Product inspection support system and product inspection support method

Abstract

To rapidly specify causes of abnormality.SOLUTION: A product inspection support system 10 for supporting work for successively inspecting multiple object products includes a sound pickup section 11 and a determination section 12. The sound pickup section 11 picks up an operation sound from an object product. The determination section 12 performs determination that determines whether the object product is abnormal and includes determination of a type of abnormality in the object product determined to be abnormal from the operation sound picked up by the pickup section 11. The determination section 12 performs the determination by using a determination model generated by machine learning using an operation sound for a normal object product and an operation sound for each type concerned for an abnormal object product.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a technology for assisting the work in the product inspection process.

Background Art

[0002] There is known a technology for performing a noise inspection of the operating sound of a product based on reference operating sound data created from the operating sound of a good product (see, for example, Patent Document 1). Further, there is known a technology for determining the abnormality of an object to be inspected based on the vibration level, determining whether the cause of the abnormality is a resonance system abnormality, and when it is determined to be another cause, obtaining the cause based on the frequency axis component by a three-layer determination (see, for example, Patent Document 2).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a method of identifying the true cause of an abnormality by repeating the estimation of the cause of the product abnormality and the determination of the correctness of the estimation, it may take time to identify the true cause of the abnormality.

Means for Solving the Problems

[0005] In one embodiment, a product inspection support system that supports the operation of sequentially inspecting a plurality of target products includes a sound collection unit and a determination unit. The sound collection unit collects the operating sound from the target product. The determination unit performs the determination as to whether there is an abnormality in the target product, including discrimination of the type of abnormality of the target product determined to have an abnormality, based on the operating sound. This determination unit performs the determination using a determination model generated by machine learning that uses the operating sound of a normal target product and the operating sound for each type of the target product with an abnormality.

Advantages of the Invention

[0006] According to the above aspect, it is possible to quickly identify the cause of the abnormality.

Brief Description of the Drawings

[0007]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Modes for Carrying Out the Invention

[0008] Hereinafter, embodiments will be described in detail with reference to the drawings.

[0009] First, the configuration of the product inspection support system 10 as an example of an embodiment will be described with reference to FIG. 1.

[0010] The product inspection support system 10 in FIG. 1 supports the product inspection work at the time of shipment of a plurality of target products sequentially manufactured through manufacturing processes 1, 2,..., N. This product inspection support system 10 includes a sound collection unit 11, a determination unit 12, a notification unit 13, and a storage unit 14, and is connected to a communication network 20. Output devices 30-1, 30-2,..., 30-N are also connected to the communication network 20. The output devices 30-1, 30-2,..., 30-N are respectively installed at the sites where the work of each manufacturing process of the manufacturing processes 1, 2,..., N is performed, and are used by the process responsible persons of each manufacturing process.

[0011] The sound collection unit 11 is for collecting the operating sound from the target product, and is, for example, a microphone. The operating sound collected by the sound collection unit 11 is sent to the determination unit 12 via the communication network 20.

[0012] In this embodiment, the sound collection unit 11 is installed at the product shipment test site where product inspections for a plurality of manufactured target products are sequentially performed, and the other components of the product inspection support system 10 are installed at a location different from the sound collection unit 11. For this reason, in FIG. 1, the sound collection unit 11 and the other components of the product inspection support system 10 are connected via the communication network 20.

[0013] Note that the other components of the product inspection support system 10 may also be installed at the product shipment test site, and the sound collection unit 11 and the other components of the product inspection support system 10 may be directly connected without passing through the communication network 20. In the case of this configuration, the operating sound collected by the sound collection unit 11 is directly sent to the determination unit 12.

[0014] The determination unit 12 determines whether there is an abnormality in the target product from the operation sound collected by the sound collection unit 11. In this determination, the type of abnormality of the target product determined to have an abnormality is also discriminated. The method of this determination will be described later.

[0015] The notification unit 13 issues a notification of the occurrence of an abnormality in response to the appearance of the target product determined to have an abnormality.

[0016] The storage unit 14 includes a notification determination master DB 14a and a countermeasure master DB 14b. Note that "DB" is an abbreviation for Database.

[0017] The notification of the occurrence of an abnormality by the notification unit 13 is performed using the information shown in the notification determination master DB 14a and the countermeasure master DB 14b, respectively. The details of the notification determination master DB 14a and the countermeasure master DB 14b, as well as the details of the method of notifying the occurrence of an abnormality by the notification unit 13, will be described later.

[0018] The product inspection support system 10 includes each of the above-described components, and directly determines the type of abnormality of the target product determined to have an abnormality from the operation sound as well as determines whether there is an abnormality in the target product, so that the cause of the abnormality can be quickly identified.

[0019] Note that the product inspection support system 10 may further include a detection unit 15 and a control unit 16 as shown in FIG. 1.

[0020] The operation sound collected by the sound collection unit 11 is also sent to the detection unit 15 together with the determination unit 12. The detection unit 15 detects a temporary abnormal sound generated by a sound source other than the target product from the operation sound collected by the sound collection unit 11. The sound of a chime emitted from a speaker installed at the manufacturing site of the target product as a signal for starting, ending, pausing, or resuming work is an example of a temporary abnormal sound generated by a sound source other than the target product.

[0021] When the volume of the operating sound picked up by the sound collection unit 11 is greater than a predetermined threshold value, for example, the detection unit 15 outputs a detection result indicating that an abnormal sound has been detected from the operating sound. Further, when the abnormal sound is a known one, for example, an acoustic feature amount is obtained in advance from the frequency spectrum of this abnormal sound. In this case, the detection unit 15 determines whether the acoustic feature amount of the abnormal sound is included in the frequency spectrum of the operating sound picked up by the sound collection unit 11, and outputs a detection result indicating that an abnormal sound has been detected from the operating sound according to this determination result.

[0022] The control unit 16 causes the detection unit 15 to detect an abnormal sound, and controls the determination unit 12 according to the detection of the abnormal sound by the detection unit 15, so that the determination unit 12 determines whether there is an abnormality between the operating sound picked up by the sound collection unit 11 again from the target product where the operating sound with the detected abnormal sound was picked up and the operating sound picked up.

[0023] The abnormal sound detected by the detection unit 15 may reduce the accuracy of the determination by the determination unit 12. By performing the above-described control by the control unit 16, the use of the determination result of the determination unit 12, which may have its accuracy reduced due to the influence of the abnormal sound, is avoided.

[0024] Note that the details of the control method by the control unit 16 will be described later.

[0025] Next, a method for determining the presence or absence of an abnormality by the determination unit 12 will be described with reference to FIG. 2.

[0026] The sound collection unit 11 picks up the operating sound from the target product. The operating sound signal representing the picked-up operating sound is sent from the sound collection unit 11 to the determination unit 12.

[0027] In this embodiment, the determination unit 12 determines whether there is an abnormality in the target product using a determination model. The determination model is a model that determines whether there is an abnormality in the target product using the signal of the operating sound of the target product, and classifies the target product determined to have an abnormality into each type of abnormality. Here, the type of abnormality is, for example, the cause of the abnormality. That is, classifying the target product determined to have an abnormality according to the cause of the abnormality is an example of classifying the target product according to the type of abnormality.

[0028] In this embodiment, as the determination model, for example, a machine learning model configured using a three-layer neural network is used. For this machine learning model, machine learning is performed using, as teacher data, a combination of the signal of the operating sound recorded from the target product, the presence or absence of an abnormality in the target product, and the type of abnormality when there is an abnormality, thereby generating a determination model. The determination unit 12 classifies the target product to be inspected into either a normal target product or each target product according to the type of abnormality according to the output from the determination model for the input of the signal of the operating sound of the target product to be inspected.

[0029] Note that in the example of FIG. 2, the determination unit 12 is shown to classify the target product into any one of "normal" and three types of abnormalities, "abnormality A", "abnormality B", and "abnormality C", but the types of abnormalities to be classified are not limited to three.

[0030] In addition, the determination model used in the determination unit 12 may be a model that extracts the feature amount of the operation sound from the signal of the operation sound of the product to be inspected. For this model, machine learning is performed using the signal of the operation sound of a normal target product and the signal of the operation sound for each type of abnormality of an abnormal target product, and the feature amount of the operation sound for each of the normal target product and the target product for each type of abnormality is extracted. When using this determination model, the determination unit 12 causes the determination model to extract the feature amount of the operation sound from the signal of the operation sound of the product to be inspected, and selects the one that is most similar among the normal target product and the target product for each type of abnormality from the extracted feature amounts. As the index of similarity used for this selection, any of various similarity indexes such as the cosine similarity may be used. The determination unit 12 classifies the product to be inspected into any one of the normal target product and each target product for each type of abnormality according to the selection result based on the output of this determination model.

[0031] Next, the details of the notification determination master DB 14a and the countermeasure master DB 14b, as well as the details of the method for notifying the occurrence of an abnormality by the notification unit 13, will be described.

[0032] FIG. 3 shows an example of the registered data in the notification determination master DB 14a. This registered data represents the appearance method of the target products determined by the determination unit 12 to have the same type of abnormality in the sequential inspection of a plurality of target products. The notification unit 13 issues a notification of the occurrence of an abnormality when a predetermined condition determined by this appearance method is satisfied.

[0033] In each record of the notification determination master DB 14a, the data in the "abnormality classification" field is associated with the data in the "continuous" and "ratio" fields respectively. Note that a number for identifying each record is registered in the "No." field.

[0034] In the "abnormality classification" field, information on the name for identifying the type of abnormality of the target product with an abnormality is registered.

[0035] In the "Continuous" field, the number of times that target products, which are determined by the determination unit 12 to have the same type of abnormality as the type identified by the "Abnormality Classification" information, continuously appear in the sequential inspection of a plurality of target products is registered. The value of this number is equal to the value of the number of target products that are continuously determined to have the same type of abnormality. The notification unit 13 is configured to issue a notification of abnormality when the condition (the first condition) that target products determined to have the same type of abnormality appear continuously for that number is satisfied.

[0036] In the "Ratio" field, the appearance rate of target products, which are determined by the determination unit 12 to have the same type of abnormality as the type identified by the "Abnormality Classification" information, in the sequential inspection of a plurality of target products is registered as information on the conditions for issuing a notification of abnormality. The notification unit 13 is configured to issue a notification of abnormality when the condition (the second condition) that target products determined to have the same type of abnormality appear at an appearance rate equal to or higher than the value indicated in the "Ratio" is satisfied, even if the above-described first condition is not satisfied.

[0037] In the example of FIG. 3, for example, in the records where "No." is "1", the fields of "Abnormality Classification", "Continuous", and "Ratio" respectively have registered data of "Abnormality A", "3", and "15.0". This record represents, as the conditions for issuing a notification of abnormality, the case where target products determined to have an abnormality of the type "Abnormality A" continuously appear "3" times in the sequential inspection, or the case where the appearance rate of the target products in the sequential inspection is 15.0% or more.

[0038] A specific example in which the notification unit 13 issues a notification of abnormality according to the conditions indicated by the registered data of the notification determination master DB14a illustrated in FIG. 3 will be described. FIG. 4 shows an example of the determination results by the determination unit 12 in the sequential inspection of a plurality of target products.

[0039] In the table shown in FIG. 4, each row represents the result of each inspection, and "number of inspections" and "abnormality classification" are associated with each other in each row. Among these, the "number of inspections" is the number of target products that have completed the inspection. The "abnormality classification" indicates information of a name for identifying the type of abnormality of the target product as the determination result by the determination unit 12 for the target product to be inspected. However, when it is determined that there is no abnormality in the target product, the word "normal" is indicated.

[0040] In the table of FIG. 4, at the stage where the "number of inspections" is "15", two consecutive target products with abnormalities where the "abnormality classification" is "abnormality A" appear. Also, at this stage, the appearance rate of the target products with abnormalities where the "abnormality classification" is "abnormality A" is 20.0 percent. The appearance rate is defined as the ratio of the number of target products with abnormalities to the number of the most recent predetermined number of target products at each stage. In this embodiment, this predetermined number is set to 10. This appearance rate satisfies "15.0"% or more, which is the value of "ratio" when the "abnormality classification" is "abnormality A" in the notification determination master DB14a shown in FIG. 3. Therefore, at this stage, the notification unit 13 issues a notification of the occurrence of an abnormality.

[0041] In this way, instead of immediately issuing a notification of the occurrence of an abnormality every time a target product determined to have an abnormality appears, by issuing the notification when a predetermined condition shown in the registered data of the notification determination master DB14a is satisfied, false notifications due to incorrect determinations are reduced.

[0042] Note that instead of setting the condition for the notification unit 13 to issue a notification as satisfying either the first condition or the second condition described above, a condition (third condition) of satisfying both the first condition and the second condition may be set as the condition for the notification unit 13 to issue a notification.

[0043] Next, the countermeasure master DB14b will be described. FIG. 5 shows an example of the registered data of the countermeasure master DB14b. This registered data indicates the information of the notification destination to which the notification unit 13 issues a notification of the occurrence of an abnormality and the information regarding the occurred abnormality.

[0044] In each record of the countermeasure master DB 14b, the data in the "abnormality classification" field is associated with the data in the "notification destination", "cause", and "countermeasure" fields respectively. Note that a number for identifying each record is registered in the "No." field.

[0045] In the "abnormality classification" field, similar to the notification determination master DB 14a, information on the name for identifying the type of abnormality of the target product with an abnormality is registered.

[0046] In the "notification destination" field, information on the notification destination when notifying of the occurrence of an abnormality is registered. For example, when "manufacturing process 1" is shown as the "notification destination", in FIG. 1, it indicates that the output device 30-1 installed in "manufacturing process 1" is the notification destination for the notification of the occurrence of an abnormality. Note that as this "notification destination", information indicating the manufacturing process in which the type of abnormality shown in the "abnormality classification" has occurred in the target product is registered.

[0047] The "cause" field shows information on the cause of the type of abnormality shown in the "abnormality classification" occurring in the target product.

[0048] The "countermeasure" field shows information on the countermeasure for eliminating the occurrence of the type of abnormality shown in the "abnormality classification" in the target product.

[0049] Each record in the countermeasure master DB 14b has the above information, and the occurrence location where the type of abnormality shown in the "abnormality classification" occurs in the target product, the cause of the occurrence of the abnormality, and the countermeasure to the abnormality are respectively associated as the "notification destination", "cause", and "countermeasure". These information are obtained from the know-how of reporters and skilled process workers regarding abnormalities that have occurred in the past.

[0050] For example, when the notification unit 13 issues a notification of an abnormality occurrence for an abnormality where the type of abnormality of the target product is "Abnormality A", in the example of FIG. 5, the information of the record where "Abnormality A" is shown as "Abnormality Classification" and "No." is "1" is referred to. In this record, as the information of "Notification Destination", "Cause", and "Countermeasure", the data of "Manufacturing Process 1", "Temperature Abnormality", and "Set the temperature within the preset range value" are registered respectively. At this time, the notification unit 13 issues a notification including the information of "Abnormality Classification" and these pieces of information of "Notification Destination", "Cause", and "Countermeasure" to the output device 30-1 installed in "Manufacturing Process 1".

[0051] Here, FIG. 6 will be described. FIG. 6 shows an example screen of an abnormality occurrence notification screen that is displayed as output on the output devices 30-1, 30-2,..., 30-N that have received the above-mentioned notification from the notification unit 13.

[0052] The example screen of FIG. 6 is what is displayed when the output device 30-1 receives the notification, and each piece of information of "Abnormality Classification", "Notification Destination", "Cause", and "Countermeasure" included in the notification of abnormality occurrence from the notification unit 13 is displayed together with the display of "Abnormality Occurrence". Note that the information of "Notification Destination" included in the notification from the notification unit 13 is displayed in the column of "Occurrence Process" in this example screen.

[0053] Due to the display on the output device 30-1 of this screen, information is presented to the process responsible person in Manufacturing Process 1 that an abnormality of type "Abnormality A" has occurred in the target product due to the "Temperature Abnormality" in "Manufacturing Process 1", and it is necessary to take the countermeasure of "Set the temperature within the preset range value". As a result, the process responsible person can promptly take various measures necessary to eliminate the abnormality that has occurred in the target product, such as stopping Manufacturing Process 1, arranging replacement parts, and planning and executing repairs, by referring to this information notified by the notification unit 13.

[0054] Note that the notification unit 13 may attach a document in which information on measures for eliminating the abnormality of the target product according to the type of abnormality, such as a trouble shooting report, is described in detail to the notification of the occurrence of the abnormality. Further, hyperlink information for accessing the information may be included in the notification of the occurrence of the abnormality. Further, the information may be displayed together on the abnormality occurrence notification screen displayed on the output devices 30-1, 30-2, …, 30-N that have received the notification from the notification unit 13. Alternatively, as in the screen example of FIG. 6, a button for receiving an instruction to access the "trouble shooting report" may be arranged on the abnormality occurrence notification screen, and access to the information may be enabled in response to a pressing operation on the button.

[0055] Next, a hardware configuration example of the information processing apparatus 40 (computer) will be described with reference to FIG. 7. The product inspection support system 10 of FIG. 1 may be configured using this information processing apparatus 40.

[0056] In FIG. 1, the information processing apparatus 40 includes a CPU 41, a memory 42, an input device 43, a display device 44, an auxiliary storage device 45, and a communication I / F 46. All of these components are connected to an internal bus 47 and can exchange data with each other. Note that "CPU" is an abbreviation for Central Processing Unit, and "I / F" is an abbreviation for Interface.

[0057] The CPU 41 provides various functions that the determination unit 12, the notification unit 13, the detection unit 15, and the control unit 16 in the product inspection support system 10 respectively have by executing a program using the memory 42.

[0058] The memory 42 is, for example, a semiconductor memory and includes a RAM area and a ROM area. The RAM area is used as a storage area for temporarily storing various data when the CPU 41 executes various programs. The ROM area is an area where various programs and data executed by the CPU 41 are stored, and for example, a non-volatile memory is used. Note that "RAM" is an abbreviation for Random Access Memory, and "ROM" is an abbreviation for Read Only Memory.

[0059] The input device 43 is, for example, a keyboard, a pointing device, etc., and is used for inputting instructions or information from the user.

[0060] The display device 44 is used, for example, for inquiries or instructions to the user and for displaying and outputting processing results.

[0061] The auxiliary storage device 45 is a magnetic disk device, an optical disk device, a magneto-optical disk device, etc., and may be, for example, a hard disk drive or a flash memory. The information processing device 40 can store programs and data in the auxiliary storage device 45 and load them into the memory 42 for use. When the product inspection support system 10 is configured using the information processing device 40, the auxiliary storage device 45 provides the function as the storage unit 14 that stores the notification determination master DB 14a and the countermeasure master DB 14b.

[0062] The communication I / F 46 transmits and receives data via the communication network 20, for example, according to an instruction sent from the CPU 41 as necessary. When the product inspection support system 10 is configured using the information processing device 40, the communication I / F 46 receives the signal data of the operation sound of the target product sent from the sound collection unit 11 and transmits the notifications made by the notification unit 13 to the output devices 30-1, 30-2,..., 30-N.

[0063] The information processing apparatus 40 shown in FIG. 7 has the above-described hardware configuration. Note that the information processing apparatus 40 does not necessarily include all of the components shown in FIG. 7, and some components may be omitted according to the application or conditions. For example, when configuring the product inspection support system 10 using the information processing apparatus 40, the display device 44 may be omitted.

[0064] Next, the product inspection support process performed by the product inspection support system 10 in FIG. 1 will be described. FIG. 8 is a flowchart showing the processing contents of an example of the product inspection support process.

[0065] When configuring the product inspection support system 10 using the information processing apparatus 40 shown in FIG. 7, a product inspection support program describing the processing contents of the product inspection support process may be created, and the CPU 41 may be caused to execute the program.

[0066] When the process in FIG. 8 is started, first, in S101, a process of acquiring a signal of the operation sound sent from the sound collection unit 11 that has collected the operation sound of the product to be inspected via the communication network 20 is performed.

[0067] In S102, a process of inputting the signal of the operation sound collected in the process of S101 to the determination model generated as described above to determine the presence or absence of an abnormality in the product to be inspected is performed. Then, in the subsequent S103, a process of determining whether the output of the determination model indicates that there is an abnormality in the target product is performed. In this determination process, when it is determined that there is an abnormality in the target product (when the determination result is YES), the process proceeds to S104, and when it is determined that the target product is normal (when the determination result is NO), the process proceeds to S109.

[0068] In S104, a process of obtaining the abnormality classification (type of abnormality) of the target product indicated as having an abnormality from the output of the determination model is performed.

[0069] In S105, with reference to the notification determination master DB14a in the storage unit 14, a process is performed to determine whether the conditions for issuing a notification of the occurrence of an abnormality, as indicated by the registration data of the record including the abnormality classification obtained by the process of S104, are satisfied. In this determination process, when it is determined that the conditions for issuing a notification of the occurrence of an abnormality are satisfied (when the determination result is YES), the process proceeds to S106, and when it is determined that the conditions for issuing a notification of the occurrence of an abnormality are not satisfied (when the determination result is NO), the process proceeds to S109.

[0070] In S106, with reference to the countermeasure master DB14b in the storage unit 14, a process is performed to obtain the registration data of each field of "notification destination", "cause", and "countermeasure" in the record including the abnormality classification obtained by the process of S104.

[0071] In S107, a process is performed to create an abnormality occurrence notification screen as illustrated in FIG. 6, including the abnormality classification information obtained by the process of S104 and the information of each of the notification destination, cause, and countermeasure obtained by the process of S106.

[0072] In S108, a process is performed to transmit the abnormality occurrence notification screen created by the process of S107 to the notification destination obtained by the process of S106 via the communication network 20.

[0073] In S109, the product to be inspected, that is, the product for which the sound collection unit 11 collects the operating sound, is sequentially changed to the next product to be inspected in the sequential inspection, and then the process returns to S101. In the process of S101 performed after this process, a process is performed to obtain the signal of the operating sound of the next product to be inspected.

[0074] The above processes are the product inspection support processes. In the product inspection support system 10 of FIG. 1, the processes of S102 and S103 in FIG. 8 are the processes performed by the determination unit 12, and the processes from S104 to S108 in FIG. 8 are the processes performed by the notification unit 13.

[0075] Next, the control process by the control unit 16 will be described. The flowchart shown in FIG. 9 shows the processing content of an example of this control process.

[0076] Note that the processes from S111 to S113 shown in the flowchart of FIG. 9 are processes that are executed instead of the process of S101 as part of the product inspection support process shown in FIG. 8.

[0077] When the product inspection support process is started, first, in S111 of FIG. 9, the process of acquiring the signal of the operating sound sent from the sound collection unit 11 that has collected the operating sound of the product to be inspected via the communication network 20 is performed in the same manner as the process of S101 in FIG. 8.

[0078] In S112, a process is performed to cause the detection unit 15 to detect abnormal sounds. The detection unit 15 attempts to detect temporary abnormal sounds generated by a sound source other than the target product from the signal of the operating sound acquired by the process of S111, for example, by the method described above.

[0079] In S113, a process is performed to determine whether an abnormal sound has been detected by the detection unit 15 as a result of the process of S112. In this determination process, when it is determined that an abnormal sound has been detected (when the determination result is YES), the process returns to S111, and a process is performed to acquire the operating sound newly collected by the sound collection unit 11 from the target product whose operating sound with the detected abnormal sound has been collected. On the other hand, in this determination process, when it is determined that no abnormal sound has been detected (when the determination result is NO), the process proceeds to S102 in FIG. 8 and the processes after S102 are performed. After that, after the process of S109 is completed, the process of S111 in FIG. 9 is performed.

[0080] By performing the above control process by the control unit 16, the determination unit 12 determines the presence or absence of an abnormality from the operating sound newly collected by the sound collection unit 11 from the target product from which the operating sound with the detected abnormal sound has been collected.

[0081] Although the above-described embodiments and their advantages have been described in detail, those skilled in the art will be able to make various changes, additions, and omissions without departing from the scope of the invention clearly described in the claims.

[0082] For example, the records in the notification determination master DB 14a may be added at any time in response to the addition of the type of abnormality, and the registered data of each record in the notification determination master DB 14a may be changed at any time.

[0083] Also, in the countermeasure master DB 14b illustrated in FIG. 5, although one notification destination is registered for each type of abnormality, there may be a case where there are a plurality of manufacturing processes in which the abnormality of the type indicated by "abnormality classification" occurs in the target product. Therefore, in order to cope with such a case, a plurality of notification destinations may be registered in the countermeasure master DB 14b for one type of abnormality.

Explanation of Reference Numerals

[0084] 10 Product inspection support system 11 Sound collection unit 12 Determination unit 13 Notification unit 14 Storage unit 14a Notification determination master DB 14b Countermeasure master DB 15 Detection unit 16 Control unit 20 Communication network 30-1, 30-2, 30-N Output devices 40 Information processing device 41 CPU 42 Memory 43 Input device 44 Display device 45 Auxiliary storage device 46 Communication I / F 47 Internal bus

Claims

1. A product inspection support system that supports a task of sequentially inspecting a plurality of target products, A sound collection unit that collects operation sounds from the target product; a determination unit that performs a determination of whether the target product has an abnormality, including a determination of a type of abnormality of the target product that has been determined to have an abnormality, based on the operation sound, the determination unit performing the determination using a determination model generated by machine learning that uses the operation sound of the target product that is normal and the operation sound of each type of the target product that has an abnormality; A product inspection support system comprising:

2. 2. The product inspection support system according to claim 1, further comprising a notification unit that notifies the occurrence of an abnormality in response to the appearance of the target product determined to have an abnormality.

3. The product inspection support system according to claim 2, characterized in that the notification unit issues the notification when the manner in which the target product determined to have the same type of abnormality appears in the sequential inspection satisfies a predetermined condition.

4. The product inspection support system according to claim 3, characterized in that the specified condition includes any one of a first condition that the target products determined to have the same type of abnormality appear consecutively a specified number of times, a second condition that the target products determined to have the same type of abnormality appear at an appearance rate equal to or greater than a specified threshold, and a third condition that both the first condition and the second condition are satisfied.

5. 4. The product inspection support system according to claim 3, wherein the notification includes information indicating the type of the target product determined to have the same type of abnormality.

6. A storage unit that stores information for each type that identifies a notification destination of the notification, The notification unit issues the notification to the notification destination identified from the information in the storage unit according to the type of the target product determined to have the same type of abnormality.

4. The product inspection support system according to claim 3.

7. The storage unit further stores information on the cause of the abnormality for each type, The notification includes information about the cause of the type of the target product determined to have the same type of abnormality.

7. The product inspection support system according to claim 6.

8. The storage unit further stores information on measures to be taken against the abnormality for each type, The notification includes information about the countermeasure for the target product of the type determined to have the same type of abnormality.

7. The product inspection support system according to claim 6.

9. A detection unit that detects a temporary abnormal sound generated by a source other than the target product from the operation sounds collected by the sound collection unit; a control unit that controls the determination unit in response to the detection of the abnormal sound, and causes the determination unit to perform the determination from the operation sound that is newly picked up by the sound pickup unit from the target product in which the operation sound from which the abnormal sound was detected is picked up; 9. The product inspection support system according to claim 1, further comprising:

10. A product inspection support method for supporting a task of sequentially inspecting a plurality of target products, comprising: Collect operating sounds from the target product, A determination is made on whether or not the target product has an abnormality, including a determination of a type of abnormality in the target product that has been determined to have an abnormality, based on the operating sound; The determination is performed using a determination model generated by machine learning using the operation sound of the target product that is normal and the operation sound of the target product for each type that is abnormal. A product inspection support method comprising:

Images