WO2022249475A1 - Plant operation support system, plant operation support method, and plant operation support program - Google Patents

Abstract

This plant operation support system (2) is provided with an event recognition unit (13), an operation support information creation unit (15), an operation pattern recognition unit (19), an evaluation unit (20), and an updating unit (21). The event recognition unit (13) recognizes an event that has occurred in a plant (4), on the basis of an event recognition model (40). The operation support information creation unit (15) causes a display unit (30) to display a screen of operation support information, on the basis of an operation support information creation model (43, 44) and the event recognized by the event recognition unit (13). The operation pattern recognition unit (19) recognizes an operation pattern on the screen by an operator (U), on the basis of an operation model (45) and operation log data. The evaluation unit (20) evaluates the event recognition accuracy of the event recognition unit and the operation pattern recognition accuracy of the operation pattern recognition unit (19). The updating unit (21) updates the event recognition model (40), on the basis of the evaluation result from the evaluation unit (20).

Description

Plant operation support system, plant operation support method, and plant operation support program

The present disclosure relates to a plant operation support system, a plant operation support method, and a plant operation support program that support the operation of a plant that is monitored and controlled by a supervisory control system.

Conventionally, monitoring and control systems that monitor and control plants such as water and sewage plants, power plants, chemical plants, and steel plants are known. Such a monitoring and control system collects measurement data output from measuring devices installed at important points of equipment in a plant, and displays process data such as collected measurement data or data based on the measurement data on a display unit. The operator judges the state of the plant from the information displayed on the display unit, and inputs data such as set values to the supervisory control system so as to optimize it.

　Since it may be difficult to determine the state of the plant, a plant operation support system may be used to support the operation of the plant. Such a plant operation support system is equipped with various prediction functions, various simulations, anomaly detection functions, setting value guidance functions, etc., and the plant operation support system assists operators in making decisions.

There are restrictions on the number and duration of process data that can be displayed on operation monitoring screens such as those of supervisory control systems or plant operation support systems. The operator must switch the process data according to the events that occur in the plant.

Therefore, in Patent Document 1, the switching operation of the operation monitoring screen when an alarm occurs is recorded, and when a new alarm occurs, the switching operation of the operation monitoring screen performed by the operator in the past regarding the new alarm is extracted and displayed. Accordingly, techniques have been proposed to reduce the labor and time required to grasp the state of the plant.

JP 2012-113586 A

However, there are a huge number of events that can occur in plants, such as water quality abnormalities, quantity abnormalities, equipment failures, inspections, or plant modifications due to construction work, and it is difficult to anticipate all of them in advance. In the technique described in Patent Document 1, it is necessary to define in advance an event that generates an alarm before introducing the plant operation support system. difficult to provide.

The present disclosure has been made in view of the above, and an object thereof is to obtain a plant operation support system capable of providing operation support information for supporting plant operation regarding events that are not defined at the time of system introduction. and

In order to solve the above-described problems and achieve the object, the plant operation support system of the present disclosure is a plant operation support system that supports operation by an operator of a plant that is monitored and controlled by a supervisory control system, A monitoring control data acquisition unit, an event recognition unit, a driving support information creation unit, an operation pattern recognition unit, an evaluation unit, and an update unit are provided. The supervisory control data acquisition unit acquires supervisory control data including time-series measurement data output by a measuring device installed in a plant. The event recognition unit recognizes an event that has occurred in the plant based on an event recognition model for recognizing an event that has occurred in the plant from the supervisory control data and the supervisory control data. The operation support information creation unit creates operation support information based on the event recognized by the event recognition unit and the operation support information creation model that determines the operation support information, which is the information to support operation from the events that occurred in the plant. and display the created driving support information screen on the display unit. The operation pattern recognition unit recognizes the operation pattern based on the operation log data and the operation model for recognizing the operator's operation pattern from the operation log data including the operation history of the screen by the operator. The evaluation unit evaluates the event recognition accuracy by the event recognition unit and the operation pattern recognition accuracy by the operation pattern recognition unit based on the event recognized by the event recognition unit and the operation pattern recognized by the operation pattern recognition unit. do. The updating unit updates the event recognition model based on the evaluation result by the evaluating unit.

　According to the present disclosure, it is possible to provide operation support information for supporting plant operation regarding events that are not defined at the time of system introduction.

1 is a diagram showing an example of a configuration of a processing system including a supervisory control system and a plant operation support system according to a first embodiment; FIG. 4 is a diagram showing an example of logical expression information included in an event recognition model stored in the event recognition model storage unit according to the first embodiment; FIG. 4 is a diagram showing an example of propositional logic definition information included in the event recognition model stored in the event recognition model storage unit according to the first embodiment; FIG. FIG. 4 is a diagram showing an example of an event recognition result by an event recognition unit of the plant operation support system according to the first embodiment; FIG. FIG. 4 is a diagram showing an example of an operation support information creation model for a trend graph stored in an operation support information creation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram showing an example of an operation support information creation model for scatter diagrams stored in an operation support information creation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of an operation support screen created by an operation support information creation unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of an operation support screen created by an operation support information creation unit of the plant operation support system according to the first embodiment and displayed on the display unit; FIG. 4 is a diagram showing an example of operation log data definition information stored in the operation log data storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of operation pattern definition information included in an operation model stored in an operation model storage unit of the plant operation support system according to the first embodiment; FIG. FIG. 4 is a diagram schematically showing an example of operation purpose definition information in an operation model stored in an operation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of first definition information included in display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of second definition information included in display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of third definition information included in display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of display period pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment; FIG. 4 is a diagram schematically showing an example of an operator's operation pattern recognized by the operation pattern recognition unit according to the first embodiment; FIG. FIG. 4 is a diagram showing an example of evaluation information used by the evaluation unit according to the first embodiment; 1 schematically represents an example of logical expression information included in an event recognition model updated by an update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low; figure An example of the propositional logic definition information included in the event recognition model updated by the update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of the event recognition is low is schematically illustrated. diagram to represent 1 schematically represents an example of an operation support information creation model for a trend graph updated by an update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low; figure 1 schematically represents an example of an operation support information creation model for a scatter diagram updated by an update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low; figure 4 is a flowchart showing an example of processing by a processing unit of the plant operation support system according to the first embodiment; 4 is a flowchart showing an example of evaluation update processing by the processing unit of the plant operation support system according to the first embodiment; 1 is a diagram showing an example of a hardware configuration of a plant operation support system according to a first embodiment; FIG. FIG. 11 is a diagram showing an example of a configuration of a processing system according to a second embodiment; FIG. 8 is a diagram schematically showing an example of an operation support screen created by an operation support information creation unit of the plant operation support system according to the second embodiment and displayed on the display unit; FIG. 11 is a diagram schematically showing another example of the operation support screen created by the operation support information creation unit of the plant operation support system according to the second embodiment and displayed on the display unit; FIG. 9 is a diagram schematically showing an example of event name editing information stored in an event name editing information storage unit of the plant operation support system according to the second embodiment; FIG. 11 is a diagram showing an example of a configuration of a processing system according to a third embodiment; FIG. 11 is a diagram showing an example of a configuration of a processing system according to a fourth embodiment; FIG. 12 is a diagram schematically showing an example of an operation support screen created by an operation support information creation unit of the plant operation support system according to the fourth embodiment and displayed on the display unit; A diagram showing an example of a configuration of a processing system according to a fifth embodiment. FIG. 14 is a diagram schematically showing an example of an operation support screen created by an operation support information creation unit of the plant operation support system according to the fifth embodiment and displayed on the display unit; FIG. 11 is a diagram schematically showing an example of selected feature amount information stored in a selected feature amount information storage unit of the driving support system according to the fifth embodiment; FIG. 12 is a diagram schematically showing an example of feature amount templates stored in a feature amount template storage unit of the driving support system according to the fifth embodiment;

The plant operation support system, plant operation support method, and plant operation support program according to the embodiment will be described in detail below with reference to the drawings.

Embodiment 1.
FIG. 1 is a diagram illustrating an example of a configuration of a processing system including a monitoring control system and a plant operation support system according to a first embodiment; As shown in FIG. 1, the processing system 100 according to the first embodiment includes a monitoring control system 1, a plant operation support system 2, and an input/output device 3. In the processing system 100 , the input/output device 3 may be included in the supervisory control system 1 , or the supervisory control system 1 and the plant operation support system 2 may have different input/output devices 3 .

The monitoring and control system 1 monitors and controls the plant 4. The plant 4 is a water and sewage plant, a power plant, a chemical plant, or a steel plant, but is not limited to these examples. The plant 4 is provided with a plurality of measuring instruments, and the monitoring and control system 1 collects measurement data measured by the plurality of measuring instruments.

For example, if plant 4 is a sewage plant, plant 4 includes equipment such as settling basins, pump wells, water pumps, primary sedimentation basins, treatment tanks, blowers, final sedimentation basins, return pumps, and chlorination basins. be The settling basin removes relatively large debris and sand contained in the inflowing sewage. The pump well stores sewage coming from the settling basin. The water pump lifts sewage from the pump well to the primary sedimentation basin.

The primary sedimentation basin settles the relatively small debris and sand contained in the sewage to the bottom and removes the relatively small debris and sand from the sewage. The blower adds activated sludge to the sewage in the treatment tank and blows air into it to aerate the sewage, thereby removing organic matter from the sewage.

The final sedimentation tank separates the activated sludge mixture flowing from the treatment tank into supernatant water and activated sludge. The return pump is provided in a pipe connecting the final sedimentation tank and the treatment tank, and returns the activated sludge from the final sedimentation tank to the treatment tank. The chlorination basin adds chlorine to the supernatant water flowing from the final sedimentation basin for sterilization.

When the plant 4 is a sewage treatment plant, the plant 4 is provided with a flowmeter, a water level meter, an airflow meter, various concentration meters, etc. as the above-described measurement equipment. Flowmeters are, for example, a flowmeter that measures the flow rate of sewage flowing into a pump well, a flowmeter that measures the flow rate of a water pump, and a flowmeter that measures the flow rate of water containing activated sludge that is returned from the final sedimentation tank to the treatment tank. It is a flow meter, or a flow meter that measures the flow rate of treated water discharged from the plant 4, or the like.

A water level gauge is, for example, a water level gauge that measures the water level of a pump well. The air flow meter is, for example, an air flow meter that measures the amount of aeration air in the treatment tank. The densitometer is a dissolved oxygen amount sensor that detects the amount of dissolved oxygen in the treatment tank, an active microorganism concentration sensor that detects the concentration of active microorganisms in the treatment tank, a BOD sensor that detects BOD (Biochemical Oxygen Demand) in the treatment tank, or plant 4 It is a nitrogen concentration meter that measures the discharged nitrogen concentration, which is the concentration of nitrogen in the water discharged from.

The input/output device 3 includes a display section 30 and an input section 31 . The display unit 30 is, for example, an LCD (Liquid Crystal Display) or an organic EL (Electro-Luminescence) display. The input unit 31 includes, for example, a keyboard, mouse, keypad, or touch panel, and is operated by the user of the processing system 100 . A user of the treatment system 100 is an operator U who manages the operation of the plant 4 .

The monitoring control system 1 controls the equipment of the plant 4 or the equipment provided in the equipment based on the control input data output from the input/output device 3 . The control input data is, for example, set value data for the flow rate of a water pump, data for set value data for the aeration air volume of a blower, or data for set value data for the flow rate of water containing activated sludge.

The plant operation support system 2 causes the display unit 30 of the input/output device 3 to display an operation support screen, which is a screen of operation support information including process data such as measurement data collected from the plant 4 or data based on the measurement data. In addition, the plant operation support system 2 selects an operation support screen, selects a signal to be displayed on the operation support screen, and selects the period of the signal to be displayed on the operation support screen based on the operation of the input unit 31 by the operator U. A process of changing the contents of the screen displayed on the display unit 30 is performed in response to an operation such as selection, selection of a time of interest, or input of a set value.

The plant operation support system 2 has a function to support the operation of the plant 4 by the operator U. The plant operation support system 2 includes a monitoring control data acquisition unit 10, a monitoring control data storage unit 11, an event recognition model storage unit 12, an event recognition unit 13, an operation support information creation model storage unit 14, and an operation support and an information creation unit 15 .

The plant operation support system 2 also includes an operation data acquisition unit 16, an operation log data storage unit 17, an operation model storage unit 18, an operation pattern recognition unit 19, an evaluation unit 20, and an update unit 21. A processing unit 22 is configured to include the monitoring control data acquisition unit 10 , the event recognition unit 13 , the driving support information creation unit 15 , the operation data acquisition unit 16 , the operation pattern recognition unit 19 , the evaluation unit 20 and the update unit 21 .

The monitoring control data acquisition unit 10 acquires monitoring control data from the monitoring control system 1 and stores the acquired monitoring control data in the monitoring control data storage unit 11 . The monitoring control data includes time-series measurement data output by measuring equipment installed in the equipment of the plant 4 and control input data, which is control data input by the operator U.

The event recognition model storage unit 12 stores an event recognition model 40 that recognizes events occurring in the plant 4 from the monitoring control data. The event recognition model 40 includes logical formula information representing each event by a logical formula of propositional logic, and propositional logic definition information defining the propositional logic. FIG. 2 is a diagram depicting an example of logical expression information included in an event recognition model stored in an event recognition model storage unit according to the first embodiment;

The logical formula information 41 shown in FIG. 2 includes an event ID (IDentifier) 411, an event name 412, and a judgment formula 413 for each event. The event ID 411 is unique identification information for each event. The event name 412 is information indicating the name of the event. A determination expression 413 is information for determining an event. The "judgment formula" is information indicating the logic formula of the propositional logic to be judged by the supervisory control data.

In the logical expression information 41 shown in _FIG . 2, for example, the event ID 411 is “a ₁ ” and the event name 412 is “midnight on a _sunny day”. ₆ p ₈ p ₁₀ ) (p ₁ p ₄ p ₆ p ₉ p ₁₁ )”. "Sunny day midnight" indicates that the event is at midnight on a sunny day.

Further, in the logical expression information 41 shown in FIG. 2, for example _, the event ID 411 is "a ₂ " and the event name 412 is "sunny day rating" _. p ₆ ∧ p ₈ ∧ p ₁₁ ''. A "sunny day rating" indicates that the event is a sunny day and that plant 4 operating conditions are normal.

Further, in the logical expression information 41 shown in FIG. 2, for example, the event ID 411 is "a ₃ " and the event name 412 is "clear day peak response", and the determination formula 413 of the event is "p ₂ ∧ _{p 5} ∧ p ₇ ∧ p ₉ ∧ p ₁₂ ”. "Sunny day peak correspondence" indicates that the event is a sunny day and the plant 4 operating condition is at peak condition.

FIG. 3 is a diagram showing an example of propositional logic definition information included in an event recognition model stored in the event recognition model storage unit according to the first embodiment. The propositional logic definition information 42 shown in FIG. 3 includes a propositional logic ID 421, target time-series data ID 422, target time-series data name 423, unit 424, and judgment formula 425 for each event.

The propositional logic ID 421 is unique identification information for each propositional logic. The target time-series data ID 422 is identification information unique to each target time-series data that is time-series measurement data that is the target of the propositional logic. The target time-series data name 423 is information indicating the name of the target time-series data. The unit 424 is information indicating the unit of the target time-series data. The judgment formula 425 is information for judging the propositional logic.

In the propositional logic definition information 42 shown in FIG. 3, for example, the propositional logic with the propositional logic ID 421 of “p ₁ ” has the target time-series data ID 422 of “x ₁ ” and the target time-series data name 423 of “pump flow rate , the unit 424 of the target time-series data is "m ³ /hr", and the determination expression 425 is "x ₁ ≦y ₁ ¹ ". The target time-series data whose target time-series data ID 422 is "x ₁ " is the time-series data of the measurement data of the measuring device that measures the flow rate of the water pump. “y ₁ ¹ ” is a threshold value, and when the hourly flow rate of the water pump is equal to or less than “y ₁ ¹ ”, the propositional logic with the propositional logic ID 421 of “p ₁ ” is satisfied.

Further, in the propositional logic definition information 42 shown in FIG. 3, for example, the propositional logic with the propositional logic ID 421 of "p ₂ " has the target time-series data ID 422 of "x ₁ " and the target time-series data name 423 of "Pumping The unit 424 of the target time-series data is "m ³ /hr", and the determination expression 425 is "y ₁ ¹ <x ₁ ≦y ₁ ² ". “y ₁ ¹ ” and “y ₁ ² ” are thresholds, and when the hourly flow rate of the water pump is greater than “y ₁ ¹ ” and less than or equal to “y ₁ ² ”, the propositional logic ID 421 is “p ₂ ” satisfies the propositional logic.

Further, in the propositional logic definition information 42 shown in FIG. 3, for example, the propositional logic with the propositional logic ID 421 of "p ₃ " has the target time-series data ID 422 of "x ₁ " and the target time-series data name 423 of "Pumping The unit 424 of the target time-series data is "m ³ /hr", and the determination formula 425 is "y ₁ ² <x ₁ ". “y ₁ ² ” is a threshold, and the propositional logic with the propositional logic ID 421 of “p ₃ ” is satisfied when the hourly flow rate of the water pump is greater than “y ₁ ² ”.

Further, in the propositional logic definition information 42 shown in FIG. 3, for example, the propositional logic with the propositional logic ID 421 of "p ₄ " has the target time-series data ID 422 of "x ₂ " and the target time-series data name 423 of "pump well water level”, the unit 424 of the target time-series data is “m”, and the determination formula 425 is “x ₂ ≦y ₂ ¹ ”. The target time-series data whose target time-series data ID is "x ₂ " is the time-series data of the measurement data of the measuring device that measures the water level of the pump well. “y ₂ ¹ ” is a threshold value, and the propositional logic with the propositional logic ID 421 of “p ₄ ” is satisfied when the water level of the pump well is equal to or less than “y ₂ ¹ ”.

Further, in the propositional logic definition information 42 shown in FIG. 3, for example, the propositional logic with the propositional logic ID 421 of "p ₅ " has the target time-series data ID 422 of "x ₂ " and the target time-series data name 423 of "pump The unit 424 of the target time-series data is "m", and the determination formula 425 is "y ₂ ² <x ₂ ". “y ₂ ² ” is a threshold, and the propositional logic with propositional logic ID 421 of “p ₅ ” is satisfied when the water level of the pump well is greater than “y ₂ ² ”.

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The event recognition unit 13 of the plant operation support system 2 recognizes an event occurring in the plant 4 from the monitoring control data stored in the monitoring control data storage unit 11 and the event recognition model 40 stored in the event recognition model storage unit 12. recognize.

For example, assume that the logical formula information 41 is the information shown in FIG. 2 and the propositional logic definition information 42 is in the state shown in FIG. In this case, the event recognizing unit 13 uses the determination formula 413 “(p ₁ ∧ p ₄ ∧ _{p 6} ∧ p ₈ ∧ _{p 10} ) ∨ (p ₁ ∧ p ₄ ∧ _{p 6} ∧ p ₉ ∧ p ₁₁ )”, it recognizes that the event that occurred in the plant 4 is “midnight on a sunny day”.

Further, the event recognizing unit 13 determined that the determination formula 413 “p ₂ ∧ _{p 4} ∧ _{p 6} ∧ p ₈ ∧ _{p 11} ” is satisfied based on the monitoring control data stored in the monitoring control data storage unit 11 . In this case, the event that occurred in plant 4 is recognized as "sunny day rating". Further, the event recognizing unit 13 determined that the determination formula 413 “p ₂ ∧ _{p 5} ∧ _{p 7} ∧ p ₉ ∧ _{p 12} ” is satisfied based on the monitoring control data stored in the monitoring control data storage unit 11 . In this case, the event that occurred in the plant 4 is recognized as "clear day peak response".

FIG. 4 is a diagram illustrating an example of an event recognition result by an event recognition unit of the plant operation support system according to the first embodiment; In the example shown in FIG. 4, the event recognition unit 13 recognizes an event with an event ID of "a ₁ ", an event with an event ID of "a ₂ ", and an event with an event ID of "a ₆ " in this order. It is shown that The event recognition unit 13 can also recognize multiple events at the same time.

In the above example, the propositional logic definition information 42 is defined with the time-series measurement data included in the supervisory control data as the target data. In addition to the series of measurement data, the definition of the propositional logic using the control input data contained in the supervisory control data may be included. In this case, the event recognition unit 13 recognizes the event based on the time-series measurement data and the control input data.

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The operation support information creation model storage unit 14 of the plant operation support system 2 stores the operation support information creation models 43 and 44 for determining the operation support information, which is the information for assisting the operation of the operator U from the events occurring in the plant 4. do. The driving support information creation model 43 is a trend graph driving support information creation model, and the driving support information creation model 44 is a scatter graph driving support information creation model.

FIG. 5 is a diagram showing an example of an operation support information creation model for trend graphs stored in the operation support information creation model storage unit of the plant operation support system according to the first embodiment. As shown in FIG. 5, the driving support information creation model 43 for the trend graph converts information of an event ID 431, an event name 432, a conditional probability 433 in units of display time-series data, and a conditional probability 434 in units of display period to an event. Including every

The event ID 431 is identification information unique to each event and is the same as the event ID 411 included in the logical expression information 41. The event name 432 is information indicating the name of the event, and is the same as the event name 412 included in the logical expression information 41 . The conditional probability 433 for each display time-series data is information representing the appropriateness of displaying each time-series data in the trend graph when the event recognition unit 13 recognizes an event. The display period unit conditional probability 434 is information representing the appropriateness of the display period for the trend graph of the time-series data when the event recognition unit 13 recognizes an event.

In the driving support information creation model 43 shown in FIG. 5, when the event recognition unit 13 recognizes an event whose event ID 431 is "a ₁ ", time-series data whose target time-series data ID 422 is "x ₁ " The conditional probability indicating the appropriateness of being displayed on the trend graph is "P(x ₁ |a ₁ )", and the conditional probability indicating the appropriateness of the display period being "l ₁ " is "P( l ₁ |a ₁ )”.

Further, in the driving support information creation model 43 shown in FIG. 5, when an event whose event ID 431 is "a ₂ " is recognized by the event recognition unit 13, time-series data whose target time-series data ID 422 is "x ₁ " is displayed on the trend graph is "P(x ₁ |a ₂ )", and the conditional probability that the display period is "l ₂ " is "P (l ₂ |a ₂ )”.

FIG. 6 is a diagram showing an example of an operation support information creation model for scatter diagrams stored in the operation support information creation model storage unit of the plant operation support system according to the first embodiment. As shown in FIG. 6, the driving support information creation model 44 for the scatter diagram includes information on an event ID 441, an event name 442, a conditional probability 443 for each combination of display time-series data, and a conditional probability 444 for each display period. for each event.

The event ID 441 is unique identification information for each event and is the same as the event ID 411 included in the logical expression information 41. The event name 442 is information indicating the name of the event, and is the same as the event name 412 included in the logical expression information 41 . The conditional probability 443 for each combination of display time-series data is information representing the appropriateness of displaying each time-series data in the scatter diagram when the event recognition unit 13 recognizes an event. The display period unit conditional probability 444 is information representing the appropriateness of the display period for the scatter diagram of the time-series data when the event recognition unit 13 recognizes the event.

In the driving support information creation model 44 shown in FIG. 6, when the event recognition unit 13 recognizes an event whose event ID 441 is "a ₁ ", the target time-series data ID 422 is "x ₁ ", "x ₂ ". The conditional probability indicating the appropriateness of displaying two pieces of time-series data in the scatter diagram is “P(x ₁ , x ₂ |a ₁ )”, and the appropriateness of the display period being “l ₁ ”. It is shown that the conditional probability that indicates the probability is "P(l ₁ |a ₁ )".

Further, in the driving support information creation model 44 shown in FIG. 6, when the event recognition unit 13 recognizes an event whose event ID 441 is "a ₂ ", the target time-series data ID 422 is "x ₁ ", "x _{2 "} . ” is “P(x ₁ , x ₂ |a ₂ )”, and the display period is “l ₂ ”. It is shown that the conditional probability of some adequacy is "P(l ₂ |a ₂ )".

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. Based on the event recognized by the event recognition unit 13 and the driving support information creation models 43 and 44 stored in the operation support information creation model storage unit 14, the operation support information creation unit 15 of the plant operation support system 2 Driving assistance information is created, and the created driving assistance information is displayed on the display unit 30 of the input/output device 3 .

FIG. 7 is a diagram schematically showing an example of an operation support screen created by the operation support information creation unit of the plant operation support system according to the first embodiment. As shown in FIG. 7, the driving support screen 60 created by the driving support information creating unit 15 includes a time-series data selection area 601, a message display area 602, a period selection area 603, a trend graph display area 604, and a It includes a scatter diagram display area 605 and a setting value input area 606 . The driving support screen 60 is displayed on the display unit 30 by the driving support information creation unit 15 .

The time-series data selection area 601 includes a check box 601a for trend graph display and a check box 601b for scatter graph display. The check box 601a for trend graph display is a condition representing the appropriateness of displaying time-series data regarding an event recognized by the event recognition unit 13 according to the number of trend graphs that can be displayed in the trend graph display area 604. Time-series data with high probability is checked. For example, if there are three trend graphs that can be displayed, the three time-series data with the highest conditional probabilities for the recognized event are checked.

In the scatter diagram display checkbox 601b, two time-series data constituting the time-series data set with the highest conditional probability representing the adequacy of displaying the time-series data set for the recognized event are checked. enters. The message display area 602 displays the time when the event recognition unit 13 newly recognized the event and the event name of the newly recognized event.

In the period selection area 603, the time series data up to a preset time in the past is displayed for the data for which the check box for trend graph display in the time series data selection area 601 is checked. The past preset time is, for example, a period selected from the specified period. The designated period is a period designated as a target for event recognition and operation pattern recognition, and is designated within an overlapping period between the acquisition period for monitoring control data and the acquisition period for operation log data, which will be described later.

The acquisition period of the supervisory control data is, for example, the period from the time when the oldest supervisory control data among the supervisory control data stored in the plant operation support system 2 is obtained to the time when the latest supervisory control data is obtained. be. Further, the acquisition period of the operation log data is, for example, the period from the time when the oldest operation data among the operation data stored in the plant operation support system 2 is acquired to the time when the newest operation data is acquired. .

Also, in the period selection area 603, the display period with the largest conditional probability representing the adequacy of the display period of the time-series data for the event recognized by the event recognition unit 13 is selected retroactively from the current time. is displayed.

In the trend graph display area 604, the time series data of the period indicated by the selection period range 603a in the period selection area 603 is displayed as a trend for the data for which the checkbox 601a for trend graph display in the time series data selection area 601 is checked. Display time-series data units in graph format.

In the scatter diagram display area 605, the time series data of the period indicated by the selection period range 603a in the period selection area 603 is scattered for the data for which the check box 601b for scatter diagram display in the time series data selection area 601 is checked. Displayed in the form of diagrams. A setting value input area 606 displays a box 606a for selecting target time-series data for inputting a setting value and a box 606b for inputting a setting value.

FIG. 8 is a diagram schematically showing an example of an operation support screen created by the operation support information creation unit of the plant operation support system according to the first embodiment and displayed on the display unit. As shown in FIG. 8, the driving assistance screen 61 created by the driving assistance information creating unit 15 and displayed on the display unit 30 includes a time-series data selection area 611, a message display area 612, a period selection area 613, and a trend graph. It includes a display area 614 , a scatter plot display area 615 and a setpoint input area 616 .

The time-series data selection area 611 includes a check box 611a for trend graph display and a check box 611b for scatter graph display. For example, it is assumed that the operator U operates the input unit 31 to change the time-series data to be checked in the check box 611a. In this case, the driving support information creation unit 15 displays the driving support screen 61 in which the time-series data displayed in the period selection area 613 and the trend graph display area 614 are changed to the time-series data whose check boxes 611a are checked. 30.

In the example shown in FIG. 8, the trend graph of the time-series data whose time-series data ID is "x ₁ ", the trend graph of the time-series data whose time-series data ID is "x ₂ ", and the time-series data ID of A trend graph of time-series data that is 'x ₄ ' is shown.

It is also assumed that the operator U has changed the time-series data to be checked in the check box 611b by operating the input unit 31. In this case, the driving support information creation unit 15 causes the display unit 30 to display the driving support screen 61 in which the time-series data displayed in the scatter diagram display area 615 is changed to the time-series data with the check box 611b checked.

The period selection area 613 includes a selection period range 613a and a period selection bar 613b. A selection period range 613a indicates the period of data to be displayed in each of the trend graph display area 614 and the scatter diagram display area 615. A period selection bar 613b is a bar for changing the selection period range 613a. 613a and is movable in the selection period range 613a.

When the operation of the input unit 31 by the operator U is to move the period selection bar 613b, the driving support information creation unit 15 moves the selection period range 613a with the period selection bar 613b as the central time, and displays the trend graph The period of the time-series data displayed in each of the display area 614 and the scatter diagram display area 615 is changed to the period corresponding to the selection period range 613a.

In the trend graph display area 614, the trend graph of the time-series data whose check box 611a is checked is displayed. The trend graph display area 614 also includes a time-of-interest selection bar 614a. The attention time selection bar 614a is a bar for changing the attention time, which is the center time of the selection period range 613a, in conjunction with the period selection bar 613b.

When the operator U operates the input unit 31 to move the attention time selection bar 614a, the driving support information creation unit 15 moves the attention time selection bar 614a together with the period selection bar 613b and the selection period range 613a. .

The driving support information creation unit 15 displays in the trend graph display area 614 the value of the data at the time indicated by the attention time selection bar 614 a among the time-series data. In the example shown in FIG. 8, the values of the time data indicated by the time-of-interest selection bar 614a are the time-series data value "y ₁ ^* " whose time-series data ID is "x ₁ " and the time-series data ID of " A time-series data value “y ₂ ^* ” whose time-series data ID is “x ₂ ” and a time-series data value “y ₄ *” whose time-series data ID is “x ₄ ^” are shown.

In addition, the driving support information creation unit 15 highlights the data of the time selected by the attention time selection bar 614a or the period selection bar 613b in the scatter diagram display area 615 as the attention data 615a. The highlighting is performed, for example, by making the color, size, shape, or the like of the attention data 615a plotted in the scatter diagram different from other data.

Further, in the scatter diagram display area 615, when the attention data 615a, which is the data that the operator U pays attention to, is selected by the operation of the input unit 31 by the operator U, the driving support information creation unit 15 selects the attention data 615a. The target time selection bar 614a and the period selection bar 613b are moved to the position of the measured time.

Further, the driving support information creating unit 15 displays the value of the attention data 615 a in the scatter diagram display area 615 . In the example shown in FIG. 8, the value of the data of interest 615a is the time-series data value "y ₂ ^* " whose time-series data ID is "x ₂ " and the time-series data value whose time-series data ID is "x ₄ ". The data value is " _y4 ^* ".

The setting value input area 616 includes a box 616a for selecting the target time-series data for inputting the setting value and a box 616b for inputting the setting value. The operator U can select the time-series data name of the time-series data to be the target time-series data from among the time-series data names of the plurality of time-series data contained in the box 616a by operating the input unit 31 . Further, the operator U can input the setting value in the box 616b by operating the input unit 31. FIG.

In the driving support screen 61 shown in FIG. 8, the target time-series data whose target time-series data ID 422 is "x ₅ " is selected in box 616a, and "y ₅ ^* " is selected as the setting value in box 616b. It is shown that there are

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The operation data acquisition unit 16 of the plant operation support system 2 acquires from the input/output device 3 operation data, which is data indicating the operation of the input unit 31 by the operator U, and stores the acquired operation data in the operation log data storage unit 17. be memorized. The operation log data storage unit 17 stores operation log data including a plurality of past time-series operation data.

The operation log data includes first operation log time-series data, second operation log time-series data, third operation log time-series data, and fourth operation log time-series data. For the first operation log time-series data, for example, the type of time-series data selected to be displayed in the trend graph display area 614 by operating the driving support information creation unit 15 or the input unit 31 from the operator U is time. This is time-series data represented by series.

For the second operation log time-series data, for example, the type of time-series data selected to be displayed in the scatter diagram display area 615 by operating the driving support information creation unit 15 or the input unit 31 from the operator U is time. This is time-series data represented by series.

The third operation log time-series data is, for example, data represented in time series by the display period of each trend graph and the display period of each scatter diagram on the driving support screen 61 . The fourth operation log time-series data is time-series data in which the time of interest selected on the driving support screen 61 is represented in time series.

In addition to the operation log data described above, the operation log data storage unit 17 stores operation log data definition information. 9 is a diagram of an example of operation log data definition information stored in an operation log data storage unit of the plant operation support system according to the first embodiment; FIG.

The operation log data definition information 58 shown in FIG. 9 includes an operation log time-series data ID 581 and a time-series data name 582 for each operation log time-series data. The operation log time-series data ID 581 is unique identification information for each operation log time-series data. The time-series data name 582 is information indicating the name of the operation log time-series data.

In the operation log data definition information 58 shown in FIG. 9, the name of the operation log time-series data with the operation log time-series data ID 581 "z ₁ " is "selected time-series_trend", and the operation log time-series data ID 581 "z 1 " is the name of the operation log time-series data. ₂ ” is the name of the operation log time-series data “selected time-series_scatter diagram”.

Further, in the operation log data definition information 58, the name of the operation log time-series data with the operation log time-series data ID 581 “z ₃ ” is “display period”, and the operation log time-series data with the operation log time-series data ID 581 “z ₄ ” is named “display period”. The name of the time-series data is "selected time".

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The operation model storage unit 18 of the plant operation support system 2 stores an operation model 45 that recognizes the operation pattern of the operator U from the operation log data stored in the operation log data storage unit 17 .

The operation model 45 includes operation pattern definition information that defines operation patterns, operation purpose definition information that defines operation purposes, display signal pattern definition information that defines display signal patterns, and display period pattern definitions that define display period patterns. including information.

FIG. 10 is a diagram schematically showing an example of operation pattern definition information included in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. Operation pattern definition information 46 included in operation model 45 shown in FIG. 10 includes operation pattern ID 461, operation purpose ID 462, display signal pattern ID 463, display period pattern ID 464, and conditional probability 465.

The operation pattern ID 461 is unique identification information for each operation pattern. The operation purpose ID 462 is unique identification information for each operation purpose. The display signal pattern ID 463 is unique identification information for each combination of time-series data displayed on the driving support screen 61 . The display period pattern ID 464 is identification information unique to each display period of the time-series data selected from the time-series data in the specified period.

The conditional probability 465 represents the probability of recognizing each operation pattern ID 461 when the display signal pattern ID 463 and the display period pattern ID 464 are recognized. In the example shown in FIG. 10, when the operation purpose of the operation purpose ID 462 “q ₁ ”, the display signal pattern of the display signal pattern ID 463 “r ₁ ”, and the display period pattern of the display period pattern ID 464 “s ₁ ” are recognized, The conditional probability 465 representing the appropriateness of recognizing the operation pattern of the operation pattern ID 461 “b ₁ ” is “P(b ₁ |q ₁ ,r ₁ ,s ₁ )”.

Further, for example, when the operation purpose of the operation purpose ID 462 “q ₁ ”, the display signal pattern of the display signal pattern ID 463 “r ₂ ”, and the display period pattern of the display period pattern ID 464 “s ₂ ” are recognized, the operation pattern ID 461 The conditional probability 465 representing the adequacy of recognizing the operation pattern of " _b2 " is "P( _b2 | _q1 , _r2 , s2) _" .

FIG. 11 is a diagram schematically showing an example of operation purpose definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. The operational purpose definition information 47 shown in FIG. 11 includes an operational purpose ID 471, an operational purpose 472, and a determination expression 473 for each operational purpose.

The operation purpose ID 471 is unique identification information for each operation pattern, and is the same as the operation pattern ID 461 shown in FIG. The operation purpose 472 is information indicating the operation purpose, and is the same as the operation purpose ID 462 shown in FIG. The determination formula 473 is information of a determination formula for determining the purpose of operation.

In the determination formula 473, “z ₁ ”, “z ₂ ”, “z ₃ ”, and “z ₄ ” are operation log data, which are time-series data of operation data. In the determination formula 473, "w ₁ ", "w ₂ ", " ^w ₃ ", "w ₄ ", . is a function that calculates an index for determining the purpose of operation from operation log data.

For example, “w ₁ ” is a condition that the number of times the time-series data displayed on the driving assistance screen 61 has been changed in the past hour is 10 or more, and “w ₂ ” is a condition displayed on the driving assistance screen 61. The condition is that the number of times the selected period range 613a has been changed is 30 or more. “w ₃ ” is, for example, a condition that the number of times the time of interest has been changed by operating the time of interest selection bar 614a or the period selection bar 613b is 50 or more.

In the operation purpose definition information 47 shown in FIG. 11, the determination formula 473 of the operation purpose 472 of "set value change amount confirmation" whose operation purpose ID 471 is " _q1 " is " ^fq ( _z1 , _z2 , _z3 , z ₄ )≡{w ₁ , w ₂ }” and the operation purpose ID 471 is “q ₂ ”, the determination formula 473 of the operation purpose 472 of “abnormality factor identification” is “f ^q (z ₁ , z ₂ , z ₃ , z ₄ ) ≡ {w ₃ , w ₄ , w ₅ }'.

Further, in the operation purpose definition information 47 shown in FIG. 11, the determination expression 473 of the operation purpose 472 of "confirm influence propagation" whose operation purpose ID 471 is "q ₄ " is "f ^q (z ₁ , z ₂ , z ₃ , z ₄ )≡{w ₆ }” and the operation purpose ID 471 is “q ₅ ”, the determination formula 473 of the operation purpose 472 of “check increase/decrease trend” is “f ^q (z ₁ , z ₂ , z ₃ , z ₄ ) ≡ {w ₆ , w ₇ }'.

FIG. 12 is a diagram schematically showing an example of first definition information included in display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. The first definition information 48a shown in FIG. 12 includes a display signal pattern ID 481 and a determination expression 482 for each display signal pattern.

The display signal pattern ID 481 is unique identification information for each combination of time-series data displayed on the driving support screen 61, and is the same as the display signal pattern ID 463 shown in FIG. The determination formula 482 is information on the determination formula for determining the display signal pattern.

In the first definition information 48a shown in FIG. 12, the determination expression 482 of the display signal pattern whose display signal pattern ID 481 is "r ₁ " is "r ₁ ^t ∧ _{r 1} ^s ", and the display signal pattern ID 481 is "r ₂ ” is “r ₁ ^t ∧ r ₂ ^s ”.

FIG. 13 is a diagram schematically showing an example of second definition information included in display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. The second definition information 48b shown in FIG. 13 includes a display signal pattern ID 483 and a determination expression 484 for each display signal pattern of the trend graph.

The display signal pattern ID 483 is information indicating the first display signal pattern, which is the combination pattern of the time-series data displayed in the trend graph display area 614 of the driving support screen 61 among the time-series data in the specified period. The determination formula 484 is information on the determination formula for determining the first display signal pattern.

In the determination formula 484 , “x ₁ ”, “x ₂ ”, “x ₃ ”, . Also, in the determination formula 484, "z ₁ " is the first operation log time-series data described above.

In the determination formula 484, "f ^rt (·)" calculates the first display signal pattern, which is the combination pattern of the time-series data displayed in the trend graph display area 614, from the operation log data "z ₁ ". It is a function for

In the second definition information 48b shown in FIG. 13, for example, the determination formula 484 of the display signal pattern whose display signal pattern ID 483 is " _r1t " is " ^frt ( ^z1 )≡{ _{x1 , x2} ,x ₃ }” and the display signal pattern ID 483 is “r ₂ ^t ”, the determination expression 484 of the display signal pattern is “ ^frt (z ₁ )≡{x ₁ , x ₂ , x ₄ }”.

FIG. 14 is a diagram schematically showing an example of the third definition information included in the display signal pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. The third definition information 48c shown in FIG. 14 includes a display signal pattern ID 485 and a determination expression 486 for each display signal pattern of the scatter diagram.

The display signal pattern ID 485 is information indicating a second display signal pattern, which is a combination pattern of time-series data displayed in the scatter diagram display area 615 of the driving support screen 61 among the time-series data in the specified period. A determination formula 486 is information on a determination formula for determining the second display signal pattern.

In the determination formula 486 shown in FIG. 14, "z ₂ " is the above-described second operation log time-series data. In addition, in the determination expression 486 shown in FIG. 14, "f ^rs (·)" is the pattern of the combination of the time-series data displayed in the scatter diagram display area 615 from the operation log data "z ₂ ". It is a function for calculating the signal pattern.

In the third definition information 48c shown in FIG. 14, for example, the determination formula 486 of the display signal pattern whose display signal pattern ID 485 is "r ₁ ^s " is "f ^rs (z ₂ )≡{x ₁ , x ₂ }" and the determination expression 486 of the display signal pattern whose display signal pattern ID 485 is "r ₂ ^s " is "f ^rs (z ₂ )≡{x ₁ , x ₃ }".

FIG. 15 is a diagram schematically showing an example of display period pattern definition information in the operation model stored in the operation model storage unit of the plant operation support system according to the first embodiment. The display period pattern definition information 49 shown in FIG. 15 includes a display period pattern ID 491 and a determination expression 492 for each display period pattern.

The display period pattern ID 491 is information indicating the combination pattern of time-series data displayed on the driving support screen 61, and is the same as the display period pattern ID 464 shown in FIG. A determination formula 492 is information of a determination formula for determining a display period pattern.

In the determination formula 492 shown in FIG. 15, "z ₃ " is the above-described third operation log data, and "l ₁ ", "l ₂ ", "l ₃ ", . . . indicate display periods. In addition, in the determination formula 492 shown in FIG. 15, "f ^s (·)" is selected from the operation log time-series data, which is the time-series operation data whose operation log data ID is "z ₃ ", within the specified period. This function is used to calculate the display period for trend graphs and scatter graphs.

In the display period pattern definition information 49 shown in FIG. 15, for example, the determination expression 492 of the display period pattern whose display period pattern ID 491 is "s ₁ " is "f ^s (z ₃ )≡l ₁ ", and the display period The determination formula 492 of the display period pattern whose pattern ID 491 is " _s2 " is " ^fs ( _z3 )? _l2 ".

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The operation pattern recognition unit 19 of the plant operation support system 2 recognizes the operator based on the operation log time-series data stored in the operation log data storage unit 17 and the operation model 45 stored in the operation model storage unit 18. Recognize U's operation pattern.

Specifically, the operation pattern recognition unit 19 recognizes the operation purpose based on the operation log data stored in the operation log data storage unit 17 and the operation purpose definition information 47 stored in the operation model storage unit 18. judge. Further, the operation pattern recognition unit 19 determines the display signal pattern based on the operation log data stored in the operation log data storage unit 17 and the display signal pattern definition information stored in the operation model storage unit 18. .

Further, the operation pattern recognition unit 19 determines the display period pattern based on the operation log data stored in the operation log data storage unit 17 and the display period pattern definition information 49 stored in the operation model storage unit 18. do. Then, the operation pattern recognition unit 19 recognizes the operation pattern of the operator U based on the determined operation purpose, display signal pattern, display period pattern, and the operation pattern definition information 46 stored in the operation model storage unit 18 . to recognize

16 is a diagram schematically illustrating an example of an operator's operation pattern recognized by the operation pattern recognition unit according to the first embodiment; FIG. In the example shown in FIG. 16, the operation patterns of the operator U recognized by the operation pattern recognition unit 19 are the operation pattern with the operation pattern ID " _b1 ", the operation pattern with the operation pattern ID " _b2 ", and the operation pattern with the operation pattern ID "b1". b ₆ ” and the operation pattern ID “b ₇ ” are displayed in that order. Note that the operation pattern recognition unit 19 can also recognize a plurality of operation patterns at the same time.

In the above example, the operation pattern recognition unit 19 determines the type of time-series data included in the trend graph displayed on the driving assistance screen 61, the display period of the trend graph, the display period of the scatter diagram, and the selection on the driving assistance screen 61. The operation pattern of the operator U is recognized based on the attention time and the like, but the operation pattern is not limited to this example.

For example, in addition to these, the operation pattern recognition unit 19 recognizes the operation of the operator U based on various screen operation information such as operation guidance, setting value guidance, operation summary, similar pattern search, or addition of abnormality detection. Can recognize operation patterns.

In this case, the operation pattern definition information of the operation model 45 includes, in addition to the presentation information such as the operation purpose, the display signal pattern, and the display period pattern, the display pattern of the operation guidance, the display pattern of the set value guidance, and the operation summary. It also includes presentation information such as display patterns, search patterns for similar pattern searches, and additional patterns for anomaly detection. Patterns added in this way are defined by definition information in the operation model 45 .

Returning to FIG. 1, the explanation of the configuration of the plant operation support system 2 will be continued. The evaluation unit 20 of the plant operation support system 2 evaluates the event recognition accuracy and operation pattern recognition by the event recognition unit 13 based on the event recognized by the event recognition unit 13 and the operation pattern recognized by the operation pattern recognition unit 19. The recognition accuracy of the operation pattern by the unit 19 is evaluated.

The evaluation unit 20 uses the evaluation information, for example, to evaluate the event recognition accuracy of the event recognition unit 13 and the operation pattern recognition accuracy of the operation pattern recognition unit 19 . 17 is a diagram depicting an example of evaluation information used by the evaluation unit according to the first embodiment; FIG.

The evaluation information 50 shown in FIG. 17 includes an event ID 501, an operation pattern ID 502, an event occurrence frequency 503, an operation pattern occurrence frequency 504, and an event and operation pattern occurrence frequency 505. The event ID 501 is unique identification information for each event and is the same as the event ID 411 shown in FIG. The operation pattern ID 502 is identification information unique to each operation pattern, and is the same as the operation pattern ID 461 shown in FIG.

The event occurrence frequency 503 is information indicating how often an event occurs, and the evaluation information 50 includes the occurrence frequency 503 of each event. _The event occurrence frequency 503 represents the length of time that an event that occurred in the plant ₄ was recognized. is represented.

The operation pattern occurrence frequency 504 is information indicating the frequency of occurrence of the operation pattern, and the evaluation information 50 includes the occurrence frequency 504 of each operation pattern. The operation pattern occurrence frequency 504 represents the length of time during which it is recognized that a _certain operation pattern occurs. b ₁ )”.

The occurrence frequency 505 of an event and an operation pattern is information indicating the frequency with which a certain event and a certain operation pattern occur simultaneously, and the evaluation information 50 includes the occurrence frequency 505 for each combination of an event and an operation pattern. be The frequency of simultaneous occurrence of the event with the event ID 501 of "a ₁ " and the operation pattern with the operation pattern ID 502 of "b ₁ " is expressed as "g(a ₁ , b ₁ )".

The evaluation unit 20 calculates the conditional entropy based on the occurrence frequency 505 of the event and the operation pattern, and determines the reliability of event recognition and the reliability of operation pattern recognition based on the calculated conditional entropy. . The event recognition reliability is an example of event recognition accuracy, and the operation pattern recognition reliability is an example of operation pattern recognition accuracy.

The evaluation unit 20 calculates the conditional entropy Ha regarding the event whose event ID 501 is "a _k ". Here, assuming that the number of events indicated by the evaluation information 50 is "M", "k" is a value from "1" to "M". Conditional entropy Ha is represented by the following formula (1), for example.

Then, the evaluation unit 20 determines the reliability of the event having the event ID 501 of "a _k " depending on whether the calculated conditional entropy Ha is equal to or greater than the threshold value H _TH or less than the threshold value H TH .

In Equation (1) above, P(b _l |a _k ) is a conditional probability, which is represented by Equation (2) below, for example. In the following formula (2), "M" is the number of events indicated by the evaluation information 50, and "N" is the number of operation patterns indicated by the evaluation information 50.

When the conditional entropy Ha is equal to or greater than the threshold value H _TH , even if an event having the event ID 501 of “a _k ” is determined, the evaluation unit 20 determines that the uncertainty of the operation pattern is large, and the reliability of uniquely determining the operation pattern is is low, it is determined that the recognition reliability of the event with the event ID 501 of "a _k " is low. Further, when the conditional entropy Ha is less than the threshold value H _TH , the evaluation unit 20 determines that the reliability of the recognition of the event having the event ID 501 of “a _k ” is high.

The evaluation unit 20 also calculates the conditional entropy Hb for the operation pattern whose operation pattern ID 502 is "b _l ". Assuming that the number of events indicated by the evaluation information 50 is M and the number of operation patterns indicated by the evaluation information 50 is "N", "l" is a value from "1" to "N".

Then, the evaluation unit 20 determines the reliability of the operation pattern having the operation pattern ID 502 of " _bl " depending on whether the calculated conditional entropy Hb is equal to or greater than the threshold value _HTH or less than the threshold value HTH. Conditional entropy Hb is represented by the following formula (3), for example.

In Equation (3) above, P(a _k |b _l ) is a conditional probability, which is represented by Equation (4) below, for example.

When the conditional entropy Hb is equal to or greater than the threshold H _TH , the uncertainty of the event is large even if the operation pattern whose operation pattern ID 502 is “b _l ” is determined, and the reliability for uniquely determining the event is is low, it is determined that the recognition reliability of the operation pattern with the operation pattern ID 502 of "b _l " is low. Further, when the conditional entropy Hb is less than the threshold value H _TH , the evaluation unit 20 determines that the recognition reliability of the operation pattern having the operation pattern ID 502 of “b ₁ ” is high.

The evaluation unit 20 performs the above-described calculations for all events and operation patterns, and determines events with low recognition reliability and operation patterns with low recognition reliability.

The update unit 21 determines whether or not to update at least one of the event recognition model 40 and the operation model 45 based on the evaluation result by the evaluation unit 20 . When determining to update the event recognition model 40, the updating unit 21 uses the monitoring control data stored in the monitoring control data storage unit 11 to, for example, divide events with low recognition reliability into two or more events. The event recognition model 40 is updated by dividing and creating event names and judgment formulas for two or more divided events.

Here, update processing of the event recognition model 40 and the operation model 45 by the update unit 21 when the evaluation unit 20 determines that the reliability of recognition of the event whose event ID 501 is "a ₆ " is low will be described.

FIG. 18 shows an example of logical expression information included in the event recognition model updated by the update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low. It is a figure represented typically.

When the update unit 21 determines that the event with the event ID 411 of "a ₆ " should be divided into two events, the updating unit 21 reallocates the event IDs 411 of the events with the event ID 411 of "a ₇ " and later, Split events that are 'a ₆ ' into two events. Then, the update unit 21 assigns “a ₆ ” as the event ID 411 to one of the two split events, and assigns “a ₇ ” as the event ID 411 to the other event. Further, the updating unit 21 assigns event names 412 to the two divided events, and also assigns determination expressions 413 corresponding to the details of the division to the two divided events.

In the logical formula information 41 shown in FIG. 18, the event with the event ID 411 of "a ₆ " is divided into the event with the event ID 411 of "a ₆ " and the event with the event ID 411 of "a ₇ ". is shown. Furthermore, the event whose event ID 411 is “a ₆ ” is assigned the event name 412 of “Rainy day peak support-1”, and the event whose event ID 411 is “a ₇ ” is assigned the event name 412 of “Rainy day peak Response-2” is assigned.

Further, in the logical expression information 41 shown in FIG. 18, the event whose event ID 411 is "a ₆ " is assigned "p ₃ ∧ p ₅ ∧ p ₇ ∧ p ₈ ∧ p ₁₀ " as the determination formula 413 , and the event ID 411 is “p ₄ ∧ _{p 5} ∧ _{p 7} ∧ _{p 8} ∧ p ₁₀ ” is assigned to the event “a ₇ ” as the determination expression 413 .

FIG. 19 shows an example of propositional logic definition information included in the event recognition model updated by the update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low. It is a figure which represents typically.

When the event is divided by classifying the pump flow rate more finely, the update unit 21 changes the proposition logic ID 421 of the proposition logic whose proposition logic ID 421 is "p ₄ " or later to "p ₅ '' and later propositional logic ID 421. In addition, the update unit 21 updates _the determination formula 425 of the propositional logic with the propositional logic ID 421 of "p ₃ ", the determination formula 425 of the propositional logic with the propositional logic ID 421 of "p 3 ", and the proposition with the propositional logic ID 421 of "p ₄ ". is replaced with the logical judgment expression 425.

In the propositional logic definition information 42 shown in FIG. 19, the determination expression 425 of the propositional logic of the newly replaced propositional logic ID 421 “p ₃ ” is “y ₁ ² <x ₁ ≦y ₁ ³ ”, and the newly replaced The determination formula 425 of the propositional logic of the propositional logic ID 421 “p ₄ ” obtained is “y ₁ ³ <x ₁ ”.

When updating the logical formula information 41 and the propositional logic definition information 42, the updating unit 21 divides the data to be divided by a clustering algorithm such as k-means, and creates a temporary evaluation table based on the new divided result. , and iteratively divides and evaluates until the conditional entropy is less than the threshold _HTH . When performing clustering, the updating unit 21 randomly determines an initial value and the like so that the clustering result is different each time.

In addition, the update unit 21 updates the driving support information creation models 43 and 44 when the evaluation unit 20 divides an event with low recognition reliability into two or more events. FIG. 20 shows an example of an operation support information creation model for a trend graph updated by the update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low. It is a figure represented typically.

As _shown in FIG. 20, the updating unit 21 updates the _conditional probability 433 and Recalculate the conditional probability 434 per display period. Such recalculation is performed based on the occurrence frequency of events, the display frequency of time-series data, and the display frequency of the display period. The update unit 21 updates the driving support information creation model 43 based on the recalculated conditional probabilities 433 and 434 .

FIG. 21 shows an example of an operation support information creation model for a scatter diagram updated by the update unit when the evaluation unit of the plant operation support system according to the first embodiment determines that the reliability of event recognition is low. It is a figure represented typically.

As shown in FIG. 21, the update unit 21 updates the conditional probability of each combination of display time-series data for a new event with an event ID 431 of "a ₆ " and a new event with an event ID 431 of "a ₇ ". 443 and conditional probabilities per display period 444 are recalculated. Such recalculation is performed based on the occurrence frequency of events, the display frequency of time-series data, and the display frequency of the display period. The updating unit 21 updates the driving support information creation model 44 based on the recalculated conditional probabilities 443 and 444 .

Further, the update unit 21 divides an operation pattern with a low reliability into two or more operation patterns, and creates an operation pattern name and a determination formula for the divided two or more operation patterns, thereby updating the operation model. Update 45.

Even when the evaluation unit 20 determines that the reliability of the operation pattern is low, the update unit 21 performs the same processing as when the evaluation unit 20 determines that the reliability of the event is low to determine that the reliability is low. The definition of an appropriate operation pattern is calculated by repeating clustering and evaluation of the clustering result for the obtained operation pattern data.

Further, the update unit 21 newly defines operation patterns based on the operation log data stored in the operation log data storage unit 17, and defines the first definition information 48a, the second definition information 48b, and the third definition information. 48c.

The updating unit 21 also calculates the occurrence frequency of the newly defined operation pattern, and updates the conditional probability 465 in the operation pattern definition information 46 included in the operation model 45 based on the calculated occurrence frequency of the operation pattern. .

In this way, the plant operation support system 2 updates the event recognition model 40, the operation model 45, and the operation support information creation models 43 and 44 when it is determined that the accuracy of event recognition and operation pattern recognition is insufficient. can be done. As a result, the plant operation support system 2 can display the operation support screen 61 based on the appropriate operation support information on the display unit 30 even when an unexpected event occurs.

In the above example, the updating unit 21 updates the logical formula information 41 and the propositional logic definition information 42 by dividing the event, but is not limited to such an example. Formula information 41 and propositional logic definition information 42 can also be updated. Further, the update unit 21 updates the logical formula information 41 or the propositional logic definition information 42 by changing the propositional logic included in the judgment formula 425 for judging an event or by shifting the threshold values between a plurality of propositional logics. You can also

Next, processing by the processing unit 22 of the plant operation support system 2 will be described using a flowchart. 22 is a flowchart illustrating an example of processing by a processing unit of the plant operation support system according to the first embodiment; FIG.

As shown in FIG. 22, the processing unit 22 of the plant operation support system 2 determines whether or not the monitoring control data has been acquired from the monitoring control system 1 (step S10). When the processing unit 22 determines that the monitoring control data has been acquired (step S10: Yes), the processing unit 22 stores the acquired monitoring control data in the monitoring control data storage unit 11 (step S11).

Next, the processing unit 22 recognizes events occurring in the plant 4 based on the monitoring control data stored in the monitoring control data storage unit 11 (step S12). Then, the processing unit 22 creates driving assistance information based on the event recognized in step S12 and the driving assistance information creation models 43 and 44 (step S13), and based on the created driving assistance information, the driving assistance information is displayed on the display unit 30 (step S14).

When the process of step S14 is completed, or when it is determined that the monitoring control data has not been acquired (step S10: No), the processing unit 22 determines whether or not the operation data indicating the operation of the operator U has been acquired. Determine (step S15). When determining that the operation data has been acquired (step S15: Yes), the processing unit 22 stores the acquired operation data in the operation log data storage unit 17 (step S16).

When the processing of step S16 is completed, or when it is determined that the operation data has not been acquired (step S15: No), the processing unit 22 determines whether or not the evaluation timing has come (step S17). The evaluation timing is, for example, timing that arrives at a preset cycle. When the processing unit 22 determines that the evaluation timing has come (step S17: Yes), the processing unit 22 performs evaluation update processing (step S18). Step S18 is the processing of steps S20 to S27 shown in FIG. 23, and will be described in detail later.

When the processing of step S18 is completed, or when it is determined that the evaluation timing has not come (step S17: No), the processing unit 22 determines whether or not the operation end timing has arrived (step S19). For example, when the processing unit 22 determines that the power supply (not shown) of the plant operation support system 2 is turned off or determines that an operation to end the operation is performed on the input unit 31, it is determined that it is time to end the operation. judge.

If the processing unit 22 determines that it is not the time to end the operation (step S19: No), the process proceeds to step S10. 22 ends.

FIG. 23 is a flowchart showing an example of evaluation update processing by the processing unit of the plant operation support system according to the first embodiment. As shown in FIG. 23, the processing unit 22 determines the recognition accuracy of the event recognition model 40 (step S20). The processing unit 22 also determines the recognition accuracy of the operation model 45 (step S21).

Next, the processing unit 22 determines whether the recognition accuracy of the event recognition model 40 is low (step S22). When the processing unit 22 determines that the recognition accuracy of the event recognition model 40 is low (step S22: Yes), it determines whether the recognition accuracy of the operation model 45 is low (step S23).

When the processing unit 22 determines that the recognition accuracy of the operation model 45 is low (step S23: Yes), it updates the event recognition model 40, the operation model 45, and the driving support information creation models 43 and 44 (step S24). When the processing unit 22 determines that the recognition accuracy of the operation model 45 is not low (step S23: No), the processing unit 22 updates the event recognition model 40 and the driving support information creation models 43 and 44 (step S25).

When the processing unit 22 determines that the recognition accuracy of the event recognition model 40 is not low (step S22: No), it determines whether the recognition accuracy of the operation model 45 is low (step S26). When the processing unit 22 determines that the recognition accuracy of the operation model 45 is low (step S26: Yes), it updates the operation model 45 (step S27).

The processing unit 22 completes the process of step S24, completes the process of step S25, completes the process of step S27, or determines that the recognition accuracy of the operation model 45 is not low (step S26: No), the process of FIG. 23 is terminated.

FIG. 24 is a diagram showing an example of the hardware configuration of the plant operation support system according to the first embodiment. As shown in FIG. 24, the plant operation support system 2 includes a computer having a processor 101, a memory 102, a communication device 103, a display device 104, an input device 105, and a bus .

The processor 101 , the memory 102 , the communication device 103 , the display device 104 and the input device 105 can transmit and receive information to and from each other via the bus 106 . The monitor control data storage unit 11 , the event recognition model storage unit 12 , the driving support information creation model storage unit 14 , the operation log data storage unit 17 and the operation model storage unit 18 are realized by the memory 102 . The processor 101 executes the functions of the processing unit 22 by reading and executing programs stored in the memory 102 . The processor 101 is an example of a processing circuit, for example, and includes one or more of a CPU (Central Processing Unit), a DSP (Digital Signal Processor), and a system LSI (Large Scale Integration).

The memory 102 includes one or more of RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable Read Only Memory), and EEPROM (registered trademark) (Electrically Erasable Programmable Read Only Memory). include. The memory 102 also includes a recording medium in which a computer-readable program is recorded. Such recording media include one or more of nonvolatile or volatile semiconductor memories, magnetic disks, flexible memories, optical disks, compact disks, and DVDs (Digital Versatile Discs). The processing unit 22 of the plant operation support system 2 may include integrated circuits such as ASIC (Application Specific Integrated Circuit) and FPGA (Field Programmable Gate Array).

The plant operation support system 2 may be composed of a server device, or may be composed of a client device and a server device. When the plant operation support system 2 is composed of two or more devices, each of the two or more devices has the hardware configuration shown in FIG. 24, for example. Note that communication between two or more devices is performed via the communication device 103 . Also, the plant operation support system 2 may be composed of two or more server devices. For example, the plant operation support system 2 may be composed of a processing server and a data server.

As described above, the plant operation support system 2 according to the first embodiment is a plant operation support system that supports the operation of the operator U of the plant 4 that is monitored and controlled by the supervisory control system 1. A data acquisition unit 10 , an event recognition unit 13 , a driving support information creation unit 15 , an operation pattern recognition unit 19 , an evaluation unit 20 and an update unit 21 are provided. The supervisory control data acquisition unit 10 acquires supervisory control data including time-series measurement data output by measuring equipment installed in the plant 4 . The event recognition unit 13 recognizes an event that has occurred in the plant 4 based on the monitoring control data and an event recognition model 40 that recognizes an event that has occurred in the plant from the monitoring control data. The driving support information creation unit 15 is based on the driving support information creation models 43 and 44 that determine the driving support information, which is the information for supporting the operation from the events occurring in the plant 4, and the events recognized by the event recognition unit 13. , the driving assistance information is created, and a driving assistance screen 61, which is a screen of the created driving assistance information, is displayed on the display unit 30. FIG. The operation pattern recognition unit 19 recognizes the operation pattern based on the operation model 45 for recognizing the operation pattern of the operator U from the operation log data including the operation history of the driving support screen 61 by the operator U and the operation log data. . The evaluation unit 20 evaluates the event recognition accuracy of the event recognition unit 13 and the operation pattern accuracy of the operation pattern recognition unit 19 based on the event recognized by the event recognition unit 13 and the operation pattern recognized by the operation pattern recognition unit 19. Evaluate the recognition accuracy. The update unit 21 updates the event recognition model 40 based on the evaluation result by the evaluation unit 20. FIG. As a result, the plant operation support system 2 can provide operation support information for supporting the operation of the plant regarding events that were not defined when the system was installed.

In addition, the update unit 21 updates the operation model 45 based on the evaluation result by the evaluation unit 20. As a result, the plant operation support system 2 can more accurately provide operation support information for assisting plant operation regarding events that are not defined when the system is installed.

In addition, the update unit 21 updates the driving support information creation models 43 and 44 based on the evaluation result by the evaluation unit 20. As a result, the plant operation support system 2 can more accurately provide operation support information for assisting plant operation regarding events that are not defined when the system is installed.

Embodiment 2.
In the processing system according to the second embodiment, when an operator designates an event name using an input/output device, the event name designated by the operator can be displayed when the same event is recognized at subsequent timings. It is different from the processing system 100 according to the first embodiment in this respect. In the following, constituent elements having functions similar to those of the first embodiment are given the same reference numerals, and descriptions thereof are omitted, and differences from the processing system 100 of the first embodiment are mainly described.

FIG. 25 is a diagram showing an example of the configuration of the processing system according to the second embodiment. As shown in FIG. 25, the processing system 100A according to the second embodiment differs from the processing system 100 in that it includes a plant operation support system 2A instead of the plant operation support system 2. FIG. The plant operation support system 2A differs from the plant operation support system 2 in that it includes a processing unit 22A in place of the processing unit 22 and further includes an event name edited information storage unit 23 .

The processing unit 22A differs from the processing unit 22 in that it includes a driving support information creating unit 15A instead of the driving support information creating unit 15 and further includes an event name editing unit 24. The driving support information creation unit 15A causes the display unit 30 to display a driving support screen 61A, which allows the event name to be designated by the operation of the input unit 31 by the operator U, instead of the driving support screen 61. It differs from the information creation unit 15 .

FIG. 26 is a diagram schematically showing an example of an operation support screen created by the operation support information creation unit of the plant operation support system according to the second embodiment and displayed on the display unit. As shown in FIG. 26, a driving assistance screen 61A displayed on the display unit 30 by the driving assistance information creation unit 15A includes a message display area 612A instead of the message display area 612. Therefore, the driving assistance screen 61A shown in FIG. It differs from screen 61 .

In the message display area 612A, an event name display frame 617 displaying the event name is displayed in addition to the event occurrence time and the event ID of the event. In the example shown in FIG. 26, the event name of the event whose event ID is “a ₇ ” is displayed in the event name display frame 617 as “rainy day peak response-2”. The operator U can edit the event name displayed in the event name display frame 617 by operating the input unit 31 .

FIG. 27 is a diagram schematically showing another example of the operation assistance screen created by the operation assistance information creating unit of the plant operation assistance system according to the second embodiment and displayed on the display unit. In the message display area 612A shown in FIG. 27, the event name of the event whose event ID is " _a7 " in the event name display frame 617 is changed to "typhoon response" by the operation of the input unit 31 by the operator U. ing. The event name specified by the operator U is stored in the event name edited information storage unit 23 shown in FIG.

FIG. 28 is a diagram schematically showing an example of event name editing information stored in the event name editing information storage unit of the plant operation support system according to the second embodiment. The event name edit information 51 shown in FIG. 28 includes an event ID 511, an event name 512, and an event name 513 specified by the operator U. FIG.

The event ID 511 is unique identification information for each event and is the same as the event ID 411 included in the logical expression information 41. The event name 512 is information indicating the name of the event and is the same as the event name 412 included in the logical expression information 41 . The event name 513 is information indicating the event name specified by the operator U. FIG.

The event name editing unit 24 shown in FIG. 25 changes the event names 432 and 442 of events having the event name 513 specified by the operator U in the event name editing information 51 in the driving support information creation models 43 and 44 . For example, when the driving support information creation models 43 and 44 are in the states shown in FIGS. 20 and 21, the operator U designates the event name of the event whose event ID 431 and 441 is " _a7 " as "typhoon response". In this case, the event name editing unit 24 changes the event name of the event whose event ID 431, 441 is “a ₇ ” in the driving support information creation models 43, 44 from “rainy day peak response-2” to “typhoon response”. change.

As a result, when the operator U designates an event name using the input/output device 3, the driving support information creation unit 15A displays the event name designated by the operator U when the same event is recognized at subsequent timings. can do. Therefore, the plant operation support system 2A recognizes a new event and names it according to a predetermined rule, even if the event name differs from the event name preferred by the operator U. An event name that the operator U likes can be displayed on the display unit 30 .

A hardware configuration example of the plant operation support system 2A according to the second embodiment is the same as the hardware configuration of the plant operation support system 2 shown in FIG. The processor 101 can execute the functions of the processing unit 22A by reading and executing the programs stored in the memory 102 .

As described above, the plant operation support system 2A according to the second embodiment includes the event name editing section 24 that changes the event name of the event recognized by the event recognition section 13 to the event name specified by the operator U. Thereby, the plant operation support system 2A can cause the display unit 30 to display the event name that the operator U prefers.

Embodiment 3.
The processing system according to the third embodiment differs from the processing system 100 according to the first embodiment in that it is possible to change the evaluation criteria regarding the detail of event recognition. In the following, constituent elements having functions similar to those of the first embodiment are given the same reference numerals, and descriptions thereof are omitted, and differences from the processing system 100 of the first embodiment are mainly described.

FIG. 29 is a diagram showing an example of the configuration of a processing system according to the third embodiment. As shown in FIG. 29, the processing system 100B according to the third embodiment differs from the processing system 100 in that it includes a plant operation support system 2B instead of the plant operation support system 2. FIG. The plant operation support system 2B is different from the plant operation support system 2 in that the processing unit 22 is replaced with a processing unit 22B.

The processing section 22B differs from the processing section 22 in that it further includes an evaluation criteria changing section 25 . When the operator U operates the input unit 31 and receives a request to change the threshold value _HTH , which is the evaluation criterion for the fineness of event recognition by the evaluation unit 20, the evaluation criterion change unit 25 changes the value included in the change request. The threshold value H _TH of the evaluation unit 20 is changed to .

When the threshold value _HTH is fixed, the determined event may be classified too coarsely or too finely compared to the event recognition desired by the operator U. However, the plant operation support system 2B , the operator U can change the threshold value _HTH, which is an evaluation criterion regarding the detail of event recognition. Therefore, the plant operation support system 2B can classify the events with fineness according to the event recognition desired by the operator U.

Note that the change request for the threshold H _TH is not limited to including the threshold H _TH itself, and may include information indicating a method for changing the threshold H _TH . The change request for the threshold H _TH may include information that can specify whether to increase the threshold H _TH or decrease the threshold H _TH .

A hardware configuration example of the plant operation support system 2B according to the third embodiment is the same as the hardware configuration of the plant operation support system 2 shown in FIG. The processor 101 can execute the functions of the processing unit 22B by reading and executing the programs stored in the memory 102 .

The plant operation support system 2B further includes an event name editing information storage unit 23 and an event name editing unit 24, similarly to the plant operation support system 2A. may be provided.

As described above, the plant operation support system 2B according to the third embodiment includes the evaluation criterion change unit 25 that changes the evaluation criterion used for evaluating the recognition accuracy in the evaluation unit 20 based on a request from the operator U. As a result, the plant operation support system 2B can classify the events with fineness according to the event recognition desired by the operator U.

Embodiment 4.
The processing system according to the fourth embodiment recognizes an operation pattern by using, as operation log data, operation data excluding operation data for a set period of operation data of the driving support screen by the operator. 1 differs from the processing system 100 according to 1. In the following, constituent elements having functions similar to those of the first embodiment are given the same reference numerals, and descriptions thereof are omitted, and differences from the processing system 100 of the first embodiment are mainly described.

FIG. 30 is a diagram showing an example of the configuration of a processing system according to the fourth embodiment. As shown in FIG. 30, a processing system 100C according to the fourth embodiment differs from the processing system 100 in that a plant operation support system 2C is provided instead of the plant operation support system 2. FIG. The plant operation support system 2C differs from the plant operation support system 2 in that it includes a processing unit 22C instead of the processing unit 22. FIG.

The processing unit 22C differs from the processing unit 22 in that it includes a driving support information creation unit 15C and an operation data acquisition unit 16C instead of the driving support information creation unit 15 and the operation data acquisition unit 16. Instead of the driving support screen 61, the driving support information creating unit 15C provides a setting for not storing the operation log data in the operation log data storage unit 17 by the operation of the input unit 31 by the operator U. It is different from the driving support information creating unit 15 in that it is displayed on the display unit 30 .

FIG. 31 is a diagram schematically showing an example of an operation support screen created by the operation support information creation unit of the plant operation support system according to the fourth embodiment and displayed on the display unit. As shown in FIG. 31, a driving assistance screen 61C displayed on the display unit 30 by the driving assistance information creating unit 15C differs from the driving assistance screen 61 shown in FIG. 8 in that it further has a learning mode selection area 618.

The learning mode selection area 618 is an area for selecting whether to turn on or off the learning mode, and the operator U operates the box shown in the learning mode selection area 618 by operating the input unit 31. You can choose to turn learning mode on or off.

The operation data acquisition unit 16C collects data indicating operations of the input unit 31 by the operator U during the period from when the operator U selects to turn on the learning mode until when the operator U selects to turn off the learning mode. is acquired from the input/output device 3 and the acquired operation data is stored in the operation log data storage unit 17 .

Further, the operation data acquisition unit 16C prevents the operator U from operating the input unit 31 until the operator U selects to turn off the learning mode after the operator U selects to turn off the learning mode. The operation data, which is the data shown, is not acquired from the input/output device 3 .

In this way, the operation data acquisition unit 16C stores the operation data excluding the operation data for the period set by the operator U's operation as the operation log data. Therefore, the operation pattern recognition unit 19 can recognize the operation pattern based on the operation log data excluding the operation data for the period set by the operator U's operation.

Therefore, for example, when the operator U does not want to use the operation log data for a certain period of time to determine the accuracy of the event recognition, the operator U does not use the operation log data for the certain period of time and asks the plant operation support system 2C for event recognition. Accuracy can be determined. A case in which it is not desired to use operation log data for a certain period of time to determine the accuracy of event recognition is, for example, to ask a new operator U who does not understand the plant 4 to understand the plant 4. This is a case where the plant operation support system 2C is temporarily operated for educational purpose. This is because, in this case, an operation pattern different from an efficient operation pattern performed by a skilled operator U may occur.

A hardware configuration example of the plant operation support system 2C according to the fourth embodiment is the same as the hardware configuration of the plant operation support system 2 shown in FIG. The processor 101 can execute the functions of the processing unit 22C by reading and executing programs stored in the memory 102 .

The plant operation support system 2C further includes an event name editing information storage unit 23 and an event name editing unit 24, similarly to the plant operation support system 2A. may be provided. Also, the plant operation support system 2C may be configured to include the evaluation criteria changing unit 25, similarly to the plant operation support system 2B.

As described above, in the plant operation support system 2C according to the fourth embodiment, the operation history of the operation support screen 61C by the operator U, excluding the operation history for the set period, is used as the operation log data. An operation data acquisition unit 16C that stores data in the data storage unit 17 is provided. The operation pattern recognition unit 19 recognizes operation patterns based on the operation log data stored in the operation log data storage unit 17 and the operation model 45 . As a result, the plant operation support system 2C can recognize the operation pattern using the operation log data excluding the operation history of the period that the operator U does not want to use for determining the accuracy of event recognition.

Embodiment 5.
The processing system according to the fifth embodiment calculates the feature amount of the time-series measurement data specified by the operator, performs classification reflecting the calculated feature amount, and updates the event recognition model. It differs from the processing system 100 according to the first form. In the following, constituent elements having functions similar to those of the first embodiment are given the same reference numerals, and descriptions thereof are omitted, and differences from the processing system 100 of the first embodiment are mainly described.

FIG. 32 is a diagram showing an example of the configuration of a processing system according to the fifth embodiment. As shown in FIG. 32, the processing system 100D according to the fifth embodiment differs from the processing system 100 in that it includes a plant operation support system 2D instead of the plant operation support system 2. FIG. The plant operation support system 2D differs from the plant operation support system 2 in that it includes a processing unit 22D instead of the processing unit 22, and further includes a selected feature amount information storage unit 26 and a feature amount template storage unit 27.

The processing unit 22D differs from the processing unit 22 in that it includes a driving support information creation unit 15D and an update unit 21D instead of the driving support information creation unit 15 and the update unit 21, and further includes a feature amount calculation unit 28. . The driving support information creation unit 15D replaces the driving support screen 61 with a driving support screen on which the operation log time-series data and its feature amount can be specified by the operation of the input unit 31 by the operator U. It is different from the driving support information creation unit 15 in that it is displayed on the .

FIG. 33 is a diagram schematically showing an example of an operation support screen created by the operation support information creation unit of the plant operation support system according to the fifth embodiment and displayed on the display unit. As shown in FIG. 33, a driving assistance screen 61D displayed on the display unit 30 by the driving assistance information creating unit 15D differs from the driving assistance screen 61 shown in FIG.

The feature amount selection area 619 is an area for selecting the feature amount of the time series data, which is the time series measurement data used for event classification processing by the update unit 21D. and a selection box 619b for selecting a feature amount.

By operating the input unit 31, the operator U can select the time-series data in the selection box 619a and select the feature amount in the selection box 619b. In the selection box 619a, the time-series data selected in the time-series data selection area 611 is displayed as a selection candidate, and the operator U selects from the time-series data selected in the time-series data selection area 611. Select time-series data for selecting features.

Information on the feature amount of the time-series data selected by the operator U is stored in the selected feature amount information storage unit 26 . 34 is a diagram schematically illustrating an example of selected feature amount information stored in a selected feature amount information storage unit of the driving support system according to the fifth embodiment; FIG. The selected feature amount information 52 shown in FIG. 34 includes a target time-series data ID 521, a target time-series data name 522, a unit 523, and a selected feature amount 524 for each target time-series data.

The target time-series data ID 521 is identification information unique to each target time-series data, which is time-series measurement data to be selected as a feature amount. The target time-series data name 522 is information indicating the name of the target time-series data. The unit 523 is information indicating the unit of the target time-series data. A selected feature quantity 524 is a feature quantity selected by the operator U.

In the example shown in FIG. 34, the feature amount selected by the operator U as the feature amount of the target time-series data whose target time-series data ID 521 is "x ₅ " is "integration [1 day]". “Integration [1 day]” is a feature quantity obtained by integration on an hourly basis. When different feature amounts are selected for the same target time series data, the selected feature amount 1, the selected feature amount 2, . . . are added in order to the selected feature amount information 52.

The feature amount calculation unit 28 acquires the time-series data selected by the operator U from the monitoring control data storage unit 11, and calculates the feature amount of the time-series data selected by the operator U from the selected feature amount information storage unit 26. Get information. The feature amount calculation unit 28 calculates the feature amount selected by the operator U based on the acquired time-series data and information on the feature amount. The driving support information creation unit 15D uses the feature amount calculated by the feature amount calculation unit 28 to display the time-series data selected by the operator U in the period selection area 613, the trend graph display area 614, and the scatter diagram display area 615. The display unit 30 is caused to display the driving support screen 61D shown.

The feature amount calculation unit 28 uses the feature amount template stored in the feature amount template storage unit 27 to calculate the feature amount of the time-series data. 35 is a schematic diagram of an example of a feature template stored in a feature template storage unit of the driving support system according to the fifth embodiment; FIG.

The feature quantity template 53 shown in FIG. 35 includes a feature quantity name 531 and a calculation formula 532 for each feature quantity. The feature name 531 is information indicating the name of the feature. The calculation formula 532 is information on the calculation formula for calculating the feature quantity. In the example shown in FIG. 35, the calculation formula 532 is a calculation formula when the measuring equipment provided in the plant 4 outputs the measurement data in units of one minute. It is information such as a difference calculation formula, an hourly difference calculation formula, an integral calculation formula for 10 minutes, an hourly integral calculation formula, or a daily integral calculation formula.

When the event recognition model 40 is updated, the update unit 21D updates the time-series data, which is the time-series measurement data stored in the monitoring control data storage unit 11, when the feature amount is calculated by the feature amount calculation unit 28. For the series data, the feature amount calculated by the feature amount calculation unit 28 is used in place of the time-series measurement data stored in the monitoring control data storage unit 11, and the classification criteria are recalculated to create an event recognition model. 40 updates.

The operator U may recognize the event using the feature amount of the time-series data such as the slope, maximum value, or minimum value of the graph when the time-series data is displayed on the driving support screen 61D. In such a case, recalculation of the classification criteria using only the time-series data stored in the supervisory control data storage unit 11 may not result in classification that takes into consideration the feature amount that the operator U takes into consideration. be.

In the plant operation support system 2D according to the fifth embodiment, the operator U is allowed to select a feature amount as a way of looking at the time-series data, and the time-series data processed into the selected feature amount is used to recalculate the classification criteria. Therefore, it is possible to perform classification that reflects the feature amount that the operator U takes into consideration when checking the time-series data.

A hardware configuration example of the plant operation support system 2D according to the fifth embodiment is the same as the hardware configuration of the plant operation support system 2 shown in FIG. The processor 101 can execute the functions of the processing unit 22D by reading out and executing programs stored in the memory 102 .

The plant operation support system 2D further includes an event name editing information storage unit 23 and an event name editing unit 24, similarly to the plant operation support system 2A. may be provided. Also, the plant operation support system 2D may be configured to include the evaluation criteria changing unit 25, similarly to the plant operation support system 2B. Further, the plant operation support system 2D may be configured to include an operation data acquisition unit 16C instead of the operation data acquisition unit 16, like the plant operation support system 2C.

As described above, the plant operation support system 2D according to the fifth embodiment includes the feature amount template storage unit 27, the selected feature amount information storage unit 26, and the feature amount calculation unit 28. The feature amount template storage unit 27 stores a feature amount template 53 including a plurality of calculation formulas for calculating different feature amounts from time-series measurement data. The selected feature amount information storage unit 26 stores information on the time-series measurement data selected by the operator U from among the time-series measurement data included in the monitoring control data and the time-series measurement data selected by the operator U. and the information of the feature amount selected by the operator U as the feature amount of . The feature amount calculation unit 28 calculates the feature amount selected by the operator U in the time-series measurement data selected by the operator U using the calculation formula included in the feature amount template 53, and updates the calculated result. Output to the section 21D. As a result, the plant operation support system 2D can perform classification that reflects the feature amount that the operator U takes into consideration when checking the time-series data.

The configurations shown in the above embodiments are only examples, and can be combined with other known techniques, or can be combined with other embodiments, without departing from the scope of the invention. It is also possible to omit or change part of the configuration.

1 Supervisory control system, 2, 2A, 2B, 2C, 2D Plant operation support system, 3 Input/output device, 4 Plant, 10 Supervisory control data acquisition part, 11 Supervisory control data storage part, 12 Event recognition model storage part, 13 Event Recognition unit 14 Driving support information creation model storage unit 15, 15A, 15C, 15D Driving support information creation unit 16, 16C Operation data acquisition unit 17 Operation log data storage unit 18 Operation model storage unit 19 Operation pattern recognition Section 20 Evaluation Section 21, 21D Update Section 22, 22A, 22B, 22C, 22D Processing Section 23 Event Name Editing Information Storage Section 24 Event Name Editing Section 25 Evaluation Criteria Changing Section 26 Selected Feature Amount Information Storage Section, 27 Feature template storage section, 28 Feature calculation section, 30 Display section, 31 Input section, 40 Event recognition model, 41 Logical formula information, 42 Proposition logic definition information, 43, 44 Driving support information creation model, 45 Operation model, 46 operation pattern definition information, 47 operation purpose definition information, 48a first definition information, 48b second definition information, 48c third definition information, 49 display period pattern definition information, 50 evaluation information, 51 event name editing information, 52 Selected feature amount information, 53 Feature amount template, 58 Operation log data definition information, 60, 61, 61A, 61C, 61D Operation support screen, 100, 100A, 100B, 100C, 100D Processing system.

Claims (9)

1. A plant operation support system that supports operation by an operator of a plant that is monitored and controlled by a supervisory control system,
   A supervisory control data acquisition unit that acquires supervisory control data including time-series measurement data output by a measuring device installed in the plant;
   an event recognition unit that recognizes an event that has occurred in the plant based on an event recognition model that recognizes an event that has occurred in the plant from the monitoring control data and the monitoring control data;
   creating the driving support information based on the event recognized by the event recognition unit and a driving support information creation model for determining the driving support information, which is the information for supporting the operation from the event occurring in the plant; a driving support information creation unit that displays the created driving support information screen on a display unit;
   an operation pattern recognition unit that recognizes the operation pattern based on the operation log data and an operation model for recognizing the operation pattern of the operator from operation log data including the operation history of the screen by the operator;
   Accuracy of event recognition by the event recognition unit and recognition of the operation pattern by the operation pattern recognition unit based on the event recognized by the event recognition unit and the operation pattern recognized by the operation pattern recognition unit an evaluation unit for evaluating accuracy;
   and an update unit that updates the event recognition model based on the evaluation result of the evaluation unit.
2. The updating unit
   The plant operation support system according to claim 1, wherein the operation model is updated based on an evaluation result by the evaluation unit.
3. The updating unit
   The plant operation support system according to claim 1 or 2, wherein the operation support information creation model is updated based on the evaluation result by the evaluation unit.
4. The plant according to any one of claims 1 to 3, further comprising an event name editing unit that changes an event name of the event recognized by the event recognition unit to an event name designated by the operator. driving assistance system.
5. 5. The plant according to any one of claims 1 to 4, further comprising an evaluation criterion changing unit that changes the evaluation criterion used for evaluating the recognition accuracy in the evaluation unit based on a request from the operator. driving assistance system.
6. an operation data acquisition unit configured to store, as the operation log data, in an operation log data storage unit an operation history excluding an operation history for a set period from among operation histories of the screen by the operator;
   The operation pattern recognition unit
   The plant operation support according to any one of claims 1 to 5, wherein the operation pattern is recognized based on the operation log data stored in the operation log data storage unit and the operation model. system.
7. a feature amount template storage unit that stores feature amount templates including a plurality of calculation formulas for calculating different feature amounts from time-series measurement data;
   Information of the time-series measurement data selected by the operator from among the time-series measurement data included in the monitoring control data, and the operator as a feature amount of the time-series measurement data selected by the operator A selected feature amount information storage unit that stores information on the feature amount selected by
   The feature amount selected by the operator in the time-series measurement data selected by the operator is calculated using a calculation formula included in the feature amount template, and the calculated result is output to the updating unit. The plant operation support system according to any one of claims 1 to 6, further comprising: a quantity calculation unit.
8. A plant operation support method for supporting operation by an operator of a plant monitored and controlled by a supervisory control system,
   a supervisory control data acquisition step of acquiring supervisory control data including time-series measurement data output by a measuring device installed in the plant;
   an event recognition step of recognizing an event that has occurred in the plant based on an event recognition model that recognizes an event that has occurred in the plant from the supervisory control data and the supervisory control data;
   creating the driving support information based on the event recognized by the event recognition step and the driving support information creation model for determining the driving support information, which is the information for supporting the operation from the event occurring in the plant; A driving support information creating step of displaying the created driving support information screen on a display unit;
   an operation pattern recognition step of recognizing the operation pattern based on the operation log data and an operation model for recognizing the operation pattern of the operator from operation log data including the operation history of the screen by the operator;
   Accuracy of event recognition by the event recognition step and recognition of the operation pattern by the operation pattern recognition step based on the event recognized by the event recognition step and the operation pattern recognized by the operation pattern recognition step an evaluation step for evaluating accuracy;
   and an update step of updating the event recognition model based on the evaluation result of the evaluation step.
9. a supervisory control data acquisition step of acquiring supervisory control data including time-series measurement data output by a measuring device installed in a plant that is monitored and controlled by a supervisory control system;
   an event recognition step of recognizing an event that has occurred in the plant based on an event recognition model that recognizes an event that has occurred in the plant from the supervisory control data and the supervisory control data;
   Based on the event recognized by the event recognition step and an operation support information creation model for determining operation support information, which is information for supporting operation by an operator of the plant, from the event occurring in the plant, the operation support a driving support information creating step of creating information and displaying a screen of the created driving support information on a display unit;
   an operation pattern recognition step of recognizing the operation pattern based on the operation log data and an operation model for recognizing the operation pattern of the operator from operation log data including the operation history of the screen by the operator;
   Accuracy of event recognition by the event recognition step and recognition of the operation pattern by the operation pattern recognition step based on the event recognized by the event recognition step and the operation pattern recognized by the operation pattern recognition step an evaluation step for evaluating accuracy;
   and an update step of updating the event recognition model based on the evaluation result of the evaluation step.

Images