Operation monitoring support system and operation monitoring support method

Abstract

To provide an operation monitoring support system capable of reducing a burden on an operator.SOLUTION: An operation monitoring support system includes a calculation unit and a storage unit storing a learning model. The arithmetic unit includes an event specifying unit that specifies an occurred event name using a learning model on the basis of input time information, alarm information, and process value information, an event detailed information management unit that associates the occurred event name output by the learning model with information related to the occurred event name, and an alarm information display unit that outputs screen information for collectively displaying the occurred event name and the information associated with the occurred event name.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an operation monitoring support system and an operation monitoring support method for facilities such as nuclear power plants. [Background technology]

[0002] During plant operation, such as at a nuclear power plant, operators on shift monitor various instruments. If an abnormality or transient event occurs in operating equipment, an alarm is automatically issued. When an alarm is issued, operators must perform numerous tasks, such as checking various parameters, confirming status, reporting, and taking appropriate action. For this reason, a nuclear power plant alarm monitoring support system has been proposed to reduce the operator's workload (see, for example, Patent Document 1). This system supports operators in predicting and determining countermeasures for alarm issuance and subsequent events after the alarm is issued. For example, the system calculates past and future time series trends of operating parameters related to the alarm and displays a list of various parameters, including the results of the time series trend calculation, and the progress of subsequent events. By displaying the results of the time series trend calculation, operators can interpret changes in operating parameters related to the alarm. As a result, the nuclear power plant alarm monitoring support system can support operators in identifying the cause of the alarm and predicting the progress of events that are expected to occur after the alarm is issued. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2017-15592 Summary of the Invention [Problem to be solved by the invention]

[0004] However, while the changes in the operating parameters are displayed as the results of the time-series trend calculations of the operating parameters mentioned above, identifying the cause of an event requires the operator to make a comprehensive judgment based on the alarms that have been issued and the fluctuations in the parameters. Therefore, the confirmation of the alarm and the response operations related to identifying the cause of an event depend on the operator's ability. In addition, when an alarm is issued, the operator refers to the relevant manuals and safety regulations to determine the appropriate countermeasures for the alarm and the event. In this way, when monitoring the operation of a plant, etc., the amount of information that is generated when an alarm is issued places a heavy burden on the operator. In addition, the initial response and the determination of the cause must be based on knowledge, and these decisions depend on the operator's ability.

[0005] In order to solve the above-mentioned problems, the present invention provides an operation monitoring support system and an operation monitoring support method that can reduce the burden on operators.

[0006] The above and other objects of the present invention and novel features of the present invention will become apparent from the description of this specification and the accompanying drawings. [Means for solving the problem]

[0007] The operation monitoring support system of the present invention includes a calculation unit that outputs plant control data from input plant operation data, and a storage unit that stores a learning model trained with a data set including pre-prepared time information, alarm information, process values, and names of occurring events. The calculation unit includes an event identification unit that uses the learning model to identify the name of the occurring event based on newly acquired time information, alarm information, and process value information, and a detailed event information management unit that links the occurring event name output by the learning model with information related to the occurring event name. The calculation unit also includes an alarm information display unit that outputs screen information that collectively displays the occurring event name and information linked to the occurring event name.

[0008] Furthermore, the operation monitoring support method of the present invention uses a learning model trained with a data set including pre-prepared time information, alarm information, process values, and names of occurred events. The operation monitoring support method uses the learning model to identify names of occurred events based on information about time information, alarm information, and process values ​​newly acquired from plant operation data. The operation monitoring support method then links the names of occurred events output by the learning model with information related to the names of occurred events, and outputs screen information that collectively displays the names of occurred events and the information linked to the names of occurred events. [Effects of the Invention]

[0009] According to the present invention, it is possible to provide an operation monitoring support system and an operation monitoring support method that can reduce the burden on operators.

[0010] Problems, configurations, and effects other than those described above will become clear from the following description of the embodiments. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 is a diagram illustrating a configuration of a driving monitoring support system. [Figure 2] FIG. 2 is a diagram showing an example of a history of operation and monitoring by an operator, which is stored in an operation and monitoring history database. [Figure 3] FIG. 10 is a diagram showing an example of each piece of information linked to the "occurring event name" stored in the event detail information database 9. [Figure 4] FIG. 10 is a diagram illustrating an example of a dataset stored in a dataset database. [Figure 5] 1 is a flowchart of a driving monitoring support method. DETAILED DESCRIPTION OF THE INVENTION

[0012] An example of a driving monitoring support system and a driving monitoring support method according to an embodiment of the present invention will be described below with reference to the drawings. Note that the present invention is not limited to the following example. In each of the drawings described below, common components are given the same reference numerals. Furthermore, in the drawings used in this specification, identical or corresponding components are given the same reference numerals, and repeated explanations of these components may be omitted. The explanation will be given in the following order. 1. Embodiment of driving monitoring support system 2. Embodiment of operation monitoring support method

[0013] 1. Embodiment of the driving monitoring support system Specific embodiments of the operation monitoring support system will be described below. In the following description, a case where the operation monitoring support system is applied to a nuclear power plant will be described as a specific example of the operation monitoring support system. Note that the application of the operation monitoring support system is not limited to nuclear power plants. The operation monitoring support system can be applied to any facility, without any particular limitations, as long as it requires work from alarm issuance to event identification.

[0014] Fig. 1 is a conceptual diagram showing the configuration of an operation monitoring support system according to this embodiment. The nuclear power plant operation monitoring support system 101 shown in Fig. 1 includes an input unit 102, a calculation unit 103, an output unit 104, and a storage unit 125. The operation monitoring support system 101 is connected to plant operation data 121 and plant control data 122.

[0015] The input unit 102 receives all input information input to the operation monitoring support system 101 and inputs it to the calculation unit 103. The input unit 102 inputs information input to the operation monitoring support system 101 from external input devices such as a mouse or keyboard, and plant operation data 121 to the calculation unit 103. The plant operation data 121 includes process values ​​and alarm information indicating the plant status, and device information of devices connected to the operation monitoring support system 101.

[0016] The output unit 104 outputs information on the processing results by the calculation unit 103 to the plant control data 122. The plant control data 122 is equipment operation information for outputting plant control instructions by operators. The output unit 104 outputs various information processed by the calculation unit 103 to the outside. The input unit 102 and the output unit 104 are configured with a communication interface or the like. The communication interface is configured with, for example, a network interface card (NIC) or a modem, and establishes a connection with a communication partner device via a network such as a LAN, and transmits and receives various data. Each arithmetic device is connected to other arithmetic devices via the communication interface, and transmits and receives information.

[0017] The calculation unit 103 is connected to the input unit 102 and the output unit 104. The calculation unit 103 has the functions of a plant state monitoring unit 111, an operation monitoring history management unit 112, an alarm information display unit 113, an event identification unit 114, an event detail information management unit 115, an equipment operation output instruction unit 116, a function screen development unit 117, a system management unit 118, an identification accuracy evaluation unit 119, and a data set management unit 120. The calculation unit 103 performs calculation processing in each of the above functions based on the information acquired from the input unit 102, and outputs the calculation results to the output unit 104. Furthermore, the calculation unit 103 acquires various pieces of information used for calculation in each of the above functions from each database stored in the storage unit 125.

[0018] The calculation unit 103 is configured by a known calculation device, etc. The calculation device is configured by, for example, a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The CPU is an example of a calculation unit related to the driving monitoring support system 101. The CPU centrally controls the operation of each unit of the driving monitoring support system 101. The CPU reads out program code of software related to various processes of the driving monitoring support system 101 stored in a ROM (an example of a recording medium) and loads it into the RAM. The CPU then controls the driving monitoring support system 101 in accordance with the loaded program. Note that the driving monitoring support system 101 may include another calculation device, such as an MPU (Micro Processing Unit), as a calculation unit instead of a CPU.

[0019] The storage unit 125 is configured with a known storage device such as a hard disk drive (HDD), a semiconductor memory, etc. The storage unit 125 stores various information used by the calculation unit 103 for calculations and the calculation results calculated by the calculation unit 103. The storage unit 125 includes a plant state database 105, an operation monitoring history database 106, a screen information database 107, a system information database 108, an event detail information database 109, and a data set database 110 as databases for storing the above-mentioned various information and calculation results. The storage unit 125 also stores a learning model 123 for identifying an event that occurs when an alarm is issued.

[0020] [Calculation unit and memory unit] Next, the functional configuration of the calculation unit 103 and the configuration of the database in the storage unit 125 will be described. The plant state monitoring unit 111 monitors information acquired from the plant operation data 121 via the input unit 102. Then, the plant state monitoring unit 111 stores the information acquired from the plant operation data 121 in the plant state database 105 of the storage unit 125. The information acquired from the plant operation data 121 and stored in the plant state database 105 is, for example, process values ​​(operation parameters, plant operation data) that quantitatively express the operating state of the equipment and facilities of the nuclear power plant. Specific process values ​​include, for example, the emergency stop signal state (on / off), reactor pressure, reactor pressure change rate, reactor containment vessel pressure, reactor power, reactor water level, reactor water level change rate, main steam pressure, main steam temperature, main steam isolation valve position, main steam isolation valve state (open / closed), main steam flow rate, and control rod insertion depth.

[0021] The operation monitoring history management unit 112 acquires the history of operation monitoring by the operator in the operation monitoring support system 101 and stores it in the operation monitoring history database 106. An example of the history of operation monitoring by the operator stored in the operation monitoring history database 106 is shown in Fig. 2. As shown in Fig. 2, the operation monitoring history includes the date and time when the operation monitoring occurred, the input point of the monitored object, the message content, the state of the input point, etc.

[0022] The alarm information display unit 113 outputs the alarm information to the output unit 104. The alarm information display unit 113 displays the alarm information on an output device such as an external display device via the output unit 104, and notifies the operator of the alarm information. The alarm information includes an alarm name that is a specific cause of the event occurrence. In addition to the alarm name, the alarm information may also include each process value linked to the alarm name, the name of the event that caused the alarm to be issued (the name of the occurring event), and the like.

[0023] The event identifying unit 114 acquires newly input "time information, alarm information, process value" from the plant operation data 121 when an alarm is issued. The event identifying unit 114 may also acquire new "time information, alarm information, process value" at the time of alarm issuance from the plant state database 105. Then, the event identifying unit 114 uses the learning model 123 based on the newly acquired input values ​​of "time information, alarm information, process value" to identify the occurring event that caused the alarm to be issued.

[0024] The event detailed information management unit 115 associates information such as alarm information, time information, process values, event severity, similar event names, related system screens, countermeasure methods, and countermeasure method effectiveness with each "occurring event name" of the occurring event identified by the event identification unit 114. The event detailed information management unit 115 then stores the "occurring event name" with each piece of associated information in the event detailed information database 109. Note that the association of the "occurring event name" with the countermeasure method and the effectiveness of the countermeasure method reflects the results of the operator's countermeasures. In other words, the operator determines whether the occurring event identified by the event identification unit 114 is correct as an event that caused the issuance of an alarm. If the occurring event identified by the event identification unit 114 is incorrect, the operator changes the "occurring event name" stored in the event detailed information database 109 to the correct event name.

[0025] An example of each piece of information linked to the "occurring event name" stored in the event detail information database 109 is shown in Fig. 3. Fig. 3 is, for example, a list of each piece of information linked to each "event name" by the event detail information management unit 115, and is an example of an output screen displayed on an external display of the operation monitoring support system 101. For example, the output screen shown in Fig. 3 is generated by the function screen deployment unit 117 based on the "occurring event name" generated by the event detail information management unit 115 and each piece of information linked to the "occurring event name". Then, the alarm information display unit 113 outputs the generated screen information to an external display device or the like via the output unit 104.

[0026] 3, an event details screen 301 collectively displays an event name 302 that displays the "name of the occurring event," and detailed information that is pre-associated with the event name 302, such as a related alarm list 303, an event severity 304, a similar event name 305, a related system screen 306, and a countermeasure 307. The event detail information management unit 115 reads out the information to be displayed as the related alarm list 303 on the event details screen 301 by referring to the "related alarm name" and "related process value" stored in the event detail information database 109. In addition, the event detail information management unit 115 reads out the information to be displayed as the countermeasure 307 by preferentially reading out the "countermeasure effectiveness" stored in the event detail information database 109, starting with the most effective.

[0027] By checking the event details screen 301 displayed on the display, the operator can check all the information used to determine whether the "occurring event name" identified by the event identification unit 114 is correct. If the "occurring event name" needs to be corrected, the operator can correct the information on the event details screen 301. The event detail information management unit 115 stores the event name corrected by the operator and each piece of information linked to the event name in the event detail information database 109. In this way, the operator's judgment is reflected in the event name determined by the operation monitoring support system 101. As a result, the operator's know-how accumulated in the operation monitoring support system 101 can be shared with other operators. In addition, the accuracy of event identification can be improved by issuing an alarm by the operation monitoring support system 101. Furthermore, the output display from the operation monitoring support system 101 allows the operator to check the "event name" and related information all at once. This reduces the burden on the operator of checking alarm-related information, process values, etc., compared to the past, and allows the operator to take appropriate action in a short amount of time. Therefore, the operation monitoring support system 101 can reduce the burden on the operator.

[0028] When an operator operates an equipment, the equipment operation output instructing unit 116 outputs the information to the plant control data 122 via the output unit 104. For example, the equipment operation output instructing unit 116 reads out information about the operator's operation, which information has been received by the operation monitoring history management unit 112 and stored in the operation monitoring history database 106, from the operation monitoring history database 106. Then, the equipment operation output instructing unit 116 outputs the information about the operator's operation history to the plant control data 122 via the output unit 104.

[0029] The function screen development unit 117 selects each function that can be used in the driving monitoring support system 101 from the screen information database 107. Then, the function screen development unit 117 develops each selected function as screen information. The screen information database 107 stores the format and style required for the screen display output to the output unit 104. The function screen development unit 117 reads data related to the format and style of each function of the driving monitoring support system 101 from the screen information database 107 as information for screen display.

[0030] The system management unit 118 performs various settings for the driving monitoring support system 101. The system management unit 118 automatically performs various settings for the driving monitoring support system 101 based on applications stored in advance in a storage unit or the like of the driving monitoring support system 101. For example, the system management unit 118 performs grouping of monitoring targets, setting thresholds for the monitoring targets, monitoring the set thresholds, and customizing the monitoring method. In addition, the system management unit 118 stores customized information for the monitoring method in the system information database 108.

[0031] The identification accuracy evaluation unit 119 evaluates whether the occurred event identified by the event identification unit 114 is correct or incorrect. The determination of whether the event is correct or incorrect is made by the operator who checks the event details screen 301 output by the above-mentioned event detail information management unit 115. The identification accuracy evaluation unit 119 evaluates whether the occurred event identified by the event identification unit 114 is correct or incorrect based on the "event name" and associated information stored in the event detail information database 109, and the information on the operator's operation monitoring stored in the operation monitoring history database 106. If the identification accuracy evaluation unit 119 detects an operation to correct the "event name" by the operator in the operator's operation monitoring stored in the operation monitoring history database 106, it evaluates the occurred event identified by the event identification unit 114 as incorrect. If the identification accuracy evaluation unit 119 evaluates the occurrence event identified by the event identification unit 114 as incorrect, it stores the "time information, alarm information, process value, and occurrence event name" in the dataset database 110 as additional dataset information. The data set management unit 120 changes the incorrect part "occurring event name" of the additional data set information written by the identification accuracy evaluation unit 119 to the correct event name. The data set management unit 120 changes the additional data set information by using the "occurring event name" input by the operator in the information of the operator's operation monitoring history stored in the operation monitoring history database 106. In addition, if the dataset management unit 120 makes changes to the dataset stored in the dataset database 110, the identification accuracy evaluation unit 119 re-trains the learning model 123 for the changed parts and updates the learning model 123.

[0032] [Learning model] Next, the learning model 123 will be described. The learning model 123 is a deep learning model trained with a dataset of "time information, alarm information, process value, and name of the occurring event" stored in the dataset database 110. The learning model 123 outputs the "name of the occurring event" from the input information of "time information, alarm information, and process value" by using a deep learning model trained with a pre-prepared dataset of "time information, alarm information, process value, and name of the occurring event."

[0033] The algorithm of the machine learning model for creating the learning model 123 is not particularly limited, and known methods such as random forest and k-nearest neighbor method are used. The learning model 123 is generated, for example, from a dataset of "time information, alarm information, process value, name of the occurring event" stored in advance in the dataset database 110, by machine learning using "time information, alarm information, process value" as learning data and "name of the occurring event" as teacher data.

[0034] FIG. 4 shows an example of a dataset 401 of "time information, alarm information, process value, and event name" stored in the dataset database 110. As shown in FIG. 4, the dataset 401 includes an alarm name associated with each event name, time-series information based on past alarm issuance cases as time information, and a process value associated with each alarm name. For example, in FIG. 4, the event names include a main steam isolation valve closure and an emergency shutdown (scram). Furthermore, alarm names associated with a main steam isolation valve closure include a drop in reactor water level, a drop in main steam pressure, a closure of the main steam isolation valve, and an increase in containment vessel pressure. Alarm names associated with an emergency shutdown (scram) include an occurrence of an emergency shutdown (scram) and an increase in reactor pressure. Furthermore, process values ​​associated with a drop in reactor water level include, for example, the reactor water level and the reactor water level change rate. Process values ​​associated with a drop in main steam pressure include, for example, the main steam pressure, the main steam temperature, and the main steam flow rate. Process values ​​associated with the closure of the main steam isolation valve include, for example, the main steam isolation valve state (open / closed) and the main steam isolation valve position. Process values ​​associated with the reactor containment vessel pressure include, for example, the reactor containment vessel pressure. Process values ​​associated with the occurrence of an emergency shutdown (scram) include, for example, the emergency shutdown signal state (on / off), the control rod insertion depth, and the reactor power. Process values ​​associated with the reactor pressure include, for example, the reactor pressure and the reactor pressure change rate.

[0035] 4, the data set database 110 stores in advance, for each event name, "associated alarm name, associated process value, event severity, similar event name, associated system screen, countermeasure method, and countermeasure effectiveness" in association with each other. This makes it possible to collectively display information related to the event that has occurred based on the information stored in the data set database 110.

[0036] Furthermore, if the event name identified by the learning model 123 is incorrect, the learning model 123 is retrained using additional dataset information in which the operator has input the correct event name as described above. At this time, the operation monitoring support system 101 automatically retrains the changed parts using the identification accuracy evaluation unit 119 and the dataset management unit 120, and the learning model 123 is updated. This can improve the accuracy of identifying event names by the learning model 123. Furthermore, by having the operator appropriately edit the countermeasure methods and the effectiveness of the countermeasures linked to the event names in the data set 401, the operator's know-how accumulated in the operation monitoring support system 101 can be shared with other operators.

[0037] 2. Embodiment of Operation Monitoring Assistance Method Next, an embodiment of a driving monitoring support method using the above-mentioned driving monitoring support system will be described. Fig. 5 shows a flowchart of the event identification support process and the event identification accuracy improvement process in the driving monitoring support method using the driving monitoring support system. The flowchart shown in Fig. 5 shows the process of realizing event identification support and event identification accuracy improvement at a certain fixed period.

[0038] First, the plant state monitoring unit 111 acquires the plant operation data 121 and determines whether an alarm is being issued (step S201). If an alarm is being issued (Yes in step S201), the event identifying unit 114 acquires the plant operation data "time information, alarm information, process value" at the time of alarm issuance from the plant operation data 121 (step S202). Then, the event identifying unit 114 identifies the occurring event that is the cause of the alarm by using the plant operation data "time information, alarm information, process value" acquired from the plant operation data 121 and the learning model 123 (step S203).

[0039] Next, the alarm information display unit 113 outputs the name of the occurred event identified by the event identification unit 114 to an output device such as an external display device via the output unit 104 (step S204). At this time, the event detail information management unit 115 links information such as the alarm name, time information, process value, event severity, similar event name, related system screen, countermeasure method, and effectiveness of the countermeasure method to the name of the occurred event identified by the event identification unit 114. Then, the alarm information display unit 113 outputs each piece of information linked to the occurred event name by the event detail information management unit 115 together with the occurred event name.

[0040] Next, the identification accuracy evaluation unit 119 evaluates whether the identification result of the occurred event name output by the event identification unit 114 was incorrect (step S205). At this time, the identification accuracy evaluation unit 119 checks the operation monitoring of the operator stored in the operation monitoring history database 106, and if it detects an operation to correct the "event name", it evaluates the identification result as incorrect. If the identification result of the occurring event name is incorrect (Yes in step S205), the identification accuracy evaluation unit 119 saves the "time information, alarm information, process value, and occurring event name" as additional dataset information in the dataset database 110 (step S206). At this point, the occurring event name included in the additional dataset is the occurring event name output by the event identification unit 114, and is incorrect.

[0041] Next, the data set management unit 120 changes the name of the occurring event in the additional data set stored in the data set database 110 to the correct occurring event name entered by the operator (step S207). Next, the identification accuracy evaluation unit 119 performs re-learning of the learning model 123 using the additional dataset in which the dataset management unit 120 has changed the names of the occurring events (step S208). Then, the identification accuracy evaluation unit 119 updates the existing learning model 123 to the learning model 123 after re-learning (step S209). If an alarm is not being issued (No in step S201), if the result of identifying the name of the occurring event is not incorrect (No in step S205), or after updating the learning model 123, the processing according to this flowchart ends.

[0042] As described above, the operation monitoring support system 101 automates the process from issuing an alarm to identifying the event that has occurred. This reduces the burden on operators when monitoring the operation of a plant or the like. Furthermore, the operation monitoring support system 101 makes it possible to identify the event that has occurred without depending on the ability of the operator. Furthermore, the operation monitoring support system 101 displays all information related to the event that has occurred in one place and provides guidelines for countermeasures, enabling operators to take prompt action. Furthermore, the operation monitoring support system 101 performs data analysis using a learning model that uses deep learning, and is therefore able to grasp complex patterns and interrelationships of process values ​​related to the operation of a plant, etc. Therefore, the operation monitoring support system 101 can analyze a large amount of real-time plant operation data quickly and efficiently. The learning model automatically learns the characteristics of the data without relying on specific parameters or conditions. This eliminates the need for manual adjustment of the judgment threshold. Furthermore, by evaluating the results of identifying the event that occurred in the operation monitoring support system 101 and having it learn again, the accuracy of event identification by the learning model can be improved.

[0043] It should be noted that the present invention is not limited to the above-described embodiments and various modifications are possible. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to an embodiment including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment. It is also possible to add the configuration of another embodiment to the configuration of one embodiment. It is also possible to delete part of the configuration of each embodiment, or to add or replace other configurations. [Explanation of symbols]

[0044] 101 Operation monitoring support system, 102 Input section, 103 Calculation section, 104 Output section, 105 Plant status database, 106 Operation monitoring history database, 107 Screen information database, 108 System information database, 109 Event detailed information database, 110 Data set database, 111 Plant status monitoring section, 112 Operation monitoring history management section, 113 Alarm information display section, 114 Event identification section, 115 Event detailed information management section, 116 Equipment operation output instruction section, 117 Function screen deployment section, 118 System management section, 119 Identification accuracy evaluation section, 120 Data set management section, 121 Plant operation data, 122 Plant control data, 123 Learning model, 125 Memory section, 301 Event details screen, 302 Event name, 303 Related alarm list, 304 Event severity, 305 Similar event name, 306 Related system screen, 307 Countermeasures, 401 datasets

Claims

1. a calculation unit that outputs plant control data based on input plant operation data; a memory unit in which a learning model trained with a data set including prepared time information, alarm information, process values, and names of occurring events is stored; The calculation unit an event identification unit that identifies a name of the occurring event using the learning model based on the newly acquired time information, the alarm information, and the process value information; an event detail information management unit that links the name of the occurring event output by the learning model with information related to the name of the occurring event; an alarm information display unit that outputs screen information that collectively displays the name of the occurring event and information linked to the name of the occurring event; Driving monitoring support system.

2. The calculation unit an operation monitoring history management unit that acquires a history of operation monitoring by an operator; an identification accuracy evaluation unit that evaluates the name of the occurring event identified by the event identification unit as incorrect when an operation to correct the name of the event by an operator is detected during the operation monitoring. The driving monitoring support system according to claim 1 .

3. When the identification accuracy evaluation unit evaluates the name of the occurred event identified by the event identification unit as incorrect, the identification accuracy evaluation unit stores the newly acquired time information, the alarm information, and the process value, as well as the name of the occurred event identified by the event identification unit, in the storage unit as additional data set information. The driving monitoring support system according to claim 2 .

4. The calculation unit a data set management unit that changes the name of the occurring event in the additional data set information stored in the storage unit to a correct event name; The driving monitoring support system according to claim 3 .

5. The data set management unit changes the name of the occurring event in the additional data set information by using the name of the occurring event corrected during the operation and monitoring by the operator. The driving monitoring support system according to claim 4.

6. The identification accuracy evaluation unit re-trains the learning model using the additional dataset information in which the occurrence event name has been changed by the dataset management unit. The driving monitoring support system according to claim 5 .

7. The alarm information display unit outputs screen information that collectively displays a list of related alarms, an event severity, a name of a similar event, a related system screen, and a countermeasure method as information linked to the name of the occurring event. The driving monitoring support system according to claim 1 .

8. An operation monitoring support method using a learning model trained with a data set including prepared time information, alarm information, process values, and names of occurring events, identifying a name of the occurring event using the learning model based on the time information, the alarm information, and the process value information newly acquired from plant operation data; Linking the occurrence event name output by the learning model with information related to the occurrence event name; outputting screen information that collectively displays the name of the occurring event and information linked to the name of the occurring event; Driving monitoring support method.

Images