JP2022015794A - Work evaluation system and method - Google Patents

Abstract

To provide a technique which can define with respect to evaluation and determination of work operation in a work process of a production flow, an appropriate criterion about the work operation .SOLUTION: A work evaluation system 1 comprises: a detection device 2 which detects work operation of a worker W on an object OB in a production process P of a production flow 200, as sensor data D2; and a processing apparatus 10. The processing apparatus 10 acquires the sensor data D2 and a quality inspection result of the object OB in an inspection process after the production process P, determines whether the production process P is good or not on the basis of determination of standard operation with respect to the work operation in the production process P in accordance with the sensor data D2 and a criterion 7 for determination about the standard operation in the production process P, determines whether or not to update the criterion 7 in the production process P on the basis of a determination result of the production process P and the quality inspection result, and updates the criterion 7 in accordance with a determination result.SELECTED DRAWING: Figure 1

Description

The present invention relates to a technique such as an information processing system, and more particularly to a technique for evaluating a work in a manufacturing process of a manufacturing flow.

At the site such as a production line of a factory, the manufacturing flow is often managed by using a system such as a Manufacturing Execution System (MES). The MES manages the manufacturing flow and gives instructions to workers for each manufacturing process. In addition, MES inspects the quality of products manufactured in the manufacturing flow and grasps the inspection results.

In each work process corresponding to each manufacturing process in the manufacturing flow, a predetermined work is performed by a worker. For example, in a certain work process, a worker performs manual work while looking at and touching an object. As an example of the work operation, it is possible to insert a connector into an object and pull the inserted connector to confirm that it cannot be pulled out.

While the working population is declining, manufacturing quality requirements are increasing, and quality assurance that does not depend on the skills of skilled workers is required. Therefore, a mechanism for evaluating the work process by grasping the work operation by the worker in the work process as digital data (in other words, sensor data or work data) using sensors and information processing is being studied.

In the system of the prior art example, the standard operation is specified as a reference for the work operation by the operator in the manufacturing process. Then, the system detects the work operation as sensor data, and evaluates / determines whether the standard operation is performed or is close to the standard operation.

As an example of the prior art related to the above technique, Japanese Patent Application Laid-Open No. 2012-198644 (Patent Document 1) can be mentioned. Patent Document 1 describes that, as a work support device or the like, it is possible to continuously change the moving image information as a model according to the proficiency of the worker, and the following. The work support device acquires work image information from the camera and compares the acquired work image information with the model work image information to determine whether the work is improved from the model work.

Japanese Unexamined Patent Publication No. 2012-198644

Conventionally, there are various types and contents of work in a work process, and standardization of standard operation standards for work operations in a work process, in other words, standardization, is not always performed. If the standard operation standard of the work process is not specified, the work process cannot be evaluated. When formulating new standards, it was sometimes difficult to establish appropriate standards.

In addition, when the standard of standard operation of the work process is specified, the standard is not always appropriate. If the criteria are inappropriate, the evaluation of the work process cannot be done properly. Even when reviewing the standards, it was sometimes difficult to establish appropriate standards.

In the prior art example, only the evaluation / judgment of the work operation in each work process is limited. In the example of Patent Document 1, the quality of the motion is determined by comparing the motion of a person in the work process with the motion of a model image based on the image captured and detected by the camera. The evaluation / judgment of the conventional work process does not consider the quality of the product obtained as a result of the work process, the quality of the product obtained as the final inspection result, and the like. In the past, even if each work process was standardized as on-site know-how, it was not associated with the quality of the final product. Standards in a work process may not be appropriate, such as being too strict or too easy for the operator. Therefore, the result of evaluating the work operation in a certain work process in comparison with the standard may not be appropriate.

Further, in the prior art example, for example, a threshold value of the amount of movement in the model is set as a reference in the work process. Such a standard is basically a constant value, but it is unclear whether it is suitable. It is also unclear if the criteria are suitable for all workers.

As described above, conventionally, there is inefficiency in the manufacturing flow as a whole, and there is insufficient study on the provision of appropriate standards for the work operation of the work process, and there is room for improvement.

It is an object of the present invention that an appropriate standard regarding a work operation can be defined with respect to a technique such as an information processing system for evaluating / determining a work operation of a work process of a manufacturing flow, and as a result, improvement of the accuracy of the work process and manufacturing can be defined. It is to provide technology that can improve the quality of goods and the productivity of production lines.

The work evaluation system according to a typical embodiment of the present invention includes a detection device that detects a work operation on an object by a worker in a manufacturing process of a manufacturing flow as sensor data, and a processing device connected to the detection device. The processing apparatus acquires the sensor data and the inspection result of the quality of the object in the inspection step after the manufacturing step of the manufacturing flow, and obtains the sensor data and the said. Based on the criteria for determining the standard operation of the manufacturing process, the propriety of the manufacturing process is determined based on the determination of the standard operation for the working operation in the manufacturing process, and the determination result of the manufacturing process and the above Based on the inspection result, it is determined whether to update the standard in the manufacturing process, and the standard is updated according to the determination result.

According to a typical embodiment of the present invention, an appropriate standard for work operation can be defined with respect to a technique such as an information processing system for evaluating / determining work operation in a work process of a manufacturing flow, and as a result, as a result. It is possible to improve the accuracy of the work process, the quality of the product, the productivity of the production line, and the like.

The configuration of the work evaluation system of Embodiment 1 of this invention is shown. In the first embodiment, a configuration example of a processing device that is a server is shown. In the first embodiment, a configuration example of the manufacturing flow is shown. In the first embodiment, a detailed configuration example of the manufacturing flow is shown. In the first embodiment, examples of work, detection, and inspection in the process are shown. In the first embodiment, the main processing flow of the processing apparatus is shown. In the first embodiment, a configuration example of a production control table which is management information is shown. In the first embodiment, the histogram in the reference update processing example is shown. The first example of the GUI screen in the modification 1 of Embodiment 1 is shown. A second example of the GUI screen in the second modification of the first embodiment is shown. An example of using the camera in the third modification of the first embodiment is shown. A configuration example of the production control table in the modified example 4 of the first embodiment is shown. A configuration example of the production control table in the modified example 5 of the first embodiment is shown. The configuration of the work evaluation system of the second embodiment is shown. A control example in the work evaluation system of the third embodiment is shown. A configuration example of the standard for each worker in the modified example of the third embodiment is shown. The processing flow and the like in the work evaluation system of Embodiment 4 are shown. A configuration example of a manufacturing flow and a production control table in the work evaluation system of the fifth embodiment is shown. The histogram of the reference update processing example in Embodiment 5 is shown.

<Embodiment 1>
The work evaluation system and method according to the first embodiment of the present invention will be described with reference to FIGS. 1 to 8 and the like. The work evaluation method of the first embodiment is a method having steps executed by the work evaluation system of the first embodiment. The work evaluation system and method of the first embodiment detect as sensor data using a sensor with respect to the work operation by the worker in each manufacturing process (corresponding work process) of the manufacturing flow. This system determines the propriety of each work process based on the sensor data and the standard operation standard for the work operation of each work process. This determination is based on an evaluation such as the presence or absence of standard operation or whether it is close to standard operation. Then, this system updates the standard of the work process based on the comparison between the judgment result of the work process and the inspection result regarding the quality of the product in the final inspection process (or the inspection result for each manufacturing process). Judge the necessity and update the standard of the work process so that it becomes a more appropriate value.

The work evaluation system and method of the first embodiment updates the work process standard according to the combination of the judgment result of the work process and the inspection result after that. One of the features of this system is that it updates the standard of operation of the previous work process based on the quality of the inspection result. In particular, this system updates the criteria to match the values of the final inspection result, for example, when the values of the determination result of the work process do not match.

According to this system, it is possible to specify suitable standards for the work of the manufacturing process according to the quality of the product in the inspection result. According to this system, while ensuring the quality of the final product, the standards of the work process can be made suitable for the operator so as not to be too strict or too easy, and the manufacturing flow. Overall efficiency can be achieved. According to this system, suitable standards can be specified in either case of reviewing existing standards in the manufacturing process or setting new standards in the manufacturing process. According to this system, work data can be collected and accumulated for a period or a plurality of objects, and can be repeatedly updated to make a more suitable standard.

[Work evaluation system]
FIG. 1 shows the configuration of the work evaluation system 1 of the first embodiment. The work evaluation system 1 includes a detection device 2 provided for each manufacturing process P of the manufacturing process flow 201 of the manufacturing flow 200, a processing device 10 which is a server, and a manufacturing execution system (MES) 20. The MES 20 includes a process inspection device 3 provided for each manufacturing process P, a final inspection device 4 provided in the final inspection process 202, a management device 5, and a database (DB) 6. The detection device 2 is connected to the processing device 10 via the communication network 9. The communication network 9 is, for example, a LAN. The processing device 10 is implemented, for example, in a server device, but is not limited thereto. The processing device 10 may be located remotely from the manufacturing site.

The detection device 2 includes a sensor 8. The detection device 2 detects the work operation by the worker W who is the target person in the manufacturing process P (corresponding work process) of the manufacturing flow 200 as the sensor data D2 by using the sensor 8. The detection device 2 outputs the sensor data D2 from the sensor 8. The server 10 receives and acquires the sensor data D2 from the detection device 2. In addition to the sensor 8, the detection device 2 may include, for example, an analog-to-digital conversion circuit, a communication interface circuit, or the like.

MES5 is a system that manages and executes manufacturing in the manufacturing flow 200. The MES 5 manages and grasps the state of each manufacturing process P in the manufacturing flow 200, and gives an instruction or the like to the worker W of each manufacturing process P. The management device 5 is a computer that controls manufacturing execution of the MES 20, for example, a MES server. Various data and information related to manufacturing execution control are stored in the DB 6.

The process inspection device 3 inspects (may be described as process inspection) the object OB (for example, an intermediate product) obtained as a result of the associated manufacturing process P. The process inspection device 3 outputs inspection data D3 including the inspection result of the manufacturing process P. Depending on the manufacturing process P, there may or may not be a process inspection.

The final inspection device 4 inspects (may be described as final inspection) the final product OBL (for example, a finished product) obtained as a result of the associated final inspection step 202. The final inspection device 4 outputs inspection data D4 including the inspection result of the final inspection step 202. The inspection data D3 and the inspection data D4 may be temporarily stored in the DB 6.

The work evaluation system 1 of the first embodiment may be linked with the existing MES5 provided in a factory or the like. Note that MES20 is not essential. The server 10 may acquire the inspection data D3 and D4 by another method. Further, in the first embodiment, the inspection result of the final inspection step 202 is used as indispensable. The process inspection, the process inspection device 3, and the inspection data D3 for each manufacturing process P are not indispensable.

Specifically, the DB 6 is, for example, a sensor data DB that stores the sensor data D2 for each manufacturing process P, a process inspection DB that stores the inspection data D3 for each manufacturing process P, and a final that stores the inspection data D4 for the final inspection process 202. Includes inspection DB, etc. The sensor data DB and the process inspection DB may be integrated into one DB (“manufacturing DB”), or the process inspection DB and the final inspection DB may be integrated into one DB (“inspection DB”). ..

The processing device 10 which is a server is a computer constituting the main controller of the work evaluation system 1 of the first embodiment. The server 10 realizes the function related to the update of the reference 7. The server 10 has a function of managing and updating the standard 7 for the standard operation in each manufacturing process P of the target manufacturing flow 200. The server 10 has a standard operation determination unit 11, a process quality assurance unit 12 (in other words, an inspection result acquisition unit), a determination standard update unit 13, and a standard operation determination result output as functional blocks realized by software program processing or the like. It has a unit 14 and a determination reference output unit 15. Some of the functions may be implemented in a dedicated circuit. As a system, a form in which the processing device 10 is merged with the MES 20 is also possible.

The sensor data D2 may be temporarily collected and stored in the DB 6 of the MES 20 from the detection device 2. In this case, the server 10 may refer to and acquire the sensor data D2 from the DB 6 of the MES 20. In this example, the standard operation determination unit 11 in the server 10 processes the sensor data D2 acquired from the detection device 2 (or DB6) (for example, feature quantity extraction) to determine whether or not there is a standard operation in the manufacturing process P. do. Not limited to this, the processing unit in the detection device 2 may process the sensor data D2 and then transmit the processed data to the server 10. Further, the processing unit (processing unit corresponding to the standard operation determination unit 11) in the detection device 2 may determine whether or not there is a standard operation based on the sensor data D2, and transmit the determination result data to the server 10. Alternatively, a processing unit in the MES 20, for example, a process inspection device 3, may determine whether or not there is a standard operation in the manufacturing process P based on the processing of the sensor data D2.

The server 10 stores the result of determining the presence or absence of the standard operation based on the reference 7 for the work operation of each manufacturing process P represented by the sensor data D2 as the standard operation determination result data D11. Further, the server 10 acquires each inspection result (inspection data D13, D14) of the manufacturing flow 200. Then, the server 10 makes a determination regarding the update of the reference 7 based on the comparison between the determination result of the manufacturing process P and the inspection result.

The standard operation determination unit 11 determines whether or not the work operation of each manufacturing process P is possible based on the sensor data D2 from the detection device 2 and the information of the reference 7 for each manufacturing process P set in the reference data D7. do. This determination is to compare the work operation represented by the sensor data D2 with the reference 7 and determine whether or not there is a standard operation. The work operation represented by the sensor data D2 is represented by a value such as a force or a movement amount as a feature amount (in other words, a variable) based on the sensor value. The standard operation reference 7 is expressed by a condition including, for example, a threshold value related to a value such as a force or a movement amount. The determination of the presence or absence of this standard operation may be determined as to whether or not the work operation represented by the sensor data D2 is sufficiently close to or similar to the standard operation. In the above determination, for example, when the standard operation determination unit 11 determines that there is a standard operation that satisfies the reference 7, the determination result of the manufacturing process P is OK (possible), and it is determined that there is no standard operation. , NG (No). The determination results of the manufacturing process P and the standard operation are used for determining the update of the reference 7. The standard operation determination unit 11 stores and outputs the standard operation determination result data D11.

The standard operation determination unit 11 grasps a predetermined feature amount representing a work operation for each manufacturing process P based on the sensor value of the sensor data D2. At this time, the standard operation determination unit 11 may calculate the feature amount from the sensor value. If the sensor value itself represents the feature amount, the calculation is unnecessary. An example of the feature amount is the fingertip pressure, and the sensor value itself of the pressure sensor of the glove represents the force. Another example of the feature amount is the amount of movement and the position of the hand extracted by calculation from the image which is the sensor data D2.

In other words, the process quality assurance unit 12 is an inspection result acquisition unit, and collects and acquires the inspection data D3 of each manufacturing process P of the manufacturing process flow 201 and the inspection data D4 of the final inspection process 202. The process quality assurance unit 12 confirms whether or not the quality of the final product OB corresponding to the object OB is guaranteed (in other words, whether or not the quality is possible) as a result of the final inspection process 202, at least based on the inspection data D4. Make a judgment. Further, if the process quality assurance unit 12 has the inspection data D3, the process quality assurance unit 12 may confirm or determine whether or not the quality of the intermediate product, which is the result of each manufacturing process P, is guaranteed based on the inspection data D3. good. In this example, each inspection result (D3, D4) is guaranteed by MES20, and it is confirmed that there is no problem in the inspection work of each inspection. Therefore, the process quality assurance unit 12 can omit the determination for quality assurance of each inspection result.

The process quality assurance unit 12 comprehensively determines each inspection result (D3, D4) of the manufacturing flow 200, confirms or determines the quality, and stores and outputs the result as the process quality assurance data D12. As a comprehensive judgment, the process quality assurance unit 12 detects the case where the inspection result of the manufacturing process P is NG, the case where the inspection result of the final inspection process 202 is NG, and the like.

The reference update unit 13 uses the reference data D7, the standard operation determination result data D11, and the process quality assurance data D12 to perform a process of updating the standard operation reference 7 in the manufacturing process P as necessary. The reference updating unit 13 confirms the state of the combination of the inspection result of the final inspection process 202 and the determination result of each manufacturing process P of the manufacturing process flow 201 based on the process quality assurance data D12. The reference updating unit 13 distinguishes and extracts cases where the determination result of the manufacturing process P and the inspection result of the final inspection process 202 match or do not match in combination. For example, in the case where the final inspection result is NG and the determination result of the manufacturing process P is OK, the standard updating unit 13 does not match as a whole, so the standard 7 of the manufacturing process P is not appropriate. It can be estimated that it was.

Therefore, the standard updating unit 13 targets the standard 7 of the manufacturing process P as an update target, determines a more appropriate value for the standard 7 of the manufacturing process P based on the data related to the combination, and determines the standard. Update 7. The server 10 holds in the memory information about the reference 7 of each manufacturing process P necessary for processing as the reference data D7. When updating the reference 7, the server 10 updates the information of the reference 7 in the reference data D7. The reference update unit 13 updates the reference 7 by reflecting the reference update information D13 in the reference data D7. The reference update unit 13 may transmit the updated reference data D7 or reference update information D13 to the MES 20. The MES 20 may hold reference data regarding the reference 7 of each manufacturing process P. In that case, the server 10 may refer to or update the reference data of the MES 20.

The standard operation determination result output unit 14 outputs the standard operation determination result information based on the standard operation determination result data D11 by the standard operation determination unit 11. The standard operation determination result output unit 14 displays, for example, the standard operation determination result information on the display screen together with the GUI (FIG. 9 described later). For example, a user U1 (FIG. 2) such as a worker W or a manager of a manufacturing flow 200 can see and confirm the standard operation determination result information on the screen. Further, this output may be an output instructing the worker W of the manufacturing process P to retry the work when the determination result of the manufacturing process P is NG (FIG. 10 described later). The output is not limited to the display, and audio output, lamp emission, or the like may be used. If the work retry is allowed according to the provisions of the manufacturing flow 200, it is also effective to output the work retry instruction.

The reference output unit 15 outputs the information of the reference 7 before the update and the reference 7 after the update based on the reference data D7 and the reference update information D13. This output is, for example, a display on a display screen. For example, a user U1 such as a worker W or a manager of a manufacturing flow 200 can see and confirm the information of the reference 7 on the screen (FIG. 9 described later).

[Processing device]
FIG. 2 shows a configuration example of the hardware and software of the processing device 10. The processing device 10 is implemented as a computer system including, for example, a server computer and an input device 105, a display device 106, and the like connected to the server computer. The processing device 10 includes a processor 101, a memory 102, a communication interface device 103, an input / output interface device 104, a bus connecting them to each other, and the like. An input device 105 such as a keyboard or a mouse and a display device 106 such as a liquid crystal display are connected to the input / output interface device 104, for example. The communication interface device 103 is also connected to the detection device 2 and the MES 20 of FIG. 1 by communication, and communicates with each other by a predetermined communication interface. The communication interface device 103 may communicate with another external device.

The processor 101 is composed of, for example, a CPU, a ROM, a RAM, or the like, and constitutes a controller. The processor 101 realizes a predetermined function based on software program processing based on the OS, the control program 102A, and the like. This function includes a function of managing and updating the standard 7 of the manufacturing process P.

The memory 102 is composed of a non-volatile storage device or the like, and stores various data and information used by the processor 101 and the like. The memory 102 stores the control program 102A, the setting information 102B, the reference data D7, the standard operation determination result data D11, the process quality assurance data D12, the production control table data D20, and the like. The control program 102A is a program for realizing the function. The setting information 102B is the setting information of the control program 102A and the setting information by the user U1. The reference data D7 includes setting information of the reference 7 of each manufacturing process P of the manufacturing flow 200. The production control table data D20 is data in which various data and information that can be acquired at the time of manufacturing execution including the work according to the manufacturing flow 200 are collectively recorded. The server 10 may acquire the production control table data D20 from the MES 20.

[Manufacturing flow (1)]
In the first embodiment, the pre-standardized manufacturing flow 200 can be applied as a target. The standardized manufacturing flow 200 here means that the standard 7 of the standard operation is set for the work operation of the manufacturing process P. That is, in the first embodiment, it is premised that the standard 7 of the manufacturing process P has been established once. Not limited to this, the first embodiment can be similarly applied to the case where the manufacturing flow 200 which has not been standardized yet, that is, the standard 7 of the manufacturing process P has not been formulated yet. The system of the first embodiment can update the set standard 7 to a more appropriate standard 7, or create and set a new suitable standard 7 for a manufacturing process in which the standard 7 is not set.

FIG. 3 shows a configuration example of the manufacturing flow 200 of FIG. The factory, which is the site, has a manufacturing flow 200 corresponding to the production line of the product. The manufacturing flow 200 includes a manufacturing process flow 201 and a subsequent final inspection step 202. The manufacturing process flow 201 is composed of a plurality of manufacturing processes P having a relationship such as an order. For example, as a plurality of (M pieces) manufacturing process P, there are manufacturing processes P1 (A), P2 (B), ..., PM (C). In this example, the first first manufacturing process P1 is simply the manufacturing process A, the second second manufacturing process P2 is the manufacturing process B, and the middle is omitted, and the final M manufacturing process PM is It is assumed that the manufacturing process is C. The manufacturing process flow 201 may have a plurality of manufacturing processes P branched in parallel at the same time.

In each manufacturing process P, a worker W who is a target person is assigned as a person in charge. For example, a worker WA is assigned to the manufacturing process A, a worker WB is assigned to the manufacturing process B, and a worker WC is assigned to the manufacturing process C. Not limited to this, the same worker W may be assigned to each manufacturing process P.

The object OB (referred to as OB1 in the initial state) is introduced into the manufacturing process flow 201. For example, the object OB1 such as a member is put into the manufacturing process P1. As a result of the work in the manufacturing process A, the object OB1 becomes an object OB2 such as an intermediate, and is input to the next manufacturing process P2. An object OBM is put into the final manufacturing process PM, and as a result of the work, a finished product OBL is generated. The final product OBL is finally inspected in the final inspection step 202.

The final inspection step 202 is composed of one or more inspection steps 205 and is a step of inspecting the quality of the final product OBL. In this example, the final inspection step 202 has an inspection step X, which is a first inspection step, and an inspection step Y, which is a second inspection step thereafter.

More specifically, each manufacturing step P has a working step 203 and a subsequent inspection step 204. The inspection step 203 may not be performed as described above. In the work process 203, the worker W performs a predetermined predetermined work on the object OB. A detection device 3 is associated with each work step 203. In the work process 203, the work operation is detected as the sensor data D2 by the sensor 8 of the detection device 2.

The inspection step 204 is a step in which the inspector or the process inspection device 3 inspects the quality of the object OB which is the result of the work step 204. The process inspection device 3 is associated with the inspection process 204. The inspector or the process inspection device 3 inspects the quality of the object OB obtained as a result of the work process 203, and outputs the inspection result as the inspection data D3. The inspection result includes, for example, a quality value. The inspection data D3 is, in other words, process inspection result information. The inspection may be performed automatically by the process inspection device 3 unattended, or may be performed by the inspector using the process inspection device 3.

For example, the manufacturing process A has a working process A and an inspection process A. Sensor data A is output from the work process A. The inspection data A is output from the inspection step A. Similarly, in the manufacturing process B, the sensor data B and the inspection data B are output. In the manufacturing process C, the sensor data C and the inspection data C are output.

In the final inspection step 202, the final inspection device 4 is associated with each inspection step 205. The final inspection device 4 performs the final inspection and outputs the inspection data D4 as the inspection result. The sensor data D2, the inspection data D3, and the inspection data D4 are collected and stored in, for example, the DB 6 of the MES 20. For example, the inspection device X in the inspection step X outputs the inspection data X. The inspection device Y in the inspection step Y outputs the inspection data Y. The final inspection may be an inspection work by a predetermined inspector, or may be an automatic inspection using the final inspection device 4. Examples of final inspections include appearance evaluation by an inspector and measurement of characteristic variation in product characteristics.

In the example of FIG. 3, each device is provided in a one-to-one relationship in each process, but the present invention is not limited to this. For example, one detection device 2 or one process inspection device 3 may be associated with a plurality of manufacturing processes P. A plurality of detection devices 2 and a plurality of process inspection devices 3 may be associated with one manufacturing process P. The detection device 2 and the process inspection device 3 may be integrated. These correspondences are set in advance in the MES 20 or the server 10 as the configuration information of the manufacturing flow 200.

[Manufacturing flow (2)]
FIG. 4 shows a more detailed configuration example of the manufacturing flow 200 of FIG. 3, and also shows a plurality of manufacturing process Ps and operation specific examples of the manufacturing process flow 201. In this example, manufacturing processes A, B, and C are included as the manufacturing process P of the manufacturing process flow 201. An example of the manufacturing flow 200 is an assembly line of a motor product as an object OB, and an example of the manufacturing process P is an assembly process. Each manufacturing process P is connected through a transport mechanism 400. A plurality of object OBs (for example, OB1, OB2, ..., OBm) are sequentially conveyed on the transfer mechanism 400. Each object OB is managed by adding identification information (ID). FIG. 4 focuses on one object OB having a certain object ID, and schematically illustrates a case where the object OB (each state is OBA, OBB, OBC, OBL) goes through each process. ing. Further, in this example, the worker W is different for each manufacturing process P, and the inspector is different for each inspection process.

In this example, the work step A is a step of incorporating the component A into a predetermined portion of the object OB (OBA) as a work by the worker WA. In this work, for example, the tip of the cable is inserted into the object OBA as a part A held by the worker WA, and after the insertion, the cable is pulled to confirm that the cable cannot be pulled out. For this work operation, the fingertip pressure at the time of inserting the component A is measured by the sensor 8 mounted on the hand of the worker WA, and the sensor data D2 is obtained. The work process B is a step of incorporating the component B into a predetermined portion of the object OB (OBB) as a work by the worker WB. The work of the work process B is the same as the work of the work process A, for example, and the tip of the cable as the part B held by the worker WB is inserted into the object OBB, and after the insertion, the cable is not pulled out. It is a work to confirm. For this work operation, the fingertip pressure at the time of inserting the component B is measured by the sensor 8 mounted on the hand of the worker WB, and the sensor data D2 is obtained. The work process C is a process of combining the parts A and the parts B in the object OB (OBC) to complete the product as the final product OBL.

The inspection step A is a step of inspecting the quality of the object OBA as a result of the work step A by the inspector KA or the process inspection device A. For example, the leakage current is measured using an ammeter and the leakage current is equal to or less than the threshold value. This is the process of confirming that. The inspection step B is a step of inspecting the quality of the object OBB as a result of the work step B by the inspector KB or the process inspection device B, and is a step of confirming the current as in the inspection step A, for example. Inspection steps A and B may not be available. The inspection steps A and B may be an inspection by a subjective evaluation by an inspector. In this inspection, for example, the appearance of the object OB may be observed to confirm that there are no scratches or unevenness.

The final inspection step 202 is a step of inspecting the quality of the final product OBL by a predetermined inspector K or the final inspection device 4, for example, the inspector K measures the leakage current of the final product OBL using an ammeter. This is a process for confirming whether the leakage current is below the threshold value. If the leakage current is equal to or less than the threshold value, the inspection result is acceptable (OK), and if the leakage current exceeds the threshold value, the inspection result is negative (NG). The conditions in the final inspection step 202 are preset as information such as "the measured current value is equal to or less than the threshold value".

The server 10 determines whether or not there is a standard operation for the work process A by using the sensor data A and the reference 7 (for example, the reference A) of the manufacturing process A. The sensor value of the sensor data A is the fingertip pressure (referred to as FA) when the component A is inserted. Criteria A include a threshold A for its fingertip pressure. For example, the reference A is a condition that "the fingertip pressure FA is equal to or higher than the threshold value A". The server 10 compares the fingertip pressure FA with the threshold value A, and if FA ≧ A, the determination result is OK (possible), and if FA <A, the determination result is NG (no). do.

[Work / Detection]
FIG. 5 shows specific examples of work, detection, and inspection in the manufacturing process P of the manufacturing flow 200 of FIG. (A) shows an example of the work by the worker W and the detection using the sensor 8 at the work place of the work step 203 of a certain manufacturing process P. As an example of the sensor 8 constituting the detection device 2, the glove 8A and the camera 8B are included. The worker W1 wears a glove 8A on his hand and a camera 8B on his head (for example, a hat or a helmet). The glove 8A has a built-in pressure sensor, for example, at the fingertip. The glove 8A may be provided with an acceleration sensor, a gyro sensor, a position sensor (for example, a GPS receiver, a beacon, etc.) and the like. An example of the standard 7 regarding the work of (A) is that "the fingertip pressure when inserting a part into an object is a predetermined force or more" and the like.

The worker W may wear at least one of the glove 8A or the camera 8B, and may wear both. The camera 8B is a viewpoint camera that obtains an image by photographing the direction in which the head or eyes of the worker W1 are facing. This image is an image obtained by roughly taking an image of an object OB, a member held in a hand, or the like from the viewpoint of the worker W. The camera 8B may be a smart glass, a head-mounted display, or the like. Not limited to this example, various sensor devices may be attached to the worker W1. At the work place, there is an object OB on the transport mechanism 400, the workbench 500, and the like. The worker W performs a predetermined work on the object OB. The glove 8A and the camera 8B in the detection device 2 detect the work operation at this time as a target. The pressure sensor of the glove 8A measures the fingertip pressure. The camera 8B captures an image including the movement of the hand.

In (B), as an example of another sensor 8, a fixed camera 8C is installed in a work place. As described above, the sensor 8 of the detection device 2 may be fixedly installed in the vicinity of the worker W1 or the like. The fixed camera 8C obtains an image taken in a predetermined direction, for example, in a certain direction of the object OB on the workbench 500. In this example, if there is a movement of the object OB, the hand of the worker W, or the like in the fixed image frame, it can be detected.

(C) shows an example of inspection work by an inspector. In one inspection step, one inspector connects an ammeter 504 to the object OB, measures the leakage current, and confirms the amount of leakage current.

[Processing flow]
FIG. 6 shows the main processing flow by the server 10 in the first embodiment. This flow has steps S1 to S6. In step S1, the server 10 sets or inputs information including the worker ID of the worker W in charge of each manufacturing process P as the information of the manufacturing flow 200 to be processed and the target object OB. As a result, the object OB and the worker W of each manufacturing process P are clarified. If this information has already been set as manufacturing flow information in MES 20, the server 10 may refer to the information from MES 20 and use it. User U1 such as an administrator may input and set the information to the server 10. When the sensor data D2 and the inspection data D3 include information such as an object ID and a worker ID, the server 10 may use such information.

In addition, when the same standard 7 is uniformly applied to all the workers W without distinguishing the workers W in the manufacturing process P, such information such as the worker ID is not indispensable. Further, as the information to be used, not only the worker ID but also the attributes of the worker W, such as age and skill level, may be used.

In step S2, the server 10 detects and acquires the work operation of the manufacturing process P as sensor data D2 by using the detection device 2. The server 10 collects and acquires the sensor data D2 once stored in the DB 6 of the MES 20 from the detection device 2, for example. The server 10 calculates a defined feature amount based on the sensor value of the sensor data D2.

In step S3, the standard operation determination unit 11 of the server 10 determines whether or not there is a standard operation for the work operation of the manufacturing process P represented by the feature amount, using the reference 7 set for each manufacturing process P. .. For example, when the feature amount is equal to or more than the threshold value of the reference 7, the server 10 determines that there is a standard operation, and when the feature amount is less than the threshold value of the reference 7, it determines that there is no standard operation. When the server 10 has the standard operation, the determination result of the manufacturing process P is acceptable (OK), and when there is no standard operation, the determination result of the manufacturing process P is rejected (NG). The server 10 stores the determination result in the memory as the standard operation determination result data D11. Further, the standard operation determination result output unit 14 may output (for example, screen display) the standard operation determination result information to the user U1 based on the standard operation determination result data D11.

In step S4, the process quality assurance unit 12 of the server 10 acquires inspection data D3 and inspection data D4, at least inspection data D4 related to the final inspection process 202, from MES20, and makes a process quality assurance determination. The process quality assurance determination includes grasping and guaranteeing the quality of the final product OBL as the inspection result of the final inspection process 202. When there is the inspection data D3, the process quality assurance determination is grasping and guaranteeing the inspection result by the combination of the inspection result of each manufacturing process P and the inspection result of the final inspection process 202. The process quality assurance unit 12 stores the process quality assurance data D12, which is the result of the process quality assurance determination, in the memory.

The processes of steps S2 to S4 are similarly repeated for each object OB identified by the object ID and for each worker W identified by the worker ID. Sensor data D2, standard operation determination result data D11, process quality assurance data D12, etc. are obtained for each object OB, and these data are also organized and stored in the production control table data D20.

In step S5, the server 10 confirms whether the processed data number N regarding the data for each object OB has reached the predetermined data number NX, and if it reaches (Y), proceeds to step S6 and reaches. If not, the process returns to step S2 and repeats in the same manner. The predetermined number of data NX is a set value that defines the number of data to be referred to when determining the update of the reference 7, and the timing of updating. This number of data NX is set in advance as a part of the setting information 202B. The number of data NX may be changed by the user U1 on the setting screen.

In step S6, the reference update unit 13 of the server 10 sets the reference 7 of the manufacturing process P of the manufacturing flow 200 based on the reference data (for example, the data group of the production control table for the object OB for the number of data NX). Update. The reference update unit 13 determines, for example, one manufacturing process P to be updated, and determines an update value such as a threshold value constituting the reference 7 of the manufacturing process P. The specific update method will be described later. The reference update unit 13 generates the reference update information D13, updates the information of the reference 7 described in the reference data D7 of the memory, and transmits the reference update information D13 or the updated reference data D7 to the MES20. The MES 20 updates the information of the reference 7 managed by the MES 20 according to the information from the server 10.

[Production control table]
FIG. 7 shows a configuration example of a production control table for the manufacturing flow 200. In this example, in the configuration information of the manufacturing flow 200, a certain manufacturing process P is associated with a certain worker W in charge. In this production control table, data on actual results such as work on a plurality of object OBs in the production flow 200 are collectively recorded. Although not shown in the production control table, information such as information on workers and inspectors in charge of the process, date and time, and the like may be described. In this example, it is assumed that the standard 7 of a certain manufacturing process P is the same standard 7 without distinguishing even if there is a difference in the worker W in charge.

The production control table of FIG. 7 has an object ID, a product name, a manufacturing process, and a final inspection process as items. The object ID item stores the ID of the object OB, for example, a serial number. In the product name item, the product name (for example, the motor M1) of the manufactured product corresponding to the object OB is stored. The manufacturing process item is an item for storing information of the manufacturing process P. The manufacturing process item has an item of each manufacturing process P of the manufacturing process flow 201, but in this example, only one manufacturing process P1 is shown. The manufacturing process item has a sensor value, a standard, and a determination result as lower-level items. The sensor value item stores the sensor value of the sensor data D2 or a feature amount based on the sensor value (for example, fingertip pressure). The reference item stores the information of the reference 7. The determination result item stores the value of the determination result for the standard operation of the manufacturing process P. An example of the sensor value item in the manufacturing process P1 is the fingertip pressure (unit [Pa]) that can be detected by the glove 8A in FIG. An example of a reference item is a lower limit of fingertip pressure as a threshold. In other words, this criterion 7 is a condition such as "50 Pa or more". The determination result item has two values, OK (OK) or NG (No).

The final inspection process item is an item for storing information of the final inspection process 202, and has measured values, conditions, and inspection results as lower-level items. The measured value item stores the measured value (for example, the current value) in the inspection. The condition item stores a condition for determination in the inspection, for example, an upper limit value of the leakage current as a threshold value of the current, in other words, a condition such as "20 mA or less". The inspection result item has two values, OK (OK) or NG (No), as the value of the inspection result.

[Data status]
Based on the data of the production control table obtained according to the manufacturing flow 200, for example, when the inspection result of the final inspection step 202 is negative (NG), the server 10 tells the object OB the process of the manufacturing process flow 201. It can be estimated that there was something wrong with it. The server 10 compares the data values with the combination of the determination result of the manufacturing process P and the inspection result of the inspection process (process inspection or final inspection) after the manufacturing process P, and determines the update of the standard 7. I do. The server 10 pays attention to the case where the result values do not match in the combination. That is, a case where the judgment result of the manufacturing process P is acceptable (OK) but the inspection result is negative (NG), or a case where the judgment result is negative (NG) but the inspection result is acceptable (OK). Is. The server 10 updates the reference 7 of the manufacturing process P so that the result values of the combinations match based on the data group related to the case. The server 10 generates a tentative reference value for updating based on the related data group, for example, based on statistics and the like. When the server 10 applies the tentative reference value to the related data group and the result values in each combination match, the tentative reference value is used as the update reference value.

When considering the combination of the determination result of the standard operation of the manufacturing process P and the inspection result of the final inspection process 202 as the state of the data in the production control table, there are some patterns as follows. In the production control table of FIG. 7, the first row in which the object ID = 1001 and the fourth row in which 1004 is 1004 are examples of the first pattern. The first pattern is a combination in which the determination result of the manufacturing process P is OK and the final inspection result after that is OK. The second line in which the object ID = 1002 is an example of the second pattern. The second pattern is a case where the determination result of the manufacturing process P is NG and the final inspection result after that is NG. The third line in which the object ID = 1003 is an example of the third pattern. The third pattern is a case where the determination result of the manufacturing process P is OK and the final inspection result after that is NG.

In the case of the first pattern and the second pattern, the value of the determination result and the value of the inspection result are in agreement, and they are in agreement as a whole. In the case of the third pattern, the value of the determination result and the value of the inspection result do not match, and do not match as a whole. In the example of the third pattern of FIG. 7, in the manufacturing process P1, the sensor value is 80 Pa and satisfies the reference 7 “50 Pa or more”, so that the determination result is OK. On the other hand, in the final inspection step, the measured value is 50 mA, which does not satisfy the condition "20 mA or less", so the inspection result is NG. In the first embodiment, the server 10 pays particular attention to the case of such a third pattern and determines the update of the reference 7.

When the third pattern occurs, it can be estimated that the server 10 is not suitable for the standard 7 (“50 Pa or more”) for the determination of the standard operation in the manufacturing process P1, and there is room for improvement. In other words, it can be estimated that the criterion 7 was too easy, specifically the threshold was too small. Therefore, the server 10 reviews the standard 7 of the manufacturing process P1 and updates the standard 7 so as to have a stricter standard 7, specifically, a larger threshold value.

For the update, specifically, the server 10 refers to the data group related to the case of this third row, that is, the data of each object in each row related to the same manufacturing process P1 from this production control table. .. In this example, it is assumed that the related data group includes the data of the object ID = 1001 to 1004. The server 10 tries to change the value of the reference 7 of the manufacturing process P1 as a tentative value, so that the value of the determination result and the value of the inspection result match in all the related data groups. To find out. The server 10 uses the value as the update value of the reference 7. In this example, for example, "100 Pa or more" can be found as the update value of the reference 7 in such a state.

More specifically, the update value of the reference 7 can be obtained by the following processing example. In order to match the judgment result and the final inspection result in all the related data groups of the manufacturing process P1, if the update value of the standard 7 is set to a value within the range of "81 Pa or more" to "120 Pa or more". I know it's good. This range is a reference candidate range. For example, if the reference 7 is assumed to be "81 Pa or more", the "130 Pa" in the first line and the "120 Pa" in the fourth line satisfy the tentative values, so that the determination result is OK and the first pattern is obtained. .. Since "40 Pa" on the second line and "80 Pa" on the third line do not satisfy the tentative value, the determination result is NG, which is the second pattern. For example, the server 10 selects "100 Pa or more" as a rough midpoint from this reference candidate range and adopts it as an update value of the reference 7.

When the standard 7 of the manufacturing process P1 is updated from the original "50 Pa or more" to this "100 Pa or more", the sensor value "130 Pa" in the first row is OK, and "40 Pa" in the second row is NG. "80Pa" in the third line is NG, and "120Pa" in the fourth line is OK. That is, in any line, the determination result and the inspection result are in agreement, in other words, the state corresponds to the first pattern or the second pattern.

As a result of updating the reference 7 of the manufacturing process P1, the determination of the standard operation in the manufacturing process P1 becomes more strict. Therefore, in the future, there is a high possibility that a consistent combination will occur in which the determination result of the manufacturing process P is OK and the inspection result of the final inspection step 202 is also OK. Therefore, the reference 7 after the update can be regarded as more suitable than before the update. As described above, the optimization of the standard 7 can be realized by updating the standard 7.

[Standard update method]
FIG. 8 shows an example in which a histogram relating to a sensor value is used as a method and a processing example of updating the reference 7 in the first embodiment. The graph of the histogram in FIG. 8 is a plot of the frequency value for each sensor value according to the OK (OK) / NG (No) value, which is the inspection result of the quality of the final product OBL in the final inspection step 202. .. Here, the sensor value and the feature amount are the same. The horizontal axis of the graph is the sensor value (for example, fingertip pressure [Pa]), and the vertical axis is the frequency. The histogram 801 of the black plot is based on the data group in the case of NG (no), and the histogram 802 of the gray plot (indicated by the dot pattern) is based on the data group in the case of OK (possible). This example corresponds to the case where the sensor value of the sensor data D2 is one-dimensional data such as fingertip pressure. The server 10 creates such a histogram for the value (OK / NG) of the final inspection result from a large number of sensor data D2, for example, a data group having a predetermined number of data or more. The average value 803 is an average value of the OK value histogram 801, that is, a sensor value having an average frequency value. The value 804 is an example of the sensor value associated with the updated new reference 7.

As one way of thinking, it is desirable to set the value to be the reference 7 after the update near the middle of the distribution between the histogram 801 of the OK value and the histogram 802 of the NG value. As an example of the update process based on this idea, the server 10 obtains a value 804 obtained by subtracting a predetermined value α from the average value 803 of the OK value histogram 801 as shown in the figure. The predetermined value α is, for example, a value obtained by multiplying the standard deviation of the histogram 801 of the OK value by a predetermined coefficient. The server 10 uses this value 804 as an update value of the reference 7.

Other examples of update processing include the following. The server 10 obtains a value that does not include most of the data (for example, the number of data is 98%) of the histogram 802 of the NG value, and uses that value as the update value of the reference 7.

Other examples of update processing include the following. The server 10 uses a machine learning algorithm used for outlier detection and abnormality detection, for example, One Class Support Vector Machine. The server 10 uses the machine learning algorithm to obtain the discrimination boundary value obtained by learning the data group of the OK value, and sets the discrimination boundary value as the update value of the reference 7.

As described above, in the first embodiment, when the server 10 calculates the updated value of the reference 7 for each manufacturing process P, the updated value is calculated by a predetermined algorithm based on the value of the inspection result and the sensor value. To calculate.

[Screen example]
The server 10 may display a production control table as shown in FIG. 7 on the screen together with the GUI based on the operation of the user U1 of FIG. The user U1 such as the worker W and the manager can confirm the sensor value representing the work operation of the manufacturing process P, the reference 7, the determination result, and the inspection result in a state of comparison on the screen. Further, the server 10 may display a histogram graph as shown in FIG. 8 on the screen together with the GUI based on the operation of the user U1. The user U1 can confirm the histogram and the value 804 representing the updated reference 7 on the screen.

[Effects (1)]
As described above, according to the work evaluation system of the first embodiment, an appropriate standard 7 regarding the work operation of the manufacturing process P can be defined, and as a result, the accuracy of the manufacturing process P is improved, the quality of the product is improved, and so on. It is possible to improve the productivity of the production line. According to the first embodiment, in the entire manufacturing flow 200, the standard 7 of the manufacturing process P is updated to a more suitable value based on the inspection result of the quality of the product obtained as a result of the work. According to the first embodiment, it is possible to formulate a standard 7 based on the detection of work by a skilled technician, or to formulate and visualize a standard operation that is tacit knowledge in the field as the standard 7. .. As a result, it can be utilized for skill training of unskilled persons and suitable staffing of manufacturing flow.

Further, according to the first embodiment, it is possible to collect and analyze new data related to the evaluation / judgment of the work operation for the manufacturing flow 200, and as a result, it can be used for the judgment of the overall quality and the securing of evidence. Further, according to the first embodiment, it is possible to point out / prevent a work error or the like by determining the deviation from the standard 7 in the manufacturing process P. If the existing standard 7 is found to be unnecessary, it can be deleted.

[Transformation example 1-Screen example]
The following is also possible as a modification of the first embodiment. When the reference 7 is updated based on the first embodiment, the server 10 displays a screen for confirming the update of the reference 7 to the worker W and the like. When the reference 7 is updated, the server 10 may notify the user U1 such as the worker W of the update.

FIG. 9 shows an example of the screen in the first modification. On this screen, for example, information such as a product name, a manufacturing process name, a worker name, and a date and time is displayed at the upper part. Further, this screen has a process instruction column 901 and a process history column 902. In the process instruction column 901, the reference information 911 before the update (for example, "Please insert at a fingertip pressure of 80 Pa or more"), the update indicator 912, and the reference information 913 after the update (for example, "insert at 100 Pa or more from the next"). Please. ") And have. In the process history column 902, the history of data related to the manufacturing execution in the past period is displayed in, for example, a table format based on the above-mentioned production control table. In this table, for example, information such as an object ID, a sensor value and a determination result of a manufacturing process P (for example, an assembly process), a measured value and an inspection result of an inspection process, and the like are displayed in comparison.

When the reference 7 is updated, the server 10 lights up the update indicator 12 on the screen and also displays the updated reference information 913. As a result, it is notified to the worker W and the like that the standard 7 has been updated. A user U1 such as a worker W or an administrator can easily confirm the updated contents of the standard 7 on such a screen. Such display or notification may be performed at any time during the work by the worker W, or may be confirmed by an administrator, an inspector, or the like in units of, for example, one day or one week.

[Transformation example 2-Screen example]
FIG. 10 shows a screen example in the second modification. Based on the first embodiment, when the determination result in a certain manufacturing process P is NG, the server 10 displays a screen as shown in FIG. This screen is a screen that informs the worker W or the like in charge of the manufacturing process P that the determination result is NG and instructs the worker to retry the work. When the determination result based on the criterion 7 is NG, the server 10 not only determines the update of the criterion 7 described above, but also immediately displays the information to the effect that the determination result is NG on the screen of FIG. Notify the worker W with. The notification destination may be the manager of the manufacturing flow 200 or the like.

On the screen of FIG. 10, in the process history column 1002, the latest object, for example, with ID = 5, is highlighted when the determination result of the assembly process is NG. When the determination result is NG, the server 10 lights up the retry indicator 1013 in the process instruction column 1001 and instructs the process to retry the work. Item 1011 displays information on the current reference 7 and work instructions. In item 1014, the sensor value as a result of the current work operation and the information of the instruction for retrying the work (for example, "It was 30 Pa now. Please try the insertion again.") Is displayed.

Further, the server 10 may set a range that allows deviation from the reference 7 as a retry allowable range for the feature amount based on the sensor value of the work operation at the time of retrying the work. Item 1012 is a field for setting the degree of deviation within the allowable range of retries, and the user U1 can variably set the set value of the degree of deviation within the allowable range of retries by using a GUI component such as a slide bar. The retry tolerance range is, for example, a predetermined range from "easy" to "severe", and a value can be selected and set from the range. "Strict" is a value close to the standard 7, and "easy" is a value far from the standard 7. This set value can be set as a numerical value indicating the degree of deviation, for example, 0%, 10%, 20% from the value of the reference 7. For example, when the standard 7 is "80 Pa or more" at first and this set value is 20%, the standard 7 at the time of retry is from 80- (80 x 0.2) = 64 to "64 Pa or more". ". The reference 7 at the time of retry is not an update value, but a correction reference for retry, and is a value obtained by correcting the value of reference 7 based on a set value within the allowable range of retry. The server 10 may display the value of the correction reference on the screen.

The server 10 determines the degree of deviation from the reference 7 by comparing the feature amount based on the sensor value detected when the worker W retries the work with the correction standard, and the feature amount satisfies the correction standard. In that case, the result of the retry is OK (OK), and if the correction criteria are not satisfied, the result of the retry is NG (No). In the case of NG, the server 10 similarly instructs further retries. For example, in a certain work process, the standard 7 is "80 Pa or more". The result of the first work operation is "30 Pa", and the determination result is NG. The instruction for the first retry is given, and the correction standard is "64 Pa or more". If the result of the work operation of the first retry is, for example, "55 Pa", the judgment result is NG, and if the result of the work operation of the first retry is, for example, "70 Pa", the judgment result is OK. Will be.

As described above, the server 10 may control the retry of the work so as to allow some deviation from the reference 7. In this case, based on a certain standard 7, it is possible to take measures according to the correction standard according to the worker W. When it is difficult for some workers W to perform the standard operation with the reference 7, it is possible to deal with the retry with the correction reference. If the set value within the retry allowable range is too strict, retries may occur frequently depending on the worker W, and the work efficiency may deteriorate. Therefore, the manager may set the optimum value for the set value within the retry allowable range for each manufacturing process P.

As another control example, the server 10 may detect and record the number of times of work including retries with respect to the work by the worker W in the manufacturing process P by using the detection device 2 or the like, and use it for control. .. The server 10 sets, for example, a threshold value regarding the number of work retries. For example, when the number of work retries by the worker W exceeds the threshold value, the server 10 takes measures such as notifying the administrator in consideration of the possibility that the standard 7 is inappropriate.

[Modification 3-Camera]
FIG. 11 shows an explanatory diagram regarding the modified example 3. In the modification 3, as a specific example of the sensor 8 of the detection device 2, the glove 8A and the camera 8B are used for control in cooperation with each other. For example, as in (A) of FIG. 5, in a certain manufacturing process P, the worker W performs the work with the glove 8A and the camera 8B attached. The detection device 2 obtains and outputs the sensor data D2 of the glove 8A and the sensor data D2 of the camera 8B. The sensor data D2 in which these are combined represents the date and time, the number of times, good or bad, and other contents of the work by the worker W as the work data. The sensor data D2 of the glove 8A is, for example, time-series data of fingertip pressure as a sensor value. As the sensor of the glove 8A, an acceleration sensor or the like may be used, and in that case, the sensor data D2 of the glove 8A is data representing the state of movement of the hand 1100. The sensor data D2 of the camera 8B is time-series data of images taken of the object OB, the hand 1100 (corresponding glove 8A), etc. from the viewpoint of the worker W, and is data representing the state of the hand 1100 with respect to the object OB. Is. The server 10 uses the sensor data D2 to combine the sensor value of the globe 8A and the sensor value of the camera 8B, which have a corresponding relationship in time series, to make a determination regarding the update of the reference 7.

FIG. 11A shows an example of the image 1105 by the camera 8B mounted on the head of the worker W. In the image 1105, the connectors 1101 and 1102 of two cables (referred to as A and B) are inserted into two places of the object OB corresponding to the motor by the hand 1100 of the worker W in a certain manufacturing process. The state is reflected. The glove 8B attached to the hand 1100 has a pressure sensor 1103 built into the fingertip. The pressure sensor 1103 can measure the force applied to the fingertip at the time of insertion as the fingertip pressure.

The server 10 grasps the working state by image analysis of the moving image obtained as the sensor data D2. For example, the server 10 can specify which of the two cables the connectors 1101 and 1102 are inserted by the worker W's hand 1100 into the two locations of the object OB. The server 10 can specify, for example, the fingertip pressure when the connector of the cable is inserted with high accuracy based on the fingertip pressure which is the sensor value of the glove 8A together with the image analysis.

FIG. 11B shows a production control table corresponding to the example of the work of FIG. 11A. The table of this example stores data obtained from the manufacturing process in which the work (A) is performed and the inspection process immediately after that. The manufacturing process item has a sensor value and a determination result for each of the cable A and the cable B. The inspection process item has a measured value and an inspection result for each of the cable A and the cable B. Although not shown, the reference 7 is, for example, "50 Pa or more". Although not shown, the inspection conditions are, for example, "20 mA or less". In this example, in the second line where the object ID = 002, the sensor value (“40 Pa”) does not satisfy the standard 7 for the operation of inserting the cable A in the manufacturing process, so the judgment result is NG. ing. In the corresponding test result, the measured value (“50 mA”) does not satisfy the condition, so the test result is NG.

[Modification example 4-fourth pattern]
In the fourth modification, the server 10 is the first in the case where the determination result is NG and the inspection result is OK as the state of the combination of the determination result of the manufacturing process and the inspection result of the inspection process in the data of the production control table. Focusing on the four patterns, the decision to update the standard 7 is made.

FIG. 12 shows an example of a production control table including an example of the fourth pattern. In the data of the object ID = 403 in the third line, for example, the sensor value in the manufacturing process P1 is "80 Pa", the standard 7 is "100 Pa or more", and the judgment result is NG, and the measured value in the final inspection process. Is "10mA", the condition is "20mA or less", and the inspection result is OK. In this fourth pattern, although the determination result of the manufacturing process P1 in the previous stage is NG, the final inspection result in the latter stage is OK.

When the fourth pattern occurs, it can be estimated that the work operation in the manufacturing process P1 does not lower the quality of the object OB, and the reference 7 of the manufacturing process P1 is too strict. Therefore, the server 10 updates the reference 7 of the manufacturing process P1 so as to be easy. In this case, for example, the same method as described above can be applied as a method for determining the update value of the reference 7. The server 10 finds an updated value of the reference 7 such that the determination result and the inspection result match in all the data groups related to the manufacturing process P1. For example, a range of "41 Pa or more" to "80 Pa or more" can be considered as a candidate. From this range, the server 10 adopts "60 Pa or more" as the updated reference 7 as an intermediate point, for example. Criteria 7 is updated from "100 Pa or more" to "60 Pa or more". This makes it easier to determine the standard operation, and there is a high possibility that a consistent result will occur in which the determination result will be OK and the final inspection result will also be OK in the future.

[Modification 5-Training]
In the modified example 5, it is applied not to the actual production line but to the training field related to the work process. The mechanism using the above-mentioned standard 7 can also be applied to the training and education of the worker W. A training ground that imitates the manufacturing process of the manufacturing flow 200 is provided. At the training ground, workers W, especially unskilled workers, are trained in the work of the manufacturing process. In advance, the standard 7 of the manufacturing process is set based on the detection of the work operation by the expert. The system of the modified example 5 compares the feature amount based on the sensor value of the acquired sensor data D2 with the standard 7 based on the expert in the training work by the unskilled person as described above. Whether or not it is possible is judged from the presence or absence of standard operation. The server 10 displays the training determination result on the screen.

FIG. 13 shows a case where the training result is displayed as a production control table as an example of the above screen. This table is a table summarizing the data regarding the work of, for example, the manufacturing process P1 regarding the trainee having a certain worker ID. The manufacturing process items in the table include trainer sensor value items, expert standard items, determination result items, and deviation items. In the expert standard item, the standard set based on the data of the work operation of the expert is displayed. The expert standard is, for example, "50 Pa or more". In the deviation item, the degree of deviation between the trainer sensor value and the expert standard, for example, the difference between the trainer sensor value and the threshold value of the expert standard is displayed.

For example, in the data on the first line where the object ID = 501, the trainee sensor value is "140 Pa" and the determination result is OK because the expert standard is satisfied. The data in the second line, which is 502, has a trainer sensor value of "25 Pa" and does not satisfy the expert standard, so the determination result is NG. The value of the dissociation changes, for example, from the first line to the fifth line, such as +90, -25, +45, -5, +5, and gradually decreases.

Trainers and managers can easily recognize the difference between skilled and unskilled people on such a screen. The trainee can grasp, for example, how much force the fingertip pressure at the time of inserting a part is suitable for the work of the manufacturing process as a quantitative numerical value. As described above, in the modification 5, the training and the like can be made more efficient by visualizing the difference between the skilled person and the unskilled person involved in the work operation of the manufacturing process based on the data.

<Embodiment 2>
The work evaluation system of the second embodiment will be described with reference to FIG. The basic configuration in the second embodiment and the like is the same as that in the first embodiment, and the components different from the first embodiment in the second embodiment and the like will be mainly described below. The second embodiment is a modification of the first embodiment, and shows another configuration example relating to the cooperation with the MES 20.

[MES]
FIG. 14 shows the configuration of the work evaluation system of the second embodiment. In this work evaluation system, in the MES 20, the management device 5 which is a MES server includes a process quality assurance unit 22, a standard operation assurance unit 23, and a comprehensive quality display unit 24 as functional blocks realized by program processing and the like. Have. The process quality assurance unit 22 has the same function as the process quality assurance unit 12 described above (FIG. 1), performs a process quality assurance determination, and outputs process quality assurance data D22. On the other hand, the processing device 10 which is a server does not include the above-mentioned process quality assurance unit 12.

The process quality assurance unit 22 collects and acquires the inspection data D3 of the process inspection device 3 and the inspection data D4 of the final inspection device 4. This acquisition may be via DB6. The process quality assurance unit 22 grasps the quality of the object OB as a result of each manufacturing process P and the quality of the final product OB in the final inspection process 202 based on the acquired inspection data D3 and D4. The process quality assurance unit 22 makes a process quality assurance determination based on those qualities, creates process quality assurance data D22, and sends it to the comprehensive quality display unit 24. The standard update unit 13 of the server 10 acquires the process quality assurance data D22 from the process quality assurance unit 22, and makes a determination regarding the update of the above-mentioned standard 7.

The standard operation guarantee unit 23 collects and acquires the standard operation determination result data D11 from the standard operation determination unit 11 of the server 10. This standard operation determination result data D11 is new data that is not handled by the conventional MES. The standard operation guarantee unit 23 grasps the quality related to the work operation of the manufacturing process P based on the standard operation determination result data D11, creates the standard operation guarantee data D23 representing the result, and sends it to the comprehensive quality display unit 24. In the standard operation determination result data D11, if it is an OK value, it indicates that there is a standard operation. Therefore, in that case, the standard operation guarantee unit 23 can set the value of the standard operation guarantee data D23 as a value indicating that the quality related to the work operation is guaranteed.

The comprehensive quality display unit 24 uses the process quality assurance data D22 and the standard operation assurance data D23 to create information indicating the overall quality of the manufacturing flow 200, and displays the information on the screen. The user U1 can confirm the overall quality by looking at the information on the screen.

According to the second embodiment, the MES 20 acquires the standard operation determination result data D11 as new data in addition to the process quality based on the inspection data D3 and D4. Then, the MES 20 can output information on the total quality of the manufacturing flow 200 created from the process quality assurance data D21 and the standard operation assurance data D23.

<Embodiment 3>
The work evaluation system of the third embodiment will be described with reference to FIG. The third embodiment has a function of controlling the update timing of the reference 7. In the third embodiment, the process of the server 10 of FIG. 1 is basically the same as the flow of FIG. 6, and some processes such as step S5 are different. It is possible to control the update timing of the reference 7 according to the definition of the processing content in step S5.

[First control example]
The first control example in the third embodiment is as follows. FIG. 15A shows a first control example. In the first control example, the server 10 updates the reference 7 by using the data number N or the like of the object OB when the work is started by the worker W in charge of the manufacturing process P, for example, the worker W is replaced. Determine the timing of. Consider a case where a shift shift system is used with respect to the arrangement of a plurality of workers W in the manufacturing flow 200. For example, it is assumed that a plurality of (m) workers W such as workers W1, W2, ..., Wm take turns in charge of a certain manufacturing process A. In this case, the server 10 controls the update timing of the reference 7 applied to the manufacturing process A as follows. In step S5, the server 10 grasps the data number N of the object OB after the start of the work of the manufacturing process P for each worker W. When the number of data N becomes the predetermined number of data NX, the server 10 proceeds to step S6 to update the reference 7. The number of data NX is set to a relatively small value such as 10. The server 10 determines the update value of the reference 7 by, for example, by the above-mentioned method, using the data acquired in the period corresponding to the number of data NX.

In the example of (A) of FIG. 15, for the manufacturing process A, the worker W1 is assigned to the time from 9:00 to 12:00 on June 1. First, it is assumed that the standard 7 of the manufacturing process A is the standard A1. After the work is started, the worker W1 is sequentially working on the object OB having the number of data N1 = 10 such as the object ID = OB1, OB2, ..., OB10. Criterion A1 is applied in the determination of those operations. Next, based on the determination in step S5, when the cumulative number of data N from the start of work reaches the predetermined number of data NX (for example, 10), the reference 7 is updated from the reference A1 to the reference A2. There is. For example, subsequently, in the work on the object OB having the number of data N2 = 10 such as the object ID = OB11, OB12, ..., OB20, the reference A2 is applied. Next, it is assumed that the worker W1 is replaced by the worker W2. Immediately after the start of work by the worker W2, the standard 7 is inherited from the standard A2 at the time of the previous worker W1. After the work is started, the worker W2 is sequentially working on the object OB having the number of data N3 = 10 such as the object ID = OB21, OB22, ..., OB30. When the number of data N from the start of the work of the worker W2 reaches the predetermined number of data NX (for example, 10), the reference 7 is updated from the reference A2 to the reference A3. For example, subsequently, in the work of the object OB having the number of data N4 = 10 such as the object ID = OB31, OB32, ..., OB40, the reference A3 is applied.

As described above, in the first control example, even if the worker W in charge of the manufacturing process P is changed, the reference 7 is updated relatively immediately based on the number of data NX after the work is started by the worker W. Thereby, in the first control example, the reference 7 can be immediately updated to a more suitable value so as to reflect the influence of the individual difference of the worker W. In the above example, the case where the timing of updating the reference 7 after the start of work is determined has been shown, but it is possible to control the frequency and cycle of updating the reference 7 based on the same idea.

[Second control example]
The second control example is as follows. FIG. 15B shows a second control example. In the second control example, the server 10 determines the update timing (for example, frequency) of the reference 7 in consideration of the skill level of the worker W in the manufacturing process P. In step S1 of FIG. 6, the server 10 also grasps the skill level of the worker W who is in charge of the manufacturing process P. For example, there are workers W1, W2, and W3 for the work of a certain manufacturing process P. The worker W1 has a skill level L1, the worker W2 has a skill level L2, and the worker W3 has a skill level L3. It is assumed that the skill level is defined as L1, L2, L3 from the lower side to the higher side, for example. The worker W1 has a low skill level, and the worker W3 has a high skill level. In the case of the highly skilled worker W3, since he is accustomed to the work, there is a high possibility that the standard operation can be performed. On the other hand, in the case of the worker W1 with a low skill level, there is a high possibility that the standard operation cannot be performed because he / she is not accustomed to the work. The proficiency level may be set by a user U1 such as an administrator, or may be automatically determined and set by the server 10 or MES20.

The server 10 changes the frequency of updating the standard 7 according to the difference in the skill level of the worker W even for the same manufacturing process P. The frequency can be controlled by the number of data NX in step S5. For example, in the case of skill level L1, the number of data is NX1 (for example, 10), in the case of skill level L2, the number of data is NX2 (for example, 50), in the case of skill level L3, the number of data is NX3 (for example, 100), and so on. It may be set in advance (NX1 <NX2 <NX3). When the worker W1 having a low skill level starts the work, the server 10 determines the update timing of the reference 7 by using the number of data NX1. As a result, in the case of the worker W1 having a low skill level, the frequency of updating the reference 7 can be increased, in other words, the time until the reference 7 is updated can be shortened. As described above, in the second control example, the standard 7 can be updated according to the skill level of the individual. The frequency is not limited to the number of data NX, and can be specified by time.

[Third control example]
The third control example is as follows. In the third control example, the timing of updating the reference 7 is controlled in consideration of the physical condition of the worker W in charge of the manufacturing process P. A wearable terminal or the like is used as the sensor 8 attached to the body of the worker W. For example, a sensor of a wearable terminal has a function of measuring vital values such as body temperature, heart rate, and blood pressure. The server 10 acquires the sensor data D2 including the vital value (at least one type of vital value). The server 10 determines the physical condition of the worker W based on the vital value. The physical condition may be expressed by the vital value itself, or may be defined by a value such as good, normal, or bad. In step S5, the server 10 adjusts the timing for updating the reference 7 according to the physical condition of the worker W. In step S5, the server 10 reflects (for example, addition or multiplication) a value for adjustment according to the state of physical condition in, for example, the number of data NX. For example, when the physical condition is good, the positive value is reflected in the data number NX so that the data number NX becomes larger, and when the physical condition is bad, the data number NX becomes smaller. , Reflect a negative value in the number of data NX. As a result, if the patient is in poor physical condition, the standard 7 can be updated more quickly. As described above, in the fourth control example, it is possible to update the standard 7 adapted to the physical condition of the individual.

[Fourth control example]
The fourth control example is as follows. In the fourth control example, the timing of updating the reference 7 is further controlled in consideration of the environmental value in the work place of the manufacturing process P. Examples of the environmental value include temperature, humidity, volume of noise, and the like. Sensors for measuring environmental values are installed in the work area. The server 10 acquires the environmental value as sensor data from the sensor. The server 10 acquires and refers to the environmental value when the worker W in the manufacturing process P works at a specified time such as at the start of the work. For example, the server 10 adjusts the update timing of the reference 7 according to the environment value. For example, the server 10 determines when the temperature is higher or lower than the standard, and determines the update timing by adjusting the number of data NX in step S5 according to each case. For example, when the temperature of the server 10 is higher than the standard, it can be estimated that the standard operation is difficult, so that the update timing of the reference 7 is earlier than usual. Each of the above-mentioned control examples can also be applied in combination.

[Modification example-Standards for each worker]
The following is also possible as a modification of the third embodiment. In this modification, how one or more criteria 7 are defined or updated when the worker W in charge of a certain manufacturing process P can change, in other words, when a plurality of workers W exist as candidates. I devised something about it. In this modification, the work operation for each manufacturing process P is updated to be a suitable standard 7 in consideration of individual differences for each worker W. More specifically, in this modification, as two types of standards, the base standard (described as "base standard") regardless of the difference in the worker W and the individual difference for each worker W are reflected. It is used in combination with the standard (described as "individual standard").

FIG. 16 shows a configuration example of the reference 7 for each worker W in the modified example. Similar to (A) in FIG. 15, it is assumed that a plurality of (m) workers W (W1, ..., Wm) exist as candidates for the worker W in charge of a certain manufacturing process P. For example, the worker W1 has a skill level L1 and is a standard T1 as an applied standard 7. The worker W2 has a skill level L2 and is a standard T2 as an applied standard 7. The worker W3 has a skill level L3 and is a standard T3 as an applied standard 7.

The server 10 sets the standard 7 for each worker W for the manufacturing process P by the following method. It is assumed that the reference 7 is represented by the symbol T. Let Tb be the base reference. The personal standard is Tp. The initial value of the base reference Tb is set in advance. The personal standard Tp is set for each worker W. The server 10 creates or updates the personal standard Tp of the worker W based on the data acquired in response to the work by the worker W (for example, data of a predetermined number of data). The server 10 similarly creates or updates an individual standard Tp for each worker W of a plurality of workers (W1 to Wm). From these results, the server 10 can grasp the tendency of the individual standard Tp of each worker W. The server 10 calculates statistical values such as the average of all individual reference Tp and the deviation of each individual reference Tp.

After that, the server 10 updates the base reference Tb using the statistical value of the personal reference Tp at a predetermined timing. For example, when the average value of the personal reference Tp is +10, the server 10 updates the base reference Tb to the same value as the average value (+10). As a result, a suitable base reference Tb that reflects the tendency of all workers W can be obtained.

When the standard 7 is applied to the actual manufacturing flow 200, the server 10 may apply the base standard Tb when a plurality of workers W are in charge of each manufacturing process P. Further, the server 10 may be applied so as to switch the personal reference Tp according to the identification of the worker W for each manufacturing process P. Since there are individual differences in the worker W even in the same manufacturing process P, it is considered that the optimum standard 7 is different. According to this modification, a suitable reference 7 that reflects the individual difference of the worker W can be set.

For example, the worker W1 has an average fingertip pressure of 70 Pa as a work operation in which the quality of the object OB is ensured. The worker W2 has an average fingertip pressure of 60 Pa. The worker W3 has an average fingertip pressure of 50 Pa. Based on these, the reference T1 (= Tp1) is set to, for example, "65Pa or more", the reference T2 (= Tp2) is set to, for example, "55Pa or more", and the reference T3 (= Tp3) is set to, for example, "45Pa or more". Is said to be. Efficiency can be improved by switching and applying the personal standard Tp according to the worker W.

In the above example, two types of criteria (Tb, Tp) are considered, but in another way of thinking, the following may be considered. As a basic formula for the provision of the standard 7, T = Tb + Tp or T = Tb × Tp. The idea in this formula is to reflect the individual difference for each worker W as Tp by addition or multiplication with respect to the base reference Tb. The personal standard Tp in this case corresponds to a correction coefficient or the like. For example, the reference T1 for the worker W1 can be determined as T1 = Tb + Tp1 using the individual reference Tp1. The reference T2 for the worker W2 can be determined as T2 = Tb + Tp2 using the individual reference Tp2. When determining the work, the server 10 determines the standard 7 for the worker W by reflecting the personal standard Tp of the worker W according to the identification of the worker W, and updates it as needed. Set. For example, when the base reference Tb is "50 Pa or more" and the individual reference Tp1 of the worker W1 is +5, the reference T1 for the worker W1 is determined to be "55 Pa or more" from T1 = Tb + Tp1. ..

<Embodiment 4>
The work evaluation system of the fourth embodiment will be described with reference to FIG. In the fourth embodiment, the reference 7 is defined by using machine learning as the reference updating method. A supervised learning method can be used for the discriminative model in this machine learning. This method includes, for example, Support Vector Machine, Random Forest, Logistic Regression, and Neural Network.

[Machine learning]
FIG. 17 shows a flow of a processing example of the server 10 corresponding to the reference update method using machine learning in the fourth embodiment. FIG. 17A shows a processing flow at the time of initial learning, and FIG. 17B shows a processing flow at the time of operation. In (A), at the time of the first learning, first, in steps S401 and S402, the server 10 obtains the data related to the detection of the work on the object OB for each manufacturing process P described above with a predetermined number of data NA for the first learning. , Measure / acquire. In step S403, the server 10 learns the discriminative model based on the data of NA pieces, using the value (OK / NG) of the result of the final inspection step 202 as the correct label.

In (B), at the time of operation, first, in steps S311 and S312, the server 10 measures and acquires the above data with a predetermined number of data NB for operation. In step S413, the server 10 discriminates and determines NB data by the discriminative model.

FIG. 17C is a table in which the identification result and the inspection result of the final inspection step 202 are expressed by a mixed matrix. In this table, the rows are the determination results of the standard operation of the manufacturing process P, and the values corresponding to the above identification results (OK / NG), and the columns are the values of the inspection results of the final inspection process 202 (OK / NG). NG). The cells of the matrix are the number of data according to the combination of those values. For example, in the combination corresponding to the above-mentioned third pattern, the determination result of the manufacturing process is OK and the final inspection result is NG, the number of data is five. For example, when the number of data NA = 100, the correct answer rate is (50 + 42) / (50 + 5 + 3 + 42) = 92%. The rate corresponding to the third pattern is 5%.

Next, in step S414, the server 10 determines whether the value of the evaluation index related to machine learning exceeds the threshold value. If it exceeds (Y), the flow ends. If the value of the evaluation index is below the threshold value (N), the process proceeds to step S415. In step S415, the server 10 relearns the discriminative model using NB data. Criterion 7 is updated by re-learning. The calculation method of the evaluation index includes, for example, a correct answer rate, a precision rate, a recall rate, and an F value, all of which are applicable. As described above, according to the fourth embodiment, the reference 7 can be efficiently updated by using machine learning.

<Embodiment 5>
The work evaluation system of the fifth embodiment will be described with reference to FIGS. 18 to 19. The fifth embodiment is an extended configuration example based on the first embodiment. In the fifth embodiment, the manufacturing process P to be updated is determined for the plurality of manufacturing processes P and the inspection process of the manufacturing flow 200, and the reference 7 of the manufacturing process P is updated. The reference updating method described in the first embodiment can be basically applied to a plurality of manufacturing processes P in the same manner. In the manufacturing flow 200 of FIGS. 3 and 4 described above, it is assumed that a determination result is obtained in each manufacturing process P in the manufacturing process flow 201, and an inspection result is obtained in an inspection process for each manufacturing process P. In the fifth embodiment, the server 10 is a manufacturing process to be updated based on the inspection result of the final inspection process 202 and the state of the combination of the inspection result for each manufacturing process P and the determination result for each manufacturing process P. P is determined, and the standard 7 of the manufacturing process P is updated.

[Multiple processes]
FIG. 18A shows a configuration example of a manufacturing flow 200 corresponding to a production line of a certain motor product. The plurality of manufacturing processes P of the manufacturing process flow 201 are, in order, the first assembly process as the first manufacturing process P1, the first inspection process K1, the coating process as the second manufacturing process P2, the second inspection process K2, and the third. It has a second assembly process as a manufacturing process P3 and a third inspection process K3. Subsequent final inspection step 202 is, for example, a visual inspection step, and then shipping. The first inspection step K1 is an inspection step relating to the quality of the object OB as a result of the first manufacturing step P1. The same applies to the second inspection step K2 and the third inspection step K3.

FIG. 18B shows an example of a production control table in the case of the manufacturing flow 200 of FIG. 18A. This table shows a combination of judgment results in each work process (P1, P2, P3) of each manufacturing process P, inspection results in each inspection process (K1, K2, K3), and inspection results in the final inspection process 202. An example is shown, and items such as the sensor value described above are omitted.

For example, in the second line where the object ID = 5002, the determination result of the painting process, which is the second manufacturing process P2, and the final inspection result are NG, and the others are OK. In the case of such a combination, it can be presumed that the reason why the final inspection result is NG is that the server 10 has a high possibility of coating unevenness in the coating process which is the second manufacturing process P2.

On the other hand, for example, in the third line where the object ID = 5003, the final inspection result is OK, the result of the third inspection step K3 is NG, but the others are OK. In the case of such a combination, the server 10 can presume that the reference 7 is too easy in at least one of the manufacturing steps (P1, P2, P3) prior to the third inspection step K3. In other words, the judgment result was OK because the standard 7 in a certain manufacturing process was too easy, but the work operation in that manufacturing process affected the quality of the object OB, and the result of the third inspection process K3 was NG. It can be estimated that it became. Therefore, the server 10 targets these three manufacturing processes (P1, P2, P3) as the reference update target. The server 10 updates, for example, each standard 7 of the three manufacturing processes (P1, P2, P3) so as to be stricter. As a result, the entire manufacturing flow 200 can be set to a more appropriate standard 7.

In this way, when the inspection result in the inspection process (inspection process for each manufacturing process P or final inspection process 202) in the latter stage of the manufacturing flow 200 is NG, the inspection process in the earlier stage is obtained. If there is data of the inspection result of, the cause can be investigated and estimated quickly and more accurately. The server 10 can update the presumed cause process to a more appropriate criterion 7. In the prior art example, it was affected by all the latter stages of the manufacturing flow, and it took a long time to investigate the cause as described above. In the fifth embodiment, as described above, the server 10 is a combination of the inspection result of the final inspection step 202, the inspection result of each manufacturing process P, and the determination result of each manufacturing process P, and the inspection result is NG. The manufacturing process P that causes the problem is estimated, and the standard 7 of the manufacturing process P is updated. In the fifth embodiment, the accuracy of the reference 7 can be gradually improved by updating the reference 7 while operating such a mechanism in the actual manufacturing flow 200.

[Cause process estimation method]
Examples of the method for estimating the manufacturing process P that causes the above inspection result to be NG in the fifth embodiment include the following. The server 10 examines one or more manufacturing processes P in front of the process that has become NG as a candidate. The server 10 creates a histogram regarding the sensor values for the data of the candidate manufacturing process P in the same manner as described above (FIG. 8). In the table of FIG. 18B, in the example of the third row in which the object ID = 5003, when the inspection result of the third inspection step K3 is NG, each manufacturing step (P1) prior to that. , P2, P3), the determination result is OK. Therefore, the server 10 sets each of these manufacturing processes (P1, P2, P3) as a cause candidate. The server 10 creates a histogram for each manufacturing process of the cause candidate, and determines the manufacturing process estimated to have the greatest influence as the cause from the analysis result of the histogram. Then, the server 10 updates the reference 7 of the manufacturing process P.

FIG. 19 shows a processing example using a histogram in the cause process estimation method. In this example, a processing example for the data in the third row in which the object ID = 5003 is shown. The server 10 creates a histogram for each of the NG value of the inspection result of the third inspection step K3 and the OK value of the determination result of the three manufacturing processes (P1, P2, P3). The method of creating the histogram is the same as that described above (FIG. 8), and the histogram of the OK value and the histogram of the NG value are arranged in parallel.

(A) shows a histogram 1911 of the NG value of the inspection result of the third inspection step K3 and the histogram 1912 of the OK value of the determination result of the first manufacturing step P1 which is the first assembly step. The sensor values on the horizontal axis are, for example, fingertip pressures. (B) shows a histogram 1921 of the NG value of the inspection result of the third inspection step K3 and the histogram 1922 of the OK value of the determination result of the third manufacturing step P3 which is the second assembly step. (C) shows a histogram 1931 of the NG value of the inspection result of the third inspection step K3 and the histogram 1932 of the OK value of the determination result of the second manufacturing step P2 which is the painting step.

Regarding the first assembly step (A) and the second assembly step (B), there is a difference in peak between the histogram of the NG value (1911, 1921) and the histogram of the OK value (1912, 1922). do not have. For example, in the histogram 1911 of the NG value and the histogram 1912 of the OK value, each peak is at "40 to 50 Pa". On the other hand, in the coating process (C), there is a difference in peak between the NG value histogram 1931 and the OK value histogram 1932. The peak of the NG value histogram 1931 is at "20-30 Pa", and the peak of the OK value histogram 1932 is at "50-60 Pa".

From the histogram analysis result as described above, it can be estimated that the cause of the inspection result of the third inspection step K3 being NG is highly likely to be the second manufacturing process P2 which is the painting process. In this case, the server 10 determines the second manufacturing process P2 presumed to be the cause as the reference update target.

As described above, according to the fifth embodiment, each standard 7 for the plurality of manufacturing processes P of the manufacturing flow 200 can be updated to an appropriate value, and overall efficiency can be realized.

Although the present invention has been specifically described above based on the embodiments, the present invention is not limited to the above-described embodiments and can be variously modified without departing from the gist.

1 ... work evaluation system, 2 ... detection device, 3 ... process inspection device, 4 ... final inspection device, 5 ... management device, 6 ... DB, 7 ... standard, 8 ... sensor, 9 ... communication network, 10 ... processing device ( Server), 11 ... Standard operation judgment unit, 12 ... Process quality assurance unit, 13 ... Standard update unit, 14 ... Standard operation judgment result output unit, 15 ... Standard output unit, 20 ... Manufacturing execution system (MES), 200 ... Manufacturing Flow, 201 ... Manufacturing process flow, 202 ... Final inspection process, P ... Manufacturing process, W ... Worker, OB ... Object, OBL ... Final product, D2 ... Sensor data, D3 ... Inspection data, D4 ... Inspection data, D7 ... Reference data, D11 ... Standard operation judgment result data, D12 ... Process quality assurance data, D13 ... Standard update data.

Claims (18)

A detection device that detects the work operation of an object by a worker in the manufacturing process of the manufacturing flow as sensor data, and a detection device.
The processing device connected to the detection device and
Equipped with
The processing device is
The sensor data and the inspection result of the quality of the object in the inspection process after the manufacturing process of the manufacturing flow are acquired.
Based on the sensor data and the criteria for determining the standard operation of the manufacturing process, the propriety of the manufacturing process is determined based on the determination of the standard operation for the working operation in the manufacturing process.
Based on the determination result of the manufacturing process and the inspection result, it is determined whether to update the standard of the manufacturing process, and the standard is updated according to the determination result.
Work evaluation system. In the work evaluation system according to claim 1,
Based on the data of the result of the work for a plurality of objects in the manufacturing flow, the processing apparatus matches the case where the value of the judgment result of the manufacturing process and the value of the possibility of the inspection result do not match. The standard of the manufacturing process is updated so as to be in a state of
Work evaluation system. In the work evaluation system according to claim 2,
The processing apparatus updates the case where the determination result of the manufacturing process is acceptable and the inspection result is negative so as to tighten the standard of the manufacturing process.
Work evaluation system. In the work evaluation system according to claim 2,
The processing apparatus updates the case where the determination result of the manufacturing process is negative and the inspection result is acceptable so as to facilitate the standard of the manufacturing process.
Work evaluation system. In the work evaluation system according to claim 1,
The processing device displays the updated reference on the screen.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus displays the determination result of the manufacturing process on the screen.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus outputs an instruction to retry the work when the determination result of the manufacturing process is negative.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus is connected to a manufacturing execution system that manages the manufacturing flow and acquires the inspection result, acquires the inspection result from the manufacturing execution system, and provides the determination result of the manufacturing process to the manufacturing execution system.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus determines the update of the standard when the number of data related to the object reaches a predetermined number after the worker in charge of the manufacturing process starts the work.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus controls the timing of updating the standard according to the skill level of the worker in charge of the manufacturing process.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus grasps the physical condition based on the vital value of the worker in charge of the manufacturing process, and controls the timing of updating the reference according to the physical condition.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus grasps the environmental value of the work site of the manufacturing process and controls the update timing of the standard according to the environmental value.
Work evaluation system. In the work evaluation system according to claim 1,
When a plurality of workers exist as candidates for the worker in the manufacturing process, the processing apparatus sets a standard reflecting individual differences for each worker with respect to the standard in the manufacturing process.
Work evaluation system. In the work evaluation system according to claim 1,
The processing device is
The sensor data in each manufacturing process of a plurality of manufacturing processes in the manufacturing flow, the inspection result in each inspection process after each manufacturing process, and the final inspection result in the final inspection process are acquired.
Whether or not each of the manufacturing processes is possible is determined, and
Based on the determination result of each manufacturing process, the inspection result in each inspection process, and the inspection result in the final inspection process, the manufacturing process to be updated with the standard is determined, and the determination result is determined. Update the above criteria,
Work evaluation system. In the work evaluation system according to claim 1,
The detection device detects the work operation as the sensor data by using the sensor mounted on the worker's hand.
The processing device determines the presence or absence of the standard operation by comparing the feature amount according to the sensor data with the reference that defines the conditions for the feature amount.
Work evaluation system. In the work evaluation system according to claim 1,
The detection device detects an image of the work operation as the sensor data by using a camera attached to the worker or installed at the work site.
The processing device determines the presence or absence of the standard operation by comparing the feature amount according to the sensor data with the reference that defines the conditions for the feature amount.
Work evaluation system. In the work evaluation system according to claim 1,
The processing apparatus creates a first histogram based on data for which the inspection result is negative and a second histogram based on data for which the inspection result is acceptable as a histogram of the values of the sensor data for the manufacturing process. The reference is updated with a value near the middle between the first histogram and the second histogram.
Work evaluation system. It is a work evaluation method in a system including a detection device that detects a work operation on an object by a worker in a manufacturing process of a manufacturing flow as sensor data, and a processing device connected to the detection device.
As a step performed by the processing device,
A step of acquiring the sensor data and the inspection result of the quality of the object in the inspection step after the manufacturing step of the manufacturing flow.
Based on the sensor data and the criteria for determining the standard operation of the manufacturing process, the step of determining the propriety of the manufacturing process based on the determination of the standard operation of the working operation in the manufacturing process.
A step of determining whether to update the standard of the manufacturing process based on the determination result of the manufacturing process and the inspection result, and updating the standard according to the determination result.
Has a work evaluation method.

Images