Operation sound evaluation method for evaluating operation sound of rotary machine, and operation sound evaluation device therefor

Abstract

To provide a technology that can properly evaluate an operation sound of a rotary machine.SOLUTION: An operation sound evaluation method is configured to calculate an overall sound quality evaluation index, using at least a high-speed modulation sound evaluation index and dominant sound evaluation index. A calculation of the high-speed modulation sound evaluation index includes: performing an envelope analysis to the operation sound to determine intensity of a modulation frequency component; and multiplying the intensity of the modulation frequency component with a weighted coefficient of roughness. A calculation of the dominant sound evaluation index includes: respectively calculating a plurality of scores of a dominant level for a plurality of 1 / N octave components within a 1 / N octave band spectrum of the operation sound; and determining the dominant sound evaluation index from the plurality of scores.SELECTED DRAWING: Figure 19

Description

[Technical field]

[0001] The present invention relates to a technique for evaluating the operating sounds of rotating machines such as pumps, blowers, compressors, turbines, and electric motors. [Background technology]

[0002] The operating noise of rotating machinery such as pumps, blowers, compressors, turbines, and electric motors is generally expressed in noise levels in dB(A). However, even low-noise rotating machinery can be criticized for producing noise (abnormal sounds) that is loud, bothersome, or harsh.

[0003] Sound quality evaluation indices (e.g., loudness, sharpness, roughness, tonality, fluctuation intensity, etc.) for evaluating listening comfort have already been put to practical use. However, there are phenomena in which it is difficult to evaluate abnormal sounds of rotating machinery using these sound quality evaluation indices. For example, high-pitched sounds (sharpness) and humming sounds (fluctuation intensity) of rotating machinery can be evaluated using sound quality evaluation indices. In contrast, prominent sounds (prominent sounds) and high-speed modulated sounds (buzzing, buzzing, etc.) of rotating machinery are difficult to evaluate using the above sound quality evaluation indices.

[0004] Technologies for diagnosing abnormalities and failures in rotating machinery such as pumps using sound have already been put to practical use, such as narrowband analysis (FFT method) and octave band analysis. However, existing technologies diagnose failures using OA (Overall) or POA (Partial Overall), and failure diagnosis using sound quality evaluation indices that can directly express the characteristics of abnormal sounds from the above rotating machinery has not yet been put to practical use. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 10-253440 Summary of the Invention [Problem to be solved by the invention]

[0006] Therefore, the present invention provides a technique that can properly evaluate the operating sound of a rotating machine. [Means for solving the problem]

[0007] In one aspect, an operation noise evaluation method for evaluating an operation noise of a rotary machine includes calculating a comprehensive sound quality evaluation index using at least a high-speed modulation sound evaluation index and a protrusive sound evaluation index, and calculating the high-speed modulation sound evaluation index includes performing an envelope analysis on the operation noise to determine the intensity of a modulation frequency component of the operation noise and multiplying the intensity of the modulation frequency component by a roughness weighting coefficient, and calculating the protrusive sound evaluation index includes performing a 1 / N octave band analysis on the operation noise to generate a 1 / N octave band spectrum showing the relationship between the frequency and the sound pressure level for each 1 / N octave, setting a frame within the 1 / N octave band spectrum, and calculating the protrusive sound evaluation index based on the energy sum of a plurality of sound pressure levels of a plurality of 1 / N octave components included in the frame. calculating a reference level based on a plurality of sound pressure levels, calculating a plurality of prominence levels for the frame by subtracting the reference level from each of the plurality of sound pressure levels, determining a plurality of prominence levels for each of a plurality of frames in the 1 / N octave band spectrum by setting the frame, calculating the reference level, and calculating the plurality of prominence levels while shifting the frame by 1 / N octave at a time, calculating a plurality of scores for a plurality of 1 / N octave components in the 1 / N octave band spectrum based on the plurality of prominence levels determined for each of the plurality of frames, and determining the prominence sound evaluation index from the plurality of scores, wherein N is a number greater than 1.

[0008] In one aspect, calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index and the protrusive sound evaluation index means calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the sharpness of the operating sound. In one aspect, calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index and the protrusive sound evaluation index means calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operating sound. In one embodiment, the frame is a 1 / 1 octave band frame. In one embodiment, the reference level is the energy sum of the plurality of sound pressure levels minus a predetermined value. In one aspect, calculating the multiple scores for each of the multiple 1 / N octave components in the 1 / N octave band spectrum involves classifying the multiple prominence levels determined for each of the multiple frames into multiple groups for each 1 / N octave, and performing statistical processing on the multiple prominence levels in each group, thereby calculating the multiple scores for each of the multiple 1 / N octave components in the 1 / N octave band spectrum. In one aspect, the window is a window having a critical bandwidth, the critical bandwidth varying according to the frequency at which the window is set. In one aspect, the method further comprises generating an alarm signal when the overall sound quality assessment index falls below a lower limit or exceeds an upper limit.

[0009] In one aspect, an operation sound evaluation device for evaluating operation sounds of a rotating machine includes a microphone for acquiring the operation sounds of the rotating machine, and an index calculation device for calculating a comprehensive sound quality evaluation index using at least a high-speed modulated sound evaluation index and a protrusive sound evaluation index, the index calculation device is configured to perform an envelope analysis on the operation sound to determine the intensity of a modulation frequency component of the operation sound, and calculate the high-speed modulated sound evaluation index by multiplying the intensity of the modulation frequency component by a roughness weighting coefficient, and the index calculation device performs a 1 / N octave band analysis on the operation sound, generates a 1 / N octave band spectrum showing the relationship between frequency and sound pressure level for each 1 / N octave, sets a frame within the 1 / N octave band spectrum, and calculates a plurality of 1 / N octave components included in the frame. a reference level is calculated based on an energy sum of a plurality of sound pressure levels per frame, and a plurality of prominence levels is calculated for the frame by subtracting the reference level from each of the plurality of sound pressure levels; the setting of the frame, the calculation of the reference level, and the calculation of the plurality of prominence levels are performed while shifting the frame by 1 / N octave, thereby determining a plurality of prominence levels for each of a plurality of frames in the 1 / N octave band spectrum; a plurality of scores is calculated for a plurality of 1 / N octave components in the 1 / N octave band spectrum based on the plurality of prominence levels determined for each of the plurality of frames; and the prominence sound evaluation index is determined from the plurality of scores, where N is a number greater than 1.

[0010] In one aspect, the index calculation device is configured to calculate the overall sound quality evaluation index using at least the high-speed modulated sound evaluation index, the protrusive sound evaluation index, and the sharpness of the operating sound. In one aspect, the index calculation device is configured to calculate the overall sound quality evaluation index using at least the high-speed modulated sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operation sound. In one embodiment, the frame is a 1 / 1 octave band frame. In one aspect, the index calculation device is configured to calculate the reference level by subtracting a predetermined value from the energy sum of the plurality of sound pressure levels. In one embodiment, the index calculation device is configured to classify the multiple prominence levels determined for each of the multiple frames into multiple groups for each 1 / N octave, and perform statistical processing on the multiple prominence levels within each group, thereby calculating the multiple scores for each of the multiple 1 / N octave components in the 1 / N octave band spectrum. In one aspect, the window is a window having a critical bandwidth, the critical bandwidth varying according to the frequency at which the window is set. In one aspect, the index calculation device is configured to generate an alarm signal when the overall sound quality assessment index falls below a lower limit or exceeds an upper limit. Effect of the Invention

[0011] The comprehensive sound quality evaluation index can be used to properly evaluate abnormal sounds such as prominent sounds (prominent sounds) from rotating machinery and high-speed modulation sounds, which could not be evaluated using conventional sound quality evaluation indexes. [Brief description of the drawings]

[0012] [Figure 1] 1 is a schematic diagram showing an embodiment of an operation sound evaluation device for evaluating operation sounds of a rotating machine; [Diagram 2] 11 is a graph showing an example of an operation sound acquired by a microphone. [Diagram 3] 3 is a graph showing a waveform of an envelope obtained by performing an envelope analysis on the operation sound shown in FIG. 2. [Figure 4] This is a power spectrum obtained by applying Fourier transform processing to the envelope waveform shown in FIG. [Diagram 5] 11 is a graph showing an example of a relationship between a roughness weighting coefficient and a modulation frequency. [Figure 6] 1 is a flow chart illustrating an embodiment of calculating a protrusive sound evaluation index. [Figure 7] FIG. 13 is a diagram illustrating a generated 1 / 24 octave band spectrum. [Figure 8] FIG. 2 is an enlarged schematic diagram showing a portion of the generated 1 / 24 octave band spectrum. [Figure 9] FIG. 8 is a diagram illustrating a schematic diagram of a first frame set in the 1 / 24 octave band spectrum of FIG. [Figure 10] FIG. 2 is an enlarged schematic diagram of the first box in the 1 / 24 octave band spectrum. [Figure 11] FIG. 13 is a diagram showing a schematic diagram of the prominence level of the 1 / 24 octave component included in each frame. [Figure 12] FIG. 8 is a diagram illustrating a schematic diagram of a second frame set in the 1 / 24 octave band spectrum of FIG. 7. [Figure 13] FIG. 2 is an enlarged schematic diagram of the second box in the 1 / 24 octave band spectrum. [Figure 14] FIG. 13 is a diagram illustrating the prominence level of the 1 / 24 octave component. [Figure 15] FIG. 13 is a diagram illustrating the prominence level of the 1 / 24 octave component. [Figure 16] FIG. 8 is a diagram showing a schematic diagram of the 1 / 24 octave band spectrum shown in FIG. 7 together with the score S. [Figure 17] FIG. 1 is a diagram showing the relationship between frequency f and critical bandwidth Δf. [Figure 18] FIG. 1 is a diagram showing the relationship between frequency f and critical bandwidth Δf according to an approximation formula. [Figure 19] 11 is a graph showing an example of the relationship between an average preference level and a comprehensive sound quality evaluation index. [Figure 20] 13 is a graph showing another example of the relationship between the average preference level and the overall sound quality evaluation index. [Figure 21] 13 is a graph showing yet another example of the relationship between the average preference level and the overall sound quality evaluation index. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, an embodiment of an operation sound evaluation method and an operation sound evaluation device for evaluating the operation sound of a rotating machine will be described with reference to the drawings. 1 is a schematic diagram showing one embodiment of an operation sound evaluation device for evaluating the operation sound of a rotating machine. The operation sound evaluation device includes a microphone 1 that acquires the operation sound of a rotating machine 100, and an index calculation device 5 that calculates a comprehensive sound quality evaluation index from the operation sound of the rotating machine 100. The microphone 1 is electrically connected to the index calculation device 5, and the operation sound of the rotating machine 100 acquired by the microphone 1 is sent to the index calculation device 5. The rotating machine 100 is a machine having a rotating body, and examples of the rotating machine 100 include a pump, a blower, a compressor, a turbine, and an electric motor.

[0014] The index calculation device 5 includes a storage device 5a in which a program is stored, and a processing device 5b that executes calculations according to instructions included in the program. The index calculation device 5 includes at least one computer. The index calculation device 5 may include multiple computers, such as a combination of an edge computer and a cloud server. The storage device 5a includes a main storage device such as a random access memory (RAM) and an auxiliary storage device such as a hard disk drive (HDD) or a solid state drive (SSD). Examples of the processing device 5b include a CPU (central processing unit) and a GPU (graphic processing unit). However, the specific configuration of the index calculation device 5 is not limited to these examples.

[0015] The index calculation device 5 is configured to calculate a comprehensive sound quality evaluation index using a high-speed modulation sound evaluation index and a protrusive sound evaluation index. The high-speed modulation sound evaluation index is an index capable of evaluating high-speed modulation sounds (such as buzzing and popping). The protrusive sound evaluation index is an index capable of evaluating protrusive sounds (prominent sounds). The comprehensive sound quality evaluation index is an index for correctly evaluating the operating sounds of the rotating machine 100 (i.e., determining whether or not the operating sounds include abnormal sounds).

[0016] The index calculation device 5 is configured to calculate the high-speed modulated sound evaluation index as follows. That is, the index calculation device 5 performs an envelope analysis on the operation sound acquired from the microphone 1 to determine the intensity of the modulation frequency component of the operation sound. Furthermore, the index calculation device 5 multiplies the intensity of the modulation frequency component of the operation sound by a weighting coefficient of roughness.

[0017] Fig. 2 is a graph showing an example of an operation sound captured by the microphone 1, Fig. 3 is a graph showing an envelope waveform obtained by performing an envelope analysis on the operation sound shown in Fig. 2, and Fig. 4 is a power spectrum obtained by applying a Fourier transform process to the envelope waveform shown in Fig. 3. In the graphs of Fig. 2 and Fig. 3, the vertical axis represents the sound pressure level [dB] and the horizontal axis represents time. In the graph of Fig. 4, the vertical axis represents the strength of the modulation frequency component and the horizontal axis represents the modulation frequency.

[0018] As can be seen from Figures 2 to 4, the index calculation device 5 performs envelope analysis on the waveform of the operation sound shown in Figure 2 to generate the envelope waveform shown in Figure 3, and performs Fourier transform processing (specifically, fast Fourier transform processing) on ​​the envelope waveform to generate the power spectrum shown in Figure 4. Then, the index calculation device 5 determines the strength T1 of the modulation frequency component of the operation sound, which is the magnitude of the peak of the power spectrum, and the modulation frequency U1 of the operation sound, which is the frequency of the peak of the power spectrum.

[0019] Sound roughness is a sound quality index that is maximum when the modulation frequency is 70 Hz. Figure 5 is a graph showing the relationship between the roughness weighting coefficient and the modulation frequency. As shown in Figure 5, the roughness weighting coefficient is maximum when the modulation frequency is 70 Hz. This is because humans feel the roughness of the sound most when they hear a sound with a modulation frequency of 70 Hz.

[0020] The index calculation device 5 stores correlation data showing the relationship between the weighting coefficient of roughness and the modulation frequency shown in Fig. 5 in advance in the storage device 5a. The correlation data may be a polynomial function or a data table. The index calculation device 5 determines the weighting coefficient of roughness corresponding to the modulation frequency U1 of the operation sound from the correlation data shown in Fig. 5, and calculates the high-speed modulation sound evaluation index by multiplying the intensity T1 of the modulation frequency component of the operation sound by the determined weighting coefficient of roughness.

[0021] Next, the protrusive sound evaluation index will be described. Fig. 6 is a flowchart showing an embodiment for calculating the protrusive sound evaluation index. In the embodiment described below, 1 / 24 octave band analysis is used as an example of 1 / N octave band analysis. N is a number greater than 1. "1 / 24" described below is an example of "1 / N", and examples of N include natural numbers such as 3, 12, and 24.

[0022] First, the operating sound of a rotating machine (for example, a pump or a blower) 100 to be evaluated is picked up by the microphone 1 shown in Fig. 6 (step S1). As a result, the sound pressure level for each frequency is obtained.

[0023] Next, the index calculation device 5 performs 1 / 24 octave band analysis on the acquired operation sound to generate a 1 / 24 octave band spectrum (step S2).

[0024] Fig. 7 is a diagram showing a schematic diagram of the generated 1 / 24 octave band spectrum. Fig. 8 is a diagram showing a schematic diagram of an enlarged portion of the generated 1 / 24 octave band spectrum. As shown in Fig. 8, the 1 / 24 octave band spectrum shows sound pressure levels P1, P2, P3, etc. at frequencies f1, f2, f3, etc. for each 1 / 24 octave.

[0025] 6, the i-th 1 / 1 octave band frame (hereinafter, simply referred to as "frame") is set in the generated 1 / 24 octave band spectrum (step S3A). The first frame is set so that its lower end (lowest frequency) includes a desired frequency (for example, the lowest frequency of the acquired operation sound).

[0026] Fig. 9 is a diagram showing a schematic diagram of a first frame set in the 1 / 24 octave band spectrum of Fig. 7. Fig. 10 is a diagram showing a schematic diagram of an enlarged view of the first frame of the 1 / 24 octave band spectrum.

[0027] In order to set a frame with a width of 1 / 1 octave band in the 1 / 24 octave band spectrum, the frame contains 24 1 / 24 octave components. Hereinafter, the frequencies corresponding to the 24 1 / 24 octave components contained in this first frame are referred to as frequencies f1 to f24, in order from the lowest frequency side.

[0028] Returning to Fig. 6, the index calculation device 5 calculates a reference level for the set first frame (step S3B). The reference level is calculated based on a plurality of sound pressure levels of a plurality of 1 / 24 octave components included in the frame. In the following description, the sound pressure level of the 1 / 24 octave component corresponding to the frequency fk (k = 1 to 24) is represented as Pk, and the reference level for the i-th frame is represented as Ai.

[0029] If the sound pressure levels P1 to P24 are the same, the sound pressure level (energy) of each 1 / 24 octave component will be 1 / 24 of the sound pressure level of 1 / 1 octave (equivalent to -13.8 dB). Therefore, as an example, the reference level A1 for the first frame can be calculated by the following formula (1). Reference level A1 = 10 × log(Σ10 Pk / 10 )-10×log24 =10×log(Σ10 Pk / 10 )-13.8 ···(1)

[0030] That is, the index calculation device 5 calculates the sum of the energies of the sound pressure levels P1 to P24 of the 1 / 24 octave components f1 to f24 included in the frame, 10×logΣ10 Pk / 10 The reference level A1 is determined by subtracting a predetermined value (13.8) from the reference level A1.

[0031] Next, the index calculation device 5 calculates the prominence level for each of the 24 1 / 24 octave components included in the first frame (step S3C). In the following description, the prominence level of the 1 / 24 octave component corresponding to the frequency fk in the i-th frame is represented as Qk_i.

[0032] As an example, the index calculation unit 5 calculates the prominence level Qk_1 of the 1 / 24 octave component corresponding to the frequency fk (k=1 to 24) in the first frame by the following formula (2). Qk_1 = Pk - reference level A1 (2)

[0033] In other words, the prominence level of each 1 / 24 octave component included in the set frame is the sound pressure level of each 1 / 24 octave component minus the reference level.

[0034] This completes the process for the set first frame. Fig. 11 diagrammatically shows prominence levels Q1_1 to Q24_1 of 1 / 24 octave components corresponding to frequencies f1 to f24, respectively, in the first frame.

[0035] When the sound pressure level of the 1 / 24 octave component corresponding to a certain frequency fk is high, if there are other 1 / 24 octave components with high sound pressure levels in the first box, the prominence level Qk_1 will be relatively low. This is because the reference level A1 will be high. In this way, when there are other frequencies with high sound pressure levels around the peak frequency fk, the prominence level Qk_1 that reflects the fact that the sound of the frequency fk is not heard very prominently will be obtained.

[0036] On the other hand, even if the sound pressure level of the 1 / 24 octave component corresponding to the frequency fk is high, if there are no other 1 / 24 octave components with high sound pressure levels in the first frame, the prominence level Qk_1 will be relatively high. This is because the reference level A1 will be low. In this way, if there are no other frequencies with high sound pressure levels around the peak frequency fk, the prominence level Qk_1 will be obtained, which reflects that the sound of the frequency fk sounds more prominent.

[0037] In this embodiment, the reference level A1 is determined by the sound pressure levels of the multiple 1 / 24 octave components included in the frame. Therefore, the prominence level Qk_1 incorporates the influence of the sound pressure levels of the surrounding bands and reflects the auditory masking effect.

[0038] Next, the index calculation device 5 shifts the first frame by 1 / 24 octave, and sets the next second frame to the 1 / 24 octave band spectrum (YES in step S3D, S3A).

[0039] Fig. 12 is a diagram showing a schematic diagram of the second frame set in the 1 / 24 octave band spectrum of Fig. 7. Fig. 13 is a diagram showing a schematic diagram of an enlarged second frame of the 1 / 24 octave band spectrum. Comparing Fig. 10 with Fig. 13, since the frame is shifted by 1 / 24 octave, the 1 / 24 octave component corresponding to the frequency f1 included in the first frame is not included in the second frame, and the 1 / 24 octave component corresponding to the frequency f25 not included in the first frame is included in the second frame.

[0040] Then, the index calculation device 5 calculates the reference level A2 for the second frame (step S3B). The reference level A2 can be calculated by the following formula (3), similar to the reference level A1 for the first frame. Reference level A2 = 10 × log(Σ10 Pk / 10 )-10×log24 =10×log(Σ10 Pk / 10 )-13.8 ···(3)

[0041] The calculation formula itself is the same as the reference level A1 in the first frame. However, the reference level may vary depending on the sound pressure levels of the multiple 1 / 24 octave components included in the frame. Specifically, the higher the overall sound pressure levels of the multiple 1 / 24 octave components included in the frame, the higher the reference level will be.

[0042] Next, the index calculation unit 5 calculates a plurality of prominence levels for the 24 1 / 24 octave components included in the second frame (step S3C). In the second frame, the prominence level Qk_2 of the 1 / 24 octave component corresponding to the frequency fk (k=2 to 25) can be calculated by the following formula (4). Qk_2=Pk-Reference level A2...(4)

[0043] This completes the process for the set second frame. Fig. 11 shows schematic diagrams of prominence levels Q2_2 to Q25_2 of the 1 / 24 octave components corresponding to the frequencies f2 to f25, respectively.

[0044] The index calculation device 5 performs the processes of steps S3A to S3D while shifting the frame by 1 / 24 octave at a time until the upper end (maximum frequency) of the set frame reaches a desired frequency (for example, the maximum frequency of the acquired operation noise), thereby determining multiple prominence levels for each of multiple frames in the 1 / 24 octave band spectrum.

[0045] In this way, the prominence level of each 1 / 24 octave component is calculated for each frame. Because the reference level is calculated for each frame, the prominence level of the 1 / 24 octave component in one frame may differ from the prominence level of the 1 / 24 octave component in another frame. This will be explained using Figures 14 and 15.

[0046] For example, as shown in Fig. 14, a certain m-th frame includes a frequency fx. In the m-th frame, the peak frequency is fx, and since there is no frequency with a high sound pressure level in the m-th frame, the reference level Am is a small value. Therefore, the prominence level Qx_m of the 1 / 24 octave component corresponding to the frequency fx is a relatively large value. In this way, since there are no other frequencies with high sound pressure levels around the peak frequency fx, a prominence level Qx_m is obtained that indicates that the sound of this frequency fx sounds more prominent.

[0047] On the other hand, as shown in Fig. 15, another n-th frame also includes frequency fx. However, in this n-th frame, there are multiple peaks with high sound pressure levels at multiple frequencies including fx, so the reference level An is a large value. Therefore, the prominent level Qx_n of the 1 / 24 octave component corresponding to the frequency fx is a relatively small value.

[0048] In this manner, in this embodiment, the index calculation device 5 calculates a prominence level that reflects the sound pressure levels of other frequencies included in the frame (that is, in the surrounding bands).

[0049] Returning to FIG. 6, the index calculation device 5 calculates the prominence levels for all frames, and then calculates a plurality of scores for the plurality of 1 / 24 octave components in the 1 / 24 octave band spectrum based on the plurality of prominence levels of the plurality of 1 / 24 octave components calculated in each of the 1 / 1 octave band frames (step S4). More specifically, as shown in FIG. 11, the index calculation device 5 classifies the plurality of prominence levels determined for each of the plurality of frames into a plurality of groups G (indicated by dotted lines) for each 1 / 24 octave. The index calculation device 5 performs statistical processing on the plurality of prominence levels in each group G to calculate a plurality of scores for the plurality of 1 / 24 octave components in the 1 / 24 octave band spectrum. Examples of statistical processing include summation, arithmetic mean, and energy mean. In the following description, the score of frequency fk is represented as sk.

[0050] For example, as shown in Fig. 11, the 1 / 24 octave components corresponding to group G of frequency f24 are included in the 1st to 24th boxes, and prominence levels Q24_1 to Q24_24 are calculated. Therefore, the score S24 of the 1 / 24 octave components corresponding to group G of frequency f24 is obtained by statistically processing (for example, summation, arithmetic average, or energy average) the prominence levels Q24_1 to Q24_24. The score obtained in this way indicates the prominence of the sound reflecting the sound pressure levels of the surrounding bands.

[0051] Then, the index calculation device 5 graphs the scores as necessary (step S5). FIG. 16 is a diagram showing the score S in schematic form together with the 1 / 24 octave band spectrum shown in FIG. 7. This graph shows a plurality of scores corresponding to a plurality of 1 / 24 octave components in the 1 / 24 octave band spectrum. These scores allow, for example, an operator to easily identify frequencies that may be the cause of unpleasant noise. As a specific example, frequencies with large scores can be determined to be the frequencies that cause unpleasant noise.

[0052] The index calculation device 5 determines a protrusive sound evaluation index from a plurality of scores corresponding to a plurality of 1 / 24 octave components in the 1 / 24 octave band spectrum (step S6). For example, the index calculation device 5 calculates a protrusive sound evaluation index that is a sum or average of the plurality of scores. In another example, the index calculation device 5 determines a protrusive sound evaluation index that is a maximum value of the plurality of scores.

[0053] As described above, in this embodiment, a frame having a width of 1 / 1 octave band is set for the 1 / 24 octave band spectrum, and a reference level is calculated for each frame. Then, based on this reference value, the prominence level of each 1 / 24 octave component is calculated. Therefore, it is possible to calculate a score for each 1 / 24 octave component that reflects the sound pressure level of the surrounding band (i.e., reflects the auditory masking effect). Based on this score, it is possible to easily identify frequencies that may cause unpleasant sounds.

[0054] Next, another embodiment for calculating the prominence sound evaluation index will be described. Details of this embodiment that are not particularly described are the same as those of the embodiment described with reference to FIGS. 6 to 16, and thus redundant descriptions thereof will be omitted.

[0055] In this embodiment, in step S3A of FIG. 6, instead of setting a 1 / 1 octave band frame, a frame based on the critical bandwidth Δf is set. The critical bandwidth Δf corresponds to the frequency range of auditory masking. The relationship between the frequency f and the critical bandwidth Δf is predetermined as shown in FIG. 17, for example, and the critical bandwidth Δf increases non-linearly as the frequency f increases. Therefore, if i < i', the frame set at the i'-th position has a wider critical bandwidth than the frame set at the i-th position. The relationship between the frequency f and the critical bandwidth Δf is stored in advance in the storage device 5a of the index calculation device 5.

[0056] For example, in FIG. 10, the first frame is set for the frequency f1. As an example, if the i-th frame is set for the frequency 50 Hz, referring to FIG. 17, the critical bandwidth of that frame is 100 Hz. On the other hand, if the i'-th frame is set for the frequency 1,000 Hz, referring to FIG. 17, the critical bandwidth of that frame is 160 Hz.

[0057] However, in FIG. 17, the relationship between the frequency f and the critical bandwidth Δf is defined discretely. Therefore, depending on the set frequency, the critical bandwidth Δf may not be predetermined. For example, the critical bandwidth Δf corresponding to the frequency 900 Hz is not defined in FIG. 17.

[0058] Therefore, an approximate expression (for example, a polynomial) showing the relationship between the frequency f and the critical bandwidth Δf may be prepared. Then, the index calculation device 5 may calculate the critical bandwidth Δf corresponding to the set frequency f and set the frame. The approximate expression showing the relationship between the frequency f and the critical bandwidth Δf is stored in advance in the storage device 5a of the index calculation device 5.

[0059] The approximation formula may be, for example, the following cubic formula: The relationship between frequency f and critical bandwidth Δf based on this approximation formula is shown in FIG. Δf=a×f 3 +b×f 2 +c×f+d a=-0.0000000003 b=0.0000173187 c=0.0831250103 d=76.1402217580

[0060] In the embodiment described above, a 1 / 1 octave band frame is set, so that the frame contains 24 1 / 24 octave components. In contrast, in this embodiment, a frame is set based on a critical bandwidth Δf that increases with increasing frequency, so that the number of 1 / 24 octave components contained in the frame is not constant, but increases with increasing frequency.

[0061] Even in this case, the reference level Ai in step S3B of Fig. 6 can be calculated in the same manner as in the embodiment described above. Specifically, if the number of 1 / 24 octave components included in the i-th frame is Mi, the reference level Ai is expressed by the following formula. Reference level Ai = 10 × log(Σ10 Pk / 10 )-10×log(Mi)

[0062] 6, the index calculation unit 5 calculates the prominence level for each of the Mi 1 / 24 octave components included in the frame. The other processes may be the same as those in the embodiment described above.

[0063] In this manner, in the present embodiment, a frame having a critical bandwidth is set, so that frequencies that may cause unpleasant sounds can be easily identified with greater accuracy. Furthermore, since a frame based on a critical bandwidth that corresponds to the frequency range of auditory masking is set, accuracy is improved.

[0064] Next, an embodiment will be described in which a comprehensive sound quality evaluation index is calculated using the high-speed modulated sound evaluation index and the protrusive sound evaluation index described above. As described above, the high-speed modulated sound evaluation index is an index obtained by multiplying the envelope analysis result by a roughness weighting coefficient, and is an index that can evaluate the roughness of high-speed modulated sound that can become a noise problem in a rotating machine. The protrusive sound evaluation index is an index that can evaluate a protrusive sound (prominent sound) included in the operating sound of a rotating machine.

[0065] The index calculation device 5 is configured to calculate the overall sound quality evaluation index using the following calculation formula (5) which includes the high-speed modulated sound evaluation index and the protrusive sound evaluation index as variables. Overall sound quality evaluation index = E1 x h + E2 x p + C (5) Here, h represents a high-speed modulated sound evaluation index, p represents a protrusive sound evaluation index, E1 and E2 represent predetermined coefficients, and C represents a predetermined constant. The calculation formula (5) is stored in advance in the storage device 5a of the index calculation device 5.

[0066] The index calculation device 5 determines the coefficients E1, E2 and constant C by performing multiple regression analysis on the average preference level, which is the result of the sensory evaluation of the operating sound of the rotating machine, and the high-speed modulation sound evaluation index and the protrusive sound evaluation index calculated for the operating sound. The average preference level is the objective variable, and the high-speed modulation sound evaluation index and the protrusive sound evaluation index are explanatory variables. The average preference level of the operating sound of the rotating machine is obtained by previously conducting a sensory test on the operating sound of the rotating machine. The average preference level is previously stored in the storage device 5a of the index calculation device 5.

[0067] 19 is a graph showing an example of the relationship between the average preference score and the overall sound quality evaluation index calculated by the formula (5). The vertical axis of the graph represents the average preference score, and the horizontal axis of the graph represents the overall sound quality evaluation index. The coefficient of determination R 2 is an index showing how close the overall sound quality evaluation index calculated by formula (5) is to the average preference score. The coefficient of determination R 2 The closer to 1, the closer the overall sound quality evaluation index is to the average preference score.

[0068] In one embodiment, the index calculation device 5 may be configured to calculate the overall sound quality evaluation index using the following calculation formula (6), which includes the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the sharpness of the operating sound of the rotating machine as variables. Overall sound quality evaluation index = E1 x h + E2 x p + E3 x s + C (6) Here, h represents the high-speed modulation sound evaluation index, p represents the prominence sound evaluation index, s represents the sharpness of the operation sound of the rotating machine, E1, E2, and E3 represent predetermined coefficients, and C represents a predetermined constant. The calculation formula (6) is stored in advance in the storage device 5a of the index calculation device 5.

[0069] Generally, sound sharpness is a sound quality index that indicates the pitch of a sound. The sharpness of the operating sound of a rotating machine is calculated according to a known technique, and therefore a detailed description of the calculation of the sharpness is omitted.

[0070] The index calculation device 5 performs multiple regression analysis on the average preference level, which is the result of the sensory evaluation of the operating sound of the rotating machine, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the sharpness of the operating sound, which are calculated for the operating sound, to determine the coefficients E1, E2, E3, and the constant C. The average preference level is the objective variable, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the sharpness of the operating sound are explanatory variables.

[0071] Fig. 20 is a graph showing an example of the relationship between the average preference score and the overall sound quality evaluation index calculated by formula (6). The vertical axis of the graph represents the average preference score, and the horizontal axis of the graph represents the overall sound quality evaluation index. As can be seen from a comparison between Fig. 19 and Fig. 20, the coefficient of determination R obtained when formula (6) is used is 2 is the coefficient of determination R obtained using formula (5) 2 Therefore, the overall sound quality evaluation index calculated by formula (6) is closer to 1 than the average preference score, which is the result of a sensory evaluation of the operating sound of a rotating machine.

[0072] In one embodiment, the index calculation device 5 may be configured to calculate the overall sound quality evaluation index using the following calculation formula (7), which includes the high-speed modulation sound evaluation index, the protrusive sound evaluation index, the sharpness of the operation sound of the rotating machine, and the fluctuation intensity of the operation sound of the rotating machine as variables. Overall sound quality evaluation index = E1 x h + E2 x p + E3 x s + E4 x v + C (7) Here, h represents the high-speed modulation sound evaluation index, p represents the protrusive sound evaluation index, s represents the sharpness of the rotating machine operation sound, v represents the fluctuation intensity of the rotating machine operation sound, E1, E2, E3, and E4 represent predetermined coefficients, and C represents a predetermined constant. The calculation formula (7) is stored in advance in the storage device 5a of the index calculation device 5.

[0073] Generally, the sound fluctuation intensity is a sound quality index that indicates the fluctuation of the sound, and is maximum when the modulation frequency is 4 Hz. The fluctuation intensity of the operation sound of the rotating machine is calculated according to a known technique, so a detailed description of the calculation of the fluctuation intensity is omitted.

[0074] The index calculation device 5 performs multiple regression analysis on the average preference level, which is the result of the sensory evaluation of the operating sound of the rotating machine, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, the sharpness of the operating sound, and the fluctuation intensity of the operating sound, which are calculated for the operating sound, to determine the coefficients E1, E2, E3, E4, and a constant C. The average preference level is the objective variable, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, the sharpness of the operating sound, and the fluctuation intensity of the operating sound are explanatory variables.

[0075] Fig. 21 is a graph showing an example of the relationship between the average preference score and the overall sound quality evaluation index calculated by formula (7). The vertical axis of the graph represents the average preference score, and the horizontal axis of the graph represents the overall sound quality evaluation index. As can be seen from a comparison between Fig. 20 and Fig. 21, the coefficient of determination R obtained when formula (7) is used is 2 is the coefficient of determination R obtained using formula (6) 2 Therefore, the overall sound quality evaluation index calculated by formula (7) is closer to 1 than the average preference score, which is the result of a sensory evaluation of the operating sound of a rotating machine.

[0076] In one embodiment, the index calculation device 5 may be configured to calculate the overall sound quality evaluation index using the following calculation formula (8), which includes the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operation sound of the rotating machine as variables. Overall sound quality evaluation index = E1 x h + E2 x p + E4 x v + C (8) Here, h represents the high-speed modulation sound evaluation index, p represents the protrusive sound evaluation index, v represents the fluctuation intensity of the operation sound of the rotating machine, E1, E2, and E4 represent predetermined coefficients, and C represents a predetermined constant. The calculation formula (8) is stored in advance in the storage device 5a of the index calculation device 5.

[0077] The index calculation device 5 performs multiple regression analysis on the average preference level, which is the result of the sensory evaluation of the operating noise of the rotating machine, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operating noise, which are calculated for the operating noise, to determine the coefficients E1, E2, E4, and the constant C. The average preference level is the objective variable, and the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operating noise are explanatory variables.

[0078] Although not shown, similarly to formulas (5) to (7), the overall sound quality evaluation index calculated by formula (8) approximates the average preference degree, which is the result of a sensory evaluation of the operating sound of a rotating machine.

[0079] In one embodiment, the index calculation device 5 is configured to generate an alarm signal when the overall sound quality evaluation index calculated by any one of the calculation formulas (5) to (8) falls below a lower limit value. When the overall sound quality evaluation index drops significantly, some abnormality may occur in the rotating machine. Examples of the alarm signal are not particularly limited, but include, for example, an electrical signal for displaying on a display device a message informing the user that an abnormality has occurred in the rotating machine.

[0080] Depending on the type of rotating machine or the type of abnormality, the occurrence of an abnormality in the rotating machine may cause the overall sound quality evaluation index to increase significantly. Therefore, in one embodiment, the index calculation device 5 may be configured to generate an alarm signal when the overall sound quality evaluation index calculated by any one of the calculation formulas (5) to (8) exceeds an upper limit value. Furthermore, in one embodiment, the index calculation device 5 may be configured to compare the overall sound quality evaluation index calculated by any one of the calculation formulas (5) to (8) with a lower limit value and an upper limit value, respectively, and generate an alarm signal when the overall sound quality evaluation index falls below the lower limit value or exceeds the upper limit value.

[0081] The index calculation device 5 can calculate a comprehensive sound quality evaluation index that can directly capture abnormal phenomena caused by relatively high frequencies that do not contribute to OA (Overall) or POA (Partial Overall) and abnormal phenomena accompanied by modulated sounds that do not appear in OA or POA. Therefore, the index calculation device 5 can properly judge abnormalities in the rotating machine based on the comprehensive sound quality evaluation index.

[0082] The above-described embodiments have been described for the purpose of enabling a person having ordinary skill in the art to practice the present invention. Various modifications of the above-described embodiments are naturally possible for a person skilled in the art, and the technical idea of ​​the present invention can be applied to other embodiments. Therefore, the present invention is not limited to the described embodiments, but is to be interpreted in the broadest scope according to the technical idea defined by the claims. [Explanation of symbols]

[0083] 1 microphone 5 Indicator calculation device 100 Rotating Machinery

Claims

1. An operation noise evaluation method for evaluating an operation noise of a rotary machine, comprising the steps of: Calculate a comprehensive sound quality evaluation index using at least the high-speed modulation sound evaluation index and the prominence sound evaluation index; The calculation of the high-speed modulated sound evaluation index includes performing an envelope analysis on the operation sound to determine the intensity of a modulation frequency component of the operation sound, and multiplying the intensity of the modulation frequency component by a roughness weighting coefficient; The calculation of the protrusive sound evaluation index is A 1 / N octave band analysis is performed on the operation sound to generate a 1 / N octave band spectrum showing the relationship between frequency and sound pressure level for each 1 / N octave; A frame is set within the 1 / N octave band spectrum, Calculating a reference level based on an energy sum of a plurality of sound pressure levels of a plurality of 1 / N octave components included in the frame; calculating a plurality of salience levels for the frame by subtracting the reference level from each of the plurality of sound pressure levels; The setting of the frame, the calculation of the reference level, and the calculation of the plurality of prominence levels are performed while shifting the frame by 1 / N octave, thereby determining a plurality of prominence levels for each of the plurality of frames in the 1 / N octave band spectrum; Calculating a plurality of scores for a plurality of 1 / N octave components in the 1 / N octave band spectrum based on the plurality of prominence levels determined for each of the plurality of frames; determining the prominence assessment index from the plurality of scores; A method wherein N is a number greater than one.

2. Calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index and the protrusive sound evaluation index is calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the sharpness of the driving sound. The driving sound evaluation method according to claim 1.

3. Calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index and the protrusive sound evaluation index is calculating the overall sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protrusive sound evaluation index, and the fluctuation intensity of the operation sound. The driving sound evaluation method according to claim 1.

4. The driving noise evaluation method according to claim 1 , wherein the frame is a 1 / 1 octave band frame.

5. The driving noise evaluation method according to claim 1 , wherein the reference level is a value obtained by subtracting a predetermined value from the energy sum of the plurality of sound pressure levels.

6. Calculating the scores for each of the 1 / N octave components in the 1 / N octave band spectrum includes: The plurality of prominence levels determined for each of the plurality of frames are classified into a plurality of groups for each 1 / N octave; The driving noise evaluation method according to claim 1, wherein the plurality of scores are calculated for the plurality of 1 / N octave components in the 1 / N octave band spectrum by performing statistical processing on the plurality of prominence levels in each group.

7. The driving noise evaluation method according to claim 1 , wherein the frame is a frame having a critical bandwidth, and the critical bandwidth varies according to a frequency at which the frame is set.

8. The driving sound evaluation method according to claim 1 , further comprising: generating an alarm signal when the comprehensive sound quality evaluation index falls below a lower limit value or exceeds an upper limit value.

9. An operation sound evaluation device for evaluating an operation sound of a rotary machine, comprising: A microphone for acquiring an operation sound of the rotating machine; An index calculation device is provided for calculating a comprehensive sound quality evaluation index by using at least a high-speed modulation sound evaluation index and a protrusive sound evaluation index; The index calculation device is configured to perform an envelope analysis on the operation sound to determine the intensity of a modulation frequency component of the operation sound, and to calculate the high-speed modulation sound evaluation index by multiplying the intensity of the modulation frequency component by a roughness weighting coefficient; The index calculation device is A 1 / N octave band analysis is performed on the operation sound to generate a 1 / N octave band spectrum showing the relationship between frequency and sound pressure level for each 1 / N octave; A frame is set within the 1 / N octave band spectrum, Calculating a reference level based on an energy sum of a plurality of sound pressure levels of a plurality of 1 / N octave components included in the frame; calculating a plurality of salience levels for the frame by subtracting the reference level from each of the plurality of sound pressure levels; The setting of the frame, the calculation of the reference level, and the calculation of the plurality of prominence levels are performed while shifting the frame by 1 / N octave, thereby determining a plurality of prominence levels for each of the plurality of frames in the 1 / N octave band spectrum; Calculating a plurality of scores for a plurality of 1 / N octave components in the 1 / N octave band spectrum based on the plurality of prominence levels determined for each of the plurality of frames; determining the prominence evaluation index from the plurality of scores; An operating noise evaluation device, wherein N is a number greater than 1.

10. The driving sound evaluation device according to claim 9, wherein the index calculation device is configured to calculate the comprehensive sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protruding sound evaluation index, and the sharpness of the driving sound.

11. The driving sound evaluation device according to claim 9, wherein the index calculation device is configured to calculate the comprehensive sound quality evaluation index using at least the high-speed modulation sound evaluation index, the protruding sound evaluation index, and the fluctuation intensity of the driving sound.

12. The driving sound evaluation device according to claim 9 , wherein the frame is a 1 / 1 octave band frame.

13. The driving sound evaluation device according to claim 9 , wherein the index calculation device is configured to calculate the reference level by subtracting a predetermined value from the energy sum of the plurality of sound pressure levels.

14. The index calculation device is The plurality of prominence levels determined for each of the plurality of frames are classified into a plurality of groups for each 1 / N octave; The driving sound evaluation device according to claim 9, configured to calculate the scores for the plurality of 1 / N octave components in the 1 / N octave band spectrum by performing statistical processing on the plurality of prominence levels in each group.

15. The driving sound evaluation device according to claim 9 , wherein the frame is a frame having a critical bandwidth, and the critical bandwidth varies according to a frequency at which the frame is set.

16. The driving sound evaluation device according to claim 9 , wherein the index calculation device is configured to generate an alarm signal when the comprehensive sound quality evaluation index falls below a lower limit value or exceeds an upper limit value.

Images