Management support system, management support method, and management support program

Abstract

To manage a manufacturing process including a plurality of manufacturing processes suitably.SOLUTION: A management support system manages a manufacturing process including a first process and a second process of which manufacturing units are different from each other. The management support system obtains a first process data in regard to a first processing at the first process, a second process data in regard to a second processing at the second process and a quality data showing a quality of the final product obtained by the manufacturing process, associates the first process data with the second process data mutually by referring to a trace data showing the relation between a first manufacturing unit at the first process and a second manufacturing unit at the second process, executes a machine learning based on the first process data with the second process data mutually associated and the quality data and generates a regression model for calculating an estimation value of the quality of the final product from the process data in regard to the processing at the manufacturing process.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] One aspect of the present disclosure relates to a management support system, a management support method, and a management support program.

Background Art

[0002] Conventionally, systems for managing manufacturing processes have been known. For example, Patent Document 1 describes a process abnormality analyzer that detects process abnormalities for each unit target product based on process data obtained during process execution in a manufacturing system consisting of a plurality of manufacturing processes.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In one aspect of the present disclosure, a mechanism for appropriately managing a manufacturing process including a plurality of manufacturing steps is desired.

Means for Solving the Problems

[0005] A management support system relating to one aspect of this disclosure is a management support system for managing a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and comprises at least one processor, the at least one processor acquires first process data relating to a first process in the first process, second process data relating to a second process in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process, relates the first process data and the second process data to each other by referring to trace data indicating the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process, and performs machine learning based on the related first process data and second process data and quality data to generate a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process.

[0006] A management support method relating to one aspect of this disclosure is a management support method performed by a management support system that manages a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and which has at least one processor, and includes the steps of: acquiring first process data relating to a first process in the first process, second process data relating to a second process in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process; relating the first process data and the second process data to each other by referring to trace data indicating the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process; and generating a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process by performing machine learning based on the related first process data and second process data and quality data.

[0007] A management support program relating to one aspect of this disclosure is a management support program that causes a computer to function as a management support system for managing a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and causes the computer to perform the following steps: acquiring first process data relating to a first process in the first process, second process data relating to a second process in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process; relating the first process data and the second process data to each other by referring to trace data indicating the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process; and generating a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process by performing machine learning based on the related first process data and second process data and quality data.

[0008] In this respect, the process data from the first and second processes, which have different manufacturing units, are correlated based on trace data that shows the relationship between the two manufacturing units. Then, machine learning is performed based on the correlated process data and quality data related to the final product to generate a regression model. Thus, it is possible to generate a regression model based on process data spanning multiple processes, while absorbing the differences in manufacturing units between processes. By using this regression model, it is possible to appropriately manage a manufacturing process that includes multiple manufacturing processes. [Effects of the Invention]

[0009] According to one aspect of this disclosure, a manufacturing process that includes multiple manufacturing steps can be properly managed. [Brief explanation of the drawing]

[0010] [Figure 1] This figure shows an example of the functional configuration of a management support system. [Figure 2] This figure shows an example of the folder structure of external storage. [Figure 3] This is a flowchart showing an example of generating a regression model. [Figure 4] This figure shows an example of process data association. [Figure 5] This flowchart shows an example of quality estimation using a regression model. [Figure 6] This figure shows an example of a user interface for displaying estimation results. [Modes for carrying out the invention]

[0011] The embodiments described herein will be described in detail below with reference to the attached drawings. In the description of the drawings, the same or equivalent elements are denoted by the same reference numerals, and redundant descriptions are omitted.

[0012] [System Overview] The management support system 10 according to this embodiment is a computer system for supporting the management of a manufacturing process that includes multiple manufacturing steps. A manufacturing process refers to a series of processes for producing a certain product. In this disclosure, the product ultimately obtained by the manufacturing process is also referred to as the "final product." A manufacturing step refers to a component of the manufacturing process. In each manufacturing step, a process specific to that step is executed.

[0013] In one example, the management support system 10 manages a manufacturing process that includes two manufacturing processes, designated as the first and second processes, where the manufacturing units are different from each other. A manufacturing unit refers to a group of products manufactured under the same conditions, and is also called a lot. Each manufacturing unit is identified by an identifier such as a lot number. The manufacturing process managed by the management support system 10 may include two or more manufacturing processes that have the same manufacturing unit. Alternatively, the manufacturing units may differ among the multiple manufacturing processes that constitute the manufacturing process.

[0014] The management support system 10 can be used for various purposes. For example, the management support system may assist in the management of the manufacturing process of functional films. Functional films are films that have added value by coating a substrate with a polymer substance having various functions. Functional films can be used, for example, for surface protection, contamination prevention, improved lubricity, anti-reflective properties, protection during processing or transport of thin products (e.g., ITO films, flexible printed circuit boards, metal foils), or improved chemical resistance. The manufacturing process of a functional film includes, for example, a synthesis step to produce materials used in the manufacture of a coating solution, a compounding step to produce the coating solution, and a coating step to produce a film by coating the coating solution onto a given substrate. The manufacturing process may further include an inspection step to actually inspect the quality of the functional film, which is the final product. In one example, the management support system 10 manages a manufacturing process in which at least two of these synthesis, compounding, and coating steps are managed in different manufacturing units.

[0015] In one example, the management support system 10 generates a regression model for estimating the quality of the final product using machine learning. Machine learning is a method that autonomously discovers laws or rules by iteratively learning based on given information. The management support system 10 may use this regression model to estimate the quality of the final product. The generation of the regression model corresponds to the learning phase, and the estimation using that regression model corresponds to the operation phase or monitoring phase. In this disclosure, the operation phase or monitoring phase is also referred to as the "estimation process."

[0016] In one example, the management support system 10 correlates individual process data related to processing in each manufacturing process to generate process data representing the entire manufacturing process as total process data. The management support system 10 then performs machine learning based on this total process data and quality data representing the quality of the final product to generate a regression model. As described above, since the manufacturing process includes a first and second process where the manufacturing units are different from each other, simply referring to individual process data does not allow for the correlation between the first process data relating to the first processing in the first process and the second process data relating to the second processing in the second process. The management support system 10 performs this correlation by referring to trace data that shows the relationships between multiple manufacturing processes corresponding to multiple manufacturing processes. The trace data shows at least the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process. Therefore, by using trace data, it becomes possible to generate total process data, and as a result, it becomes possible to generate a regression model that takes the entire manufacturing process into account.

[0017] A regression model is a computational model that calculates an estimate of the quality of the final product from all process data. All process data is an example of process data related to the processing in the manufacturing process. All process data corresponds to the explanatory variables in the regression model, and the estimated quality of the final product corresponds to the dependent variable. With x as the explanatory variable and y as the dependent variable, the regression model can be expressed as y = f(x).

[0018] [System Configuration] FIG. 1 is a diagram showing an example of the functional configuration of the management support system 10. The management support system 10 includes a processor 101 and a memory 102 as hardware devices. The processor 101 is, for example, a CPU (Central Processing Unit), a DSP (Digital Signal Processor), or a GPU (Graphics Processing Unit). The memory 102 is composed of a storage device such as a flash memory or a hard disk, for example. However, the types of hardware devices constituting the management support system 10 are not limited to these and may be arbitrarily selected. Each function of the management support system 10 is realized by the processor 101 executing a program stored in the memory 102.

[0019] The management support program for causing a computer to function as the management support system 10 includes program codes for realizing each functional module of the management support system 10. This management support program may be provided after being non-temporarily recorded on a tangible recording medium such as a CD-ROM, a DVD-ROM, or a semiconductor memory. Alternatively, the management support program may be provided via a communication network as a data signal superimposed on a carrier wave. The provided management support program is stored in the memory 102, for example.

[0020] The management support system 10 may be composed of one computer or a set of multiple computers, that is, a distributed system. The computers used for the management support system 10 are not limited. For example, various types of computers such as personal computers, workstations, tablet terminals, and smartphones can be used as at least a part of the management support system 10. When multiple computers are used for the management support system 10, these computers are connected via a communication network such as the Internet or an intranet, thereby logically constructing one management support system 10.

[0021] In one example, the management support system 10 connects to external storage 20 and user terminals 30 via a communication network. This communication network may be the internet, an intranet, or a combination thereof. The communication network may be a wired network, a wireless network, or a combination thereof.

[0022] The external storage 20 is a device or recording medium that stores various types of data used for processing in the management support system 10. The external storage 20 may be a component of the management support system 10, or it may be located in a computer system separate from the management support system 10.

[0023] Figure 2 shows an example of the folder structure of the external storage 20. In this example, the folder structure includes an output folder that stores estimation results regarding the quality of the final product, and a dataset folder that stores various data used for the estimation. The output folder stores the estimation results for each manufacturing unit (each lot). The dataset folder includes a sensor folder that stores sensor data for each process, a trace folder that stores trace data, and a process quality folder that stores process quality data for each process. Sensor data is data obtained from one or more sensors installed in the manufacturing equipment that performs the manufacturing process. Process quality data is data that indicates the quality of the product obtained in the manufacturing process. Both sensor data and process quality data can constitute at least a part of the process data. The sensor folder includes multiple subfolders corresponding to multiple manufacturing processes, for example, a synthesis process folder, a compounding process folder, and a coating process folder. Each subfolder stores process data for each manufacturing unit (each lot) obtained in the corresponding manufacturing process. The process quality folder stores quality data of the product manufactured in each process.

[0024] In one example, sensor data and process quality data are generated based on data obtained from various sensors within the manufacturing equipment and stored in external storage 20. At least one of the sensor data and process quality data may include feature quantities described later. In one example, trace data is pre-created by the user of the management support system 10 and stored in external storage 20. Estimation results may be automatically stored in external storage 20 by the management support system 10, or they may be stored based on user instructions.

[0025] The user terminal 30 is a computer operated by a user of the management support system 10. For example, the user may be a manager or manufacturer of the manufacturing process. The user terminal 30 may be any computer, such as a personal computer, workstation, tablet, smartphone, or wearable device.

[0026] In one example, the processor 101 functions as a data preparation unit 11, a learning unit 14, and an estimation unit 15. The data preparation unit 11 includes an association unit 12 and a feature calculation unit 13.

[0027] The data preparation unit 11 is a functional module that accesses external storage 20 in both the learning phase and the operation phase to prepare the data to be used in each phase. In the learning phase, the data preparation unit 11 acquires from external storage 20 multiple process data corresponding to multiple manufacturing processes that constitute the manufacturing process, and quality data indicating the quality of the final product obtained by the manufacturing process. In the operation phase, the data preparation unit 11 acquires from external storage 20 multiple process data corresponding to multiple manufacturing processes. The process data in the learning phase is part of the training data. The process data in the operation phase is used to estimate unknown quality.

[0028] The association unit 12 is a functional module that associates multiple process data corresponding to multiple manufacturing processes based on trace data. In other words, the association unit 12 generates all process data.

[0029] The feature calculation unit 13 is a functional module that calculates feature quantities based on time-series data related to a predetermined process in a predetermined manufacturing process, and sets these feature quantities as at least a part of the total process data. Feature quantities refer to indicators that represent the characteristics of that time-series data. The feature calculation unit 13 may update the process data by adding the calculated feature quantities to the corresponding process data (sensor data or process quality data) stored in the external storage 20.

[0030] The learning unit 14 is a functional module that performs machine learning based on the interrelated process data, namely the total process data and the quality data, to generate a regression model 19 for estimating the quality of the final product. In other words, the learning unit 14 performs the learning phase.

[0031] The estimation unit 15 is a functional module that estimates the quality of the final product using a regression model 19. In other words, the estimation unit 15 executes the operational phase. The estimation unit 15 inputs the interconnected process data, i.e., all process data, into the regression model 19 and calculates an estimated value of the final product quality.

[0032] [System operation] Referring to Figures 3 to 6, an example of processing by the management support system 10 will be explained, along with an example of the management support method according to this embodiment. Figure 3 is a flowchart showing an example of the generation of a regression model 19 as processing flow S1. Figure 4 is a diagram showing an example of process data association. Figure 5 is a flowchart showing an example of quality estimation using the regression model 19 as processing flow S2. Figure 6 is a diagram showing an example of a user interface for displaying the estimation results.

[0033] The learning phase will be explained with reference to Figure 3. In step S11, the data preparation unit 11 acquires process data for each manufacturing process, final product quality data, and trace data from the external storage 20. The final product quality data may be represented by the process quality data of the final manufacturing process, or in the example in Figure 2, by the process quality data of the coating process. Alternatively, the final product quality data may be prepared based on data obtained automatically or manually in the actual inspection process.

[0034] In step S11, the feature calculation unit 13 may calculate features based on time-series data relating to processing in at least one manufacturing process and set these features as data items for the process data. For example, the feature calculation unit 13 calculates statistical values ​​of the time-series data of a certain data item as features. Examples of such statistical values ​​include maximum value, minimum value, mean, standard deviation, median, mode, and quartiles. For example, the feature calculation unit 13 may calculate features for each data item of the process data. The feature calculation unit 13 may perform standardization on each feature. This standardization is a scaling method that sets the mean to 0 and the variance to 1. This standardization can eliminate or reduce the effects of differences in units or scales between data items of the process data.

[0035] In step S12, the association unit 12 associates the process data of each manufacturing process with each other based on the trace data. Multiple process data corresponding to multiple manufacturing processes include first process data relating to the first processing in the first process and second process data relating to the second processing in the second process. The trace data shows at least the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process. The association unit 12 identifies the correspondence between the first manufacturing unit and the second manufacturing unit by referring to the trace data in the external storage 20, and associates the first process data and the second process data with each other based on that correspondence. If the manufacturing process includes three or more manufacturing processes in which the manufacturing units are different from each other, the trace data shows the relationships between the three or more manufacturing units corresponding to those three or more manufacturing processes. The association unit 12 selects two of the three or more manufacturing processes as the first and second processes and associates the two process data corresponding to these two manufacturing processes with each other. The association unit 12 performs the association while modifying the two manufacturing processes selected as the first and second processes based on the trace data, and finally generates all process data.

[0036] As described above, at least one process data corresponding to at least one manufacturing process may contain features. In this case, the association unit 12 consequently sets the features as at least a part of the total process data. For example, the association unit 12 may set the features as at least a part of the first process data and the second process data that are associated with each other.

[0037] An example of process data association will be explained with reference to Figure 4. Figure 4 shows process data for each of the synthesis, blending, and coating processes that constitute the manufacturing process of functional films, as well as trace data showing the relationship between the synthesis and blending processes, and trace data showing the relationship between the blending and coating processes.

[0038] Each process data set includes a lot number and at least one data item based on data obtained during the manufacturing process. Each data item of the process data may be represented directly by sensor data or process quality data, or by calculated feature quantities. Examples of data items in the synthesis process include the temperature or pressure in the tank and the temperature or quality of the resulting material. Examples of data items in the compounding process include the temperature in the tank, the current value of the agitator, and the quality of the resulting coating solution. Examples of data items in the coating process include the viscosity or density of the coating solution, the ambient temperature, and a given physical property of the functional film.

[0039] Each record in the trace data showing the relationship between the synthesis process and the compounding process includes the lot number of the synthesis process and the lot number of the compounding process. Each record in the trace data showing the relationship between the compounding process and the coating process includes the lot number of the compounding process and the lot number of the coating process.

[0040] In the example shown in Figure 4, the association unit 12 refers to trace data records 231 and 241 and associates the synthesis process data record 201 with the compounding process data record 211 with the coating process data record 221 to generate the total process data record 251. The association unit 12 further refers to trace data records 231 and 242 and associates the synthesis process data record 201 with the compounding process data record 211 with the coating process data record 222 to generate the total process data record 252. The association unit 12 also refers to trace data records 232 and 243 and associates the synthesis process data record 202 with the compounding process data record 212 with the coating process data record 223 to generate another total process data record 253.

[0041] Returning to Figure 3, in step S13, the learning unit 14 generates a regression model 19 by performing machine learning based on the interrelated process data, i.e., the total process data and the quality data. In this machine learning, the learning unit 14 uses each record of the total process data as an explanatory variable and the estimated quality of the final product, indicated by each record of the quality data, as the dependent variable. Examples of machine learning include linear regression, ridge regression, support vector machine, neural network, and random forest regression. Alternatively, the learning unit 14 may generate the regression model 19 by combining two or more of these methods.

[0042] In step S14, the learning unit 14 stores the generated regression model 19 in memory 102.

[0043] In one example, the management support system 10 executes a processing flow S1 for each type of final product to generate a regression model 19 for each type.

[0044] The operation phase will be explained with reference to Figure 5. In step S21, the data preparation unit 11 acquires process data and trace data for each manufacturing process. In one example, the data preparation unit 11 acquires process data obtained in real time. The process data acquired in step S21 is new process data that is different from the process data acquired in step S11. In step S21, the feature calculation unit 13 may calculate feature quantities based on time-series data relating to processing in at least one manufacturing process using the same method as in step S11, and set these feature quantities as data items for the new process data.

[0045] In step S22, the association unit 12 associates the process data (new process data) of each manufacturing process based on the trace data to generate new total process data. The association unit 12 performs this association using the same method as in step S12. In one example, the new total process data is represented by a single record, such as one of records 251 to 253 shown in Figure 4.

[0046] In step S23, the estimation unit 15 inputs the associated process data (new process data), i.e., all new process data, into the regression model 19 to calculate an estimate of the final product quality. The estimation unit 15 reads the regression model 19 corresponding to the final product for which it wants to calculate the quality estimate from the memory 102, inputs all of its process data into the regression model 19, and obtains the estimate calculated by the regression model 19.

[0047] In step S24, the estimation unit 15 transmits the estimation results, including the estimated values, to the user terminal 30. The user terminal 30 displays the estimation results in various formats, such as text, graphs, tables, and computer graphics. In other words, the estimation unit 15 displays the calculated estimated values ​​on the user terminal 30. This allows the user to understand what the quality of the final product will be.

[0048] In addition to or instead of its transmission process, the estimation unit 15 may store the estimation results in a predetermined storage device, such as external storage 20. In relation to the example in Figure 2, the estimation unit 15 may record the estimation results in a predetermined data file in the output folder. The user terminal 30 may access the external storage 20 directly or via the management support system 10 based on user instructions to obtain and display the estimation results.

[0049] As shown in step S25, the management support system 10 may repeat the processes in steps S21 to S24. In this iteration, the management support system 10 acquires even newer process data for each manufacturing process, calculates new estimates based on that process data, and sends the new estimation results, including those estimates, to the user terminal 30. The user terminal 30 displays the new estimation results. In one example, the management support system 10 repeatedly performs the estimation process to generate time-series data of the calculated estimates. The processing flow S2 terminates when the user selects to end the process or in response to the completion of the manufacturing process.

[0050] In one example, the management support system 10 executes the processing flow S2 for each type of final product and calculates an estimate for each type.

[0051] Referring to Figure 6, an example of the estimation results displayed on the user terminal 30 will be explained. In this example, the user terminal 30 displays the estimation results on the user interface 300. The user interface 300 includes a report 310 that displays the date and time, process name, product name, estimated quality, and process improvement suggestions, a report 320 that shows time-series data of the estimated quality, and a report 230 that shows the estimated quality for each type of final product.

[0052] The date and time in Report 310 indicates the date and time the data was acquired. The process name is the name of the manufacturing process. The product name is the name of the final product. The estimated quality indicates the estimated quality of the final product. The process improvement suggestion is displayed when an anomaly is detected in the manufacturing process and provides instructions for resolving that anomaly. The user can take action to improve the manufacturing process based on this Report 310.

[0053] Report 320 presents time-series data of estimated quality using a line graph 321 where the horizontal axis represents elapsed time and the vertical axis represents estimated quality. Report 320 also displays pre-set upper and lower specification limits. This report 320 allows users to grasp at a glance the changes in estimated quality over time and whether the current estimated quality is within specifications. For example, if the estimated quality falls outside specifications, it indicates an abnormality in the manufacturing process. Line graph 322 shows the estimated quality of the final product at a relatively later point in time, i.e., the estimated quality in the relatively distant future. The management support system 10 may calculate its estimated quality using a regression model 19, or it may calculate it using a different regression model or method.

[0054] Report 330 presents the estimated quality for each type of final product using a bar graph. This graph may show the most recent estimated quality. Report 330 also displays the upper and lower limit of the specification. This report 330 allows the user to understand the estimated quality of each final product.

[0055] [effect] As described above, the management support system relating to one aspect of this disclosure is a management support system for managing a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and acquires first process data relating to the first processing in the first process, second process data relating to the second processing in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process, relates the first process data and the second process data to each other by referring to trace data showing the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process, and performs machine learning based on the related first process data and second process data and quality data to generate a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process.

[0056] A management support method relating to one aspect of this disclosure is a management support method performed by a management support system that manages a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and which has at least one processor, and includes the steps of: acquiring first process data relating to a first process in the first process, second process data relating to a second process in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process; relating the first process data and the second process data to each other by referring to trace data indicating the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process; and generating a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process by performing machine learning based on the related first process data and second process data and quality data.

[0057] A management support program relating to one aspect of this disclosure is a management support program that causes a computer to function as a management support system for managing a manufacturing process including a first process and a second process in which the manufacturing units are different from each other, and causes the computer to perform the following steps: acquiring first process data relating to a first process in the first process, second process data relating to a second process in the second process, and quality data indicating the quality of the final product obtained by the manufacturing process; relating the first process data and the second process data to each other by referring to trace data indicating the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process; and generating a regression model that calculates an estimate of the quality of the final product from process data relating to the processing in the manufacturing process by performing machine learning based on the related first process data and second process data and quality data.

[0058] In this respect, the process data from the first and second processes, which have different manufacturing units, are correlated based on trace data that shows the relationship between the two manufacturing units. Then, machine learning is performed based on the correlated process data and quality data related to the final product to generate a regression model. Thus, it is possible to generate a regression model based on process data spanning multiple processes, while absorbing the differences in manufacturing units between processes. By using this regression model, it is possible to appropriately manage a manufacturing process that includes multiple manufacturing processes. For example, by obtaining estimates using this regression model, it becomes possible to understand the quality of the final product itself, or the state of the manufacturing processes that affect that quality.

[0059] When multiple manufacturing processes, each with different manufacturing units, are managed individually, it becomes difficult to identify and isolate the causes of defects in the final product or manufacturing process. While an overview of the entire manufacturing process facilitates management, simply referring to data from individual manufacturing processes does not easily provide a comprehensive picture. Therefore, maintaining a desired yield level is challenging. For example, to address this, multiple process data corresponding to multiple manufacturing processes are automatically correlated to generate process data for the entire manufacturing process, i.e., total process data. Then, a regression model is generated using machine learning with this total process data. Using this regression model, it becomes possible to easily understand the state within the manufacturing process and manage it appropriately. Ultimately, this can lead to the maintenance of a certain level of yield, or even an improvement in yield.

[0060] In management support systems relating to other aspects, at least one processor may calculate features based on time-series data related to at least one of the first and second processes, and set the calculated features as at least a portion of the first and second process data associated with each other. Since features that more clearly represent the processing state in the manufacturing process are used in machine learning, a more accurate regression model can be generated.

[0061] In management support systems relating to other aspects, at least one processor may perform an estimation process that calculates estimates using a regression model, and the estimation process may include the steps of: acquiring new first process data relating to the first process in the first step and new second process data relating to the second process in the second step; relating the new first process data and new second process data to each other by referring to trace data; and inputting the related new first process data and new second process data into a regression model to calculate estimates. With this configuration, the quality of the final product can be estimated using the generated regression model.

[0062] In management support systems relating to other aspects, at least one processor may repeatedly perform estimation processing to generate time-series data of the calculated estimates. This configuration allows for obtaining changes in the estimates along a time series.

[0063] In management support systems relating to other aspects, at least one processor may display the calculated estimates on the user terminal. This configuration allows the user to be presented with estimated results regarding the quality of the final product.

[0064] In management support systems related to other aspects, the final product may be a functional film. In this case, the system can appropriately manage the manufacturing process of the functional film, which may involve multiple manufacturing steps.

[0065] [Differentiation] The technology relating to this disclosure has been described in detail above based on various examples. However, this disclosure is not limited to the embodiments described above. The technology relating to this disclosure can be modified in various ways without departing from its essence.

[0066] The feature calculation unit 13 may calculate feature quantities based on time-series data relating to processing in at least one manufacturing process after all process data has been generated by the association unit 12, and set these feature quantities as predetermined data items of the process data. In other words, in the processing flow S1 described above, the calculation of feature quantities may be performed in either step S11 or S12. Similarly, in the processing flow S2 described above, the calculation of feature quantities may be performed in either step S21 or S22.

[0067] In the above embodiment, the management support system is configured as a client-server system such as a cloud system, but the management support system may also be implemented by a standalone computer.

[0068] The processing steps of a method executed by at least one processor are not limited to the examples in the above embodiments. For example, some of the steps (processes) described above may be omitted, or each step may be executed in a different order. Also, any two or more of the steps described above may be combined, or some of the steps may be modified or deleted. Alternatively, other steps may be executed in addition to each of the above steps.

[0069] In comparing the magnitudes of two numbers in this disclosure, either of the two criteria, "greater than or equal to" and "greater than," may be used, or either of the two criteria, "less than or equal to" and "less than," may be used. The choice of such criteria does not change the technical significance of the process of comparing the magnitudes of two numbers.

[0070] In this disclosure, the expression "at least one processor executes a first process, a second process, ... and the nth process," or a corresponding expression, refers to a concept that includes cases where the entity executing the n processes from the first process to the nth process changes along the way. In other words, this expression refers to a concept that includes both cases where all n processes are executed by the same processor and cases where the processor changes at an arbitrary rate for the n processes. [Explanation of Symbols]

[0071] 10...Management support system, 11...Data preparation unit, 12...Association unit, 13...Feature calculation unit, 14...Learning unit, 15...Estimation unit, 19...Regression model, 20...External storage, 30...User terminal, 101...Processor, 102...Memory, 300...User interface.

Claims

1. A management support system for managing a manufacturing process of a functional film, which includes a first and second process in which the manufacturing units are different from each other, Equipped with at least one processor, The at least one processor, for each of the two or more types of the functional film, First process data relating to the first treatment in the first step corresponding to the functional film of the said type, second process data relating to the second treatment in the second step corresponding to the functional film of the said type, and quality data indicating the quality of the functional film of the said type obtained by the manufacturing process of the functional film of the said type are obtained. By referring to trace data showing the relationship between the first manufacturing unit in the first process and the second manufacturing unit in the second process, the acquired first process data and the acquired second process data are correlated to generate total process data. Machine learning is performed based on the aforementioned total process data and the acquired quality data to generate a regression model that calculates an estimate of the quality of the functional film of that type from process data relating to the processing in the manufacturing process corresponding to the functional film of that type. The at least one processor performs an estimation process to calculate the estimated value for each of the two or more types of the functional film using the regression model corresponding to that type of functional film. Management support system.

2. The at least one processor, for each of the two or more types of the functional film, Feature quantities are calculated based on time-series data related to at least one of the first treatment and the second treatment corresponding to the functional film of the said type. The calculated feature quantities are set as at least a part of the total process data corresponding to the functional film of that type. The management support system according to claim 1.

3. The estimation process is performed for each of the two or more types of the functional film, A step of obtaining new first process data relating to the first treatment in the first step corresponding to the functional film of the said type, and new second process data relating to the second treatment in the second step corresponding to the functional film of the said type, The steps include: referencing the trace data to generate new total process data by relating the new first process data and the new second process data to each other; The steps include inputting the new total process data into the regression model corresponding to the type of functional film and calculating the estimated value, including, The management support system according to claim 1 or 2.

4. The at least one processor repeatedly performs the estimation process for each of the two or more types of the functional film to generate time-series data of the calculated estimated values. The management support system according to claim 3.

5. The at least one processor displays the calculated estimated value on the user terminal for each of the two or more types of the functional film. The management support system according to claim 3 or 4.

6. A management support method for managing a manufacturing process of a functional film, which includes a first and second process in which the manufacturing units are different from each other, and which is performed by a management support system equipped with at least one processor, For each of the two or more types of the aforementioned functional film, A step of obtaining first process data relating to a first treatment in the first step corresponding to the functional film of the said type, second process data relating to a second treatment in the second step corresponding to the functional film of the said type, and quality data indicating the quality of the functional film of the said type obtained by the manufacturing process of the functional film of the said type, A step of generating total process data by relating the acquired first process data and the acquired second process data to each other, by referring to trace data showing the relationship between the first manufacturing unit in the first step and the second manufacturing unit in the second step, The steps include: generating a regression model that calculates an estimated quality of a functional film of a certain type from process data relating to the processing in the manufacturing process corresponding to that type of functional film by performing machine learning based on the aforementioned total process data and the acquired quality data; The steps include: performing an estimation process to calculate the estimated value for each of the two or more types of functional film using the regression model corresponding to that type of functional film; Management support methods including

7. A management support program that enables a computer to function as a management support system for managing a manufacturing process of a functional film, which includes a first and second process in which the manufacturing units are different from each other, For each of the two or more types of the aforementioned functional film, A step of obtaining first process data relating to a first treatment in the first step corresponding to the functional film of the said type, second process data relating to a second treatment in the second step corresponding to the functional film of the said type, and quality data indicating the quality of the functional film of the said type obtained by the manufacturing process of the functional film of the said type, A step of generating total process data by relating the acquired first process data and the acquired second process data to each other, by referring to trace data showing the relationship between the first manufacturing unit in the first step and the second manufacturing unit in the second step, The steps include: generating a regression model that calculates an estimated quality of a functional film of a certain type from process data relating to the processing in the manufacturing process corresponding to that type of functional film by performing machine learning based on the aforementioned total process data and the acquired quality data; The steps include: performing an estimation process to calculate the estimated value for each of the two or more types of functional film using the regression model corresponding to that type of functional film; A management support program that causes the aforementioned computer to execute the following.

Images