JP6605409B2 - Sound inspection device - Google Patents

Description

  The present invention relates to a hammering inspection apparatus.

  2. Description of the Related Art A hitting test is performed on a cast or the like based on an acoustic signal obtained by hitting a cast or the like with a hammer. When performing a hammering test, pass / fail judgment is performed by extracting a value of a predetermined characteristic parameter from the hammering sound data and comparing it with a characteristic parameter value of a healthy product that has been measured in advance. . As examples of the characteristic parameter, the convergence time of the amplitude of the sound waveform and the maximum peak frequency of the sound frequency spectrum are known (see Patent Document 1).

JP 2010-261816 A

  When the above-described method described in Patent Document 1 is used, a determination threshold for determining how far the characteristic parameter value of a sound product as a reference deviates from that is determined as a defective product. It was necessary to set values in advance. This determination threshold value must be set appropriately in consideration of the variation range of characteristic parameter values that differ for each type of product to be inspected.

According to the hammering inspection apparatus of the present invention, based on the sounding data of a plurality of sound products, a variation allowable width setting unit that sets a variation allowable width allowed as a sound product, and the sounding data of the inspection target product are the variation. A determination unit that determines whether or not the inspection target product is a healthy product by determining whether or not the product is within an allowable range, and the sounding data includes a time-series amplitude waveform. The variation allowable width setting unit extracts each frequency spectrum from the amplitude waveform data collected by the plurality of healthy products, and sets the variation allowable width from the extracted contour line of the frequency spectrum. To do .

  According to the present invention, it is possible to simplify the work of presetting a judgment threshold value for sound hit inspection.

It is a system configuration figure of a hammering inspection device. It is a figure which shows the production | generation screen of the healthy goods dispersion | variation range. It is a flowchart which produces | generates the healthy product dispersion | variation range. It is a figure which shows the metadata table of sound goods hitting sound data group. It is a figure which shows the amplitude waveform chart of the sound goods hitting sound data group. It is a figure which shows the procedure of an outline extraction. It is a figure which shows the procedure which extracts the upper limit value and lower limit value of an outline. It is a figure which shows a defect presence / absence determination screen. It is a flowchart which shows the process sequence of defect presence / absence determination. It is a figure which shows the metadata table of historical data. It is a figure which shows the raw data table of a tapping sound. It is a system block diagram of the sound-inspection apparatus in 2nd Embodiment. It is a figure which shows the production | generation screen of the healthy product dispersion | variation range in 2nd Embodiment. It is a flowchart which produces | generates the healthy product dispersion | variation range in 2nd Embodiment. It is a figure which shows a frequency spectrum and its outline. It is a figure which shows the procedure of an outline extraction. It is a figure which shows the procedure which extracts the upper limit curve of an outline. It is a figure which shows the defect presence / absence determination screen in 2nd Embodiment. It is a flowchart which shows the process sequence of the defect presence determination in 2nd Embodiment. It is a figure which shows the production | generation screen of the healthy product dispersion | variation range in 3rd Embodiment. It is a figure which shows the defect presence / absence determination screen in 3rd Embodiment.

(First embodiment)
FIG. 1 is a system configuration diagram of a hammering inspection apparatus according to the present embodiment. In the present embodiment, an example in which the hammering inspection apparatus is realized by a computer system is shown.

  As shown in FIG. 1, the sound inspection device 1 includes a sound inspection computer 100, an impact control signal output device 200, and an A / D converter 300. The computer 100 for hammering test includes an arithmetic processing device 101, a display device 102, and a storage device 103. The storage device 103 includes a history data storage unit 103a and a healthy product variation range storage unit 103b.

  The sound inspection device 1 collects sound data (time change of sound pressure) when the metal product 2 is struck by the sound device 3 using the microphone 4 and converts it into a digital value by the A / D converter 300. And get. Further, the shape dimension data of the metal product 2 is acquired from the shape dimension measuring device 8. The shape dimension measuring instrument 8 measures the dimensions such as the angle and the shape of the metal product 2 with a necessary measurement accuracy by, for example, a projection method, a laser scanning method, or an image method. In the present embodiment, shape dimension data will be described as an example of the sound variation factor. Further, the hammering test apparatus 1 collects product data of the metal product 2 from an external manufacturing execution system (MES) 5 via the network 7.

  Prior to performing the hammering inspection of the metal product 2, the hammering inspection apparatus 1 first uses the hammering sound data groups of the plurality of metal products 2 that have been confirmed to be healthy products in advance. A “sound product variation range” representing the variation range of the hit sound data is generated and stored in association with the product type data and the shape dimension data in the sound product variation range storage unit 103b. The metal product 2 in which the hitting data is used to generate the healthy product variation range may be a healthy product other than the hitting test by the hitting test apparatus 1 (offline inspection, manual sensory test, etc.). It has been confirmed in advance. A method for generating the healthy product variation range will be described in detail later.

  When performing the hammering inspection of the metal product 2, the hammering inspection apparatus 1 stores the healthy product variation range associated with the product type data and shape dimension data to which the metal product 2 to be inspected belongs from the healthy product variation range storage unit 103b. Extract. And the presence or absence of the defect of the metal product 2 is determined by determining whether or not the outline of the hitting data obtained for the metal product 2 to be inspected deviates from the healthy product variation range. The sound inspection device 1 transmits a determination result to a quality control system (QCS) 6 and displays the determination result on the display device 102 as necessary.

  The hammering inspection apparatus 1 associates the hammering sound data, product type data, shape data, and determination data obtained by the hammering test on various metal products 2 with each other and stores them in the history data storage unit 103a. In addition, the hammering inspection apparatus 1 extracts these data obtained by past hammering inspections from the history data storage unit 103a as necessary, and among them, a plurality of metal products that are determined to be healthy products. The sound product variation range can be regenerated using the second hit sound data group. Using the newly generated healthy product variation range, the sound hitting inspection apparatus 1 updates the healthy product variation range stored in the healthy product variation range storage unit 103b until then.

-Sound product variation range generation screen-
FIG. 2 is a diagram illustrating an example of a sound quality variation range generation screen displayed on the display device 102. Note that the display device 102 includes a touch panel on the surface thereof, and also has a function as an input device for the arithmetic processing device 101. As shown in FIG. 2, the healthy product variation range generation screen includes a product specification box 191A, a file specification box 192A, a shape dimension classification specification box 199A, a healthy product variation range generation button 193, and a healthy product variation range display field for amplitude waveforms. 196A.

-Process for generating healthy product variation range-
FIG. 3 is a flowchart showing a processing procedure for generating a healthy product variation range by the hammering inspection apparatus 1. The process shown in this flowchart shows the process executed by the arithmetic processing unit 101 of the computer 100 for hammering test.

  When the healthy product variation range generation button 193 is designated on the generation screen of FIG. 2 displayed on the display device 102, the processing shown in FIG. 3 is started. At this time, based on the product data of the metal product 2 that is a healthy product acquired from the manufacturing execution system 5, the product number (P-a) is displayed in the product specification box 191A. A product type may be specified by inputting a product type number in the product type specification box 191A. Furthermore, the shape dimension classification in the product type number is designated in the shape dimension classification designation box 199A. The shape / dimension classification is obtained by dividing the angle of forming the metal product 2 and the range of variation of the size of the metal product 2 into a plurality of ranges although the shape / dimension varies within the dimensional tolerance even in the same type. .

  In step S11, the tapping sound inspection apparatus 1 acquires a sound hitting sound data group of sound products having the same product type and the same shape and size category as the designated product (P-a) and shape and size category from the history data storage unit 103a. FIG. 4 shows a metadata table of the acquired sound hitting sound data group, and FIG. 5 shows an amplitude waveform chart thereof. These are examples of sound data sets of sound products that have been acquired. The metadata table of FIG. 4 represents metadata related to each sound hit data included in the sound hit data group of the healthy product.

  In the metadata table of FIG. 4, “sound data ID” in the first column represents the ID number of each sound data, and “product number” in the second column is the product type of the metal product 2 corresponding to each sound data. In the third column, “shape dimension classification” represents the shape dimension classification of the metal product 2, and “judgment result” in the fourth column represents the determination result of the presence or absence of defects in the metal product 2 corresponding to each hitting data. Represents. In step S11, as the sound hitting sound data group of the sound product, as shown in the fourth column of the metadata table of FIG. 4, the sound product and the same shape and size category, that is, the same size category (D2) determined as having no defect. It is acquired by narrowing down the sound data (Data-1, Data-4, Data-5) of the metal product 2. Note that, as described above, the sound hitting sound data group used to generate the sound product variation range first is a sound product using a method other than the sound hitting inspection performed by the sound hitting inspection apparatus 1. The sound hitting data of the plurality of metal products 2 that have been determined to be used are employed. For this reason, the “determination result” in the metadata of the sound hitting data group of the healthy product at this time stores the determination result of the presence / absence of a defect made by the other method.

  In the amplitude waveform chart shown in FIG. 5, “hitting sound data ID” in the first column represents the ID number of each hitting data as in FIG. 4, and “raw sound data” in the second column represents each hitting data. Represents raw data. In this embodiment, as shown in FIG. 5, time-series amplitude waveform data 201A, 204A, and 205A with time (elapsed time after impact) on the horizontal axis and sound (sound pressure) on the vertical axis Obtained as raw sound data. The “contour line extraction” in the third column represents the contour lines 201B, 204B, and 205B calculated in the process of step S12 described later for the amplitude waveform data 201A, 204A, and 205A, respectively.

  Next, in step S12, the hammering test apparatus 1 calculates a contour line of each hammering sound data (amplitude waveform data). In step S13, it is determined whether or not the contour calculation of all hit sound data has been completed. If not, the process returns to step S12. By repeating this processing, contour lines 201B, 204B, and 205B shown in the third column are calculated from the amplitude waveform data 201A, 204A, and 205A in the “raw sound data” in the second column of FIG.

  After it is determined in step S13 that the contour calculation of all the hitting sound data has been completed, the hitting sound inspection apparatus 1 proceeds to the process of step S14. In step S14, the upper and lower limits of the contour line group combining all the contour lines are extracted. The procedure for extracting the upper and lower limits of the contour group will be described with reference to FIGS. FIG. 6 is a chart in which outlines 201B, 204B, and 205B are overwritten. In step S14, the upper limit curve 205C and the lower limit curve shown in FIG. 7 are extracted by sequentially extracting the upper limit value and the lower limit value at each time of these contour lines while scanning the time axis which is the horizontal axis of FIG. 204C can be extracted.

  In the examples of FIGS. 6 and 7, the contour lines 201B, 204B, and 205B of the contour group do not intersect each other, so that the contour lines 205B and 204C remain as the upper limit curve 205C and the lower limit curve 204C of the contour line, respectively. Has been extracted. However, the present invention is not limited to this, and even when a plurality of contour lines included in the contour line group intersect each other, the upper limit curve and the lower limit curve can be extracted by the procedure described above.

In this embodiment, the upper limit curve and the lower limit curve are directly extracted from the contour group. However, at a stage where a large amount (for example, 50 or more) of healthy product history data is accumulated, The upper limit curve and the lower limit curve of the contour line group may be determined by sequentially determining the upper limit value and the lower limit value from the point group corresponding to the value by statistical processing. An example of statistical processing is shown in (Formula 1).

Where L _upper (t) is the upper limit curve, L _lower (t) is the lower limit curve, x _i (t) is the i-th contour line, t is the time, α is the coefficient of the point group standard deviation, and N is the contour line Is the number of The upper limit value and the lower limit value are determined by multiplying the standard deviation of the point group by a coefficient. Further, the healthy product variation range may be learned by extracting the upper limit curve and the lower limit curve in a predetermined cycle such as daily, weekly, or monthly.

  In the next step S15, the hammering inspection apparatus 1 stores the upper limit curve and the lower limit curve generated in step S14 in the healthy product variation range storage unit 103b. At this time, the upper limit curve and the lower limit curve that define the specifications of the healthy product variation range with the file name (for example, Criteria_P-a_D2.csv) input to the file designation box 192A on the generation screen of FIG. 2 displayed on the display device 102. Is recorded in the healthy product variation range storage unit 103b. At this time, the shape dimension classification based on the shape dimension acquired from the shape dimension measuring device 8 is stored in association with the product type. Alternatively, when the shape dimension measuring instrument 8 is not installed, the shape dimension classification designated in the shape dimension classification designation box 199A is stored in association with the product type.

  In step S15, the hammering test apparatus 1 displays the charts of the upper limit curve and the lower limit curve generated in step S14 in the healthy product variation range display field 196A of the amplitude waveform on the generation screen illustrated in FIG. . Thereafter, the healthy product variation range generation process shown in the flowchart of FIG. 3 is terminated. Although the upper limit curve and lower limit curve data tables for defining the healthy product variation range are not shown, they are the same as the percussion sound data tables shown in FIGS. 10 and 11 to be described later, and the product number (Pa) and the shape dimension classification (D2). And memorize it. In this way, based on the sound hit data of a plurality of healthy products, the allowable range of variation allowed as healthy products, that is, the healthy product variation range by the upper and lower contour curves extracted from the amplitude waveform data is set. To do.

-Screen for judging whether defects exist-
FIG. 8 is a diagram illustrating an example of a defect presence / absence determination screen displayed on the display device 102 by the hammering inspection apparatus 1. As shown in FIG. 8, the defect presence / absence determination screen is displayed from a product specification box 191B, a file specification box 192B, a shape dimension classification specification box 199B, a defect presence / absence determination button 194, a determination result display box 195, and a healthy product variation range of the amplitude waveform. The deviation determination result display field 196B.

-Defect existence judgment procedure-
FIG. 9 is a flowchart showing a processing procedure of defect presence / absence determination by the sound inspection apparatus 1. The process shown in this flowchart shows the process executed by the arithmetic processing unit 101 of the computer 100 for hammering test.

  When the defect presence / absence determination button 194 is designated from the screen shown in FIG. 8 of the display device 102, the processing shown in FIG. 9 is started. At this time, based on the product data and shape / dimension category of the metal product 2 that is the inspection target product acquired from the manufacturing execution system 5, the product type number (Pa) is set in the product specification box 191B and the shape / size category (eg, D2) is formed It is displayed in the dimension classification designation box 199B. At the same time, the file name (for example, Criteria_P-a_D2.csv) that defines the specifications of the healthy product variation range that conforms to the displayed product number (Pa) and shape and dimension classification (D2) is stored in the healthy product variation range. The file is searched from the section 103b and displayed in the file designation box 192B. By inputting a file name in the file designation box 192B, a file that defines the specification of the healthy product variation range may be designated.

  In step S <b> 21 of FIG. 9, the sound inspection device 1 outputs a batting control signal (hitting command) to the batting device 3 using the batting control signal output device 200. Thereby, the striking device 3 strikes the metal product 2 to be inspected.

  In the next step S22, the hitting sound inspection device 1 aligns the timing of the hitting of the hitting device 3, and A / D-converts the sound pressure data measured from the microphone 4 by the A / D converter 300. Start collecting data. In step S23, it is determined by a timer in the hammering test computer 100 whether or not a predetermined time interval has passed. The hitting data collected during a predetermined time interval is temporarily stored in the RAM in the hitting test computer 100.

  If the predetermined time interval has elapsed in step S23, the A / D conversion is terminated, and the process proceeds to step S24. In step S24, the collected hitting data is stored in the history data storage unit 103a. At that time, the product type number of the metal product 2 collected from the manufacturing execution system 5 and the shape dimension classification based on the shape dimension acquired from the shape dimension measuring device 8 are stored in association with the hitting data. Alternatively, when the shape dimension measuring instrument 8 is not installed, the shape dimension classification designated in the shape dimension classification designation box 199B is stored. 10 and 11 show examples of data tables in the history data storage unit 103a.

  FIG. 10 is a metadata table representing metadata relating to each hitting sound data stored in the history data storage unit 103a. As in FIG. 4, in the metadata table of FIG. 10, the “sound data ID” in the first column represents the ID number of each sound data, and the “product number” in the second column corresponds to each sound data. The “shape dimension classification” in the third column represents the classification of the shape dimension of the metal product 2, and the “judgment result” in the fourth column represents the metal product 2 corresponding to each sound data. The determination result of the presence or absence of defects is represented. These metadata are recorded every time each hitting sound data is measured.

  FIG. 11 is a raw data table representing the content of each hitting data stored in the history data storage unit 103a. In the raw data table of FIG. 11, “sounding data ID” in the first column represents the ID number of each percussion data, and “time 1” to “time N” in the second column and thereafter represent the percussion sound at each time. Represents the sound pressure level. These data are recorded every time each hitting sound data is measured. The metadata shown in FIG. 10 and the live sound data shown in FIG. 11 are linked by the ID number represented by each “sound data ID”.

  Next, in step S25, the hammering inspection apparatus 1 calculates the contour line of the amplitude waveform data represented by the hammering data from the hammering data of the metal product 2 to be inspected stored in the history data storage unit 103a in step S24. To do. Subsequently, in step S26, the sound hitting inspection apparatus 1 searches and acquires the file having the file name displayed in the file designation box 192B from the healthy product variation range storage unit 103b. In the file, information on the upper limit curve and the lower limit curve that define the specification of the range of healthy product variation is stored in association with the product type and the shape dimension classification. It is possible to determine an upper limit curve and a lower limit curve that define a healthy product variation range of the amplitude waveform data corresponding to the metal product 2 to be inspected.

  In step S27, the hammering inspection apparatus 1 causes the contour of the amplitude waveform data as the hammering data of the metal product 2 to be inspected calculated in step S25 to deviate above the upper limit curve that defines the healthy product variation range ( Judgment is made as to whether or not there is a deviation (excess deviation) below the lower limit curve. If it deviates to either, it is determined that the metal product 2 to be inspected is defective, and the process proceeds to step S28. In step S28, a flag with a defect is set. On the other hand, if it does not deviate from either, it is determined that the metal product 2 to be inspected is a healthy product, and the process proceeds to step S29. In step S29, a flag indicating no defect (healthy product) is set. As described above, when the amplitude waveform data, which is the hitting data of the inspection target product, is within the range of allowable variation as the healthy product variation range, the metal product 2 to be inspected is determined to be a healthy product.

  Finally, in step S30, the sound hitting inspection apparatus 1 sends the determination result to the quality management system 6 and displays the determination result in the determination result display box 195 of the determination screen shown in FIG. Further, a chart of determination results is displayed in the deviation determination result display field 196B from the healthy product variation range of the amplitude waveform. In the display column 196B of FIG. 8, the contour line 206 of the amplitude waveform data that is insufficiently deviated downward from the lower limit curve 204C is displayed together with the upper limit curve 205C and the lower limit curve 204C, when the metal product 2 to be inspected is a defective product. As an example. The determination result sent to the quality management system 6 is reflected in the “determination result” in the fourth column in the metadata table of FIG. 10 and stored in the history data storage unit 103a.

-Effects of the first embodiment-
As described above, according to the sound hitting inspection apparatus 1 of the present embodiment, the sound quality variation range of the amplitude waveform data due to sound hitting is automatically calculated from the sound hit sound data group of the sound product, and the range is calculated from the range. The hammering test is performed by determining the deviation of the amplitude waveform data of the product to be inspected. Therefore, the following effects can be obtained.

That is, in the hammering test apparatus 1 of this embodiment, it is not necessary to manually set a judgment threshold for hammering test for each product type, so that it is possible to reduce man-hours or labor costs. In addition, since there is no need to prepare a defective product in advance and set the judgment threshold by comparing the sound data of sound products with the sound data of defective products, the management cost of defective products or simulated defects Product manufacturing costs are eliminated. In addition, since the determination threshold value setting based on the defective product assumed in advance is not performed, it is possible to detect any defective product that deviates from the healthy product variation range.
Furthermore, since the defect on the metal product material affects the strength of the sound pressure of the hitting sound, the hitting sound inspection apparatus 1 of the present embodiment uses such a material from the deviation of the amplitude waveform on the time axis from the hit to the decay. Defects can be detected.
Furthermore, even within the same type of metal product, the shape dimensions vary within the dimensional tolerances. However, since the sound inspection is performed for each shape and size category according to the sound variation factor, the detection accuracy of the sound detection is improved. .

  In the first embodiment, for the metal product 2 to be inspected, the product type data is divided by the shape dimension data, and it is determined whether or not the contour line of the hitting data deviates from the healthy product variation range. Then, the presence or absence of defects of the metal product 2 was determined. However, with respect to the metal product 2 to be inspected, the product type data is divided by the shape and dimension data, and the presence / absence of a defect is determined using the frequency spectrum of the hitting data as described in the second embodiment below. May be.

  In the first embodiment, for the metal product 2 to be inspected, by classifying product type data by shape size data, and determining whether or not the hitting sound data deviates from the healthy product variation range, The presence or absence of defects in the metal product 2 was determined. However, the presence / absence of a defect in the metal product 2 may be determined for the metal product 2 to be inspected without classifying the product data by the shape / dimension data. Such a determination can be applied to the case where the metal product 2 has a uniform shape dimension, or the case where the hitting data does not depend on the shape dimension.

(Second Embodiment)
FIG. 12 is a system configuration diagram of the hammering inspection apparatus according to the present embodiment. In the present embodiment, an example in which the hammering inspection apparatus is realized by a computer system is shown. In the second embodiment, an example will be described in which the presence / absence of a defect is determined for the entire frequency spectrum of the hitting data. Furthermore, in the present embodiment, the manufacturing route data will be described as an example of the sound variation factor.
In this embodiment, a number reader 9 is provided instead of the shape measuring device 8 in the system configuration diagram shown in FIG. 1 as the first embodiment. Other configurations are the same as those in FIG.

  The computer 100 for hammering inspection reads the number 2N affixed to the metal product 2 with the number reader 9, and acquires the number 2N. The number 2N is an individual number that identifies an individual of the metal product 2 or an individual group number that identifies an individual group of the metal product 2, and is associated with a unique manufacturing path identification number for each manufacturing path of the metal product 2. Yes. The computer 100 for hammering inspection can use the manufacturing traceability information recorded in the manufacturing execution system 5 to assign a manufacturing path identification number corresponding to the acquired number 2N. Alternatively, the manufacturing path identification number may be included in the number 2N, and the manufacturing path identification number may be directly read by the number reader 9. The number reader 9 is, for example, a bar code reader or an image recognition device according to the number 2N marking method.

  A manufacturing path identification number is a number that identifies a manufacturing line when the same product is manufactured on a plurality of manufacturing lines, or a manufacturing apparatus that is identified when a plurality of manufacturing apparatuses are used to manufacture the same product on a single manufacturing line. Or a number for identifying a jig when a single product is manufactured using a jig such as a plurality of molds or a combination thereof.

  Even if the same product is the same product type, the shape and size of the metal product 2 may vary within the dimensional tolerance due to the difference in the manufacturing path. In the present embodiment, the presence / absence of a defect is determined based on the hitting sound data by using a manufacturing path section in which the range of variation is divided into a plurality of sections.

-Sound product variation range generation screen-
FIG. 13 is a diagram illustrating an example of a sound quality variation range generation screen displayed on the display device 102 by the sound hitting inspection apparatus 1 illustrated in FIG. 12. Components different from the screen of FIG. 2 described in the first embodiment are provided with a healthy product variation range display column 197A of the frequency spectrum instead of the healthy product variation range display column 196A of the amplitude waveform. Further, a manufacturing path segment designation box 200A is provided in place of the shape dimension category designation box 199A.

-Process for generating healthy product variation range-
FIG. 14 is a flowchart showing a processing procedure for generating a healthy product variation range by the sound hitting inspection apparatus 1. The process shown in this flowchart shows the process executed by the arithmetic processing unit 101 of the computer 100 for hammering test.

  When the healthy product variation range generation button 193 is designated on the generation screen shown in FIG. 13 displayed on the display device 102, the processing shown in FIG. 14 is started. At this time, based on the product data of the metal product 2 that is a healthy product acquired from the manufacturing execution system 5, the product number (P-a) is displayed in the product specification box 191A. A product type may be specified by inputting a product type number in the product type specification box 191A. Further, the manufacturing route classification (for example, R3) in the product type number is displayed or designated in the manufacturing route classification designation box 200A.

  In step S41, the tapping sound inspection apparatus 1 acquires a sound hitting data group of sound products having the same product route classification and the same product type as the designated product type (Pa) and production route category (R3) from the history data storage unit 103a. To do. In step S42, the frequency spectrum of each hit sound data is calculated. For example, the frequency spectrum 211A shown in the second column of FIG. 15 is calculated from the amplitude waveform data 201A as the hit sound raw data shown in the second column of FIG.

  In the frequency spectrum 211A of FIG. 15, the horizontal axis is frequency (kHz), and the vertical axis is spectrum intensity (dB). Subsequently, in step S43, the contour line of the frequency spectrum calculated in step S42 is calculated. For example, as shown in the third column of FIG. 15, a contour 211B is calculated for the frequency spectrum 211A. In step S44, it is determined whether or not the contour calculation of the frequency spectrum of all hit sound data has been completed. If not completed, the process returns to steps S42 and S43. By repeating this process, the frequency spectra 214A and 215A of FIG. 15 are calculated from the amplitude waveform data 204A and 205A as the hitting sound data of FIG. 5, and the contour lines 214B and 215B of the respective spectra are calculated.

  After it is determined in step S44 that the calculation of the contour lines of the frequency spectrum for all the hitting sound data has been completed, the hitting sound inspection apparatus 1 sets the upper limit of the contour line group including the contour lines of all the frequency spectra in step S45. Extract. The procedure for extracting the upper limit of the contour line group of the frequency spectrum will be described with reference to FIGS. FIG. 16 is a chart in which contour lines 211B, 214B, and 215B of the frequency spectrum are overwritten. In step S45, the upper limit curve 216C shown in FIG. 17 is extracted by sequentially extracting the upper limit value at each frequency of the contour line of these frequency spectra while scanning the frequency axis which is the horizontal axis of FIG. be able to. The upper limit curve 216C of the contour group of the frequency spectrum extracted in this way has a shape combining a trapezoid and a rectangle that cross each adjacent peak.

  In this embodiment, the lower limit curve of the contour group of the frequency spectrum is omitted, but the lower limit curve can also be obtained. In the present embodiment, the upper limit curve is extracted directly from the contour group of the frequency spectrum. However, as in the first embodiment, the history data of healthy products is accumulated in large quantities (for example, 50 or more). Then, the upper limit curve of the frequency spectrum outline group may be determined by sequentially determining the upper limit value by statistical processing from the point group corresponding to each frequency value of the frequency spectrum outline group.

  Finally, in step S46, the sound hitting inspection apparatus 1 stores the upper limit curve of the frequency spectrum generated in step S45 in the healthy product variation range storage unit 103b. At this time, information on the upper limit curve that defines the specification of the healthy product variation range with the file name (for example, Criteria_P-a_R3.csv) input in the file designation box 192A on the generation screen of FIG. 13 displayed on the display device 102 is displayed. And recorded in the healthy product variation range storage unit 103b. At this time, the manufacturing route classification based on the number 2N acquired from the number reader 9 is stored in association with the hitting data ID. Alternatively, when the number reader 9 is not installed, the manufacturing route segment designated in the manufacturing route category designation box 200A is stored in association with the hitting sound data ID.

  Further, in step S46, the sound hitting inspection apparatus 1 further displays a chart of the upper limit curve generated in step S45 in the healthy product variation range display column 197A of the frequency spectrum of the generation screen shown in FIG. Thereafter, the generation process of the healthy product variation range in FIG. 14 is terminated. In this way, based on the sound hit data of a plurality of healthy products, a variation allowable range permitted as a healthy product, that is, a healthy product variation range based on the upper limit curve of the contour line extracted from the frequency spectrum is set.

-Screen for judging whether defects exist-
FIG. 18 is a diagram illustrating an example of a defect presence / absence determination screen displayed on the display device 102 by the hammering inspection apparatus 1. The components different from the screen of FIG. 8 described in the first embodiment are the deviation determination result display field from the healthy product variation range of the frequency spectrum instead of the deviation determination result display column 196B from the healthy product variation range of the amplitude waveform. 197B is provided. Further, a manufacturing path segment designation box 200B is provided instead of the shape dimension category designation box 199B.

-Defect existence judgment procedure-
FIG. 19 is a flowchart showing a processing procedure of defect presence / absence determination by the sound inspection apparatus 1. The process shown in this flowchart shows the process executed by the arithmetic processing unit 101 of the computer 100 for hammering test.

  When the defect presence / absence determination button 194 on the screen of the display device 102 shown in FIG. 18 is designated, the processing shown in FIG. 19 is started. At this time, based on the product data and manufacturing route classification of the metal product 2 that is the inspection object product obtained from the manufacturing execution system 5, the product number (Pa) is set in the product specification box 191B, and the manufacturing route classification (R3) is set in the manufacturing route. Displayed in the category designation box 200B. At the same time, the file name (for example, Criteria_P-a_R3.csv) that defines the specifications of the healthy product variation range that conforms to the displayed product number (Pa) and manufacturing route classification (R3) is stored in the healthy product variation range storage. The file is searched from the section 103b and displayed in the file designation box 192B. By inputting a file name in the file designation box 192B, a file that defines the specification of the healthy product variation range may be designated.

  The processing in steps S21 to S24 shown in FIG. 19 is the same as that in steps S21 to S24 shown in FIG. 9 and will not be described in detail. However, the hitting data is acquired, and the history data is stored as shown in FIGS. It is stored in the history data storage unit 103a. At that time, the product number of the metal product 2 collected from the manufacturing execution system 5 and the manufacturing route classification based on the number 2N acquired from the number reader 9 are stored in association with the sound hitting data. Alternatively, if the number reader 9 is not installed, the manufacturing route classification designated in the production route classification designation box 200B is stored.

  Next, in step S51, the hammering inspection apparatus 1 calculates the frequency spectrum of the amplitude waveform data represented by the hammering data from the hammering data of the metal product 2 to be inspected stored in the history data storage unit 103a in step S24. calculate. In step S52, the contour line of the frequency spectrum is calculated.

  Subsequently, in step S53, the sound hitting inspection apparatus 1 acquires the file having the file name displayed in the file designation box 192B from the healthy product variation range storage unit 103b, and reads this to read the metal product 2 to be inspected. The upper limit curve that defines the range of sound product variation in the frequency spectrum corresponding to is read. In the file, information defining the specification of the range of healthy product variation is stored in association with the product type and the manufacturing route classification. By reading this file, the hammering inspection apparatus 1 can detect the metal product 2 to be inspected. It is possible to determine the upper and lower limits that define the healthy product variation range of the frequency spectrum corresponding to.

  In step S54, the hammering inspection apparatus 1 causes the contour line of the frequency spectrum as the hammering data of the metal product 2 to be inspected calculated in step S52 to exceed the upper limit curve defining the healthy product variation range. Determine if it is deviating. If there is a deviation, a flag indicating that there is a defect is set in step S55, and if there is no deviation, a flag indicating that there is no defect (healthy product) is set in step S56. As described above, when the hit sound data of the inspection target product is within the range of the allowable variation range expressed as the healthy product variation range, the metal product 2 to be inspected is determined as a healthy product.

  Finally, in step S56, the sound hitting inspection apparatus 1 sends the determination result to the quality management system 6, and displays the determination result in the determination result display box 195 of the determination screen shown in FIG. Further, a chart of determination results is displayed in the deviation determination result display column 197B from the healthy product variation range of the frequency spectrum. In the display column 197B of FIG. 18, the contour line 217C of the frequency spectrum that deviates from the upper limit curve 216C is shown as an example when the metal product 2 to be inspected is a defective product together with the upper limit curve 216C.

-Effects of the second embodiment-
As described above, according to the hammering sound inspection apparatus 1 of the present embodiment, the sound product variation range of the sounding frequency spectrum is automatically calculated from the sounding sound data group of the sound product, and the inspection target is calculated from the range. The hammering test is performed by determining the deviation of the frequency spectrum from the hammering data of the product. Therefore, the following effects can be obtained.

  That is, in the sound inspection apparatus 1 of the present embodiment, a frequency peak to be focused on for defect inspection is manually determined from a number of frequency peaks existing on the frequency spectrum, and the sound inspection test for the frequency peak is performed. Since it is not necessary to manually set a judgment threshold for each product type, it is possible to reduce man-hours or labor costs. The frequency peak of a complex-shaped metal product may be several tens of major ones, and the frequency varies depending on the product type. Also, it is necessary to prepare a defective product in advance, determine the frequency peak by comparing the frequency spectrum of sound data of sound product and the frequency spectrum of sound data of defective product, and set the judgment threshold. Since it is not necessary, the management cost of the defective product or the manufacturing cost of the simulated defective product becomes unnecessary. Further, since the frequency peak determination and determination threshold value setting based on the defective product assumed in advance are not performed, it is possible to detect any defective product that deviates from the healthy product variation range of the frequency spectrum.

Further, since the structural defect of the metal product influences the frequency of the hitting sound, the hitting sound inspection apparatus 1 according to the present embodiment has such a structural defect from the deviation due to the shift of the peak frequency on the frequency axis in the entire frequency spectrum. Can be detected.
Furthermore, even if the same product is the same product type, there are variations in the shape dimensions depending on the metal product manufacturing path within the dimensional tolerances. The detection accuracy of sound determination is improved.

  In the second embodiment, for the metal product 2 to be inspected, the product type data is divided by the production path data, and the presence / absence of a defect is determined using the frequency spectrum of the hitting data. However, for the metal product 2 to be inspected, the product type data is divided by the manufacturing route data, and as described in the first embodiment, the outline of the hitting data deviates from the healthy product variation range. Whether or not the metal product 2 is defective may be determined by determining whether or not the metal product 2 is present.

  In the second embodiment, for the metal product 2 to be inspected, by classifying product type data by manufacturing route data and determining whether or not the hitting sound data deviates from the healthy product variation range, The presence or absence of defects in the metal product 2 was determined. However, the presence / absence of a defect in the metal product 2 may be determined for the metal product 2 to be inspected without classifying the product data by the manufacturing route data. Such a determination can be applied when the manufacturing route data of the metal product 2 is uniform, or when the hitting data does not depend on the manufacturing route.

(Third embodiment)
The system configuration diagram of the hammering inspection apparatus according to the third embodiment is the same as that shown in FIG. In the third embodiment, the presence / absence of a defect is determined for both the amplitude waveform from the hit of the hit sound data to the decay and the entire frequency spectrum. The sound variation factor will be described using an example of the shape and dimension, but the same applies to the manufacturing path.
The processing procedure for generating the healthy product variation range and the processing procedure for determining the presence / absence of defects are combined or parallel processing of the flowcharts shown in the first embodiment and the second embodiment, and thus description thereof is omitted.

  As described in the first embodiment and the second embodiment, the hammering inspection apparatus 1 generates a healthy product variation range for both the amplitude waveform and the frequency spectrum of the hammering data. Then, as shown in FIG. 20, on the generation screen of the display device 102, the respective charts are displayed in both the healthy product variation range display column 196A of the amplitude waveform and the healthy product variation range display column 197A of the frequency spectrum.

  As described in the first embodiment and the second embodiment, the tapping sound inspection apparatus 1 determines both the deviation of the amplitude waveform from the healthy product variation range and the deviation of the frequency spectrum from the healthy product variation range. To do. As shown in FIG. 21, in the defect presence / absence determination screen of the display device 102, the deviation determination result display field 196B from the healthy product variation range of the amplitude waveform and the deviation determination result display field 197B from the healthy product variation range of the frequency spectrum are displayed. The chart of each determination result is displayed.

-Effects of the third embodiment-
As described above, according to the hammering test apparatus 1 of the present embodiment, the healthy product variation range of the amplitude waveform data due to the hammering sound and the healthy product variation range of the frequency spectrum from the sounding data group of the healthy product automatically. Both are calculated, and the hammering test is performed by determining the deviation of the amplitude waveform and frequency spectrum from the hammering data of the product to be inspected from that range. Therefore, the following effects can be obtained.

  That is, in the sound inspection apparatus 1 of this embodiment, the type of defect is automatically detected by examining whether the deviation is detected in the amplitude waveform or the frequency spectrum, or both are detected. Enable. For example, since a defect in the material of the metal product affects the strength of the sound pressure of the hitting sound, the material defect can be detected from the deviation of the amplitude waveform on the time axis from the hit to the decay. On the other hand, since the defect in the structure of the metal product affects the frequency of the hitting sound, the structural defect can be detected from the deviation due to the shift of the peak frequency on the frequency axis in the entire frequency spectrum. Detailed detection information of amplitude waveform and frequency spectrum (time zone and deviation direction of amplitude waveform deviation, peak frequency value of deviation of frequency spectrum, etc.) and detailed defect type (material defect level, structural defect location, etc.) A detailed defect type can be detected by learning by collating.

  In the third embodiment, the sound variation factor has been described as an example of the shape dimension. However, for the metal product 2 to be inspected, the metal product is not divided into the shape data and the manufacturing route data without classifying the product data. The presence or absence of two defects may be determined. Such a determination can be applied when the metal product 2 has a uniform shape dimension, when the manufacturing path data is uniform, or when the hitting data does not depend on the shape dimension or the manufacturing path.

  The inspection target product is not limited to a single-material metal product, but includes various products and parts formed of a material that generates sound when hit. For example, it may be a composite material of metal and other materials such as ceramics, artificial stone (concrete), or the like.

According to the embodiment described above, the following operational effects can be obtained.
(1) The arithmetic processing unit 101 included in the hammering sound inspection apparatus 1 sets a healthy product variation range representing a variation allowable range allowed as a healthy product based on the sounding data of a plurality of healthy products (step S14, S15, steps S45, S46), it is determined whether or not the inspection target product is a healthy product by determining whether or not the hitting data of the inspection target product is within the range of the allowable variation range (step S45). S27, step S54). This simplifies the task of presetting the judgment threshold for hammering sound inspection.

(2) The hit sound data is time-series amplitude waveform data, and in the first embodiment, the arithmetic processing unit 101 extracts each contour line from the amplitude waveform data collected by a plurality of healthy products (step S12) The range between the upper limit and the lower limit of the extracted contour line is set as a variation allowable width (steps S14 and S15). In the second embodiment, the arithmetic processing unit 101 extracts each frequency spectrum from the amplitude waveform data collected by a plurality of healthy products (step S42), and allows variation from the contour line of the extracted frequency spectrum. A width is set (steps S45 and S46). Further, in the third embodiment, a variation allowable width is set by combining these. Since it did in this way, variation tolerance width can be set up appropriately.

(3) In the first embodiment, the arithmetic processing unit 101 determines that the inspection target product is a material defect when the contour line of the amplitude waveform data of the inspection target product is not within the range of the allowable variation range. Yes (steps S27 and S28). In the second embodiment, the arithmetic processing unit 101 can determine that the inspection target product is a structural defect when the contour line of the frequency spectrum of the inspection target product does not fall within the range of the allowable variation range ( Steps S54 and S55). Moreover, in 3rd Embodiment, the kind of defect of a test object can be determined combining these. Thereby, when it is determined that the product to be inspected is not a healthy product but has a defect, the type of the defect can also be determined.

(4) In the first embodiment, a sound product variation range is set for each dimension category by providing a means for measuring the shape dimension of the product. It is possible to improve the inspection accuracy by determining whether or not is within the range of allowable variation.

(5) In the second embodiment, a sound product variation range is set for each production path category by providing means for reading the product individual number, and the sound hit data of the product to be inspected for each production route category which is a sound variation factor It is possible to improve the inspection accuracy by determining whether or not is within the range of allowable variation.

  The present invention is not limited to the above-described embodiment, and other forms conceivable within the scope of the technical idea of the present invention are also included in the scope of the present invention as long as the characteristics of the present invention are not impaired. .

DESCRIPTION OF SYMBOLS 1 Tapping sound inspection apparatus 2 Metal product 3 Impact apparatus 4 Microphone 5 Manufacturing execution system (MES)
6 Quality control system (QCS)
7 Network 8 Shape measuring instrument 9 Number reader 100 Computer for hammering inspection 101 Arithmetic processing device 102 Display device 103 Storage device

Claims (5)

Based on the sound hit data of a plurality of healthy products, a variation allowable width setting unit that sets a variation allowable width allowed as a healthy product,
A determination unit that determines whether or not the inspection target product is a healthy product by determining whether or not the hitting data of the inspection target product is within the range of the variation allowable width ; and
The sound hitting data is time-series amplitude waveform data,
The variation allowable width setting unit extracts each frequency spectrum from the amplitude waveform data collected by the plurality of sound products, and sets the variation allowable width from the extracted outline of the frequency spectrum. apparatus. The sound inspection device according to claim 1 ,
The variation permissible width setting unit sets a permissible variation permissible as a sound product for each sound variation factor based on the sound hit data of a plurality of sound products,
The determination unit determines whether or not the inspection target product is a healthy product by determining whether or not the hitting data of the inspection target product is within the range of the variation allowable width for each of the sound variation factors. Sound inspection device for determining The sound inspection device according to claim 2 ,
The sound variation factor is a predetermined shape dimension of the inspection target product,
The variation permissible width setting unit sets a permissible variation permissible as the healthy product for each shape dimension category,
The determination unit is a hammering inspection apparatus that determines whether or not the hammering data of the inspection target product is within the range of the allowable variation range according to the shape and dimension. The sound inspection device according to claim 2 ,
The sound variation factor is a manufacturing path classification of the inspection target product,
The variation permissible width setting unit sets a permissible variation permissible as the healthy product for each manufacturing path category,
The determination unit is a hammering inspection apparatus that determines whether or not the hammering data of the inspection target product is within the range of the variation allowable width, according to the manufacturing path classification. The sound inspection device according to claim 1 ,
The determination unit is a hammering inspection apparatus that determines the inspection target product as a structural defect when the contour line of the frequency spectrum of the inspection target product does not fall within the range of the allowable variation range.

Images