Abnormal sound specifying device in abnormal sound diagnosis system

Abstract

To allow for more accurately specifying the presence or absence of an abnormal sound and its frequency band in an abnormal sound diagnosis system configured to detect an abnormal sound from among sounds emitted from a mechanical apparatus and collected by a microphone and diagnose a source of the abnormal sound, by confirming whether an abnormal vibration actually occurs in a part of the mechanical apparatus in a frequency band where an abnormal sound is estimated to occur on a power spectrum of sounds collected by the microphone or not when a worker hears the sound with his / her ears and confirms sounds extracted in this frequency band.SOLUTION: An abnormal sound specifying device generates a power spectrum of sounds collected by a microphone, extracts a component of a frequency band set by a worker from sound data collected by the microphone and reproduces this component as a sound which the worker can hear with his / her ears, and extracts a component in the same band as a frequency band defined for the sound data from acceleration value data measured by an acceleration sensor installed in a mechanical apparatus and reproduces this component as a sound which the worker can hear with his / her ears.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a system (abnormal sound diagnosis system) that detects abnormal sounds generated in vehicles or other machinery and equipment and diagnoses their causes, and more specifically, to a device that enables an operator of the system to identify abnormal sounds. [Background technology]

[0002] In the past, abnormal vibrations during the operation of vehicles and other machinery and equipment were detected by workers who would place a machine or automotive stethoscope on various parts of the machinery and equipment to listen to the vibrations as sounds and search for the source of the abnormal vibrations, but detection of abnormal vibrations using such a stethoscope required experienced and skilled techniques.In recent years, therefore, with the aim of enabling even inexperienced users to detect abnormal vibrations in machinery and equipment, development has been underway of abnormal sound diagnosis systems that detect the presence or absence of abnormal vibrations and their frequency bands from sounds collected by machinery and equipment, and diagnose the source of the detected abnormal vibrations using AI (artificial intelligence). As such a system, for example, Patent Document 1 proposes a configuration in which a microphone that collects sounds in a band from about 20 Hz to about 10 kHz is installed at an appropriate location on the object of abnormal sound diagnosis (vehicle or other machine or equipment), the power spectrum of the sound measured by the microphone is calculated, and an operator extracts and plays sounds in an arbitrarily defined frequency band while referring to the power spectrum, and confirms the presence of a sound (abnormal sound) corresponding to abnormal vibration, thereby detecting abnormal sounds, and further, an AI diagnoses the cause of the abnormal sound based on the frequency characteristics of the detected abnormal sound. Patent Document 2 proposes a configuration in which the frequency characteristics of an audio input signal are analyzed as an acoustic processing to improve the clarity of any audio signal, and signals of different frequency bands are given to the left and right ears based on the analysis results. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2023-108581 [Patent Document 2] Japanese Patent Application Laid-Open No. 2006-87018 [Summary of the Invention] [Problems to be Solved by the Invention]

[0004] In the abnormal sound diagnosis system as described above, in the process of identifying an abnormal sound from among sounds collected by a microphone in a machine tool, that is, a sound generated from a part of the machine tool that is causing abnormal vibration, more specifically, first, a power spectrum of the sound (with frequency as a variable) is created from the data of the sound collected by the microphone using STFT (short-time Fourier transform) or the like, then, an operator refers to the power spectrum and defines a frequency band where the abnormal sound is estimated to be generated, extracts the sound of the defined frequency band using a band-pass filter and plays it back, and the operator listens to the played back sound with his / her ears to determine whether or not an abnormal sound actually exists (and, when the occurrence of an abnormal sound is confirmed, information on the abnormal sound is sent to AI, and its cause is diagnosed). In the process of identifying an abnormal sound by the operator's ear, the sound collected by the microphone actually includes not only sounds generated from each part of the machine tool, but also various environmental sounds superimposed thereon, so it may be difficult to determine whether the sound heard in the defined frequency band where the abnormal sound is estimated to be generated on the power spectrum is indeed due to abnormal vibration in a part of the machine tool. In addition, when an operator listens to a sound extracted in a frequency band defined on a power spectrum where an abnormal sound is estimated to be generated, if the volume of the abnormal sound is not sufficiently large compared to the environmental sound, it may be difficult to confirm the presence of the abnormal sound. Therefore, when an operator listens to a sound extracted and played back in a frequency band defined on a frequency spectrum to confirm the presence of an abnormal sound, if it can be confirmed that a part of the machine actually vibrates abnormally in the same frequency band, the presence of the abnormal sound can be more accurately determined. In other words, when a sound extracted in a frequency band defined on a power spectrum where an abnormal sound is estimated to be generated and a vibration of a part of the machine can be confirmed in the same frequency band, the sound heard as an abnormal sound can be more accurately determined to be an "abnormal sound". In addition, whether or not a part of the machine is vibrating in a certain frequency band can be confirmed by installing an appropriate acceleration sensor on the machine, extracting a signal in the certain frequency band from the detection data of the acceleration sensor, playing it back as a sound, and listening to the sound with the operator's ears.

[0005] In view of the above circumstances, a main object of the present invention is to enable an operator to listen to and confirm with his or her ears the sound extracted and played back in a frequency band defined on the power spectrum as the frequency band in which the abnormal sound is presumed to be occurring in an abnormal sound in an abnormal sound diagnosis system such as that described above, at the stage of confirming the presence of an abnormal sound from among sounds picked up by a microphone, and to confirm whether or not abnormal vibrations are actually occurring in a part of the machine or tool in the same frequency band, thereby enabling the presence or absence of an abnormal sound and its frequency band to be more accurately identified. [Means for solving the problem]

[0006] According to the present invention, the above problem is solved by providing an apparatus for allowing an operator to specify the presence or absence of an abnormal sound and its frequency band in an abnormal sound diagnosis system configured to detect the presence or absence of an abnormal sound from sounds collected by a microphone in a machine or tool and diagnose the source of the abnormal sound when the abnormal sound is present, A microphone for collecting sound installed on the machine or device; a frequency analysis means for performing a frequency analysis of the sound data collected by the microphone to generate a power spectrum; A frequency band definition means for defining a frequency band in the power spectrum set by the operator; a first bandpass filter that extracts components in the frequency band set by the operator from the sound data collected by the microphone; a first sound reproducing means for reproducing the sound data components extracted by the first band pass filter as sounds audible to the worker; An acceleration sensor installed on the machine or tool; a second band-pass filter that extracts components in the frequency band set by the operator from the acceleration value data measured by the acceleration sensor; a second sound reproducing means for reproducing the acceleration data component extracted by the second band pass filter as a sound audible to the worker; This is achieved by an apparatus comprising:

[0007] In the above configuration, the "machine" may be a vehicle such as an automobile or any other machine. The "microphone" may be, for example, any type of microphone that collects sounds in the audible range. In the present invention, the "abnormal sound" is, as already described, a sound generated from a part of the machine that is vibrating abnormally, and does not include abnormal environmental sounds. The "abnormal sound diagnosis system" is, as described above, a system configured to detect the presence or absence of an abnormal sound from a sound collected by a microphone in the machine and to diagnose the source of the abnormal sound when the abnormal sound is present. The "frequency analysis" may be an analysis that generates a power spectrum with frequency as a variable by using STFT or FFT on the time series sound intensity value (sound data). The "acceleration sensor" may be any sensor that can capture the vibration of a part as a change in acceleration value. The bandpass filter is a filter that extracts components of a specified frequency band from sound data or acceleration data. The first and second sound reproducing means may each be a speaker (which may be a speaker of a headphone or an earphone) that reproduces the intensity value or acceleration value of the sound sent from the bandpass filter as sound.

[0008] As described above, the device of the present invention is, in short, a device (abnormal sound identification device) used by an operator to identify an "abnormal sound" from among sounds collected by a microphone in the above-mentioned "abnormal sound diagnosis system." In the conventional abnormal sound diagnosis system, for example, as described in Patent Document 1, an operator defines a frequency band in which an abnormal sound is estimated to exist on a power spectrum generated by frequency analysis of data on sounds collected by a microphone, plays back sound components in the defined frequency band, and judges the presence or absence of an abnormal sound by listening to the played back sound. However, as already described, since various environmental sounds are also superimposed on the sounds collected by the microphone, it was sometimes difficult to determine whether an abnormal sound corresponding to abnormal vibrations in a part of a machine or tool is occurring and to identify the abnormal sound from the sound alone.

[0009] Therefore, the device of the present invention is configured so that the worker can hear the frequency band components in which the abnormal sound is estimated to exist, extracted from the sound collected by the microphone, and at the same time, hear the frequency band components in which the abnormal sound is estimated to exist, extracted from the acceleration value detected by the acceleration sensor of the machine. Here, the acceleration value detected by the acceleration sensor is a value that directly captures the vibration of a part of the machine as an acceleration change. With this configuration, the worker can detect whether or not abnormal vibration is occurring in any part of the machine corresponding to the sound that is estimated to be an abnormal sound, from the sound collected by the microphone, and thus it is possible to more accurately identify the presence or absence of an abnormal sound corresponding to the abnormal vibration in the machine and its frequency band than before.

[0010] In the above-mentioned configuration of the device of the present invention, the processing of the frequency analysis means and the bandpass filter may be performed using digital data obtained by A / D (analog-digital) conversion of the sound data or acceleration data, and the sound reproduction means may be configured to reproduce the analog data obtained by D / A (digital-analog) conversion of the digital data output from the bandpass filter as a voice through a speaker. The installation positions of the microphone and the acceleration sensor in the machine tool may be appropriately selected through experiments, etc. Furthermore, the frequency analysis means, the frequency band definition means, the first and second bandpass filters, and the first and second sound reproduction means according to the present invention may be realized in a mobile terminal (smartphone, tablet) that can be operated by the worker. In an embodiment, a configuration may be adopted in which a component extracted from the sound collected by the microphone is reproduced on one side of the headphones, and a component extracted from the data detected by the acceleration sensor is reproduced on the other side of the headphones. Effect of the Invention

[0011] Thus, according to the device of the present invention, in a so-called abnormal sound diagnosis system, at the stage where an operator confirms the presence of an abnormal sound, the presence or absence of vibration of a part of the machine tool obtained from the acceleration sensor can be confirmed together with the sound collected by the microphone, so the presence or absence of an abnormal sound and its frequency band can be more accurately identified. In the case of the configuration of the device of the present invention, the components of the acceleration sensor are hardly affected by environmental sounds, so it is easier to determine whether the sound heard is noise such as environmental sounds compared to when only sound is present, and since the presence or absence of vibration of a part can be determined, it is also possible to detect minute abnormal sounds whose peaks are difficult to see in the results of frequency analysis.

[0012] Other objects and advantages of the present invention will become apparent from the following description of preferred embodiments of the invention. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a schematic diagram of an abnormal sound diagnosis system including an abnormal sound identification device according to this embodiment. [Diagram 2] FIG. 2 is a block diagram showing the configuration of the abnormal sound identification device according to this embodiment. [Diagram 3] FIG. 3 is a schematic diagram of a power spectrum of a sound, and is a diagram for explaining a process in which a frequency band in which an abnormal sound is estimated to exist is defined by an operator on the power spectrum of the sound in the abnormal sound identification device according to this embodiment. [Explanation of symbols]

[0014] 10...vehicle, 12...microphone, 13...communication device, 14...acceleration sensor, 15...communication device, 20...mobile terminal, 22...headphone, 23L, R...speaker, 30...diagnosis device BEST MODE FOR CARRYING OUT THEINVENTION

[0015] The present invention will now be described in detail with reference to some preferred embodiments thereof with reference to the accompanying drawings, in which like reference numerals indicate like parts.

[0016] Configuration of abnormal sound diagnosis system In an abnormal sound diagnosis system in which the abnormal sound identification device according to this embodiment is employed, as shown in Fig. 1, a sound collection microphone 12 and an acceleration sensor 14 are installed in appropriately selected locations (not necessarily at the same location) in a machine or equipment such as a vehicle 10, and the voice data collected by the microphone 12 is transmitted by a communication device 13, and the acceleration value data detected by the acceleration sensor 14 is transmitted by a communication device 15 to a terminal 20 (which may be a smartphone, tablet, or the like) held by an operator, by short-range wireless communication technology. The transmitted data is then processed in the terminal 20 in a manner described in detail later, and is reproduced as sound by speakers 23L and 23R of a headphone 22. When the operator listens to the data and detects an abnormal sound in the voice data from the microphone 12, information relating to the characteristics of the abnormal sound is transmitted in an arbitrary manner to a diagnosis device 30 (which may be a computer device that operates according to program commands) by short-range wireless communication technology, etc., and the diagnosis device 30 executes a diagnosis of the source of the abnormal sound by a discriminator using AI technology, etc. Here, the microphone 12 may be a microphone capable of collecting sounds in the human audible range (range of about 20 Hz to 10 kHz) that is employed in conventional abnormal sound diagnosis systems (such as those described in Patent Document 1, etc.). The acceleration sensor 14 may be any sensor capable of detecting vibrations occurring in a part of the machine or instrument as changes in acceleration value.

[0017] Configuration of the device for identifying abnormal noise In the abnormal sound identification device of this embodiment employed in the abnormal sound diagnosis system described above, the configuration for reproducing the sound that the operator hears from the sound data collected by the microphone 12 may be the same as that of a conventional abnormal sound diagnosis system. Specifically, as shown in the left side of Fig. 2, the configuration includes an A / D converter that converts the sound data, which is an analog signal collected by the microphone 12, into a digital signal, a frequency analyzer that performs STFT or FFT on the sound data converted into a digital signal to generate a power spectrum with frequency as a variable, a display that displays the power spectrum, a processor (frequency band definition processor) that defines a frequency band specified by the operator on the power spectrum, a bandpass filter that extracts components of the defined frequency band from the sound data collected by the microphone 12 and converted into a digital signal, a D / A converter that converts the components extracted by the bandpass filter into analog, and a speaker that reproduces the sound data converted into analog by the D / A converter. In this embodiment, a configuration is further provided for reproducing data extracted from the acceleration value data obtained by the acceleration sensor 14 as a sound to be heard by the worker. Specifically, as shown in the right side of Fig. 2, this configuration includes an A / D converter for converting the acceleration value data obtained by the acceleration sensor 14 into a digital signal, a bandpass filter for extracting components in the acceleration value data in the same frequency band as the frequency band defined by the frequency band definition processor for the sound data, a D / A converter for converting the components extracted by the bandpass filter into analog, and a speaker for reproducing the acceleration data converted into analog by the D / A converter. In addition, in order to check the frequency characteristics of the acceleration value data, a frequency analyzer may be provided for performing STFT or FFT on the acceleration value data to generate a power spectrum with frequency as a variable.

[0018] The above series of configurations may be realized by operations according to a program of a computer device in a mobile terminal carried by a worker. The speaker that reproduces as sound the components extracted from the voice data collected by the microphone and the speaker that reproduces as sound the components extracted from the acceleration value data measured by the acceleration sensor may be, for example, left and right speakers of headphones (or earphones) as shown in FIG.

[0019] Activation of the device for identifying abnormal noise In the operation of the abnormal sound identification device, first, the microphone 12 and the acceleration sensor 14 of the machine tool simultaneously measure the voice and acceleration value, respectively, and the data are converted into digital signals by an A / D converter (the A / D conversion process may be performed after transmission to the terminal 20), and then transmitted to the worker's mobile terminal. In the mobile terminal, the transmitted voice data is frequency analyzed to generate a power spectrum, and the power spectrum PS (horizontal axis f: frequency, vertical axis P: power) of the voice data is displayed on the display unit of the mobile terminal, as shown in FIG. 3. Here, the worker refers to the power spectrum PS and operates the screen of the display unit to define and identify a frequency band Δ in which an abnormal sound is estimated to exist (frequency band definition processor). In this case, the width w, upper limit Lu, and lower limit Lb of the frequency band Δ may be set by operating the screen with a finger, for example. Thus, when the frequency band Δ is specified, the components of the voice data in the band Δ are extracted by the band pass filter, and the components of the transmitted acceleration value data in the band Δ are extracted by the band pass filter. The extracted components are converted into analog signals by the D / A converter and reproduced as voice by the speaker. Then, the worker listens to the voice and acceleration vibration in the frequency band Δ reproduced by the speaker with his ears to check whether or not there is an abnormal sound, i.e., abnormal vibration, originating from a part of the machine or tool in the band Δ. After that, when checking whether or not there is an abnormal sound in another frequency band, another band is defined on the power spectrum of the voice displayed on the display unit, and similarly, the voice data and acceleration value data in the defined band are extracted and reproduced as voice. By the operation of the abnormal sound identification device described above, it is possible to check whether or not there is an abnormal sound, and if there is an abnormal sound, to identify the frequency band.

[0020] According to the configuration of the present embodiment, as already described, not only the sound collected by the microphone but also the presence or absence of vibration in parts of the machine or tool can be confirmed by the acceleration sensor, so it is easy to determine whether the sound heard or the sound of interest is due to the vibration of the machine or tool or noise such as environmental sound, and it is expected that the presence or absence of abnormal sounds and their frequency bands can be detected more accurately. Also, it is advantageous that even a minute abnormal sound whose peak is difficult to see on the power spectrum of the sound can be easily diagnosed as an abnormal sound if there is a vibration in the acceleration value at the same frequency.

[0021] The above description has been given in relation to the embodiment of the present invention, but it will be apparent to those skilled in the art that many modifications and changes can be easily made thereto, and that the present invention is not limited to the embodiment exemplified above, but can be applied to various devices without departing from the concept of the present invention.

Claims

[Claim 1] An abnormal sound diagnostic system that detects the presence or absence of abnormal sounds from sounds collected by a microphone in a machine or tool, and diagnoses the source of the abnormal sounds when the abnormal sounds are present, and allows an operator to specify the presence or absence of the abnormal sounds and their frequency bands, A microphone for collecting sound installed on the machine or tool; a frequency analysis means for performing a frequency analysis of the sound data collected by the microphone to generate a power spectrum; a frequency band definition means for defining a frequency band in the power spectrum that is set by the operator; a first bandpass filter that extracts components in the frequency band set by the operator from the sound data collected by the microphone; a first sound reproducing means for reproducing the sound data components extracted by the first band pass filter as sounds audible to the worker; An acceleration sensor installed on the machine or tool; a second band-pass filter that extracts components in the frequency band set by the operator from the acceleration value data measured by the acceleration sensor; a second sound reproducing means for reproducing the acceleration data component extracted by the second band pass filter as a sound audible to the worker; An apparatus comprising:

Images