WO2025158965A1 - Support method, support device, and support program - Google Patents

Abstract

Provided is a method for supporting implementation of a model related to operation of plant equipment. The method is executed by a processor, and includes: acquiring equipment identification information of plant equipment or process identification information of a process executed in the plant equipment; acquiring parameter identification information of a parameter acquired in operation of the plant equipment; acquiring time-series data of the parameter; selecting a model recommended to be implemented from among a plurality of models related to operation of the plant equipment on the basis of the equipment identification information or the process identification information, the parameter identification information, and the time-series data of the parameter; and outputting information of the selected model.

Description

Support method, support device, and support program

The present invention relates to a method, device, and program for supporting the development of models related to the operation of plant equipment.

Various models are used to design, construct, operate, and manage plants. Models are developed for each plant to suit its type, size, location, etc.

Japanese Patent Application Laid-Open No. 2023-016369

Plant managers do not have specialized knowledge about models, so it is difficult for them to know what models they should implement in order to optimally operate the plants they manage. The inventors recognized the need for technology that can accurately select models that are recommended for implementation in accordance with the plant and propose them to plant managers and others.

The present invention was made in light of these circumstances, and its purpose is to provide technology to support the development of models related to the operation of plant equipment.

In order to solve the above problem, one aspect of the present invention is a method for supporting the implementation of a model related to the operation of plant equipment. The method is executed by a processor and includes the steps of acquiring equipment identification information for the plant equipment or process identification information for a process executed in the plant equipment, acquiring parameter identification information for parameters acquired during the operation of the plant equipment, acquiring time-series data for the parameters, selecting a model recommended for implementation from among a plurality of models related to the operation of the plant equipment based on the equipment identification information or process identification information, the parameter identification information, and the time-series data for the parameters, and outputting information about the selected model.

Another aspect of the present invention is a support device. This device includes an information acquisition unit that acquires equipment identification information for plant equipment or process identification information for processes executed in the plant equipment, parameter identification information for parameters acquired during operation of the plant equipment, and time-series data for the parameters; a selection unit that selects a model recommended for implementation from among multiple models related to the operation of the plant equipment based on the equipment identification information or process identification information, the parameter identification information, and the time-series data for the parameters; and an output unit that outputs information about the selected model.

In addition, any combination of the above components, and any transformation of the present invention into a method, device, system, recording medium, computer program, etc., are also valid aspects of the present invention.

The present invention provides technology to support the development of models related to the operation of plant equipment.

FIG. 1 is a diagram schematically illustrating an example of a method for developing a model related to the operation of a plant facility. 1 is a diagram illustrating a configuration of a development support system according to a first embodiment. 1 is a diagram illustrating a configuration of a construction support apparatus according to a first embodiment. 3 is a flowchart showing the procedure of a support method according to the first embodiment. FIG. 10 illustrates an example of a data structure of a processing element storage unit. FIG. 10 illustrates an example of a data structure of a processing element storage unit. 1 is a diagram illustrating a configuration of a development support device according to a first embodiment; 3 is a flowchart showing the procedure of a support method according to the first embodiment. FIG. 2 is a diagram illustrating an example of a screen displayed on a display device of the development support device. FIG. 2 is a diagram illustrating an example of a screen displayed on a display device of the development support device. FIG. 10 is a diagram illustrating a configuration of a development support system according to a second embodiment. FIG. 10 is a sequence diagram showing the procedure of a support method according to a second embodiment. FIG. 10 is a diagram illustrating a configuration of a development support device according to a second embodiment. FIG. 10 is a diagram illustrating an example of internal data of a recommendation level storage unit. FIG. 10 is a diagram illustrating an example of internal data of a recommendation level storage unit. FIG. 10 is a diagram illustrating an example of time-series data of parameters. FIG. 10 is a diagram illustrating an example of a screen displayed on an administrator terminal from the development support device. FIG. 10 is a diagram illustrating an example of a screen displayed on an administrator terminal from the development support device.

First, as a first embodiment of the present disclosure, we will explain a technology that, when developing a model related to the operation of plant equipment, provides a list of processing elements that includes the processing elements that should make up the development process of that model. Next, as a second embodiment of the present disclosure, we will explain a technology that, based on information related to the plant equipment, proposes a model that is recommended for implementation in order to operate the plant equipment optimally.

(First embodiment)
Figure 1 shows a schematic diagram of an example of a method for developing a model related to the operation of a plant facility. This diagram shows the steps of a method for developing an AI (artificial intelligence) that optimizes the operation of the plant facility as an example of a model related to the operation of the plant facility.

In step (1), operational data collected and accumulated from the customer's existing plant equipment is stored in a database server.

In step (2), preprocessing is performed on the driving data stored in the database server. Preprocessing may include, for example, assigning the same tag to identical or similar driving data when different tags are assigned to the driving data, aligning the units of the driving data, adjusting the offset of the driving data, processing outliers in the driving data, normalizing the driving data, supplementing missing driving data, and calculating other data representing physical properties, states, etc. from the driving data using predetermined formulas, algorithms, simulations, etc. Preprocessing may be performed by referring to a tag correspondence table that defines the correspondence between tags assigned to the driving data. The content of the preprocessing may be determined by a data scientist with specialized knowledge for organizing and analyzing data.

In step (3), the data scientist visualizes the preprocessed data by referencing design information such as a process flow diagram (PFD) and heat and material balance (H&MB). The data scientist visualizes the preprocessed data in a format that matches the plant's design information, such as a time series plot, histogram, or box plot. The visualized data may be referenced by a data scientist or process engineer in a subsequent step, or may be provided to the customer. The visualized data may be used, for example, to understand the plant's operating status (product production volume, whether or not operation has been stopped, equipment efficiency, etc.), as well as to select operating data (data type, extraction period) and to confirm the validity of the operating data preprocessing.

In step (4), a process engineer with specialized knowledge of plant equipment design and operation creates a process simulation to simulate the operation of the plant equipment based on design information such as process flow diagrams (PFDs), heat and material balances (H&MBs), and equipment performance. If an existing process simulator exists, step (4) is omitted.

In step (5), the data scientist uses the preprocessed stable region data as input values for the process simulation to simulate the operation of the plant equipment and obtain simulation results such as the plant's thermal efficiency.

In step (6), the data scientist constructs a surrogate model (substitute model) to replace the process simulation. The surrogate model inputs preprocessed stable region data and outputs simulation results without simulating the operation of the plant equipment. The surrogate model may be constructed using a neural network, for example. The data scientist trains the surrogate model using the preprocessed data and the simulation results from the process simulation as training data. For example, when preprocessed data is input into the input layer of the neural network, the middle layer of the neural network may be adjusted so that the output layer of the neural network outputs actual data corresponding to that preprocessed data, or the simulation results output when that preprocessed data is input into the process simulation. Note that the surrogate model may be an emulation model constructed based on operational data or a physical model constructed using first principles.

In step (7), the process engineer defines constraints for optimizing the operation of the plant equipment. Constraints may include upper and lower limits for various parameters. The process engineer may define constraints based on information obtained through customer interviews, equipment design information, instrumentation alarm information, etc.

In step (8), the operation optimization AI searches for values of operation parameters that optimize the operation of the plant equipment under the defined constraints. The operation optimization AI inputs a large amount of preprocessed data into a surrogate model and searches, using a specified optimization algorithm, for input data values that will optimize the value of a specified piece of output data. Because the surrogate model can instantly calculate output data from input data, it can be used to search for optimal solutions from among huge combinations of input data, even for complex processes where simulation calculations take days or even weeks.

The above-mentioned work process flow is similar when developing AI to optimize the operation of other plant equipment. Therefore, if information about the work process when developing a model is accumulated, the next time a similar model is developed, it will be possible to smoothly develop the model using a similar work process by referring to past performance. This makes it easier to formulate work plans and manage work processes for model development, thereby reducing the cost, time, and effort involved in model development.

Some of the work processes involved in the development of operation optimization AI are also performed in the development of operation optimization AI for other plant equipment and other models. For example, operation data is time-series data collected during the operation of plant equipment, and is typically expected to have continuous values. Considering this characteristic, it is difficult to imagine that the data will suddenly change to a completely different value without context. Therefore, the pre-processing of operation data involves removing outliers that have increased or decreased by more than a specified rate from the previous value. This work should also be performed in the development of operation optimization AI for other plant equipment and other models. If such common work is packaged in advance, existing packages can be used when developing new models, significantly reducing the cost, time, and effort required for model development.

Figure 2 shows the configuration of a development support system according to the first embodiment. The development support system 1 includes a development device 3, a design device 4, an operation control device 5, a construction support device 100, a development support device 200, and a communication network 2 for connecting these devices so that they can communicate with each other.

The plant equipment 10 includes facilities, equipment, devices, piping, etc. for carrying out processes in the plant. For example, in a chemical plant, the plant equipment 10 includes reactors, separation equipment, drying equipment, piping, etc.

The development device 3 develops a model related to the operation of the plant equipment 10. The model may be a simulation model that simulates the operating state or fluid state of the plant equipment 10, a reaction model that simulates chemical reactions in the plant equipment 10, a proxy model that replaces the simulation model or reaction model, AI for verifying scale-up or scale-down in the design of the plant equipment 10, AI for optimizing the operation of the plant equipment 10, AI for detecting or predicting abnormalities during the operation of the plant equipment 10, etc.

The design device 4 designs the plant equipment 10 using the model developed by the development device 3.

The operation control device 5 controls the operation of the plant equipment 10 using the model developed by the development device 3.

The development support device 200 supports the development of models in the development device 3. The development support device 200 acquires equipment identification information for the plant equipment 10 or process identification information for the process executed in the plant equipment 10, and target output information for the model, and outputs a processing element list including the processing elements that should make up the model's development process based on the acquired equipment identification information or process identification information and target output information. This allows developers to develop models using processing element lists of models developed in the past, significantly reducing development costs, time, and effort.

The target output information of a model is information that the model outputs ultimately or intermediately in accordance with the model's purpose. For example, the target output information of a simulation model is the simulation results, and the target output information of an operation optimization AI is the operating parameters for optimizing the operation of the plant equipment 10.

The construction support device 100 accumulates past performance information that the development support device 200 references to output a processing element list, and supports the construction of a development support environment. The construction support device 100 acquires equipment identification information for the plant equipment 10 that is the target of the developed model, or process identification information for the process executed in the plant equipment 10, and target output information for the developed model, acquires the processing elements that make up the development process of the developed model, and stores the processing elements in association with the equipment identification information or process identification information, and/or target output information.

FIG. 3 shows the configuration of a construction support device 100 according to the first embodiment. The construction support device 100 includes a communication device 101, a display device 102, an input device 103, a processing device 120, and a storage device 130.

The communication device 101 controls wireless or wired communication. The display device 102 displays a screen generated by the processing device 120. The display device 102 may be a liquid crystal display device, an organic EL display device, or the like. The input device 103 transmits instructions input by the user of the construction support device 100 to the processing device 120. The input device 103 may be a mouse, keyboard, touchpad, or the like. The display device 102 and input device 103 may be implemented as a touch panel.

The storage device 130 stores data and computer programs used by the processing device 120. The storage device 130 includes a processing element holding unit 131.

The processing element holding unit 131 holds processing elements that should constitute the development process of a model related to the operation of the plant equipment 10, in association with the equipment identification information or process identification information of the plant equipment 10, and/or the target output information of the model. The processing element holding unit 131 may also hold source code and models of modules that should constitute the development process of the model. Furthermore, the processing element holding unit 131 may hold variables, constants, parameters, etc. included in the source code.

Processing device 120 comprises information acquisition unit 121, processing acquisition unit 122, processing division unit 123, and processing element registration unit 124. In terms of hardware components, these components can be realized by any circuit, a computer CPU, memory, programs loaded into memory, etc., but here we have depicted functional blocks realized by the cooperation of these components. Therefore, those skilled in the art will understand that these functional blocks can be realized in various ways using hardware alone, software alone, or a combination of these.

FIG. 4 is a flowchart showing the steps of the support method according to the first embodiment. The steps for supporting the construction of a development support environment using the construction support device 100 will be explained with reference to FIGS. 3 and 4. The model that has already been developed is referred to as the "first model," and the newly developed model is referred to as the "second model."

The information acquisition unit 121 acquires equipment identification information for the plant equipment that is the subject of the developed first model, or process identification information for the process performed in the plant equipment (S10). The equipment identification information may include information such as the type, scale, size, and performance of the plant equipment. The process identification information may include information such as the type, time, and conditions of raw materials, final products, intermediate products, reactions, and processing. The information acquisition unit 121 may acquire equipment identification information or process identification information from the development device 3, design device 4, development support device 200, etc. The information acquisition unit 121 may acquire equipment identification information or process identification information from a developer, etc. via the input device 103.

The information acquisition unit 121 acquires target output information of the developed first model (S12). The target output information is information that the model outputs ultimately or intermediately, and for example, in the plant optimization model described above, includes conditions such as the values of controlled variables for optimizing plant operation. The information acquisition unit 121 may acquire the target output information from the development device 3, the design device 4, the development support device 200, etc. The information acquisition unit 121 may also acquire the target output information from a developer, etc. via the input device 103. The information acquisition unit 121 may acquire the target output information by analyzing the source code of the first model, system design drawings, etc.

The process acquisition unit 122 acquires the processes that make up the developed first model (S14). The processes may include work steps for developing the first model and components (modules) that make up the first model. The process acquisition unit 122 may acquire source code, variables, constants, parameters, etc. of the modules that make up the first model. The process acquisition unit 122 may acquire the processes from the development device 3, the design device 4, the development support device 200, etc. The process acquisition unit 122 may acquire the processes from a developer, etc. via the input device 103.

If the processing acquisition unit 122 has acquired the entire first model, the processing division unit 123 divides the acquired processing into multiple processing elements as necessary (S16). The processing division unit 123 may divide the processing into units of work processes. The processing division unit 123 may divide the processing into processing elements that are common to the development of multiple models and processing elements that differ individually depending on the type of plant, the target output information of the model, etc. Note that if the processing acquisition unit 122 is able to acquire the processing of the first model for each processing element, the processing division unit 123 does not have to divide the processing.

The processing element registration unit 124 stores processing elements in the processing element holding unit 131 in association with equipment identification information or process identification information, and/or target output information.

Figure 5 shows an example of the data structure of the processing element storage unit 131. In this example, processing elements are stored for each target plant equipment, process, and issue (target output information). In addition, processing elements that are commonly included in the model regardless of the target plant equipment, process, or issue are stored.

Figure 6 shows an example of the data structure of the processing element storage unit 131. In the example shown in this figure, a processing element storage unit 131 is provided for each work process. The processing element storage unit 131 for each work process stores processing elements in a matrix of the target plant equipment and process, and the target issue (target output information). Note that the processing element storage unit 131 is not limited to a relational database like those shown in Figures 5 and 6; for example, it may be configured as a graph database in which individual processing elements are linked by edges to the target plant equipment and process, and/or target issue.

FIG. 7 shows the configuration of a development support device 200 according to the first embodiment. The development support device 200 includes a communication device 201, a display device 202, an input device 203, a processing device 220, and a storage device 230.

The communication device 201 controls wireless or wired communication. The display device 202 displays a screen generated by the processing device 220. The display device 202 may be a liquid crystal display device, an organic EL display device, or the like. The input device 203 transmits instructions input by a user of the development support device 200 to the processing device 220. The input device 203 may be a mouse, keyboard, touchpad, or the like. The display device 202 and input device 203 may be implemented as a touch panel.

The storage device 230 stores data and computer programs used by the processing device 220. The storage device 230 includes a processing element holding unit 231.

The processing element holding unit 231 holds the processing elements that constitute the development process of a model related to the operation of the plant equipment 10, in association with the equipment identification information or process identification information of the plant equipment 10, and/or the target output information of the model. The processing element holding unit 231 may be the same as the processing element holding unit 131. The processing element holding unit 231 may be obtained from the construction support device 100 and stored in the storage device 230. If the development support device 200 accesses the processing element holding unit 131 of the construction support device 100, the processing element holding unit 231 does not need to be provided.

The processing device 220 comprises an information acquisition unit 221, a processing element list acquisition unit 222, an output unit 223, a processing element determination unit 224, a processing parameter reception unit 225, a processing order determination unit 226, a connection element reception unit 227, a processing element connection unit 228, and a processing element registration unit 229. In terms of hardware components, these components are realized by any circuit, a computer CPU, memory, a program loaded into memory, etc., but here we have depicted functional blocks realized by the cooperation of these components. Therefore, those skilled in the art will understand that these functional blocks can be realized in various ways using hardware alone, software alone, or a combination of these.

FIG. 8 is a flowchart showing the steps of the support method according to the first embodiment. The steps for supporting model development using the development support device 200 will be explained with reference to FIGS. 7 and 8.

The information acquisition unit 221 acquires equipment identification information for the plant equipment being developed or process identification information for the process being executed in the plant equipment (S20). The information acquisition unit 221 may acquire the equipment identification information or process identification information from the development device 3, design device 4, etc. The information acquisition unit 221 may also acquire the equipment identification information or process identification information from a developer, etc. via the input device 203.

The information acquisition unit 221 acquires target output information for the model (S22). The information acquisition unit 221 may acquire the target output information from the development device 3, the design device 4, etc. The information acquisition unit 221 may also acquire the target output information from a developer, etc. via the input device 203.

The processing element list acquisition unit 222 acquires a processing element list containing processing elements that should constitute the model development process based on the equipment identification information or process identification information and the target output information (S24). The processing element list acquisition unit 222 may acquire the processing element list by referencing the processing element storage unit 231 or the processing element storage unit 131. If the processing element storage unit 231 or the processing element storage unit 131 has the data structure shown in FIG. 5, the processing element list acquisition unit 222 acquires a processing element list by extracting processing elements whose target equipment, target process, and target task match the equipment identification information, process identification information, and target output information, respectively, as well as common processing elements. If the processing element storage unit 231 or the processing element storage unit 131 has the data structure shown in FIG. 6, the processing element list acquisition unit 222 acquires a processing element list by extracting processing elements whose target equipment, target process, and target task match the equipment identification information, process identification information, and target output information, respectively, as well as common processing elements, for each work process.

The output unit 223 outputs the acquired processing element list to the developer as a candidate list of processing elements that should constitute the model development process (S25). The output unit 223 may display the candidate list on the display device 202. The output unit 223 may also transmit the candidate list to a terminal device used by the developer.

The processing element determination unit 224 accepts a selection of processing elements from the candidate list that are to constitute the model development process (S26), and determines the selected processing elements as processing elements that constitute the model development process (S28). The processing element determination unit 224 may accept a selection of processing elements from the developer via the input device 203. The processing element determination unit 224 may also accept a selection of processing elements from a terminal device used by the developer. The processing element determination unit 224 may automatically determine processing elements listed in the processing element list as processing elements that constitute the model development process. The processing element determination unit 224 may automatically select processing elements based on predetermined selection criteria, and determine the selected processing elements as processing elements that constitute the model development process. When processing elements are automatically selected, the selection of processing elements may be performed by a machine learning model. In this case, the machine learning model may be a model that learns the relationship between the equipment identification information and/or process identification information, target output information, and processing elements included in the development process of previously constructed AI, as well as general AI construction flows that are not limited to the plant field, and determines the processing elements of the second model by inputting the equipment identification information, process identification information, or target output information of the second model.

The processing parameter receiving unit 225 receives input of processing parameters for the determined processing element (S30). The processing parameters may include parameters used when executing the processing element. The processing parameter receiving unit 225 sets the received processing parameters for the processing element. The processing parameter receiving unit 225 may receive processing parameters from the developer via the input device 203. The processing parameter receiving unit 225 may receive processing parameters from a terminal device used by the developer. For example, if the processing element relates to outlier removal processing in pre-processing of driving data, the processing parameter receiving unit 225 receives input of parameters that serve as a criterion for removing outliers. The parameters that serve as a criterion for removing outliers may be, for example, based on the interquartile range (hereinafter referred to as IQR) of the acquired data,
・Lower limit is (first quartile) - 1.5 x IQR
・Upper limit is (third quartile) + 1.5 x IQR
The processing parameters may be set without requiring input from the developer. In this case, the processing element may store initial setting parameters, such as the setting parameters in the first model. In this case, the processing element may be configured to allow the developer to check the initial setting parameters and edit them as necessary.

The processing order determination unit 226 determines the processing order in which the processing elements are to be executed (S32). The processing order determination unit 226 may receive the processing order from the developer via the input device 203. The processing order determination unit 226 may also receive the processing order from a terminal device used by the developer. The processing order determination unit 226 may automatically determine the processing order based on a predetermined criterion. The predetermined criterion may be, for example, the processing order in the first model, or a criterion based on the processing order in AI related to the operation of multiple plants constructed in the past, such as the first model, or a criterion based on the order in AI development in other fields, not limited to plant operation. Such processing order criteria may refer to those stored in the processing element storage unit 231, or may be learned based on the processing order of AI constructed in the past.

The combined element receiving unit 227 receives processing elements required to combine the determined processing elements (S34). In cases where the information output from the preceding processing element does not match the information to be input to the subsequent processing element, a processing element that generates input information for the subsequent processing element from the output information of the preceding processing element may be added as a combined element. The combined element receiving unit 227 may receive source code, parameters, etc. of the combined element from the developer via the input device 203. The combined element receiving unit 227 may also receive source code, parameters, etc. of the combined element from a terminal device used by the developer. The combined element receiving unit 227 may automatically generate combined elements based on the output information of the preceding processing element and the input information of the subsequent processing element.

The processing element combining unit 228 combines the determined processing elements (S36). The processing element combining unit 228 combines the determined processing elements and the combined elements in the determined processing order. This generates a second model.

The processing element registration unit 229 stores the equipment identification information or process identification information, the target output information, and the determined processing element in the processing element storage unit 231 (S38). This allows the processing elements that make up the development process of the developed model to be used in the development of subsequent models.

If a developed model is changed or added, the contents of the processing element storage unit 131 and the processing element storage unit 231 may be updated accordingly.

Figure 9 shows an example of a screen displayed on the display device 202 of the development support device 200. In this figure, a user interface screen is displayed that allows the developer to select the type of equipment to be developed and the type of model or problem. When the developer selects the type of equipment and the type of model or problem, the information acquisition unit 221 acquires the equipment identification information of the plant equipment to be developed and the target output information of the model.

Figure 10 shows an example of a screen displayed on the display device 202 of the development support device 200. In this figure, the output unit 223 displays a list of processing elements that should make up the model development process, acquired by the processing element list acquisition unit 222 according to the type of device selected by the developer and the type of model or problem.

The development support system of this embodiment utilizes knowledge and resources gained from previous model development, thereby streamlining model development, which has traditionally been done individually, and significantly reducing the cost, time, and effort required to develop new models. It also enables the division of labor in model development, allowing process engineers, data scientists, AI developers, and others to focus on their respective essential tasks.

Second Embodiment
11 shows the configuration of a development support system 1 according to the second embodiment. The development support system 1 according to the second embodiment includes an administrator terminal 300 in addition to the configuration of the development support system 1 according to the first embodiment shown in FIG. The development support system 1 according to the second embodiment supports the development of a model that is recommended to be implemented in order to operate the plant facility 10 favorably. The following mainly describes the configuration and operation that are different from the development support system 1 according to the first embodiment, and omits descriptions of the configuration and operation that are the same as those of the first embodiment as appropriate.

The administrator terminal 300 is a terminal used by the administrator of the plant equipment 10. The administrator terminal 300 may be a terminal used by the owner, designer, developer, maintainer, engineer, etc. of the plant equipment 10. The administrator terminal 300 may be any terminal device such as a computer, smartphone, or mobile phone.

FIG. 12 is a sequence diagram showing the steps of the support method according to the second embodiment. The administrator terminal 300 transmits equipment identification information for the plant equipment 10 or process identification information for a process executed in the plant equipment 10 to the development support device 200 (S10). The administrator terminal 300 transmits parameter identification information for parameters acquired during operation of the plant equipment 10 to the development support device 200 (S12). Here, the parameter identification information includes information regarding the type of parameter acquired during operation of the plant equipment 10, such as temperature, pressure, or concentration. The parameter identification information may also include information regarding the acquisition location, indicating where in the plant equipment 10 the parameter was acquired, such as the entrance to the equipment, inside the equipment, or exit from the equipment. The administrator terminal 300 transmits time-series data for the parameters acquired during operation of the plant equipment 10 to the development support device 200 (S14).

The development support device 200 calculates a recommendation level, which indicates the degree to which implementation of a model should be recommended for optimal operation of the plant equipment 10, for multiple models related to the operation of the plant equipment 10, based on the equipment identification information or process identification information, parameter identification information, and parameter time-series data acquired from the administrator terminal 300 (S16). The recommendation level is determined in advance based on, for example, past recommendation performance, past implementation performance, evaluations of models implemented in the past, and the quality of operation of the plant equipment 10 in which the model was previously implemented, and is stored in the development support device 200. The development support device 200 refers to the calculated recommendation level and selects a model recommended for implementation from among the multiple models related to the operation of the plant equipment 10 (S18). The development support device 200 outputs information about the selected model to the administrator terminal 300 (S20).

The development support device 200 presents to the administrator terminal 300 information about the dataset required to implement the selected model (S22). The development support device 200 presents to the administrator terminal 300 output information output from the model when the selected model is tested by inputting time-series parameter data into the model (S24).

The administrator references the information presented by the development support device 200, determines the model to be implemented, and requests the development support device 200 to develop the model from the administrator terminal 300 (S26).

As explained in the first embodiment, the development support device 200 acquires a processing element list including the processing elements that should make up the development process of the model for which development has been requested (S28), and outputs the acquired processing element list to the administrator terminal 300 (S30). The subsequent procedures are the same as those in the first embodiment.

FIG. 13 shows the configuration of a development support device according to the second embodiment. The development support device 200 includes a communication device 201, a display device 202, an input device 203, a processing device 220, and a storage device 230. The development support device 200 according to the second embodiment may further include the configuration of the development support device 200 according to the first embodiment shown in FIG. 7.

The storage device 230 stores data and computer programs used by the processing device 220. The storage device 230 includes a model information storage unit 261 and a recommendation level storage unit 262.

The model information storage unit 261 stores information on multiple models related to the operation of the plant equipment 10. The model information storage unit 261 stores information such as the model's target output information, input information, the dataset required to implement the model, a processing element list including the processing elements that should make up the model's development process, the time, cost, and manpower required to develop the model, past performance in recommending the model, and past performance in implementing the model, as well as the model itself.

The recommendation level holding unit 252 holds a recommendation level that indicates the degree to which implementation of a model should be recommended for each of the equipment identification information or process identification information, parameter identification information, and characteristics of the time-series data of the parameters of the plant equipment 10. Figures 14 and 15 show examples of internal data of the recommendation level holding unit 262. Figure 14 shows a table that stores the correspondence between the temperature and pressure when the plant equipment 10 is operating and the recommendation level of a model. Figure 15 shows a table that stores the correspondence between the characteristics of the time-series data of parameters acquired when the plant equipment 10 is operating and the recommendation level of a model. The table shown in Figure 15 may be held for each type of parameter. The recommendation level holding unit 252 may hold a table that stores the correspondence between the type of parameters acquired when the plant equipment 10 is operating and the recommendation level of a model.

The processing device 220 comprises an information acquisition unit 241, a time-series data analysis unit 242, a selection unit 243, an output unit 244, a recommendation performance recording unit 245, an implementation performance recording unit 246, an evaluation acquisition unit 247, a recommendation level setting unit 248, and a trial unit 249. These components are realized by hardware components such as any circuit, a computer CPU, memory, or a program loaded into memory, but the functional blocks realized by the cooperation of these components are depicted here. Therefore, those skilled in the art will understand that these functional blocks can be realized in various ways using only hardware, only software, or a combination of these.

The information acquisition unit 241 acquires, from the administrator terminal 300, equipment identification information for the plant equipment 10 or process identification information for processes executed on the plant equipment 10, parameter identification information for parameters acquired during operation of the plant equipment 10, and time series data for the parameters. The information acquisition unit 241 may acquire information about plant equipment 10 that is already in operation from the administrator terminal 300, or may acquire information about plant equipment 10 that is not yet in operation from the administrator terminal 300. The information acquisition unit 241 may acquire, from the administrator terminal 300, parameter identification information and time series data for parameters acquired when the plant equipment 10 is in operation, or may acquire, from the administrator terminal 300, parameter identification information and time series data for parameters acquired from a simulator that simulates the operation of the plant equipment 10. The information acquisition unit 241 may acquire this information from the development device 3, the design device 4, the operation control device 5, the plant equipment 10, etc. The information acquisition unit 241 may acquire equipment identification information or process identification information from a developer, etc. via the input device 203.

The time series data analysis unit 242 analyzes the time series data of the parameters acquired by the information acquisition unit 241. The time series data analysis unit 242 may calculate statistical values such as the maximum value, minimum value, average value, variance, standard deviation, and median of the time series data. The time series data analysis unit 242 may calculate the rate of change of the time series data, the rate of change of the rate of change, local maximum values, local minimum values, inflection points, amplitude, and frequency. The time series data analysis unit 242 may analyze the characteristics, aspects, and trends of the time series data over a specified period. The characteristics of the time series data may be that the time series data is stable, constantly fluctuating, gradually increasing, gradually decreasing, or that outliers are present. For example, the time series data analysis unit 242 may determine that the time series data is stable if the difference between the maximum and minimum values of the time series data over a specified period is equal to or less than a specified value, if the variance is equal to or less than a specified value, or if the average absolute value of the rate of change is equal to or less than a specified value. The time series data analysis unit 242 may determine that the time series data is constantly fluctuating if the variance of the time series data over a predetermined period is equal to or greater than a predetermined value, or if the average absolute value of the rate of change is equal to or greater than a predetermined value.

The selection unit 243 selects a model recommended for implementation from among multiple models related to the operation of the plant equipment 10, based on the equipment identification information or process identification information, parameter identification information, and parameter time series data acquired by the information acquisition unit 241. The selection unit 243 calculates the recommendation level for each of the multiple models by referring to the recommendation level storage unit. For example, the selection unit 243 may acquire the recommendation level corresponding to the equipment identification information or process identification information from the table shown in FIG. 14, acquire the recommendation level corresponding to the aspect of the parameter time series data from the table shown in FIG. 15, and calculate the product of these recommendation levels to determine the recommendation level for the model. If a table storing the correspondence between parameter identification information and recommendation levels is stored in the recommendation level storage unit 252, the selection unit 243 may further acquire the recommendation level corresponding to the parameter identification information, and calculate the product of the recommendation level corresponding to the equipment identification information or process identification information, the recommendation level corresponding to the parameter identification information, and the recommendation level corresponding to the aspect of the parameter time series data to determine the recommendation level for the model. The selection unit 243 selects a model with a large calculated recommendation level as a model recommended for implementation. The selection unit 243 may select a model based on further conditions related to equipment identification information or process identification information, parameter identification information, time-series data of parameters, the location, climate, years of operation of the plant equipment 10, or any combination thereof. The selection unit 243 may also select a model recommended for implementation based on a recommendation level output by a nonlinear model or machine learning model that outputs a recommendation level for each model corresponding to input including equipment identification information or process identification information, parameter identification information, and the aspect of the time-series data of parameters. In this case, a machine learning model may be used in which the equipment identification information or process identification information, the parameter identification information, and information related to the aspect of the time-series data of parameters are used as explanatory variables and the target output information is used as a target variable. Here, the information related to the aspect of the time-series data of parameters may be a feature obtained by dimensional compression of raw data of the time-series data of parameters as an explanatory variable. The machine learning model may be trained using the recommendation performance recorded in the recommendation performance recording unit 245 (described later), the model implementation performance, evaluations of previously implemented models acquired by the evaluation acquisition unit 247 (described later), or the quality of operation of the plant equipment 10 in which the model was previously implemented as training data.

The output unit 244 reads information about the model selected by the selection unit 243 from the model information storage unit 261 and outputs the information to the administrator terminal 300. The output unit 244 reads information about the dataset required to implement the model selected by the selection unit 243 from the model information storage unit 261 and outputs the information to the administrator terminal 300.

The trial unit 249 inputs the time-series data of the parameters acquired by the information acquisition unit 241 into the model selected by the selection unit 243, and trials the model. The output unit 244 outputs output information of the model trialed by the trial unit 249 to the administrator terminal 300.

The recommended performance recorder 245 records information about the recommended performance of models recommended by the development support device 200 to the administrator terminal 300 in the model information storage unit 261. The implementation performance recorder 246 records information about the implementation performance of models adopted by the administrator and implemented in the operation control device 5 in the model information storage unit 261.

The evaluation acquisition unit 247 acquires an evaluation of a model that was previously implemented in the operation control device 5. The evaluation acquisition unit 247 may acquire an evaluation of the model by the administrator from the administrator terminal 300. The evaluation acquisition unit 247 may evaluate the model based on the quality of operation of the plant equipment 10 in which the model was previously implemented. For example, the evaluation acquisition unit 247 may evaluate the quality of operation of the plant equipment 10 in which the model was implemented based on time-series data of parameters acquired during the operation of the plant equipment 10 in which the model was previously implemented, and determine an evaluation of the model.

The recommendation level setting unit 248 sets the recommendation level of the model and stores it in the recommendation level holding unit 252. The recommendation level setting unit 248 may set the recommendation level of the model based on the recommendation history or implementation history of the model held in the model information holding unit 261. For example, the more times a model has been recommended or implemented, the higher the recommendation level may be. Furthermore, the higher the probability that the recommended model has been implemented, the higher the recommendation level may be. The recommendation level setting unit 248 may set the recommendation level of the model based on the evaluation of the model acquired by the evaluation acquisition unit 247. For example, the higher the evaluation of the model, the higher the recommendation level of the model may be.

Immediately after the development support device 200 begins operation, the model recommendation level may be set manually, as the model's recommendation track record, implementation track record, model evaluations, etc. have not yet been accumulated. As the development support device 200 continues to operate, the recommendation level setting unit 248 updates the model recommendation level in accordance with the model's recommendation track record, implementation track record, model evaluations, etc. The model recommendation level is updated in association with the equipment identification information of the plant equipment for which the model has been recommended, implemented, or evaluated, or the process identification information of the process executed in the plant equipment, the parameter identification information acquired when the model was recommended, and the characteristics of the parameter time-series data. This makes it possible to set a model recommendation level that is more in line with reality, thereby improving the accuracy of model recommendations.

Figure 16 shows an example of time series data for parameters. The information acquisition unit 241 may acquire time series data for multiple parameters. The time series data analysis unit 242 analyzes the acquired time series data for the parameters and determines the state and trend of the time series data.

Figure 17 shows an example of a screen presented from the development support device 200 to the administrator terminal 300. The output unit 244 outputs to the administrator terminal 300 a screen displaying the equipment identification information acquired by the information acquisition unit 241, the time-series data of the parameters, information about the model selected by the selection unit 243, and the effects of implementing the model. When the button for trying out the model is pressed, the trial unit 249 inputs the time-series data of the parameters acquired by the information acquisition unit 241 into the model and tries out the model.

Figure 18 shows an example of a screen presented by the development support device 200 to the administrator terminal 300. The output unit 244 outputs to the administrator terminal 300 a screen displaying a configuration diagram of the model selected by the selection unit 243 and the trial results from the trial unit 249. When a button to request implementation of the model is pressed, the output unit 244 reads from the model information storage unit 261 a processing element list including the processing elements that should make up the development process of the model requested to be implemented, and outputs this to the administrator terminal 300.

The technology of this embodiment selects and recommends a model based on information about the operation of the plant equipment 10, making it possible to accurately recommend a model for optimal operation of the plant equipment 10. This improves the operating efficiency of the plant equipment 10. It also significantly reduces the cost and effort required to develop a model.

The present invention has been described above based on examples. These examples are merely illustrative, and those skilled in the art will understand that various modifications are possible in the combination of each component and each treatment process, and that such modifications are also within the scope of the present invention.

In the first embodiment, an example was described in which processing elements constituting a previously developed model are stored in advance in the processing element storage unit 131. In another example, information about previously developed models may be collected when the processing element list acquisition unit 222 acquires a list of processing elements. In this case, the processing element storage unit 131 may also be temporarily generated.

The present invention can be used in a method, device, and program for supporting the development of models related to the operation of plant equipment.

1 Development support system, 2 Communication network, 3 Development equipment, 4 Design equipment, 5 Operation control device, 10 Plant equipment, 100 Construction support device, 101 Communication device, 102 Display device, 103 Input device, 120 Processing device, 121 Information acquisition unit, 122 Processing acquisition unit, 123 Processing division unit, 124 Processing element registration unit, 130 Storage device, 131 Processing element holding unit, 200 Development support device, 201 Communication device, 202 Display device, 203 Input device, 220 Processing device, 221 Information acquisition unit, 222 Processing element list acquisition unit, 223 Output unit, 224 processing element determination unit, 225 processing parameter reception unit, 226 processing order determination unit, 227 combined element reception unit, 228 processing element combination unit, 229 processing element registration unit, 230 storage device, 231 processing element storage unit, 241 information acquisition unit, 242 time series data analysis unit, 243 selection unit, 244 output unit, 245 recommended performance recording unit, 246 implementation performance recording unit, 247 evaluation acquisition unit, 248 recommendation level setting unit, 249 trial unit, 252 recommendation level storage unit, 261 model information storage unit, 262 recommendation level storage unit, 300 administrator terminal.

Claims (12)

1. A method for assisting in the implementation of a model for the operation of a plant facility, comprising:
The method is executed by a processor,
acquiring equipment identification information of the plant equipment or process identification information of a process executed in the plant equipment;
acquiring parameter identification information of parameters acquired during operation of the plant equipment;
acquiring time series data of the parameter;
selecting a model recommended for implementation from among a plurality of models related to the operation of the plant equipment based on the equipment identification information or the process identification information, the parameter identification information, and time-series data of the parameters;
Outputting information about the selected model;
A method comprising: The method according to claim 1, wherein when selecting a model recommended for implementation, the method selects the model recommended for implementation by referring to a recommendation level that indicates a degree to which implementation of the model should be recommended, the recommendation level being stored in advance for each of the equipment identification information or the process identification information, the parameter identification information, or a feature of the time-series data of the parameter. The method according to claim 2 , wherein the recommendation level is determined based on past recommendation performance. The method according to claim 2 , wherein the recommendation level is determined based on past implementation results. The method according to claim 2 , wherein the recommendation level is determined based on evaluations of previously implemented models. The method according to claim 5 , wherein the evaluation is determined based on the quality of operation of plant equipment in which the model has been implemented in the past. The method according to claim 5 , wherein the evaluation is determined based on time-series data of parameters acquired during the operation of a plant facility in which the model was implemented in the past. The method of claim 1 , further comprising presenting information about a dataset required to implement the selected model. The method of claim 1 , further comprising presenting output information from the selected model when time series data of the parameters is input to the selected model. The method further includes outputting a processing element list including processing elements that are to constitute the development process of the selected model.
The method of claim 1. an information acquisition unit that acquires equipment identification information of a plant facility or process identification information of a process executed in the plant facility, parameter identification information of a parameter acquired during operation of the plant facility, and time-series data of the parameter;
a selection unit that selects a model recommended for implementation from among a plurality of models related to the operation of the plant equipment based on the equipment identification information or the process identification information, the parameter identification information, and time-series data of the parameters;
an output unit that outputs information about the selected model;
A support device comprising: Computer,
an information acquisition unit that acquires equipment identification information of a plant facility or process identification information of a process executed in the plant facility, parameter identification information of a parameter acquired during operation of the plant facility, and time-series data of the parameter;
a selection unit that selects a model recommended for implementation from among a plurality of models related to the operation of the plant equipment based on the equipment identification information or the process identification information, the parameter identification information, and time-series data of the parameters;
an output unit that outputs information about the selected model;
A support program to help it function as a