JP7774497B2 - Abnormal sound detection system and abnormal sound detection method - Google Patents

Description

The present invention relates to an abnormal sound detection system and an abnormal sound detection method.

Engineers working in plants and factories such as refineries may hear unusual operating noises while working and become aware of an abnormality in the equipment. However, plants and factories may be equipped with high-noise sources such as compressors, and when the noise level reaches 85 dB or higher, for example, engineers wear noise protection equipment such as earplugs and earmuffs. Wearing noise protection equipment can make it difficult for even experienced engineers to hear unusual noises emitted by the equipment.

Patent Document 1 discloses a system for identifying the cause of abnormal noise (for example, vehicle components such as the engine or drivetrain). This system generates FFT data by performing an FFT (Fast Fourier Transform) on sound pressure data containing abnormal noise, and performs statistical analysis to calculate the degree of abnormality. It then creates integrated FFT data by integrating FFT data with an abnormality degree above a threshold, and identifies the cause of the abnormal noise by comparing the integrated FFT data with clusters that have been clustered for each abnormality cause.

Japanese Patent Application Laid-Open No. 2021-081364

Plants and factories need a way to detect abnormal sounds and pinpoint their source, even in high-noise environments.

The object of this invention is to provide an abnormal sound detection system and method that can solve the above-mentioned problems.

According to one aspect of the present disclosure, an abnormal sound detection system includes a sound collection and detection device and an acoustic signal analysis device. The sound collection and detection device has a sound collection unit that collects external sound and outputs the collected sound as acoustic data, a positioning unit that determines position information of the location where the sound is collected, and a communication unit that transmits the sound data and the position information determined by the positioning unit. The acoustic signal analysis device has a receiving unit that receives the sound data and the position information transmitted by the communication unit, an acoustic signal analysis unit that performs frequency analysis of the sound data to calculate spectrum data of an object to be evaluated, an abnormal sound determination unit that determines whether the sound data contains an abnormal sound by comparing the spectrum data of the object to be evaluated in a predetermined frequency domain with predetermined spectrum data in a normal state, and a position estimation unit that, if it is determined that an abnormal sound is included, estimates the location of the abnormal sound based on the position information of the location where the sound data determined to contain an abnormal sound was collected.

According to one aspect of the present disclosure, an abnormal sound detection method includes the steps of receiving acoustic data of collected external sound and positional information of the position where the sound was collected; calculating spectrum data of an object to be evaluated by performing frequency analysis on the acoustic data; determining whether the acoustic data contains an abnormal sound by comparing the spectrum data of the object to be evaluated in a predetermined frequency range with predetermined spectrum data in a normal state; and, if it is determined in the determining step that an abnormal sound is contained, estimating the position where the abnormal sound is occurring based on the positional information of the position where the acoustic data determined to contain an abnormal sound was collected.

This disclosure makes it possible to detect abnormal sounds even in high-noise environments and estimate the location where the abnormal sounds are originating.

FIG. 1 is a schematic diagram illustrating an example of an abnormal sound detection system according to an embodiment. FIG. 1 is a block diagram illustrating an example of an abnormal sound detection system according to an embodiment. FIG. 10 is a diagram showing an example of spectrum data of area A according to the embodiment. FIG. 10 is a diagram showing an example of time-series spectrum data of area A according to the embodiment. FIG. 10 is a diagram illustrating an example of spectrum data of area B according to the embodiment. FIG. 10 is a diagram illustrating an example of spectrum data of an area C according to the embodiment. FIG. 10 is a diagram showing an example of spectrum data of the D area according to the embodiment. FIG. 10 is a diagram showing an example of spectrum data when an abnormal noise occurs according to the embodiment. FIG. 1 is a first diagram illustrating the process of estimating the abnormal noise generation position according to the embodiment. FIG. 10 is a second diagram illustrating the process of estimating the abnormal noise generation position according to the embodiment. 10 is a flowchart illustrating an example of an abnormal sound detection process according to the embodiment.

<Embodiment>
(System configuration)
An abnormal sound detection system according to an embodiment of the present disclosure will be described below with reference to FIGS. 1 to 10 . FIG. 1 is a schematic diagram illustrating an example of an abnormal sound detection system according to an embodiment. The abnormal sound detection system 100 uses a sound collection and detection device 10 worn by a worker M working within a facility G, such as a plant or factory, to collect acoustic data at various locations within the facility G, determine whether an abnormal sound is occurring, and, if an abnormal sound is occurring, estimate where within the facility G the abnormal sound is coming from. The abnormal sound detection system 100 includes the sound collection and detection device 10, a wireless communication device 5, a network NW, an acoustic signal analysis device 20, a display device 30, and an external storage device 40. The sound collection and detection device 10 is incorporated into noise protection equipment, such as earplugs or earmuffs, worn by the worker M. Alternatively, the sound collection and detection device 10 may be attached to an armband or work clothes worn by the worker M, or may be incorporated into a wristwatch, tool, or various devices carried by the worker M. Furthermore, detection of sounds including abnormal sounds may be performed by sound collection and detection devices 10 worn by multiple workers M working at various locations in the facility G, or by the sound collection and detection devices 10 worn by workers M patrolling the facility G for monitoring, transporting materials, etc. Furthermore, in addition to being worn by the workers M, the sound collection and detection devices 10 may also be attached to a mobile object such as a transport device used for monitoring, transporting, etc. The sound data collected by the sound collection and detection devices 10 is transmitted to the acoustic signal analysis device 20 via the wireless communication device 5 and the network NW. The acoustic signal analysis device 20 receives the transmitted acoustic data and performs frequency analysis such as FFT (fast Fourier transform) to determine whether the acoustic data contains abnormal sounds. If abnormal sounds are detected, the acoustic signal analysis device 20 estimates the source of the abnormal sounds. The acoustic signal analysis device 20 then outputs the estimated location of the abnormal sounds to the display device 30. For example, facility G has areas A to D as shown in the figure, and the acoustic signal analyzer 20 determines whether an abnormal sound is occurring by comparing it with sounds collected in the vicinity of areas A to D under normal circumstances, and if an abnormal sound is occurring, estimates the location of the abnormal sound based on acoustic data collected at multiple locations. Below, an example will be described in which an abnormal sound is detected based on sounds collected in the vicinity of areas A to D, but there is no need to limit the area, and the abnormal sound detection system 100 can detect an abnormal sound by comparing it with sounds generated under normal circumstances from any location in facility G, and identify the location of its occurrence.

FIG. 2 is a block diagram illustrating an example of an abnormal sound detection system according to the embodiment.
The sound collection and detection device 10 includes a sound collection unit 11, a communication unit 12, and a positioning unit 13. The sound collection unit 11 collects surrounding sounds and outputs the collected sounds as acoustic data (electrical signals). The sound collection unit 11 is, for example, a microphone. The communication unit 12 is a communication module that enables wireless communication such as Wi-Fi (registered trademark) or wireless LAN. The positioning unit 13 includes, for example, a Global Navigation Satellite System (GNSS) receiver and measures position information of the sound collection and detection device 10. The positioning unit 13 may acquire position information using an RFID (Radio Frequency Identifier) tag. The positioning unit 13 may acquire position information based on signal strength from wireless communication devices such as Wi-Fi (registered trademark) or wireless LAN. Alternatively, the positioning unit 13 may acquire position information by incorporating a gyro, acceleration sensor, and an autonomous navigation calculation module. The sound collection/detection device 10 transmits the acoustic data collected by the sound collection unit 11 and the position information of the sound collection/detection device 10 determined by the positioning unit 13 to the acoustic signal analysis device 20 using the communication unit 12 at a predetermined control period.

The wireless communication device 5 mediates communication between the sound collection and detection device 10 and the acoustic signal analysis device 20. For example, the wireless communication device 5 may include an access point (AP) capable of communicating with the sound collection and detection device 10 and a router capable of communicating with the acoustic signal analysis device 20. The wireless communication device 5 may communicate directly with the sound collection and detection device 10, or may communicate with the sound collection and detection device 10 via a mobile terminal such as a smartphone connected to the sound collection and detection device 10 via Bluetooth (registered trademark). The sound collection and detection device 10 communicates with the acoustic signal analysis device 20 located on a network NW such as the Internet via the wireless communication device 5. Specifically, the sound collection and detection device 10 transmits acoustic data and location information to the acoustic signal analysis device 20.

The acoustic signal analysis device 20 is configured by a server device and includes a receiving unit 21, an input unit 22, a control unit 23, and a storage unit 28.
The receiving unit 21 receives the acoustic data and location information transmitted by the sound collection and detection device 10, and records these in the storage unit 28 in association with the time of reception.
The input unit 22 is configured using input devices such as a keyboard, a mouse, a touch panel, buttons, etc., and accepts user operations using the input devices. For example, the input unit 22 accepts settings regarding which frequency range of the spectrum obtained by FFT analysis of acoustic data should be focused on for abnormal sound detection.
The control unit 23 analyzes the acoustic data collected by the sound collection and detection device 10, and controls the process of detecting abnormal sounds and the process of estimating the position where the abnormal sound is generated. The control unit 23 includes an acoustic signal analysis unit 24, an abnormal sound determination unit 25, a position estimation unit 26, and an output unit 27.
The acoustic signal analysis unit 24 performs frequency analysis on the acoustic data received by the receiving unit 21 using FFT or the like, and records the analyzed spectrum data in the storage unit 28 .
The abnormal sound determination unit 25 compares the spectrum data calculated by the acoustic signal analysis unit 24 with the spectrum of the sound emitted from each area of the facility G, and determines whether or not the acoustic data contains an abnormal sound.
The position estimation unit 26 estimates the location where the abnormal sound has occurred, using spectrum data that has been determined to contain an abnormal sound by the abnormal sound determination unit 25. For example, the position estimation unit 26 estimates the location where the abnormal sound has occurred by a three-point positioning method, based on acoustic data collected by the sound collection and detection device 10 at least three locations.
The output unit 27 outputs the determination result by the abnormal sound determination unit 25 and the abnormal sound generation position estimated by the position estimation unit 26 to the display device 30.
The storage unit 28 stores various data received by the receiving unit 21, spectrum data calculated by the acoustic signal analysis unit 24, and normal spectrum data for each area used by the abnormal sound detection unit 25 to detect abnormal sounds.

The display device 30 is configured using a liquid crystal display or the like, and is connected to the acoustic signal analysis device 20. The display device 30 displays the information output by the output unit 27.
The external storage device 40 is a storage device provided outside the acoustic signal analysis device 20. The external storage device 40 may be provided in a so-called cloud computing system. The acoustic signal analysis device 20 and the external storage device 40 can communicate with each other via a network NW, and the acoustic signal analysis device 20 can record data in the external storage device 40 and read data from the external storage device 40. In the abnormal sound detection system 100, the external storage device 40 can be used in place of the storage unit 28. For example, the receiving unit 21 of the acoustic signal analysis device 20 may record various received data in the external storage device 40. The acoustic signal analysis unit 24 of the acoustic signal analysis device 20 may record calculated spectrum data in the external storage device 40. The external storage device 40 may also store normal spectrum data for each area. In the following description, the storage unit 28 is used as an example to record various data, but the external storage device 40 can be used in place of the storage unit 28.

(Method for detecting abnormal noise)
In each of areas A to D of facility G, sounds with a spectrum specific to that area are output, even under normal conditions. Spectrum data of sound collected from each area under normal conditions (when no abnormal sounds are occurring) is pre-registered in the memory unit 28 as reference data to be used for comparison. The abnormal sound determination unit 25 detects abnormal sounds by comparing the normal spectrum data prepared for each area with spectrum data obtained by analyzing the sound data transmitted from the sound collection and detection device 10. Examples of normal spectrum data for each area are shown in Figures 3 to 8. The vertical axis of the graphs shown in Figures 3 to 8 represents the sound pressure level (dB), and the horizontal axis represents the frequency (kHz).

Figure 3 is a diagram showing an example of spectrum data for area A according to an embodiment. Figure 3 shows spectrum data w0 obtained from acoustic data collected in area A at a certain time. Box w1 shows the waveform in the frequency range above 600 kHz, and box w2 shows the peak value in the frequency range above 600 kHz.

Figure 4 is a diagram showing an example of time-series spectrum data for area A according to an embodiment. Figure 4 shows spectrum data w0 obtained from acoustic data collected in area A from times t1 to t7. As in Figure 3, boxes w1 and w2 each show the waveform and peak value in the frequency range above 600 kHz. As shown in Figure 4, in the frequency range below 600 kHz, the waveform changes significantly over time. In contrast, in the frequency range above 600 kHz, there is little fluctuation in the waveform over time, and the peak value and the frequency at which the peak value appears remain almost unchanged.

Similarly, Figures 5 to 7 each show an example of spectrum data w0 for areas B to D under normal conditions. In each figure, box w1 shows the waveform in the frequency range above 600 kHz, and box w2 shows the peak value. Although not shown, analysis results confirm the same thing as described using Figure 4 for areas B to D. That is, even in areas B to D, the waveform fluctuates significantly over time in the frequency range below 600 kHz, but fluctuates little over time in the frequency range above 600 kHz. Furthermore, the peak value and the frequency at which the peak value appears do not change. In other words, because the waveform fluctuates significantly in the frequency range below 600 kHz even under normal conditions, it is difficult to determine whether or not an abnormal noise is occurring by focusing on this frequency range. Conversely, if we focus on the frequency range above 600 kHz, where there is little waveform fluctuation, and a change in the waveform occurs in this frequency range, we can conclude that an abnormal noise is occurring.

As such, it is known that in each of areas A through D, the waveform fluctuates significantly in frequency ranges below a predetermined value (600 kHz in this example), while there is little change in the waveform and peak value above that frequency. Taking advantage of this property, this embodiment focuses on the frequency range above a predetermined value (e.g., 600 kHz) and determines that abnormal noise has occurred when waveform fluctuations not observed under normal conditions occur in this frequency range. Examples of waveform fluctuations not observed under normal conditions include a change in the frequency at which peak values appear, or an increase in overall sound pressure to the point where peak values are hidden (generally, when abnormal noise occurs due to an equipment malfunction, the sound does not become quieter, but rather becomes louder). The abnormal sound determination unit 25 compares the waveform of the spectrum data of the acoustic data transmitted from the sound collection and detection device 10 with the waveform of normal spectrum data for areas A through D, limited to the predetermined frequency range, to determine whether the acoustic data transmitted from the sound collection and detection device 10 contains abnormal sound. In this example, the frequency range above 600 kHz is uniformly focused on, but the frequency range focused on for each area may be different, such as 600 kHz or above for area A, 500 kHz or above for area B, etc.

(Example of abnormal noise)
FIG. 8 shows an example of spectrum data of acoustic data collected when an abnormal noise occurs in Area C. As in FIGS. 3 to 7 , w0 represents the spectrum data, w1 represents the frequency range above 600 kHz, and w2 represents the frequency at which a peak value appears. Furthermore, waveform w0' in FIG. 8 shows the spectrum of acoustic data collected at a certain time when a malfunction occurred in a device in Area C. As shown in the figure, spectrum data w0' exceeds normal spectrum data w0 in the frequency range above 600 kHz, and the frequency components within frame w2 of spectrum data w0' exceed the normal peak value. Furthermore, unlike normal cases, spectrum data w0' exhibits a peak value in the frequency range indicated by frame w2'. When a waveform state different from normal occurs, such as when a peak value appears in a frequency range different from normal cases or when the sound pressure in the frequency range of interest (within frame w1) increases overall compared to normal cases (the sound becomes louder in the treble range), the abnormal sound detector 25 determines that an abnormal sound has occurred.

(Estimation of the location of the abnormal noise)
Next, the process of estimating the location of an abnormal sound when it is determined that an abnormal sound has occurred will be described. As shown in FIG. 9A , assume that an abnormal sound is detected in the acoustic data collected by the sound collection/detection device 10 when a worker M patrolling within a facility G is located at points a, b, and c. The position estimation unit 26 then estimates the location of the abnormal sound by triangulation using the location information of points a, b, and c. The triangulation method is a well-known algorithm used, for example, in situations where the location of a terminal transmitting radio waves is calculated based on the radio wave intensity measured by three wireless LAN access points. By replacing the radio wave intensity in this algorithm with sound pressure level, the location of the abnormal sound can be estimated using the triangulation method. For example, the location information of point a is (p1, q1), the location information of point b is (p2, q2), and the location information of point c is (p3, q3). Furthermore, let the position information of the abnormal sound source be (x, y), the distance between point a and the abnormal sound source be R1, the distance between point b and the abnormal sound source be R2, and the distance between point c and the abnormal sound source be R3. Then, the following equations (1) to (3) are established as shown in FIG. 9B.
(x-p1) ² + (y-q1) ² = R1 ² ...(1)
(x-p2) ² + (y-q2) ² = R2 ² ...(2)
(x-p3) ² + (y-q3) ² = R3 ² ...(3)

(p1, q1), (p2, q2), and (p3, q3) are known because they are measured by the sound collection and detection device 10. Therefore, if R1 to R3 are known, the position information (x, y) of the abnormal sound source can be found using equations (1) to (3). R1 to R3 are calculated, for example, as follows. Generally, if the sound level measured at a point r1 (m) (for example, r1 = 1 (m)) from a certain noise source is Lr1 (dB), and the sound level measured at a point r2 (m) from the noise source is Lr2 (dB), then the following equation (4) holds.
Noise attenuation (Lr1-Lr2)=10·log ₁₀ (r1/r2) (4)
When equation (4) is solved for Lr2, the following equation (4') is obtained.
Lr2=Lr1-10・log ₁₀ (r1/r2)
=Lr1-10・(log _e (r1/r2)/log ₁₀ )...(4')
Let the sound pressure level (dB) of the abnormal sound detected at each of points a to c be noise Lr2 (dB), and let the distance between each of points a to c and the location where the abnormal sound is generated be r2 (i.e., R1, R2, and R3, respectively). Equation (4') is obtained for each of points a to c. From these three equations (4'), R1, R2, and R3 are calculated so that Lr1 is equal. Then, using equations (1) to (3) above, the approximate range where the abnormal sound source (x, y) is located can be estimated. Point P in Figure 9A is an example of the location of the abnormal sound estimated using triangular positioning. In this way, the location of the abnormal sound can be estimated by collecting acoustic data at various locations in facility G and analyzing the collected acoustic data.

(operation)
Next, the abnormal sound detection process of this embodiment will be described with reference to Fig. 10. As mentioned above, the following description will also be given taking as an example a case where the storage unit 28 is used as the data recording destination, but the external storage device 40 may be used instead of the storage unit 28.
FIG. 10 is a flowchart illustrating an example of an abnormal sound detection process according to the embodiment.
As a premise, the memory unit 28 stores normal spectrum data for each area, as exemplified in FIGS. 3 and 5 to 7 . The normal spectrum data may be updated with spectrum data obtained by frequency analysis of the acoustic signal analyzer 24 using acoustic data detected by the sound collection and detection device 10 near each of areas A to D while the worker M is patrolling the facility G daily, or may be selected by an engineer from the spectrum data obtained by frequency analysis of the acoustic signal analyzer 24. The frequency range of interest (e.g., 600 kHz or higher) and the frequency of the peak value in that frequency range may be set by an engineer with knowledge, or the acoustic signal analyzer 24 may analyze the time-dependent fluctuations in normal spectrum data obtained over a predetermined period of time to detect and automatically set the frequency range in which the waveform fluctuations fall within a predetermined range, the peak value in that frequency range, and the frequency at which the peak value appears. Information on normal spectrum data for each area, the frequency range of interest, the peak value, and the frequency at which the peak value appears is registered in the memory unit 28.

Furthermore, worker M wearing sound collection and detection device 10 periodically patrols facility G, measuring acoustic data and location information at various locations within facility G. Sound collection and detection device 10 pairs acoustic data and location information measured at the same time (the same time period that can be considered the same time) and transmits them in real time to acoustic signal analysis device 20. This data is transmitted to acoustic signal analysis device 20 via network NW. In parallel with the measurement and transmission of acoustic data, etc. by sound collection and detection device 10, the following processing is repeatedly executed at a predetermined control cycle on the acoustic signal analysis device 20 side.

In the acoustic signal analysis device 20, when the receiver 21 receives a set of acoustic data and location information (step S1), it associates these with the time of reception and records them in the memory unit 28. Next, the acoustic signal analyzer 24 performs frequency analysis on the recorded acoustic data using FFT or the like (step S2), and records the spectrum data resulting from the analysis (referred to as the spectrum data to be evaluated) in the memory unit 28, associating it with the acoustic data, location information, and reception time information. Next, the abnormal sound detector 25 determines whether the acoustic data contains an abnormal sound (step S3). The abnormal sound detector 25 compares the spectrum data to be evaluated with normal spectrum data to determine whether the received acoustic data contains an abnormal sound (detects an abnormal sound). For example, the abnormal sound determination unit 25 compares a region above a predetermined frequency (e.g., 600 kHz or higher) in any of the spectrum data for areas A to D when normal, or in a waveform obtained by combining multiple pieces of spectrum data for areas A to D when normal, with the same frequency region in the spectrum data to be evaluated, and if the waveform shape, sound pressure level, peak value magnitude, and frequency at which the peak value appears match, or if the difference between the two falls within a predetermined range that can be considered to match, determines that an abnormal sound is not included in the acoustic data transmitted from the sound collection and detection device 10. Furthermore, if, as a result of the comparison, the spectrum data to be evaluated does not match in characteristics with either the spectrum data for areas A to D when normal, or the waveform data obtained by combining multiple pieces of spectrum data for areas A to D when normal, the abnormal sound determination unit 25 determines that the spectrum data to be evaluated contains an abnormal sound. For example, if the average sound pressure level in the frequency range of interest of the spectrum data to be evaluated minus the average sound pressure level in the frequency range of interest of the spectrum data for each area under normal conditions exceeds a predetermined threshold, or if a peak value appears at a frequency not present in the spectrum data for each area under normal conditions, the abnormal sound detector 25 determines that the spectrum data to be evaluated contains an abnormal sound. If an abnormal sound is determined to be present, the abnormal sound detector 25 records flag information indicating that an abnormal sound has been detected in the storage unit 28, in association with the spectrum data containing the abnormal sound, acoustic data, reception time information, and location information. The abnormal sound detector 25 also records the sound pressure (dB) of the peak value detected in the frequency range of interest in the storage unit 28, in association with the spectrum data containing the abnormal sound. The sound pressure of the peak value is used as Lr2 (dB) in the three-point positioning method described above.

If the abnormal sound detection unit 25 determines that no abnormal sound is included (step S4; No), the output unit 27 outputs information such as "normal" or "no abnormal sound detected" to the display device 30 along with information about the location and time (reception time) at which the acoustic data was collected (step S7).

If the abnormal sound determination unit 25 determines that an abnormal sound is included (Step S4; Yes), the position estimation unit 26 performs a position estimation. First, the position estimation unit 26 searches the memory unit 28 to determine whether a total of three or more pieces of spectrum data, peak sound pressure, acoustic data, location information, and reception time information, each marked with flag information indicating that an abnormal sound has been detected, are recorded. At this time, the unit may select only acoustic data collected within a predetermined time period based on the reception time information of the acoustic data determined in Step S3 to contain an abnormal sound (e.g., data from several days ago may be excluded because it is too old and different from the currently detected abnormal sound). Alternatively, the range of location information may be limited (e.g., acoustic data containing an abnormal sound collected only a few meters away a few seconds ago may be excluded because it is not very useful for use in triangulation calculations). If three or more pieces of acoustic data with flag information indicating that an abnormal sound has been detected, including the acoustic data determined to contain an abnormal sound in step S3, cannot be found (step S5; No). Therefore, location estimation using the triangulation method is not possible. For example, a supervisor or other person conducting monitoring using the acoustic signal analyzer 20 instructs the worker M to collect more acoustic data. The worker M patrols the facility G and measures acoustic data and location information using the sound collection and detection device 10. The sound collection and detection device 10 transmits the acoustic data and location information to the acoustic signal analyzer 20. The processing from step S1 onward is repeated until acoustic data determined to contain an abnormal sound is collected at three or more locations. At this time, the output unit 27 may output to the display device 30 a message indicating that an abnormal sound has been detected, that there is insufficient data indicating that an abnormal sound has been detected, and the location information of the locations where the abnormal sound has been detected to date. Upon seeing this output, the supervisor may instruct the on-site worker M to go to an appropriate location and collect acoustic data using the sound collection and detection device 10.

If data determined to contain an abnormal sound exists from three or more locations (Step S5; Yes), the location estimation unit 26 estimates the location of the abnormal sound using triangulation, using acoustic data detected at three different locations (Step S6). With triangulation, if the location of the abnormal sound does not lie within the area enclosed by the three selected locations, no solution is found. Therefore, the acoustic data determined to contain an abnormal sound in Step S3 is used, and for the remaining two data points, any two points may be selected from acoustic data collected at different locations. The location estimation using triangulation may be repeated while varying the combination of the three locations containing the data determined to contain the abnormal sound to estimate the location of the abnormal sound. The location estimation unit 26 records the estimated location of the abnormal sound in the memory unit 28. Note that the method for estimating the location of the abnormal sound is not limited to triangulation and other methods may be used. For example, if acoustic data containing an abnormal sound is collected around a certain area, it may be estimated that the source of the abnormal sound is roughly located in that area. Next, the output unit 27 outputs the estimated location of the abnormal sound to the display device 30 (step S7). For example, as shown in FIG. 9A, the output unit 27 may generate an image showing the three points used in the estimation and the location P of the abnormal sound on a map of facility G, and output this image to the display device 30. If multiple locations of the abnormal sound are estimated, all of the locations of the abnormal sound may be displayed.

As described above, the abnormal sound detection system 100 of this embodiment is equipped with a sound collection and detection device 10 including a sound collection unit 11 for external environmental sounds, a communication unit 12, and a positioning unit 13, mounted on noise protection equipment worn by a worker. The system compares external sounds collected by the sound collection unit 11 with normal sounds, and if there is a difference between the two, determines that an abnormal sound has been detected. The comparison is performed focusing on frequency ranges that exhibit characteristics specific to each device and location and exhibit little fluctuation over time, thereby enabling accurate detection of abnormal sounds. Furthermore, the location of the abnormal sound can be estimated based on the location information of the location where the abnormal sound was detected, allowing for the rapid identification and treatment of abnormal occurrences within the vast facility G. Thus, according to this embodiment, abnormal sounds can be detected and their location identified early simply by worker M wearing noise protection equipment such as earplugs while patrolling his or her normal patrol route, or by worker M simply performing normal work in his or her workplace while wearing the noise protection equipment.

The processing steps in the sound collection and detection device 10 and the acoustic signal analysis device 20 can be realized by a processor, such as a CPU (Central Processing Unit), included in each of the sound collection and detection device 10 and the acoustic signal analysis device 20 executing a program. The programs executed by the sound collection and detection device 10 and the acoustic signal analysis device 20 may be recorded on a computer-readable recording medium and implemented by reading and executing the program recorded on the recording medium. The sound collection and detection device 10 and the acoustic signal analysis device 20 include hardware such as an operating system (OS) and peripheral devices. Examples of computer-readable recording media include portable media such as flexible disks, optical magnetic disks, read-only memories (ROMs), and compact disk read-only memories (CD-ROMs), as well as storage devices such as hard disks built into the sound collection and detection device 10 and the acoustic signal analysis device 20. Computer-readable recording media may also include those that dynamically store programs for a short period of time, such as communication lines when transmitting programs via networks like the Internet or communication lines like telephone lines, or those that store programs for a fixed period of time, such as volatile memory within a computer system that serves as a server or client in such cases. The above programs may also be programs that realize some of the functions described above, or may be programs that can realize the functions described above in combination with programs already stored in the computer system.

(Appendix 1)
an acoustic signal analyzer that performs frequency analysis of the acoustic data to calculate spectrum data of an object to be evaluated; an abnormal sound determination unit that determines whether the acoustic data includes an abnormal sound by comparing the spectrum data of the object to be evaluated in a predetermined frequency range with predetermined spectrum data in a normal state; and a position estimation unit that, if it is determined that an abnormal sound is included, estimates the position of occurrence of the abnormal sound based on the position information of the position where the acoustic data determined to include an abnormal sound was collected.

(Appendix 2)
the abnormal sound determination unit determines that the sound data includes an abnormal sound when a sound pressure level of the spectrum data to be evaluated in the predetermined frequency range is greater than a sound pressure level of the spectrum data in a normal state by a predetermined threshold or more.

(Appendix 3)
The abnormal sound detection system according to claim 1 or 2, wherein the abnormal sound determination unit determines that the sound data includes an abnormal sound when a frequency of a peak value of the spectrum data of the evaluation target in the predetermined frequency range differs from a frequency of a peak value of the spectrum data in a normal state.

(Appendix 4)
4. The abnormal sound detection system according to claim 1, wherein the position estimation unit estimates the position where the abnormal sound is generated by a triangulation method.

(Appendix 5)
5. The abnormal sound detection system according to claim 1, wherein the sound collection and detection device is provided in hearing protection equipment.

(Appendix 6)
an abnormal sound detection method comprising the steps of: receiving acoustic data of collected external sound and positional information of the position where the sound was collected; calculating spectrum data of an object to be evaluated by frequency analyzing the acoustic data; determining whether or not the acoustic data includes an abnormal sound by comparing the spectrum data of the object to be evaluated in a predetermined frequency range with predetermined spectrum data in a normal state; and, if it is determined in the determining step that an abnormal sound is included, estimating the position where the abnormal sound is occurring based on the positional information of the position where the sound was collected in the acoustic data determined to include the abnormal sound.

In addition, the components in the above-described embodiments may be replaced with well-known components as appropriate, provided that the spirit of the present invention is not deviated from. Furthermore, the technical scope of this invention is not limited to the above-described embodiments, and various modifications may be made without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 5: Wireless communication device 10: Sound collection/detection device 11: Sound collection unit 12: Communication unit 13: Positioning unit 20: Acoustic signal analysis device 21: Receiving unit 22: Input unit 23: Control unit 24: Acoustic signal analysis unit 25: Abnormal sound determination unit 26: Position estimation unit 27: Output unit 28: Storage unit 30: Display device 40: External storage device 100: Abnormal sound detection system

Claims (6)

A sound collection and detection device,
an acoustic signal analyzer;
Including,
The sound collection and detection device is
a sound collection unit that collects external sounds and outputs the collected sounds as acoustic data;
a positioning unit that measures position information of a position where the sound is collected;
a communication unit that transmits the acoustic data and the position information measured by the positioning unit;
and
The acoustic signal analysis device
a receiving unit that receives the acoustic data and the position information transmitted by the communication unit;
an acoustic signal analysis unit that performs frequency analysis on the acoustic data to calculate spectrum data of an object to be evaluated;
an abnormal sound determination unit that determines whether the acoustic data includes an abnormal sound by comparing spectrum data of the evaluation target in a predetermined frequency range with predetermined spectrum data in a normal state;
a position estimation unit that, when it is determined that an abnormal sound is included, estimates the position where the abnormal sound is generated based on the position information of the position where the acoustic data determined to include the abnormal sound was collected; and
An abnormal noise detection system having: the abnormal sound determination unit determines that the sound data includes an abnormal sound when a sound pressure level of the spectrum data to be evaluated in the predetermined frequency range is greater than a sound pressure level of the spectrum data in a normal state by a predetermined threshold or more.
The abnormal noise detection system according to claim 1 . the abnormal sound determination unit determines that the sound data includes an abnormal sound when a frequency of a peak value of the spectrum data to be evaluated in the predetermined frequency range is different from a frequency of a peak value of the spectrum data in a normal state.
The abnormal noise detection system according to claim 1 or 2. the position estimation unit estimates the generation position of the abnormal noise by a triangular positioning method;
The abnormal noise detection system according to claim 1 or 2. The sound collection and detection device is provided in hearing protection equipment.
The abnormal noise detection system according to claim 1 or 2. receiving acoustic data of collected external sounds and location information of the locations where the sounds are collected;
a step of performing frequency analysis on the acoustic data to calculate spectrum data of an object to be evaluated;
a step of determining whether or not the acoustic data includes an abnormal sound by comparing spectrum data of the evaluation target in a predetermined frequency range with predetermined spectrum data in a normal state;
a step of estimating a position where an abnormal sound is generated based on the position information of the position where the sound is collected in the acoustic data determined to include an abnormal sound in the determining step;
An abnormal sound detection method comprising: