Manufacturing line work improvement support device, manufacturing line work improvement support method and program

Abstract

To properly specify an improvement object process for improving a work.SOLUTION: A calculation part 212 obtains a process work time on the basis of imaging data for imaging a work for each process of a manufacturing line and obtains a characteristic value showing characteristics of the work time for each process on the basis of the process work time. An improvement object process specification part 213 specifies, as an improvement object process, a process required to be improved preferentially than other processes on the basis of the characteristics.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a manufacturing line operation improvement support device, a manufacturing line operation improvement support method, and a program. [Background technology]

[0002] There are known techniques for extracting problems in a manufacturing line and providing support for improving them. For example, Patent Document 1 discloses a method for extracting problems in a manufacturing line from data obtained by a manufacturing simulation and presenting solutions to the extracted problems. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2003-280730 Summary of the Invention [Problem to be solved by the invention]

[0004] When improving problems on a production line, it is desirable to identify the bottleneck process among all processes and target that process for improvement in order to increase the efficiency of the improvement. However, it is difficult to objectively identify the bottleneck process. For this reason, applying the method described in Patent Document 1 to identify the bottleneck process is considered. However, in an actual production line, the work time for each process changes daily due to factors such as worker changes, making it difficult to obtain actual work times that match the simulation. For this reason, even if a problematic process is identified based on data obtained from a simulation, it may lead to the identification of a process in the actual production line that is different from the problematic process.

[0005] The present disclosure has been made in consideration of the above-mentioned problems, and aims to more appropriately identify a process in a manufacturing line that is a target for work improvement. [Means for solving the problem]

[0006] To achieve the above object, a manufacturing line work improvement support device according to the present disclosure includes a calculation unit and an improvement target process identification unit. The calculation unit calculates a process work time based on image data of work performed at each process on the manufacturing line, and calculates multiple characteristic values ​​indicating the characteristics of the work time for each process based on the calculated process work time. The improvement target process identification unit uses the calculated multiple characteristic values ​​to identify a process that requires priority improvement over other processes as an improvement target process. The multiple characteristic values ​​include an average process work time for each process and a dispersion indicating the variation in process work time, and the improvement target process identification unit inputs the combination of the average process work time and the dispersion calculated by the calculation unit into a trained model obtained by learning the combination of the average process work time and the dispersion and the process that needs to be improved as a priority when that combination is obtained, and presents the output improvement target process. [Effects of the Invention]

[0007] According to the present disclosure, it is possible to easily and accurately obtain performance-based work time based on photographic data of work performed at each process on a production line. Furthermore, it is possible to objectively obtain characteristic values ​​that indicate the characteristics of the actual work time at each process based on the obtained multiple work times. Furthermore, it is possible to objectively identify processes that require work improvement based on the characteristic values. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a block diagram illustrating a configuration example of a production line work improvement support system according to a first embodiment of the present disclosure. [Figure 2] Flowchart of process for identifying improvement targets [Figure 3] (a) A table showing the average process work time and standard deviation for each process, (b) A graph showing the average process work time and standard deviation for each process [Figure 4] A diagram showing an example of data showing a process to be improved and an improvement implementation example. [Figure 5] FIG. 1 is a diagram illustrating an example of a hardware configuration of a manufacturing line operation improvement support device according to a first embodiment of the present disclosure. [Figure 6] A block diagram showing a configuration example of a production line work improvement support system according to a second embodiment. [Figure 7] FIG. 10 is a diagram showing an overview of a neural network according to a second embodiment. [Figure 8] Flowchart showing learning processing according to the second embodiment [Figure 9] Flowchart showing inference processing according to the second embodiment DETAILED DESCRIPTION OF THE INVENTION

[0009] (Embodiment 1) A first embodiment of the present disclosure will be described in detail with reference to the drawings.

[0010] The manufacturing line work improvement support system 1 according to the first embodiment supports work improvement by taking videos of work at a production site where various process tasks are performed, identifying processes that should be improved as a priority, in other words, bottleneck processes, and outputting improvement implementation examples for the bottleneck processes. Note that a process that needs to be improved as a priority over other processes will be referred to as an improvement target process or a bottleneck process hereinafter. Here, process tasks broadly include, for example, transportation, conveyance, processing, machining, assembly, packaging, and carrying out.

[0011] As shown in FIG. 1, the production line work improvement support system 1 includes a photographing device 100 that photographs video of workers' work, and a production line work improvement support device 200 that identifies and outputs processes to be improved based on the work video.

[0012] In the following description, it is assumed that there are six processes in the production line, and each process has a different process work area and a different worker in charge. One image capture device 100 is placed in each process work area. That is, the number of image capture devices 100 installed corresponds to the number of process work areas, and in this embodiment, six image capture devices 100 are installed.

[0013] The image capturing device 100 is any image capturing device, such as a video camera, that captures images of a task including a worker who is performing the task and a work object that is the target of the task. The image capturing device 100 is installed above the process work area in a position where it can capture images of the work object. In this embodiment, six video cameras are installed above each worker in the six process work areas.

[0014] The image capture device 100 continuously captures images of the production line while it is in operation, generating live-view images. The image capture device 100 stores in its internal memory the external shapes of the worker and the workpiece at the start of the work and at the end of the work. The image capture device 100 determines that the work has started when a pre-registered external shape at the start of the work is detected in the live-view image, and starts recording the live-view image while recording the time the work started based on the value of its built-in timer. The image capture device 100 determines that the work has been completed when a pre-registered external shape at the end of the work is detected in the live-view image, and stops recording the live-view image while recording the time the work finished based on the value of its built-in timer. The recorded video of the actual work is stored in an internal storage device as an actual work video file, along with the identification number of the image capture device 100 installed for each process and the start and end times measured by the clock built into the image capture device 100.

[0015] The imaging device 100 is communicably connected to the production line work improvement support device 200 and transmits the stored video files to the production line work improvement support device 200 at a predetermined timing. In the first embodiment, the predetermined timing is, for example, when a production line operation manager starts up the production line work improvement support device 200 at a predetermined time in a day. The predetermined timing may also be any timing in a week.

[0016] In addition, the photographing device 100 may simply supply live view images to the production line work improvement support device 200, and the production line work improvement support device 200 may perform processing such as detecting the external shape at the time the work starts and the external shape at the time the work is completed in the live view images, recording the time the work starts and the time the work is completed, recording images from the time the work starts to the time the work is completed, and adding an identification number.

[0017] The manufacturing line work improvement support device 200 includes a processing unit 210 that executes processing to acquire or generate various data based on the video captured by the imaging device 100, a memory unit 220 that stores the data acquired or generated by the processing unit 210, and a display unit 230 that displays the information acquired or generated by the processing unit 210.

[0018] The arithmetic processing unit 210 is an arbitrary arithmetic processing device, and includes, for example, a CPU (Central Processing Unit). The arithmetic processing unit 210 executes a program stored in the storage unit 220 to function as an actual work video acquisition unit 211 that saves video files in the storage unit 220, a calculation unit 212 that calculates a process evaluation value for each process indicating whether the process is a bottleneck or not, i.e., the degree to which the process is an improvement target process that should be improved, an improvement target process identification unit 213 that identifies the improvement target processes, and an output unit 214 that outputs the improvement target processes and improvement implementation examples. In this embodiment, the process evaluation value is a value obtained by multiplying the average work time for each process during the analysis period by the standard deviation of the process work times.

[0019] The storage unit 220 includes any storage device, such as a nonvolatile semiconductor memory including a flash memory or an EPROM (Erasable Programmable Read Only Memory). The storage unit 220 may also include a nonvolatile memory including a magnetic disk, a flexible disk, an optical disk, a compact disk, a minidisk, or a DVD (Digital Versatile Disc).

[0020] The storage unit 220 has an actual work file storage unit 221 that stores actual work video files, a calculated value storage unit 222 that stores calculated values ​​calculated by the calculation unit 212, an improvement target process storage unit 223 that stores improvement target processes, and an improvement activity data storage unit 224 that stores improvement performance data for past improvement measures or improvement activity data indicating expected improvement activities. The storage unit 220 also stores a program to be executed by the calculation processing unit 210. The improvement measures and improvement activities are examples of improvements related to the present disclosure.

[0021] The various data in the storage unit 220 of the production line work improvement support device 200 may be stored in a cloud-type server external to the production line work improvement support device 200. The files stored in the actual work file storage unit 221 may be compressed before storage. The storage capacity of the actual work file storage unit 221 may be managed by automatically deleting old files after a period determined at each work site.

[0022] The actual work video acquisition unit 211 of the calculation processing unit 210 acquires an actual work video file capturing the work of each product from start to finish from the image capturing device 100. The product is, for example, a refrigerator. Note that the product is not limited to a refrigerator and may be any other type of electrical appliance. Note that the actual work video file is an example of captured data according to the present disclosure.

[0023] The calculation unit 212 calculates the process work time per product based on video data acquired from the actual work file storage unit 221 of the storage unit 220. The calculation method is to calculate the time from the start to completion of the process work for each product as one cycle of process work time. The process time is calculated from the difference between the time when the process work is completed and the time when the process work is started. Based on the process time, the calculation unit 212 calculates a process evaluation value that indicates whether the process is a bottleneck.

[0024] The improvement target process identifying unit 213 identifies, from among the series of processes, an improvement target process that is desirable to be improved with priority over other processes, based on the process evaluation value calculated by the calculating unit 212.

[0025] The output unit 214 outputs the identified improvement target process and the improvement performance data corresponding to the improvement process to the display unit 230.

[0026] The display unit 230 displays the information output by the arithmetic processing unit 210. The display unit 230 includes, for example, a liquid crystal display.

[0027] Next, the process for identifying a process to be improved will be described with reference to Fig. 2. For example, when a person in charge of production line work starts up the production line work improvement support device 200 at a predetermined time in a day, the process for identifying a process to be improved shown in Fig. 2 is executed.

[0028] When the process starts, the actual work video acquisition unit 211 of the production line work improvement support device 200 acquires an actual work video file from the camera device 100 and additionally stores it in the actual work file storage unit 221 (step S201). As a result, the actual work video, the identification number of the camera device 100, and the start and completion times of the work are additionally stored in the actual work file storage unit 221. If the current time is within the analysis period, the process ends here.

[0029] When the analysis period ends and the actual work file storage unit 221 stores video files for the total number of units produced during the analysis period, the calculation unit 212 calculates the process work time for the total number of units produced during the analysis period and calculates the average process work time for each process (step S202). Specifically, the calculation unit 212 calculates the process work time for each task based on the actual work file storage unit 221, by calculating the difference between the completion time and start time of the task. The calculation unit 212 stores the calculated process work time together with the identification number of the image capture device 100 installed for each process in the calculated value storage unit 222. The calculation unit 212 adds up the process work times associated with the same identification number stored in the calculated value storage unit 222. The calculation unit 212 calculates the average process work time by dividing the added process work time by the total number of units produced during the analysis period.

[0030] The calculation unit 212 calculates the standard deviation of the process work time for each process (step S203). The average process work time and the standard deviation of the process work time are examples of characteristic values ​​that indicate the characteristics of the work time of a process.

[0031] The calculation unit 212 calculates the multiplication value of the average process work time and the standard deviation (step S204). The multiplication value calculated by the calculation unit 212 is stored in the calculated value storage unit 222. This multiplication value is an example of a process evaluation value, and is an index value indicating whether or not a process corresponds to a bottleneck, that is, a process that should be improved with priority.

[0032] The improvement target process identification unit 213 identifies the process with the maximum multiplication value calculated for each process in step S204 as the improvement target process (step S205). Figure 3 shows example values ​​and a graph showing the average process work time and standard deviation for each process. The average process work time and standard deviation for each process are shown for a production volume of 6,000 units from January 1 to March 31, 2022. In Figure 3, process 3, which has the largest multiplication value, corresponds to the improvement target process. Here, a large average process work time indicates that the process work time is long on average and there is a possibility of a problem with the process work. Furthermore, a large standard deviation indicates that the process work time varies greatly and there is a possibility of disruption in the process work. A large multiplication value indicates that one or both of these are large and there is a high possibility of improvement. The identified improvement target processes are recorded in the improvement target process storage unit 223.

[0033] The output unit 214 outputs the improvement target process and past improvement execution examples for the process identified as the improvement target process to the display unit 230 (step S206). The improvement execution examples are stored in the improvement activity data storage unit 224. FIG. 4 is a diagram showing the improvement target process and past improvement execution examples. A past improvement execution example corresponding to process 3 identified as the improvement target process is shown. FIG. 4 shows four improvement perspectives: people, things, equipment, and methods, but the present disclosure is not limited to these. Additional improvement perspectives may be added, or only one may be shown. Also, only the improvement execution example may be shown without indicating the improvement perspective. For example, an improvement execution example may be, "When fastening part A, using the middle finger of the left hand shortens the work time."

[0034] Next, an example of the hardware configuration of the production line operation improvement support device 200 will be described with reference to Fig. 5. The production line operation improvement support device 200 in Fig. 5 is realized by a computer such as a personal computer or a microcontroller.

[0035] The production line work improvement support device 200 comprises a processor 1001 that executes an operating program which is a program for the operation of the production line work improvement support device 200, a memory 1002 that serves as the main storage area, an interface 1003 that realizes the communication function of the production line work improvement support device 200, and a secondary storage device 1004 that stores the operating program for executing processing, all of which are connected to each other via a bus 1000.

[0036] The processor 1001 is, for example, a CPU (Central Processing Unit). The processor 1001 reads an operation program stored in a secondary storage device 1004 into a memory 1002 and executes the program, thereby realizing each function of the production line operation improvement support device 200.

[0037] The memory 1002 is a main storage device configured, for example, by a RAM (Random Access Memory). The memory 1002 stores the operating program that the processor 1001 reads from the secondary storage device 1004. The memory 1002 also functions as a work memory when the processor 1001 executes the operating program.

[0038] The interface 1003 is an I / O (Input / Output) interface such as a serial port, a USB (Universal Serial Bus) port, a network interface, etc. The interface 1003 realizes the communication function of the production line operation improvement support device 200.

[0039] The secondary storage device 1004 is, for example, a flash memory, a hard disk drive (HDD), or a solid state drive (SSD). The secondary storage device 1004 stores the operating programs executed by the processor 1001.

[0040] As described above, in the production line work improvement support system 1 according to the first embodiment, the calculation unit 212 calculates the multiplication value of the average process work time and the standard deviation based on the actual work video file, and the improvement target process identification unit 213 identifies the process that shows the maximum value of the multiplication value for each process as the improvement target process. The output unit 214 outputs the improvement target process and past improvement execution examples to the display unit 230. By calculating the process work time based on the video data from the actual site, it becomes possible to identify the improvement target process based on the process work time at the actual site. This makes it possible to more appropriately identify the improvement target process among the processes on the production line that most needs to be improved.

[0041] Furthermore, due to time constraints, some factories can only calculate the work times for some of the work processes. However, analysis results based on the work times for some of the work processes are unreliable, and the identification of improvement target processes based on unreliable analysis results may not reflect the actual situation. Identification of improvement target processes that does not reflect the actual situation makes it difficult to implement efficient improvement activities. However, the production line work improvement support device 200 collects all work times for all work processes and performs the work process of identifying improvement target processes. Therefore, it is possible to present improvement target processes that reflect the actual situation to the user.

[0042] (Modification of the first embodiment) Note that determining when a task is started and when it is completed by detecting the external shape in an image is merely one example, and the start and completion of a task may be determined by any other method. For example, a task may be sent to a worker from a previous process, and a sensor may be attached to the location where the task is picked up, and the task may be determined to have started when the sensor output detects that the task has been picked up. Also, a sensor may be attached to the location where the completed task is sent to a subsequent process, and the task may be determined to have been completed when the sensor output detects that the completed task has been sent.

[0043] The process work time may be the recording time of the operation in the video file, or may be calculated from the captured video footage using any analysis software.

[0044] Furthermore, the improvement target process storage unit 223 and the improvement activity data storage unit 224 may be stored in a device that can be shared among multiple departments. For example, by devising a way to hold a work item in each process in a factory, it may be possible to prevent collisions between the work item and factory equipment such as conveyors or workbenches, thereby reducing the occurrence of inefficient work. Therefore, by storing the improvement target process and improvement activity data in a form that can be shared among multiple departments, effective improvement target processes and improvement activity results can be spread throughout the factory, and improvement activities can be carried out efficiently.

[0045] Furthermore, data is output as information indicating the process to be improved and the improvement activity implementation examples, but this is not limited to data and may be audio data explaining the process to be improved and the improvement activity implementation examples.

[0046] Although the product is an electrical appliance in the example, the product is not limited to an electrical appliance and may be any type of product that can be a target for improvement of process operations on a manufacturing line.

[0047] Although the process evaluation value is a value obtained by multiplying the average task time for each process during the analysis period by the standard deviation of the process task times, the present disclosure is not limited to this. The process evaluation value may also be a value obtained by multiplying the average task time for each process during the analysis period by the variance of the process task times. The standard deviation and variance are examples of indicators that indicate the degree of dispersion of the process task times.

[0048] (Embodiment 2) 6 to 9, a manufacturing line work improvement support device 200 according to a second embodiment will be described. The second embodiment differs from the first embodiment in that the manufacturing line work improvement support device 200 presents processes to be improved using AI (Artificial Intelligence) based on the average process work time, standard deviation, and a trained model. The following describes the second embodiment, focusing on the differences from the first embodiment.

[0049] 6, the production line operation improvement support device 200 according to this embodiment includes an improvement target process presenting unit 215 instead of the improvement target process identifying unit 213 of the production line operation improvement support device 200 according to embodiment 1. The improvement target process presenting unit 215 presents improvement target processes based on the average process task time, the standard deviation, and an implemented model.

[0050] The storage unit 220 further includes a trained model storage unit 225.

[0051] The trained model storage unit 225 stores trained models. The trained models are generated using a known machine learning library, for example, by a "supervised learning method" in which a set of data in which training data is labeled with a correct answer is trained. Note that the trained models may also be generated by learning using other learning methods, such as an "unsupervised learning method," a "reinforcement learning method," or a "semi-supervised learning method."

[0052] The learning algorithm can be a known learning algorithm such as a neural network. The neural network has an input layer composed of a plurality of nodes to which different input parameters are input, an intermediate layer to which signals output from each node of the input layer are input, and an output layer to which signals output from the intermediate layer are input and which outputs output parameters. Note that the neural network may be composed of one or more intermediate layers.

[0053] An overview of a neural network will now be described with reference to FIG. 7. A neural network is composed of an input layer, a hidden layer, and an output layer, each containing a plurality of neurons. In the second embodiment, the hidden layer is assumed to be one layer. For example, in the three-layer neural network shown in FIG. 7, when a plurality of inputs are input to neurons X1 to Xn in the input layer, the values ​​are multiplied by weights W11 to Wnm and input to neurons Y1 to Ym in the hidden layer. The results input to neurons Y1 to Ym in the hidden layer are further multiplied by weights V11 to Vkm and output from neurons Z1 to Zk in the output layer. The output results from neurons Z1 to Zk in the output layer vary depending on the values ​​of weights W11 to Wnm and V11 to Vkm. Here, n is 12, k is 6, and m is an integer equal to or greater than 3. For example, neuron X1 in the input layer receives the average process task time of process 1, neuron X2 receives the standard deviation of process 1, neuron X3 receives the average process task time of process 2, neuron X4 receives the standard deviation of process 2, etc. Neurons Z1 to Z6 in the output layer are assigned to processes 1 to 6, respectively.

[0054] The neurons X1 to X12 in the input layer are supplied with the average process work time, which is the longest of the six process average work times normalized to a value of 1, and the standard deviation, which is the largest of the six process standard deviations normalized to a value of 1.

[0055] The learning data is created based on a combination of the average process work time and standard deviation for each of the six processes, which are obtained in advance by an expert through analysis for each product, and the process to be improved. A learning process for the process to be improved is performed based on the learning data. The structure of the neural network, including the weights W11 to Wnm, V11 to Vkm, values ​​n, k, etc. obtained through the learning process, is stored as a learned model in the learned model storage unit 225.

[0056] The improvement target process presentation unit 215 presents the improvement target process to the worker using the trained model.

[0057] As shown in FIG. 6, the improvement target process presenting unit 215 includes a learning unit 215a that generates a trained model, and an inference unit 215b that infers the improvement target process using the trained model.

[0058] The learning process performed by the learning unit 215a will be described below. The learning process shown in Fig. 8 starts when an operation unit (not shown) receives an operation from a user instructing execution of the learning process.

[0059] As described above, the learning unit 215a receives training data including multiple combinations of the average process work times and standard deviations of six pre-designed processes and the process to be improved (step S301). The training data is input to the learning unit 215a via an input unit (not shown).

[0060] The learning unit 215a repeatedly learns using the training data (step S302) and adjusts the weights W11 to Wnm and V11 to Vkm. The learning unit 215a stores the learned model in the learned model storage unit 225 (step S303) and ends the learning process.

[0061] The inference unit 215b uses the trained model to execute the inference process shown in Fig. 9. The inference process is executed at a timing preset by the user. The timing for executing the inference process may be set, for example, on a weekly or monthly basis.

[0062] The inference unit 215b applies normalized values ​​of the average process work times and standard deviations of the six processes as input parameters to the trained model (step S401).

[0063] The inference unit 215b uses the trained model to perform inference based on the input parameters (step S402), outputs data indicating the process to be improved as output parameters (step S403), and ends the inference process.

[0064] The improvement target process presenting unit 215 acquires the output parameters output from the inference unit 215b as improvement target process data. Upon acquiring the improvement target process data, the improvement target process presenting unit 215 stores the improvement target process data in the improvement target process storage unit 223 shown in FIG.

[0065] When the improvement target process presentation unit 215 acquires the improvement target process data, it presents the improvement target work process to the worker based on the acquired improvement target process data. Specifically, the improvement target process presentation unit 215 outputs the improvement target process data to the display unit 230, and the improvement target process data is displayed on the display unit 230, thereby presenting the improvement target process data to the worker.

[0066] The improvement target process presentation unit 215 may, for example, use an application such as a spreadsheet software based on the improvement target process data to display the improvement target process on the display unit 230. Furthermore, the output unit 214 may display the improvement execution cases stored in the improvement activity data storage unit 224 together with the improvement target process data on the display unit 230. The improvement execution cases can be used as training materials for new employees.

[0067] As described above, according to the manufacturing line work improvement support device 200 of the second embodiment, the improvement target process presentation unit 215 presents the improvement target process to the worker using a trained model trained with data including the average process work time, standard deviation, and the improvement target process. The worker can carry out efficient improvement activities by implementing improvements to the improvement target process presented by the improvement target process presentation unit 215. As a result, the occurrence of work processes with low work efficiency is reduced, and the burden of manufacturing line work improvement support is lightened. Therefore, the efficiency of manufacturing line work improvement support can be improved.

[0068] In the second embodiment, the learning target of the trained model used by the improvement target process presentation unit 215 may be multiple types of products. Furthermore, the learning unit 215a may learn about actual work performed in one area, or may learn about actual work performed independently in multiple different areas.

[0069] Furthermore, a product to be trained for a trained model may be added or removed from the training target during the training process. Furthermore, any trained model may be applied to a trained model for another model, or the other model may be retrained based on the trained model. Furthermore, the training data may, for example, be a combination of the average process work time and the standard deviation as input, and the process for which the product of these is the largest as output.

[0070] (Modification of the second embodiment) In the second embodiment, the learning unit 215a uses a neural network as a learning algorithm, but the learning unit 215a may use other known methods, such as genetic programming, functional logic programming, or support vector machines, as a learning algorithm.

[0071] In the second embodiment, the learning unit 215a and the inference unit 215b are provided in the production line operation improvement support device 200, but the learning unit 215a and the inference unit 215b may be provided in a device different from the production line operation improvement support device 200. The improvement-target process presenting unit 215 may acquire, as improvement-target process data, output parameters output from the inference unit 215b provided in a device different from the production line operation improvement support device 200.

[0072] The trained model resulting from the learning process performed elsewhere may be stored in the trained model storage unit 225, and the learning unit 215a may not be included in the improvement target process presentation unit 215.

[0073] Furthermore, the improvement-target process presenting unit 215 may use data that combines the average process work time and standard deviation of not only the same model but also similar models with the improvement-target process. That is, the improvement-target process presenting unit 215 can also use data that combines the average process work time and standard deviation of similar models with the improvement-target process as learning data.

[0074] In the second embodiment, the process to be improved is presented based on the average process work time and standard deviation. However, the present invention is not limited to this, and the most effective improvement viewpoint may be presented based on the average process work time and standard deviation. The process to be improved and the most effective improvement viewpoint may be presented together.

[0075] In this case, for example, the combination of the average process work time, standard deviation, the process to be improved, and the most effective improvement viewpoint is prepared as learning data from an analysis of past performance data. Also, neurons assigned to each improvement viewpoint are placed in the output stage of the neural network shown in Figure 7.

[0076] In the inference stage, a set of average process work time and standard deviation is input to the input stage, and the process to be improved is identified from the firing state of the neuron to which the process is assigned in the output stage, and the most effective improvement perspective is identified from the firing state of the neuron to which the improvement perspective is assigned.

[0077] In the above embodiment, an example was shown in which the manufacturing line operation improvement support device 200 is configured by a computer having a processor and a memory, but the device configuration is arbitrary. For example, all or part of the processing executed by the processor may be executed by an ASIC (Application Specific Integrated Circuit), FPGA (Field-Programmable Gate Array), etc.

[0078] Although the preferred embodiments have been described in detail above, the present invention is not limited to the above-described embodiments, and various modifications and substitutions can be made to the above-described embodiments without departing from the scope of the claims.

[0079] Various aspects of the present disclosure are summarized below as appendices.

[0080] (Appendix 1) A production line work improvement support device that supports improvement of work on a production line that manufactures products, a calculation unit that calculates a process operation time based on photographic data of work performed in each process of the production line, and calculates a plurality of characteristic values ​​that indicate characteristics of the operation time for each process based on the calculated process operation time; an improvement target process identifying unit that identifies a process that needs to be improved with priority over other processes as an improvement target process using the obtained plurality of characteristic values; A manufacturing line work improvement support device comprising: (Appendix 2) an output unit that outputs the improvement target process and improvement activity data that presents improvements to the process identified as the improvement target process; 2. A manufacturing line work improvement support device according to claim 1. (Appendix 3) the plurality of characteristic values ​​include an average process work time for each process and a dispersion indicating a variation in the process work time, the calculation unit calculates a value obtained by multiplying the average process work time by the dispersion degree as a process evaluation value for evaluating the work time for each process; 3. The manufacturing line work improvement support device according to claim 1, wherein the improvement target process identifying unit identifies, as the improvement target process, a process for which a maximum process evaluation value is obtained from among the calculated process evaluation values. (Appendix 4) The dispersion includes one of a variance and a standard deviation. 4. A manufacturing line work improvement support device according to claim 3. (Appendix 5) the improvement target process identifying unit presents the improvement target process based on the average process work time and the degree of dispersion using a combination of the average process work time and the degree of dispersion and a trained model that has been trained for the improvement target process. (Appendix 6) The improvement target process specifying unit a learning unit that generates a trained model based on training data including the combination and the improvement target process; an inference unit that infers the process to be improved using the trained model based on the combination; having 6. A manufacturing line work improvement support device according to claim 5. (Appendix 7) the improvement target process identifying unit uses a trained model that has learned about a combination of the average process work time and the dispersion and about the improvement target process and perspectives for improvement to present the improvement target process and perspectives for improvement that are considered to be effective for improving the improvement target process, based on the average process work time and the dispersion. (Appendix 8) The work time for each process on the line is calculated, and a characteristic value indicating the characteristics of the calculated work time is calculated. Identifying a process that needs to be improved with priority over other processes based on the characteristic values; A method for supporting improvements in manufacturing line work. (Appendix 9) On the computer, The work time for each process on the line is calculated, and a characteristic value indicating the characteristics of the calculated work time is calculated. Identifying a process that needs to be improved with priority over other processes based on the characteristic values; A program that executes a process. [Explanation of symbols]

[0081] 1 Manufacturing line work improvement support system, 100 imaging device, 200 manufacturing line work improvement support device, 210 calculation processing unit, 211 actual work video acquisition unit, 212 calculation unit, 213 improvement target process identification unit, 214 output unit, 215 improvement target process presentation unit, 215a learning unit, 215b inference unit, 220 memory unit, 221 actual work file memory unit, 222 calculated value memory unit, 223 improvement target process memory unit, 224 improvement activity data memory unit, 225 learned model memory unit, 230 display unit, 1000 bus, 1001 processor, 1002 memory, 1003 interface, 1004 secondary storage device.

Claims

1. A production line work improvement support device that supports improvement of work on a production line that manufactures products, a calculation unit that calculates a process operation time based on photographic data of work performed in each process of the production line, and calculates a plurality of characteristic values ​​that indicate characteristics of the operation time for each process based on the calculated process operation time; an improvement target process identifying unit that identifies a process that needs to be improved with priority over other processes as an improvement target process using the obtained plurality of characteristic values; Equipped with the plurality of characteristic values ​​include an average process work time for each process and a dispersion indicating a variation in the process work time, the improvement target process identification unit inputs the combination of the average process work time and the dispersion calculated by the calculation unit into a trained model obtained by learning a combination of the average process work time and the dispersion and a process that needs to be improved with priority when the combination is obtained, and presents the output improvement target process. A manufacturing line work improvement support device.

2. The dispersion includes one of a variance and a standard deviation. The manufacturing line operation improvement support device according to claim 1.

3. The improvement target process specifying unit a learning unit that generates the trained model based on training data including the combination and the improvement target process; an inference unit that applies the combination to the trained model to infer the process to be improved; having 3. The manufacturing line operation improvement support device according to claim 1.

4. The trained model is obtained by learning a combination of an average process work time and a dispersion degree, a process that needs to be improved as a priority when that combination is obtained, and an improvement perspective that is considered to be effective for improving the process, the improvement target process identification unit inputs a combination of the average process work time and the dispersion calculated by the calculation unit into the trained model, and presents the output improvement target process and an improvement perspective that is considered to be effective for improving the improvement target process.

3. The manufacturing line operation improvement support device according to claim 1.

5. A manufacturing line work improvement support method executed by a computer, comprising: A process operation time is calculated based on photographic data obtained by photographing work for each process on a production line for manufacturing a product, and a plurality of characteristic values ​​indicating characteristics of the operation time for each process are calculated based on the calculated process operation time; Using the obtained multiple characteristic values, a process that needs to be improved with priority over other processes is identified as a process to be improved; the plurality of characteristic values ​​include an average process work time for each process and a dispersion indicating a variation in the process work time, The combination of the average process work time and the dispersion is input into a trained model obtained by learning a combination of the average process work time and the dispersion and a process that needs to be improved with priority when the combination is obtained, and the output process to be improved is presented. A method for supporting improvements in manufacturing line work.

6. On the computer, A process operation time is calculated based on photographic data obtained by photographing work for each process on a production line for manufacturing a product, and a plurality of characteristic values ​​indicating characteristics of the operation time for each process are calculated based on the calculated process operation time; Using the obtained multiple characteristic values, a process that needs to be improved with priority over other processes is identified as a process to be improved; the plurality of characteristic values ​​include an average process work time for each process and a dispersion indicating a variation in the process work time, The combination of the average process work time and the dispersion is input into a trained model obtained by learning a combination of the average process work time and the dispersion and a process that needs to be improved with priority when the combination is obtained, and the output process to be improved is presented. A program that executes a process.

Images