JP7486672B2 - Inspection device, inspection system, inspection method, and inspection program - Google Patents

Description

The present disclosure relates to an inspection apparatus, an inspection system, an inspection method, and an inspection program.

Patent Document 1 discloses an apparatus that measures the vibration intensity of an excited test object and analyzes the measurement data to diagnose whether or not the test object has a structural abnormality such as a crack (i.e., the presence or absence of a structural abnormality). This apparatus measures the vibration intensity for each frequency of the excited test object to detect the frequencies of multiple vibration peaks, calculates a correlation coefficient between the detected frequency and the frequencies of multiple vibration peaks detected in advance for a reference object confirmed to be normal, and diagnoses the presence or absence of a structural abnormality in the test object based on this correlation coefficient.

JP 2008-232763 A

However, the above-mentioned conventional devices diagnose the presence or absence of abnormalities using multiple vibration peaks of a reference object and multiple vibration peaks of the object being inspected, which results in a problem that the accuracy of diagnosing the presence or absence of structural abnormalities deteriorates when the frequency interval between the multiple vibration peaks is narrow or when the measured vibration intensity contains noise components.

The present disclosure aims to provide an inspection device, an inspection system, an inspection method, and an inspection program that enable the presence or absence of abnormalities in an object to be inspected with high diagnostic accuracy.

The inspection device disclosed herein is characterized by having a data collection unit that collects, as measurement data, the measurement values for each vibration frequency for N inspection targets (N is an integer equal to or greater than 3) as the object from a vibration detection unit that detects vibrations of an object excited by a vibration unit and outputs measurement values of the vibrations; a data analysis unit that calculates the similarity of the measurement data for each combination of two inspection targets among the N inspection targets and calculates an evaluation value, which is a statistic of the similarity, for each of the N inspection targets; and a diagnosis unit that compares the evaluation values of the N inspection targets with each other and diagnoses whether or not each of the N inspection targets has an abnormality based on the results of the comparison.

The inspection method disclosed herein is characterized by comprising the steps of: collecting, as measurement data, the measurement values for each vibration frequency for N inspection objects (N is an integer equal to or greater than 3) from a vibration detection unit that detects vibrations of an object excited by a vibration unit and outputs measurement values of the vibrations; performing a process of calculating the similarity of the measurement data for two of the N inspection objects for each combination of two of the N inspection objects, and calculating an evaluation value, which is a statistic of the similarity, for each of the N inspection objects; and comparing the evaluation values of the N inspection objects with each other and diagnosing whether or not each of the N inspection objects has an abnormality based on the result of the comparison.

According to the present disclosure, it is possible to test for the presence or absence of abnormalities in the test subject with high diagnostic accuracy.

1 is a schematic diagram showing a configuration of an inspection device and an inspection system according to a first embodiment; 11A to 11D are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by the data collection unit of the inspection device of embodiment 1 as frequency/vibration intensity characteristics. 5 is a diagram showing a process performed by a data analysis unit of the inspection device according to the first embodiment; FIG. FIG. 4 is a frequency distribution diagram showing the relationship between the average correlation coefficient and the frequency shown in FIG. 3 . 4 is a flowchart showing an operation of the inspection device according to the first embodiment. 1 is a diagram showing an example in which a vibration applying unit and a vibration detecting unit of an inspection system according to a first embodiment are mounted on a moving body; FIG. 11 is a schematic diagram showing a configuration of an inspection device and an inspection system according to a second embodiment. 13A and 13B are diagrams showing a vibration mode of an inspection object excited by a vibrator of an inspection system according to a second embodiment, and a vibration detector. 9 is a diagram showing the relationship between a vibrator of an inspection system according to a second embodiment and the vibration modes shown in FIGS. 8A and 8B. FIG. 13A to 13D are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by a data collection unit of an inspection device according to embodiment 2 as frequency/vibration intensity characteristics. 13 is a diagram showing a process performed by a data analysis unit of an inspection device according to a second embodiment. FIG. FIG. 12 is a frequency distribution diagram showing the relationship between the average and frequency of the correlation coefficients shown in FIG. 11 . FIG. 11 is a schematic diagram showing a configuration of an inspection device and an inspection system according to a third embodiment. 13A to 13D are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by a data collection unit of an inspection device according to embodiment 3 as frequency/vibration intensity characteristics. 11A to 11D are diagrams showing the maximum vibration intensity values of the measurement data of inspection objects #1 to #4 collected by the data collection unit of the inspection device of embodiment 3 as frequency-vibration intensity characteristics. FIG. 13 is a schematic diagram showing a configuration of an inspection device and an inspection system according to a fourth embodiment. 13A and 13B are diagrams showing a vibration mode of an inspection object excited by a vibrator of an inspection system according to a fourth embodiment, and a vibration detector. 18 is a diagram showing the relationship between a vibration detector of an inspection system according to a fourth embodiment and the vibration modes shown in FIGS. 17A and 17B. FIG. 13A to 13D are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by a data collection unit of an inspection device according to embodiment 4 as frequency/vibration intensity characteristics. 13 is a diagram showing a process performed by a data analysis unit of an inspection device according to a fourth embodiment; FIG. FIG. 21 is a frequency distribution diagram showing the relationship between the average and frequency of the correlation coefficients shown in FIG. 20 . FIG. 13 is a schematic diagram showing a configuration of an inspection device and an inspection system according to a fifth embodiment. FIG. 1 is a diagram illustrating an example of a hardware configuration of an inspection device and an inspection system according to first to fifth embodiments. FIG. 13 is a diagram illustrating another example of the hardware configuration of the inspection device and the inspection system according to the first to fifth embodiments.

Below, an inspection device, an inspection system, an inspection method, and an inspection program according to embodiments will be described with reference to the drawings. The following embodiments are merely examples, and it is possible to combine the embodiments as appropriate and to modify each embodiment as appropriate. Note that in each figure, the same or similar configurations are given the same reference numerals.

<<1>> First embodiment
1-1 Configuration FIG. 1 is a schematic diagram showing the configuration of an inspection device 11 and an inspection system 1 according to the first embodiment. As shown in FIG. 1, the inspection system 1 includes a vibration unit 21, a vibration detection unit 31, and an inspection device 11. The inspection device 11 includes a data collection unit 41, a data analysis unit 51, and a diagnosis unit 61. The inspection device 11 is a device capable of implementing the inspection method according to the first embodiment. The inspection device 11 performs an inspection of whether or not there is an abnormality in an inspection target 100 as an object (i.e., the presence or absence of an abnormality). The inspection of the presence or absence of an abnormality is mainly an inspection of the presence or absence of a structural abnormality. The structural abnormality is, for example, a crack or fissure (i.e., a crack) in the inspection target 100, a missing part of the inspection target 100, a deformation of the inspection target 100, and an incomplete installation of a part constituting the inspection target 100.

The vibration unit 21 has a vibration exciter 21a. The vibration exciter 21a is composed of, for example, a piezoelectric element that can vibrate the inspection object 100 at a frequency of 1 kHz or more. The vibration exciter 21a, for example, contacts the inspection object 100 to vibrate the inspection object 100.

The vibration detection unit 31 has a vibration detector 31a. The vibration detection unit 31 is composed of, for example, a piezoelectric element capable of detecting vibrations of the test object 100 at a frequency of 1 kHz or more. The vibration detector 31a comes into contact with the test object 100 and measures the vibration of the test object 100.

The data collection unit 41 collects measurement values for each vibration frequency as measurement data for each of N objects (N is an integer equal to or greater than 3) to be inspected from the vibration detection unit 31, which detects the vibration of the object excited by the excitation unit 21 and outputs the measurement value of the vibration. The measurement data is also called the measurement result. The data collection unit 41 collects measurement data related to the vibration intensity detected by the vibration detection unit 31.

The data analysis unit 51 executes the process of calculating the correlation coefficient of the measurement data of two of the N test subjects for each combination of two of the N test subjects, and calculates an evaluation value that is the sum or average of the correlation coefficients for each of the N test subjects. In the first embodiment, the data analysis unit 51 calculates the average of the correlation coefficients with the measurement data of the N test subjects.

The diagnosis unit 61 compares the evaluation values of the N test objects with each other, and diagnoses whether or not there is an abnormality in each of the N test objects based on the results of the comparison (for example, diagnoses whether or not there is a structural abnormality). In other words, the diagnosis unit 61 compares and analyzes the sum or average of the correlation coefficients for each test object calculated by the data analysis unit 51, diagnoses whether or not there is an abnormality for each test object, and outputs the results.

The inspection device 11 is configured, for example, by a processing circuit. The processing circuit may be configured by a processor such as a CPU (Central Processing Unit) that executes a program as software stored in a memory. The processor may be a processing circuit such as a system LSI. In addition, multiple processors and multiple memories may work together to realize the above functions.

<<1-2>> Operation The number of inspection objects may be three or more, but in the first embodiment, an example in which the number of inspection objects is four will be described as an example. The inspection objects are parts (i.e., objects) manufactured with the same design, and are represented as inspection object #1, inspection object #2, inspection object #3, and inspection object #4. Inspection objects #1 to #3 are normal products (e.g., products with no structural abnormalities), and inspection object #4 is an abnormal product (e.g., a product with a structural abnormality). An abnormal product is, for example, an inspection object in which cracks on the order of several hundred μm to several mm or more have occurred due to aging or the like.

The vibration unit 21 vibrates each inspection object at a predetermined frequency band of 1 kHz or more. The position of vibration is not particularly limited as long as it is the same position for inspection objects #1 to #4. However, if there is a vibration position at which the difference in correlation coefficient between normal and abnormal products is significantly evident depending on the inspection object, it is desirable to vibrate at that position.

The vibration detection unit 31 detects the vibration intensity when each test object is vibrated. The data collection unit 41 collects the vibration intensity detected by the vibration detection unit 31 as measurement data.

Figures 2 (A) to (D) are diagrams showing examples of measurement data 201 to 204 of inspection objects #1 to #4 collected by the data collection unit 41 of the inspection device 11 according to embodiment 1 as frequency/vibration intensity characteristics. As shown in Figures 2 (A) to (C), the measurement data 201 to 203 of inspection objects #1 to #3, which are normal products, generally show the same tendency. On the other hand, as shown in Figure 2 (D), the measurement data 204 of inspection object #4, which is an abnormal product, has a different frequency of vibration peaks or a different number of vibration peaks compared to the measurement data 201 to 203 of inspection objects #1 to #3.

The data analysis unit 51 analyzes the measurement data 201-204 collected by the data collection unit 41. For each of the measurement data 201-204 of test objects #1-#4, the data analysis unit 51 calculates the correlation coefficient with the measurement data of the other test objects, and calculates the average of the correlation coefficients. In the embodiment, the Pearson product-moment correlation coefficient shown in the following formula (1) is used as the correlation coefficient r.

In formula (1), A and B represent two test objects, and A _i and B _i are measurement data collected by the data collection unit 41 at frequency i. The frequency range is an arbitrary range within the frequency band of the measurement data. A _ave is the average value of A _i in the frequency range, and B _ave is the average value of B _i in the frequency range.

Figure 3 is a diagram showing the processing performed by the data analysis unit 51 of the inspection device 11 according to the first embodiment. The data analysis unit 51 calculates the correlation coefficients of the measurement data 202 to 204 of the inspection objects #2 to #4 with the measurement data 201 of the inspection object #1, and the calculated correlation coefficients are 0.99, 0.98, and 0.60, respectively. The average is 0.86. The data analysis unit 51 performs the same processing for the inspection objects other than the inspection object #1. The average of the correlation coefficients for the inspection objects #1 to #3, which are normal products, is in the range of 0.83 to 0.86, while the average of the correlation coefficients for the inspection object #4, which is an abnormal product, is 0.59, which is significantly different from the average of the correlation coefficients for the inspection objects #1 to #3.

Figure 4 is a frequency distribution diagram showing the relationship between the average and frequency of the correlation coefficients shown in Figure 3. The diagnosis unit 61 diagnoses the presence or absence of an abnormality in the test object based on the analysis result of the data analysis unit 51, and outputs the diagnosis result. In Figure 4, the width of the distribution is 0.1. The width of the distribution may be other values. The average of the correlation coefficients for the test objects #1 to #3, which are normal products, belongs to a frequency 402 of 0.8 to 0.9. On the other hand, the average of the correlation coefficients for the test object #4, which is an abnormal product, belongs to a frequency 401 of 0.5 to 0.6. In the distribution of the average of the correlation coefficients for the test objects #1 to #4, the average of the correlation coefficients for the test object #4, which is an abnormal product, is an outlier. Therefore, it is possible to detect the average of the test object #4 as an outlier by a clustering method such as the mean shift method or hierarchical clustering. The test object #4 detected as an outlier is determined to be abnormal. On the other hand, the test objects #1 to #3 are determined to be normal products because they are distributed in the same position. The above results are output as the diagnosis result of the presence or absence of abnormality. Note that if all the test objects are normal, all the frequencies will be distributed in approximately the same position, so they will be classified into the same class by clustering, and all the test objects will be determined to be normal.

Figure 5 is a flowchart showing the operation of the inspection device 11 according to the first embodiment. The data collection unit 41 collects measurement values for each vibration frequency for each of N objects (N is an integer equal to or greater than 3) from the vibration detection unit 31, which detects the vibration of an object excited by the excitation unit 21 and outputs a measurement value of the vibration (step S1). The data analysis unit 51 calculates the correlation coefficient of the measurement data for each combination of two objects among the N objects, and calculates an evaluation value that is the sum or average of the correlation coefficients for each of the N objects (step S2). The diagnosis unit 61 compares the evaluation values of the N objects with each other, and diagnoses whether or not there is an abnormality in each of the N objects based on the results of the comparison (step S3).

In the inspection device or inspection method according to the first embodiment, it is not necessary to calculate the frequency of each vibration peak because the correlation coefficient for the measurement data of each inspection target is calculated. Therefore, high diagnostic accuracy can be maintained even when the frequency interval of each vibration peak is narrow or the measurement data contains noise.

Furthermore, in the inspection device or inspection method according to embodiment 1, when the majority of test objects are normal, the average correlation coefficient for the normal test objects will be large. On the other hand, the average correlation coefficient for the abnormal test objects will be small. Therefore, even if there are no objects that have been confirmed in advance as normal, it is possible to diagnose the presence or absence of abnormalities by determining the presence or absence of outliers from the average correlation coefficients for each test object.

In addition, in the inspection device or inspection method according to the first embodiment, it is desirable to set the frequency of the vibration of the inspection target detected by the vibration detection unit 31 to a predetermined frequency band of 1 kHz or more. In this case, the following first to third effects are obtained.

The first effect is that by detecting vibrations of several kHz or more, it becomes possible to detect cracks on the order of several hundred μm to several mm. This is described, for example, in the following non-patent document 1.

Hiroki Mine, Jiang Wu, Yosuke Mizuno, and Kentaro Nakamura, “Non-destructive testing of T-shaped metal object by considering vibration mode”, Proceedings of Symposium on Ultrasonic Electronics, Vol. 39 (2018).

The second effect is that when the frequency band in which the response change due to cracks is large is known in advance, the diagnostic accuracy can be improved by detecting vibrations in a frequency band of several kHz or more. The third effect is that the vibration-to-noise ratio can be improved by not detecting frequencies below 1 kHz, which are not very suitable for crack detection.

By using piezoelectric elements for the vibrator 21a and vibration detector 31a, high-frequency vibration of several kHz or more can be realized, making it possible to detect cracks on the order of a few mm or less. In addition, the use of a signal generator or the like stabilizes the vibration force, making it possible to improve measurement accuracy compared to hammering tests, which generally have low stability of the vibration force.

In addition, the inspection device or inspection method according to embodiment 1 has the following fourth to sixth advantages compared to the technology described in non-patent document 2, which diagnoses the presence or absence of abnormalities in parts by ultrasonic testing when inspecting parts installed in a narrow space of a machine, such as a technology for inspecting a rotor wedge of a generator.

H. Katayama, F. Sato, Y. Gunji, T. Ishikawa, A. Matsuzaki, and H. Shimada, "Toshiba Robotic Inspection Technology for Turbo-Generator", CIGRE SC A1 Meeting and International Tutorials & Colloquium-CIGRE India (2019).

First, the fourth effect will be explained. In ultrasonic flaw detection, it is necessary to measure every position of the part to be inspected, which results in a problem of long inspection time. On the other hand, by using the inspection device or inspection method according to embodiment 1, inspection can be performed on a part-by-part basis, thereby shortening the inspection time.

Next, the fifth effect will be described. The ultrasonic sensor used for ultrasonic flaw detection is generally several tens of millimeters thick or more, and therefore there is a problem that the ultrasonic sensor cannot be used to inspect spaces that have gaps equal to or less than the thickness of the ultrasonic sensor. On the other hand, the inspection device or inspection method according to embodiment 1 can use a piezoelectric element, the thickness of which is several millimeters or less, and therefore can be applied to a larger number of inspection objects.

Next, the sixth effect will be explained. In ultrasonic flaw detection, if the test object is contaminated with dust or the like, ultrasonic waves are significantly attenuated between the sensor and the test object, which may result in a deterioration in test accuracy. On the other hand, in the test device or test method according to embodiment 1, the presence or absence of dust does not affect the test accuracy in a measurement principle in which the test object is vibrated by a piezoelectric element. Therefore, high test accuracy can be maintained even if the test object is contaminated with dust or the like.

<<1-4>> Modifications Fig. 6 is a diagram showing an example in which the excitation unit 21 and vibration detection unit 31 of the inspection system 1 according to the first embodiment are mounted on a moving body 71. The inspection system 1 may be one in which the excitation unit 21 and vibration detection unit 31 are mounted on a moving body 71 such as a mobile robot, and measurements can be taken remotely. With the above configuration, it is possible to realize inspection in places that cannot be accessed by humans. Furthermore, by performing automatic inspection using a moving body, it is possible to realize labor saving.

In technology for inspecting the rotor wedge of a generator, the gap into which a moving body can be inserted is approximately 100 mm or less. Therefore, by making the thickness of the vibration unit 21 and the vibration detection unit 31 100 mm or less, a configuration that allows the rotor wedge to be inspected can be realized.

The data analysis unit 51 may use the correlation coefficient with the measurement data of each test subject as the average of the correlation coefficients calculated in each of a plurality of predetermined frequency ranges. When there is a vibration mode with a significantly large vibration intensity compared to other vibration modes, the correlation coefficient is greatly affected by this vibration mode, and is less susceptible to response changes due to the presence or absence of abnormalities in other vibration modes. Therefore, by calculating the correlation coefficient for each of a plurality of frequency bands, it is possible to calculate a correlation coefficient that is normalized in each of the frequency bands with large and small vibration intensity of the vibration mode. Therefore, it is possible to calculate a correlation coefficient that is equally affected regardless of the vibration intensity of each vibration mode.

The information detected by the vibration detection unit 31 may not be vibration intensity, but may be information related to phase, such as the difference between the phase of the signal with which the vibration unit 21 vibrates the test object and the phase of the vibration detected by the vibration detection unit 31. The information detected by the vibration detection unit 31 may also be information that combines vibration intensity and the phase of the signal with which the vibration unit 21 vibrates the test object or the phase of the vibration detected by the vibration detection unit 31.

In addition, the data analysis unit 51 may perform the process of calculating the correlation coefficient of the measurement data of two of the N test objects on each of the two test objects, rather than on all of the two of the N test objects.

The data analysis unit 51 may calculate other similarities instead of the correlation coefficient. For example, the data analysis unit 51 may use a Euclidean distance calculated by taking the vibration intensity of each frequency of the measurement data 201 to 204 as the length of each dimension. The similarity is the correlation coefficient or the Euclidean distance.

In addition, the data analysis unit 51 may calculate other statistics such as sums, medians, trimmed means, etc., instead of the average similarity.

In other words, the data analysis unit 51 performs a process of calculating the similarity of the measurement data of two of the N test objects for each combination of two of the N test objects, and calculates an evaluation value, which is a statistical measure of the similarity, for each of the N test objects.

In addition, the diagnosis unit 61 may diagnose the presence or absence of anomalies by threshold judgment rather than by clustering.

In addition, the vibrator 21a may be configured to strike the test object with a hammer or the like, and the vibration detector 31a may be configured to detect vibrations with a microphone or the like and measure the vibration intensity at each frequency.

In addition, an electrodynamic vibrator or the like may be used as the vibrator 21a.

The vibration detector 31a may be an acceleration sensor or the like. The vibrator 21a may not vibrate the test object by contact, but may vibrate the test object in a non-contact manner using a laser or the like. The vibration detector 31a may not detect the vibration of the test object by contact, but may detect the vibration in a non-contact manner using a laser or the like.

<<2>> Second embodiment
<<2-1>> Configuration Fig. 7 is a schematic diagram showing the configuration of the inspection device 12 and the inspection system 2 according to the second embodiment. As shown in Fig. 7, the inspection system 2 has a vibration unit 22, a vibration detection unit 31, and the inspection device 12. The inspection device 12 has a data collection unit 42, a data analysis unit 52, and a diagnosis unit 62. The inspection device 12 is a device capable of carrying out the inspection method according to the second embodiment. The inspection device 12 diagnoses the presence or absence of an abnormality in the inspection object.

The vibration unit 22 has multiple vibration exciters. In the second embodiment, a case where the vibration unit 22 has two vibration exciters 22a and 22b is described. The two vibration exciters 22a and 22b excite different positions of the inspection object 100.

The vibration detection unit 31 in embodiment 2 is the same as that described in embodiment 1.

The data collection unit 42 collects measurement data regarding the vibration intensity detected by the vibration detection unit 31 when the vibrator 22a vibrates the inspection object and when the vibrator 22b vibrates the inspection object.

The data analysis unit 52 calculates correlation coefficients between the measurement data of three or more test objects measured when the vibrator 22a vibrates the test object and when the vibrator 22b vibrates the test object, and the measurement data of other test objects, and calculates the average of the correlation coefficients.

The diagnosis unit 62 compares and analyzes the average correlation coefficients for each test object calculated by the data analysis unit 52 when the vibrator 22a vibrates the test object and when the vibrator 22b vibrates the test object, diagnoses the presence or absence of abnormalities for each test object, and outputs the results.

<<2-2>> Operation The number of test objects may be three or more, but in the second embodiment, an example in which four test objects #1 to #4 are tested will be described.

8(A) and (B) are diagrams showing vibration modes 301, 302 of an inspection object excited by the vibrators 22a, 22b of the excitation unit 22 of the inspection system 2 according to embodiment 2, and a vibration detector 31a. The vibrators 22a, 22b of the excitation unit 22 perform excitation in a predetermined frequency band of 1 kHz or more. When one vibrator is vibrating the inspection object, the other vibrator does not vibrate the inspection object.

Figure 9 is a diagram showing the relationship between the vibrators 22a and 22b of the inspection system 2 according to the second embodiment and the vibration modes 301 and 320 shown in Figures 8(A) and (B). A case will be described in which two types of vibration modes, 301 and 302, are present as part of the vibration modes of the inspection object 100. The vibration mode 301 can be excited by the vibrator 22a whose installation location is not a node of the vibration mode 301, while it cannot be excited by the vibrator 22b whose installation location is a node of the vibration mode 301. The vibration mode 302 cannot be excited by the vibrator 22a whose installation location is a node of the vibration mode 302, while it can be excited by the vibrator 22b whose installation location is not a node of the vibration mode 302. Figure 9 shows in a table format the vibration modes that can be excited by each of the vibrators 22a and 22b (shown by a circle mark "○") and the vibration modes that cannot be excited (shown by a cross mark "X"). The vibration mode 301 can be excited by the vibrator 22a, and the vibration mode 302 can be excited by the vibrator 22b, and by using both the vibrators, it is possible to excite both the vibration modes 301 and 302. The vibration detection unit 31 detects the vibration intensity when each inspection object is vibrated in each of the cases where the vibrator 22a vibrates the inspection object and where the vibrator 22b vibrates the inspection object.

Figures 10 (A) to (D) are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by the data collection unit 42 of the inspection device 12 according to embodiment 2 as frequency/vibration intensity characteristics. The data collection unit 42 collects measurement data related to vibration intensity detected by the vibration detection unit 31 in each of the cases where the vibrator 22a vibrates the inspection object and where the vibrator 22b vibrates the inspection object. When the vibrator 22a vibrates the inspection object, the data collection unit 42 acquires measurement data 201 to 204 shown in Figures 2 (A) to (D). When the vibrator 22b vibrates the inspection object, the data collection unit 42 acquires measurement data 205 to 208 shown in Figures 10 (A) to (D). The measurement data 205 to 207 of the inspection objects #1 to #3, which are normal products, generally show the same tendency. On the other hand, the measurement data 208 of the inspection object #4, which is an abnormal product, has a different frequency of vibration peaks or a different number of vibration peaks compared to the measurement data 205-208 of the inspection objects #1-#3. In addition, when the vibration exciters 22a and 22b each excite the inspection object, the measurement data is different because the vibration modes that can be excited are different between the two. Since the effect of a crack on the vibration mode differs depending on the state, it may not affect one vibration mode while greatly affecting the other vibration mode. Therefore, by collecting measurement data excited by both the vibration exciters 22a and 22b, it becomes easier to obtain measurement data in which the response is significantly different between normal and abnormal products.

Figure 11 is a diagram showing the processing performed by the data analysis unit 52 of the inspection device 12 according to the second embodiment. The data analysis unit 52 performs analysis in the same manner as in the first embodiment, in each of the cases where the vibrator 22a vibrates the inspection object and where the vibrator 22b vibrates the inspection object. As shown in the first embodiment, when the vibrator 22a vibrates the inspection object, the average correlation coefficient for the inspection objects #1 to #3 is in the range of 0.83 to 0.86. The average correlation coefficient for the inspection object #4 is 0.59. On the other hand, the analysis result when the vibrator 22b vibrates the inspection object is as shown in Figure 11. The average correlation coefficient for the measurement data #1 to #3 is in the range of 0.80 to 0.81, while the average correlation coefficient for the inspection object #4 is 0.48, which is significantly different from the average correlation coefficient for the inspection objects #1 to #3.

Figure 12 is a frequency distribution diagram showing the relationship between the average and frequency of the correlation coefficients shown in Figure 11. The diagnosis unit 62 diagnoses the presence or absence of an abnormality based on the analysis results of the data analysis unit 52 in each of the cases where the vibrator 22a vibrates the inspection object and where the vibrator 22b vibrates the inspection object, and outputs the diagnosis result. The judgment result of the presence or absence of an abnormality when the vibrator 22a vibrates the inspection object is as described in the first embodiment. Figure 12 shows the distribution of the average correlation coefficients for the inspection objects #1 to #4 when the vibrator 22b vibrates the inspection object. The average correlation coefficients for the inspection objects #1 to #3, which are normal products, belong to the frequency 404 of 0.8 to 0.9. On the other hand, the average correlation coefficient for the inspection object #4, which is an abnormal product, belongs to the frequency 403 of 0.4 to 0.5. As in the case of the first embodiment, the inspection object #4 is judged to be abnormal by the clustering method. It can be seen that the difference in the average correlation coefficient between the normal inspection objects #1 to #3 and the abnormal inspection object #4 is larger when the vibration exciter 22b vibrates the inspection object. In other words, the accuracy of diagnosing the presence or absence of an abnormality is higher when the vibration exciter 22b vibrates the inspection object. As described above, the accuracy of diagnosis can be improved by using multiple vibration exciters.

<<2-3>> Effects When the inspection device 12 and the inspection method according to the second embodiment are used, in addition to the effects described in the first embodiment, the use of multiple vibrators makes it possible to collect measurement data that is more sensitive to abnormalities and to more easily obtain results with different correlation coefficients between normal products and abnormal products, thereby improving the accuracy of diagnosis.

In this example, the presence or absence of an abnormality could be correctly diagnosed when both vibrators 22a and 22b vibrated the inspection object, but if there is a large amount of noise in the measurement data, the diagnostic accuracy may deteriorate. Due to the above effects, when vibrator 22a vibrates the inspection object, there is a risk that the presence or absence of an abnormality cannot be correctly determined and an abnormal product cannot be detected. For this reason, the inspection device 12 and inspection method according to embodiment 2 are suitable for use when there is little noise. In addition, when vibrator 22b vibrates the inspection object, the difference in correlation coefficient between normal and abnormal products is larger than when vibrator 22a vibrates the inspection object, so the presence or absence of an abnormality can be correctly determined.

In the second embodiment, two vibrators are used, but three or more vibrators may be used. Also, one vibrator may be used to perform measurements multiple times at different positions to obtain data similar to that obtained when multiple vibrators are used.

<3> Third embodiment
<<3-1>> Configuration Fig. 13 is a schematic diagram showing the configuration of the inspection device 13 and the inspection system 3 according to the third embodiment. As shown in Fig. 13, the inspection system 3 has a vibration unit 22, a vibration detection unit 31, and the inspection device 13. The inspection device 13 has a data collection unit 43, a data analysis unit 53, and a diagnosis unit 63. The inspection device 13 is a device capable of carrying out the inspection method according to the third embodiment. The inspection device 13 diagnoses the presence or absence of an abnormality in the inspection target.

The vibration unit 22 and the vibration detection unit 31 are the same as those in the second embodiment. The data analysis unit 53 calculates the average correlation coefficient using the frequency-vibration intensity characteristics created by selecting the maximum value of the vibration intensity at each frequency from multiple measurement data measured on the same test object. The diagnosis unit 63 is the same as the diagnosis unit 61 in the first embodiment.

<<3-3>> Operation The number of inspection objects may be three or more, but in the third embodiment, an example in which four inspection objects #1 to #4 are inspected will be described. The vibrators 22a and 22b of the vibration unit 22 perform vibration in a predetermined frequency band of 1 kHz or more, as in the second embodiment. The vibration detection unit 31 detects the vibration intensity when each inspection object is vibrated, in each of the cases where the vibrator 22a vibrates the inspection object and the vibrator 22b vibrates the inspection object, as in the second embodiment.

As in the case of embodiment 2, the data collection unit 43 collects measurement data regarding the vibration intensity detected by the vibration detection unit 31 when the vibrator 22a vibrates the inspection object and when the vibrator 22b vibrates the inspection object.

The data analysis unit 53 calculates the average correlation coefficient using the frequency/vibration intensity characteristic created by selecting the maximum value of the vibration intensity at each frequency for multiple measurement data measured for the same test object. Figures 14 (A) to (D) are diagrams showing examples of measurement data of test objects #1 to #4 collected by the data collection unit of the inspection device according to embodiment 3 as frequency/vibration intensity characteristics. The measurement data 201 to 204 of test objects #1 to #4 are measurement data collected by the data collection unit 43 when the vibrator 22a vibrates the test object. The measurement data 205 to 208 of test objects #1 to #4 are measurement data collected by the data collection unit 43 when the vibrator 22b vibrates the test object. While the vibration intensity of each frequency of both detects similar vibration peaks, there are also cases where different vibration peaks are detected.

Figures 15 (A) to (D) are diagrams showing the maximum vibration intensity values of the measurement data 209 to 212 of inspection objects #1 to #4 collected by the data collection unit 43 of the inspection device 13 according to embodiment 3 as frequency/vibration intensity characteristics. This makes it possible to detect vibration peaks detected by both sets of measurement data and capture vibration peaks whose response changes due to abnormal parts. Next, the average correlation coefficient for each frequency/vibration intensity characteristic is calculated using the same method as in embodiment 1 for the frequency/vibration intensity characteristics that have been created.

As in the first embodiment, the diagnosis unit 63 diagnoses whether or not there is an abnormality based on the analysis results of the data analysis unit 53 and outputs the diagnosis result.

<<3-3>> Effects When the inspection device 13 and the inspection method according to the third embodiment are used, in addition to the effects described in the first embodiment, by selecting the maximum value of multiple measurement results at each frequency, it is possible to suppress missed detection of vibration modes that occur depending on the excitation position and measurement position, thereby improving the diagnostic accuracy.

In embodiment 3, two vibrators are used, but three or more may be used.

<4> Fourth embodiment
<<4-1>> Configuration Fig. 16 is a schematic diagram showing the configuration of the inspection device 14 and the inspection system 4 according to the fourth embodiment. As shown in Fig. 16, the inspection system 4 has a vibration applying unit 21, a vibration detecting unit 32, and the inspection device 14. The inspection device 14 has a data collecting unit 44, a data analyzing unit 54, and a diagnosing unit 64. The inspection device 14 is a device capable of carrying out the inspection method according to the fourth embodiment. The inspection device 14 diagnoses whether or not an abnormality exists in the inspection object. The inspection device 14 is composed of the vibration applying unit 21, the vibration detecting unit 32, the data collecting unit 44, the data analyzing unit 54, and the diagnosing unit 64, and diagnoses whether or not an abnormality exists in the inspection object.

The vibration unit 21 is the same as that described in embodiment 1. The vibration detection unit 32 has multiple vibration detectors. In embodiment 4, the vibration detection unit 32 has vibration detectors 32a and 32b.

The data collection unit 44 collects measurement data regarding vibration intensity when the vibration detector 32a detects vibration and when the vibration detector 32b detects vibration.

The data analysis unit 54 calculates the correlation coefficient between each of the measurement data of three or more test subjects and the measurement data of other test subjects when the vibration detector 32a detects vibration and when the vibration detector 32b detects vibration, and calculates the average of the correlation coefficients.

The diagnosis unit 64 compares and analyzes the average correlation coefficients for each test object calculated by the data analysis unit 54 when the vibration detector 32a detects vibration and when the vibration detector 32b detects vibration, diagnoses the presence or absence of an abnormality for each test object, and outputs the results.

<<4-2>> Operation The number of inspection objects may be three or more, but in the embodiment 4, an example in which four inspection objects #1 to #4 are inspected will be described as in the same manner as in the embodiment 1. The vibrator 21a of the vibration unit 21 performs vibration in a predetermined frequency band of 1 kHz or more as in the embodiment 1.

17(A) and (B) are diagrams showing the vibration mode of the inspection object vibrated by the vibrator 21a of the inspection system 4 according to the fourth embodiment and the vibration detectors 32a and 32b. Each of the vibration detectors 32a and 32b of the vibration detection unit 32 detects the vibration intensity when each inspection object is vibrated. As in the case of the second embodiment, a case will be described in which two types of vibration modes, vibration mode 301 and vibration mode 302, are present as part of the vibration mode of the inspection object 100. The vibration mode 301 can be detected by the vibration detector 32a whose installation location is not the node of the vibration mode 301, but cannot be detected by the vibration detector 32b whose installation location is the node of the vibration mode 301. The vibration mode 302 cannot be detected by the vibration detector 32a whose installation location is the node of the vibration mode 302, but can be detected by the vibration detector 32b whose installation location is not the node of the vibration mode 302.

Figure 18 is a diagram showing the relationship between the vibration detectors of the inspection system according to embodiment 4 and the vibration modes shown in Figures 17(A) and (B). Vibration modes 301 and 302 can be detected only by vibration detectors 32a and 32b, respectively, and both vibrations of vibration modes 301 and 302 can be detected by using both vibration detectors 32a and 32b.

Figures 19 (A) to (D) are diagrams showing examples of measurement data of inspection objects #1 to #4 collected by the data collection unit 44 of the inspection device according to embodiment 4 as frequency/vibration intensity characteristics. The data collection unit 44 collects measurement data related to vibration intensity when vibration detector 32a detects vibration and when vibration detector 32b detects vibration. When vibration detector 32a detects vibration, measurement data 201 to 204 shown in Figures 2 (A) to (D) are obtained. When vibration detector 32b detects vibration, measurement data 213 to 216 shown in Figures 19 (A) to (D) are obtained.

The measurement data 213-215 of the inspection objects #1-#3, which are normal products, generally show the same tendency. On the other hand, the measurement data 216 of the inspection object #4, which is an abnormal product, differs in the frequency of the vibration peak or the number of vibration peaks compared to the measurement data 213-215 of the inspection objects #1-#3. In addition, when each of the vibration detectors 32a and 32b detects vibration, the measurement data is different because the vibration modes that can be detected are different. Since the effect of a crack on the vibration mode differs depending on the state, it may not affect one vibration mode while greatly affecting the other vibration mode. Therefore, by collecting the measurement data detected by both the vibration detectors 32a and 32b, it becomes easier to obtain measurement data that responds significantly differently between normal and abnormal products.

Figure 20 is a diagram showing the processing performed by the data analysis unit 54 of the inspection device 14 according to the fourth embodiment. The data analysis unit 54 performs analysis in the same manner as in the first embodiment in each of the cases where the vibration detector 32a detects vibration and where the vibration detector 32b detects vibration. As shown in the first embodiment, when the vibration detector 32a detects vibration, the average correlation coefficient for the inspection objects #1 to #3 is in the range of 0.83 to 0.86. The average correlation coefficient for the inspection object #4 is 0.59. On the other hand, the analysis result when the vibration detector 32b detects vibration is shown in Figure 20. The average correlation coefficient for the measurement data #1 to #3 is 0.81, while the average correlation coefficient for the inspection object #4 is 0.48, which is significantly different from the average correlation coefficient for the inspection objects #1 to #3.

Figure 21 is a frequency distribution diagram showing the relationship between the average and frequency of the correlation coefficients shown in Figure 20. The diagnosis unit 64 diagnoses the presence or absence of an abnormality based on the analysis results of the data analysis unit 54 in each of the cases where the vibration detector 32a detects vibration and where the vibration detector 32b detects vibration, and outputs the diagnosis result. The judgment result of the presence or absence of an abnormality when the vibration detector 32a detects vibration is as described in the first embodiment. As shown in Figure 21, the average of the correlation coefficients for the inspection objects #1 to #3, which are normal products, belongs to the frequency 406 of 0.8 to 0.9. On the other hand, the average of the correlation coefficients for the inspection object #4, which is an abnormal product, belongs to the frequency 405 of 0.4 to 0.5. As in the case of the first embodiment, the inspection object #4 is judged to be abnormal by the clustering method. It can be confirmed that the difference in the average correlation coefficients between the inspection objects #1 to #3, which are normal products, and the inspection object #4, which is an abnormal product, is larger when the vibration detector 32b detects vibration. In other words, the accuracy of diagnosing the presence or absence of an abnormality is higher when the vibration detector 32b detects vibration. As described above, by using a plurality of vibration detectors, the accuracy of diagnosis can be improved.

4-3 Effects According to the inspection device 14 and the inspection method of the fourth embodiment, in addition to the effects described in the first embodiment, by using multiple vibration detectors, it is possible to collect measurement data with higher sensitivity to abnormalities and obtain results with different correlation coefficients between normal and abnormal products, thereby improving the accuracy of diagnosis. Note that, although an example has been described in which the presence or absence of an abnormality can be correctly diagnosed in both cases where the vibration detectors 32a and 32b detect vibration, if there is a large amount of noise in the measurement data, the accuracy of diagnosis may deteriorate. For this reason, the inspection device 14 and the inspection method of the fourth embodiment are suitable for use when there is little noise.

Due to the above effects, when vibration detector 32a detects vibration, there is a risk that it will not be able to correctly determine whether there is an abnormality and will not be able to detect an abnormal product. On the other hand, when vibration detector 32b detects vibration, the difference in correlation coefficient between normal and abnormal products is larger than when vibration detector 32a detects vibration, so it is still possible to correctly determine whether there is an abnormality.

<4-4> Modifications Although two vibration detectors are used in the fourth embodiment, three or more vibration detectors may be used. Although one vibration exciter is used in the fourth embodiment, two or more vibration exciters may be used as shown in the second embodiment. In this case, measurement data according to the number of combinations of two or more vibration exciters and two or more vibration detectors can be obtained. Moreover, by performing measurements multiple times at different positions using one vibration exciter and one vibration detector, data similar to that obtained when multiple vibration exciters and vibration detectors are used may be obtained.

In addition, in the fourth embodiment, as in the third embodiment, the data analysis unit 54 may diagnose the presence or absence of an abnormality by calculating an average correlation coefficient using a frequency-vibration intensity characteristic created by selecting the maximum value of vibration intensity at each frequency of multiple measurement data measured on the same test object in the measurement data when each of the vibration detectors 32a and 32b detects an abnormality. In this case, the diagnosis unit 64 is configured to diagnose the presence or absence of an abnormality based on the analysis result of the data analysis unit 54 and output the diagnosis result, as in the first embodiment.

<5> Fifth embodiment
<<5-1>> Configuration Fig. 22 is a schematic diagram showing the configuration of the inspection device 15 and the inspection system 5 according to the fifth embodiment. As shown in Fig. 22, the inspection system 5 has a vibration unit 21, a vibration detection unit 31, and the inspection device 15. The inspection device 15 has a data collection unit 45, a data analysis unit 55, and a diagnosis unit 65. The inspection device 15 is a device capable of carrying out the inspection method according to the fifth embodiment. The inspection device 15 diagnoses the presence or absence of an abnormality in the inspection target.

In the inspection system 5, the vibrator 21a of the vibration unit 21 and the vibration detector 31a of the vibration detection unit 31 are installed on a conveying line that conveys the inspection objects 101 to 106 in the conveying direction. The inspection device 15 has a data collection unit 45, a data analysis unit 55, and a diagnosis unit 65, and diagnoses the presence or absence of an abnormality in the inspection object. The vibration unit 21, the vibration detection unit 31, and the data collection unit 45 are the same as those described in the first embodiment. The data analysis unit 55 executes a process of calculating an evaluation value of the correlation coefficient of a predetermined number of inspection objects close to the current time, that is, a predetermined number of inspection objects collected most recently, among the measurement data collected by the data collection unit 45. The number of inspection objects may be three or more, but the data analysis unit 55 analyzes the measurement data for the four inspection objects for which data was collected most recently, in the same manner as in the first embodiment. The diagnostic unit 65 diagnoses the presence or absence of abnormalities for the four most recently collected test objects based on the analysis results of the data analysis unit 55 in the same manner as in the first embodiment, and outputs the diagnosis results.

In the fifth embodiment, an example of inspecting inspection objects 101 to 106 on a conveyor line will be described. The excitation unit 21 excites each inspection object traveling on the conveyor line with a predetermined frequency band of 1 kHz or more, as in the first embodiment.

The vibration detection unit 31 detects the vibration intensity when each inspection object traveling on the conveying line is vibrated, as in embodiment 1.

As in embodiment 1, the data collection unit 45 collects measurement data regarding the vibration intensity detected by the vibration detection unit 31.

The data analysis unit 55 analyzes the measurement data collected most recently for the four test objects in the same manner as in embodiment 1. Thus, when measurement data for test object 104 is collected, the most recently obtained measurement data for test objects 101 to 104 is analyzed.

The diagnosis unit 65 diagnoses the presence or absence of abnormalities in the four most recently collected test objects based on the analysis results of the data analysis unit 55, as in embodiment 1, and outputs the diagnosis results.

By performing the above process each time measurement data is collected for each inspection object, the presence or absence of abnormalities in the inspection objects traveling on the conveyor line can be diagnosed.

<<5-3>> Effects By using the inspection device 15 and the inspection method according to embodiment 5, in addition to the effects described in embodiment 1, it is possible to realize in-line inspection since it becomes possible to determine the presence or absence of an abnormality each time measurement data of the inspection target is collected.

In addition, in embodiment 5, the diagnosis of the presence or absence of abnormalities was performed using the four most recently collected test objects, but the number of test objects to be diagnosed for the presence or absence of abnormalities may be any number as long as it is three or more.

<<6>> Hardware Configuration Fig. 23 is a diagram showing an example of a hardware configuration of the inspection devices 11 to 15 and the inspection systems 1 to 5 according to the first to fifth embodiments. In the example of Fig. 23, the inspection devices 11 to 15 according to the first to fifth embodiments have a processor 501 such as a CPU, a memory 502 as a storage device, and an interface 503. The excitation unit 21 (or 22) and the vibration detection unit 31 (or 32) are connected to the interface 503. The processor 501 executes a program (e.g., the inspection program according to the first to fifth embodiments) as software stored in the memory 502. The inspection devices 11 to 15 may be computers.

Figure 24 is a diagram showing another example of the hardware configuration of the inspection devices 11 to 15 and the inspection systems 1 to 5 according to embodiments 1 to 5. In the example of Figure 24, the inspection devices 11 to 15 according to embodiments 1 to 5 have a processing circuit 504 and an interface 503. The vibration unit 21 (or 22) and the vibration detection unit 31 (or 32) are connected to the interface 503. The processing circuit 504 is, for example, a semiconductor integrated circuit, a system LSI, an FPGA (field-programmable gate array), etc. that constitute the inspection devices 11 to 15 according to embodiments 1 to 5. In addition, one or more processors, one or more memories, and a processing circuit may work together to realize the functions of the inspection devices 11 to 15.

1-5 inspection system, 11-15 inspection device, 21, 22 vibration unit, 21a, 22a, 22b vibration exciter, 31, 32 vibration detection unit, 31a, 32a, 32b vibration detector, 41-45 data collection unit, 51-55 data analysis unit, 61-65 diagnosis unit, 71 moving body, 100-106 inspection object, 201-216 measurement data, 301, 302 amplitude of vibration mode, 401-406 frequency of distribution of average correlation coefficient.

Claims (17)

a data collection unit that collects, as measurement data, the measurement values for each frequency of vibration for each of N (N is an integer of 3 or more) test objects as the object from a vibration detection unit that detects vibration of the object excited by a vibration excitation unit and outputs measurement values of the vibration;
a data analysis unit that executes a process of calculating a similarity of the measurement data of two of the N inspection objects for each combination of two of the N inspection objects, and calculates an evaluation value that is a statistic of the similarity for each of the N inspection objects;
a diagnosis unit that compares the evaluation values of the N test objects with each other and diagnoses whether or not each of the N test objects has an abnormality based on a result of the comparison;
An inspection device comprising: The inspection device according to claim 1 , wherein the similarity is a correlation coefficient or a Euclidean distance. 3. The inspection apparatus according to claim 1, wherein the statistic is a sum, an average, a median, or a trimmed mean. 4. The inspection device according to claim 1, wherein the data collection unit collects at least one of an intensity and a phase of the vibration for each frequency as the measurement value. The inspection device according to claim 1 , wherein the diagnosis unit performs the diagnosis based on a frequency distribution indicating a relationship between a range of the evaluation value for each of the N inspection objects and a frequency of the similarity included in the range. 6. The inspection device according to claim 1, wherein the data collection unit collects, as the measurement data, the measurement values output from the vibration detection unit that detects vibrations in a predetermined frequency band. 7. The inspection device according to claim 1, wherein the data collection unit collects the measurement values for each frequency of vibration applied to different positions of each of the N inspection objects by a plurality of vibration exciters included in the vibration exciter as measurement data for each of the plurality of vibration exciters. 8. The inspection device according to claim 7, wherein the data analysis unit selects the measurement data in which the vibration intensity is large for each vibration frequency from the measurement data of each of the multiple vibrators, and executes the process of calculating the similarity using a frequency/vibration intensity characteristic created from the selected measurement data for each frequency. 8. The inspection device according to claim 1 , wherein the data collection unit collects, as the measurement data, the measurement values measured at different positions of each of the N inspection objects by a plurality of vibration detectors included in the vibration detection unit. The inspection device according to claim 9, characterized in that the data analysis unit selects the measurement data in which the vibration intensity is large for each frequency of the vibration from the measurement data of each of the multiple vibration detectors, and performs the process of calculating the similarity using a frequency/vibration intensity characteristic created from the selected measurement data for each frequency. 11. The inspection device according to claim 1, wherein the data analysis unit executes a process of calculating the evaluation value of the degree of similarity of a predetermined number of inspection objects close to a current point in time from the measurement data collected by the data collection unit. An inspection device according to any one of claims 1 to 11;
The vibration unit,
The vibration detection unit;
An inspection system comprising: The vibration detector further includes a moving body that holds the vibration excitation unit and the vibration detection unit,
The inspection system according to claim 12 , wherein a position at which vibration is applied by the vibration applying unit and a position at which vibration is detected by the vibration detecting unit are changed by the movement of the moving body. The thickness of the vibration part is 100 mm or less,
14. The inspection system according to claim 12, wherein the vibration detection unit has a thickness of 100 mm or less. The vibration unit includes a piezoelectric element,
The inspection system according to claim 12 , wherein the vibration detection unit includes a piezoelectric element. a step of collecting, as measurement data, the measurement values for each frequency of vibration for each of N (N is an integer of 3 or more) test objects as the object from a vibration detection unit that detects vibration of the object excited by a vibration excitation unit and outputs measurement values of the vibration;
a step of calculating a similarity of the measurement data of two of the N inspection objects for each combination of two of the N inspection objects, and calculating an evaluation value that is a statistic of the similarity for each of the N inspection objects;
comparing the evaluation values of the N test objects with each other, and diagnosing whether or not there is an abnormality in each of the N test objects based on a result of the comparison;
An inspection method comprising the steps of: a step of collecting, as measurement data, the measurement values for each frequency of vibration for each of N (N is an integer of 3 or more) test objects as the object from a vibration detection unit that detects vibration of the object excited by a vibration excitation unit and outputs measurement values of the vibration;
a step of calculating a similarity of the measurement data of two of the N inspection objects for each combination of two of the N inspection objects, and calculating an evaluation value that is a statistic of the similarity for each of the N inspection objects;
comparing the evaluation values of the N test objects with each other, and diagnosing whether or not there is an abnormality in each of the N test objects based on a result of the comparison;
An inspection program characterized by causing a computer to execute the above steps.

Images