Computer system, information processing method, and program

Abstract

This computer system receives blending information of raw materials, calculates a first prediction that is a result of predicting a flow characteristic of a mixture generated from a plurality of raw materials by using the blending information of the raw materials, and repeatedly executes search processing for searching for a device configuration and an operating condition of a process device that generates the mixture by performing processing including a step for mixing the plurality of raw materials. The search processing includes: processing for setting the device configuration and the operating condition of the process device; processing for calculating a second prediction that is a result of predicting a first physical quantity representing the state of the mixture in the process device on the basis of the first prediction, the device configuration of the process device, and the operating condition of the process device; processing for calculating a first correction for correcting the second prediction on the basis of the operating condition of the process device; and processing for correcting the second prediction using the first correction.

Description

Computer System, Information Processing Method, and Program Incorporation by Reference

[0001] This application claims the priority of Japanese Patent Application No. 2024-163593 filed on September 20, 2024, and incorporates its content by reference into this application.

[0002] The present invention relates to a system for predicting various physical quantities in a process device for manufacturing materials.

[0003] In the manufacture of materials such as resins, there is a process of determining the mixing ratio of raw materials input into the process device and the operating conditions of the process device, etc. in order to manufacture a material having predetermined physical properties. As a technique for reducing the number of trial and errors in this process, the technique described in Patent Document 1 is known.

[0004] Patent Document 1 describes that "the simulation device is a simulation device that simulates the behavior of an extruder, and includes an acquisition unit that acquires data related to the behavior of the extruder from a control device that controls the operation of the extruder, a sensor provided in the extruder, or an imaging device that images components constituting the extruder, and a calculation unit that calculates a physical quantity indicating the behavior of the extruder based on the acquired data."

[0005] Japanese Unexamined Patent Application Publication No. 2023-151219

[0006] In conventional simulators, there are physical phenomena that are not considered for simplification of calculations. For example, heat exchange between the process device and the outside, and heat generation due to chemical reactions of multiple raw materials. Because the physical phenomena as described above are not considered, conventional simulators have a problem of low prediction accuracy.

[0007] A typical example of the invention disclosed in this application is as follows: a computer system comprising a processor, a memory device connected to the processor, and a connection device connected to the processor, wherein the processor receives raw material blending information via the connection device, calculates a first prediction which is a prediction result of the flow characteristics of a mixture produced from a plurality of raw materials using the received raw material blending information, and repeatedly performs a search process to search for the device configuration and operating conditions of a process apparatus that produces the mixture by processing including a step of mixing the plurality of raw materials, the search process includes a first process of setting the device configuration and operating conditions of the process apparatus, a second process of calculating a second prediction which is a prediction result of a first physical quantity representing the state of the mixture in the process apparatus based on the first prediction, the device configuration of the process apparatus, and the operating conditions of the process apparatus, a third process of calculating a first correction for correcting the second prediction based on the operating conditions of the process apparatus, and a fourth process of correcting the second prediction using the first correction.

[0008] According to the present invention, the configuration and operating conditions of the process apparatus can be determined with high precision. Other problems, configurations, and effects will be clarified by the following description of the embodiments.

[0009] This figure shows an example of the system configuration of Example 1. This is a flowchart illustrating an example of the error model generation process performed by the prediction system of Example 1. This figure shows an example of the error model of Example 1. This figure shows an example of the error model of Example 1. This figure illustrates the flow of the search process performed by the prediction system of Example 1. This is a flowchart illustrating an example of the search process performed by the prediction system of Example 1. This figure shows an example of the screen displayed on the terminal of Example 1. This figure shows an example of the system configuration of Example 2. This figure illustrates the flow of the search process performed by the prediction system of Example 2. This is a flowchart illustrating an example of the search process performed by the prediction system of Example 2. This figure shows an example of the data structure of the prediction result data recorded by the prediction system of Example 2. This figure shows an example of the screen displayed on the terminal of Example 2. This figure shows an example of the heat generation model of Example 3. This figure shows an example of the heat generation model of Example 3. This figure illustrates the flow of the search process performed by the prediction system of Example 3. This is a flowchart illustrating an example of the search process performed by the prediction system of Example 3. This figure shows an example of the screen displayed on the terminal of Example 3.

[0010] The embodiments of the present invention will be described below with reference to the drawings. However, the present invention is not to be construed as being limited to the embodiments described below. It will be readily apparent to those skilled in the art that the specific configuration can be modified without departing from the spirit or intent of the present invention.

[0011] In the configuration of the invention described below, identical or similar components or functions are denoted by the same reference numerals, and redundant descriptions are omitted.

[0012] The designations "First," "Second," "Third," etc., used in this specification are for the purpose of identifying constituent elements and do not necessarily limit their number or order.

[0013] The positions, sizes, shapes, and ranges of each component shown in the drawings, etc., may not represent the actual positions, sizes, shapes, and ranges, etc., in order to facilitate understanding of the invention. Therefore, the present invention is not limited to the positions, sizes, shapes, and ranges, etc., disclosed in the drawings, etc.

[0014] Figure 1 shows an example of the system configuration of Example 1.

[0015] The system consists of a prediction system 100, a terminal 101, and a process apparatus 102. The prediction system 100 connects to the terminal 101 via a network such as a LAN (Local Area Network). The prediction system 100 also connects to a sensor 103 installed in the process apparatus 102 via the network.

[0016] The process apparatus 102 is a device for processing materials, such as an extruder. The process apparatus 102 includes a screw for kneading the materials, a heating cylinder for heating the materials, and the like. The processing performed by the process apparatus 102 includes at least a step of kneading multiple materials.

[0017] Terminal 101 is a terminal operated by a user utilizing the prediction system 100. Examples of terminals include general-purpose computers, smartphones, and tablet devices.

[0018] The prediction system 100 simulates physical quantities that represent the state of the mixture in the process apparatus 102. These physical quantities include, for example, the temperature of the mixture and the pressure acting on it. In Example 1, the temperature of the mixture in the process apparatus 102 is used as an example of a physical quantity representing the state of the mixture. The prediction system 100 simulates the temperature distribution of the mixture in the process apparatus 102. The temperature distribution of the mixture in the process apparatus 102 refers to the temperature distribution along the z-axis in Figure 1.

[0019] The prediction system 100 comprises a arithmetic unit 110, a storage device 111, and a connection device 112 as its hardware configuration. Each hardware element is connected via a bus 113.

[0020] The arithmetic unit 110 is a processor or the like, and executes programs stored in the memory device 111. By executing processing according to the program, the arithmetic unit 110 operates as a functional unit (module) that realizes a specific function. In the following description, when the processing is described with a functional unit as the subject, it indicates that the arithmetic unit 110 is executing the program that realizes that functional unit.

[0021] The storage device 111 is a memory or the like, and stores the program executed by the arithmetic unit 110 and the information used by the program. The storage device 111 is also used as a work area. The prediction system 100 may also be equipped with a large-capacity storage medium such as an HDD (Hard Disk Drive) and an SSD (Solid State Drive).

[0022] The connection device 112 is, for example, a network interface and connects to an external device.

[0023] The calculation unit 110 of Embodiment 1 functions as a search unit 120, an error model generation unit 121, a flow characteristic prediction unit 122, a state prediction unit 123, an error prediction unit 124, and a correction unit 125.

[0024] The flow characteristics prediction unit 122 uses mathematical formulas or models such as neural networks to predict physical quantities that represent the flow characteristics of a mixture produced by kneading multiple materials, based on the raw material blending information. In the following description, the physical quantities that represent the flow characteristics of the mixture will simply be referred to as "flow characteristics."

[0025] The state prediction unit 123 predicts physical quantities representing the state of the mixture in the process apparatus 102 using mathematical formulas or models such as neural networks. In Example 1, the temperature distribution of the mixture in the process apparatus 102 is predicted. In the following description, the physical quantities representing the state of the mixture in the process apparatus 102 will simply be referred to as "state".

[0026] The error model generation unit 121 generates an error model to predict the error in the state predicted by the state prediction unit 123 when physical phenomena not considered by the state prediction unit 123 are taken into account. As will be described later, the prediction system 100 corrects the state using the error. In Example 1, the temperature error distribution of the mixture in the process apparatus 102 is predicted. In the following description, the output of the error model will be referred to as the predicted error.

[0027] One physical phenomenon not considered in conventional simulators is the heat exchange between the process apparatus 102 and the outside. This physical phenomenon is correlated with the configuration of the process apparatus 102. Therefore, in Example 1, as an error model that reflects the heat exchange between the process apparatus 102 and the outside, an error model that predicts the temperature error distribution of the mixture inside the process apparatus 102 will be explained as an example.

[0028] The search unit 120 searches for a device configuration and operating conditions that satisfy the user's requirements.

[0029] The error prediction unit 124 calculates the predicted error using the error model. The correction unit 125 corrects the state output by the state prediction unit 123 using the predicted error output by the error prediction unit 124.

[0030] Furthermore, the functional units of the prediction system 100 may be combined into a single functional unit, or a single functional unit may be divided into multiple functional units according to its function.

[0031] The storage device 111 in Embodiment 1 stores the data management DB 130 and the model management DB 131.

[0032] The data management DB130 stores a dataset for each device configuration. A dataset is a collection of training data that includes operating conditions and prediction errors.

[0033] Here, the equipment configuration refers to the configuration of the processing equipment in the process apparatus 102. For example, the screw length and shape. The operating conditions refer to the operating conditions of the process apparatus 102. For example, the pressure inside the apparatus, the temperature, and the rotational speed of the screw.

[0034] The Model Management DB 131 is a database that stores various models. The Model Management DB 131 stores multiple error models associated with the device configuration.

[0035] Figure 2 is a flowchart illustrating an example of the error model generation process performed by the prediction system 100 of Example 1. Figures 3A and 3B show an example of the error model of Example 1.

[0036] When the prediction system 100 receives an execution instruction or receives training data, it executes the error model generation process.

[0037] The error model generation unit 121 starts loop processing of the device configuration (step S101). Specifically, the error model generation unit 121 sets an arbitrary device configuration. In Embodiment 1, it is assumed that the configurable device configurations are registered as patterns. It is also assumed that each device configuration is assigned an identification ID.

[0038] The error model generation unit 121 starts loop processing of operating conditions (step S102). Specifically, the error model generation unit 121 sets arbitrary operating conditions. In Embodiment 1, it is assumed that the configurable operating conditions are registered as patterns. In addition, it is assumed that the operating conditions are assigned an identification ID.

[0039] The error model generation unit 121 operates the process apparatus 102, which reflects the selected apparatus configuration and operating conditions, and measures its state using the sensor 103 (step S103). At this time, the error model generation unit 121 acquires raw material blending information via the terminal 101. The measurement results are stored in the data management DB 130.

[0040] The error model generation unit 121 acquires the state predicted by the state prediction unit 123 (step S104). Specifically, the error model generation unit 121 acquires the flow characteristics by inputting the raw material blending information to the flow characteristics prediction unit 122. The error model generation unit 121 also acquires the state by inputting the flow characteristics, the selected equipment configuration, and the selected operating conditions to the state prediction unit 123.

[0041] The error model generation unit 121 calculates the state error acquired in the processes of step S103 and step S104 (step S105). The error model generation unit 121 stores the learning data in the data management DB130.

[0042] The error model generation unit 121 determines whether to end the loop process of the operation conditions (step S106). For example, when the processing for all selectable operation conditions is completed, the error model generation unit 121 determines to end the loop process of the operation conditions.

[0043] When not ending the loop process of the operation conditions, the error model generation unit 121 selects a new operation condition and returns to step S103.

[0044] When ending the loop process of the operation conditions, the error model generation unit 121 generates an error model f ｋ (p) (step S107).

[0045] Here, k represents the device configuration, and p represents the operation condition. The error model f ｋ (p) is a model with the operation condition as a parameter, and outputs the distribution of the error of the temperature of the mixture in the z-axis direction. The error model f ｋ (p) may be generated using the method of multiple regression analysis or may be generated using the method of machine learning. When the error model f ｋ (p) is a function, a group of functions as shown in FIGS. 3A and 3B is stored in the model management DB131.

[0046] The error model generation unit 121 determines whether to end the loop process of the device configuration (step S108). For example, when the processing for all selectable device configurations is completed, the error model generation unit 121 determines to end the loop process of the device configuration.

[0047] When not ending the loop process of the device configuration, the error model generation unit 121 selects a new device configuration and returns to step S102.

[0048] When ending the loop process of the device configuration, the error model generation unit 121 ends the error model generation process.

[0049] Figure 4 is a diagram illustrating the flow of the search process performed by the prediction system 100 of Example 1. Figure 5 is a flowchart illustrating an example of the search process performed by the prediction system 100 of Example 1. Figure 6 is a diagram showing an example of the screen displayed on the terminal 101 of Example 1.

[0050] The user operates terminal 101 to access the prediction system 100 in order to search for the device configuration and operating conditions that will achieve a predetermined state.

[0051] When the prediction system 100 receives an access request from terminal 101, it displays screen 600 on terminal 101. Screen 600 includes input fields 601, 602, 603, buttons 604, 605, 606, and a display field 607.

[0052] Input field 601 is for entering raw material blending information. Input field 602 is for entering the initial equipment configuration. Input field 602 may display selectable equipment configurations as a dropdown list. Input field 603 is for entering the initial operating conditions. Input field 603 may display selectable operating conditions as a dropdown list. A checkbox may be provided to prohibit changes to the equipment configuration and operating conditions during the search process.

[0053] Button 604 is for sending a search request. Button 605 is for sending a retry request. Button 606 is for sending an output request for the searched device configuration and operating conditions. Display field 607 is for displaying the simulation results of physical quantities.

[0054] The user enters information into input fields 601, 602, and 603 and presses button 604. Terminal 101 sends a search request to prediction system 100, which includes the raw material blending information, initial equipment configuration, and initial operating conditions entered into input fields 601, 602, and 603.

[0055] Note that input fields 602 and 603 may be left blank. In this case, the prediction system 100 sets the initial device configuration and initial operating conditions.

[0056] The search unit 120 acquires the raw material blending information, initial equipment configuration, and initial operating conditions included in the search request (step S201). If the search request does not include the initial equipment configuration and initial operating conditions, the search unit 120 sets the initial equipment configuration and initial operating conditions.

[0057] The search unit 120 calculates the flow characteristics by inputting the raw material blending information into the flow characteristics prediction unit 122 (step S202).

[0058] The search unit 120 calculates the state by inputting the flow characteristics, equipment configuration, and operating conditions to the state prediction unit 123 (step S203). In Example 1, the temperature at each position of the process apparatus 102 is calculated.

[0059] The search unit 120 calculates the predicted error by inputting the device configuration and operating conditions to the error prediction unit 124 (step S204). In Example 1, the temperature error at each position of the process apparatus 102 is calculated.

[0060] The error prediction unit 124, upon receiving input from the search unit 120, performs the following processing: The error prediction unit 124 obtains an error model corresponding to the device configuration from the model management DB 131. The error prediction unit 124 calculates the predicted error by inputting the operating conditions into the error model.

[0061] The search unit 120 corrects the state by inputting the state and prediction error to the correction unit 125 (step S205). The correction unit 125, for example, adds the error to the temperature at each position of the process apparatus 102 output by the state prediction unit 123.

[0062] The search unit 120 outputs the corrected state to the terminal 101 (step S206). At this time, the search unit 120 may also output the state before correction.

[0063] Terminal 101 displays the corrected state in the display area 607 of screen 600. If the pre-correction state has been received, terminal 101 also displays the pre-correction state in the display area 607. The user refers to the display area 607 and determines whether the corrected state meets predetermined conditions. If the predetermined conditions are not met, the user presses button 605; if the predetermined conditions are met, the user presses button 606.

[0064] When the search unit 120 receives a request from the terminal 101, it determines whether or not the request is an output request (step S207).

[0065] If the received request is a request for re-execution, the search unit 120 changes at least one of the device configuration and operating conditions (step S208), and then returns to step S203.

[0066] Since changing the device configuration requires considerable effort, in Embodiment 1, the search unit 120 prioritizes changing the operating conditions, and only changes the device configuration if changing the operating conditions is insufficient.

[0067] If the received request is an output request, the search unit 120 outputs the current device configuration and configuration conditions to the terminal 101 (step S209), and then terminates the search process.

[0068] The user may pre-set ideal statistical values, and the search unit 120 may output the device configuration and operating conditions if the statistical values ​​calculated from the physical quantities and prediction errors at each location are smaller than the ideal statistical values. The statistical values ​​may be, for example, the sum, mean, and variance. Alternatively, the user may pre-set ideal values ​​for the physical quantities at each location, and the search unit 120 may output the device configuration and operating conditions after processing in step S205 if the corrected physical quantities at all locations match the ideal values.

[0069] According to Example 1, the prediction system 100 can determine the device configuration and operating conditions with high accuracy by correcting the predicted state using an error model.

[0070] In Example 2, the prediction system 100 predicts the properties of the material produced by the process apparatus 102, and determines the apparatus configuration and operating conditions based on the results of the prediction. The following describes Example 2, focusing on the differences from Example 1.

[0071] Figure 7 shows an example of the system configuration of Example 2.

[0072] The system configuration of Example 2 is the same as that of Example 1. In Example 2, the functional configuration of the prediction system 100 is different. Specifically, the prediction system 100 has a material property prediction unit 126. The material property prediction unit 126 receives physical quantities representing the state of the mixture as input and uses a mathematical formula or a model such as a neural network to predict physical quantities representing the properties of the material produced by the process apparatus 102. The model used by the material property prediction unit 126 is stored in the model management DB 131. Examples of physical quantities representing material properties include strength. In the following description, physical quantities representing material properties will simply be referred to as material properties.

[0073] Figure 8 is a diagram illustrating the flow of the search process performed by the prediction system 100 of Example 2. Figure 9 is a flowchart illustrating an example of the search process performed by the prediction system 100 of Example 2. Figure 10 is a diagram illustrating an example of the data structure of the prediction result data recorded by the prediction system 100 of Example 2. Figure 11 is a diagram illustrating an example of the screen displayed on the terminal 101 of Example 2.

[0074] The user operates terminal 101 to access the prediction system 100 in order to search for an apparatus configuration and operating conditions that will achieve a predetermined material property.

[0075] When the prediction system 100 receives an access request from terminal 101, it displays screen 1100 on terminal 101. Screen 1100 includes input fields 1101, 1102, 1103, buttons 1104, 1105, 1106, 1108, selection field 1107, and display field 1109.

[0076] Input fields 1101, 1102, and 1103 are the same as input fields 601, 602, and 603. Buttons 1104, 1105, and 1106 are the same as buttons 604, 605, and 606.

[0077] The selection field 1107 is for selecting operating conditions. The button 1108 is for sending a display request. The display field 1109 is for displaying the results of the processing performed by the prediction system 100.

[0078] The user enters information into input fields 1101, 1102, and 1103 and presses button 1104. Terminal 101 sends a search request to prediction system 100, which includes the raw material blending information, initial equipment configuration, and initial operating conditions entered into input fields 1101, 1102, and 1103.

[0079] Note that input fields 1102 and 1103 may be left blank. In this case, the prediction system 100 will set the initial equipment configuration and initial operating conditions.

[0080] The search unit 120 acquires the raw material composition information, initial equipment configuration, and initial operating conditions included in the search request (step S201).

[0081] The search unit 120 calculates the flow characteristics by inputting the raw material blending information into the flow characteristics prediction unit 122 (step S202).

[0082] The search unit 120 calculates the state by inputting the flow characteristics, equipment configuration, and operating conditions to the state prediction unit 123 (step S203).

[0083] The search unit 120 calculates the predicted error by inputting the device configuration and operating conditions to the error prediction unit 124 (step S204).

[0084] The search unit 120 corrects the state by inputting the state and prediction error to the correction unit 125 (step S205).

[0085] The search unit 120 calculates the material properties by inputting the corrected state to the material property prediction unit 126 (step S251).

[0086] At this time, the search unit 120 generates prediction result data 1000 with a data structure as shown in Figure 10, and stores the prediction result data 1000 in the data management DB 130.

[0087] The prediction result data 1000 includes ID 1001, device configuration 1002, operating conditions 1003, state 1004, and material properties 1005.

[0088] ID 1001 is a field that stores the ID of the prediction result data 1000. Device configuration 1002 is a field that stores the device configuration. Operating conditions 1003 is a field that stores the operating conditions. State 1004 is a field that stores the corrected state. Material properties 1005 is a field that stores the material properties.

[0089] The search unit 120 outputs prediction result data 1000 to the terminal 101 (step S252).

[0090] The user sets the operating conditions in the selection field 1107 and presses the button 1108. The terminal 101 displays the prediction results in the display field 1109 based on the prediction result data 1000 in which the selected operating conditions 1003 are set. Figure 6 shows the mapping result of the prediction results to the feature space with state and material properties as parameters.

[0091] If a desired prediction result exists, the user selects one of the prediction results displayed in the display field 1109 and presses button 1106. If no desired prediction result exists, the user may change the operating conditions in the selection field 1107 and press button 1108, or press button 1105.

[0092] When the search unit 120 receives a request from the terminal 101, it determines whether or not the request is an output request (step S207).

[0093] If the received request is a request for re-execution, the search unit 120 changes at least one of the device configuration and operating conditions (step S208), and then returns to step S203.

[0094] If the received request is an output request, the search unit 120 outputs the device configuration and configuration conditions selected by the user to the terminal 101 (step S253), and then terminates the search process.

[0095] The user may pre-set ideal values ​​for material properties, and after processing in step S251, the search unit 120 may output the device configuration and operating conditions if the material properties match the ideal values, or if the error between the material properties and the ideal values ​​is smaller than a threshold.

[0096] According to Example 2, the prediction system 100 can determine the equipment configuration and operating conditions based on the material properties predicted using the corrected state. This allows for the determination of the equipment configuration and operating conditions with higher accuracy than conventional methods.

[0097] Example 3 differs from Example 1 in that it uses multiple error models that take into account different physical phenomena, and measures obtained from the sensor 103 while operating the process apparatus 102. Below, Example 3 will be described focusing on the differences from Example 1.

[0098] The system configuration of Example 3 is the same as that of Example 1.

[0099] In Example 1, an error model was used that took into account heat exchange between the process apparatus 102 and the outside. In Example 3, the aforementioned error model and an error model (exothermic model) that takes into account the heat generated by the chemical reaction due to the mixing of materials are used.

[0100] Figures 12A and 12B show an example of the exothermic model for Example 3. As shown in Figures 12A and 12B, the exothermic model is a graph representing the distribution of heat generation from the mixture in the process apparatus 102. Multiple exothermic models exist for each apparatus configuration.

[0101] In Example 3, multiple exothermic models are stored in the model management DB 131. Since the exothermic reaction caused by mixing materials does not correlate with the apparatus configuration, the prediction system 100 selects the exothermic model to use based on measured values.

[0102] Figure 13 is a diagram illustrating the flow of the search process performed by the prediction system 100 of Example 3. Figure 14 is a flowchart illustrating an example of the search process performed by the prediction system 100 of Example 3. Figure 15 is a diagram showing an example of the screen displayed on the terminal 101 of Example 3.

[0103] The user operates terminal 101 to access the prediction system 100 in order to search for the device configuration and operating conditions that will achieve a predetermined state.

[0104] When the prediction system 100 receives an access request from the terminal 101, it displays screen 600 on the terminal 101. Since screen 600 is the same as in Embodiment 1, its description is omitted. The display area 607 shows the prediction results so that the physical quantities at the location of the sensor 103 can be understood.

[0105] The search unit 120 acquires the raw material composition information, initial equipment configuration, and initial operating conditions included in the search request (step S201).

[0106] The search unit 120 operates the process apparatus 102, which reflects the apparatus configuration and operating conditions, and acquires measured values ​​(temperature) from the sensor 103 (step S261). The measured values ​​also acquire information indicating the installation location of the sensor 103, i.e., the location of the process apparatus 102.

[0107] The search unit 120 calculates the flow characteristics by inputting the raw material blending information into the flow characteristics prediction unit 122 (step S202).

[0108] The search unit 120 calculates the state by inputting the flow characteristics, equipment configuration, and operating conditions to the state prediction unit 123 (step S203).

[0109] The search unit 120 calculates the predicted error by inputting the measured value, physical quantity, device configuration, and operating conditions to the error prediction unit 124 (step S262). The processing performed by the error prediction unit 124 in Example 3 differs in part from that in Example 1.

[0110] (Procedure 1) The error prediction unit 124 obtains an error model corresponding to the device configuration from the model management DB 131. The error prediction unit 124 calculates the error of a physical quantity (first predicted error) by inputting the operating conditions into the error model.

[0111] (Procedure 2) The error prediction unit 124 calculates the error (second prediction error) of the physical quantity corresponding to the position of the sensor 103 for each heat generation model. That is, the error prediction unit 124 reads out the value corresponding to the position of the sensor 103 on the graph.

[0112] (Step 3) The error prediction unit 124 corrects the physical quantity at the position of the sensor 103 using the physical quantity, the first prediction error, and the second prediction error. For example, the error prediction unit 124 adds the first prediction error and the second prediction error to the physical quantity. The error prediction unit 124 identifies a heat generation model in which the error between the corrected mixed physical quantity and the measured value is small.

[0113] (Step 4) The error prediction unit 124 calculates a second prediction error using the identified heat generation model.

[0114] (Procedure 5) The error prediction unit 124 outputs the first predicted error and the second predicted error.

[0115] The search unit 120 corrects the state by inputting a physical quantity, a first prediction error, and a second prediction error to the correction unit 125 (step S263). The correction method is the same as in (step 3).

[0116] The search unit 120 outputs the corrected state to the terminal 101 (step S206).

[0117] When the search unit 120 receives a request from the terminal 101, it determines whether or not the request is an output request (step S207).

[0118] If the received request is a request for re-execution, the search unit 120 changes at least one of the device configuration and operating conditions (step S208), and then returns to step S203.

[0119] If the received request is an output request, the search unit 120 outputs the current device configuration and configuration conditions to the terminal 101 (step S209), and then terminates the search process.

[0120] The user may pre-set ideal statistical values, and the search unit 120 may output the device configuration and operating conditions if the statistical values ​​calculated from the physical quantities and prediction errors at each location are smaller than the ideal statistical values. The statistical values ​​may be, for example, the sum, mean, and variance. Alternatively, the user may pre-set ideal values ​​for the physical quantities at each location, and the search unit 120 may output the device configuration and operating conditions after processing in step S205 if the corrected physical quantities at all locations match the ideal values.

[0121] The search unit 120 may select multiple heat generation models and display multiple correction results corrected using each heat generation model.

[0122] According to Example 3, the prediction system 100 can determine the device configuration and operating conditions with high accuracy by correcting physical quantities using multiple error models.

[0123] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are included. Furthermore, for example, the embodiments described above are detailed explanations of the configuration in order to clearly illustrate the present invention, and are not necessarily limited to those having all the configurations described. In addition, some of the configurations in each embodiment can be added to, deleted from, or replaced with other configurations.

[0124] Furthermore, each of the above-mentioned configurations, functions, processing units, processing means, etc., may be implemented in hardware, either partially or entirely, by designing them as integrated circuits, for example. The present invention can also be implemented by software program code that realizes the functions of the embodiment. In this case, a storage medium on which the program code is recorded is provided to a computer, and the processor of that computer reads the program code stored in the storage medium. In this case, the program code read from the storage medium itself realizes the functions of the embodiment described above, and the program code itself and the storage medium on which it is stored constitute the present invention. Examples of storage media used to supply such program code include flexible disks, CD-ROMs, DVD-ROMs, hard disks, SSDs (Solid State Drives), optical disks, magneto-optical disks, CD-Rs, magnetic tapes, non-volatile memory cards, ROMs, and the like.

[0125] Furthermore, the program code that implements the functions described in this embodiment can be implemented in a wide range of programming or scripting languages, such as assembler, C / C++, Perl, Shell, PHP, Python, and Java.

[0126] Furthermore, the program code for the software that implements the functions of the embodiment may be distributed via a network and stored in a storage means such as a computer's hard disk or memory, or in a storage medium such as a CD-RW or CD-R, and the computer's processor may read and execute the program code stored in the storage means or storage medium.

[0127] In the above-described embodiment, the control lines and information lines shown are those deemed necessary for explanation and do not necessarily represent all control lines and information lines in the actual product. All components may be interconnected.

Claims

1. A computer system comprising a processor, a storage device connected to the processor, and a connection device connected to the processor, wherein the processor receives raw material blending information via the connection device, calculates a first prediction which is a prediction result of the flow characteristics of a mixture produced from a plurality of raw materials using the received raw material blending information, and repeatedly performs a search process to search for the device configuration and operating conditions of a process apparatus that produces the mixture by performing a process including a step of mixing a plurality of materials, wherein the search process includes: a first process of setting the device configuration and operating conditions of the process apparatus; a second process of calculating a second prediction which is a prediction result of a first physical quantity representing the state of the mixture in the process apparatus based on the first prediction, the device configuration of the process apparatus, and the operating conditions of the process apparatus; a third process of calculating a first correction for correcting the second prediction based on the operating conditions of the process apparatus; and a fourth process of correcting the second prediction using the first correction.

2. A computer system according to claim 1, wherein the storage device stores: a first model for calculating the first prediction from raw material blending information; a second model for calculating the second prediction from flow characteristics, the device configuration of the process apparatus, and the operating conditions of the process apparatus; and a third model for calculating the error of the second prediction as the first correction when physical phenomena not considered in the second model are taken into account from the operating conditions of the process apparatus; the processor calculates the first prediction using the received raw material blending information and the first model; the processor calculates the second prediction in the second process using the first prediction, the device configuration of the process apparatus, the operating conditions of the process apparatus, and the second model; and the processor calculates the first correction in the third process using the operating conditions of the process apparatus and the third model.

3. A computer system according to claim 2, wherein the first prediction is the distribution of the first physical quantity at the location of the process apparatus, the first correction is the distribution of errors of the first physical quantity at the location of the process apparatus, and the processor receives ideal statistical values ​​via the connection device, and if the statistical values ​​calculated using the first physical quantity at each location of the process apparatus included in the second prediction and the errors of the first physical quantity at each location of the process apparatus included in the first correction are smaller than the ideal statistical values, the computer system terminates the search process and outputs the apparatus configuration of the process apparatus and the operating conditions of the process apparatus set in the search process.

4. A computer system according to claim 3, wherein the system is connected to a sensor that measures the first physical quantity at an arbitrary position of the process apparatus, the storage device stores a plurality of fourth models for calculating a second correction to correct the second prediction, the fourth model is a graph representing the distribution of errors of the first physical quantity at the position of the process apparatus when considering physical phenomena that are not considered in the second model and different from the third model, the search process includes a process of operating the process apparatus that reflects a set apparatus configuration and set operating conditions of the process apparatus and acquiring the first physical quantity from the sensor, and the processor, in the fourth process, corrects the first physical quantity at the position where the sensor is installed, which is included in the second prediction, using the first correction and the second correction calculated using an arbitrary fourth model, identifies a fourth model in which the error between the corrected first physical quantity and the first physical quantity acquired from the sensor is small, and corrects the second prediction using the first correction and the second correction calculated using the identified fourth model.

5. A computer system according to claim 2, wherein the storage device stores a fifth model for calculating a fourth prediction, which is a prediction result of a second physical quantity representing the properties of the mixture; the fifth model is a model for calculating the fourth prediction from the second prediction; the search process includes a process for calculating the fourth prediction by inputting the corrected second prediction into the fifth model; the processor receives a second ideal value of the properties of the material via the connection device; the search process is terminated if the fourth prediction matches the second ideal value, or if the error between the fourth prediction and the second ideal value is less than a threshold; and the system outputs the device configuration of the process apparatus and the operating conditions of the process apparatus set in the search process.

6. A computer system according to claim 2, wherein the storage device stores a plurality of the third models, the third models are managed in association with the device configuration of the process apparatus, and the processor, in the third process, obtains the third model corresponding to the set device configuration of the process apparatus from the storage device.

7. A computer system according to claim 2, wherein the process apparatus is an extruder.

8. A computer system according to claim 2, characterized in that the first physical quantity is at least one of temperature and pressure.

9. A computer system according to claim 2, wherein the storage device stores learning data including the operating conditions of the process apparatus and the first correction, and the processor generates the third model by performing regression analysis or machine learning using the learning data.

10. An information processing method to be executed by a computer system, wherein the computer system comprises a processor, a storage device connected to the processor, and a connection device connected to the processor, wherein the storage device stores: a first model for calculating a first prediction, which is a prediction result of the flow characteristics of a mixture produced from a plurality of raw materials; a second model for calculating a second prediction, which is a prediction result of a first physical quantity representing the state of the mixture in a process apparatus that produces the mixture by processing including a step of mixing the plurality of raw materials; and a third model for calculating a first correction for correcting the second prediction, wherein the first model is a model for calculating the first prediction from raw material blending information; the second model is a model for calculating the second prediction from the flow characteristics, the apparatus configuration of the process apparatus, and the operating conditions of the process apparatus; the third model is a model for calculating the error of the second prediction as the first correction when considering physical phenomena not considered in the second model, based on the operating conditions of the process apparatus, and the information processing method comprises: the step of the processor receiving raw material blending information via the connection device; and the step of the processor calculating the first prediction using the received raw material blending information and the first model. Information processing method comprising: a step of the processor repeatedly performing a search process for searching for the device configuration of the process apparatus and the operating conditions of the process apparatus, wherein the search process includes: a step of the processor setting the device configuration of the process apparatus and the operating conditions of the process apparatus; a step of the processor calculating the second prediction using the first prediction, the device configuration of the process apparatus, the operating conditions of the process apparatus, and the second model; a step of the processor calculating the first correction using the operating conditions of the process apparatus and the third model; and a step of the processor correcting the second prediction using the first correction.

11. A program to be executed on a computer having a processor, a storage device connected to the processor, and a connection device connected to the processor, wherein the storage device stores: a first model for calculating a first prediction, which is a prediction result of the flow characteristics of a mixture produced from a plurality of raw materials; a second model for calculating a second prediction, which is a prediction result of a first physical quantity representing the state of the mixture in a process apparatus that produces the mixture by processing including a step of mixing the plurality of raw materials; and a third model for calculating a first correction for correcting the second prediction, wherein the first model is a model for calculating the first prediction from raw material blending information; the second model is a model for calculating the second prediction from the flow characteristics, the apparatus configuration of the process apparatus, and the operating conditions of the process apparatus; the third model is a model for calculating the error of the second prediction as the first correction when considering physical phenomena not considered in the second model, based on the operating conditions of the process apparatus; the program comprises: a procedure for receiving the raw material blending information via the connection device; and a procedure for the processor to calculate the first prediction using the received raw material blending information and the first model. A program characterized in that the processor causes the computer to execute a procedure for repeatedly performing a search process to search for the device configuration and operating conditions of the process apparatus, the search process comprising: a procedure for setting the device configuration and operating conditions of the process apparatus; a procedure for calculating the second prediction using the first prediction, the device configuration of the process apparatus, the operating conditions of the process apparatus, and the second model; a procedure for calculating the first correction using the operating conditions of the process apparatus and the third model; and a procedure for correcting the second prediction using the first correction.

Images