity of a sound pressure signal for each frequency, and emphasizes the intensity of the sound pressure signal of a frequency that is estimated to be included in the abnormal sound of the target device among the sound pressure signals for each frequency, the storage unit stores an estimated feature quantity indicating a feature related to an intensity of a sound pressure signal of the abnormal sound information as the abnormal sound information, The abnormal sound detection device according to claim 2 , wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound feature amount and the estimated feature amount.

7. An abnormal sound detection device that detects abnormal sounds contained in operating sounds of a target device to be inspected, A sound acquisition unit (10) that acquires the operation sound of the target device as a sound pressure time signal, which is a time signal indicating a change in intensity of a sound pressure signal over time; A signal conversion unit (22) that converts the sound pressure time signal into a frequency characteristic that indicates the intensity of the sound pressure signal for each frequency; an abnormal sound emphasis unit (23) that obtains an abnormal sound emphasis signal by emphasizing the intensity of a sound pressure signal of a frequency related to an abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics based on the frequency characteristics; an abnormal sound determination unit (25) that determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound emphasis signal; a sound analysis unit (24) for determining an allophone feature quantity indicating a feature related to the intensity of a sound pressure signal in the allophone emphasis signal; A storage unit (21) for storing the allophone feature amount; a change amount setting unit (40) that sets an amount of change when the abnormal sound emphasis unit changes the sound pressure signal for each frequency in the frequency characteristic in order to obtain the abnormal sound emphasis signal, The target device has a plurality of constituent devices (M1, M2, C), the abnormal sound emphasis unit obtains the abnormal sound emphasis signal for each of the plurality of constituent devices by emphasizing an intensity of a sound pressure signal of a frequency included in an operation sound when each of the plurality of constituent devices operates in a normal state, among the sound pressure signals for each frequency in the frequency characteristics; the sound analysis unit determines the allophone feature amount for each of the plurality of constituent devices based on the allophone emphasis signal determined by the allophone emphasis unit for each of the plurality of constituent devices; The abnormal sound detection device, wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on the abnormal sound feature amounts for each of the plurality of component devices in the past stored in the memory unit and the abnormal sound feature amounts for each of the plurality of component devices currently determined by the sound analysis unit.

8. 8. The abnormal sound detection device according to claim 7, wherein the abnormal sound emphasis unit determines the abnormal sound emphasis signal by performing at least one of an increasing adjustment for increasing intensity of a sound pressure signal of a frequency estimated to be included in the abnormal sound of the target device among the sound pressure signals for each frequency in the frequency characteristics, and a decreasing adjustment for decreasing intensity of a sound pressure signal of a frequency different from the frequency estimated to be included in the abnormal sound of the target device.

9. 9. The abnormal sound detection device according to claim 7, wherein the abnormal sound determination unit determines whether or not an abnormal sound is included in the operation sound of the target device based on a change over time in the abnormal sound feature amount calculated based on a difference between the abnormal sound feature amount for each of the plurality of component devices in the past stored in the storage unit and the abnormal sound feature amount for each of the plurality of component devices currently determined by the abnormal sound emphasis unit.

10. a signal inverse conversion unit (26) for converting the allophone emphasis signal into an emphasis sound signal which is a time signal indicating a change over time in the intensity of a sound pressure signal; and a sound generating unit (50) that generates a sound that emphasizes an abnormal sound of the target device relative to an operation sound of the target device based on the emphasis sound signal.

Images