TWI884432B - Support device, support method and support program - Google Patents

Abstract

[Topic] Provide a support device, support method and support program for determining the operation status of a factory with high accuracy. [Solution] A support device (10) for supporting the operation of a factory, comprising: a data acquisition unit, which acquires simulation data related to the operation status of the factory calculated by inputting data containing a predetermined condition list into a physical model simulator, and in the predetermined condition list, factors that determine the operation status of the factory and indicators for the factors are associated; a statistical model generation unit (14c), which generates a first statistical model based on the simulation data; and an operation status determination unit (14f), which refers to the first statistical model and determines the operation status of the factory based on the process data of the factory.

Description

Support device, support method and support program

The present invention relates to a support device, a display device, a support method and a support program for supporting the operation of a factory.

In the central monitoring room of the factory, the alarms issued by the monitoring of process data and DCS (distributed control system) are always monitored. The operation status of the factory is judged based on the statistical model, and the results are notified to the operation personnel if they are judged to be abnormal. The statistical model used in the judgment can be generated by learning the process data obtained by the actual operation of the factory.

[Prior Art Literature] [Patent Literature]

[Patent Document 1] Japanese Patent Publication No. 2011-253275

However, when a factory is just starting to operate and there is not enough process data accumulated, learning cannot be performed sufficiently, and the accuracy of the judgment of the operating status based on the statistical model will decrease.

One of the exemplary purposes of one aspect of the present invention is to provide a support device, support method and support program for determining the operating status of a factory with high accuracy.

In order to solve the above-mentioned problem, a supporting device of one aspect of the present invention is a supporting device for supporting the operation of a factory, comprising: a data acquisition unit, which acquires simulation data related to the operation state of the factory calculated by inputting data including a predetermined condition list into a physical model simulator, and in the predetermined condition list, factors determining the operation state of the factory and indicators for the factors are associated; a statistical model generation unit, which generates a first statistical model based on the simulation data; and an operation state determination unit, which refers to the first statistical model and determines the operation state of the factory based on the process data of the factory.

According to the above, since the data used in learning the statistical model is obtained through physical model simulation, it is possible to generate a statistical model that can determine the operating status with high accuracy even if there is not enough process data actually obtained from the factory.

Another aspect of the present invention is a statistical model generation device. The device is a device for generating a statistical model for determining the operating state of a factory, and comprises: a data acquisition unit, which acquires simulation data related to the operating state of the aforementioned factory by inputting data containing a predetermined condition list into a physical model simulator, and the predetermined condition list determines the factors of the operating state of the aforementioned factory and establishes a correlation with the indicators for the aforementioned factors; and a statistical model generation unit, which generates a first statistical model based on the aforementioned simulation data.

Another aspect of the present invention is a support method. The method is a support method for supporting the operation of a factory, comprising the following steps: a data acquisition step, which is to obtain simulation data related to the operation state of the aforementioned factory calculated by inputting data containing a predetermined condition list into a physical model simulator, and in the aforementioned predetermined condition list, factors that determine the operation state of the aforementioned factory and indicators for the aforementioned factors are associated; a statistical model generation step, which is to generate a first statistical model based on the aforementioned simulation data; and an operation state determination step, which is to refer to the aforementioned first statistical model and determine the operation state of the aforementioned factory based on the process data of the aforementioned factory.

Another aspect of the present invention is a support program. The program is a support program for supporting the operation of a factory, and the support program enables the computer to function as the following means: data acquisition means, which acquires simulation data related to the operation state of the aforementioned factory by inputting data containing a predetermined condition list into a physical model simulator, and the factors that determine the operation state of the aforementioned factory and the indicators for the aforementioned factors are associated in the aforementioned predetermined condition list; statistical model generation means, which generates a first statistical model based on the aforementioned simulation data; and operation state determination means, which refers to the aforementioned first statistical model and determines the operation state of the aforementioned factory based on the process data of the aforementioned factory.

In addition, any combination of the above components or the components of the present invention or the mutual replacement between the components or the methods, devices, systems, computer programs, data structures, recording media, etc., are also valid as aspects of the present invention.

According to the present invention, the operating status of the factory can be determined with high precision.

1: Factory

10: Support devices

12:Input part

14: Processing Department

14a: Control Department

14b: Process data acquisition department

14c: Statistical model generation unit

14e: Physical model simulator department

14f: Operation status determination unit

14g: Display control unit

16: Display unit

18: Memory Department

106: External memory device

108: User Interface

110: Display

112: Communication interface

600: Operation diagnosis screen

660: Alarm list display area

700: First statistical model comparison screen

800: Second statistical model comparison screen

[Figure 1] is a diagram showing the structure of a support device 10 according to an embodiment of the present invention.

[Figure 2] is a diagram showing an example of the hardware configuration of the support device 10.

[Figure 3] is a diagram showing an example of a condition list.

[Figure 4] is a diagram showing an example of a condition list.

[Figure 5] is a flowchart showing an example of a support method based on the support device 10.

[Figure 6] is a flowchart showing an example of a support method based on the support device 10.

[Figure 7] is a diagram showing an example of the generation process of a statistical model.

[Figure 8] is a diagram showing an example of the generation process of a statistical model.

[Figure 9] is a diagram for explaining the supporting method of the supporting device 10.

[Figure 10] is a diagram for explaining the supporting method of the supporting device 10.

[Figure 11] is a diagram for explaining the support method of the support device 10.

The present invention is described below with reference to the drawings and through the embodiments of the invention, but the following embodiments do not limit the scope of the invention application, and all combinations of features described in the embodiments are not necessarily required by the solution of the invention. The same or equivalent components, components, and processes shown in each drawing are marked with the same component symbols, and repeated descriptions are appropriately omitted.

Figures 1 to 4 are diagrams of a support device 10 for explaining an implementation of the present invention. Specifically, Figure 1 is a diagram showing the configuration of a support device 10 of an implementation of the present invention, and Figure 2 is a diagram showing an example of a hardware configuration of the support device 10. Figures 3 and 4 are diagrams showing an example of a condition list stored in a condition list storage unit 18c.

The support device 10 is a device that supports the operation of the factory 1. As an example of the factory 1, a power plant, an incineration plant, or a chemical plant can be cited. Any process data is used in the factory 1. The process data is, for example, data of process values such as temperature, pressure, air volume, concentration, or composition.

The support device 10 is connected to the factory 1 via the DCS (distributed control system) 2 to obtain the process data of the sensors installed in the factory 1. According to the support device 10, even if the process data is not sufficiently accumulated, it is possible to make a high-precision judgment on the operating status based on the statistical model.

The support device 10 is pre-installed with a predetermined program required for executing the support method of the present embodiment, and an example of its hardware configuration is shown in FIG2 . Specifically, the support device 10 can use a general-purpose or dedicated computer equipped with a CPU 100, a ROM 102, a RAM 104, an external memory device 106, a user interface 108, a display 112, and a communication interface 110. Based on the information input by the operator of the factory through the user interface 108, the CPU 100 performs calculations and outputs the calculation results to the display 112. The operator can recognize the output while inputting the required information to the support device 10 through the user interface 108.

The support device 10 can be composed of a single computer or a plurality of computers distributed on a network. In the support device 10, for example, by the CPU executing a predetermined program (a program that specifies the support method of this embodiment) stored in the above-mentioned ROM, RAM, external storage device, etc. or downloaded via a communication network, the support device 10 can function as various functional blocks or various steps to be described later.

Below, various functional blocks of the support device 10 are described using Figure 1.

The support device 10 of this embodiment includes: an input unit 12 having an input receiving unit 12a and an operation receiving unit 12b, a processing unit 14, a display unit 16 having a display 16a, and a memory unit 18. Here, the processing unit 14 includes a control unit 14a, a process data acquisition unit 14b, a statistical model generation unit 14c, a random number generation unit 14d, a physical model simulator unit 14e, an operation state determination unit 14f, and a display control unit 14g. In addition, the memory unit 18 includes a process data storage unit 18a, a simulation data storage unit 18b, a condition list storage unit 18c, a distinction information storage unit 18d, a statistical model storage unit 18e, a physical model storage unit 18f, and a determination result storage unit 18g.

The control unit 14a is composed of a microcomputer including a CPU and a semiconductor memory, and performs general calculations that are not performed by other functional blocks included in the processing unit 14. For example, it performs time synchronization with an external device through the communication interface 110 or DNS (Domain Name System) name resolution, etc.

The process data acquisition unit 14b acquires the process data of the factory 1 from the DCS 2. The process data is data indicating the operation status of the factory, and can also be called operation data. The process data is, for example, data indicating the gradual change of the measured value of the sensor. In this case, the process data can also be the change of the measured value of the continuous sensor in a predetermined time interval. The process data can also be multidimensional data. The process data acquired by the process data acquisition unit 14b is stored in the process data storage unit 18a.

The statistical model generation unit 14c generates a statistical model by learning predetermined data. The predetermined data refers to, for example, process data or simulation data output by a physical simulation to be described later. The process data used in learning can also be process data for learning that is distinguished from the process data for evaluation used in the operation state determination. The generated statistical model is stored in the statistical model storage unit 18e.

Furthermore, generating a statistical model includes generating a machine learning model, which outputs a predetermined judgment when a certain data is input. The machine learning model can also be generated by any method. The algorithm of the machine learning model can be, for example, support vector machine, logistic regression, neural network, deep neural network, k-means method, etc., and its type is not particularly limited.

The random number generating unit 14d generates random numbers according to a predetermined probability distribution. The predetermined probability distribution is, for example, a regular distribution with a mean M and a standard deviation σ. Furthermore, M and σ are arbitrary real numbers. In addition, the probability distribution can also be a uniform distribution, an exponential distribution, a gamma distribution, etc.

The physical model simulator unit 14e performs physical model simulation according to the physical model stored in the physical model storage unit 18f and the condition list stored in the condition list storage unit 18c to obtain simulation data. The simulation data is stored in the simulation data storage unit 18b.

Simulation data refers to simulation process data calculated by physical model simulation. Simulation data can also have the same data structure or dimension as process data obtained from the factory. Specifically, simulation data can be processed in the same way as process data in the statistical model generation unit 14c without being distinguished from process data. More specifically, the statistical model generation unit 14c can generate a corresponding statistical model regardless of whether it learns simulation data or process data.

Furthermore, the physical model simulation based on the physical model simulator unit 14e can be implemented regardless of the operating status of the factory. For example, it can be executed even before the factory starts operating, during operation, or during a temporary stop of operation.

The condition list stored in the condition list storage unit 18c includes information related to the execution conditions of the physical model simulation. An example of the condition list will be described later using Figures 3 and 4.

The operation state determination unit 14f refers to the statistical model to determine whether the process data to be determined is within the range of the threshold value. The threshold value, for example, refers to the value determined at the same time as the statistical model is generated, which is information stored in the statistical model storage unit 18e. The determination result is stored in the determination result storage unit 18g.

The display control unit 14g displays information on the display 16a, the information including the determination result stored in the determination result storage unit 18g and the distinction information stored in the distinction information storage unit 18d. An example of the display content will be described later using FIG. 10.

The input accepting unit 12a and the operation accepting unit 12b accept input and operation from the user through the user interface 108. The display control unit 14g can also change the display content of the display 16a according to the accepted input and operation. An example of changing the display content will be described later using Figure 10. In addition, the condition list can also be created or changed according to the accepted input or operation.

Next, an example of a condition list is explained using Figures 3 and 4.

The condition list refers to a set of execution conditions for physical model simulation. Specifically, the condition list includes one or more factors, and includes information that one or more indicators are associated with each factor. Here, factors, for example, refer to elements that affect the operating status of the factory. As an example, the condition list shown in Figure 3 includes factor 202 (fuel), factor 212 (external temperature) and factor 222 (boiler load), and seven indicators 204 (particle 100%, particle 50% + PKS50%, etc.) are associated with factor 202, two indicators 214 (summer average temperature, annual average temperature) are associated with factor 212, and five indicators 216 (100%, 90%, etc.) are associated with factor 222. The condition list is input to the physical model simulator to specify the execution conditions of the simulation.

In addition, as shown in FIG4 , the condition list may also be information including multiple records. Here, the record includes an indicator associated with one or more factors. For example, record 322 is associated with factor 310b (fuel), factor 310c (external temperature), and factor 310d (boiler load) respectively, and indicators 322b (particles 100%), 322c (average summer temperature), and 322d (100%) are established. The condition list is information consisting of 8 records 320 including record 322.

The physical model simulation is performed according to the execution conditions of the physical model simulation included in the condition list. An example of a specific execution method will be described later using Figure 6.

Part of the information included in the condition list may also have no substantial impact on the results of the physical model simulation. For example, in the condition list of FIG4 , the column numbered 310a only represents the serial number for each record and has no impact on the results of the physical model simulation.

The condition list may also be stored in the condition list storage unit 18c in the following manners: receiving sensor data input from the user via the user interface 108; receiving from other devices via the communication interface 110; or reading via the RAM 104 or the external storage device 106.

Furthermore, at least a portion of the indicators in the condition list may be set based on random numbers generated in the random number generation unit 14d according to a predetermined probability distribution. For example, each indicator associated with the factor 310d (boiler load) may be set based on a value not exceeding 100% of the random numbers in a normal distribution with an average of 70% and a standard deviation of 5%.

The condition list is not limited to being expressed in the list format of FIG. 3 or the table format of FIG. 4. For example, it can also be expressed in multiple tables that can be equivalent to the table of FIG. 4 by appropriately combining tables. In addition, it can also be expressed in a data structure other than a table, such as a tree structure, JSON (JavaScript Object Notation: programming language) format, YAML format (YAML Ain’t Markup Language: YAML is not a markup language) and XML (Extensible Markup Language: Extensible Markup Language) format.

As an embodiment of the support device of the present invention, an example of the operation of the support device is described below using FIG. 5 to FIG. 8. The flowchart of FIG. 5 is an example of the entire operation of the support device of the present embodiment. The flowchart of FIG. 6 is an example of a method for generating a statistical model based on the statistical model generating unit 14c using simulation data. FIG. 7 and FIG. 8 are examples of a process for generating a statistical model by learning process data and simulation data, respectively.

FIG5 is a flowchart showing an example of the operation of the support device 10.

In FIG5 , first, the process data of the factory 1 is acquired by the process data acquisition unit 14b and stored in the process data storage unit 18a (S10). The acquisition of process data can be directly acquired from the DCS2, or can be performed by accepting the input of sensor data from the user through the user interface 108, or can be received from other devices through the communication interface 110, or can be performed by reading from the RAM 104 or the external storage device 106. In addition, the process data can be acquired successively (for example, every second), or the process data of a certain period of time (for example, one day) can be acquired at once.

Next, a first statistical model is generated as a statistical model generated based on simulation data, and it is determined whether it is in a state that can be used (S12). If it has not been generated (S12 No), the first statistical model is generated in the statistical model generation unit 14c, and the information related to it is stored in the statistical model storage unit 18e (S14). An example of a method for generating the first statistical model will be described later using the flowchart of Figure 6.

Furthermore, since the first statistical model is generated based on simulation data, it can be generated and used to determine the operating status regardless of the operating status of the plant. For example, the first statistical model can be generated in the support device even before the plant starts operating, during operation, or during a temporary suspension of operation.

When the first statistical model has been generated (S12 is) or the first statistical model is generated in step S14, the operating state of the factory is determined in the operating state determination unit 14f (S16). The operating state is determined based on the process data stored in the process data storage unit 18a and by referring to the first statistical model stored in the statistical model storage unit 18e. The determination result is stored in the determination result storage unit 18g.

The situation where the first statistical model has been generated refers to the situation where the statistical model generation unit 14c of the support device 10 has already generated the first statistical model, for example, the situation where information related to the first statistical model is received through the communication interface 110 before the factory starts operating, or the situation where the information is read through the RAM 104 or the external storage device 106.

Next, it is determined whether the second statistical model, which is a statistical model generated based on the process data, is in a usable state (S18). Information related to the second statistical model and information related to the first statistical model are separated and stored in the statistical model storage unit 18e.

Since the second statistical model improves accuracy by fully learning process data, it cannot accurately determine the operating status when process data under the operating status has not been sufficiently accumulated, such as when the factory has just started operating or when the factory's operating conditions have just been significantly changed.

In the above-mentioned situation, when the use of the second statistical model is not practical (S18 No), in the statistical model generation unit 14c, the second statistical model is updated by learning the process data stored in the process data storage unit 18a (S20). By processing step S20, the information related to the second statistical model stored in the statistical model storage unit 18e is updated.

Furthermore, even when the second statistical model is updated (S20), sufficient process data may not be learned, so the preparation for using the second statistical model to determine the operation status of the factory may not be completed. Therefore, in this embodiment, when the second statistical model is updated (S20), the operation status determination (S22) of the second statistical model is not referred to, but the operation status determination result is displayed (S24).

On the other hand, when the second statistical model has learned enough process data to be in a state where it can be used (S18 is), the operation state of the factory is determined in the operation state determination unit 14f (S22). The operation state is determined based on the process data stored in the process data storage unit 18a and by referring to the second statistical model. The determination result is distinguished from the determination result stored in step S16 and stored in the determination result storage unit 18g.

The types of judgment results are, for example, "normal" and "abnormal", and "abnormal" can also have stages such as "caution", "warning" and "monitoring". The so-called abnormality is not limited to the state that requires the factory to stop operating, but also includes the state that can continue to operate but has deviated from the good operating state.

Finally, according to the judgment result stored in step S16 or step S22, the display control unit 14g displays information related to the operating state on the display 16a (S24). The displayed information can be changed according to the input from the user input receiving unit 12a or the input to the operation receiving unit 12b. An example of the display screen is further described using FIG. 10. After step S24, the support device 10 ends the action.

In one example of the present embodiment, the operation status of the factory can be determined with high accuracy even when process data is not sufficiently accumulated. Specifically, the second statistical model is not practical for determining the operation status before the process data is fully learned, but if the first statistical model generated based on the simulation data is used, the operation status can be determined with high accuracy. More specifically, even when the preparation for the use of the second statistical model is not complete and further learning of process data is required (S20), the operation status can be determined (S16) in the first statistical model, and the determination result can be prompted to the operation personnel (S24).

According to this effect, for example, in the past, in order to accumulate process data for learning statistical models over a long period of time, it was necessary to conduct a trial operation of the factory. However, according to the support device shown in one example of this embodiment, it is possible to refer to the statistical model generated by the simulation data and determine the operating status immediately after the start of operation, so this trial operation is not necessary.

Furthermore, as another effect, the first statistical model and the second statistical model can be compared to determine the effectiveness of learning the process data. Specifically, in the statistical model generated from the process data, if there is a situation where the process data that originally occurred due to a temporary failure of a factory in the factory is learned as normal process data, the operation personnel can find the data by comparing it with the statistical model generated from the simulation data and delete it from the learning object.

The above-mentioned actions of the implementation method of the present invention are only examples and are not limited thereto. For example, the action can also be repeatedly executed during the operation of the support device 10. That is, it can be restarted after the action is completed. In addition, the order of each step can be changed within the range where the action does not conflict, and a part of it can be repeatedly executed (for example, after step S24, it does not end but returns to step S10, etc.).

In addition, some steps can also be performed in parallel. For example, after obtaining process data in step S10, the processing related to the first statistical model (corresponding to S12~S16) and the processing related to the second statistical model (corresponding to S18~S22) can be performed in parallel. In this case, the user can also be prompted to choose whether to perform processing related to either the first statistical model or the second statistical model.

In addition, the determination of the operating status can be temporarily performed without using either the first statistical model or the second statistical model. For example, when the second statistical model has learned enough process data and is ready for use, the processing related to the first statistical model can be temporarily stopped (S12~S16).

In addition, the process data used in the determination of the operating state (S16, S22) and the learning of the second statistical model (S20) are not limited to the process data obtained in the processing corresponding to the immediately preceding step S10, but may be the process data memorized in the process data storage unit 18a before the processing. For example, when the process data is obtained every second (S10), the process data used in the updating of the second statistical model (S20) may be the process data obtained one day before the start of the updating processing.

In addition, the process data can be divided into evaluation process data and learning process data according to arbitrary rules for processing. For example, in the process data storage unit 18a, the process data acquired within the last week can be divided into evaluation process data, and the process data before that can be divided into learning process data, and only the learning process data can be used to learn (S20) the second statistical model, and only the evaluation process data can be used to determine (S16, S22) the operating status.

In addition, a statistical model can be created by learning both process data and simulation data. For example, the first statistical model can be generated by running a physical model simulation before the plant starts operating, and the first statistical model can be made to learn process data after the plant starts operating.

Next, an example of a method for generating the first statistical model accompanying the physical model simulation (S14) is described using the flowchart of FIG6 and an example of a condition list of FIG4.

First, in the control unit 14a, the initial value of the variable N is set to "1" (S14a). The variable N is a variable that increases by repeated processing.

Next, a record having the same value as the condition number 310a and the variable N is read from the condition list stored in the condition list storage unit 18c (S14b). For example, when the variable N is 1, the record 322 having the condition number "1" is read.

Next, the read record 322 is input to the physical model simulator unit 14e, and the physical model simulation is performed according to the factors and indicators of the record (S14c). In this case, the physical model simulation is performed according to the factors and indicators of "particle 100% (indicator 322b)" as the fuel (factor 310b), "summer average temperature (indicator 322c)" as the outside temperature (factor 310c), and "100% (indicator 322d)" as the boiler load (factor 310d). The simulation process data calculated as the execution result, that is, the simulation data, is stored in the simulation data storage unit 18b.

Furthermore, the so-called physical model simulation, for example, inputs a predetermined action parameter and reproduces the operation state of the factory in the action parameter in the computer. The physical model simulator that performs physical model simulation includes the program used when designing the factory.

Afterwards, the simulation data stored in the simulation data storage unit 18b is input to the statistical model generation unit 14c, and the first statistical model is made to learn the data. Along with this, the information related to the first statistical model stored in the statistical model storage unit 18e is updated (S14d).

The above steps S14b to S14d are executed respectively according to multiple records. Specifically, it is determined whether the value stored in the variable N is the last number of the condition number 310a (S14e). If not (S14e No), the variable N is incremented by 1 (S14f) and the process returns to step S14b. For example, when the variable N is "1", since it is not the last number "8" of the condition number 310a, the variable N is incremented to "2" and the process returns to step 14b.

In step 14b, the record with the same value of condition number 310a and variable N in the condition list stored in the condition list storage unit 18c is read, so the record 324 with condition number "2" is read.

By repeating the above process in the same way and executing the process of step S14b to step S14d according to all the records, the generation of the first statistical model is completed.

The above-mentioned method for generating the first statistical model is only an example and is not limited to this. For example, physical model simulations based on multiple records can also be performed in parallel.

Next, an example of the learning process of the first statistical model and the second statistical model is described using Figures 7 and 8.

Figure 7 is an example of the process of setting the ideal line and threshold of the second statistical model based on process data.

Graph 416 of FIG. 7 is a scatter plot of multiple process data with process variable A on the horizontal axis and process variable B on the vertical axis. Process variable A and process variable B are variables included in the process data, such as heat absorption or boiler load. Multiple dots represented by white hollow circles are drawn corresponding to one process data. As shown in graph 416, if there is not enough process data accumulated, it is impossible to determine a practical ideal line and threshold value, and thus it cannot be used to determine the operating status.

In contrast, graph 418 is a scatter plot when process data has been accumulated to a certain extent over a certain period of time. By learning the accumulated data, the ideal line 418a in the second statistical model can be set. The ideal line is set, for example, by a mathematical model generation method such as nonlinear least squares method.

Graph 420 is a scatter plot when process data is sufficiently accumulated after a certain period of time. In addition to the ideal line, thresholds 420a and 420b can be set. When the process data exceeds the value, it can be determined that there is an abnormality in the operation state. Furthermore, the threshold can be set, for example, so that 95% of the process data used for learning are judged as "no abnormality". In addition, for example, it is also possible to assume that the process data is distributed according to a predetermined distribution and set the threshold according to the distribution.

The second statistical model sets the ideal line and threshold by learning process data, so the factory operation status cannot be determined immediately after the operation starts (for example, the state of graph 416), and the factory operation status can only be determined with high accuracy after a certain period of time has passed since the start of the operation (for example, the state of graph 420).

Figure 8 is an example of the process of setting the ideal line and threshold of the first statistical model based on simulation data.

Graph 516 of FIG8 is a scatter plot of multiple simulation data with process variable A on the horizontal axis and process variable B on the vertical axis. Multiple dots represented by squares are drawn corresponding to one simulation data. Here, since one simulation data is calculated based on one record in the condition list, one dot can also correspond to one record in the condition list. For example, dot 516a can correspond to record 322, and dot 516b can correspond to record 324.

In addition, either process variable A or process variable B may be a variable set as an input to the physical model simulator. For example, when the input to the physical model simulator includes a value related to "boiler load" and the simulation data output includes a value related to "heat absorption", process variable A may be set to boiler load and process variable B may be set to heat absorption.

Graph 518 is a scatter plot of the calculated simulation data after more physical model simulations are performed compared to Graph 516. By learning the simulation data, the ideal line 518a in the first statistical model can be set.

Chart 520 is a scatter plot of the calculated simulation data after more physical model simulations are performed compared to Chart 518. By learning the simulation data calculated based on sufficient combinations of execution conditions, not only can the ideal line be set, but also the threshold 520a and the threshold 520b can be set. When the process data exceeds the value, it can be determined that there is an abnormality in the operation state. Furthermore, the threshold can be set to 95% of the learning data to determine "no abnormality". In addition, for example, it can also be assumed that the process data is distributed according to a predetermined distribution and the threshold can be set according to the distribution.

When calculating simulation data, part of the execution conditions can also be shared in part of the physical model simulation. For example, the dot group 520c composed of 3 dot points all share the process variable A (for example, the shared "boiler load" is 70%), which is the simulation data of the simulation performed by changing other execution conditions.

Generally, it is difficult to strictly predict the actual conditions of the factory to plan the physical model simulation. Therefore, there are often cases where the simulation data of the physical model simulation deviates from the actual process data, and the operation status of the factory cannot necessarily be correctly determined based on the execution results of the simulation.

Therefore, as shown in this embodiment, by executing the physical model simulation based on the condition list which is a set of execution conditions corresponding to various operating conditions and accumulating simulation data, it is possible to generate a statistical model that takes into account the changes of various factors that can occur during actual plant operation.

The generation process of the first statistical model and the second statistical model and the method of judging process data illustrated in FIG7 and FIG8 are examples and do not limit the applicable objects of the present invention. That is, the present invention can be implemented using any statistical model as long as it is a statistical model that can learn data and judge the data. For example, the statistical model can use more than three process variables to judge the operating state, and can also judge the operating state based on the model generated by deep learning.

In addition, multiple thresholds can be set for each statistical model depending on the type of judgment result. For example, a threshold for setting a judgment result as "caution" and a threshold for setting a judgment result as "to be monitored" can be set separately.

Next, an example of a display screen of the support device 10 is described using FIG. 9 to FIG. 11. FIG. 9 is an example of a screen displaying the result of the determination of the operating state, and FIG. 10 and FIG. 11 are examples of screens displaying multiple statistical models at the same time. Each screen is displayed on the display 16a by the display control unit 14g. The display content can be displayed based on the information stored in the storage unit 18, and can be changed based on the input and operation received by the input receiving unit 12a and the operation receiving unit 12b.

First, the operation diagnosis screen 600 is explained using FIG9. The operation diagnosis screen 600 includes a timing chart display area 610, an operation status determination result display area 620, a model comparison button 630, a legend display area 640, a guide display area 650, and an alarm list display area 660.

In the timing chart display area 610, a timing chart is displayed together with a scroll bar 616, with the horizontal axis being the measurement time 614 and the vertical axis being the measurement value 612 of the process variable B. The process variable B may be a process variable that is particularly important when determining the operating status of the factory.

The legend display area 640 displays the legend of each information displayed in the operation status determination result display area 620.

The operation status determination result display area 620 displays the ideal line and threshold of the statistical model (first statistical model) generated based on the simulation data together with the learning data and the evaluation data. The type of the statistical model currently being referenced is displayed in the distinction information 624. In this example, since the statistical model being referenced is generated based on the simulation data, the learning data is the simulation data and the evaluation data is the process data.

The display content of the operation state determination result display area 620 can also be changed according to the user's operation. For example, the display of the learning data in the operation state determination result display area 620 can be switched by accepting the mouse operation from the user through the user interface 108 and pressing the learning data 642 part of the legend display area 640. In addition, the statistical model (second statistical model) generated by the process data can be switched and displayed by the method described later using Figures 10 and 11.

The guidance display area 650 displays the method 654 for viewing the chart and the method 656 for coping when the operating state is determined to be abnormal according to the display content of the timing chart display area 610. Although omitted in FIG. 9, articles that prompt the user are displayed in the method 654 for viewing the chart and the method 656 for coping.

The alarm list display area 660 displays the date and time of the most recent alarm (notification of abnormality in the operating state) and the date and time when the abnormality that caused the alarm was resolved. For example, abnormalities that have not been resolved after the alarm can be highlighted.

Next, using Figures 9 to 11, an example of a method for comparing the first statistical model and the second statistical model according to the user's operation and switching the statistical model displayed in the operating state determination result display area 620 is described.

The first example is a method of comparing the first statistical model and the second statistical model side by side. Specifically, in the operation diagnosis screen 600 of FIG. 9 , by pressing the model comparison button 630 , the screen is transitioned to the first statistical model comparison screen 700 of FIG. 10 .

FIG. 10 includes: a display setting area 710, a statistical model comparison display area 740 including an operation state determination result graph 742 and an operation state determination result graph 746, a legend display area 720 corresponding to the operation state determination result area 742, and a legend display area 730 corresponding to the operation state determination result area 746.

The operation state judgment result graph 742 is a scatter diagram when the operation state is judged by referring to the statistical model (second statistical model) generated based on the process data. On the other hand, the operation state judgment result graph 746 is a scatter diagram when the operation state is judged by referring to the statistical model (first statistical model) generated based on the simulation data.

The second example is a method of overlaying and comparing the first statistical model and the second statistical model. Specifically, in the operation diagnosis screen 600 of FIG. 9 , by pressing the model comparison button 630 , the screen is transitioned to the second statistical model comparison screen 800 of FIG. 11 .

FIG. 11 includes a display setting area 810, a statistical model comparison display area 830, and a legend display area 820.

The statistical model comparison display area 830 displays the following scatter plots in an overlay: a scatter plot when judging the operating state by referring to the statistical model (second statistical model) generated based on process data, and a scatter plot when judging the operating state by referring to the statistical model (first statistical model) generated based on simulation data.

The statistical model displayed in the operation state determination result display area 620 can be switched by pressing the formal model setting button 714d and the formal model setting button 716d. For example, in the initial state, as shown in FIG9 , when a statistical model generated based on simulation data is displayed in the operation state determination result display area 620, by pressing the formal model setting button 714d, the displayed chart can be switched to a statistical model generated based on process data.

As shown in the statistical model comparison display area 740 of FIG. 10 and the statistical model comparison display area 830 of FIG. 11 , by comparing the statistical models and displaying them, the user can determine the time to switch from the first statistical model to the second statistical model. Specifically, the comparison display is used as follows.

First, for the period when process data has not been sufficiently accumulated, as mentioned above, the statistical model (first statistical model) generated from the simulation data can determine the operating status with higher accuracy. On the other hand, if process data is sufficiently accumulated, the statistical model (second statistical model) generated from the process data can determine the operating status with higher accuracy, and the process data also includes effects that cannot be expressed in the physical model simulation such as aging and degradation of the factory. Therefore, it is necessary to switch the statistical model used from the first statistical model to the second statistical model at an appropriate time point, but by comparing and displaying the statistical models, this time point can be easily determined.

The present invention is not limited to the above-mentioned implementation method, and can be used in various modifications. In the operation of the support device 10, it is not limited to all automation through computer calculation processing, but also includes at least a part of manual operation performed by operators. In addition, in the above-mentioned implementation method, the display screen described in Figures 9 to 11 is only an example and is not limited to this.

The implementation modes described in the above-mentioned implementation modes can be appropriately combined, modified or improved for use according to the purpose, and the present invention is not limited to the description of the above-mentioned implementation modes.

This application claims priority based on Japanese Patent Application No. 2022-052955 filed on March 29, 2022. The entire contents of the Japanese application are incorporated by reference in this specification.

1: Factory

2: DCS (Distributed Control System)

10: Support devices

12:Input part

12a: Input acceptance department

12b: Operation acceptance department

14: Processing Department

14a: Control Department

14b: Process data acquisition department

14c: Statistical model generation unit

14d: Random number generation department

14e: Physical model simulator department

14f: Operation status determination unit

14g: Display control unit

16: Display unit

16a: Display

18: Memory Department

18a: Process data storage unit

18b: Analog data memory unit

18c: Condition list memory

18d: Differentiated information memory unit

18e: Statistical model memory unit

18f: Physical model memory

18g: Judgment result storage unit

Claims (12)

A support device for supporting the operation of a factory, comprising: a data acquisition unit, which acquires simulation data related to the operation state of the aforementioned factory by inputting data including a predetermined condition list into a physical model simulator; a statistical model generation unit, which generates a first statistical model by learning the aforementioned simulation data; and an operation state determination unit, which refers to the aforementioned first statistical model and determines the operation state of the aforementioned factory according to the process data of the aforementioned factory; wherein, in the aforementioned predetermined condition list, factors that determine the operation state of the aforementioned factory and indicators for the aforementioned factors are associated. A support device as in claim 1, wherein the predetermined condition list includes one or more of the aforementioned factors, and one or more of the aforementioned indicators are established in association with each of the aforementioned factors. A support device as in claim 2, wherein: the aforementioned predetermined condition list includes a plurality of records; in the aforementioned records, one aforementioned indicator is established in association with each of the aforementioned factors. A support device as in any one of claim items 1 to 3, further comprising: an input receiving unit that receives input of the aforementioned indicator from a user; in the aforementioned predetermined condition list, at least a portion of the aforementioned indicator is set based on the aforementioned input. A support device as in any one of claim items 1 to 3, wherein, in the predetermined condition list, at least a portion of the indicators are set according to random numbers according to a predetermined probability distribution. A support device as in any one of claim items 1 to 3, further comprising: A display unit for displaying the first determination result determined by the operation state determination unit based on the first statistical model. A support device as claimed in claim 6, wherein: the statistical model generation unit further generates a second statistical model based on process data related to the operation status of the factory; the operation status determination unit further refers to the second statistical model to determine the operation status of the factory; the display unit displays at least one of the first determination result and the second determination result determined by the operation status determination unit based on the second statistical model. A support device as claimed in claim 7, wherein the display unit associates the distinction information for distinguishing the first determination result and the second determination result with at least one of the first determination result and the second determination result and displays the distinction information. The support device of claim 7 further comprises: an operation accepting unit that accepts an operation from a user related to switching of a display method of a judgment result; the display unit displays both the first judgment result and the second judgment result according to the operation from the user. A statistical model generating device is a device for generating a statistical model for determining the operating state of a factory, comprising: a data acquisition unit, which acquires simulation data related to the operating state of the aforementioned factory by inputting data including a predetermined condition list into a physical model simulator; and a statistical model generating unit, which generates a first statistical model by learning the aforementioned simulation data; wherein, in the aforementioned predetermined condition list, factors determining the operating state of the aforementioned factory and indicators for the aforementioned factors are associated. A method for supporting the operation of a factory comprises the following steps: A data acquisition step, which is to obtain simulation data related to the operation state of the aforementioned factory by inputting data containing a predetermined condition list into a physical model simulator; A statistical model generation step, which is to generate a first statistical model by learning the aforementioned simulation data; and An operation state determination step, which is to refer to the aforementioned first statistical model and determine the operation state of the aforementioned factory according to the process data of the aforementioned factory; Wherein, in the aforementioned predetermined condition list, the factors that determine the operation state of the aforementioned factory and the indicators for the aforementioned factors are associated. A support program for supporting the operation of a factory, the support program enables a computer to function as the following means: Data acquisition means, which acquires simulation data related to the operation state of the aforementioned factory by inputting data containing a predetermined condition list into a physical model simulator; Statistical model generation means, which generates a first statistical model by learning the aforementioned simulation data; and Operation state determination means, which refers to the aforementioned first statistical model and determines the operation state of the aforementioned factory based on the process data of the aforementioned factory; Wherein, in the aforementioned predetermined condition list, factors that determine the operation state of the aforementioned factory and indicators for the aforementioned factors are associated.

Images