Plant model construction device, plant model construction method and program

Abstract

To provide a plant model construction device that can achieve appropriate operation support of plants.SOLUTION: Provided is a plant model construction device comprising a common model construction unit for constructing a common model, which is a physical model based on plant data from a plurality of plants, and an individual model construction unit for constructing an individual model to control a prescribed target plant on the basis of the common model. Provided is also a plant model construction method comprising a step of constructing a common model, which is a physical model based on plant data from a plurality of plants, and a step of constructing an individual model to control a prescribed target plant on the basis of the common model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a plant model construction device, a plant model construction method, and a program. [Background technology]

[0002] Patent Document 1 describes an information processing method that "collects sensor value data from manufacturing equipment from multiple companies to prepare multiple learning models that are rich in variation and have excellent detection accuracy, and allows each company that is the user of the information processing equipment to appropriately select and use the learning model that is required or the most suitable learning model." [Prior art document] [Patent documents] [Patent Document 1] JP 2022-128069 A [Patent Document 2] International Publication No. 2019 / 159883 [Patent Document 3] JP 2011-118574 A Summary of the Invention

[0003] A first aspect of the present invention provides a plant model construction device including: a common model construction unit that constructs a common model, which is a physical model based on plant data from a plurality of plants; and an individual model construction unit that constructs an individual model for controlling a predetermined target plant based on the common model.

[0004] The plant model construction device may include a model updating unit that updates the common model based on the plant data from the plurality of plants and updates the individual model based on the plant data from the target plant.

[0005] In any of the plant model construction devices described above, the model update unit may update the common model when a predetermined first update condition is satisfied, and update the individual models in accordance with the update of the common model.

[0006] In any of the above plant model construction devices, the model update unit may update the individual models without updating the common model when a predetermined second update condition is satisfied.

[0007] Any of the above-described plant model construction devices may include an input unit that inputs instructions from an operator of the target plant, and the model update unit may update the individual model in response to the input instructions from the operator.

[0008] Any of the above-described plant model construction devices may include an evaluation result receiving unit that receives, from the plurality of plants, an evaluation result of evaluating accuracy of the individual models of the plurality of plants. The model updating unit may update the common model or the individual models based on the evaluation result.

[0009] In any of the above plant model construction devices, the individual model may include a hybrid model of a physical model and a machine learning model.

[0010] In any of the above-described plant model construction devices, the individual model construction unit may construct the individual model in which internal parameters of the common model have been changed by executing machine learning using the plant data of the target plant.

[0011] In any of the above plant model construction devices, the individual model may include a machine learning model.

[0012] In any of the above-described plant model construction devices, the individual model construction unit may construct the individual model that outputs an operation value for controlling the target plant in response to input of the output of the common model by performing machine learning using the plant data of the target plant and the output of the common model.

[0013] In any of the above-described plant model construction devices, the individual model construction unit may execute machine learning using the plant data of the target plant and an output of the common model to construct the individual model that outputs a correction value for correcting an operating value for controlling the target plant, the operating value being output by the common model in response to input of the plant data of the target plant.

[0014] In any of the above plant model construction devices, the individual model construction unit may construct a plurality of individual models for controlling the target plant.

[0015] In any of the above plant model construction devices, the common model may include a plurality of common models corresponding to the plurality of individual models.

[0016] In any of the plant model construction devices described above, the common model construction unit may construct the common model based on the plant data from the plurality of plants and raw material information.

[0017] In any of the above plant model construction devices, the individual model construction unit may construct the individual model based on the common model and the raw material information.

[0018] In any of the above-described plant model construction devices, the raw material information may include at least one of information on the ingredients of the raw material, information on the strength, purity, color or grade of the raw material, information on the raw material provider, information on the original product of the raw material, information on the usage status of the original product of the raw material, and information on the storage of the raw material.

[0019] In any of the above-described plant model construction devices, the common model may be a model that outputs an operation value for controlling the target plant in response to input of the plant data of the target plant, and the common model may calculate and output an intermediate variable that associates the plant data of the target plant with the operation value.

[0020] In any of the above-described plant model construction devices, the individual model may be a model that outputs an operation value for controlling the target plant in response to input of the plant data of the target plant, and the individual model may calculate and output an intermediate variable that associates the plant data of the target plant with the operation value.

[0021] A second aspect of the present invention provides a plant model construction method including: a step of constructing a common model, which is a physical model based on plant data from a plurality of plants; and a step of constructing individual models for controlling predetermined target plants based on the common model.

[0022] In a third aspect of the present invention, there is provided a program that is executed by a computer to cause the computer to operate as a common model construction unit that constructs a common model, which is a physical model based on plant data from a plurality of plants, and an individual model construction unit that constructs individual models for controlling predetermined target plants based on the common model.

[0023] The above summary of the invention does not list all of the features of the present invention, and subcombinations of these features may also be inventions. [Brief explanation of the drawings]

[0024] [Figure 1] An example of the configuration of the plant model construction device 100 is shown together with a plant 200. [Figure 2A] An example of a method for constructing the individual model 122 will be described below. [Figure 2B] An example of a method for constructing the individual model 122 will be described below. [Figure 2C] An example of a method for constructing the individual model 122 will be described below. [Figure 3A] An example of an intermediate variable is shown below. [Figure 3B] An example of an intermediate variable is shown below. [Figure 4]An example of a method for constructing a plant model will be shown below. [Figure 5] 1 shows an example of a flow of a method for constructing a plant model. [Figure 6] An example of a plant model operation method is shown below. [Figure 7A] 1 shows an example of a flow of a plant model operation method. [Figure 7B] 1 shows an example of a flow of a plant model operation method. [Figure 8] A modified example of the plant model construction device 100 is shown together with a plant 200. [Figure 9] An example of a method for updating a plant model will be described. [Figure 10] A modified example of the plant model construction device 100 is shown together with a plant 200, a raw material information provider 300, and a raw material provider 400. [Figure 11] 1 illustrates an example computer 1000 in which aspects of the present invention may be embodied in whole or in part. DETAILED DESCRIPTION OF THE INVENTION

[0025] The present invention will be described below through embodiments of the invention, but the following embodiments do not limit the scope of the invention according to the claims. Furthermore, not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0026] FIG. 1 shows an example of the configuration of a plant model construction device 100 together with a plant 200. Note that the blocks shown in the figure are functionally separated functional blocks and do not necessarily correspond to the actual device configuration. That is, a block shown as one block in this figure does not necessarily have to be configured by one device. Also, blocks shown as separate blocks in this figure do not necessarily have to be configured by separate devices.

[0027] In this example, N plants are shown, including plant 200-1, plant 200-2, ..., plant 200-N. These N plants 200 may be of the same type. In the following description, when there is no need to particularly distinguish between the plants 200, they may be simply referred to as plant 200, and when there is a need to distinguish between the plants 200, they may be referred to as plant 200-1, etc. In other words, a feature described simply referring to plant 200 may be a feature common to multiple plants 200-1, 200-2, ..., 200-N, and a feature described referring to plant 200-1, etc. may be a feature related to a specific plant 200.

[0028] The plant 200 may be a recycling plant for reusing or regenerating waste, an industrial plant such as a chemical or bio plant, a plant that manages and controls wellheads or their surrounding areas of gas fields or oil fields, a plant that manages and controls power generation such as hydroelectric, thermal, and / or nuclear power, a plant that manages and controls energy generation such as solar or wind power, a plant that manages and controls water supply and sewage systems or dams, etc. The plant 200 may include an acquisition unit 210, a control unit 220, and a model storage unit 230.

[0029] The acquisition unit 210 may acquire plant data of the plant 200. The plant data may include data indicating various conditions (physical quantities) inside and outside the plant 200, may include data indicating operation values ​​for controlling the plant 200, may include operation data indicating the operating state as a result of controlling the plant 200, may include consumption data indicating the amount of energy or raw material consumed in the plant 200, may include disturbance environment data indicating physical quantities that may act as disturbances to the control of the plant 200, etc. However, the types of data included in the plant data are not limited to these.

[0030] For example, the acquisition unit 210 may include one or more sensors capable of measuring various conditions inside and outside the plant 200, and acquire data indicating various conditions inside and outside the plant 200 via the one or more sensors. The acquisition unit 210 may acquire data indicating operation values ​​for controlling the plant 200 from the control unit 220. In this way, the acquisition unit 210 may include a configuration for acquiring plant data depending on the type of data included in the plant data, and may acquire plant data from another configuration having plant data.

[0031] The physical quantities included in the plant data may include, for example, the temperature, pressure, flow rate, level, speed, weight, molecular weight, and pH of an object in the process of the plant 200. However, the physical quantities included in the plant data are not limited to these.

[0032] The acquiring unit 210 may supply the acquired plant data to the plant model construction device 100. The acquiring unit 210 may supply the acquired plant data to the control unit 220.

[0033] The control unit 220 controls the plant 200. The control unit 220 may control the plant 200 based on the plant data acquired by the acquisition unit 210 and the common model 112 and / or the individual models 122 constructed by the plant model construction device 100. Details of the common model 112 and the individual models 122 will be described later. The control unit 220 may control the plant 200 based on the output results of the common model 112 and / or the individual models 122.

[0034] The control unit 220 may control the plant 200 using the individual model 122 corresponding to the plant 200 in which the control unit 220 is provided. For example, the control unit 220 provided in the plant 200-1 controls the plant 200-1 using the individual model 122-1 corresponding to the plant 200-1, the control unit 220 provided in the plant 200-2 controls the plant 200-2 using the individual model 122-2 corresponding to the plant 200-2, and the control unit 220 provided in the plant 200-N controls the plant 200-N using the individual model 122-N corresponding to the plant 200-N. The control units 220 provided in each of the plants 200-1, 200-2, ..., 200-N may use the same common model 112 to control each of the plants 200-1, 200-2, ..., 200-N.

[0035] The control unit 220 may control the plant 200 by controlling a controlled object. The controlled object may be a device provided in the plant 200 that is to be controlled. For example, the controlled object is an actuation element that controls a physical quantity in a process of the plant 200, and includes actuators such as valves, heaters, motors, fans, and switches. The control unit 220 may execute a given operation according to an actuation value. The controlled object may also be a controller that controls the actuation element. In other words, the term "control" used in this specification may be broadly interpreted to include not only direct control of an actuation element but also indirect control of an actuation element via a controller.

[0036] The model store 230 may store the common model 112 and / or the individual models 122. When only the common model 112 is used to control the plant 200, the model store 230 may store only the common model 112, or may store both the common model 112 and the individual models 122. When only the individual models 122 are used to control the plant 200, the model store 230 may store only the individual models 122, or may store both the common model 112 and the individual models 122. When both the common model 112 and the individual models 122 are used to control the plant 200, the model store 230 may store the common model 112 and the individual models 122.

[0037] The model storage unit 230 may store a plurality of common models 112 and / or a plurality of individual models 122. The control unit 220 may select a common model 112 and / or an individual model 122 from the plurality of common models 112 and / or the plurality of individual models 122 stored in the model storage unit 230 according to the operating state of the plant 200, and control the plant 200 using the selected common model 112 and / or individual model 122. The control unit 220 may input plant data to the common model 112 and / or the individual model 122.

[0038] The plant model construction device 100 generates a plant model based on plant data from multiple plants 200. The plant model construction device 100 may be a computer such as a personal computer (PC), a tablet computer, a smartphone, a workstation, a server computer, or a general-purpose computer, or may be a computer system in which multiple computers are connected. Such a computer system is also considered a computer in a broad sense. The plant model construction device 100 may be implemented using one or more virtual computer environments executable within a computer. Alternatively, the plant model construction device 100 may be a dedicated computer designed for constructing a plant model, or may be dedicated hardware implemented using dedicated circuits. If the plant model construction device 100 is connectable to the Internet, it may be implemented using cloud computing.

[0039] The plant model construction device 100 includes a common model construction unit 110 and an individual model construction unit 120. The plant model construction device 100 may include a storage unit .

[0040] The common model construction unit 110 constructs a common model 112, which is a physical model based on plant data from the multiple plants 200. The common model 112 may be a model that outputs an operation value for controlling a target plant in response to input of plant data for the target plant. The target plant is the plant 200 that is the control target among the multiple plants 200. For example, when plant 200-1 is to be controlled, the target plant is plant 200-1. In this case, the common model 112 may output an operation value for controlling plant 200-1 in response to input of plant data for plant 200-1. However, the target plant is not limited to this.

[0041] For example, the common model 112 may estimate or predict the state of the target plant based on plant data of the target plant, and may output information necessary for supporting the operation of the target plant.

[0042] The common model construction unit 110 of this example constructs the common model 112 based on plant data from multiple plants 200. This allows the common model 112 to be constructed from a large amount of data, making it possible to construct a more accurate common model 112 compared to constructing a physical model based on plant data from a single plant 200. Whether the common model 112 is used in any of the plants 200-1, 200-2, ... 200-N, appropriate operation support for the plant 200 can be realized.

[0043] The common model 112 may calculate and output intermediate variables that associate plant data of the target plant with the operation values. The intermediate variables may be variables that indicate the process up to deriving the operation values ​​from the plant data. Details of the intermediate variables will be described later.

[0044] In the plant model construction device 100 of this example, the common model construction unit 110 constructs the common model 112, which is a physical model, thereby improving interpretability in operational support for the plant 200. For example, when a machine learning model that outputs an operating value in response to input plant data is used, the process of deriving the operating value from the plant data becomes a black box, which can make it difficult for an operator to interpret why the operating value was output. In contrast, because the plant model construction device 100 of this example constructs the common model 112, which is a physical model, the operator can interpret why the operating value was output by appropriately referring to intermediate variables calculated by the common model 112. For example, when the operating value output from the common model 112 contradicts the operator's intuition based on their experience, it may be important to interpret why the operating value was output.

[0045] In this example, the common model 112 is described as a physical model, but the common model 112 may be a hybrid model of a physical model and a machine learning model. Even in this case, the operator can interpret the reason why the operation value was output by appropriately referring to the intermediate variables calculated by the common model 112.

[0046] The individual model construction unit 120 constructs an individual model 122 for controlling a predetermined target plant based on the common model 112. The individual model 122 may be a model that outputs an operation value for controlling the target plant like the common model 112, or may be a model for correcting the operation value output by the common model 112. Details of the individual model 122 will be described later.

[0047] The individual model construction unit 120 may construct multiple individual models 122, each corresponding to one of the multiple plants 200. In this example, the individual model construction unit 120 constructs multiple individual models 122-1, 122-2, ..., 122-N corresponding to the multiple plants 200-1, 200-2, ..., 200-N. The individual model construction unit 120 may construct the individual models 122 corresponding to the target plants based on the common model 112 and plant data from the target plants. For example, the individual model construction unit 120 constructs the individual model 122-1 corresponding to the plant 200-1 based on the common model 112 and the plant data from the plant 200-1, constructs the individual model 122-2 corresponding to the plant 200-2 based on the common model 112 and the plant data from the plant 200-2, and constructs the individual model 122-N corresponding to the plant 200-N based on the common model 112 and the plant data from the plant 200-N.

[0048] The plant data from the target plant used to build the individual model 122 may or may not be included in the plant data from the multiple plants 200 used to build the common model 112. For example, when the target plant is plant 200-1, the common model building unit 110 may build the common model 112 based on the plant data from the multiple plants 200-1, 200-2, ..., 200-N, or may build the common model 112 based on the plant data from the multiple plants 200-2, ..., 200-N.

[0049] The individual model construction unit 120 of this example constructs an individual model 122 for controlling the target plant based on the common model 112. This allows the plant model construction device 100 of this example to realize appropriate operation support according to the target plant.

[0050] As described above, the plant model construction device 100 of this example includes the common model construction unit 110 and the individual model construction unit 120. This makes it possible to realize appropriate operation support for the plant 200 using the highly accurate common model 112 and improve the interpretability of the operation support for the plant 200, while also realizing appropriate operation support for the target plant using the individual model 122.

[0051] In this example, operation support is described as being realized using the individual models 122, but operation support may also be implemented using only the common model 112. For example, at the beginning of operation of the target plant, there may be no plant data from the target plant or only a small amount of plant data, so that the individual models 122 cannot be constructed. Even in this case, the common model construction unit 110 constructs the common model 112 based on plant data from the multiple plants 200, so that highly accurate operation support for the target plant can be realized.

[0052] The plant model construction device 100 may supply the common model 112 and / or the individual models 122 to the plant 200. The model storage unit 230 of the plant 200 may store the common model 112 and / or the individual models 122 supplied from the plant model construction device 100. That is, the common model 112 and the individual models 122 may be constructed by the plant model construction device 100 in the construction phase and may be stored in each plant 200 in the operation phase.

[0053] In this example, the common model 112 and / or the individual models 122 are described as being stored in each plant 200 during the operation phase, but the common model 112 and / or the individual models 122 may also be stored in the plant model construction device 100 during the operation phase. In this case, the target plant (e.g., plant 200-1) may supply acquired plant data to the plant model construction device 100, which may then execute a model calculation using the corresponding individual model 122 (e.g., individual model 122-1), and supply information necessary for operation support, which is an output result of the model, to the target plant (e.g., plant 200-1).

[0054] The storage unit 130 may store plant data from a plurality of plants 200. The storage unit 130 may store the plant data from the plurality of plants 200 by linking them to each plant 200. For example, the storage unit 130 stores the plant data from the plant 200-1 together with an identifier indicating that the plant data is from the plant 200-1, stores the plant data from the plant 200-2 together with an identifier indicating that the plant data is from the plant 200-2, and stores the plant data from the plant 200-N together with an identifier indicating that the plant data is from the plant 200-N.

[0055] The common model construction unit 110 may construct the common model 112 based on the plant data stored in the storage unit 130. The individual model construction unit 120 may construct the individual model 122 based on the plant data stored in the storage unit 130. For example, when the individual model construction unit 120 constructs the individual model 122-1 corresponding to the plant 200-1, the individual model construction unit 120 constructs the individual model 122-1 based on the common model 112 and the plant data stored in association with the plant 200-1.

[0056] 2A shows an example of a method for constructing the individual model 122. The individual model 122 may include a hybrid model of a physics model and a machine learning model. For example, the individual model 122 includes a hybrid model that combines the common model 112, which is a physics model, with machine learning.

[0057] The individual model construction unit 120 may construct an individual model 122 in which the internal parameters of the common model 112 have been changed by executing machine learning using plant data of the target plant. For example, if the target plant is plant 200-1, the individual model construction unit 120 constructs an individual model 122-1 in which the internal parameters of the common model 112 have been changed by executing machine learning using plant data of plant 200-1. In this manner, the individual model 122-1 adapted to plant 200-1 is constructed. The individual model 122 may be a model that outputs an operation value for controlling the target plant in response to input of plant data of the target plant.

[0058] The individual model construction unit 120 of this example uses plant data of the target plant to construct the individual model 122. This allows the individual model 122 to more accurately represent the characteristics of the target plant, so the plant model construction device 100 of this example can realize appropriate operation support according to the target plant.

[0059] The individual model 122 may calculate and output intermediate variables that associate plant data of the target plant with operation values. In the plant model construction device 100 of this example, the individual model construction unit 120 constructs the individual model 122, which is a hybrid model of a physical model and a machine learning model, and therefore, the interpretability in operation support for the plant 200 can be improved. In other words, since the plant model construction device 100 of this example constructs the individual model 122, which is a hybrid model of a physical model and a machine learning model, an operator can interpret the reason why the operation value was output by appropriately referring to the intermediate variables calculated by the individual model 122.

[0060] Furthermore, the individual model construction unit 120 of this example constructs an individual model 122 in which the internal parameters of the common model 112 have been changed. As a result, by comparing the parameters of the common model 112 and the individual model 122, it is possible to deepen understanding of the characteristics of the plant 200, thereby improving interpretability in operational support for the plant 200. The operator can deepen his or her understanding of the process of the plant 200.

[0061] 2B illustrates an example of a method for constructing the individual model 122. The individual model 122 may include a machine learning model or a statistical model. In one example, the individual model includes a machine learning model.

[0062] The individual model construction unit 120 may construct an individual model 122 that outputs an operation value for controlling the target plant in response to input of the output of the common model 112, by performing machine learning using plant data of the target plant and the output of the common model 112. That is, the individual model 122 may be a model that receives an operation value that is the output of the common model 112 as an input and outputs an operation value that is more suitable for the target plant. For example, if the target plant is plant 200-1, the individual model construction unit 120 constructs an individual model 122-1 that outputs an operation value for controlling the plant 200-1 in response to input of the output of the common model 112, by performing machine learning using plant data of the plant 200-1 and the output of the common model 112. In this manner, the individual model 122-1 adapted to the plant 200-1 is constructed.

[0063] The individual model construction unit 120 of this example uses plant data of the target plant to construct the individual model 122. This allows the individual model 122 to more accurately represent the characteristics of the target plant, so the plant model construction device 100 of this example can realize appropriate operation support according to the target plant.

[0064] Furthermore, the individual model construction unit 120 of this example constructs an individual model 122 that outputs an operation value for controlling the target plant in response to input of the output of the common model 112. As a result, by analyzing the individual model 122, it is possible to understand the characteristics of the error of the common model 112, thereby improving the interpretability in operation support for the plant 200.

[0065] 2C illustrates an example of a method for constructing the individual model 122. The individual model 122 may include a machine learning model or a statistical model. In one example, the individual model includes a machine learning model.

[0066] The individual model construction unit 120 may perform machine learning using plant data of the target plant and the output of the common model 112 to construct an individual model 122 that outputs a correction value for correcting an operating value for controlling the target plant, the operating value being output by the common model 112 in response to input of plant data for the target plant. That is, the individual model 122 may be a model that receives input of plant data for the target plant and outputs a correction value for correcting an operating value that the common model 112 outputs when the same plant data is input, to an operating value more suitable for the target plant. For example, if the target plant is plant 200-1, the individual model construction unit 120 may perform machine learning using plant data for plant 200-1 and the output of the common model 112 to construct an individual model 122-1 that outputs a correction value for correcting an operating value for controlling plant 200-1, the operating value being output by the common model 112 in response to input of plant data for plant 200-1. In this manner, the individual model 122-1 adapted to plant 200-1 is constructed.

[0067] The individual model construction unit 120 of this example uses plant data of the target plant to construct the individual model 122. This allows the individual model 122 to more accurately represent the characteristics of the target plant, so the plant model construction device 100 of this example can realize appropriate operation support according to the target plant.

[0068] Furthermore, the individual model construction unit 120 of this example constructs an individual model 122 that outputs a correction value for correcting an operation value for controlling the target plant output by the common model 112 in response to input of plant data of the target plant. As a result, by analyzing the individual model 122, it is possible to understand the characteristics of the error of the common model 112, and it is possible to improve interpretability in operation support for the plant 200.

[0069] 2A to 2C are examples of methods for constructing the individual model 122, but the methods for constructing the individual model 122 are not limited to these. The individual model constructing unit 120 may construct the individual model 122 by a method that combines at least two of the methods for constructing the individual model 122 described in Figures 2A to 2C.

[0070] 3A shows an example of an intermediate variable. The plant 200 in this example is a waste plastic recycling plant. The plant 200 may generate molten plastic by heating the waste plastic in a decomposition layer. The vaporized oil generated by the melting of the waste plastic may be collected as liquid oil in a collection tank.

[0071] For example, the plant data input to the common model 112 includes data on the materials of the waste plastics and data indicating operating conditions related to heating. In response to the input of such plant data, the common model 112 may output operating values ​​for improving the decomposition efficiency of the waste plastics. The common model 112 calculates various intermediate variables in the process of deriving operating values ​​from the input plant data. For example, the common model 112 calculates the molecular weight distribution of oil (vapor) and the molecular weight distribution of dissolved plastics as intermediate variables.

[0072] In this way, the common model 112, which is a physical model, calculates intermediate variables that indicate the process up to deriving the operation value from the plant data. As a result, the operator can interpret the reason why the operation value was output by appropriately referring to the intermediate variables calculated by the common model 112. Therefore, the plant model construction device 100 of this example can improve interpretability in supporting the operation of the plant 200.

[0073] Note that, when the individual model 122 includes a hybrid model of a physical model and a machine learning model, the individual model 122 may also calculate intermediate variables similar to those of the common model 112. This allows the operator to interpret the reason why the operation value was output by appropriately referring to the intermediate variables calculated by the individual model 122. Therefore, the plant model construction device 100 of this example can improve interpretability in supporting the operation of the plant 200.

[0074] 3B shows an example of an intermediate variable. In this example, the common model 112 calculates a predicted value for the intermediate variable in addition to, or instead of, the current value of the intermediate variable. For example, the intermediate variable is a molecular weight distribution.

[0075] In Figure 3B, the molecular weight distribution shown by the solid line is the current value of the intermediate variable. The molecular weight distribution shown by the dotted line is the predicted value of the intermediate variable when operation is continued for a predetermined time with the current manipulated variable. The molecular weight distribution shown by the dashed-dotted line is the predicted value of the intermediate variable when operation is continued for a predetermined time with the manipulated variable output by the common model 112.

[0076] The common model 112 may calculate current values ​​of intermediate variables in response to input of plant data of the target plant. The common model 112 may calculate predicted values ​​of the intermediate variables when operation is continued for a predetermined time period with the current operation values. The common model 112 may calculate predicted values ​​of the intermediate variables when operation is continued for a predetermined time period with the operation values ​​output by the common model 112.

[0077] The common model 112 of this example calculates predicted values ​​of intermediate variables. This allows an operator to verify the validity of the operation values ​​output by the common model 112 based on the calculated predicted values, and therefore the plant model construction device 100 of this example can improve interpretability in supporting the operation of the plant 200. The same applies to the case where the individual models 122 include a hybrid model of a physical model and a machine learning model.

[0078] 4 shows an example of a method for constructing a plant model. In step S110, a common model 112, which is a physical model based on plant data from multiple plants 200, is constructed. For example, the common model construction unit 110 constructs the common model 112. In step S120, an individual model 122 for controlling a predetermined target plant is constructed based on the common model 112. For example, the individual model construction unit 120 constructs the individual model 122. The details of the construction of each model are as described above.

[0079] 5 shows an example of the flow of the plant model construction method. In step S110, the plant model construction device 100 constructs a common model 112.

[0080] In step S112, the operator of the target plant specifies plant data of the target plant for constructing the individual model 122. In step S114, the operator of the target plant specifies a construction method for the individual model 122. For example, the operator may select any one of the construction methods for the individual model 122 described in Figures 2A to 2C, or may select any combination thereof.

[0081] The order of step S112 and step S114 is not limited. The operator may specify the plant data of the target plant before specifying the construction method of the individual models 122, or may specify the plant data of the target plant after specifying the construction method of the individual models 122. Furthermore, the order of step S110, step 112, and step S114 is not limited. The operator may specify the plant data and the construction method of the individual models 122 before the common model 112 is constructed in the plant model construction device 100, or may specify the plant data and the construction method of the individual models 122 after the common model 112 is constructed in the plant model construction device 100.

[0082] In step S120, the plant model construction device 100 constructs an individual model 122. The individual model construction unit 120 may construct the individual model 122 using plant data specified by an operator in a method specified by the operator.

[0083] In step S130, the plant model construction device 100 transmits the models to the plant 200. The plant model construction device 100 may select the models to transmit depending on the construction method of the individual models 122. For example, if the individual models 122 are constructed by the method described in FIG. 2A, the plant model construction device 100 may transmit only the individual models 122 to the plant 200. As another example, if the individual models 122 are constructed by the method described in FIG. 2B or 2C, the plant model construction device 100 may transmit the common model 112 and the individual models 122 to the plant 200. Note that even if the individual models 122 are constructed by the method described in FIG. 2A, the plant model construction device 100 may transmit the common model 112 and the individual models 122 to the plant 200.

[0084] In step S132, the plant 200 receives the plant model. The model storage unit 230 may store the received plant model.

[0085] 6 shows an example of a method for operating a plant model. In step S210, plant data is acquired. For example, the acquisition unit 210 acquires plant data of the plant 200.

[0086] In step S220, an operation value for controlling the plant 200 is output. When the individual models 122 are constructed by the method described in FIG. 2A , the individual models 122 output an operation value for controlling the plant 200 in response to input of the plant data acquired by the acquisition unit 210. When the individual models 122 are constructed by the method described in FIG. 2B , the common model 112 outputs an operation value in response to input of the plant data acquired by the acquisition unit 210, and the individual models 122 output an operation value for controlling the target plant in response to input of the output of the common model 112. When the individual models 122 are constructed by the method described in FIG. 2C , the common model 112 outputs an operation value in response to input of the plant data acquired by the acquisition unit 210, and the individual models 122 output a correction value for correcting the operation value output by the common model 112 in response to input of the plant data acquired by the acquisition unit 210.

[0087] In step S230, the control unit 220 controls the plant 200. For example, the control unit 220 controls the plant 200 using the output results of the common model 112 and / or the individual models 122.

[0088] 7A shows an example of a flow of a plant model operation method. In step S210, plant data is acquired in the plant 200. For example, the acquisition unit 210 acquires the plant data of the plant 200. The acquisition unit 210 may transmit the acquired plant data to the plant model construction device 100.

[0089] In step S212, the plant data is stored. For example, the storage unit 130 stores the plant data of the plant 200. The plant data stored in the storage unit 130 may be used to build and / or update the common model 112 and / or the individual models 122. Updating of each model will be described later.

[0090] In step S220, an operation value for controlling the plant 200 is output. In step S222, an operator checks the output operation value. The operator may check the intermediate variables calculated by the common model 112 and / or the individual models 122 along with the output operation value.

[0091] In step S224, the operator inputs an operation value. The operator may input the operation value output by the model, or may input another operation value. For example, the operator checks intermediate variables in the process of outputting the operation value, and if the operator determines that the output operation value is appropriate, the operator inputs the operation value output by the model. If the operator determines that the output operation value is inappropriate, the operator inputs another operation value.

[0092] In step S230, the control unit 220 controls the plant 200. For example, the control unit 220 controls the plant 200 using an operation value input by an operator.

[0093] Fig. 7B shows an example of a flow of a plant model operation method. The flow of this example differs from the flow of Fig. 7A in that there is no step in which an operator checks an operation value. The rest of the flow may be the same as the flow of Fig. 7A.

[0094] For example, the manipulated value output by the model in step S220 is directly input in step S224 and used for the control in step S230. As an example, the flow of FIG. 7B can be used when the reliability of the manipulated value output by the plant model is sufficiently high. As another example, before the manipulated value is input in step S224, a flow for determining the validity of the manipulated value output in step S222 may be separately executed. That is, in the flow of FIG. 7A, the operator determines the validity of the manipulated value by interpreting the output manipulated value, but the flow for determining the validity of the manipulated value may be automatically executed in plant 200.

[0095] 8 shows a modified example of the plant model construction device 100 together with a plant 200. The plant model construction device 100 of this example differs from the embodiment of FIG. 1 in that it includes a model update unit 140, an input unit 150, an evaluation result receiving unit 160, and a notification unit 170. In this example, differences from the embodiment of FIG. 1 will be particularly described, and the rest may be the same as the embodiment of FIG. 1.

[0096] The model updating unit 140 may update the common model 112 based on plant data from the multiple plants 200, and may update the individual models 122 based on plant data from the target plant. The model updating unit 140 may update the common model 112 based on plant data from the multiple plants 200 acquired after the common model building unit 110 built the common model 112. The model updating unit 140 may update the individual models 122 based on plant data from the target plant acquired after the individual model building unit 120 built the individual models 122.

[0097] The model updating unit 140 may update the common model 112 when a predetermined first update condition is satisfied. The first update condition may include at least one of a condition related to the amount of plant data from the multiple plants 200 accumulated in the storage unit 130 after the common model construction unit 110 constructed the common model 112, a condition related to the amount of plant data from the multiple plants 200 accumulated in the storage unit 130 after the model updating unit 140 updated the common model 112, a condition related to the period that has elapsed since the common model construction unit 110 constructed the common model 112, a condition related to the period that has elapsed since the model updating unit 140 updated the common model 112, a condition related to the presence or absence of an instruction from an operator, or a condition related to the evaluation result of the individual model 122. However, the types of the first update condition are not limited to these.

[0098] For example, the first update condition includes a condition regarding the amount of plant data from the multiple plants 200 accumulated in the memory unit 130 after the common model construction unit 110 constructed the common model 112. The model update unit 140 may update the common model 112 when the amount of plant data from the multiple plants 200 accumulated in the memory unit 130 after the common model construction unit 110 constructed the common model 112 exceeds a predetermined reference amount.

[0099] For example, the first update condition includes a condition regarding the amount of plant data from the multiple plants 200 accumulated in the storage unit 130 after the model update unit 140 updates the common model 112. The model update unit 140 may update the common model 112 when the amount of plant data from the multiple plants 200 accumulated in the storage unit 130 after the model update unit 140 updates the common model 112 exceeds a predetermined reference amount.

[0100] In this way, after the common model 112 is constructed or updated, when a sufficient amount of data has been accumulated to update the common model 112, the model update unit 140 may update the common model 112. The reference amount after the common model 112 is constructed and the reference amount after the common model 112 is updated may be the same or different.

[0101] For example, the first update condition includes a condition regarding the period of time that has elapsed since the common model construction unit 110 constructed the common model 112. The model update unit 140 may update the common model 112 when the period of time that has elapsed since the common model construction unit 110 constructed the common model 112 exceeds a predetermined reference period.

[0102] For example, the first update condition includes a condition regarding the period of time that has elapsed since the model updating unit 140 updated the common model 112. The model updating unit 140 may update the common model 112 when the period of time that has elapsed since the model updating unit 140 updated the common model 112 exceeds a predetermined reference period.

[0103] In this way, when a sufficient period of time has passed since the common model 112 was constructed or updated, the model update unit 140 may update the common model 112. In other words, when a sufficient period of time has passed since the common model 112 was constructed or updated, it may be assumed that a sufficient amount of data has been accumulated to update the common model 112. The reference period after the common model 112 is constructed and the reference period after the common model 112 is updated may be the same or different.

[0104] For example, the first update condition includes a condition regarding the presence or absence of an instruction from an operator. The model update unit 140 may update the common model 112 when an instruction to update the common model 112 is received from the operator.

[0105] For example, the first update condition includes a condition regarding the evaluation result of the individual model 122. Details of the update based on the evaluation result of the individual model 122 will be described later.

[0106] The model update unit 140 may update the individual models 122 in response to an update of the common model 112. The individual models 122 are constructed based on the common model 112. When the common model 112 is updated, the individual models 122 may be updated based on the updated common model 112.

[0107] When a predetermined second update condition is satisfied, the model updating unit 140 may update the individual models 122 without updating the common model 112. That is, the second update condition may be a condition for updating only the individual models 122. The second update condition may include at least one of a condition related to the amount of plant data from the target plant accumulated in the storage unit 130 after the individual model construction unit 120 constructed the individual model 122, a condition related to the amount of plant data from the target plant accumulated in the storage unit 130 after the model updating unit 140 updated the individual model 122, a condition related to the period elapsed since the individual model construction unit 120 constructed the individual model 122, a condition related to the period elapsed since the model updating unit 140 updated the individual model 122, a condition related to the presence or absence of an instruction from an operator, or a condition related to an evaluation result of the individual model 122. However, the types of the second update condition are not limited to these.

[0108] For example, the second update condition includes a condition regarding the amount of plant data from the target plant that has been accumulated in the memory unit 130 after the individual model construction unit 120 constructed the individual model 122. The model update unit 140 may update the individual model 122 when the amount of plant data from the target plant that has been accumulated in the memory unit 130 after the individual model construction unit 120 constructed the individual model 122 exceeds a predetermined reference amount.

[0109] For example, the second update condition includes a condition regarding the amount of plant data from the target plant that is accumulated in the storage unit 130 after the model update unit 140 updates the individual model 122. The model update unit 140 may update the individual model 122 when the amount of plant data from the target plant that is accumulated in the storage unit 130 after the model update unit 140 updates the individual model 122 exceeds a predetermined reference amount.

[0110] In this way, after the individual model 122 is constructed or updated, if a sufficient amount of data has been accumulated to update the individual model 122, the model update unit 140 may update the individual model 122. The reference amount after the individual model 122 is constructed and the reference amount after the individual model 122 is updated may be the same or different.

[0111] The reference amount for updating the common model 112 may be different from the reference amount for updating the individual models 122. The reference amount for updating the common model 112 may be greater than the reference amount for updating the individual models 122. As an example, when the amount of plant data from the multiple plants 200 is less than the reference amount for updating the common model 112, but the amount of plant data from the target plant is greater than the reference amount for updating the individual models 122, the common model 112 is not updated, and the individual model 122 corresponding to the target plant is updated.

[0112] For example, the second update condition includes a condition regarding the period of time that has elapsed since the individual model construction unit 120 constructed the individual model 122. The model update unit 140 may update the individual model 122 when the period of time that has elapsed since the individual model construction unit 120 constructed the individual model 122 exceeds a predetermined reference period.

[0113] For example, the second update condition includes a condition regarding the period of time that has elapsed since the model update unit 140 updated the individual model 122. The model update unit 140 may update the individual model 122 when the period of time that has elapsed since the model update unit 140 updated the individual model 122 exceeds a predetermined reference period.

[0114] In this way, when a sufficient period of time has passed since the individual model 122 was constructed or updated, the model update unit 140 may update the individual model 122. In other words, when a sufficient period of time has passed since the individual model 122 was constructed or updated, it may be assumed that a sufficient amount of data has been accumulated to update the individual model 122. The reference period after the individual model 122 is constructed and the reference period after the individual model 122 is updated may be the same or different.

[0115] The reference period for updating the common model 112 and the reference period for updating the individual models 122 may be the same or different. The common model 112 is updated based on plant data from the multiple plants 200, and the individual models 122 are updated based on plant data from the target plant. Therefore, the rate at which plant data from the multiple plants 200 for updating the common model 112 is accumulated differs from the rate at which plant data from the target plant for updating the individual models 122 is accumulated, and therefore the reference period may be set appropriately for each.

[0116] For example, the second update condition includes a condition regarding the presence or absence of an instruction from an operator. The model update unit 140 may update the common model 112 when there is an instruction from an operator to update the individual model 122.

[0117] For example, the second update condition includes a condition regarding the evaluation result of the individual model 122. Details of the update based on the evaluation result of the individual model 122 will be described later.

[0118] The input unit 150 may input instructions from an operator of the target plant. For example, the input unit 150 may display a screen for inputting instructions from the operator on a display device or the like in the target plant, and allow the operator to input instructions.

[0119] The model update unit 140 may update the common model 112 in response to an instruction input from an operator. The model update unit 140 may update the individual model 122 in response to an instruction input from an operator. That is, the first update condition and / or the second update condition may include a condition regarding the presence or absence of an instruction from the operator.

[0120] The plant model construction device 100 of this example can improve the interpretability in supporting the operation of the plant 200. This allows the operator of the target plant to appropriately instruct the updating of the model based on the interpretation result.

[0121] The evaluation result receiving unit 160 may receive, from the multiple plants 200, evaluation results obtained by evaluating the accuracy of each of the individual models 122 of the multiple plants 200. For example, the evaluation result receiving unit 160 receives, from the plant 200-1, an evaluation result obtained by evaluating the accuracy of the individual model 122-1, from the plant 200-2, an evaluation result obtained by evaluating the accuracy of the individual model 122-2, and from the plant 200-N, an evaluation result obtained by evaluating the accuracy of the individual model 122-N.

[0122] The plant 200 may include an evaluation unit 240 that evaluates the accuracy of the individual model 122. The evaluation unit 240 may evaluate the accuracy of the individual model 122 by comparing the output of the individual model 122 with the actual operating state of the target plant. The evaluation unit 240 may transmit an evaluation result of the evaluation of the accuracy of the individual model 122 to the evaluation result receiving unit 160.

[0123] Based on the evaluation result, the model update unit 140 may update the common model 112 or the individual model 122. The first update condition and the second update condition may include a condition regarding the evaluation result of the individual model 122.

[0124] For example, when the proportion of evaluation results evaluating the accuracy of the individual models 122 of the multiple plants 200 in which the accuracy of the individual models 122 is low is higher than a predetermined reference proportion, the model updating unit 140 may update the common model 112. That is, since the individual models 122 of the multiple plants 200 are constructed based on the common model 112, when the proportion of evaluation results evaluating the accuracy of the individual models 122 in which the accuracy is low is high, the accuracy of the common model 112 may be low. Therefore, when the proportion of evaluation results evaluating the accuracy of the individual models 122 in which the accuracy is low is high, the model updating unit 140 may update the common model 112.

[0125] For example, when, among the evaluation results of evaluating the accuracy of the individual models 122 of the multiple plants 200, the proportion of evaluation results evaluating the accuracy of the individual models 122 as high is higher than a predetermined reference proportion, and the accuracy of the individual model 122 of the target plant is evaluated as low, the model updating unit 140 may update the individual model 122 of the target plant. In other words, when the proportion of evaluation results evaluating the accuracy of the individual model 122 as high is high, it is estimated that the accuracy of the common model 112 is also high, and therefore, when the accuracy of the individual model 122 of the target plant is evaluated as low, the model updating unit 140 may update the individual model 122 evaluated as having low accuracy.

[0126] When the common model 112 or the individual model 122 is updated, the notification unit 170 may notify the operator that an update has occurred. For example, the notification unit 170 notifies the operator that an update has occurred. This can reduce confusion that may occur when the common model 112 or the individual model 122 is updated.

[0127] The plant model construction device 100 of this example includes a model update unit 140 that updates the common model 112 and the individual models 122. This makes it possible to maintain each model in an updated state based on the amount of accumulated data, the period over which the data has been accumulated, instructions from an operator, the accuracy of the individual models 122, and the like. Furthermore, because each model can be maintained in an updated state, the plant model construction device 100 of this example can realize appropriate operation support for the plant 200.

[0128] 9 shows an example of a method for updating a plant model. In step S310, it is determined whether a predetermined first update condition is satisfied. The details of the first update condition are as described above.

[0129] If the predetermined first update condition is satisfied in step S310 (yes), the common model 112 is updated in step S312. For example, the model update unit 140 updates the common model 112. In step S314, the individual models 122 are updated. For example, the model update unit 140 updates the individual models 122 in response to the update of the common model 112. The model update unit 140 may update the individual models 122 based on the updated common model 112.

[0130] If the predetermined first update condition is not satisfied in step S310 (no), it is determined whether the predetermined second update condition is satisfied in step 320. The details of the second update condition are as described above.

[0131] If the predetermined second update condition is satisfied in step S320 (yes), then in step S322, the individual model 122 is updated. For example, the model update unit 140 updates the individual model 122 without updating the common model 112. If the predetermined second update condition is not satisfied in step S320 (no), then none of the models are updated and the flow ends.

[0132] 10 shows a modified example of the plant model construction device 100 together with a plant 200. The plant model construction device 100 of this example differs from the embodiment of FIG. 1 in that the individual model construction unit 120 constructs a plurality of individual models 122 for controlling the target plant. In this example, differences from the embodiment of FIG. 1 will be particularly described, and the rest may be the same as the embodiment of FIG. 1. Note that in this example, due to space limitations, only one plant 200-1 is shown, but the plurality of plants 200 may include a plant 200-2, ..., a plant 200-N, as in FIG. 1.

[0133] The raw material provider 400 provides raw materials to the plant 200. Different raw materials may be provided by multiple raw material providers 400. The raw material provider 400 provides raw material information to the raw material information provider 300.

[0134] The raw material information may include at least one of information on the ingredients of the raw material, information on the strength, purity, color or grade of the raw material, information on the raw material provider 400, information on the original product of the raw material, information on the usage status of the original product of the raw material, or information on the storage of the raw material, but the types of information included in the raw material information are not limited to these.

[0135] The raw material information provider 300 collects raw material information from the raw material providers 400. The raw material information provider 300 may provide the collected raw material information to the plant model construction device 100.

[0136] The individual model construction unit 120 may construct a plurality of individual models 122 for controlling the target plant. In the embodiment of FIG. 1 , the individual model construction unit 120 has also been described as constructing a plurality of individual models 122. This is because the individual model construction unit 120 constructs an individual model 122 corresponding to each of the plurality of plants 200. The individual model construction unit 120 of this example constructs a plurality of individual models 122 corresponding to the target plant. That is, the individual model construction unit 120 may construct a plurality of individual models 122 corresponding to each of the plurality of plants 200.

[0137] For example, the individual model construction unit 120 constructs M individual models 122-1-1, ..., 122-1-M corresponding to the plant 200-1, and constructs M individual models 122-N-1, ..., 122-NM corresponding to the plant 200-N. Note that the number of individual models 122 corresponding to each of the multiple plants 200 may be the same or different. That is, although the individual model construction unit 120 in this example has been described as constructing M individual models 122 for each of the plant 200-1 and the plant 200-N, the individual model construction unit 120 may construct a different number of individual models 122 for each plant 200.

[0138] The individual model construction unit 120 may construct multiple individual models 122 according to the operating conditions of the target plant. For example, the individual model construction unit 120 may construct a different individual model 122 for each raw material handled by the target plant, may construct a different individual model 122 for each raw material provider 400 of the raw materials handled by the target plant, may construct a different individual model 122 for each season in which the target plant operates, or may construct a different individual model 122 for each environment in which the target plant is installed. However, the types of multiple individual models 122 constructed by the individual model construction unit 120 are not limited to these.

[0139] In the plant model operation stage, for example, an operator of the target plant can select and operate an individual model 122 that corresponds to the operating status of the target plant from the plurality of individual models 122. As an example, when it is summer, the operator of the target plant can select and operate an individual model 122 that corresponds to summer from the plurality of individual models 122.

[0140] The common model 112 may include a plurality of common models 112 corresponding to a plurality of individual models 122. In this example, the common model 112 includes, for example, M common models 112-1, ..., 112-M corresponding to M individual models 122-1-1, ..., 122-1-M.

[0141] The common model construction unit 110 may construct multiple common models 112 according to the operating conditions of the target plant. For example, the common model construction unit 110 may construct a different common model 112 for each raw material handled by the target plant, may construct a different common model 112 for each raw material provider 400 of the raw materials handled by the target plant, may construct a different common model 112 for each season in which the target plant operates, or may construct a different common model 112 for each environment in which the target plant is installed. However, the types of multiple common models 112 constructed by the common model construction unit 110 are not limited to these.

[0142] The individual model construction unit 120 of this example constructs a plurality of individual models 122 for controlling the target plant. This allows the operator of the target plant to select and operate an appropriate individual model 122 according to the operating conditions, so the plant model construction device 100 of this example can realize appropriate operation support for the plant 200.

[0143] The common model construction unit 110 may construct a common model 112 based on plant data from multiple plants 200 and raw material information. The individual model construction unit 120 may construct an individual model 122 based on the common model 112 and the raw material information.

[0144] The plant model construction device 100 may acquire raw material information accumulated in the company's own plant 200, or may acquire raw material information accumulated in an external platform. That is, the raw material information used by the common model construction unit 110 to construct the common model 112 may include raw material information accumulated in the company's own plant 200, may include raw material information accumulated in an external platform, or may include both raw material information accumulated in the company's own plant 200 and raw material information accumulated in an external platform. The same applies to the raw material information used by the individual model construction unit 120 to construct the individual model 122.

[0145] The common model construction unit 110 of this example constructs a common model 112 based on the raw material information, and the individual model construction unit 120 constructs an individual model 122 based on the raw material information. This allows the plant model to output appropriate operation values ​​according to the raw material information, so the plant model construction device 100 of this example can realize appropriate operation support for the plant 200. Furthermore, because the raw material information of this example is provided by the raw material information provider 300, the cost for realizing appropriate operation support can be reduced compared to when raw materials are sensed in the plant 200.

[0146] The raw material information in this example includes various information such as information on the ingredients of the raw material, information on the strength, purity, color, or grade of the raw material, information on the raw material provider 400, information on the product from which the raw material was derived, information on the usage status of the product from which the raw material was derived, or information on the storage of the raw material. In other words, the raw material information also includes information that cannot be obtained by sensing the raw material. This makes it possible to achieve more appropriate operational support than when sensing the raw material in the plant 200.

[0147] In this example, the individual model construction unit 120 has been described as constructing the individual model 122 based on the ingredient information, but the individual model construction unit 120 may simply construct the individual model 122 based on the common model 112. In other words, since the common model construction unit 110 constructs the common model 112 based on the ingredient information, even if the individual model construction unit 120 simply constructs the individual model 122 based on the common model 112, it is estimated that the individual model 122 will output an operation value or a correction value that fully reflects the ingredient information.

[0148] The features described in each of Figures 1, 8, and 10 may be implemented together with the features described in the other figures. That is, the plant model construction device 100 according to the present invention may include all or any combination of the features described in Figures 1, 8, and 10.

[0149] Various embodiments of the present invention may be described with reference to flowcharts and block diagrams, where the blocks may represent (1) stages of a process in which operations are performed or (2) sections of an apparatus responsible for performing the operations. Particular stages and sections may be implemented by dedicated circuitry, programmable circuitry provided with computer-readable instructions stored on a computer-readable medium, and / or a processor provided with computer-readable instructions stored on a computer-readable medium. Dedicated circuitry may include digital and / or analog hardware circuitry, and may include integrated circuits (ICs) and / or discrete circuits. Programmable circuitry may include reconfigurable hardware circuitry, including logical AND, OR, XOR, NAND, NOR, and other logical operations, flip-flops, registers, memory elements such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), and the like.

[0150] A computer-readable medium may include any tangible device capable of storing instructions that are executed by an appropriate device, such that the computer-readable medium having instructions stored thereon comprises an article of manufacture containing instructions that can be executed to create means for performing the operations specified in the flowcharts or block diagrams. Examples of computer-readable media may include electronic, magnetic, optical, electromagnetic, and semiconductor storage media. More specific examples of computer-readable media may include floppy disks, diskettes, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), electrically erasable programmable read-only memory (EEPROM), static random access memory (SRAM), compact disc read-only memory (CD-ROM), digital versatile disc (DVD), Blu-ray (RTM) disc, memory stick, integrated circuit card, and the like.

[0151] The computer readable instructions may include either assembler instructions, Instruction Set Architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, state-setting data, or source or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk®, JAVA®, C++, etc., and conventional procedural programming languages ​​such as the “C” programming language or similar programming languages.

[0152] The computer-readable instructions may be provided to a processor or programmable circuit of a programmable data processing device, such as a computer, locally or over a wide area network (WAN) such as a local area network (LAN) or the Internet, and the computer-readable instructions may be executed to create means for performing the operations specified in the flowcharts or block diagrams. Here, the computer may be a personal computer (PC), a tablet computer, a smartphone, a workstation, a server computer, a general-purpose computer, a special-purpose computer, or the like, or may be a computer system in which multiple computers are connected. Such a computer system in which multiple computers are connected is also called a distributed computing system, and is a broad definition of computer. In a distributed computing system, the multiple computers collectively execute a program by each executing a portion of the program and passing data between the computers as needed during program execution.

[0153] Examples of processors include a computer processor, a central processing unit (CPU), a processing unit, a microprocessor, a digital signal processor, a controller, a microcontroller, etc. A computer may have one or more processors. In a multiprocessor system with multiple processors, each processor executes a portion of a program and passes data between processors as needed during program execution, allowing the multiple processors to collectively execute a program. For example, in multitasking, each of the multiple processors may execute a portion of each task in small chunks by switching tasks at time slice intervals. In this case, which portion of a program each processor executes changes dynamically. Which portion of a program each of the multiple processors executes may also be statically determined by multiprocessor-aware programming.

[0154] 11 illustrates an example of a computer 1000 in which aspects of the present invention may be embodied, in whole or in part. A program installed on the computer 1000 may cause the computer 1000 to function as or perform operations associated with an apparatus or one or more sections of the apparatus according to an embodiment of the present invention, and / or to perform a process or steps of the process according to an embodiment of the present invention. Such a program may be executed by the CPU 1012 to cause the computer 1000 to perform specific operations associated with some or all of the blocks of the flowcharts and block diagrams described herein.

[0155] The computer 1000 according to this embodiment includes a CPU 1012, a RAM 1014, a graphics controller 1016, and a display device 1018, which are interconnected by a host controller 1010. The computer 1000 also includes input / output units such as a communication interface 1022, a hard disk drive 1024, a DVD-ROM drive 1026, and an IC card drive, which are connected to the host controller 1010 via an input / output controller 1020. The computer 1000 also includes legacy input / output units such as a ROM 1030 and a keyboard 1042, which are connected to the input / output controller 1020 via an input / output chip 1040.

[0156] The CPU 1012 operates according to programs stored in the ROM 1030 and RAM 1014, thereby controlling each unit. The graphics controller 1016 acquires image data generated by the CPU 1012 into a frame buffer or the like provided in the RAM 1014 or into the graphics controller itself, and causes the image data to be displayed on the display device 1018.

[0157] The communication interface 1022 communicates with other electronic devices via a network. The hard disk drive 1024 stores programs and data used by the CPU 1012 in the computer 1000. The DVD-ROM drive 1026 reads programs or data from a DVD-ROM 1027 and provides the programs or data to the hard disk drive 1024 via the RAM 1014. The IC card drive reads programs and data from an IC card and / or writes programs and data to an IC card.

[0158] The ROM 1030 stores therein a boot program and the like that is executed by the computer 1000 upon activation, and / or programs that depend on the hardware of the computer 1000. The input / output chip 1040 may also connect various input / output units to the input / output controller 1020 via a parallel port, a serial port, a keyboard port, a mouse port, and the like.

[0159] The programs are provided by a computer-readable medium such as a DVD-ROM 1027 or an IC card. The programs are read from the computer-readable medium, installed in the hard disk drive 1024, RAM 1014, or ROM 1030, which are also examples of computer-readable media, and executed by the CPU 1012. Information processing described in these programs is read by the computer 1000, and causes cooperation between the programs and the various types of hardware resources described above. An apparatus or method may be configured by implementing information manipulation or processing in accordance with the use of the computer 1000.

[0160] For example, when communication is performed between the computer 1000 and an external device, the CPU 1012 may execute a communication program loaded into the RAM 1014 and instruct the communication interface 1022 to perform communication processing based on the processing described in the communication program. Under the control of the CPU 1012, the communication interface 1022 reads transmission data stored in a transmission buffer processing area provided in the RAM 1014, the hard disk drive 1024, the DVD-ROM 1027, or a recording medium such as an IC card, and transmits the read transmission data to the network, or writes reception data received from the network to a reception buffer processing area or the like provided on the recording medium.

[0161] The CPU 1012 may also cause all or a necessary portion of a file or database stored on an external recording medium such as a hard disk drive 1024, a DVD-ROM drive 1026 (DVD-ROM 1027), an IC card, etc. to be read into the RAM 1014, and perform various types of processing on the data on the RAM 1014. The CPU 1012 then writes back the processed data to the external recording medium.

[0162] Various types of information, such as various types of programs, data, tables, and databases, may be stored on the recording medium and may undergo information processing. The CPU 1012 may perform various types of processing on data read from the RAM 1014, including various types of operations, information processing, conditional judgment, conditional branching, unconditional branching, information search / replacement, etc., as described throughout this disclosure and specified by the instruction sequences of the programs, and write the results back to the RAM 1014. The CPU 1012 may also search for information in a file, database, etc. on the recording medium. For example, if multiple entries, each having an attribute value of a first attribute associated with an attribute value of a second attribute, are stored on the recording medium, the CPU 1012 may search for an entry that matches a condition specified by the attribute value of the first attribute from among the multiple entries, read the attribute value of the second attribute stored in the entry, and thereby obtain the attribute value of the second attribute associated with the first attribute that satisfies a predetermined condition.

[0163] The above-described programs or software modules may be stored in a computer-readable medium on or near the computer 1000. A recording medium such as a hard disk or RAM provided in a server system connected to a dedicated communication network or the Internet can also be used as a computer-readable medium, thereby providing the programs to the computer 1000 via the network.

[0164] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various modifications and improvements can be made to the above embodiments. It is clear from the claims that such modifications and improvements can also be included within the technical scope of the present invention.

[0165] It should be noted that the execution order of each process, such as operations, procedures, steps, and stages, in the devices, systems, programs, and methods shown in the claims, specifications, and drawings is not specifically stated as "before," "prior to," etc., and that the processes can be performed in any order unless the output of a previous process is used in a subsequent process. Even if the operational flow in the claims, specifications, and drawings is described using "first," "next," etc. for convenience, this does not mean that the processes must be performed in this order. [Explanation of symbols]

[0166] 100 Plant model construction device, 110 Common model construction unit, 112 Common model, 120 Individual model construction unit, 122 Individual model, 130 Memory unit, 140 Model update unit, 150 Input unit, 160 Evaluation result receiving unit, 170 Notification unit, 200 Plant, 210 Acquisition unit, 220 Control unit, 230 Model storage unit, 240 Evaluation unit, 300 Raw material information provider, 400 Raw material provider, 1000 Computer, 1010 Host controller, 1012 CPU, 1014 RAM, 1016 Graphic controller, 1018 Display device, 1020 Input / output controller, 1022 Communication interface, 1024 Hard disk drive, 1026 DVD-ROM drive, 1027 DVD-ROM, 1030 ROM, 1040 Input / output chip, 1042 Keyboard

Claims

1. a common model construction unit that constructs a common model, which is a physical model based on plant data from a plurality of plants; an individual model construction unit that constructs an individual model for controlling a predetermined target plant based on the common model; Equipped with Plant model construction device.

2. a model updating unit that updates the common model based on the plant data from the plurality of plants and updates the individual model based on the plant data from the target plant; The plant model construction device according to claim 1 .

3. The model update unit updating the common model when a first predetermined update condition is satisfied; The individual models are updated in response to updates to the common model. The plant model construction device according to claim 2 .

4. The model update unit updates the individual models without updating the common model when a predetermined second update condition is satisfied. The plant model construction device according to claim 2 .

5. an input unit for inputting instructions from an operator of the target plant; The model update unit updates the individual model in response to an instruction input from the operator. The plant model construction device according to claim 4 .

6. an evaluation result receiving unit that receives, from the plurality of plants, an evaluation result obtained by evaluating the accuracy of the individual model of each of the plurality of plants; The model update unit updates the common model or the individual model based on the evaluation result. The plant model construction device according to claim 2 .

7. The individual model includes a hybrid model of a physical model and a machine learning model. The plant model construction device according to claim 1 .

8. The individual model construction unit constructs the individual model in which parameters inside the common model are changed by executing machine learning using the plant data of the target plant. The plant model construction device according to claim 7 .

9. The individual model includes a machine learning model. The plant model construction device according to claim 1 .

10. The individual model construction unit executes machine learning using the plant data of the target plant and the output of the common model to construct the individual model that outputs an operation value for controlling the target plant in response to input of the output of the common model. The plant model construction device according to claim 9 .

11. The individual model construction unit executes machine learning using the plant data of the target plant and an output of the common model to construct the individual model that outputs a correction value for correcting an operation value for controlling the target plant, the operation value being output by the common model in response to input of the plant data of the target plant. The plant model construction device according to claim 9 .

12. The individual model construction unit constructs a plurality of individual models for controlling the target plant. The plant model construction device according to claim 1 .

13. The common model includes a plurality of common models corresponding to the plurality of individual models. The plant model construction device according to claim 12.

14. The common model construction unit constructs the common model based on the plant data from the plurality of plants and raw material information. The plant model construction device according to claim 1 .

15. The individual model construction unit constructs the individual model based on the common model and the raw material information. The plant model construction device according to claim 14.

16. The raw material information includes at least one of information on ingredients of the raw material, information on the strength, purity, color or grade of the raw material, information on the raw material provider, information on the original product of the raw material, information on the usage status of the original product of the raw material, and information on the storage of the raw material. The plant model construction device according to claim 14.

17. the common model is a model that outputs an operation value for controlling the target plant in response to input of the plant data of the target plant, The common model calculates and outputs intermediate variables that associate the plant data of the target plant with the operation values. The plant model construction device according to claim 1 .

18. the individual model is a model that outputs an operation value for controlling the target plant in response to input of the plant data of the target plant, The individual model calculates and outputs intermediate variables that associate the plant data of the target plant with the operation values. The plant model construction device according to claim 1 .

19. building a common model, which is a physical model based on plant data from multiple plants; constructing an individual model for controlling a predetermined target plant based on the common model; Equipped with How to build a plant model.

20. When executed by a computer, the computer a common model construction unit that constructs a common model, which is a physical model based on plant data from a plurality of plants; an individual model construction unit that constructs an individual model for controlling a predetermined target plant based on the common model; and make it work program.

Images