JP2023176035A - Process model construction system and process model construction method - Google Patents

Abstract

An object of the present invention is to construct a process model that requires less construction and operation load than conventional methods even when a controlled object has multiple characteristics. [Solution] A process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object, wherein the first system acquires characteristics from past operation performance data that indicates operational performance according to the characteristics of the controlled object. The processing unit, the actual value of the process to be controlled regarding the characteristics included in the past operation performance data, and the estimated value of the process to be controlled estimated by a process model constructed using a predetermined theoretical formula are set under predetermined conditions. and a second processing unit that calculates characteristic parameters for correcting the process model for each characteristic so as to satisfy the following. [Selection diagram] Figure 1

Description

The present invention relates to a process model construction system and a process model construction method.

In process operation sites, although the optimal operating conditions are known from past operating results, there are situations in which the operating method to reproduce those operating conditions is unknown. This is because various disturbances differ depending on the operation, such as reaction heat generated within the control target including control equipment such as valves, deteriorated characteristics of control equipment, and external environmental conditions that cannot be obtained as input. This is because the driving method to achieve the same driving conditions is different each time.

An existing technology for this purpose is MPC (Model Predictive Control). MPC makes it possible to predict the control amount with respect to the manipulated variable of the controlled object by modeling the characteristics of the controlled object, and thereby makes it possible to back-calculate the manipulated variable that satisfies a desired control waveform. As such existing technology, for example, there is Patent Document 1. In Patent Document 1, a first-principles model is used to simulate a batch process, and this first-principles model is used to configure a multiple-input/multiple-output control routine to control the batch process. It can be done. First principles models can generate estimates of batch variables that cannot or are not measured during actual batch process operation. An example of such a variable may be the rate of change of a component of a batch process (eg, production rate, cell growth rate, etc.). It is stated that the first principles model and its constructed multiple input/multiple output control routines can be used to facilitate control of batch processes.

JP2012-64245A

In Patent Document 1, by adding a mechanism (soft sensor) that estimates difficult-to-measure elements that exist during a batch process using a first-principles model, an appropriate operating method can be determined based on the estimated values of the first-principles model. By using the estimated values as control targets, we have realized operations that shorten batch cycles and improve yields. However, in order to control objects with different characteristics, it is necessary to manually construct different first-principles models, which makes practical application difficult in multi-product production sites. For example, if there are multiple combinations of the characteristics of the product manufactured by the controlled object and the characteristics of the control equipment to be controlled, it is necessary to construct a process model for each combination. Particularly, in a multi-product production site such as a batch process, there are many combinations of products, equipment, etc., and the number of process models to be constructed is enormous. It is difficult to construct a huge number of process models, and maintenance after construction becomes complicated, resulting in a large operational burden. Therefore, even when a controlled object has multiple characteristics, there is a need for a technique for constructing a process model that requires less construction and operational load than conventional methods.

One aspect of the present invention aims to provide a technique for constructing a process model that requires less construction and operation load than conventional methods even when a controlled object has multiple characteristics.

A process model construction system according to the present invention is a process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object, and includes past operation performance data that indicates operational performance according to the characteristics of the controlled object. a first processing unit that acquires the characteristics from the above, a performance value of the process to be controlled with respect to the characteristics included in the past operation performance data, and a process model constructed using a predetermined theoretical formula. A second processing unit that calculates characteristic parameters for correcting the process model such that the estimated value of the controlled process satisfies a predetermined condition for each of the characteristics. It is configured as a process model construction system.

According to one aspect of the present invention, even when a controlled object has a plurality of characteristics, it is possible to construct a process model that requires less construction and operation load than conventional methods.

1 is a diagram showing an example of the configuration of a process model construction system in Example 1. FIG. FIG. 1 is a diagram schematically showing a computer. FIG. 3 is a diagram showing an example of past operation performance data. FIG. 2 is a diagram illustrating an example of a process model and an example of the relationship between the process model and characteristics. FIG. 3 is a diagram showing an example of a characteristic parameter group. FIG. 3 is a diagram showing an example of characteristic information. 5 is a diagram illustrating an example of a flowchart illustrating a procedure of model construction processing performed by a model construction unit in Example 1. FIG. 8 is a diagram illustrating an example of a flowchart illustrating the processing procedure of characteristic parameter calculation processing performed in S706 illustrated in FIG. 7. FIG. 8 is a diagram illustrating an example of a flowchart illustrating the processing procedure of characteristic parameter calculation processing performed in S706 illustrated in FIG. 7. FIG. 8A and 8B are diagrams illustrating an example of a flowchart illustrating a processing procedure of a model identification process for a process model recorded in a model identification data group in the characteristic parameter calculation process shown in FIGS. 8A and 8B. FIG. 8A and 8B are diagrams illustrating an example of a flowchart illustrating a processing procedure of a model identification process for a process model recorded in a model identification data group in the characteristic parameter calculation process shown in FIGS. 8A and 8B. FIG. 5 is a diagram illustrating an example of a flowchart illustrating a processing procedure of a simulation process performed by a simulation unit in Example 1. FIG. 3 is a diagram illustrating an example of the configuration of a process model construction system in Example 2. FIG. FIG. 7 is a diagram illustrating an example of the configuration of a process model construction system in Example 3. It is a figure which shows an example of additional characteristic information. FIG. 7 is a diagram illustrating an example of the configuration of a process model construction system in Example 4. It is a figure which shows an example of model identification information. FIG. 7 is a diagram illustrating an example of the configuration of a process model construction system in Example 5. 12 is a diagram illustrating an example of a flowchart illustrating a processing procedure of simulation processing performed by a simulation unit in Example 5. FIG. FIG. 7 is a diagram illustrating an example of the configuration of a process model construction system in Example 6. FIG. 12 is a diagram illustrating an example of a flowchart illustrating a procedure of model construction processing performed by a model construction unit in Example 6; 18 is a diagram illustrating an example of a flowchart illustrating the processing procedure of characteristic model calculation processing performed in S1706 illustrated in FIG. 17. FIG. 18 is a diagram illustrating an example of a flowchart illustrating the processing procedure of characteristic model calculation processing performed in S1706 illustrated in FIG. 17. FIG. FIG. 2 is a diagram showing an example of a screen (model construction screen) displayed on a display of a computer that constitutes a process model construction system when a model construction unit executes a model construction process. FIG. 2 is a diagram showing an example of a screen (simulation execution screen) displayed on a display of a computer forming the process model construction system when the simulation unit executes simulation processing. FIG. 2 is a diagram illustrating an example of a screen (feedback execution screen) displayed on the display of a computer that constitutes the process model construction system when the control equipment input calculation unit back-calculates the operating method of the controlled object that satisfies the optimal operating conditions.

Embodiments of the present invention will be described below with reference to the drawings. The following description and drawings are examples for explaining the present invention, and are omitted and simplified as appropriate for clarity of explanation. The present invention can also be implemented in various other forms. Unless specifically limited, each component may be singular or plural.

The position, size, shape, range, etc. of each component shown in the drawings may not represent the actual position, size, shape, range, etc. in order to facilitate understanding of the invention. Therefore, the present invention is not necessarily limited to the position, size, shape, range, etc. disclosed in the drawings.

In the following description, various information may be described using expressions such as "database," "table," and "list," but various information may be expressed using data structures other than these. "XX table", "XX list", etc. are sometimes referred to as "XX information" to indicate that they do not depend on the data structure. When describing identification information, when expressions such as "identification information", "identifier", "name", "ID", and "number" are used, these expressions can be replaced with each other.

When there are multiple components having the same or similar functions, the same reference numerals may be given different suffixes for explanation. However, if there is no need to distinguish between these multiple components, the subscripts may be omitted in the description.

In addition, in the following explanation, processing performed by executing a program may be explained, but the program is executed by a processor (for example, a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit)). The processor may be the main body of the processing in order to perform the processing using appropriate storage resources (for example, memory) and/or interface devices (for example, communication ports). Similarly, the subject of processing performed by executing a program may be a controller, device, system, computer, or node having a processor. The main body of the processing performed by executing the program may be an arithmetic unit, and may include a dedicated circuit (for example, FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)) that performs specific processing. .

A program may be installed on a device, such as a computer, from a program source. The program source may be, for example, a program distribution server or a computer-readable storage medium. When the program source is a program distribution server, the program distribution server includes a processor and a storage resource for storing the program to be distributed, and the processor of the program distribution server may distribute the program to be distributed to other computers. Furthermore, in the following description, two or more programs may be realized as one program, or one program may be realized as two or more programs.

The configuration of an embodiment of the process model construction system and process model construction method according to this embodiment will be described below.

FIG. 1 is a diagram illustrating an example of the configuration of a process model construction system according to a first embodiment. As shown in FIG. 1, the process model construction system 1000 according to this embodiment includes a model construction section 1001, a simulation section 1002, a characteristic parameter group 1003, and a process model 1004.

The process model construction system 1000 is suitable for control objects that have many combinations of characteristics by setting the parameters of a process model built using theoretical formulas according to the characteristics of the process model based on past operation performance data of the control object. This is a system for building process models. This enables process simulation using a small number of process models without creating a process model for each combination of characteristics. A controlled object is an object controlled by this system, such as a control device that performs process operation. Although the following explanation assumes that the control target is equipment such as a reactor used in a plant, the invention may be applied to various equipment and devices other than plants that perform process operations. Further, although the description assumes that the characteristics of the controlled object are the type of equipment and the type of product manufactured by the controlled object, other information representing the characteristics of the controlled object may be determined as the characteristic.

Characteristics are elements that affect the behavior of a process model, such as the type of product manufactured by a controlled object performing process operation and the equipment that performs the process operation. The characteristic type represents, for example, the type of element included in a certain type of characteristic. The combination of characteristics refers to, for example, a combination of multiple types of characteristics.

The model construction unit 1001 calculates characteristic parameters (described later) for correcting the parameters of the process model for controlling the controlled object that performs the process operation, based on the past operation performance data 101 and the prepared process model 1004, The calculated characteristic parameters are stored in the characteristic parameter group 1003.

The simulation unit 1002 corrects the parameters of the process model 1004 using the correction parameters stored in the characteristic parameter group 1003 based on the input characteristic information 102. The simulation unit 1002 simulates the behavior of the process performed by the controlled object based on the simulation input information 103 and stores the result in the simulation result 104. The simulation input information 103 is information that defines a simulation method and conditions (for example, user-desired operating conditions), and includes, for example, a reference value that represents a set value for each time set for a controlled object.

The characteristic parameter group 1003 stores one or more characteristic parameters. The characteristic parameter group 1003 is a table that stores a set of correction parameters for correcting a process model indicating the characteristics of a controlled object. Here, as the characteristics of the controlled object, two characteristic parameters are stored in the characteristic parameter group 1003: a characteristic parameter indicating the type of product manufactured by the controlled object, and a characteristic parameter indicating the equipment to be controlled. exemplify. As mentioned above, the number and types of characteristic parameters may be determined as appropriate depending on the environment to which this system is applied.

The process model 1004 is a model of the characteristics of the controlled object, and is a model constructed using a predetermined theoretical formula. Although the process model 1004 can be expressed using various formulas, a model expressed by y=ax ² +bx+c will be described below as an example. In this case, a, b, and c are the parameters of the model, x is the input value of the model, and y is the output value of the model. Specific functions of the process model construction system 1000 will be described later using flowcharts and the like. Further, in this example, a case is illustrated in which the process model 1004 is represented by one logical formula related to a mathematical model, but it may be composed of a plurality of logical formulas, and furthermore, one or more of these logical formulas There may be multiple inputs to the formula or multiple outputs from the logical formula.

The process model construction system 1000 shown in FIG. 1 includes, for example, a CPU 1601, a memory 1602, an external storage device 1603 such as an HDD (Hard Disk Drive), and a CD (Compact) as shown in FIG. 2 (computer schematic diagram). A reading device 1607 that reads and writes information to and from a portable storage medium 1608 such as a disk or USB memory, an input device 1606 such as a scanner, keyboard, and mouse, and an output device 1605 such as a display are connected to a communication network. This can be realized by a general computer 1600 that includes a communication device 1604 such as a NIC (Network Interface Card) for the purpose of communication, and an internal communication line (referred to as a system bus) 1609 such as a system bus that connects these devices.

Various data (for example, characteristic parameter group 1003, process model 1004) stored in the process model construction system 1000 or used for processing can be realized by the CPU 1601 reading it from the memory 1602 or the external storage device 1603 and using it. be. In addition, each functional unit (for example, model construction unit 1001, simulation unit 1020) of each system or device is controlled by the CPU 1601 loading a predetermined program stored in the external storage device 1603 into the memory 1602 and executing it. It is possible.

The above-mentioned predetermined program is stored (downloaded) in the external storage device 1603 from the storage medium 1608 via the reading device 1607 or from the network via the communication device 1604, and then loaded onto the memory 1602. It may also be executed by the CPU 1601. Alternatively, the program may be directly loaded onto the memory 1602 from the storage medium 1608 via the reading device 1607 or from a network via the communication device 1604 and executed by the CPU 1601.

In the following, a case will be exemplified in which the process model construction system 1000 is configured by one computer, but all or part of these functions are distributed and provided in one or more computers such as a cloud, Similar functions may be realized by communicating with each other via a network. The specific processing performed by each unit constituting the process model construction system 1000 will be described later using a flowchart.

FIG. 3 is a diagram showing an example of past operation performance data 101. Past operation performance data 101 is data indicating past operation performance obtained from a controlled object that performs process operation, and is input to the model construction unit 1001. As shown in FIG. 3, past operation performance data 101 stores a data ID for identifying data, a product type and equipment indicating characteristics, and operation performance data in association with each other. The operation performance data is, for example, operation and measurement information of the controlled objects constituting the plant, and stores, for example, time series data including reference values, manipulated variables, and controlled variables. The reference value is information indicating a set value (in this example, a temperature set value) at each time set for a controlled object (for example, a reactor). The manipulated variable is information indicating the operating method of the controlled object calculated by the device to be controlled based on the reference value in order to realize control using the controlled variable. In this example, it is assumed that hot water is flowing on the wall of the reactor that is being controlled, and the temperature is changed by changing the opening/closing degree of the valve that allows the hot water to flow. The control amount is an operating condition of control actually performed by the controlled object, and is, for example, information indicating the actual temperature inside the reactor that is the controlled object.

In FIG. 3, for example, past operation performance data identified by data ID "0" includes operation performance data 301 for a controlled object with characteristics represented by product type "A" and equipment "A." The reference value, the manipulated variable, and the controlled variable are the reference value α0 [°C]: [20, 20, 50, 50,..., 20], and the manipulated variable β0 [%]: [15, 18, 80, 70,... , 0] and control amount γ0 [° C.]: [15, 18, 37, 40,..., 30], values at predetermined intervals (for example, every minute) are stored in chronological order. For example, the reference value α0 of equipment "A" that manufactures product type "A" has temperature values such as 20℃, 20℃, 50℃, 50℃, ..., 20℃ every minute. remembered.

FIG. 4 is a diagram showing an example of the process model 1004 and an example of the relationship between the process model and characteristics. The process model is expressed by a model constructed using a predetermined theoretical formula as described above. In FIG. 4, a model represented by y=ax ² +bx+c is illustrated as a process model. In this example, the initial values of parameters a, b, and c of the process model are given as 2, 1, and 0, respectively. In addition, through the processing described later, the correction amounts of process model parameters are calculated as +0.2, +0, and +0 for the product type, and as +0, -0.1, and +0 for the equipment, respectively. It shows that The results show that the parameters a, b, and c of the corrected process model were calculated as 1.2, −0.1, and 0, respectively.

FIG. 5 is a diagram showing an example of the characteristic parameter group 1003. The characteristic parameter group 1003 stores characteristic parameters according to characteristics. In this example, characteristic parameters are stored for each characteristic and for each characteristic type. As shown in FIG. 5, the characteristic parameter group 1003 stores the type of characteristic of the controlled object (product type in this example) and characteristic parameters that are correction parameters for the parameters of the process model that indicate the characteristic. ing. In FIG. 5, for example, the characteristic parameters for correcting the parameters of the process model indicated by the product type "A" are stored as a: 1.2, b: -0.1, and c: 0, respectively. I understand.

Although FIG. 5 shows an example of a table in which characteristic parameters are stored for each product type, as many tables as there are characteristics are stored in the characteristic parameter group 1003. Furthermore, in FIG. 5, since the process model is a model expressed by y=ax ² +bx+c, the explanation has been made on the assumption that the parameters are composed of three, a, b, and c. However, as described above, when the process model is expressed in another expression, the number of characteristic parameters corresponding to the format of the expression is stored.

FIG. 6 is a diagram showing an example of the characteristic information 102. The characteristic information 102 is information indicating the characteristics of the controlled object and the types of characteristics included in the characteristics. The characteristic information 102 is input when the simulation unit 1020 executes a simulation. As shown in FIG. 6, the characteristic information 102 stores characteristics (for example, product type, equipment, etc.) and characteristic types (A, B, A, B, etc.) included in each characteristic in association with each other. In FIG. 6, for example, it is shown that A, B, etc. are defined as characteristic types for varieties.

FIG. 7 is a diagram illustrating an example of a flowchart illustrating the procedure of model construction processing performed by the model construction unit 1001 in the first embodiment. This process is performed when there is an instruction or operation from the user to start the process. The model construction process is a process for calculating characteristic parameters based on past operation performance data 101, and is mainly a process for preparing for calculating characteristic parameters.

In S701, the model construction unit 1001 acquires the process model 1004 and past operation performance data 101 (S701).

In S702, the model construction unit 1001 creates a process model P using arbitrarily set initial values of the process model (S702). For example, the model construction unit 1001 creates a process model P in which initial values a=2, b=1, and c=0 are set for the parameters of the process model represented by y=ax ² +bx+c.

In S703, the model construction unit 1001 acquires the characteristic list L from the past operation performance data 101 (S703). For example, the model construction unit 1001 obtains, as the characteristic list L, a list [product type, equipment] that associates the product types and devices defined in the past operation performance data 101 with each other.

In S704, the model construction unit 1001 extracts the characteristic l without restoration from the characteristic list L acquired in S703, and acquires a characteristic type list K of the characteristic l (S704). For example, the model construction unit 1001 obtains l=product type from the above-described property list L, and obtains a property type list K=[a, b] for the product product.

In S705, the model construction unit 1001 extracts the characteristic type k without restoration from the characteristic type list K acquired in S704 (S705). For example, the model construction unit 1001 obtains the characteristic type k=I from the characteristic type list K=[A, B] described above.

In S706, the model construction unit 1001 calculates the characteristic parameter of the characteristic type k acquired in S705, and stores it in the characteristic parameter group 1003 (S706). For example, the model construction unit 1001 calculates characteristic parameters for variety I. A specific method for calculating characteristic parameters will be described later using FIGS. 8A and 8B.

In S707, the model construction unit 1001 determines whether the characteristic type list K is an empty set (S707). For example, when the model construction unit 1001 calculates the characteristic parameters for the characteristic type A of the variety, it is determined that the set is not an empty set because the characteristic type list K = [B] exists (S707; No), and the process returns to S705. . Then, the model construction unit 1001 executes S705 and S706 for the characteristic type list K=[b]. At this time, processing is being executed for each of the characteristic type list K=[A, B]. Therefore, the model construction unit 1001 determines that the characteristic type list K is an empty set (S707; Yes), and proceeds to S708.

In S708, the model construction unit 1001 determines whether the characteristic list L is an empty set (S708). For example, when the model construction unit 1001 is processing the characteristic list L=[product type], it is determined that the set is not an empty set because the characteristic list L=[device] exists (S708; No), and the process returns to S704. The model construction unit 1001 then executes S704, S705, and S706 for l=[device]. At this time, processing is being executed for each of the characteristic list L=[product type, device]. Therefore, the model construction unit 1001 determines that the characteristic list L is an empty set (S708; Yes), and ends the process.

8A and 8B are diagrams illustrating an example of a flowchart illustrating the processing procedure of the characteristic parameter calculation process performed in S706 illustrated in FIG. 7. The characteristic parameter calculation process is a process for calculating characteristic parameters. Hereinafter, a specific explanation will be given using the past operation performance data 101 shown in FIG. 3. In the explanation using FIGS. 8A and 8B, Loop 1 to Loop 4 will be specifically explained in the order of loop 1 to loop 4 in order to explain an example of a controlled object having two characteristics including the above-mentioned two types of characteristics.

In loop 1, a case will be described in which the characteristic list L=[equipment], the characteristic type list K=[b] are unprocessed, the characteristic l=product type, and the characteristic type k=a are processed.

In S801, the model construction unit 1001 extracts partial data D of characteristic l=characteristic type k from the past operation performance data 101 (S801). For example, the model construction unit 1001 extracts data of type A from the past operation performance data 101, and creates partial data D consisting of data IDs (#) of 0 and 1. At this time, among the past operation performance data 101 shown in FIG. 3, the records "0, I, A, [α0, β0, γ0...]" and " 1, A, B, [α1, β1, γ1...]” are extracted as partial data D.

In S802, the model construction unit 1001 sets a loop variable i=0 (S802).

In S803, the model construction unit 1001 creates an empty list M of model identification data groups (S803). For example, the model construction unit 1001 creates a model identification data group M=[]. The model identification data group is a set of data for identifying characteristic parameters of a process model that are corrected according to past operation performance data 101 for each characteristic and characteristic type. The model identification data group will be described later using FIGS. 9A and 9B.

In S804, the model construction unit 1001 selects the i-th data D[i] of the partial data D (S804). Here, since i=0, the model construction unit 1001 selects the record "0, i, A, [α0, β0, γ0...]" whose data ID (#) is 0 from the partial data D.

In S805, the model construction unit 1001 determines whether characteristic parameters other than characteristic l of characteristic type k have been calculated for data D[i] (S805). At present, characteristics other than product type, that is, characteristic parameters of equipment, have not been calculated. Therefore, the model construction unit 1001 determines that the characteristic parameters other than the characteristic l of the characteristic type k have not been calculated for the data D[i] (S805; No), and proceeds to S809.

In S806, the model construction unit 1001 creates a process model m by correcting the process model P using each calculated characteristic parameter (S806). Here, since the determination in S805 is No as described above, the model construction unit 1001 skips this process.

In S807, [process model m, data d[i]] is added to the model identification data group M (S807). Here, since the determination in S805 is No as described above, the model construction unit 1001 skips this process.

In S808, the model construction unit 1001 executes i=i+1 (S808).

In S809, the model construction unit 1001 determines whether i>=the number of data D (S809). Here, since i=1 and the number of data D is 2, the model construction unit 1001 determines that i>=the number of data (S809; No) and returns to S804. Thereafter, the model construction unit 1001 skips the processes of S807 and S808. In the next loop, since i=2 and i>=2, the model construction unit 1001 determines that i>=the number of data (S809; Yes) and proceeds to S810.

In S810, the model construction unit 1001 determines whether the list of model identification data group M is empty (S810). Here, the model construction unit 1001 determines that the list of model identification data group M is still empty (S810; Yes), and proceeds to S811.

In S811, the model construction unit 1001 adds [process model P, data D] to the model identification data group M (S811). For example, the model construction unit 1001 adds the record "0, I, A, [α0, β0, γ0...]" selected in the loop from S804 to S809 to the process model P using the initial values created in S702. A model identification data group M=[[process model P, data D]] is created in which each of "1, A, B, [α1, β1, γ1, . . .]" is associated with each other.

In S812, the model construction unit 1001 calculates a correction parameter C that minimizes the objective function for each model and data pair in the model identification data group M (S812). For example, the model construction unit 1001 calculates the correction parameter C of the process model P such that the difference between the measured control amounts γ0 and γ1 of each data D and the estimated control amount of the process model P is minimized. The correction parameter C is a parameter for correcting the parameters a, b, and c of the process model 1004 shown in FIG. 4, for example, C={a:+1, b=-1, c=+3}. .

In S813, the model construction unit 1001 stores the correction parameter calculated in S812 as a characteristic parameter of characteristic l and characteristic type k (S813). For example, the model construction unit 1001 records the characteristic parameters={a:+1, b=-1, c=+3} for the characteristic type "a" of the characteristic "variety". When the processing up to this point is completed, the process returns to S705, and processing is performed to calculate the characteristic parameters for the different characteristic type "B" for the same characteristic "Type" and store them in the characteristic parameter group 1003 (S706, FIGS. 8A and 8B). .

In the following loop 2, a case will be described in which the characteristic list L=[equipment], the characteristic type list K=[] are unprocessed, the characteristic l=product type, and the characteristic type k=b are processed.

In S801, the model construction unit 1001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S801). For example, the model construction unit 1001 extracts data of type RO from the past operation performance data 101, and creates partial data D having a data ID (#) of 2. At this time, among the past operation performance data 101 shown in FIG. Extracted. At this time, the model construction unit 1001 reassigns the data ID (#) to the record "0, B, B, [α2, β2, γ2...]" in order to identify the record of partial data D. Create partial data D.

In S802 and S803, the model construction unit 1001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of type A.

In S804, the model construction unit 1001 selects the i-th data D[i] of the partial data D (S804). Here, since i=0, the model construction unit 1001 selects the record "0, B, B, [α2, β2, γ2...]" whose data ID (#) is 0 from the partial data D.

In S805, the model construction unit 1001 determines whether characteristic parameters for data D[i] other than characteristic l of characteristic type k have been calculated (S805). Here, as in the case of type A, the characteristics other than the type, that is, the characteristic parameters of the device have not been calculated at this point, so the process advances to S809.

In S806 and S807, as in the case of type A, since the determination in S805 is No as described above, the model construction unit 1001 skips the process.

In S808, the model construction unit 1001 executes i=i+1 (S808).

In S809, the model construction unit 1001 determines whether i>=the number of data D (S809). Here, since i=1 and the number of data D is 1, the model construction unit 1001 determines that i>=the number of data (S809; Yes) and proceeds to S810.

In S810, the model construction unit 1001 determines whether the list of model identification data group M is empty (S810). Here, the model construction unit 1001 determines that the list of model identification data group M is still empty (S810; Yes), and proceeds to S811.

In S811, the model construction unit 1001 adds [process model P, data D] to the model identification data group M (S811). For example, the model construction unit 1001 adds the record "0, B, B, [α2, β2, γ2...]" selected in the loop from S804 to S809 to the process model P using the initial values created in S702. Create a model identification data group M=[[process model P, data D]] in which the following are associated.

In S812, the model construction unit 1001 calculates a correction parameter C that minimizes the objective function for each model and data pair in the model identification data group M (S812). For example, the model construction unit 1001 calculates the correction parameter C of the process model P such that the difference between the measured control amount γ2 of the data D and the estimated control amount of the process model P is minimized. The correction parameter C is a parameter for correcting the parameters a, b, and c of the process model 1004 shown in FIG. 4, such as C={a:+2, b=+1, c=+10}, for example.

In S813, the model construction unit 1001 stores the correction parameter calculated in S812 as a characteristic parameter of characteristic l and characteristic type k. (S813). For example, the model construction unit 1001 records the characteristic parameters={a:+2, b=+1, c=+10} for the characteristic type “b” of the characteristic “variety”. When the processing up to this point is completed, the process returns to S704, and processing is performed to calculate the characteristic parameters for the characteristic type "A" of the different characteristic "device" and store them in the characteristic parameter group 1003 (S706, FIGS. 8A and 8B).

In the following loop 3, a case will be described in which the characteristic list L=[], the characteristic type list K=[B] are unprocessed, the characteristic l=equipment, and the characteristic type k=A are processed.

In S801, the model construction unit 1001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S801). For example, the model construction unit 1001 extracts data in which the device is A from the past operation performance data 101, and creates partial data D having a data ID (#) of 0. At this time, among the past operation performance data 101 shown in FIG. Extracted.

In S802 and S803, the model construction unit 1001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of types A and B.

In S804, the model construction unit 1001 selects the i-th data D[i] of the partial data D (S804). Here, since i=0, the model construction unit 1001 selects the record "0, i, A, [α0, β0, γ0...]" whose data ID (#) is 0 from the partial data D.

In S805, the model construction unit 1001 determines whether characteristic parameters for data D[i] other than characteristic l of characteristic type k have been calculated (S805). Here, since the characteristics other than the device, that is, the characteristic parameters of the product type have been calculated, the process advances to S806.

In S806, the model construction unit 1001 creates a process model m by correcting the process model P using each calculated characteristic parameter (S806). For example, the model construction unit 1001 corrects the process model P using the characteristic parameters stored in S813 of loop 1, where the characteristic is the product type and the characteristic type is A, to create the process model m.

In S807, the model construction unit 1001 adds [process model m, data d[i]] to the model identification data group M (S807). For example, the model construction unit 1001 associates the process model m created in S806 with the records "0, I, A, [α0, β0, γ0...]" selected in S804, and the model identification data group M=[ Create [process model m, data d[0] (#0)]]. Through the processing in S807, the process model m corrected with the calculated characteristic parameters corresponds to the data D[i] (record "0, i, A, [α0, β0, γ0...]") currently being processed. and added to the list of model identification data group M.

In S808, the model construction unit 1001 executes i=i+1 (S808).

In S809, the model construction unit 1001 determines whether i>=the number of data D (S809). Here, since i=1 and the number of data D is 1, the model construction unit 1001 determines that i>=the number of data (S809; Yes) and proceeds to S810.

In S810, the model construction unit 1001 determines whether the list of model identification data group M is empty (S810). Here, the model construction unit 1001 creates model identification data for the process model m in S807, and adds the model identification data to the list of the model identification data group M. Therefore, the model construction unit 1001 determines that the list of model identification data group M is not empty (S810; No), and proceeds to S812.

In S811, the model construction unit 1001 skips the process because the determination in S810 is No as described above.

In S812, the model construction unit 1001 calculates a correction parameter C that minimizes the objective function for each model and data pair in the model identification data group M (S812). For example, the model construction unit 1001 calculates the correction parameter C of the process model m such that the difference between the measured control amount γ0 of the data d[0] and the estimated control amount of the process model m is minimized. The correction parameter C is a parameter for correcting the parameters a, b, and c of the process model 1004 shown in FIG. 4, such as C={a:-1, b=+1, c=+1}, for example. .

In S813, the model construction unit 1001 stores the correction parameter calculated in S812 as a characteristic parameter of characteristic l and characteristic type k (S813). For example, the model construction unit 1001 records the characteristic parameters={a:-1, b=+1, c=+1} for the characteristic type "A" of the characteristic "equipment". When the processing up to this point is completed, the process returns to S704, and processing is performed to calculate characteristic parameters for a different characteristic type "B" for the same characteristic "equipment" and store them in the characteristic parameter group 1003 (S706, FIGS. 8A and 8B). .

In the following loop 4, a case will be described in which the characteristic list L=[], the characteristic type list K=[] are unprocessed, the characteristic l=equipment, and the characteristic type k=B are processed.

In S801, the model construction unit 1001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S801). For example, the model construction unit 1001 extracts data for device B from the past operation performance data 101, and creates partial data D consisting of data IDs (#) of 1 and 2. At this time, among the past operation performance data 101 shown in FIG. B, B, [α2, β2, γ2...]” is extracted as partial data D. At this time, the model construction unit 1001 reassigns the ID to the record "0, I, B, [α1, β1, γ1...]" for the data ID (#) in order to identify the record of the partial data D. Partial data D is created as "1, B, B, [α2, β2, γ2...]".

In S802 and S803, the model construction unit 1001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of types A, B, and device A.

In S804, the model construction unit 1001 selects the i-th data D[i] of the partial data D (S804). Here, since i=0, the model construction unit 1001 selects the record "0, i, B, [α1, β1, γ1...]" whose data ID (#) is 0 from the partial data D.

In S805, the model construction unit 1001 determines whether characteristic parameters for data D[i] other than characteristic l of characteristic type k have been calculated (S805). Here, since the characteristics other than the device, that is, the characteristic parameters of the product type have been calculated, the process advances to S806.

In S806, the model construction unit 1001 creates a process model m by correcting the process model P using each calculated characteristic parameter (S806). For example, the model construction unit 1001 corrects the process model P using the characteristic parameters stored in S813 of loop 1, where the characteristic is the product type and the characteristic type is A, to create the process model m.

In S807, the model construction unit 1001 adds [process model m, data d[i]] to the model identification data group M (S807). For example, the model construction unit 1001 associates the process model m created in S806 with the records "0, I, B, [α1, β1, γ1...]" selected in S804, and the model identification data group M=[ Create [process model m, data d[0] (#0)]]. Through the processing in S807, the process model m corrected using the calculated characteristic parameters corresponds to the data D[i] (record "0, i, B, [α1, β1, γ1...]") currently being processed. and added to the list of model identification data group M.

In S808, the model construction unit 1001 executes i=i+1 (S808).

In S809, the model construction unit 1001 determines whether i>=the number of data D (S809). Here, since i=1 and the number of data D is 2, the model construction unit 1001 determines that i>=the number of data (S809; No) and returns to S804. Thereafter, the model construction unit 1001 skips the processes of S806 and S807. In the next loop, i=2 and i>=2, so the model construction unit 1001 determines that i>=the number of data (S809; Yes) and proceeds to S810. After completing the next loop, the model construction unit 1001 adds the record "2, B, B, [α2, β2, γ2...]" selected in S804 of the next loop to the process model m created in S806. A model identification data group M=[[process model m, data d[1] (#1)]] is created in which the data are associated with each other. Finally, model identification data group M [[process model m (a), data d [0] (#0)]], [[process model m (b), data d [1] (#1)]] is created.

In S810, the model construction unit 1001 determines whether the list of model identification data group M is empty (S810). Here, the model construction unit 1001 creates model identification data for the process model m in S807, and adds the model identification data to the list of the model identification data group M. Therefore, the model construction unit 1001 determines that the list of model identification data group M is not empty (S810; No), and proceeds to S812.

In S811, the model construction unit 1001 skips the process because the determination in S810 is No as described above.

In S812, the model construction unit 1001 calculates a correction parameter C that minimizes the objective function for each model and data pair in the model identification data group M (S812). For example, the model construction unit 1001 calculates the difference between the measured control amount γ1 of data d[0] and the estimated control amount of process model m(a), and the difference between the measured control amount γ2 of data d[1] and the process model m(ro). ) is calculated for the correction parameter C of the process model m that minimizes the difference in the estimated control amount. The correction parameter C is a parameter for correcting the parameters a, b, and c of the process model 1004 shown in FIG. 4, such as C={a:+0, b=+0, c=+1}, for example.

In S813, the model construction unit 1001 stores the correction parameter calculated in S812 as a characteristic parameter of characteristic l and characteristic type k. (S813). For example, the model construction unit 1001 records the characteristic parameters={a:+0, b=+0, c=+1} for the characteristic type “B” of the characteristic “equipment”. When the processing up to this point is completed, the characteristic parameter calculation processing shown in FIGS. 8A and 8B has been performed for all characteristics and all types of characteristics. Therefore, the process advances to S707 of the model construction process shown in FIG.

In S707, since the process has already been executed for each of the characteristic type list K=[A, B], the model construction unit 1001 determines that the characteristic type list K is an empty set (S707; Yes), and in S708 Proceed to.

In S708, since the process has already been executed for each of the characteristic list L = [product type, device], the model construction unit 1001 determines that the characteristic list L is an empty set (S708; Yes), and ends the process. do.

9A and 9B are diagrams illustrating an example of a flowchart illustrating the processing procedure of the model identification process of the process model recorded in the model identification data group in the characteristic parameter calculation process shown in FIGS. 8A and 8B. This process is performed when called from the characteristic parameter calculation process. For example, it is called during the process of S812 of the characteristic parameter calculation process. The model identification process is a process for performing identification on the process model and data pair recorded in the model identification data group in the characteristic parameter calculation process.

As explained in FIGS. 8A and 8B, the model identification data group M includes paired data in which process models and data are associated (for example, [process model P, data D], [process model m, data d]). [i]]) are stored. In this example, the model identification data group M includes [[process model m(a), data d[0](#0)]], [[process model m(b), data d[1](#1 )]] is stored. The model identification data group M includes a process model corrected with the characteristic parameters for each characteristic that has already been calculated, and the actual performance of the characteristics of the controlled object included in the past operation performance data 101 used when constructing the process model before correction. This is data associated with values. Data D and data d[i] are partial data of past operation performance data 101, and as explained above, performance values of reference value α, manipulated variable β, and controlled variable γ are recorded.

In the process described below, for each pair of data, the model construction unit 1001 first corrects the process model using characteristic parameters. Next, the model construction unit 1001 inputs the data corresponding to the corrected process model (e.g., controlled variable) and the estimated output value of the process model (e.g., the reference value/operated amount obtained by inputting the (estimated value of the control amount) is calculated using an arbitrary objective function. As an arbitrary objective function, for example, RMSE (Root Mean Squared Error) can be used.

Then, the model construction unit 1001 finally calculates the difference between the process model and each actual value for each pair of data, and calculates the average value of the differences calculated for each pair of data. If the calculated average value satisfies a predetermined criterion, such as being sufficiently small, the model construction unit 1001 ends the process and returns it as a final characteristic parameter. On the other hand, if the above-mentioned arbitrary criteria are not satisfied, the determination of the characteristic parameters is started again. Hereinafter, a detailed explanation will be given along the flowchart.

In S901, the model construction unit 1001 selects the characteristic parameter candidates and characteristic parameters calculated in S812 based on an arbitrary optimization algorithm (S901).

In S902, the model construction unit 1001 sets a loop variable j=0 (S902).

In S903, the model construction unit 1001 creates an empty list V that stores the results of the objective function (S903). For example, the model construction unit 1001 creates an objective function result list V=[].

In S904, the model construction unit 1001 acquires the process model m and data d from the j-th list M[j] recorded in the model identification data group M (S904).

In S905, the model construction unit 1001 corrects the process model m obtained in S904 with the characteristic parameter selected in S901, and obtains a process model m' (S905).

In S906, the model construction unit 1001 calculates the value v of an arbitrary objective function using the process model m' corrected in S905 and the data d corresponding to the process model m acquired in S904, and converts the value v into It is added to the objective function result list V (S906).

In S907, the model construction unit 1001 executes j=j+1 (S907).

In S908, the model construction unit 1001 determines whether the length of the model identification data group M>=j (S908). That is, the model construction unit 1001 determines whether j has reached the number of records stored as the model identification data group M.

If the model construction unit 1001 determines that j has not reached the number of records stored as the model identification data group M (S908; No), the process returns to S904 and the subsequent processes are repeated. On the other hand, if the model construction unit 1001 determines that j has reached the number of records stored as the model identification data group M (S908; Yes), the process proceeds to S909.

In S909, the model construction unit 1001 calculates the average value a of the values v of the arbitrary objective functions recorded in the objective function result list V in S906, and compares it with a predetermined standard. Evaluate a (S909).

In S910, the model construction unit 1001 determines whether the calculated average value a satisfies a predetermined criterion (S910). If the model construction unit 1001 determines that the calculated average value a does not meet the predetermined criteria (S910; No), the process returns to S901 and the characteristic parameters are selected again.

In S911, if the model construction unit 1001 determines that the calculated average value a satisfies the predetermined criteria (S910; Yes), the model construction unit 1001 determines that the characteristic parameter selected in S901 is the final characteristic parameter and outputs it. (S911).

As explained above, when the processes in Figures 7 to 9B are completed, it is possible to construct a new process model in which the parameters of the process model constructed using the theoretical formula are corrected for each characteristic based on past operational performance data. , it is possible to express the dynamic characteristics of the controlled object without creating a process model for each combination of characteristics. Therefore, it becomes possible to perform process simulation using a small number of process models even for a controlled object that has a huge number of combinations of characteristics.

FIG. 10 is a diagram illustrating an example of a flowchart illustrating the procedure of simulation processing performed by the simulation unit 1002 in the first embodiment. This process is performed when there is an instruction or operation from the user to start the process. The simulation process is a process in which a process simulation is performed using the process model built in the model building process performed by the model building unit 1001.

In S1001, the simulation unit 1002 acquires characteristic parameters that match the input characteristic information 102 from the characteristic parameter group 1003 stored in FIGS. 7 to 9B (S1001).

In S1002, the simulation unit 1002 corrects the process model P using the acquired characteristic parameters to create a process model P' (S1002).

In S1003, the simulation unit 1002 inputs the simulation input information 103 to the process model P' created in S1002, and executes a process simulation of the controlled object using the process model P' (S1003).

In S1004, the simulation unit 1002 returns the results of executing the process simulation (S1004).

As described above, in this embodiment, in a process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object, past operation performance data (for example, A first processing unit (e.g., model construction unit 1001, FIG. 7) that acquires the characteristics (e.g., product type, equipment, etc.) from the past operation performance data 101); The process to be controlled is estimated by the process model constructed from the actual value of the process to be controlled (for example, the controlled amount of the control target) and a predetermined theoretical formula (for example, the theoretical formula shown in FIG. 4). The characteristic parameters for correcting the process model (for example, the condition that the difference between the measured control amount and the estimated control amount of the process model is the minimum) The second processing unit (for example, model construction unit 1001, FIG. 8) calculates the characteristic parameters stored in the characteristic parameter group 1003 shown in FIG. 5 for each of the above characteristics. As a result, even when a controlled object has multiple characteristics, it is possible to construct a process model that requires less construction and operation load than conventional methods, and it also reduces the number of process models constructed to perform process simulation. be able to.

Further, the process is corrected using the characteristic parameter having the same characteristic as characteristic information (for example, characteristic information 102) indicating the characteristic of the controlled object and the type of characteristic included in the characteristic, which is input into the process model construction system. It has a third processing unit (for example, simulation unit 1002) that creates a model and simulates the behavior of the process to be controlled using the created process model. As a result, process simulation can be performed using a small number of process models constructed by the model construction unit 1001, even for a controlled object that has a huge number of combinations of characteristics and types of characteristics.

Although the first embodiment has been specifically described above, the following modifications can be made as appropriate depending on the environment in which this system is used.

For example, the model construction unit 1001 may correct the parameters of the process model using various methods.

In the embodiment described above, the model construction unit 1001 calculated characteristic parameters for correcting the parameters of the process model constructed using the theoretical formula for each characteristic of the controlled object based on past operation performance data. However, the model construction unit 1001 may calculate statistical values based on the calculated characteristic parameters, and use the statistical characteristic parameters based on the statistical values to correct the parameters of the process model constructed using the above theoretical formula. The above statistical values include, for example, when the model construction unit 1001 corrects the characteristic parameters stored in the characteristic parameter group 1003 by a weighted linear sum, corrects the characteristic parameters by weighted multiplication, or corrects the characteristic parameters for the initial parameters of the process model. There are statistical values obtained through processing such as correction using a correction magnification. The weighting and correction magnification described above may be determined in advance by the user.

In this way, the second processing unit (for example, model construction unit 1001, FIG. 8) calculates a statistical value based on the characteristic parameter calculated for each characteristic, and uses the statistical characteristic parameter based on the statistical value. Then, the process model constructed using the predetermined theoretical formula is corrected. Thereby, the parameters of the process model can be corrected using the characteristic parameters specified in advance by the user and corrected by each of the above-mentioned methods.

Further, the initial values of the parameters of the process model may be determined by identifying them from past operation performance data.

In the embodiments described above, the initial values of the parameters of the process model are given in advance. However, for example, the model construction unit 1001 may set, as initial values, the process model parameters corrected by the characteristic parameters identified from the past operation performance data and obtained by the model identification process shown in FIGS. 9A and 9B. You can.

In this way, the second processing unit (for example, the model construction unit 1001 in FIG. 8) processes the process model corrected using the characteristic parameters and the past operating performance used when building the process model before correction. Create model identification data that associates the actual values included in the data, and perform identification on the pair of the process model and the actual value corrected with the characteristic parameters included in the created model identification data. A model identification process (for example, the model identification process shown in FIGS. 9A and 9B) is performed, and the first processing unit uses the characteristic parameters obtained by the model identification process as initial values of the parameters of the process model. Set the parameters of the corrected process model. Thereby, the initial values of the parameters of the process model can be automatically set based on past operation performance data without the user specifying them in advance.

Further, the characteristic parameters may be manually specified by the user.

In the embodiment described above, the model construction unit 1001 calculated characteristic parameters for correcting the parameters of the process model constructed using the theoretical formula for each characteristic of the controlled object based on past operation performance data. However, the model construction unit 1001 may store the characteristic parameters specified by the user in the characteristic parameter group. Alternatively, the model construction unit 1001 may use some of the calculated characteristic parameters (for example, {a:+0, b=+0} of characteristic parameters = {a:+0, b=+0, c=+1}) and the user A characteristic parameter (for example, {c=+1} of characteristic parameters={a:+0, b=+0, c=+1}) specified from the above may be saved in the characteristic parameter group.

In this way, the process model construction system 1000 has an input section (for example, the input device 1606) that receives input of the characteristic parameters, and the second processing section (for example, the model construction section 1001, FIG. 8) At least some of the characteristic parameters are set as parameters input from the input section, and the calculation is performed for parameters other than the set parameters. This makes it possible to directly store characteristic parameters that reflect the user's intentions.

As mentioned above, the present system can be applied to various plants that perform process operations. In this case, the first processing unit (e.g., model construction unit 1001, FIG. 7) selects a first device and a The second processing unit (e.g., model construction unit 1001, FIG. 8) obtains the actual value of the process to be controlled for the first device and the first type (e.g., The control amount of the reactor) and the estimated value of the process to be controlled satisfy a predetermined condition (for example, the condition that the difference between the measured control amount of the reactor and the estimated control amount of the process model is minimum). The process of calculating characteristic parameters for each of the first device and the first product type, and generating the first product type in the first device based on at least the calculated characteristic parameters. Correct the model. As a result, it is possible to construct process models that require less construction and operation load than conventional methods in various plants that perform process operations, and it is also possible to reduce the number of process models constructed for process simulation. become able to.

In the second embodiment, a case will be described in which a simulation result by the simulation unit 1002 is used to back-calculate the operating method of the controlled object that satisfies the optimum operating conditions, and the result is fed back to the controlled object.

FIG. 11 is a diagram illustrating an example of the configuration of a process model construction system in Example 2. As shown in FIG. 11, the process model construction system 2000 according to the present embodiment has a control equipment input calculation section 2001 in addition to the configuration of the process model construction system 1000 in the first embodiment. The process model construction system 2000 also receives input of desired operating condition data 106 for the controlled object specified by the user. As in the first embodiment, the process model construction system 2000 shown in FIG. 11 can be realized by, for example, a general computer 1600 as hardware as shown in FIG. 2 (computer schematic diagram). Below, the same components as those of the process model construction system 1000 in Example 1 are given the same reference numerals, and their explanations are omitted.

The control equipment input calculation unit 2001 is a processing unit that calculates an input value to the controlled object 105 in order to satisfy the desired operating condition data 106 for a given controlled object by back calculation. The control equipment input calculation unit 2001 uses an arbitrary optimization algorithm to calculate input values to the controlled object 105 through trial and error, and performs optimization calculations to find input values that match the desired operating conditions. For example, the control equipment input calculation unit 2001 sequentially re-executes the above optimization calculation using the feedback results of the control amount obtained from the controlled object 105, and calculates the optimal input value according to the operating status of the controlled object 105. calculate. Input values to the controlled object 105 include time-series data of manipulated variables, reference values, and controlled variables (or a data set combining some or all of these). Further, as the desired operating condition data 106, for example, exemplary operation performance data selected by the user from among the operation performance data 301 of the past operation performance data 101 shown in FIG. 3 can be cited.

In this way, in this embodiment, input to the control object is performed using a predetermined algorithm (for example, an arbitrary optimization algorithm) and the feedback result of the control amount for the control object obtained from the control object. values (for example, time-series data of manipulated variables, reference values, and controlled variables (or a data set combining some or all of these)), and calculate desired operating conditions for the control object (for example, operating condition data 106). ) has a fourth processing unit (eg, control device input calculation unit 2001) that performs optimization calculation to find an input value that matches the input value. In other words, the control equipment input calculation unit 2001 recalculates the input value of the controlled object that satisfies the desired operating conditions each time based on the feedback results obtained from the controlled object, so that it can perform the optimum operation for the current controlled object. Input values that satisfy the conditions can be calculated. Therefore, as in the first embodiment, even if there are many types of characteristics or combinations of characteristics, it is possible to back-calculate the operation amount that satisfies the optimum operating conditions.

In the third embodiment, a case will be described in which characteristic parameters are estimated even for unknown characteristics (for example, product type or equipment) and process simulation is performed.

FIG. 12A is a diagram illustrating an example of the configuration of a process model construction system in Example 3. As shown in FIG. 12A, in addition to the configuration of the process model construction system 1000 in Example 1, the process model construction system 3000 according to the present example further includes a characteristic parameter model construction section 3001, and Example 1 and Example 2. It has a simulation unit 3002 and a characteristic parameter model 3003, which are different from the . The process model construction system 3000 also accepts input of additional characteristic information 107. The process model construction system 3000 shown in FIG. 12A can be realized by a general computer 1600 as hardware, for example, as shown in FIG. 2 (computer schematic diagram), as in the first and second embodiments. . In the following, the same components as those in the process model construction system 1000 and the process model construction system 2000 in Example 1 and Example 2 are given the same reference numerals, and their explanations are omitted.

The characteristic parameter model construction unit 3001 is a processing unit that uses the characteristic parameter group 1003 and the additional characteristic information 107 to construct a machine learning model that estimates characteristic parameters from the additional characteristic information 107. For example, if the characteristic parameter model construction unit 3001 knows the factors (for example, the ratio of raw materials) that affect a certain characteristic (for example, the product type manufactured by the controlled object), is acquired as additional characteristic information 107. The characteristic parameter model construction unit 3001 uses the acquired additional characteristic information 107 as an explanatory variable and the characteristic parameters as an objective variable, and uses any method such as machine learning to construct a characteristic parameter model that becomes a parameter estimator. Thereby, characteristic parameters can be estimated even for unknown characteristic types, and dynamic characteristics of the process model can be estimated.

The additional characteristic information 107 is data that is referred to when constructing the characteristic parameter model described above or estimating characteristic parameters of unknown characteristics.

FIG. 12B is a diagram illustrating an example of additional characteristic information 107. As shown in FIG. 12B, the additional characteristic information 107 includes unknown characteristics (for example, an unknown product type manufactured by the controlled object) and characteristic information corresponding to the characteristic (for example, raw materials for the type manufactured by the controlled object). are stored in association with each other. In FIG. 12B, for example, the ratio of the raw materials of type A manufactured by the controlled object is 30% for raw material 1, 20% for raw material k. Although the raw materials of the product types are exemplified here, various information representing characteristics of characteristics, such as setting values of equipment, may be defined as additional characteristic information.

When correcting the process model with characteristic parameters that match the characteristics of the characteristic information 102, the simulation unit 3002 uses the additional characteristic information 107 and the characteristic parameters described above for types of unknown characteristics (such as varieties with no past operation results). The characteristic parameters are estimated using the model. For example, in S1001 shown in FIG. 10, the simulation unit 3002 inputs the input characteristic information 102 into the characteristic parameter model constructed using the additional characteristic information 107, and inputs the characteristic parameter model that matches the characteristic information 102. Get an estimate of . Furthermore, in S1002, the simulation unit 3002 corrects the process model using the obtained estimated value of the characteristic parameter, and thereafter performs simulation in the same manner as in S1003 and S1004.

In this way, in this embodiment, the characteristics of the control target (for example, an unknown variety manufactured by the control target) and the factors that influence the characteristics (for example, characteristic information such as raw materials of the variety manufactured by the control target) By machine learning, the characteristic information (for example, additional characteristic information 107) associated with the characteristic information (for example, additional characteristic information 107) is used as an explanatory variable, and the characteristic parameter that corrects the process model of the controlled object of the characteristic is used as the objective variable. It has a fifth processing unit (for example, a characteristic parameter model construction unit 3001) that constructs a characteristic parameter model to be estimated, and the third processing unit (for example, a simulation unit 3002) adds the characteristic information to the characteristic parameter model. The behavior of the process to be controlled is simulated using the process model corrected using the estimated value of the characteristic parameter obtained by inputting. That is, the characteristic parameter model construction unit 3001 constructs a characteristic parameter model using the additional characteristic information 107 (for example, when the characteristic is a product type, the raw material ratio of the product type, etc.) and the characteristic parameter group 1003, and the simulation unit 3002 constructs a characteristic parameter model using the characteristic parameter group 1003. The process model is corrected using estimated values of the characteristic parameters obtained by inputting the characteristic information 102 into the constructed characteristic parameter model, and the simulation unit 3002 performs a simulation using the corrected process model. Therefore, it is possible to estimate characteristic parameters and perform process simulation even for unknown types and equipment. In other words, dynamic characteristics can be estimated and process simulations can be performed even for unknown types of characteristics or combinations of characteristics.

In the fourth embodiment, in a controlled object for which there is a small amount of past operation performance data, there are fewer combinations of characteristics and types of characteristics, and certain combinations may cause inconsistencies in the characteristic parameters. In the following, when a contradiction occurs in the characteristic parameters, a method for resolving the contradiction will be explained.

FIG. 13A is a diagram illustrating an example of the configuration of a process model construction system in Example 4. As shown in FIG. 13A, in addition to the configuration of the process model construction system 1000 in Example 1, the process model construction system 4000 according to the present example further includes model identification information 4001, and is different from Examples 1 to 3. It has a different model construction unit 4002. The process model construction system 4000 shown in FIG. 13A can be realized by, for example, a general computer 1600 as hardware as shown in FIG. 2 (computer schematic diagram), as in the first to third embodiments. In the following, the same components as those in the process model construction system 1000, process model construction system 2000, and process model construction system 3000 in Examples 1 to 3 are given the same reference numerals, and their explanations are omitted.

Model identification information 4001 is information that the model construction unit 4002 refers to when performing model identification as shown in FIGS. 9A and 9B, and is corrected according to past operation performance data 101 for each characteristic and characteristic type. This is a collection of data for identifying a characteristic model of a process model. In Example 1, characteristic parameters were calculated for all parameters of the process model as the characteristic parameter group 1003 shown in FIG. It is stored as a characteristic parameter that is the target of the characteristic or characteristic type.

FIG. 13B is a diagram showing an example of model identification information 4001. As shown in FIG. 13B, model identification information 4001 stores one or more characteristic parameters of a process model according to a characteristic or a type of characteristic. In FIG. 13B, for example, a correction parameter {a:+1.2, b=-0.1} which is a characteristic parameter for variety A, and a correction parameter {c=0 which is a characteristic parameter for variety B } are stored in association with each characteristic type. Here, among the parameters a, b, and c forming the characteristic parameters, the characteristic parameters for type A are associated with a and b, and the characteristic parameters for type R are stored in correspondence with c.

Next, the model identification process in this embodiment will be described. In the fourth embodiment, only the step corresponding to S901 differs from the model identification process shown in the first embodiment (FIGS. 9A and 9B), so the flowchart will be omitted. , S901 will be explained.

In S901 of the model identification process shown in FIGS. 9A and 9B, the model construction unit 4002 arbitrarily selects a characteristic parameter corresponding to the characteristic or characteristic type determined by the model identification information 4001 from among the characteristic parameters stored in S813. Select based on optimization algorithm. The characteristic parameters according to the above characteristics and characteristic types can be selected with reference to the model identification information 4001. The model identification information 4001 may be set in advance by the user with reference to the characteristic parameters stored in S813.

When the model construction unit 4002 selects the characteristic parameters determined according to the characteristics and the type of characteristics as described above, the model construction unit 4002 executes the processing from S902 onwards. At this time, in S905, the model construction unit 4002 corrects the process model m obtained in S904 with the characteristic parameter corresponding to the characteristic or characteristic type selected in the step corresponding to S901 in the fourth embodiment described above. Get '.

As described above, in this embodiment, the process model construction system 4000 has model identification information (for example, model identification information 4001) in which the characteristic parameters of the process model according to one or more characteristics are stored, The second processing unit (for example, model construction unit 4002) generates the process model corrected using the characteristic parameters calculated for each characteristic, and past operation performance data 101 used when building the process model before correction. A model that creates model identification data that associates the actual values included in the model identification data, and performs identification on the pair of the process model and the actual value corrected with the characteristic parameters included in the created model identification data. An identification process (for example, the model identification process shown in FIGS. 9A and 9B) is executed, and in the model identification process, the characteristic parameters defined in the model identification information are applied to a predetermined algorithm (for example, an arbitrary optimization algorithm). ), and the process model is corrected using the selected characteristic parameters.

That is, in this embodiment, in addition to the configuration of Embodiment 1, only the parameters of the process model corresponding to the characteristic parameters are corrected using characteristic parameters according to the characteristics and characteristic types defined in the model identification information 4001. . This makes it possible to obtain process model output values with high accuracy even when there is little data on characteristics or combinations of characteristic types and certain combinations result in inconsistencies in characteristic parameters, thereby suppressing increases in process model prediction errors. I can do it.

In the above-described example, characteristic parameters corresponding to characteristics and characteristic types are determined as model identification information 4001. However, the model construction unit 4002 sets a target value for the value v of the objective function calculated at the time of model identification in S906, and, for example, adopts a percentage error for the objective function, and when identifying the model for variety characteristics, the percentage error N %, and when identifying a device characteristic model, settings may be made to identify with a percentage error of 0% as a target. As a result, the output value of the process model can be obtained with high accuracy, as described above, without determining the model identification information 4001.

In Example 5, a method for correcting a process model when a simulation is performed and a deviation occurs between the behavior of the actual process and the behavior of the process model will be described.

FIG. 14 is a diagram illustrating an example of the configuration of a process model construction system in Example 5. As shown in FIG. 14, the process model construction system 5000 according to the present embodiment has a simulation unit 5001 that is different from the configuration of the process model construction system 1000 in the first embodiment than in the first to fourth embodiments. . Further, the process model construction system 5000 receives input of the feedback result of the control amount obtained from the controlled object 1401.

The process model construction system 5000 shown in FIG. 14 can be realized by a general computer 1600 as hardware, for example, as shown in FIG. 2 (computer schematic diagram), as in the first to fourth embodiments. In the following, the same components as in the process model construction system 1000, process model construction system 2000, process model construction system 3000, and process model construction system 4000 in Examples 1 to 4 are given the same reference numerals, and the explanation thereof will be given. It is omitted.

The simulation unit 5001 compares the simulation result and the control result of the controlled object 1401 (for example, the control amount obtained from the controlled object 1401), and determines whether the difference is temporary (for example, 10 seconds) or cumulative (for example, 1. month) to determine whether the value exceeds a certain value. When the simulation unit 5001 determines that the difference is temporarily or cumulatively greater than a certain value, the simulation unit 5001 corrects and recalculates the simulation result, and stores the result in the simulation result 104. As a method for correcting the simulation results, for example, the simulation unit 5001 identifies parameters of a process model that minimizes the difference between the control result of the controlled object 1401 and the simulation result, or a characteristic that satisfies such conditions. There is a method of reselecting parameters from a group of characteristic parameters.

FIG. 15 is a diagram illustrating an example of a flowchart illustrating the processing procedure of simulation processing performed by the simulation unit 5001 in the fifth embodiment. This process is performed when there is an instruction or operation from the user to start the process. Similar to the first embodiment, the simulation process is a process in which a process simulation is performed using the process model constructed in the model construction process performed by the model construction unit 1001.

Each process of S1501 and S1502 is the same as S1001 and S1002 shown in FIG. 10, so the steps from S1503 onward will be described.

In S1503, the simulation unit 5001 inputs the simulation input information 103 to the process model P' created in S1502, executes a simulation of the controlled object by the process model P', and obtains the result S of the executed simulation (S1503). .

In S1504, the simulation unit 5001 acquires the actual measurement value sequence T from the controlled object 1401 (S1504). The actual measurement value sequence T is time-series data included in the operation performance data, such as the actual measurement control variables γ0, γ1, γ2, etc., for example.

In S1505, the simulation unit 5001 determines whether the difference between the simulation result S and the actual measurement value sequence T is greater than or equal to an arbitrary standard that is a predetermined threshold (S1505). In S1505, if the simulation unit 5001 determines that the difference between the simulation result S and the actual measurement value sequence T is not greater than or equal to an arbitrary reference that is a predetermined threshold (S1505; No), the process proceeds to S1507. On the other hand, in S1505, if the simulation unit 5001 determines that the difference between the simulation result S and the actual measurement value sequence T is greater than or equal to an arbitrary reference that is a predetermined threshold (S1505; Yes), the process proceeds to S1506.

In S1506, the simulation unit 5001 corrects the characteristic parameters of the process model used in S1502 so that the difference between the simulation result of the controlled object 1401 using the process model and the actual measurement value sequence T of the controlled object 1401 is minimized. , execute the simulation again (S1506).

In S1507, the simulation unit 5001 returns the simulation results or the results of the re-executed simulation (S1507).

As described above, in this embodiment, when the difference between the simulation result and the control result of the controlled object exceeds a predetermined threshold, the third processing unit (for example, the simulation unit 5001) The characteristic parameters of the process model are corrected so that the values are reduced, and the simulation is re-executed.

That is, the simulation unit 5001 checks the difference between the process simulation result and the actual control amount output by the controlled object, and if the difference exceeds a certain amount, the simulation unit 5001 compares the previous simulation result and the actual control amount. Correct the characteristic parameters of the process model so that the waveforms of This makes it possible to perform process simulations using a small number of process models even for controlled objects that have a huge number of combinations of characteristics and types of characteristics. If a deviation from the actual measured value occurs, the characteristic parameter of the process model can be re-identified based on the actual measured value obtained from the controlled object.

In the above example, the simulation unit 5001 corrected and optimized the characteristic parameters so that the difference between the simulation result of the controlled object 1401 using the process model and the actual measurement value sequence T of the controlled object 1401 was minimized. The simulation may be executed again by correcting any characteristic parameter selected by the user from the characteristic parameter group.

In Example 6, a case will be described in which a machine learning model is used for the process model and characteristic parameters.

FIG. 16 is a diagram illustrating an example of the configuration of a process model construction system in Example 6. As shown in FIG. 16, the process model construction system 6000 according to the present embodiment has a model construction section 6001 and a simulation section 6002 that are different from the configuration of the process model construction system 1000 in the first embodiment from those of the first to fifth embodiments. , and has a characteristic model group 6003 instead of the characteristic parameter group 1003.

The process model construction system 6000 shown in FIG. 16 can be realized by, for example, a general computer 1600 as hardware as shown in FIG. 2 (computer schematic diagram), as in the first to fifth embodiments. In the following, the same components as in the process model construction system 1000, process model construction system 2000, process model construction system 3000, process model construction system 4000, and process model construction system 5000 in Examples 1 to 5 are given the same reference numerals. Therefore, the explanation is omitted.

The model construction unit 6001 is a processing unit that constructs a process model 1004 and a characteristic model group 6003 based on past operation performance data 101 using machine learning.

The simulation unit 6002 is a processing unit that calls the process model 1004 and the characteristic model group 6003 based on the input characteristic information 102 and simulation input information 103, and executes a simulation. The characteristic model group 6003 is a set of characteristic models constructed for each characteristic based on the difference between the process model constructed by machine learning and past operation performance data.

FIG. 17 is a diagram illustrating an example of a flowchart illustrating the procedure of model construction processing performed by the model construction unit 6001 in the sixth embodiment. This process is performed when there is an instruction or operation from the user to start the process. The model construction process is a process for calculating a characteristic model based on the past operation performance data 101, and is mainly a process for preparing for building the characteristic model. In the following description, it is assumed that the process model is average value data of past operation performance data 101.

In S1701, the model construction unit 6001 acquires the process model 1004 and past operation performance data 101 (S1701).

In S1702, the model construction unit 6001 constructs a process model P from the average value of the past operation performance data 101 (S1702). For example, the model construction unit 6001 calculates the average values of reference values, manipulated variables, and controlled variables, which are time-series data included in the operation performance data accumulated as the past operation performance data 101, and calculates these average values. Create data as a process model P. The process model P is an initial machine learning model for constructing a machine learning model described later using FIGS. 18A and 18B.

In S1703, the model construction unit 6001 acquires the characteristic list L from the past operation performance data 101, as in the case of the first embodiment.

In S1704, the model construction unit 6001 extracts the characteristic l without restoration from the characteristic list L acquired in S1703, and acquires a characteristic type list K of the characteristic l, as in the case of the first embodiment (S1703, S1704).

In S1705, the model construction unit 6001 extracts the characteristic type k without restoration from the characteristic type list K acquired in S1704, as in the case of the first embodiment (S1705).

In S1706, the model construction unit 6001 calculates the characteristic model of characteristic type k acquired in S1705, and stores it in the characteristic model group 6003 (S1706). For example, the model construction unit 6001 calculates a characteristic model for variety I. A specific method for calculating the characteristic model will be described later using FIGS. 18A and 18B.

In S1707, the model construction unit 6001 determines whether the characteristic type list K is an empty set (S1707). For example, when the model construction unit 6001 calculates the characteristic model for the characteristic type A of the variety, it is determined that the set is not an empty set because the characteristic type list K = [B] exists (S1707; No), and the process returns to S1705. . Then, the model construction unit 6001 executes S1705 and S1706 for the characteristic type list K=[b]. At this time, as in the case of the first embodiment, processing is executed for each of the characteristic type list K=[A, B]. Therefore, the model construction unit 6001 determines that the characteristic type list K is an empty set (S1707; Yes), and proceeds to S1708.

In S1708, the model construction unit 6001 determines whether the characteristic list L is an empty set (S1708). For example, when the model construction unit 6001 is processing the characteristic list L=[product type], it is determined that the set is not an empty set because the characteristic list L=[device] exists (S1708; No), and the process returns to S1704. The model construction unit 6001 then executes S1704, S1705, and S1706 for l=[device]. At this time, as in the case of the first embodiment, processing is executed for each of the characteristic list L=[product type, device]. Therefore, the model construction unit 6001 determines that the characteristic list L is an empty set (1S708; Yes) and ends the process.

18A and 18B are diagrams illustrating an example of a flowchart illustrating the processing procedure of the characteristic model calculation process performed in S1706 illustrated in FIG. 17. The characteristic model calculation process is a process for constructing a characteristic model.

In the characteristic model calculation process, the model construction unit 6001 extracts data D whose characteristic types match the characteristics of the characteristic model to be constructed from the past operation performance data 101. As explained below, the data D and the data d[i] are partial data of the past operation performance data 101, as in the case of the first embodiment, and as explained above, the reference value α, the manipulated variable β, Each actual value of the control amount γ is recorded.

If the model construction unit 6001 has not constructed a characteristic model for other characteristics, the estimated output value of the data D and the process model P (for example, the estimated value of the controlled variable obtained by inputting the reference value/operated amount) Create differential data between the two and build a machine learning model using this as the objective variable. For example, RMSE can be used as an arbitrary objective function. If the operation performance data included in the past operation performance data 101 is time series data of controlled variables, the process model P is the average time series data of controlled variables. The model construction unit 6001 creates difference data d between time series data of each control amount of data D and average time series data, and constructs a time series prediction model for the set of data d. Regarding explanatory variables, data for the first window frame may be given to obtain the actual control amount from the controlled object, or data D may be extended and dummy data for one window frame may be inserted. . Alternatively, additional information may be prepared in the past operation performance data 101.

On the other hand, if the model construction unit 6001 has already constructed other characteristic models, when calculating the difference between each data D and the process model P, the model construction unit 6001 subtracts the output of the constructed model and uses the data D. A characteristic model is constructed, and the constructed characteristic model is stored in a characteristic model group 6003.

Hereinafter, a specific explanation will be given using the past operation performance data 101 shown in FIG. 3. In the explanation using FIGS. 18A and 18B, as in the case of Embodiment 1, in order to explain the control target of two characteristics including two types of characteristics as an example, loop 1 to loop 4 will be specifically explained in order. . As mentioned above, the process model will be explained on the premise that it is a model identified using all of the past operation performance data 101, or time series data of the average value of the controlled variable constructed from the average value of the controlled variable at each time. .

In loop 1, a case will be described in which the characteristic list L=[equipment], the characteristic type list K=[b] are unprocessed, the characteristic l=product type, and the characteristic type k=a are processed.

In S1801, the model construction unit 6001 extracts partial data D of characteristic l=characteristic type k from the past operation performance data 101, as in the case of Example 1 (S801). For example, among the past operation performance data 101 shown in FIG. , i, B, [α1, β1, γ1...]” are extracted as partial data D.

In S1802 and S1803, the model construction unit 6001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of the first embodiment (S1802, S1803). Regarding the model identification data group, processing similar to the model identification processing shown in FIGS. 9A and 9B is performed.

In S1804, the model construction unit 6001 selects the i-th data D[i] of the partial data D, as in the case of the first embodiment (S1804). Here, since i=0, the model construction unit 1001 selects the record "0, i, A, [α0, β0, γ0...]" whose data ID (#) is 0 from the partial data D.

In S1805, the model construction unit 6001 determines whether a characteristic model other than characteristic l of characteristic type k has been constructed for data D[i] (S1805). At this point, the model construction unit 1001 determines that no characteristic models other than characteristic l of characteristic type k have been constructed for data D[i] (S1805; No), and proceeds to S1809.

In S1806, the model construction unit 6001 creates difference data g between the value obtained by adding the output of each constructed characteristic model to the output of the process model P and data d[i] (S1806). Here, since the determination in S1805 is No as described above, the model construction unit 6001 skips this process.

In S1807, difference data g is added to the model identification data group (S1807). Here, since the determination in S805 is No as described above, the model construction unit 6001 skips this process.

In S1808, the model construction unit 6001 executes i=i+1 (S1808).

In S1809, the model construction unit 6001 determines whether i>=the number of data D (S1809). Here, since i=1 and the number of data D is 2, the model construction unit 6001 determines that i>=the number of data (S1809; No) and returns to S1804. Thereafter, the model construction unit 1001 skips the processes of S807 and S808. In the next loop, i=2 and i>=2, so as in the case of Example 1, the model construction unit 6001 determines that i>=the number of data (S1809; Yes), and proceeds to S1810. move on.

In S1810, the model construction unit 6001 determines whether the list of model identification data group M is empty (S1810). Here, the model construction unit 6001 determines that the list of model identification data group M is still empty (S1810; Yes), and proceeds to S1811.

In S1811, the model construction unit 6001 adds difference data g between the output of the process model P and the control amount of the data D to the model identification data group M (S1811). For example, the model construction unit 6001 uses the output value of the process model P using the initial value created in S1702 and the record "0, I, A, [α0, β0, γ0" selected in the loop from S1804 to S1809. ...]" and "1, A, B, [α1, β1, γ1...]", a model identification data group M=[data g] is created.

In S1812, the model construction unit 6001 constructs a regression model s using machine learning for the data set of the model identification data group (S1812). For example, the model construction unit 6001 constructs a regression model s using the reference value/operated amount of data D as an explanatory variable and the controlled amount of data D as an objective variable.

In S1813, the model construction unit 6001 stores the regression model s calculated in S1812 as a characteristic model with characteristic l and characteristic type k (S1813). For example, the model construction unit 6001 records a characteristic model for the characteristic type “i” of the characteristic “variety”. When the processing up to this point is completed, the process returns to S1705, and a process is performed to construct a characteristic model for a different characteristic type "B" for the same characteristic "variety" and store it in the characteristic model group 6003 (S1706, FIGS. 18A and 18B) .

In the following loop 2, a case will be described in which the characteristic list L=[equipment], the characteristic type list K=[] are unprocessed, the characteristic l=product type, and the characteristic type k=b are processed.

In S1801, the model construction unit 6001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S1801). For example, the model construction unit 6001 extracts data of type RO from the past operation performance data 101, and creates partial data D having a data ID (#) of 2. At this time, among the past operation performance data 101 shown in FIG. Extracted. At this time, as in the case of the first embodiment, the model construction unit 1001 reassigns the ID and creates partial data D as a record "0, B, B, [α2, β2, γ2...]".

In S1802 and S1803, the model construction unit 6001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of type A.

In S1804, the model construction unit 6001 selects the i-th data D[i] of the partial data D (S1804). Here, since i=0, the model construction unit 1001 selects the record "0, B, B, [α2, β2, γ2...]" whose data ID (#) is 0 from the partial data D.

In S1805, the model construction unit 6001 determines whether characteristic models for data D[i] other than characteristic l of characteristic type k have been constructed (S1805). Here, as in the case of type A, a characteristic model for characteristics other than the type, that is, the characteristics of the device, has not been constructed at this point, so the process advances to S1809.

In S1806 and S1807, as in the case of type A, since the determination in S1805 is No as described above, the model construction unit 6001 skips the processing.

In S1808, the model construction unit 6001 executes i=i+1 (S1808).

In S1809, the model construction unit 6001 determines whether i>=the number of data D (S1809). Here, since i=1 and the number of data D is 1, the model construction unit 6001 determines that i>=the number of data (S1809; Yes) and proceeds to S1810.

In S1810, the model construction unit 1001 determines whether the list of model identification data group M is empty (S1810). Here, the model construction unit 6001 determines that the list of model identification data group M is still empty (S1810; Yes), and proceeds to S1811.

In S1811, the model construction unit 1001 adds difference data g between the output value of the process model P and the control amount of the data D to the model identification data group M (S1811). For example, the model construction unit 6001 uses the output value of the process model P using the initial value created in S1702 and the record "0, B, B, [α2, β2, γ2" selected in the loop from S1804 to S1809. . . ]", a model identification data group M=[data g] is created.

In S1812, the model construction unit 6001 constructs a regression model s using machine learning for the data set of the model identification data group M (S1812). For example, the model construction unit 6001 constructs a regression model s using the reference value/operated amount of data D as an explanatory variable and the controlled amount of data D as an objective variable.

In S1813, the model construction unit 6001 stores the regression model s calculated in S1812 as a characteristic model with characteristic l and characteristic type k (S1813). For example, the model construction unit 1001 records a characteristic model for the characteristic type “b” of the characteristic “variety”. When the processing up to this point is completed, the process returns to S1704, and a process is performed to construct a characteristic model for the characteristic type "A" of the different characteristic "device" and store it in the characteristic model group 6003 (S1706, FIGS. 18A and 18B).

In the following loop 3, a case will be described in which the characteristic list L=[], the characteristic type list K=[B] are unprocessed, the characteristic l=equipment, and the characteristic type k=A are processed.

In S1801, the model construction unit 6001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S1801). For example, the model construction unit 6001 extracts data of device A from the past operation performance data 101, and creates partial data D consisting of data ID (#) of 0. At this time, among the past operation performance data 101 shown in FIG. Extracted.

In S1802 and S1803, the model construction unit 6001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of types A and B.

In S1804, the model construction unit 6001 selects the i-th data D[i] of the partial data D (S1804). Here, since i=0, the model construction unit 6001 selects the record "0, i, A, [α0, β0, γ0...]" whose data ID (#) is 0 from the partial data D.

In S1805, the model construction unit 6001 determines whether characteristic models for data D[i] other than characteristic l of characteristic type k have been constructed (S1805). Here, since a characteristic model with characteristics other than the device, that is, the characteristic is the product type, has been constructed, the process advances to S1806.

In S1806, the model construction unit 6001 creates difference data g between the value obtained by adding the output of each constructed characteristic model to the output of the process model P and data d[i] (S1806). For example, the model construction unit 6001 adds the output of the characteristic model whose characteristic is product type and the characteristic type is A and the output of the process model P stored in S1813 of loop 1, and the control amount of data d[i]. The difference between the two is created as new difference data g.

In S1807, the model construction unit 6001 adds the difference data g to the model identification data group M (S1807). For example, the model construction unit 6001 sets the model identification data group M=[g]. Through the process of S1807, new difference data g, which is the difference between the constructed characteristic model and data d[i], is changed to the currently processed data D[i] (record "0, I, A, [α0 , β0, γ0...]") is added to the list of model identification data group M.

In S1808, the model construction unit 6001 executes i=i+1 (S1808).

In S1809, the model construction unit 6001 determines whether i>=the number of data D (S1809). Here, since i=1 and the number of data D is 1, the model construction unit 6001 determines that i>=the number of data (S1809; Yes) and proceeds to S1810.

In S1810, the model construction unit 6001 determines whether the list of model identification data group M is empty (S1810). Here, the model construction unit 6001 adds the new difference data g to the list of the model identification data group M in S1807. Therefore, the model construction unit 6001 determines that the list of model identification data group M is not empty (S1810; No), and proceeds to S1812.

In S1811, since the determination in S1810 is No as described above, the model construction unit 6001 skips the process.

In S1812, the model construction unit 6001 constructs a regression model s using machine learning for the data set of the model identification data group (S1812). For example, the model construction unit 6001 constructs a regression model s using the reference value/operation amount of the new difference data g as an explanatory variable and the control amount of the new difference data g as an objective variable.

In S1813, the model construction unit 6001 stores the regression model s calculated in S1812 as a characteristic model with characteristic l and characteristic type k (S1813). For example, the model construction unit 1001 records a characteristic model for characteristic type “A” of characteristic “equipment”. When the processing up to this point is completed, the process returns to S1704, and a process is performed to construct a characteristic model for a different characteristic type "B" for the same characteristic "equipment" and store it in the characteristic model group 6003 (S1706, FIGS. 18A and 18B) .

In the following loop 4, a case will be described in which the characteristic list L=[], the characteristic type list K=[] are unprocessed, the characteristic l=equipment, and the characteristic type k=B are processed.

In S1801, the model construction unit 6001 extracts data D of characteristic l=characteristic type k from the past operation performance data 101 (S1801). For example, the model construction unit 6001 extracts data for device B from the past operation performance data 101, and creates partial data D consisting of data IDs (#) of 1 and 2. At this time, among the past operation performance data 101 shown in FIG. B, B, [α2, β2, γ2...]” is extracted as partial data D. At this time, the model construction unit 6001 reassigns the IDs and records "0, I, B, [α1, β1, γ1...]", "1, B, B, [α2, Partial data D is created as "β2, γ2...]".

In S1802 and S1803, the model construction unit 6001 sets the loop variable i=0 and creates an empty list M of the model identification data group, as in the case of types A, B, and device A.

In S1804, the model construction unit 6001 selects the i-th data D[i] of the partial data D (S1804). Here, since i=0, the model construction unit 1001 selects the record "0, i, B, [α1, β1, γ1...]" whose data ID (#) is 0 from the partial data D.

In S1805, the model construction unit 6001 determines whether characteristic models for data D[i] other than characteristic l of characteristic type k have been constructed (S1805). Here, since a characteristic model with characteristics other than the device, that is, the characteristic is the product type, has been constructed, the process advances to S1806.

In S1806, the model construction unit 6001 creates difference data g between the value obtained by adding the output of each constructed characteristic model to the output of the process model P and data d[i] (S1806). For example, the model construction unit 6001 adds the output of the characteristic model whose characteristic is the product type and the characteristic type is A, stored in S1813 of loop 1, and the output of the process model P, and the control amount of data d[0]. The difference between the two is created as new difference data g.

In S1807, the model construction unit 6001 adds data g to the model identification data group M (S1807). For example, the model construction unit 6001 sets the model identification data group M=[g(a)]. Through the process of S1807, the new difference data g, which is the difference between the constructed characteristic model and the data d[0], is changed to the currently processed data D[0] (record "0, I, B, [α1 , β1, γ1...]") is added to the list of model identification data group M.

In S1808, the model construction unit 6001 executes i=i+1 (S1808).

In S1809, the model construction unit 6001 determines whether i>=the number of data D (S1809). Here, since i=1 and the number of data D is 2, the model construction unit 1001 determines that i>=the number of data (S1809; No) and returns to S1804. Thereafter, the model construction unit 6001 skips the processes of S1807 and S1808. In the next loop, i=2 and i>=2, so the model construction unit 6001 determines that i>=the number of data (S1809; Yes) and proceeds to S1810. When the next loop described above is completed, the model construction unit 6001 calculates the difference between the output of the process model P created in S1806, the output of the characteristic model of type R, and the control amount of data d[1] as data g. Create as. Finally, a model identification data group M=[g(a), g(b)] is created.

In S1810, the model construction unit 6001 determines whether the list of model identification data group M is empty (S1810). Here, the model construction unit 6001 adds the new difference data g(a) to the list of the model identification data group M in S1807. Therefore, the model construction unit 6001 determines that the list of model identification data group M is not empty (S1810; No), and proceeds to S1812.

In S1811, since the determination in S1810 is No as described above, the model construction unit 6001 skips the process.

In S1812, the model construction unit 6001 constructs a model s using machine learning for the data set of the model identification data group (S1812). For example, the model construction unit 6001 uses the reference values/operated amounts of the new difference data g(a) and g(b) as explanatory variables, and the control amounts of the new difference data g(a) and g(b) as explanatory variables. A regression model s is constructed as the objective variable.

In S1813, the model construction unit 6001 stores the regression model s calculated in S1812 as a characteristic model with characteristic l and characteristic type k (S1813). For example, the model construction unit 1001 records a characteristic model for characteristic type “B” of characteristic “equipment”. When the processing up to this point is completed, the characteristic model construction processing shown in FIGS. 18A and 18B has been performed for all characteristics and all types of characteristics. Therefore, the process advances to S1707 of the model construction process shown in FIG. 17.

In S1707, since the process has already been executed for each of the characteristic type list K=[A, B], the model construction unit 6001 determines that the characteristic type list K is an empty set (S1707; Yes), and in S1708 Proceed to.

In S1708, since the process has already been executed for each of the characteristic list L = [product type, device], the model construction unit 6001 determines that the characteristic list L is an empty set (S1708; Yes), and ends the process. do.

As explained above, in this embodiment, in a process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object, past operation performance data ( For example, a first processing unit (e.g., model construction unit 6001, FIG. 17) that acquires the characteristics (e.g., product type, equipment, etc.) from the past operation performance data 101) and the characteristics included in the past operation performance data The above control estimated by the actual value of the process to be controlled (e.g., the controlled amount of the control object) and the process model constructed by predetermined machine learning (for example, arbitrary machine learning such as linear regression) A second processing unit (for example, model construction unit 6001, FIG. 18) that calculates a characteristic model (for example, regression model s) obtained by correcting the process model for each of the above characteristics based on the difference from the target estimated value. and has. In other words, a process model built using machine learning, a characteristic model for each characteristic based on the difference between the process model and past operation performance data, and a characteristic model constructed using the calculation results of the process model built using machine learning. Correct using the calculation results. Therefore, as in the case of Example 1, even when the controlled object has multiple characteristics, it is possible to construct a process model that requires less construction and operation load than in the past. The number of process models to be created can be reduced. Furthermore, as in the case of the first embodiment, process simulation is performed using a small number of process models constructed by the model construction unit 6001, even for a controlled object with a huge number of combinations of characteristics and types of characteristics. You will be able to do this.

In this example, a case has been described in which the entire process model is replaced with a machine learning model, but a configuration may also be adopted in which some formulas of the process model are replaced with the machine learning model. Further, in this example, an example in which the control amount is used as the target for the difference data is shown, but the target is not limited to this, and other measured values or values that can be inferred from the measured values may be used as the target. Furthermore, in this example, we have explained the case where difference data is prepared as training data when constructing a device or product model. Appropriate training data (corresponding measurement values specified by the user, etc.) may be prepared for each type (for example, equipment or type).

In a seventh embodiment, an input/output interface in the model building process shown in the first embodiment will be described. As already explained, in the model construction process, the model construction unit 1001 receives the past operation performance data 101 as input data, and reads out the process model 1004 and the initial values of the parameters of the process model 1004. Then, the model construction unit 1001 outputs characteristic parameters for each characteristic in the model construction process.

FIG. 19 shows a screen (model construction screen) that the model construction unit 1001 displays on the display of the computer that constitutes the process model construction system 1000 (for example, the output device 1605 of the computer 1600 shown in FIG. 2) when the model construction unit 1001 executes the model construction process. ) is a diagram showing an example.

As shown in FIG. 19, the model construction unit 1001 constructs a model that includes input past operation performance data 101, which is input data to the model construction process, a process model 1004, and initial values of parameters of the process model. Screen W1901 is displayed. The user confirms these input data on the model construction screen and instructs execution of the model construction process by pressing an execution button (not shown) on the screen.

Further, as shown in FIG. 19, the model construction unit 1001 displays a model construction screen W1902 that includes a characteristic parameter group 1003, which is output data of the model construction process. The user checks the output data on the model construction screen and checks the characteristic parameters for each characteristic and characteristic type.

In this way, in this embodiment, past operation performance data (for example, past operation performance data 101) indicating the operation performance according to the characteristics of the controlled object (for example, product type, equipment, etc.) and information obtained from the past operation performance data are provided. Input information including a process model (for example, a process model expressed by the theoretical formula shown in FIG. 4) for modeling the characteristics of the controlled object to be controlled, and the process estimated by the operation record and the process model. Calculated for each characteristic of the controlled object so that the estimated value of the controlled object satisfies a predetermined condition (for example, the condition that the difference between the measured controlled variable and the estimated controlled variable of the process model is minimum); A model construction unit that displays output information including characteristic parameters for correcting the process model (for example, characteristic parameters stored in the characteristic parameter group 1003 shown in FIG. 5) on a computer screen as states before and after processing. (for example, a model construction unit 1001). That is, the model construction section displays the model construction screen described above when performing model construction processing. Therefore, the user can grasp at a glance the value of the characteristic parameter calculated for each characteristic or characteristic type for the input process model.

In this embodiment, a case is illustrated in which a screen is used as an interface when the input data is received by the user or when output data is presented to the user. However, if the input device 1606 includes an audio input/output device such as a microphone or a speaker, the input data can be received by voice through the device, or the output data can be presented. Other types of interfaces that can be recognized by the user may also be used.

In the eighth embodiment, an input/output interface in the simulation process shown in the first embodiment will be described. As already explained, in the simulation process, the simulation unit 1002 receives, as input data, the characteristics of the product type, equipment, etc., and the types of characteristics included therein as the characteristic information 102. Further, time series data of the manipulated variable of the controlled object is received as simulation input information 103 which is the desired operating conditions. Then, the simulation unit 1002 outputs time-series data of the control amount as a result of the simulation, for example, in the simulation process.

FIG. 20 shows a screen (simulation execution screen) that the simulation unit 1002 displays on the display of the computer that constitutes the process model construction system 1000 (for example, the output device 1605 of the computer 1600 shown in FIG. 2) when the simulation unit 1002 executes the simulation process. It is a figure showing an example.

As shown in FIG. 20, the simulation unit 1002 displays a simulation execution screen W2001 that includes simulation input information 103, which is input data to simulation processing, and characteristic information 102. The user confirms these input data on the simulation execution screen and instructs execution of the simulation process by pressing an execution button (not shown) on the screen.

Further, as shown in FIG. 20, the simulation unit 1002 displays a model construction screen W2002 that includes simulation results, which are output data of the simulation process. The user confirms the output data on the simulation execution screen and confirms the simulation results, such as time-series changes in the control amount.

As described above, in this embodiment, the third processing unit (for example, the simulation unit 1002) receives the simulation input that defines the characteristic information (for example, the characteristic information 102) and the desired operating conditions for performing the simulation. A corrected process model is created using input information including information (for example, simulation input information 103) and the characteristic parameters having the same characteristics as the characteristic information, and the simulation input information is input into the created process model. output information including the results of the simulation (for example, time-series data of controlled variables) obtained by executing the simulation are displayed on the screen of the computer as states before and after processing. That is, the simulation unit 1002 displays the above-mentioned simulation execution screen when performing simulation processing. Therefore, the user can grasp at a glance the results of simulation using the process model corrected by the characteristic parameters obtained from the input characteristic information. In this example, time-series data of the manipulated variable of the controlled object is provided, but a reference value may also be provided.

In this embodiment, a case is illustrated in which a screen is used as an interface when the input data is received by the user or when output data is presented to the user. However, as in the case of Embodiment 7, other types of interfaces that can be recognized by the user may be used.

In Example 9, an input/output interface will be described in the case where the operation method (eg, manipulated variable) of the controlled object that satisfies the optimal operating conditions shown in Example 2 is calculated backwards and the result is fed back to the controlled object. As already explained, in the second embodiment, the control equipment input calculation unit 2001 receives desired operating condition data 106 for the controlled object as input data, and characteristics of the controlled object such as product type and equipment as characteristic information 102. Receives the included property type. For example, the control equipment input calculation unit 2001 calculates the input (in this example, the manipulated variable) that matches the optimal operating condition, which is the desired operating condition data 106, and the result of the simulation executed by the simulation unit 1002, and calculates the input (in this example, the manipulated variable). Outputs the time series data as the input value of the control target.

FIG. 21 shows the display of the computer configuring the process model construction system 1000 (for example, the output of the computer 1600 shown in FIG. 16 is a diagram showing an example of a screen (feedback execution screen) displayed on the device 1605). FIG.

As shown in FIG. 21, the control device input calculation unit 2001 displays a feedback execution screen W2101 that includes desired operating condition data 106, which is input data when performing the above-mentioned feedback, and characteristic information 102. The user confirms these input data on the feedback execution screen and presses an execution button (not shown) on the screen to instruct backward calculation of the operating method of the controlled object that satisfies the optimum operating conditions.

Further, as shown in FIG. 21, the control device input calculation unit 2001 displays a feedback execution screen W2102 that includes the feedback result, which is the output data of the above-mentioned back calculation. The user confirms the output data on the feedback execution screen and confirms the feedback results, such as time-series changes in the manipulated variable.

As described above, in this embodiment, the fourth processing unit (for example, the control device input calculation unit 2001) calculates the desired operating conditions for the controlled object (for example, the operating condition data 106) and the characteristics of the controlled object. The process to be controlled is controlled by a process model corrected using input information including characteristic information (for example, characteristic information 102) indicating the characteristic type and the characteristic type included in the characteristic, and the characteristic parameter having the same characteristic as the characteristic information. Output information including input values to the controlled object (for example, time-series data of manipulated variables) that match the result of simulating the behavior of is displayed on the screen of the computer as the state before and after processing.

That is, when the control equipment input calculation unit 2001 back-calculates the operating method of the controlled object that satisfies the optimal operating conditions and feeds back the result to the controlled object, the above-mentioned feedback execution screen is displayed. Therefore, the user can grasp at a glance the feedback result of the control based on the input value to the controlled object that satisfies the optimal operating conditions for the controlled object at the current time.

In this embodiment, a case is illustrated in which a screen is used as an interface when the input data is received by the user or when output data is presented to the user. However, as in the seventh and eighth embodiments, other types of interfaces that can be recognized by the user may be used. In this example, a case has been described in which a manipulated variable is calculated backwards, but the present invention may be similarly applied to a case in which a reference value is calculated backwards.

The present invention is not limited to the above-described embodiments as they are, and in the implementation stage, the components may be modified and embodied without departing from the gist thereof, or multiple components disclosed in the above-described embodiments may be embodied. These can be implemented in appropriate combinations.

1000 to 6000 Process model construction system 1001 Model construction section 1002 Simulation section 1003 Characteristic parameter group 1004 Process model 2001 Control device input calculation section 3001 Characteristic parameter model construction section 3002 Simulation section 4001 Model identification information 4002 Model construction section 6001 Model construction section 6002 Simulation Department

Claims (15)

A process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object,
a first processing unit that acquires the characteristics from past operation performance data indicating an operation performance according to the characteristics of the controlled object;
The actual value of the process to be controlled with respect to the characteristic included in the past operation performance data and the estimated value of the process to be controlled estimated by the process model constructed using a predetermined theoretical formula meet a predetermined condition. a second processing unit that calculates, for each of the characteristics, characteristic parameters for correcting the process model such that the process model is satisfied;
A process model construction system comprising: A corrected process model is created using the characteristic parameters of the same characteristics as the characteristic information input to the process model construction system indicating the characteristics of the controlled object and the types of characteristics included in the characteristics, and the created process a third processing unit that uses a model to simulate the behavior of the controlled process;
The process model construction system according to claim 1, characterized in that it has: An input value to the controlled object is calculated using a predetermined algorithm and a feedback result of the control amount for the controlled object obtained from the controlled object, and an input value that matches a desired operating condition for the controlled object. a fourth processing unit that performs optimization calculation to find
The process model construction system according to claim 1, characterized in that it has: Machine learning is performed using characteristic information that associates the characteristics of the controlled object with elements that influence the characteristics as an explanatory variable, and the characteristic parameters that correct the process model of the controlled object for the characteristics as the objective variable. a fifth processing unit that constructs a characteristic parameter model for estimating the characteristic parameter of
The third processing unit simulates the behavior of the process to be controlled using the process model corrected using the estimated value of the characteristic parameter obtained by inputting the characteristic information into the characteristic parameter model. do,
3. The process model construction system according to claim 2. The process model construction system has model identification information in which the characteristic parameters of the process model according to one or more characteristics are stored,
The second processing unit generates model identification data that associates the process model corrected with the characteristic parameters with actual values included in the past operation performance data used when constructing the process model before correction. performing a model identification process of performing identification on a pair of the process model and the actual value that have been created and corrected with the characteristic parameters included in the created model identification data;
In the model identification process, the characteristic parameters defined in the model identification information are selected based on a predetermined algorithm, and the process model is corrected with the selected characteristic parameters.
The process model construction system according to claim 1, characterized in that: The second processing unit calculates a statistical value based on the characteristic parameter calculated for each characteristic, and uses the statistical characteristic parameter based on the statistical value to calculate the process model constructed by the predetermined theoretical formula. to correct,
The process model construction system according to claim 1, characterized in that: The second processing unit generates model identification data that associates the process model corrected with the characteristic parameters with actual values included in the past operation performance data used when constructing the process model before correction. Performing model identification processing to perform identification on a pair of the process model and the actual value that have been created and corrected with the characteristic parameters included in the created model identification data,
The first processing unit sets, as initial values of the parameters of the process model, parameters of the process model corrected by the characteristic parameters obtained by the model identification process.
The process model construction system according to claim 1, characterized in that: The process model construction system has an input unit that receives input of the characteristic parameter,
The second processing unit sets at least some of the characteristic parameters to the parameters input from the input unit, and performs the calculation for parameters other than the set parameters.
The process model construction system according to claim 1, characterized in that: When the difference between the simulation result and the control result of the controlled object exceeds a predetermined threshold, the third processing unit corrects the characteristic parameter of the process model so that the difference becomes smaller; re-running the simulation;
3. The process model construction system according to claim 2. A process model construction system that uses a computer to construct a process model that models the characteristics of a controlled object,
a first processing unit that acquires the characteristics from past operation performance data indicating an operation performance according to the characteristics of the controlled object;
Based on the difference between the actual value of the process of the controlled object regarding the characteristic included in the past operation performance data and the estimated value of the controlled object estimated by the process model constructed by predetermined machine learning, a second processing unit that calculates a characteristic model obtained by correcting the process model for each of the characteristics;
A process model construction system comprising: Input information including past operation performance data indicating operation performance according to the characteristics of the controlled object, and a process model for modeling the characteristics of the controlled object obtained from the past operation performance data;
including a characteristic parameter for correcting the process model calculated for each characteristic of the controlled object so that the operation performance and the estimated value of the controlled object estimated by the process model satisfy a predetermined condition. displaying the output information on the computer screen as the pre- and post-processing states;
A process model construction system characterized by: The third processing unit performs correction using input information including the characteristic information and simulation input information defining desired operating conditions for performing the simulation, and the characteristic parameters having the same characteristics as the characteristic information. A process model is created, and output information including the simulation results obtained by inputting the simulation input information into the created process model and executing the simulation is displayed on the screen of the computer as a state before and after processing. do,
3. The process model construction system according to claim 2. The fourth processing unit receives input information including desired operating conditions for the controlled object, characteristic information indicating the characteristics of the controlled object and the types of characteristics included in the characteristic, and input information including desired operating conditions for the controlled object, and characteristic information indicating the characteristics of the controlled object and the types of characteristics included in the characteristic information. Output information including input values to the controlled object that match the results of simulating the behavior of the controlled object process using the process model corrected using the characteristic parameters is displayed on the screen of the computer as the pre- and post-processing states. indicate,
The process model construction system according to claim 3, characterized in that: The first processing unit acquires a first device and a first product type indicating characteristics of the controlled object from the past operation performance data,
The second processing unit determines the characteristic parameter such that the actual value of the process to be controlled and the estimated value of the process to be controlled for the first device and the first product type satisfy a predetermined condition. is calculated for each of the first equipment and the first product type, and at least the process model when the first product is generated by the first equipment is calculated based on the calculated characteristic parameters. to correct,
The process model construction system according to claim 1, characterized in that: A process model construction method for constructing a process model modeling the characteristics of a controlled object using a computer, the method comprising:
Obtaining the characteristics from past operation performance data indicating operation performance according to the characteristics of the controlled object,
The actual value of the process to be controlled with respect to the characteristic included in the past operation performance data and the estimated value of the process to be controlled estimated by the process model constructed using a predetermined theoretical formula meet a predetermined condition. calculating characteristic parameters for correcting the process model for each of the characteristics, such that the process model satisfies the
A process model construction method characterized by:

Images