Production management device, production management system, production management method, and storage medium

Abstract

The present application provides a planned production management device, production management system, production management method and storage medium that can be produced with less deviation from actual production. The production management device of the embodiment acquires a first short-term plan that is produced based on a first long-term plan indicating a plan for production in a prescribed period and indicates a plan for production in a first period that is shorter than the prescribed period, acquires first actual performance data indicating actual performance in work performed along a part of the first short-term plan, acquires first prediction data indicating a prediction of actual performance in work performed along another part of the first short-term plan using the first actual performance data. The production management device corrects the first short-term plan based on the first prediction data, acquires second prediction data indicating a prediction of actual performance in production in the prescribed period using second actual performance data indicating actual performance in work in the first period, and produces a second long-term plan for production in the prescribed period using the second prediction data.

Description

Technical Field

[0001] The embodiments of the present invention relate to a production management device, a production management system, a production management method, and a storage medium. Background Technology

[0002] Previously, production schedulers were used to create production plans. The requirement now is for technology capable of creating plans that deviate less from actual production.

[0003] [Existing Technical Documents]

[0004] [Patent Literature]

[0005] [Patent Document 1] Japanese Patent Application Publication No. 6-231135 Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a production management device, a production management system, a production management method, and a storage medium capable of producing a production plan that deviates less from actual production.

[0007] The production management device of the implementation method acquires a first short-term plan, which is created based on a first long-term plan representing a production plan for a specified period, and represents a production plan for a first period shorter than the specified period. The production management device also acquires first performance data, which represents the performance of operations performed along a portion of the first short-term plan. The production management device further uses the first performance data to acquire first forecast data, which represents a forecast of the performance of the operations performed along another portion of the first short-term plan. The production management device also modifies the first short-term plan based on the first forecast data. The production management device also uses second performance data representing the performance of the operations in the first period to acquire second forecast data representing a forecast of the production performance for the specified period. The production management device further uses the second forecast data to create a second long-term plan for the specified period. Attached Figure Description

[0008] Figure 1 This is a schematic diagram illustrating the structure of the production management system implemented in this way.

[0009] Figure 2 (a) and Figure 2 (b) is a table that illustrates the production master data.

[0010] Figure 3 (a)~ Figure 3 (d) is a table that represents the master data of the production process.

[0011] Figure 4(a) is an example of a long-term plan. Figure 4 (b) is an example of a short-term plan.

[0012] Figure 5 (a) is a schematic diagram showing the situation of the operation. Figure 5 (b) is with Figure 5 The image corresponding to (a). Figure 5 (c) is a schematic diagram representing an example of the recognition result.

[0013] Figure 6 (a) is a schematic diagram representing the time-series detection data obtained by the detector. Figure 6 (b) is a schematic diagram representing template data.

[0014] Figure 7 (a) is a diagram illustrating the actual performance of a portion of a short-term plan. Figure 7 (b) is a schematic diagram representing the forecast of the other part of the short-term plan.

[0015] Figure 8 (a) is a table representing a short-term plan made in advance. Figure 8 (b) is a table representing the revised short-term plan.

[0016] Figure 9 (a) is a diagram illustrating the actual performance of a portion of the revised short-term plan. Figure 9 (b) is a schematic diagram representing the forecasts for the other part of the revised short-term plan.

[0017] Figure 10 (a) is a table representing the revised short-term plan. Figure 10 (b) is a table representing the further revised short-term plan.

[0018] Figure 11 (a) is a diagram representing a portion of the performance during the specified period. Figure 11 (b) is a schematic diagram representing the forecast for the other part of the specified period.

[0019] Figure 12 (a) is a table that represents the updated manufacturing master data. Figure 12 (b) is a table that illustrates the updated production master data.

[0020] Figure 13 This is a flowchart illustrating the production management method implemented.

[0021] Figure 14 This is a flowchart illustrating the specific processing of the production management device in the implementation method.

[0022] Figure 15This is a flowchart illustrating the specific processing of the production management device in the implementation method.

[0023] Figure 16 It is a schematic diagram representing the hardware structure. Detailed Implementation

[0024] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In this specification and the drawings, elements that are the same as those already described are labeled with the same reference numerals, and detailed descriptions are omitted where appropriate.

[0025] The production management system implemented in this manner is used to manage or support product production. Based on data obtained during the execution of production-related operations, the production management system modifies or creates production plans. These plans include short-term and long-term plans.

[0026] Figure 1 This is a schematic diagram illustrating the structure of the production management system implemented in this way.

[0027] like Figure 1 As shown, the production management system 1 of the embodiment includes a production management device 10, a storage device 20, a camera device 30, and a detector 40.

[0028] The camera device 30 is installed at the production site to capture images of workers or items being handled during operations. For example, the camera device 30 repeatedly acquires still images and stores this image data in the storage device 20. The camera device 30 can also acquire moving images. In this case, still images are repeatedly captured from the moving images.

[0029] Detector 40 detects signals generated by the operator's actions. For example, detector 40 includes one or more sensors selected from torque sensors, acceleration sensors, and angular velocity sensors, and is installed on tools such as digital torque wrenches or digital vernier calipers. Detector 40 detects signals when the operator uses the tool. Detector 40 stores the obtained time-series detection data in storage device 20.

[0030] Detector 40 can also be an accelerometer or angular velocity sensor worn on the worker's hand or foot. Detector 40 detects the acceleration or angular velocity of a part of the worker's body. Multiple detectors 40 can also be worn on the worker's body. Detector 40 stores the obtained time-series detection data in storage device 20.

[0031] In addition to storing image data and detection data, storage device 20 also stores past production performance data, production master data related to various production elements, and manufacturing master data related to various manufacturing elements.

[0032] Figure 2(a) and Figure 2 (b) is a table that illustrates the production master data.

[0033] Figure 2 (a) and Figure 2 The production master data 100A and 100B shown in (b) include a process column 101, a job column 102, a lead time column 103, and an available equipment column 104. The product is manufactured through multiple processes. The process column 101 records strings (names or IDs) used to identify each process. Each process includes more than one job. The job column 102 records strings used to identify each job. The lead time column 103 records the time required from the start to the end of a production-related job. Hereinafter, lead time will be abbreviated as "LT". In the illustrated example, in each job, the "number of devices that can be assembled within 1 hour" is recorded as LT. The available equipment column 104 records the equipment that can be used in the job.

[0034] Figure 3 (a)~ Figure 3 (d) is a table that represents the master data of the production process.

[0035] Manufacturing master data includes individual master data for each operator and equipment operating master data. Figure 3 (a)~ Figure 3 (c) is an example of individual master data 110A-110C related to operators X-Z. Individual master data 110A-110C includes operator column 111, job column 112, skill level column 113, preparation time column 114, and yield column 115. Operator column 111 records the string used to identify each operator. Job column 112 records the string used to identify each job. Skill level column 113 records the skill level (proficiency) associated with each job of the operator. Preparation time column 114 records the time taken (LT) for each job of the operator. Yield column 115 records the yield of each job of the operator.

[0036] The time-to-work (LT) recorded in the production master data is related to the LT recorded in the manufacturing master data. In the manufacturing master data, the LT for each job is recorded for each operator. The production master data records the average LT for each job with one or more operators. In the manufacturing master data, if an operator's LT improves, the LT in the production master data may also improve.

[0037] Figure 3 (d) is an example of equipment operation master data 120. Equipment operation master data 120 includes equipment column 121 and stop rate column 122.

[0038] Register the string used to identify each piece of equipment in Equipment column 121. Register the temporary stop rate for each piece of equipment in Stop Rate column 122. "Temporary stop" refers to a work stoppage of a few minutes to tens of minutes. Temporary stoppages are caused by minor equipment malfunctions, failure of raw materials to arrive at the equipment, etc.

[0039] Production management device 10 uses various data to create long-term plans. Long-term plans, compared to short-term plans (described later), represent production plans over a longer period. Long-term can be, for example, several weeks to several months. Long-term plans can also be referred to as "production schedules" or "large-scale schedules," etc. As an example, a long-term plan specifies the quantity of each product type to be produced within a period of 2 to 3 months, ranging from several days to one week, as well as the timeline from the start to the completion of production for each product.

[0040] Production management unit 10 creates short-term plans based on long-term plans. Short-term plans, compared to the aforementioned long-term plans, represent production plans for a shorter period. A short period is, for example, one day. Short-term plans can also be called "input plans" or "mini-schedule plans," etc. As an example, a short-term plan specifies the types and quantities of products to be produced within one day, the tasks to be performed, the start time of each task, and the operators performing each task. The following is an example of a production plan for a one-day period specified through a short-term plan.

[0041] The production management device 10 uses a generally available scheduler to create long-term and short-term plans. Examples of available schedulers include FLEXSCHE (registered trademark), Asprova (registered trademark), and JoyScheduler (registered trademark). In creating long-term and short-term plans, in addition to the aforementioned production master data and manufacturing master data, process master data, resource master data, product structure master data, and work schedule master data are also referenced.

[0042] The process master data records the name of each process, the sequence between processes, and the standard time limit (LT) for each process. The LT of the process master data is shared with the production master data. The resource master data records the name of the equipment used in the process, the number of each piece of equipment, the name of the operator using each piece of equipment, and the number of operators. The product structure master data records the raw materials used, processed products during production, assembled products during production, and finished products, defining the process flow until the product is completed. The work schedule master data records the working hours of each operator, the overtime hours of each operator, and the workdays of each operator. The master data used in the planning process is pre-stored in storage device 20.

[0043] Figure 4 (a) is an example of a long-term plan. Figure 4 (b) is an example of a short-term plan.

[0044] Figure 4 (a) represents the result of extracting only the plan to start operations on "February 1st" from the long-term production plan of machine type A over several months. Figure 4 As shown in long-term plan 130 (a), it is planned to receive the components related to the device in the warehouse on February 1 and ship 12 devices before February 5. Figure 4 (b) indicates based on Figure 4 It is part of a short-term plan created in accordance with (a) of the plan. Figure 4 In (b), a short-term plan for "February 2nd" is shown. This short-term plan 140 plans the number of parts or semi-finished products to be delivered to the work site, the time of each operation, the operators performing each operation, and the equipment used. Furthermore, regarding the illustrated operations, "delivery" refers to transporting parts to the work area. "Delivery" refers to using parts to produce products.

[0045] The daily work is carried out according to a pre-made short-term plan. After the work begins, the camera device 30 acquires image data, and the detector 40 acquires detection data. Alternatively, data can be continuously acquired by the camera device 30 and the detector 40 from before the work begins. The production management device 10 acquires both image data and detection data. The production management device 10 uses one or both of the image data and detection data to calculate performance data representing the actual performance of the performed work. The calculated performance data includes the work's time limit (LT). The specific method for calculating actual data based on image data or detection data will be described below.

[0046] The production management device 10 identifies items that will be the objects of the operation from image data. Through identification, it determines the type of item, the quantity of each type, and so on. Items are the objects of the operation and can be components, semi-finished products, finished products, etc. For example, individual components and combinations thereof are identified as different types of items. When a component is installed on a semi-finished product, the semi-finished product with the component installed is identified as a different type of item from the semi-finished product without the component installed.

[0047] In object recognition, a recognition model is used to identify objects reflected in an image. The recognition model preferably includes a neural network. To improve recognition accuracy, the recognition model more preferably includes a convolutional neural network (CNN). The recognition model is pre-learned to output the number of objects reflected in the image for each category based on the input image data. Image data and teaching data are used in the learning process. The teaching data teaches the location of the objects reflected in the image, the category of the objects, and the number of objects in each category.

[0048] The production management device 10 inputs image data into the recognition model to obtain the type and quantity of items reflected in the image. The production management device 10 then associates the recognition results with the image capture time and stores them in the storage device 20. Changes in the type and quantity of items correspond to the progress of the operation. For example, based on changes in the quantity of each type of item, the production management device 10 estimates the time to completion (LT) of a single operation, the progress of the operation, and the number of previously performed operations as performance data.

[0049] Figure 5 (a) is a schematic diagram showing the situation of the operation. Figure 5 (b) is with Figure 5 The image corresponding to (a). Figure 5 (c) is a schematic diagram representing an example of the recognition result.

[0050] exist Figure 5 In the example shown in (a), the operator W performs the task of assembling part 151 onto the semi-finished product 152. Figure 5 (b) indicates that the camera device 30 is a camera. Figure 5 Image data 160 is obtained from the operation shown in (a). When the production management device 10 acquires image data 160, it inputs the image data 160 into the recognition model. For example, as shown in (a)... Figure 5 As shown in (c), the recognition model outputs recognition result 161 corresponding to component 151 and recognition result 162 corresponding to semi-finished product 152. The production management device 10 identifies the items reflected in the image based on the output results from the recognition model.

[0051] Regarding the operation of assembling component 151 onto semi-finished product 152, if the quantity of component 151 decreases and the quantity of finished products in semi-finished product 152 containing component 151 increases, it can be inferred that the operation is completed. In the case of repeating the same operation, the increased quantity of finished products corresponds to the number of times the operation is performed. The rate of increase in the quantity of finished products corresponds to LT (Time Limit). The difference between the decreased quantity of component 151 and the increased quantity of finished products corresponds to the number of defective products. The production management device 10 stores these performance data inferred based on the identification results in the storage device 20.

[0052] In addition to image data, the production management device 10 uses template matching of detection data to infer the work performed by the operator, the time of work (LT), etc. For example, the detector 40 includes at least one selected from torque sensors, acceleration sensors, and angular velocity sensors, assembled on a tool (wrench or vernier caliper). When the tool is used during the work, the detector 40 detects a signal different from when the tool is not used. The detector 40 continuously detects the signal. Thus, time-series detection data is obtained.

[0053] The production management device 10 refers to pre-set work standards. These standards record the tasks to be performed, their sequence, and the standard time for each task. The production management device 10 extracts a portion of the time-series detection data. The length of the extracted data can be fixed or set based on the standard time. Here, the portion of data extracted from the entire time-series data is referred to as "partial data."

[0054] The production management device 10 compares partial data with pre-prepared template data. Template data is prepared for each job registered in the work standards. The more similar the waveform of the partial data is to the waveform of the template data, the higher the probability that the operator will execute the job corresponding to that template data. The production management device 10 calculates the similarity between the waveform of the partial data and the waveform of each template data. The similarity is calculated using methods such as Dynamic Time Warping (DTW). The production management device 10 extracts the combination of the partial data with the highest similarity and the template data. If the similarity of this combination exceeds a pre-set threshold, the production management device 10 presumes that the job corresponding to the template data has been executed.

[0055] The production management device 10 can also capture a portion of the detection data while changing the start time of capture. The production management device 10 compares multiple data segments with different start times with template data. Based on the similarity between each data segment and the template data, the operation to be executed is estimated.

[0056] Figure 6 (a) is a schematic diagram representing the time-series detection data obtained by the detector. Figure 6 (b) is a schematic diagram representing template data.

[0057] Figure 6 (a) represents detection data 200 from the detector 40 assembled with the wrench. In this case, a larger signal is detected along with the tightening of the bolt. The number of times the large signal is detected corresponds to the number of times the bolt is tightened and the number of bolts. The magnitude of the signal corresponds to the tightening strength. The interval between the large signals corresponds to the timing of the tightening. Figure 6 In example (a), the operator performs operation W2 after operation W1. In operation W2, two bolts are tightened. The operator tightens each bolt three times. Therefore, three peaks 201 to 203 are detected during the tightening of one bolt.

[0058] The production management device 10 extracts data with a time width of 205 from the detection data 200. The production management device 10 compares a portion of the data 210 with template data for a job that can be executed after job W1. Figure 6(b) represents an example of template data 220 corresponding to operation W2. Template data 220 includes the waveform when one bolt is tightened. A high degree of similarity is obtained between partial data 210 and template data 220. As a result, the production management device 10 infers that the operator is performing operation W2. Furthermore, the production management device 10 calculates the number of tightened bolts based on the matching of partial data 212, which is later than partial data 210, with the template data. For example, the number of bolts tightened in operation W2 is registered in the work standard. By comparing the counted number of bolts with the number of bolts registered in the work standard, the production management device 10 can infer the progress of operation W2.

[0059] The production management device 10 estimates the time limit (LT) based on the estimated results of the operations. The method for estimating the LT is arbitrary. For example, the time from the previously estimated operation to the next estimated operation can be used to estimate the LT of one operation. Alternatively, template data for estimating the start and end of one operation can be prepared separately, and the LT can be estimated based on the matching results using these template data. Another option is to prepare template data corresponding to the entire operation, performing template matching while varying the period of the extracted data. In this case, the length (time) of the data segment with the highest similarity is estimated as the preparation time. In the case of repeated specific operations, the average LT of one operation can be calculated by dividing the time until multiple operations are performed by the number of times the operation is performed.

[0060] Even when an operator repeatedly performs a specific task, template data can be determined for comparison with the detection data. For example, if an operator repeatedly performs task W2, the production management device 10 compares only a portion of the data with the template data 220. The production management device 10 repeatedly extracts data and matches portions of the data with the template data 220. Based on the results, the production management device 10 can estimate the progress of a task W2, the number of times task W2 has been performed so far, the time to completion (LT) of task W2, etc.

[0061] When detector 40 is an acceleration sensor or angular velocity sensor installed on the worker, the operation and time limit are estimated in the same way as described above. Multiple detectors 40 can also be installed on the worker. In this case, template data is prepared for each part of the worker's body. The production management device 10 compares the partial data obtained from each detection data point with the individual template data, and estimates that the operation with the template data that yielded the highest similarity was performed.

[0062] Time-series data can also be obtained from image data. The production management device 10 detects the posture of the worker reflected in the image. Through posture detection, the worker's skeleton is detected. The production management device 10 calculates the position of a specific part (e.g., the head) in the image. The production management device 10 continuously acquires images or acquires dynamic images. The production management device 10 calculates the position of the specific part based on multiple temporally consecutive images. Thus, for a specific part, time-series data representing continuous positional changes is obtained. This time-series data can also be used to estimate the ongoing operation and its preparation time, similar to the method described above.

[0063] Data specific to each task can also be used in the estimation. For example, for a task performed by a particular worker, the time limit (LT), task progress, and number of tasks can be estimated based on changes in the quantity of various types of items reflected in the image. For other tasks performed by other workers, the LT, task progress, and number of tasks can be estimated based on the matching of at least a portion of the detection data with the template data. Alternatively, for a single task, the LT, task progress, and number of tasks can be estimated by combining changes in the quantity of various types of items reflected in the image with template matching.

[0064] During the execution of production operations along a short-term plan, the production management device 10 repeatedly identifies items using image data and performs template matching of inspection data. As a result, the production management device 10 obtains performance data for each operator. This performance data includes time intervals (LTs), number of operations, etc., within a portion of the executed short-term plan.

[0065] The production management device 10 uses the obtained performance data for each operator to predict the performance of other parts of the short-term plan. For example, the short-term plan specifies the daily work schedule for each operator. For a given operator, the production management device 10 estimates the working time (LT) for the morning's work based on image data or inspection data obtained in the morning. Using the estimated LT, the production management device 10 predicts the LT for the afternoon's work for that operator.

[0066] The production management device 10 uses a first prediction model to predict performance. The first prediction model preferably includes a neural network. To improve prediction accuracy, the first prediction model more preferably includes a recurrent neural network (RNN). The first prediction model is pre-learned to output predictions of performance based on input data. During learning, performance data obtained from past operations is used. A portion of the actual data is used as input data, and another portion of the actual data is used as teaching data to be predicted.

[0067] In addition to actual performance data, the production management device 10 can also input other data that may affect LT into the first prediction model. For example, the production management device 10 can also input environmental data representing the work site environment and worker biological data into the prediction model. Environmental data includes one or more selected from temperature, humidity, and air pressure. Biological data includes one or more selected from body temperature, pulse rate, blood pressure, body movement, cardiac potential, perspiration, transcutaneous arterial oxygen saturation, and respiratory rate. Biological data can be obtained by wearable sensors installed on the worker.

[0068] Figure 7 (a) is a diagram illustrating the actual performance of a portion of a short-term plan. Figure 7 (b) is a schematic diagram representing the forecast of the other part of the short-term plan.

[0069] exist Figure 7 (a) and Figure 7 In (b), the horizontal axis represents time. The vertical axis represents LT. Figure 7 In (b), the solid line represents the estimated LT for operations performed along a portion of the short-term plan. The dashed line represents the projected LT for operations to be performed later along other portions of the short-term plan. Figure 7 (a) and Figure 7 As shown in (b), the production management device 10 uses estimated performance data and a prediction model to predict LT.

[0070] The production management unit 10 uses the predicted time to adjust the short-term plan. First, the production management unit 10 compares the predicted time with the short-term plan. The short-term plan specifies the number of operations to be performed. If the predicted time is longer than the standard, it may be impossible to perform the specified number of operations, and the number of operations needs to be reduced. If the predicted time is shorter than the standard, more operations can be performed.

[0071] Figure 8 (a) is a table representing a short-term plan made in advance. Figure 8 (b) is a table representing the revised short-term plan.

[0072] exist Figure 8 In example (a), planner X and operator Y assemble two units of model A each in the morning, the first half of the afternoon, and the second half of the afternoon. For example, Figure 7 (a) and Figure 7 (b) represents the actual performance and forecast of worker X. For example... Figure 7 (a) and Figure 7 As shown in (b), when a reduction in the predicted LT for operator X is anticipated, operator X is subsequently able to assemble more equipment. The production management unit 10, based on the predicted LT, as follows... Figure 8 (a) and Figure 8 As shown in (b), the number of devices assembled by worker X in the afternoon is increased.

[0073] The production management device 10 can also send notifications to pre-defined terminal devices when the short-term plan is revised. For example, the production management device 10 sends the revised number of units to be put into operation to the terminal devices held by the transport operators. By sending at least a portion of the revised short-term plan, operators can easily perform their work according to the revised short-term plan.

[0074] For example, a wearable device worn by the worker can notify the worker of revisions to the short-term plan via sound, light, or vibration. The device can be a smartphone, tablet, or smart glasses with a display. In this case, the device displays the revised short-term plan or the details of the revisions to the short-term plan to the worker.

[0075] Figure 9 (a) is a diagram illustrating the actual performance of a portion of the revised short-term plan. Figure 9 (b) is a schematic diagram representing the forecasts for the other part of the revised short-term plan.

[0076] Figure 10 (a) is a table representing the revised short-term plan. Figure 10 (b) is a table representing the further revised short-term plan.

[0077] exist Figure 9 (a) and Figure 9 In (b), the horizontal axis represents time. The vertical axis represents LT. Figure 9 In (b), the solid line represents the estimated LT for operations performed along a portion of the revised short-term plan. The dashed line represents the projected LT for operations to be performed in the future along other portions of the revised short-term plan. The revised short-term plan may also be further revised based on subsequent performance data.

[0078] Figure 9 (a) indicates the LT (Long Time Limit) when operator X performs the first half of the afternoon's work according to the revised short-term plan. Production management device 10 is based on performance in the morning and the first half of the afternoon's work, such as... Figure 9 As shown in (b), predicting LT in the latter half of the afternoon's work. In the example illustrated, with Figure 7 Compared to the prediction shown in (b), LT was not shortened. Production management device 10, based on the prediction results, such as... Figure 10 (a) and Figure 10As shown in (b), the number of devices assembled by operator X is reduced. In this way, production management device 10 can also use image data or detection data acquired in real time during the execution of the operation to repeatedly revise short-term plans.

[0079] At the end of a day's work according to the short-term plan, the production management unit 10 obtains performance data representing the actual performance of that day's work. This performance data, similar to the performance data in the aforementioned part of the short-term plan, includes time taken (LT), number of jobs performed, etc. Performance data is obtained for each operator. Using the obtained performance data and their previous past performance data, the production management unit 10 obtains long-term forecast data for each operator. This forecast data represents a prediction of work performance over a specified period (several weeks to several months). For example, the forecast data includes LT, number of jobs performed, etc., for each day within the specified period.

[0080] The production management device 10 uses a second prediction model to predict performance. The second prediction model is used to predict performance over a longer period compared to the first prediction model. The second prediction model preferably includes a neural network. To improve prediction accuracy, the second prediction model more preferably includes an RNN. The second prediction model is pre-learned to output performance predictions based on input data. Long-term past performance data is used during learning. A portion of the actual data is used as input data, and another portion of the actual data is used as teaching data to be predicted. Environmental data, biological data, etc., can also be input into the second prediction model, similar to the first prediction model.

[0081] The production management unit 10 updates the production master data and manufacturing master data based on the forecast data output from the second forecast model. First, the production management unit 10 updates the LT of the manufacturing master data for each operator based on the forecast data for each operator. The production management unit 10 then updates the LT of the production master data according to the updated LT of the manufacturing master data.

[0082] Figure 11 (a) is a diagram representing a portion of the performance during the specified period. Figure 11 (b) is a schematic diagram representing the forecast for the other part of the specified period.

[0083] exist Figure 11 (a) and Figure 11 In (b), the horizontal axis represents time, and the vertical axis represents LT. If according to... Figure 4 (b) Figure 8 (b) and Figure 10 If the short-term plan shown in (b) is executed on "February 2nd", then the actual results of the "February 2nd" operation can be obtained. (Refer to Production Management Device 10) Figure 11(a) shows the performance data for "February 2nd" and the performance data prior to that date. Production management unit 10 uses this performance data, such as... Figure 11 As shown in (b), the projected performance after "February 3rd" is presented.

[0084] Figure 12 (a) is a table that represents the updated manufacturing master data. Figure 12 (b) is a table that illustrates the updated production master data.

[0085] For example, Figure 11 (a) and Figure 11 (b) represents the actual performance of worker X and its forecast. For example... Figure 11 (a) and Figure 11 As shown in (b), in the case of a long-term prediction of a shortening of the LT of worker X, as Figure 12 As shown in (a), the production management device 10 updates the manufacturing master data LT related to operator X. Based on the update of the manufacturing master data, the production management device 10, as... Figure 12 As shown in (b), update the LT of the job associated with operator X in the production master data.

[0086] The production management unit 10 uses the updated production master data to create a new long-term plan. The creation of the long-term plan can utilize a scheduler as described above. Based on the new long-term plan and the updated manufacturing master data, the production management unit 10 creates a short-term plan for the next day.

[0087] Then, the above process is repeated. That is, production-related tasks are performed according to the new short-term plan on the next day. Based on the performance of the tasks performed along a part of the short-term plan, the subsequent performance is predicted. The short-term plan is revised based on the predicted performance. In addition, if the next day ends, a new long-term plan and a new short-term plan are created using the performance of the next day's tasks.

[0088] The examples above primarily illustrate the use of time-to-work (LT) and the number of times a job is executed as performance metrics. Besides this example, yield can also be used as a performance metric. For instance, the production management unit 10 counts the number of non-conforming products generated in jobs executed along a portion of the short-term plan. The generation of non-conforming products is determined based on inputs to specific terminal devices, the identification results of the number of items in the non-conforming product placement locations, etc. Based on the counted number of non-conforming products, the production management unit 10 predicts the number of non-conforming products in jobs executed along other portions of the short-term plan. In predicting the number of non-conforming products, a different prediction model than that used for LT prediction is employed. The production management unit 10 adjusts the short-term plan based on the predicted number of non-conforming products. Furthermore, the production management unit 10 counts the number of non-conforming products generated in a day's work and uses this count to predict the number of non-conforming products within a specified period. Based on the predicted number of non-conforming products, the manufacturing master data and production master data are updated, and new long-term and short-term plans are created.

[0089] The number of nonconforming items can also be expressed as the number of conforming items, the occurrence rate of nonconforming items, or other corresponding metrics such as yield. The production management unit 10 can also use both the number of nonconforming items and LT (Lead Time) to perform short-term plan revisions and update various master data.

[0090] Figure 13 This is a flowchart illustrating the production management method implemented.

[0091] exist Figure 13In the production management method (PM) shown, firstly, the production management device 10 creates a long-term plan (first long-term plan) representing the production plan for a specified period and a short-term plan (first short-term plan) representing the production plan for the first period (step S1). The long-term plan and the short-term plan are created based on various master data. Production-related operations are performed along a portion of the short-term plan. The production management device 10 obtains first performance data for this operation (step S2). The production management device 10 uses the first performance data to obtain first forecast data (step S3). The first forecast data includes forecasts of performance in other parts of the short-term plan's operations to be performed subsequently. The production management device 10 modifies the short-term plan based on the first forecast data (step S4). Steps S2 to S4 can also be repeated a predetermined number of times. The production management device 10 obtains second performance data representing the performance in the operations during the first period (step S5). The production management device 10 uses the second performance data and past performance data to obtain second forecast data representing forecasts of production performance for the specified period (step S6). The production management device 10 updates the manufacturing master data based on the second forecast data (step S7). The production management device 10 updates the production master data based on the updated manufacturing master data (step S8). The production management device 10 determines whether the processing termination condition is met (step S9). The termination condition is receiving a stop instruction from the user, executing the process a predetermined number of times, etc. If the termination condition is not met, step S1 is executed again.

[0092] For example, by repeating step S1, using the updated production master data and manufacturing master data, a new long-term plan (second long-term plan) and a short-term plan (second short-term plan) representing the production plan for the second period following the first period are created. Third performance data representing the performance of operations performed along a portion of the new short-term plan is obtained. Using this third performance data, third forecast data representing the forecast of the performance of operations performed along the other portion of the new short-term plan is obtained. Based on this third forecast data, the new second short-term plan is revised. Thus, using the obtained performance and the forecasts based on that performance, the revision of the short-term plan, the creation of new long-term plans, and the creation of new short-term plans are repeatedly performed.

[0093] Figure 14 and Figure 15 This is a flowchart illustrating the specific processing of the production management device in the implementation method.

[0094] Reference Figure 14 and Figure 15 ,illustrate Figure 13Here is an example of a specific method for estimating performance in step S2. First, the production management device 10 obtains data on a length of 2t from the detection data (step S21). "t" is a preset value. t can also be set based on the standard operating time registered in the work standards. The production management device 10 analyzes the obtained data (step S22). Based on the analysis results, the production management device 10 calculates various data related to performance (step S23). The production management device 10 determines whether the termination condition is met (step S24). If the termination condition is not met, step S21 is executed again.

[0095] exist Figure 15 The following describes a specific example of the processing in step S22. In this example, the analysis is performed using detection data obtained from detector 40. First, the production management device 10 segments the detection data into 2t lengths (step S22-1). For example, in temporary segmentation, the detection data is divided into pre-defined lengths. This results in multiple partial data. The production management device 10 reads unprocessed partial data from the multiple partial data (step S22-2). The production management device 10 reads unprocessed template data from multiple template data, where the similarity to the read partial data has not been calculated (step S22-3). The production management device 10 performs a dynamic programming method to establish a corresponding DTW between the template data and the partial data (step S22-4). The production management device 10 determines the average similarity of the obtained shortest path as the similarity between the partial data and the template data (step S22-5).

[0096] The production management device 10 determines whether the obtained similarity is the highest among the similarities obtained with respect to the partial data read in step S22-2. Then, the production management device 10 determines whether the obtained similarity exceeds a preset threshold (step S22-6). If the obtained similarity is the highest and exceeds the threshold, the production management device 10 presumes that a job corresponding to the template data was performed during the period of that partial data. The production management device 10 includes the action represented by this partial data in the number of jobs of the presumed job (step S22-7). In step S22-6, if the obtained similarity is not the highest or does not exceed the threshold, the production management device 10 determines whether there is unprocessed template data, which is template data for which the similarity for the partial data read in step S22-2 has not been calculated (step S22-8).

[0097] If unprocessed template data exists, step S22-3 is executed again to read the unprocessed template data. If no unprocessed template data exists, or if step S22-7 has been executed, the production management device 10 determines whether there is unprocessed partial data (step S22-9). If unprocessed partial data exists, step S22-2 is executed again to read the unprocessed partial data. If no unprocessed partial data exists, the production management device 10 ends step S22.

[0098] The advantages of the implementation method are explained.

[0099] In production, production schedulers are typically used to create both long-term and short-term plans. By utilizing production schedulers, production efficiency can be improved. In particular, production schedulers consider not only the time to delivery (LT) of tasks but also the time to transport goods when creating plans. Therefore, if the created plans can be executed, just-in-time production can be achieved.

[0100] On the other hand, deviations between plans and actual results are frequent due to factors such as short-term fluctuations in operator concentration, differences in operator skill levels, and long-term changes in operator skill levels. Furthermore, the master data required for plan creation, such as production time (LT) and yield, are maintained infrequently. These data deviate from actual values, resulting in situations where the production scheduler cannot create accurate plans. Whenever a deviation occurs, the person in charge of the plan adjusts the plan to reduce the deviation. However, this operation requires a significant amount of man-hours.

[0101] To address this technical problem, the production management device 10 in this embodiment uses the performance of tasks performed along a portion of a short-term plan to predict the performance of tasks performed along other portions of the same short-term plan. For example, LT is affected by the operator's concentration. Even if the operator's concentration changes throughout the day, the performance of subsequent tasks is predicted based on the performance of a portion of the tasks. The production management device 10 then modifies the short-term plan based on the predicted performance. By modifying the short-term plan, the deviation between the short-term plan and actual production can be reduced.

[0102] Furthermore, the production management device 10 uses past operational performance to predict production performance within a specified period. The production management device 10 uses this prediction to create new long-term and short-term plans. This reduces the deviation between long-term plans and actual production. Additionally, the short-term plans based on the new long-term plans are further revised using data from the execution of those short-term plans.

[0103] In this way, by repeatedly revising short-term plans, creating new long-term plans, and creating new short-term plans, the deviation between planned and actual production can be reduced. In particular, by analyzing and estimating operations using image data, inspection data, etc., short-term plans can be revised in real time. Short-term plans can also be revised repeatedly. According to the implementation method, plans with smaller deviations from actual production can be created. Therefore, more efficient production can be achieved compared to the past.

[0104] Furthermore, the production management device 10 updates the master data based on long-term performance forecasts. By automatically updating the master data, manual updates are no longer required, thus reducing the burden of manual master data maintenance.

[0105] Figure 16 It is a schematic diagram representing the hardware structure.

[0106] Production management device 10 includes, for example, Figure 16 The hardware structure shown. Figure 16 The computer 90 shown includes a CPU 91, ROM 92, RAM 93, storage device 94, input interface 95, output interface 96, and communication interface 97. The functions of the production management device 10 can be implemented by one computer 90 or by the cooperation of multiple computers 90.

[0107] ROM 92 stores programs used to control the computer's operations. ROM 92 contains the programs necessary for the computer to perform the aforementioned processes. RAM 93 functions as a storage area expanded from the programs stored in ROM 92.

[0108] CPU 91 includes processing circuitry. CPU 91 uses RAM 93 as its working memory and executes programs stored in at least one of ROM 92 or storage device 94. During program execution, CPU 91 controls various structures via system bus 98 and performs various processes.

[0109] Storage device 94 stores the data required for program execution and the data obtained through program execution.

[0110] The input interface (I / F) 95 connects the computer 90 and the input device 95a. The input I / F 95 is, for example, a serial bus interface such as USB. The CPU 91 can read various data from the input device 95a via the input I / F 95.

[0111] Output interface (I / F) 96 connects computer 90 and output device 96a. Output I / F 96 is, for example, a video output interface such as Digital Visual Interface (DVI) or High-Definition Multimedia Interface (HDMI (registered trademark)). CPU 91 can send data to output device 96a via output I / F 96, causing output device 96a to display images.

[0112] The communication interface (I / F) 97 connects the external server 97a to the computer 90. The communication I / F 97 is, for example, a network card such as a LAN card. The CPU 91 can read various data from the server 97a via the communication I / F 97.

[0113] Computer 90 can also communicate with terminal device 97b via communication I / F 97. Terminal device 97b is, for example, a smartphone, tablet, smart glasses, or wearable device carried by the operator.

[0114] Storage device 94 includes one or more selected from Hard Disk Drive (HDD) and Solid State Drive (SSD). Input device 95a includes one or more selected from mouse, keyboard, microphone (voice input), and touchpad. Output device 96a includes one or more selected from monitor, projector, speaker, and printer. A device that combines the functions of both input device 95a and output device 96a, such as a touchpad, may also be used. Storage device 94 may also be used as storage device 20.

[0115] The processing of the various data described above can also be recorded as programs that can be executed by a computer on a disk (floppy disk and hard disk, etc.), optical disk (CD-ROM, CD-R, CD-RW, DVD-ROM, DVD±R, DVD±RW, etc.), semiconductor memory, or other non-transitory computer-readable storage medium.

[0116] For example, information recorded on a recording medium can be read by a computer (or embedded system). The recording format (storage format) on the recording medium is arbitrary. For instance, a computer reads a program from the recording medium and, based on that program, causes the CPU to execute the instructions described in the program. In a computer, program retrieval (or reading) can also be performed via a network.

[0117] The implementation method may include the following features.

[0118] (Feature 1)

[0119] A production management device,

[0120] A first short-term plan is obtained, which is prepared based on a first long-term plan representing a production plan for a specified period, and represents a production plan for a first period shorter than the specified period.

[0121] Obtain first performance data, which represents the performance in operations performed along a portion of the first short-term plan.

[0122] Using the first performance data, first forecast data is obtained, which represents a forecast of the performance in the tasks performed along another part of the first short-term plan.

[0123] Based on the first forecast data, revise the first short-term plan.

[0124] Using second performance data representing the actual performance of the operations during the first period, second forecast data representing the forecast of production performance during the specified period is obtained.

[0125] Using the second forecast data, a second long-term plan is created for the specified period.

[0126] (Feature 2)

[0127] According to the production management device of feature 1, a second short-term plan is made based on the second long-term plan, which represents the production plan in a second period that is shorter than the specified period after the first period.

[0128] (Feature 3)

[0129] According to the production management device described in feature 2,

[0130] Obtain third performance data, which represents the performance of the operations performed as part of the second short-term plan.

[0131] Using the third performance data, third forecast data is obtained, which represents a forecast of the performance of the operations performed along another part of the second short-term plan.

[0132] The second short-term plan will be revised based on the third forecast data.

[0133] (Feature 4)

[0134] The production management device according to any one of features 1 to 3,

[0135] It also obtains additional performance data representing the performance of the operations carried out along a portion of the revised first short-term plan.

[0136] Using the aforementioned additional performance data, further forecast data is obtained to represent the forecasts of performance in the operations performed along another portion of the revised first short-term plan.

[0137] Based on the other forecast data, the revised first short-term plan is further revised.

[0138] (Feature 5)

[0139] The production management device according to any one of features 1 to 4,

[0140] The first performance data includes the preparation time for the tasks performed along the first short-term plan.

[0141] Image data reflecting the operation was repeatedly acquired.

[0142] Identify the type and quantity of items reflected in each of the plurality of image data.

[0143] The preparation time is calculated using the variations in the type and quantity of the items.

[0144] (Feature 6)

[0145] The production management device according to any one of features 1 to 4,

[0146] The first performance data includes the preparation time for the tasks performed along the first short-term plan.

[0147] Acquire time-series detection data representing signals generated by the actions of the operator performing the task.

[0148] The preparation time is calculated based on the comparison results between at least a portion of the detection data and the template data.

[0149] (Feature 7)

[0150] According to any one of features 1 to 6, in the production management device, the second performance data includes the preparation time in the operation during the first period.

[0151] (Feature 8)

[0152] The production management device according to any one of features 1 to 7,

[0153] Use the second predicted data to update the master data.

[0154] The second long-term plan is created by scheduling based on the updated master data.

[0155] (Feature 9)

[0156] According to any one of features 1 to 8, the production management device obtains the first prediction data by inputting the first performance data and environmental data representing the production environment into a prediction model including a neural network.

[0157] (Feature 10)

[0158] A production management device,

[0159] A short-term plan is obtained, which is made based on a long-term plan representing the production plan for a specified period, and a production plan for a first period shorter than the specified period.

[0160] Obtain past preparation time for operations performed as part of the aforementioned short-term plan.

[0161] Using the past preparation time, predict the future preparation time for the tasks to be performed along the other part of the short-term plan.

[0162] The short-term plan will be revised based on the estimated future preparation time.

[0163] Using the preparation time in the operations performed along the revised short-term plan, predict the preparation time for the operations during the specified period.

[0164] Using the predicted preparation time within the specified period, a new long-term plan for the specified period is created.

[0165] (Feature 11)

[0166] A production management system, comprising:

[0167] The production management device according to any one of features 1 to 10;

[0168] A camera device is used to record the operation; and

[0169] A detector that detects signals generated by the actions of the operator performing the task.

[0170] Based on the embodiments described above, a production management device, a production management system, a production management method, and a storage medium are provided that can produce a production plan with smaller deviations from actual production.

[0171] The above examples illustrate several embodiments of the present invention, but these embodiments are merely illustrative and not intended to limit the scope of the invention. These new embodiments can be implemented in a wide variety of other ways, with various omissions, substitutions, and modifications possible without departing from the spirit of the invention. These embodiments or variations thereof are included within the scope and spirit of the invention, and are also included within the scope of the invention described in the patent application and its equivalents. Furthermore, the various embodiments can be implemented in combination with each other.

Claims

1. A production management apparatus, acquires a first short-term plan made based on a first long-term plan that represents a plan for production in a prescribed period, a plan for production in a first period shorter than the prescribed period, acquires first actual performance data that represents actual performance in a work performed along a part of the first short-term plan, acquires first prediction data that represents a prediction of actual performance in the work performed along another part of the first short-term plan using the first actual performance data, corrects the first short-term plan based on the first prediction data, acquires second prediction data that represents a prediction of actual performance in the production in the prescribed period using second actual performance data that represents actual performance in the work in the first period, makes a second long-term plan for the production in the prescribed period using the second prediction data.

2. The production management apparatus according to claim 1, makes a second short-term plan that represents a plan for production in a second period after the first period and shorter than the prescribed period based on the second long-term plan.

3. The production management apparatus according to claim 2, acquires third actual performance data that represents actual performance in the work performed along a part of the second short-term plan, acquires third prediction data that represents a prediction of actual performance in the work performed along another part of the second short-term plan using the third actual performance data, corrects the second short-term plan based on the third prediction data.

4. The production management apparatus according to any one of claims 1 to 3, further acquires other actual performance data that represents actual performance in the work performed along a part of the corrected first short-term plan, acquires other prediction data that represents a prediction of actual performance in the work performed along another part of the corrected first short-term plan using the other actual performance data, further corrects the corrected first short-term plan based on the other prediction data.

5. The production management apparatus according to any one of claims 1 to 3, the first actual performance data includes a preparation time in the work performed along the part of the first short-term plan, repeatedly acquires image data that represents the work, identifies a kind and a number of articles represented in each of a plurality of the image data, calculates the preparation time using a change in the kind and the number of the articles.

6. The production management apparatus according to any one of claims 1 to 3, the first actual performance data includes a preparation time in the work performed along the part of the first short-term plan, acquires detection data that represents a time series of signals generated by an action of a worker who performs the work, calculates the preparation time based on a result of comparison of at least a part of the detection data with template data.

7. The production management apparatus according to any one of claims 1 to 3, The second performance data includes a preparation time in the work in the first period.

8. The production management apparatus according to any one of claims 1 to 3, The second prediction data is used to update production master data, The second long-term plan is made through scheduling based on the updated production master data.

9. The production management apparatus according to any one of claims 1 to 3, The first prediction data is obtained by inputting the first performance data and environment data representing an environment of the production to a prediction model including a neural network.

10. A production management apparatus, A short-term plan is obtained, the short-term plan being made based on a long-term plan representing a plan of production in a prescribed period, representing a plan of production in a first period shorter than the prescribed period, A past preparation time in a work performed along a part of the short-term plan is obtained, A future preparation time in the work performed along another part of the short-term plan is predicted using the past preparation time, The short-term plan is revised based on the future preparation time, A preparation time in the work performed along the revised short-term plan is predicted, A new long-term plan in the prescribed period is made using the predicted preparation time in the prescribed period.

11. A production management system comprising: The production management apparatus according to any one of claims 1 to 10; An imaging device that images the work; and A detector that detects a signal generated by an action of a worker performing the work.

12. A production management method, A first short-term plan is obtained, the first short-term plan being made based on a first long-term plan representing a plan of production in a prescribed period, representing a plan of production in a first period shorter than the prescribed period, First performance data representing a performance in a work performed along a part of the first short-term plan is obtained, First prediction data representing a prediction of a performance in the work performed along another part of the first short-term plan is obtained using the first performance data, The first short-term plan is revised based on the first prediction data, Second prediction data representing a prediction of a performance of production in the prescribed period is obtained using second performance data representing a performance in the work in the first period, The second prediction data is used to make a second long-term plan in the prescribed period.

13. A storage medium storing a program causing a computer to execute the production management method according to claim 12.

Images