Information processing device, alarm prediction method, and computer-readable storage medium

Abstract

This invention provides an information processing device, an alarm prediction method, and a computer-readable storage medium to improve alarm prediction accuracy and assist in safe factory operations. The information processing device uses models generated under different conditions to predict alarms that will occur under multiple virtually generated operating modes based on operator actions in a real factory. The information processing device determines the reliability of the prediction results for each of the multiple operating modes. Based on the reliability of the prediction results, the information processing device executes display control for each alarm.

Description

Technical Field

[0001] This invention relates to an information processing apparatus, an alarm prediction method, and a computer-readable storage medium. Background Technology

[0002] In various plants that utilize petroleum, petrochemicals, chemicals, gases, etc., safe operations are performed by workers (or operators). For example, workers determine the plant's operational tendencies based on measured values ​​such as temperature and pressure obtained from various sensors installed in the plant, such as temperature sensors and flow meters. Workers operate control instruments such as valves and heaters installed in the plant to maintain its operation. Furthermore, the operations described in this application also include on-site manual operations.

[0003] In recent years, real-time sensor values, measured values, control values, and other factory data have been obtained from actual factories (hereinafter sometimes referred to as real factories) to enable the operation of simulated or virtual factories. Virtual factories (hereinafter sometimes referred to as mirror factories) that follow the operation of real factories are used for operational assistance and training of workers (or operators).

[0004] Patent Document 1: Japanese Patent Publication No. 2009-9301

[0005] Patent Document 2: Japanese Patent Publication No. 2011-8756

[0006] In a mirror factory, the operational status of the real factory is predicted by simulating factory data, including manual operations on-site. However, operators may make decisions based on experience and subjective judgment according to the prediction results, potentially leading to incorrect choices regarding more efficient or safer operations. Summary of the Invention

[0007] The purpose of this invention is to improve the accuracy of alarm prediction and assist in safe factory operations.

[0008] An information processing apparatus includes: a prediction unit that uses models generated under different conditions to predict alarms that will occur under multiple virtually generated operating modes based on operations performed by operators on a real factory; a decision unit that determines the reliability of the prediction results for each of the multiple operating modes; and a display control unit that performs display control on each alarm based on the reliability of the prediction results.

[0009] One predictive method involves a computer performing the following processing: using models generated under different conditions, predicting alarms that would occur under multiple virtually generated operating modes related to operator actions at a real factory; determining the reliability of the prediction results for each of the multiple operating modes; and performing display control of each alarm based on the reliability of the prediction results.

[0010] A computer-readable storage medium stores an alarm prediction program that causes a computer to perform the following processes: using models generated under different conditions, predicting alarms that would occur under multiple virtually generated operating modes for a real factory, each based on the operations performed by operators on the real factory; determining the reliability of the prediction results for each of the multiple operating modes; and performing display control of each alarm based on the reliability of the prediction results.

[0011] One implementation method can improve the accuracy of alarm predictions and assist in safe factory operations. Attached Figure Description

[0012] Figure 1 This is a diagram illustrating an example of the overall structure of the system in Implementation Method 1.

[0013] Figure 2 This is a functional block diagram illustrating the functional structure of the information processing device in Implementation Method 1.

[0014] Figure 3 This is a diagram representing an example of information stored in the system's database.

[0015] Figure 4 This is a diagram representing an example of information stored in a relational database.

[0016] Figure 5 It is a trend chart representing the state of a real factory based on a simulation.

[0017] Figure 6 This is a diagram illustrating an example of generating multiple virtual operating modes.

[0018] Figure 7 This is a diagram illustrating the operation of multiple operating modes and displaying examples of predictive alarms.

[0019] Figure 8 This is a flowchart representing the process of trend display processing.

[0020] Figure 9 It is a flowchart representing the display processing flow of the operation mode.

[0021] Figure 10This is a diagram illustrating an example of the operation mode of Implementation Method 2.

[0022] Figure 11 This is a diagram illustrating Example 1 of the alarm in Implementation Method 3.

[0023] Figure 12 This is a diagram illustrating Example 2 of the alarm in Implementation Method 3.

[0024] Figure 13 This is a diagram illustrating an example of alarm display suppression in Implementation Method 4.

[0025] Figure 14 This is a diagram illustrating an example of collaboration between implementation method 5 and trend display.

[0026] Figure 15 This is a diagram illustrating an example of suppressing predictive alarms based on re-simulation.

[0027] Figure 16 This is a diagram illustrating an example of suppressing associated alarms based on re-simulation.

[0028] Figure 17 This is a flowchart illustrating the process of suppressing alarms based on re-simulation.

[0029] Figure 18 This is a graph illustrating the reliability of the simulation.

[0030] Figure 19 This is a diagram illustrating the display suppression of alarms based on reliability.

[0031] Figure 20 This is a flowchart representing the process of displaying and controlling alarms based on reliability.

[0032] Figure 21 This is a diagram illustrating an example of a hardware structure. Detailed Implementation

[0033] Hereinafter, embodiments of the information processing apparatus, alarm prediction method, and computer-readable storage medium disclosed in this application will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to these embodiments. Furthermore, the same reference numerals are used to label the same elements, and redundant descriptions are appropriately omitted; the embodiments can be appropriately combined within a non-contradictory scope.

[0034] (Implementation Method 1)

[0035] (Overall structure)

[0036] Figure 1 This is a diagram illustrating an example of the overall structure of the system in Implementation Method 1. (See diagram below.) Figure 1As shown, the system comprises a real factory 1 and a mirror factory 100. The mirror factory 100 constructs a virtual factory in real time, following the state of the real factory 1, to ensure the safe operation of the real factory 1. Specifically, the real factory 1 is a factory constructed in the real world using real-world instruments, while the mirror factory 100 is a virtual factory constructed in virtual space (cyberspace) using software, following the real factory 1. Furthermore, the real factory 1 and the mirror factory 100 are connected via a network, whether wired or wireless.

[0037] Real-world factory 1 is an example of various factories that utilize petroleum, petrochemicals, chemicals, gases, etc., and includes workshops and other facilities equipped with various facilities for obtaining products. Examples of products include LNG (liquefied natural gas), resins (plastics, nylon, etc.), and chemical products. Examples of facilities include factory facilities, machinery facilities, production facilities, power generation facilities, storage facilities, and facilities in well sites for drilling oil and natural gas, etc.

[0038] The actual factory 1 is constructed using distributed control systems (DCS). For example, although the diagram is omitted, the control system in the actual factory 1 uses the process data utilized by the actual factory 1 to perform various controls on the field instruments and other control instruments installed in the equipment being controlled, as well as the operating instruments corresponding to the equipment being controlled.

[0039] In addition, field instruments refer to operating instruments that have the function of measuring the operating status of the set equipment (such as pressure, temperature, flow rate, etc.) and the function of controlling the operation of the set equipment according to the input control signal (such as actuators, etc.). As a sensor, the field instrument outputs the operating status of the set equipment as process data to the controller in the control system in sequence, and the field instrument as an actuator controls the operation of the process according to the control signal calculated by the controller.

[0040] Here, process data includes measured values ​​(PV), setpoint values ​​(SV), and manipulated values ​​(MV). Furthermore, process data also includes information about the type of measured values ​​output (e.g., pressure, temperature, flow rate). Additionally, information such as the tag name assigned to the field instrument for identification is associated with the process data. Moreover, the measured values ​​output as process data can be not only measured values ​​by the field instrument acting as a sensor, but also calculated values ​​derived from the measured values, and manipulated values ​​for the field instrument acting as an actuator. The calculation of the calculated values ​​from the measured values ​​can be performed within the field instrument or by an external instrument (not shown) connected to the field instrument.

[0041] The mirror factory 100 is a virtual factory comprising a mirror model 200, a verification model 300, and a resolution model 400, and it tracks the state of the real factory 1 in real time. In addition to the instruments installed in the real factory 1, the mirror factory 100 can also virtually install instruments in locations that cannot be installed in the real factory 1, such as high temperatures and high altitudes, or virtually install instruments that are not installed due to cost reasons, thus providing effective services and enabling the real factory 1 to operate more accurately and stably. An example of the information processing device 10 executing each model has been described here, but it is not limited to this; each model can also be executed by a separate device.

[0042] The mirror model 200 simulates the actions of the real factory 1 by operating in parallel with the real factory 1, acquiring data from the real factory 1 while performing simulations. Simultaneously, it infers unmeasured state variables within the real factory 1, thus visualizing its internal structure. For example, the mirror model 200 may be a physical model that acquires process data from the real factory 1 and performs real-time simulations. In other words, the mirror model 200 visualizes the state of the real factory 1. For instance, the mirror model 200 takes process data from the real factory 1, tracks its actions, and outputs the results to the monitoring terminal 500. Furthermore, the mirror model 200 can also consider instruments not present in the real factory 1, predicting the actions of the real factory 1 after an operator performs a certain operation, and providing this information to the monitor.

[0043] To match the measured data of the mirror model 200 with that of the actual factory 1, the verification model 300 periodically infers the performance parameters of the instrument based on the data obtained from the actual factory 1. For example, the verification model 300 is a physical model used to adjust the error between the mirror model 200 and the actual factory 1. That is, the verification model 300 adjusts the parameters of the mirror model 200 as needed, either at regular intervals or when the error between the mirror model 200 and the actual factory 1 increases. For example, the verification model 300 obtains the values ​​of various parameters and variables representing performance from the mirror model 200, updates them, and outputs the updated values ​​to the mirror model 200. As a result, the values ​​of the parameters and variables in the mirror model 200 are updated. Furthermore, the values ​​of the parameters and variables include design data and operational data.

[0044] Analytical model 400 predicts the future operational state of the real factory 1 based on the actions of the real factory 1 simulated by mirror model 200. For example, analytical model 400 performs steady-state prediction, transition state prediction, and preventative diagnosis (anomaly diagnosis). As an example, analytical model 400 is a physical model that performs simulations to analyze the state of the real factory 1. That is, analytical model 400 performs future predictions of the real factory 1. For example, by using parameters and variables obtained from mirror model 200 as initial values ​​for high-speed calculations, analytical model 400 can predict the actions of the real factory 1 from the current point in time to several hours later and display this prediction as a trend chart.

[0045] In such a system, the information processing device 10, based on multiple virtually generated operating modes for operations performed by operators on a real factory 1, predicts the state transitions of each real factory 1 under different operating modes by simulating factory data. Furthermore, the information processing device 10 outputs the multiple operating modes in association with the simulated state transitions of each real factory 1. As a result, the information processing device 10 can provide operators with options for more efficient and safer operation, enabling safe and efficient use of the factory.

[0046] Furthermore, the information processing device 10 obtains information related to each alarm (predicted alarm) that is predicted to occur outside the range of a predefined state of the actual factory 1 by simulating factory data related to the operation of the actual factory 1. Moreover, the information processing device 10 performs display control on the monitoring terminal 500 monitoring the mirror factory 100 based on the relationship between each alarm and its related information. As a result, by using the mirror factory 100 that follows the actual factory 1 to predict alarm occurrences, the information processing device 10 can shorten the time spent on anomaly detection and cause identification in the actual factory 1.

[0047] (Functional Structure)

[0048] Figure 2 This is a functional block diagram illustrating the functional structure of the information processing device 10 in Embodiment 1. For example... Figure 2 As shown, the information processing device 10 includes a communication unit 11, a storage unit 12, and a processing unit 20.

[0049] The communication unit 11 is a processing unit that controls communication with other devices, and is implemented, for example, through a communication interface. For instance, the communication unit 11 controls communication with a real-world factory, and obtains factory data in real time. Furthermore, the communication unit 11 sends various information to the monitoring terminal 500 and displays various output information to the monitoring terminal 500.

[0050] Storage unit 12 is a processing unit that stores various data and programs executed by processing unit 20, and is implemented, for example, by a memory and a hard disk. Storage unit 12 stores system DB13 and associated DB14.

[0051] System DB13 is a database that stores the system structure of instruments and equipment installed within the real factory 1. For example, System DB13 stores an overview of instruments with upstream and downstream relationships based on instrument installation locations, product paths, and factory data paths. Furthermore, it is not limited to instruments installed within the real factory 1, but may also include instruments virtually installed within the mirror factory 100.

[0052] Figure 3 This is a diagram representing an example of information stored in system DB13. For example... Figure 3 As shown, system DB13 stores system 1, system 2, system 3, ..., system N. Here, larger numbers indicate that the system is located further downstream in the hierarchy. Figure 3 In the example, device A is located upstream, device B is downstream of device A, and device C is downstream of device B. Furthermore, in... Figure 3 In the example, it is shown that instrument X is located at the upstream end, instruments Y and Q are located downstream of instrument X, and instrument Z is located downstream of instrument Y.

[0053] In addition, the information stored in system DB13 can be pre-generated by administrators or automatically generated by parsing the design documents of real factory 1 and mirror factory 100.

[0054] Relational DB14 is a database that stores the relationships between process data (tags). Figure 4 This is a diagram representing an example of information stored in relational DB14. For example... Figure 4 As shown, DB14 stores "Operation Objects" and "Association Tags" in an associated manner. The "Operation Objects" stored here represent instruments operated by personnel, such as equipment temperature settings, flow meter settings, and valve opening / closing. The "Association Tags" represent instruments affected by the Operation Objects; as specific examples, they also include instruments and software sensors that are the same as the "Operation Objects." Figure 4 The example shows that along with the operation on "Action Label", "Associated Label 1", "Associated Label 2", and "Associated Label 3" were affected.

[0055] The processing unit 20 is responsible for the overall information processing device 10, and is implemented, for example, by a processor. The processing unit 20 includes a mirror processing unit 30, an identity processing unit 40, a prediction processing unit 50, and a display processing unit 60. Furthermore, the mirror processing unit 30, the identity processing unit 40, the prediction processing unit 50, and the display processing unit 60 are implemented through electronic circuits of the processor and processes executed by the processor.

[0056] The mirror processing unit 30 is a processing unit that performs state visualization of the real factory 1. Specifically, the mirror processing unit 30 obtains process data from the real factory 1 in real time, and visualizes the state of the real factory 1 by using real-time simulation of the physical model. That is, the mirror processing unit 30 uses the mirror model 200 described above.

[0057] The verification processing unit 40 is a processing unit that adjusts the errors between the simulation performed by the mirror processing unit 30 and the actual factory 1. Specifically, the verification processing unit 40 updates the values ​​of various parameters and variables used in the simulation performed by the mirror processing unit 30. That is, the verification processing unit 40 generates the aforementioned verification model 300.

[0058] The prediction processing unit 50 uses the analytical model 400 described above and is a processing unit that has a first prediction unit 51 and a second prediction unit 52, and performs a simulation to analyze the state of the real factory 1 in order to predict the future state of the real factory 1.

[0059] The first prediction unit 51 is a processing unit that predicts the actions of the actual factory 1 from the current time point, from a few minutes to a few hours later, and generates a trend chart. Specifically, the first prediction unit 51 performs action prediction simulations at any time, such as when instructed periodically by operators or other personnel, or when operations occur in the actual factory 1. In this embodiment, operators or other personnel are simply referred to as "operators or other personnel".

[0060] For example, if an operator performs the operation of "setting the temperature of equipment A to 50 degrees" in the actual factory 1 at time T, the first prediction unit 51 simulates the state of the actual factory 1 after time T by taking the operation information "the temperature of equipment A = 50 degrees" as input. The simulated state of the actual factory 1 here corresponds to the amount of products produced by the actual factory 1, the state quantities of the actual factory 1 including the pressure and temperature of a certain instrument affected by equipment A, etc.

[0061] Figure 5 This is a trend chart representing the state of a simulated real-world factory 1. For example... Figure 5 As shown, the first prediction unit 51 generates a trend graph with time as the horizontal axis and the state of the actual factory 1 as the vertical axis. Figure 5In the trend chart shown, TR110 is the actual measured value of Factory 1, and TR112 is the predicted data after the current moment.

[0062] The second prediction unit 52 is a processing unit that, based on multiple virtual operation modes generated for the operation performed by the operators on the real factory 1, predicts the state transition of each real factory 1 under the condition that multiple operation modes are executed by using simulation of factory data.

[0063] Specifically, the second prediction unit 52 predicts the state changes of the actual factory 1 when multiple operation modes have been executed since a certain point in time, at any time, such as when a new operation is performed, when instructed by operators, or when unstable behavior occurs in the actual factory 1, by using a pre-generated physical model and simulations of a model identical to the actual factory 1. At this time, the second prediction unit 52 can also further predict the number of alarms and alarms (predicted alarms) that occur in each operation mode.

[0064] Here, the processing of the second prediction unit 52 is explained in more detail. First, the second prediction unit 52 generates multiple virtual operation modes. Specifically, the second prediction unit 52 generates virtual operation modes available to operators from the current operating status of the real factory 1 up to any predetermined time afterward, based on operation guidelines, past operation history, etc., for a certain tag (operation tag). Figure 6 This is a diagram illustrating an example of generating multiple virtual operation modes. For example... Figure 6 As shown, the second prediction unit 52 generates virtual operation modes from mode 1 to mode 5.

[0065] Here, Mode 1 is the mode that executes operation A only at 12:00 from the current time. Mode 2 is the mode that executes operation B at 12:00 and operation A at 12:30 from the current time. Mode 3 is the mode that executes operation C only at 12:00 from the current time. Mode 4 is the mode that executes operation B at 12:00 and also at 12:30 from the current time. Mode 5 is the mode that executes operation B at 12:00 and operation C at 12:30 from the current time.

[0066] Next, the second prediction unit 52 uses Figure 6 The simulation of each operating mode shown predicts the temporal changes in the state of the actual factory 1. At this time, the second prediction unit 52 also predicts the number and time of alarms that occur in each operating mode, and displays the predicted state of the actual factory 1 in association with the alarms to the monitoring terminal 500.

[0067] Figure 7This is a diagram illustrating the actions (operations) of multiple operating modes and the display examples of predicted alarms (based on simulation results). Figure 7 The text shows examples of screen displays representing the timing of operations on arbitrary operation labels (such as temperature and valve opening / closing degree) for multiple operating modes. Figure 7 The horizontal axis setting time is not limited to 2-dimensional space; the dimension can also be increased by subdividing the state of the predicted object.

[0068] like Figure 7 As shown, the second prediction unit 52 displays the time series changes predicted for mode 1 as "BL111", the time series changes predicted for mode 2 as "BL112", the time series changes predicted for mode 3 as "BL113", the time series changes predicted for mode 4 as "BL114", and the time series changes predicted for mode 5 as "BL115".

[0069] Furthermore, for mode 1 "BL111", the second prediction unit 52 predicts and displays that after operation A at "12:00" is performed, an alarm will occur around "12:15", and displays the total number of alarm occurrences as "1". Similarly, for mode 2 "BL112", the second prediction unit 52 predicts that after operation B is performed at "12:00" and then operation A is performed at "12:30", three alarms will occur between "12:30" and "13:30", displaying the alarm occurrence times and displaying the total number of alarm occurrences as "3".

[0070] Furthermore, for mode 3 "BL113", the second prediction unit 52 predicts and displays that after operation C is performed at "12:00", an alarm will occur around "13:00", and displays the total alarm occurrence count as "1". Similarly, for mode 4 "BL114", the second prediction unit 52 predicts that after operation B is performed at "12:00" and also at "12:30", no alarm will be displayed, and displays the total alarm occurrence count as "0". For mode 5 "BL115", the second prediction unit 52 predicts and displays that after operation B is performed at "12:00" and operation C is performed at "12:30", an alarm will occur around "12:45", and displays the total alarm occurrence count as "1".

[0071] Thus, the second prediction unit 52 can inform operators of the timing and number of alarms occurring on operation labels such as temperature, based on the multiple operating modes available to them. As a result, operators can select the optimal operating mode with fewer alarms, which can contribute to safe operation in the actual plant 1.

[0072] return Figure 2The display processing unit 60 is a processing unit that has an acquisition unit 61 and a monitoring and control unit 62, and performs various controls when displaying the screen generated by the mirror processing unit 30 and the prediction processing unit 50.

[0073] The acquisition unit 61 is a processing unit that acquires images generated by the mirror processing unit 30 and the prediction processing unit 50. For example, the acquisition unit 61 is not limited to data format; it can also acquire trend information generated by the mirror processing unit 30 through simulation. Similarly, the acquisition unit 61 can also acquire the occurrence time and number of alarms generated through simulation for each operating mode from the prediction processing unit 50. In addition, the acquisition unit 61 outputs the acquired information to the monitoring and control unit 62.

[0074] The monitoring and control unit 62 is a processing unit that shapes various information acquired by the acquisition unit 61 and displays it on the monitoring terminal 500. For example, the monitoring and control unit 62 may highlight specific alarms, suppress the display of specific alarms, switch the display, or end the display of alarms that have been completed. Further details will be provided using the embodiments described later.

[0075] (Trend display processing flow)

[0076] Figure 8 This is a flowchart illustrating the process of trend display. For example... Figure 8 As shown, if the first prediction unit 51 obtains the latest factory data (S101: Yes), the verification model 300 infers the performance parameters of the instrument and performs verification processing of the mirror model 200 (S102). The first prediction unit 51 predicts the state of the real factory 1 after the current moment through simulation (S103).

[0077] Furthermore, the first prediction unit 51 generates a trend chart displaying the prediction results and... Figure 5 The output is displayed to the monitoring terminal 500 in the form shown (S104). In addition, the display destination can be set arbitrarily, such as the monitoring terminal of the actual factory 1, the smartphone and portable terminal of the operator, etc.

[0078] (The process of displaying and processing the operation mode)

[0079] Figure 9 This is a flowchart representing the display processing flow of the operation mode. For example... Figure 9 As shown, if the second prediction unit 52 is instructed to start processing (S201: Yes), it uses the instructions of the operator or the operation guide to determine the operation label of the simulated object (S202), and obtains multiple operation modes for the determined operation label (S203).

[0080] For example, the second prediction unit 52 can determine the next instrument to be targeted as the operation tag based on the operation steps. Furthermore, the multiple operation modes can be virtually generated operation modes or operation modes input by the operator.

[0081] Then, the second prediction unit 52 selects an operation mode from the multiple generated operation modes (S204) and performs a simulation using the selected operation mode (S205).

[0082] Subsequently, the second prediction unit 52 predicts the state of the actual factory 1 (the timing changes of the operation tags that become objects) and output alarms from the current time to a predetermined time later through simulation (S206). Then, the second prediction unit 52 counts the number of output alarms (S207).

[0083] Here, the second prediction unit 52 determines whether the prediction has ended for all operating modes (S208). If there are unpredicted operating modes (S208: no), the processing after S204 is performed for the next operating mode.

[0084] On the other hand, after the prediction of all operating modes is completed (S208: Yes), the second prediction unit 52 performs the association of operating mode, actual plant 1 status and alarm for each operating mode (S209), and uses... Figure 7 The associated output will be displayed in the form shown (S210).

[0085] (Effect)

[0086] As described above, the information processing device 10 can predict and output the status of the actual factory 1 for each operating mode, allowing operators to select operating modes with fewer alarms. As a result, the information processing device 10 reduces the likelihood of operators mistakenly choosing between more efficient and safer operating methods, enabling safe and efficient factory operation. Furthermore, unlike a simple simulator, the information processing device 10 performs simulations using a model identical to the actual factory 1, thereby obtaining more accurate results.

[0087] Furthermore, since the information processing device 10 can output the occurrence time and number of alarms in each operating mode, it can contribute to safer factory operations. Moreover, since the information processing device 10 can graphically output the state transitions of the actual factory 1 based on each operating mode, as well as the occurrence time and number of alarms, it can provide information for operators to make objective judgments, reducing the possibility of errors in selection by operators.

[0088] (Implementation Method 2)

[0089] In Implementation 1, an example of prediction for a single operation tag was described, but it is not limited to this. For example, the information processing device 10 can also predict associated tags that are associated with the operation tag at the same time.

[0090] Specifically, when multiple operating modes are executed, the information processing device 10 simultaneously predicts, through simulation, the occurrence of alarms caused by operators in the actual factory 1 for a first object (operation tag) among multiple objects, and the occurrence of alarms for at least one second object (association tag) affected by the operation of the first object.

[0091] The above example will be explained. If an operation tag is selected in S203, the information processing device 10 refers to... Figure 4 The correlation DB14 shown identifies the associated labels "Association Label 1, Association Label 2, and Association Label 3" that are associated with the operation label. Then, the information processing device 10 performs simulations of multiple operation modes for each operation label, and also performs simulations of multiple operation modes for each associated label.

[0092] Thus, the information processing device 10 predicts changes in operation tags and alarm occurrences when multiple operation modes are executed respectively, and also predicts changes in associated tags and alarm occurrences when multiple operation modes are executed respectively. Furthermore, the information processing device 10 can display and output these prediction results to the monitoring terminal 500.

[0093] Figure 10 This is a diagram illustrating an example of the operation mode of Embodiment 2. For example... Figure 10 As shown, the second prediction unit 52 of the information processing device 10 displays the predicted timing changes for mode 1 as "BL111", the predicted timing changes for mode 2 as "BL112", the predicted timing changes for mode 3 as "BL113", the predicted timing changes for mode 4 as "BL114", and the predicted timing changes for mode 5 as "BL115" for operation tags and each associated tag. Figure 10 The circular and quadrilateral marks (operations A to C) on the associated labels 1 to 3 are merely markers indicating the timing of the operation performed on the operation label, and are not operations performed on the associated labels 1 to 3. On the other hand, the diamond-shaped mark (alarm) displayed on the associated labels 1 to 3 indicates an alarm that occurred on the associated label.

[0094] For example, for mode 1 "BL111", after performing operation A at "12:00", the second prediction unit 52 predicts and displays an alarm around "12:15" in the operation label, an alarm around "12:35" in the associated label 1, an alarm around "13:00" in the associated label 2, and an alarm around "13:20" in the associated label 3. In addition, the second prediction unit 52 displays the alarm count as 1 and the total alarm count as (4) for the operation label, the alarm count as 1 for the associated label 1, the alarm count as 1 for the associated label 2, and the alarm count as 1 for the associated label 3.

[0095] Similarly, for mode 2 "BL112", after the second prediction unit 52 performs operation B at "12:00" and then performs operation A at "12:30", it predicts and displays alarms around "12:45", "13:00", and "13:20" respectively in the operation label, predicts and displays alarms around "13:00" and "13:20" respectively in the association label 1, predicts and displays alarms around "13:10" and "13:25" respectively in the association label 2, and predicts and displays alarms around "13:25" respectively in the association label 3. In addition, the second prediction unit 52 displays the alarm count as 3 and the total alarm count as (8) for the operation label, the alarm count as 2 for the association label 1, the alarm count as 2 for the association label 2, and the alarm count as 1 for the association label 3.

[0096] Similarly, for mode 3 "BL113", after the second prediction unit 52 performs operation C at "12:00", it predicts and displays an alarm around "13:15" in the operation label, an alarm around "13:15" in the associated label 1, an alarm around "13:20" in the associated label 2, and predicts no alarm in the associated label 3. In addition, the second prediction unit 52 displays the alarm count as 1 and the total alarm count as (3) for the operation label, the alarm count as 1 for the associated label 1, the alarm count as 1 for the associated label 2, and the alarm count as 0 for the associated label 3.

[0097] Similarly, for mode 4 "BL114", after the second prediction unit 52 performs operation B at "12:00" and also performs operation B at "12:30", it predicts that no alarm will occur in the operation label and associated label 1, predicts and displays an alarm around "13:10" in associated label 2, and predicts and displays an alarm around "13:20" in associated label 3. In addition, the second prediction unit 52 displays the alarm count as 0 and the total alarm count as (2) for the operation label, the alarm count as 0 for associated label 1, the alarm count as 1 for associated label 2, and the alarm count as 1 for associated label 3.

[0098] Similarly, for mode 5 "BL115", after the second prediction unit 52 performs operation B at "12:00" and operation C at "12:30", it predicts and displays an alarm around "12:50" in the operation label, and predicts no alarm in each associated label. In addition, the second prediction unit 52 displays the alarm count as 1 and the total alarm count as (1) for the operation label, the alarm count as 0 for associated label 1, the alarm count as 0 for associated label 2, and the alarm count as 0 for associated label 3.

[0099] Furthermore, the second prediction unit 52 can display each prediction screen on a single monitor, or it can switch between web screens and dedicated screens via tabs, or it can switch between screens using known switching operations such as swiping. Of course, the second prediction unit 52 is not limited to manual display switching; it can also switch automatically, like a slideshow.

[0100] As described above, the information processing device 10 can not only predict and output the prediction result for the first object (operation tag), but also predict and output the result for associated tags simultaneously. As a result, the information processing device 10 can suppress excessive information for operators and others, and can focus on providing necessary information that contributes to safe operation. Furthermore, the information processing device 10 can output judgment material to operators, allowing them to determine which operation mode to select even by simply observing the total number of alarms displayed for the operation tag (even without looking at the associated tag display). In addition, the information processing device 10 can highlight the operation mode with the fewest total alarms.

[0101] (Implementation Method 3)

[0102] In the event of a predicted high frequency of alarms, it is conceivable that the information overload would increase the visual burden on operators. Even in this case, the information processing device 10 can suppress the information overload on operators by highlighting specific alarms. Furthermore, in this embodiment, alarms occurring for a single operating mode are considered as the same type of alarm.

[0103] Specifically, the information processing device 10 performs display control on each alarm for the monitoring terminal 500 used to monitor the mirror factory 100, based on the relationship between the alarms predicted by simulation and indicating that the real factory 1 is outside a predetermined state. For example, the display processing unit 60 of the information processing device 10 displays each alarm sequentially according to the expected output order, and for multiple related alarms of the same type, highlights the first alarm of the same type that was output.

[0104] Figure 11This is a diagram illustrating Example 1 of the alarm in Implementation Method 3. Figure 11 The example shown is the same as Figure 7 The display example described herein is the same, so detailed explanation is omitted. If such a screen is displayed using the second prediction unit 52, the display processing unit 60 will only highlight the foremost alarm for the same type of alarm. Figure 11 In the example, the display processing unit 60 displays three identical alarms (alarms R1, R2, and R3) between "12:30" and "13:30" for mode 2 (BL112), thus highlighting the first alarm R1.

[0105] In addition, this kind of highlighting can also be handled for related tags. Figure 12 This is a diagram illustrating Example 2 of the alarm in Implementation Method 3. Figure 12 The example shown is the same as Figure 10 The display example described herein is the same, so detailed explanation is omitted. If such a screen is displayed by the second prediction unit 52, the display processing unit 60 will highlight only the foremost alarm of the same type among the labels. Figure 12 In the example, the display processing unit 60 highlights the top alarm for each mode in the operation label, but does not highlight other alarms in the operation label or alarms for each associated label.

[0106] In this way, by executing the above-described highlighting control, the information processing device 10 can reduce the amount of information and improve the visibility for operators and others.

[0107] In Implementation 3, an example is described where the foremost alarm is highlighted when multiple alarms are predicted to occur, but this is not the only example.

[0108] For example, the display processing unit 60 suppresses the display of the same type of alarm other than the first alarm of the same type among a plurality of the same type of alarms, or the same type of alarm after a predetermined time has elapsed since the first alarm of the same type of alarm was output.

[0109] exist Figure 11 In the example, the display shows that the processing unit 60 suppresses the display of alarms R2 and R3 out of alarms R1, R2, and R3. Furthermore, in... Figure 12 In the example, the display processing unit 60 displays the alarms initially output in each operating mode in each label, and suppresses the display of other alarms.

[0110] Furthermore, the display processing unit 60 can suppress the display of alarms for a predetermined time (e.g., 20 minutes) starting from the first alarm, or suppress the display of alarms other than the first alarm after a predetermined time has elapsed following a complete display. Moreover, the display processing unit 60 can display alarms for upstream instruments (e.g., operation tags) and suppress alarms for downstream instruments (e.g., associated tags). Additionally, the suppression of the display is not limited to complete blackout, but also includes changing the color or making the color semi-transparent.

[0111] (Implementation Method 4)

[0112] Regarding predicted alarms, typically, when their time arrives, operators or others perform some form of response in the actual factory 1 or mirror factory 100 to avoid the alarm. In this case, even if it is predicted, continuously displaying the alarm still increases the visual burden on the operators. Therefore, in Implementation 4, an example is described whereby the visual burden on operators or others is reduced by suppressing the display of alarms after the predicted alarm that has been avoided.

[0113] Specifically, after displaying multiple alarms of the same type, the information processing device 10 changes the display to not display alarms of the same type that have been executed by operators or others after the corresponding time. Figure 13 This is a diagram illustrating an example of alarm display suppression in Implementation Method 4. Figure 13 It shows Figure 7 The screen described in the text. In this state, if the display processing unit 60 detects that the time is "12:35" and performs an avoidance operation corresponding to alarm R1 of mode 2 "BL112" in the real factory 1, it sets subsequent alarms R2 and R3 to non-display. In addition, if the information processing device 10 performs or adds an operation not predetermined in BL112, it performs a re-simulation and display at this point in time.

[0114] Thus, the information processing device 10 can link the operation of the actual factory 1 with the alarm display, and can distinguish between unanswered alarms and answered alarms, thereby improving visibility for operators. For example, if an alarm is answered based on an unplanned operation, the information processing device 10 can perform a re-simulation. Furthermore, when only low-importance alarms are set to not be displayed, the information processing device 10 can also set related alarms to not be displayed. In addition, it is useful for timing the next prediction, such as when a predetermined number of alarms are cleared. Here, not displaying is not limited to changing the displayed content to another display format, but also includes ending the display.

[0115] (Implementation Method 5)

[0116] The information processing device 10 described above can also display trend displays and operation mode predictions in a comparable manner. Here, operation tags are used as an example for explanation, but associated tags can also be processed in the same way.

[0117] Figure 14 This is a diagram illustrating an example of collaboration between implementation method 5 and trend display. For example... Figure 14 As shown, the second prediction unit 52 performs simulations for the operation tag using multiple virtual operation modes (BL111 to BL115), predicts the movement of the operation tag and the occurrence of alarms, and displays the screen containing its prediction results on the monitoring terminal 500. Furthermore, the second prediction unit 52 outputs the content of the operation contained in each operation mode and the time of alarm occurrence, etc., as information related to the multiple operation modes to the first prediction unit 51.

[0118] The first prediction unit 51 simulates the overall operation of the actual factory 1 using the operational content included in each of the multiple operation modes (BL111 to BL115), and generates the expected trend. Furthermore, the first prediction unit 51, as... Figure 14 As shown in the right figure, after the prediction is completed at "12:00", the expected data for each operation mode will be displayed on the trend chart.

[0119] In this way, the information processing device 10 can provide operators with information on how each operating mode affects the overall operation of the actual factory 1. Therefore, operators can select an operating mode that makes the actual factory 1 operate more safely, thus achieving safe operation of the actual factory 1.

[0120] (Implementation Method 6)

[0121] The alarm display control method is not limited to the above-described embodiments; various criteria can be used to perform display suppression, etc. Therefore, another method for alarm display suppression will be described in Embodiment 6.

[0122] For example, regarding alarms predicted to occur through simulation of the aforementioned multiple operating modes (predicted alarms), the information processing device 10 can consciously change the parameters of the tag that generates the alarm, identify the affected tag (predicted alarm), and suppress the display of the alarm.

[0123] Specifically, the information processing device 10 determines, through simulation using the mirror model 200, an alarm predicted to occur due to the prediction process value (hereinafter referred to as the prediction process value) exceeding a threshold, and displays a simulation result screen containing the alarm. Furthermore, the information processing device 10 performs a second simulation while forcibly setting the prediction process value corresponding to the alarm to below the threshold, determines the result of the second simulation and any alarms that did not occur, and suppresses the display of the alarm from the simulation result screen. Additionally, in this embodiment, an alarm using operation tags is described as an example of a first alarm, and an alarm using association tags is described as an example of a second alarm. However, this is not limited to these examples; both the first and second alarms can be alarms related to operation tags, alarms related to association tags, or even alarms where the first alarm is an alarm with an association tag and the second alarm is an alarm with an operation tag.

[0124] Figure 15 This is a diagram illustrating an example of suppressing predictive alarms based on re-simulation. Figure 15 This indicates that it is aimed at and Figure 10 The document describes the prediction of alarm occurrences for multiple operation modes related to the operation labels. For example... Figure 15 As shown, the occurrence of alarms (Ⅰ), (Ⅱ), and (Ⅲ) is predicted in operation mode BL112 by using simulation of mirror model 200.

[0125] In this state, the predictive processing unit 50 arbitrarily selects an alarm (Ⅰ) corresponding to a specific first alarm and obtains the predictive process corresponding to alarm (Ⅰ) from the simulation results. Here, the predictive process is equivalent to the values ​​of the actual process and sensor values ​​of the actual plant 1, such as temperature, humidity, and piping flow.

[0126] Next, the prediction processing unit 50, based on setting the prediction process value corresponding to alarm (I) to be below a threshold, performs a re-simulation using the mirror model 200 and outputs the result. For example, if the output is after the predicted temperature of alarm (I) is above the threshold (e.g., 40 degrees), the prediction processing unit 50 sets the temperature to 30 degrees and performs a re-simulation.

[0127] The display processing unit 60 determines that alarm (III) will no longer be displayed by resimulating. That is, alarm (III) depends on alarm (I), and the display processing unit 60 determines that a response to alarm (III) has been made by performing a response to alarm (I).

[0128] As a result, the display processing unit 60 suppresses the display of alarm (III) from the screen displaying alarm (I), alarm (II), and alarm (III) for operation mode BL112. Furthermore, the display processing unit 60 can restore the prediction process settings, perform a re-simulation to suppress the display, or change the display result of the initial simulation.

[0129] Here, Figure 15 The instructions describe how to handle tags, but alerts for associated tags can be handled in the same way. Figure 16 This is a diagram illustrating an example of suppressing associated alarms based on re-simulation. Figure 16 This indicates that it is aimed at and Figure 10 The diagram describes the prediction of alarm occurrences for multiple operation modes related to the operation tags and alarm occurrences for associated tags. Although the attached diagram uses different reference numerals... Figure 10 Different, but the occurrence of the alarm is the same as Figure 10 same.

[0130] In this state, suppose the prediction processing unit 50 performs... Figure 15 The simulation is repeated as shown. At this time, the display processing unit 60 detects that alarm (V) associated with tag 1 and alarm (VII) associated with tag 2 are no longer displayed. That is, alarm (V) and alarm (VII) depend on alarm (I), and the display processing unit 60 determines that a response has been made for alarm (V) and alarm (VII) by performing a response to alarm (I).

[0131] That is, the display processing unit 60 determines alarms (V) and (VII) as related alarms to alarm (I). The result is as follows: Figure 16 As shown, the display processing unit 60 can also suppress the display of alarms (V) and (VII) from the display screen of the initially simulated alarm.

[0132] Additionally, in a re-simulation following a forced change to alarm (Ⅰ), if neither the alarms for the operation label nor the associated label can be cleared, other alarms can be selected for the same processing. However, since the re-simulation target can be arbitrarily selected, even if the associated alarm is cleared in a re-simulation of a certain alarm, other alarms can still be processed. Figure 15 and Figure 16 The processing.

[0133] Furthermore, the above examples illustrate the forced change of the prediction process value, but are not limited to these. The prediction process value can also be set to be below a threshold by changing parameters such as the physical model and formulas used in the calculation of the prediction process value. Moreover, the object of the forced change is not limited to the prediction process value; it can also include sensor values ​​from software used within the mirror model 200.

[0134] Figure 17 This is a flowchart illustrating the process of suppressing alarms based on re-simulation. For example... Figure 17As shown, the prediction processing unit 50 selects an alarm for investigation of related alarms (S001), forcibly resets the prediction process value of the selected alarm to a threshold range (S002), and performs a simulation in the reset state (S003).

[0135] Furthermore, the display processing unit 60 classifies the alarms that have been cleared as related alarms, except for the selected alarms (S004). Subsequently, the prediction processing unit 50 restores the prediction process value of the selected alarms and performs a simulation (S005), and the display processing unit 60 suppresses the display of alarms that have been classified as related alarms in the simulation result display (S006).

[0136] As described above, the information processing device 10 performs simulations for multiple operating modes, and can display alarms predicted to occur and identify alarms with high correlation. Furthermore, the information processing device 10 can prompt operators to indicate which alarm will be cleared, providing information that helps operators select the optimal operating mode.

[0137] (Implementation Method 7)

[0138] In mirror model 200, simulation is performed using a model (e.g., an approximation) that corresponds to the load conditions of the actual plant 1. However, it is impractical to generate a model that corresponds to all the assumed loads. Therefore, it is advisable to prepare several models in advance and perform simulations while corresponding to the load interpolation models of the predicted objects. That is, it can also be considered that the reliability of the simulation results varies depending on the interpolation conditions.

[0139] Furthermore, depending on the nature of the elements that become the modeling objects, there exist models that can be rigorously formulated, as well as examples that correspond to approximate formulas that match actual actions. That is, the accuracy of the model varies depending on the elements. Therefore, the reliability of the simulation varies depending on the simulation conditions and the accuracy of the model.

[0140] Therefore, when the information processing device 10 simulates an operation mode and displays alarms, it suppresses excessive information for operators by providing information on the reliability of the prediction results or filtering the displayed alarms. Furthermore, in this embodiment, the relationship between interpolation conditions (interpolation ratio and load, etc.) and reliability is predefined using tables or the like.

[0141] Next, with Figure 10 The reliability is illustrated using the various operating modes shown. Figure 18 This is a graph illustrating the reliability of the simulation. For example... Figure 18As shown, the information processing device 10 pre-generates and maintains a model that assumes and adjusts the load of the actual factory 1 to 50%, and a model that assumes and adjusts the load of the actual factory 1 to 80%. Here, load refers to, for example, the load of processes executed in the actual factory 1, the quantity and quality of products, the flow rate of piping, etc.

[0142] In this state, with Figure 10 Similarly, the predictive processing unit 50 obtains operating modes BL111, BL112, BL113, BL114, and BL115. Among them, operating mode BL111 is the operation content with a 50% load, operating mode BL112 is the operation content with a 60% load, operating mode BL113 is the operation content with a 20% load, operating mode BL114 is the operation content with a 75% load, and operating mode BL115 is the operation content with a 90% load.

[0143] In this case, since the operation mode BL111 is designed with a 50% load, the predictive processing unit 50 predicts the occurrence of the alarm by simulating the model with a 50% load. Therefore, the predictive processing unit 50 sets the reliability of the operation mode BL111 to 100.

[0144] Similarly, since the operation mode BL112 is designed for a 60% load, the prediction processing unit 50 predicts the occurrence of an alarm by simulating a model that interpolates between the models at 50% load and 80% load. Therefore, the prediction processing unit 50 sets the reliability of the operation mode BL112 to 90%.

[0145] Similarly, since the operation mode BL113 is designed for a 20% load, the prediction processing unit 50 predicts the occurrence of an alarm by simulating a model that interpolates the models for 50% load and 80% load. Therefore, the prediction processing unit 50 sets the reliability of the operation mode BL113 to 80%.

[0146] Furthermore, since the operation mode BL114 is designed for a 75% load operation, the prediction processing unit 50 predicts the occurrence of alarms by simulating a model that interpolates between the models at 50% load and 80% load. Therefore, the prediction processing unit 50 sets the reliability of the operation mode BL114 to 95.

[0147] Furthermore, since the operation mode BL115 is designed for a 90% load operation, the prediction processing unit 50 predicts the occurrence of alarms by simulating a model that interpolates the models for 50% load and 80% load. Therefore, the prediction processing unit 50 sets the reliability of the operation mode BL115 to 85.

[0148] Furthermore, the display processing unit 60 suppresses the display of alarms whose reliability is below a threshold (e.g., 90) among those predicted by simulation using each model. Figure 19 This is a graph illustrating the display suppression of reliability-based alarms. For example... Figure 19 As shown, the display unit 60 suppresses alarms in operation mode BL113 and operation mode BL115 whose reliability is lower than the threshold in each of the alarms of operation tag, associated tag 1, associated tag 2, and associated tag 3.

[0149] Furthermore, the display control of each alarm can be implemented by considering the importance of each alarm based on risk analysis and other factors. For example, the display processing unit 60 may display alarms that would cause a major event if they occur, even if the simulation reliability is low. As another example, the display processing unit 60 may also consider the magnitude of the deviation between the target process value (or alarm threshold) and the prediction result to determine whether to display an alarm. For example, the display processing unit 60 may suppress the display of alarms near the threshold when the simulation reliability is high, and display alarms near the threshold even when the reliability is low.

[0150] Figure 20 This is a flowchart illustrating the display control process for reliability-based alarms. For example... Figure 20 As shown, the information processing device 10 sets the alarm detection threshold (S1) by using instructions from operators or other personnel.

[0151] Next, the information processing device 10 predicts the occurrence of alarms in the mirror factory 100 and calculates their reliability (S2). Then, the information processing device 10 sets an importance level based on risk management, or calculates the deviation between the target value and the predicted value of the location where the alarm occurs (S3).

[0152] Subsequently, the information processing device 10 calculates the alarm display threshold (S4) from the reliability (and importance or the deviation between the target value and the predicted value). For example, the information processing device 10 can calculate it by multiplying the threshold set to reliability 100 by the rate of decrease from reliability 100 to reliability α and the proportion representing the deviation, etc., or it can be set arbitrarily.

[0153] Then, the information processing device 10 displays only alarms that are predicted to occur and whose alarm display threshold is above the threshold (S5), and recommends an operating mode from the total number of displayed alarms (S6). For example, the information processing device 10 recommends an operating mode with the fewest operations.

[0154] Alternatively, instead of using the reliability of the simulation, the probability of the event (alarm) occurring can be used to determine the suppression of the predicted alarm display. For example, if the process is affected by weather or other factors, the information processing device 10 can calculate the probability of the alarm occurring by taking into account factors such as the probability of precipitation one hour later.

[0155] More specifically, the information processing device 10 can also display an alarm related to processes affected by weather conditions such as dust, regardless of the reliability of the aforementioned model, when the probability of precipitation during the predicted period, i.e., one hour later, is 50% or higher. Furthermore, the information processing device 10 can add a predetermined value (e.g., 10) to the reliability of the aforementioned model, and conversely, in cases related to processes involving cooling temperatures due to rain, etc., it can subtract a predetermined value (e.g., 10) from the reliability of the aforementioned model.

[0156] Furthermore, when using machine learning models or the like, the information processing device 10 can also obtain the probability of each alarm occurring, and thus can use the aforementioned reliability and probability of occurrence to change the color and intensity of the alarm.

[0157] (Implementation Method 8)

[0158] The embodiments of the present invention have been described above. However, the present invention can be implemented in various other ways besides the embodiments described above.

[0159] (numerical values, etc.)

[0160] The screen display example, time, examples of each label, number of systems, number of associated labels, and number of alarms used in the above embodiments are only examples and can be arbitrarily changed. Furthermore, each simulation can use a pre-generated physical model. Moreover, each simulation can employ a machine learning model, which is generated using training data that correlates inputs (descriptive variables) such as operational content like temperature with outputs (objective variables) such as label values.

[0161] Furthermore, while the processes described in Embodiments 6 and 7 are illustrated using operation tags and associated tags as examples, they are not limited to this and can also be applied to various operations and settings within the factory that are the objects of operation. Additionally, the alarm prediction objects in the processes described in Embodiments 6 and 7 are not necessarily limited to operation tags and associated tags with known relationships; they can be applied to operations tags with unknown relationships, associated tags, or operations tags and associated tags with unknown relationships.

[0162] Furthermore, while Implementation 6 describes an example of changing the values ​​of the prediction process to perform a re-simulation, it is not limited to this. For example, the parameters of the simulation model and machine learning model used for prediction can also be changed to perform a re-simulation. However, the number of predicted alarms does not have to be multiple; it can be either one or zero.

[0163] Furthermore, while an example of using models with different loads was described in Embodiment 7, the invention is not limited to this; models generated under conditions different from the prediction conditions can also be used. For example, models generated according to environmental conditions such as temperature, humidity, and climate, as well as operational conditions such as the skill level of the workers, can be used, even if not limited to the factory load. Even in this case, as in Embodiment 7, the more different the generation conditions of each model are from the conditions at which each alarm is predicted, the lower the reliability of the prediction result will be. For example, the information processing device 10 can also use deviation, etc., instead of interpolation or extrapolation. An example is given: when the information processing device 10 uses a model generated under a temperature of 40 degrees Celsius to make a prediction under a prediction condition of 30 degrees Celsius, the reliability of the prediction result can be calculated as "100 × 30 / 40 = 75".

[0164] (Operation Mode)

[0165] For example, the operation mode virtually generated by the second prediction unit 52 can be an operation mode for a certain operation tag, or it can be an operation mode related to the entire real factory 1 or the entire mirror factory 100 containing multiple operation tags.

[0166] (Same type of alarm)

[0167] In the above embodiments, it is explained that alarms occurring for a single operating mode are considered as the same type of alarm, but this is not a limitation. For example, if the simulation of the second prediction unit 52 is a physical model whose cause can be predicted or determined, the same type of alarms can be grouped according to the cause.

[0168] For example, in Figure 11 In the example, when alarms R1 and R3 have the same cause, but alarm R2 has a different cause, the display processing unit 60 suppresses the display for alarm R3, but does not suppress the display for alarm R2. Furthermore, the determination based on the cause corresponds to, for example, alarms R1 and R3 occurring at temperatures above 70 degrees Celsius, and alarm R2 occurring at flow rates below 10 L / min. This processing can be applied to various embodiments. For example, the display processing unit 60 can also perform display control for alarms with the same cause between operation tags and associated tags.

[0169] (system)

[0170] The processing steps, control steps, specific names, and information containing various data and parameters shown in the above-described specification and accompanying drawings may be changed arbitrarily, unless otherwise specified.

[0171] Furthermore, the structural elements of each device illustrated are functional conceptual elements and do not necessarily have to be physically constructed as shown in the illustrations. That is, the specific ways in which the devices are distributed and integrated are not limited to those shown in the illustrations. In other words, all or part of them can be distributed and integrated in any unit, either functionally or physically, to correspond to various loads and usage conditions.

[0172] Furthermore, all or any part of the processing functions executed in each device can be implemented by a CPU (Central Processing Unit) and the program parsed and executed by the CPU, or as hardware based on wiring logic.

[0173] (hardware)

[0174] Next, an example of the hardware structure of the information processing device 10 will be described. Figure 21 This is a diagram illustrating an example of a hardware structure. For example... Figure 21 As shown, the information processing device 10 includes a communication device 10a, an HDD (Hard Disk Drive) 10b, a memory 10c, and a processor 10d. Furthermore, Figure 21 The various parts shown are interconnected by buses, etc.

[0175] Communication device 10a is a network interface card, etc., used for communication with other servers. HDD 10b provides storage. Figure 2 The functions shown are performed by the program and database (DB).

[0176] Processor 10d reads data from HDD 10b, etc., for execution and... Figure 2 The same processing program is shown in each processing unit and expanded in memory 10c, thereby enabling execution. Figure 2 The processor 10d performs the functions described in the description. For example, the process performs the same functions as each processing unit in the information processing device 10. Specifically, the processor 10d reads a program from the HDD 10b, etc., that has the same functions as the image processing unit 30, the authentication processing unit 40, the prediction processing unit 50, the display processing unit 60, etc. Moreover, the processor 10d executes a process that performs the same processing as the image processing unit 30, the authentication processing unit 40, the prediction processing unit 50, the display processing unit 60, etc.

[0177] Thus, the information processing device 10 operates as an information processing device that performs various processing methods by reading and executing programs. Furthermore, the information processing device 10 can also achieve the same functions as the embodiments described above by reading the aforementioned programs from storage media using a media reading device and executing the read programs. Additionally, the other programs described in the embodiments are not limited to those executed by the information processing device 10. For example, the present invention can also be applied to cases where other computers or servers execute programs, or where they cooperate in executing programs.

[0178] The program can be distributed via networks such as the Internet. Furthermore, the program can be recorded on computer-readable storage media such as hard disks, floppy disks (FD), CD-ROMs, MO (Magneto Optical disk), and DVDs (Digital Versatile Disc), and executed by a computer by reading it from the storage media.

Claims

1. An information processing device, characterized in that, include: The prediction unit uses models generated under different conditions to predict alarms that will occur when multiple operating modes, which are virtually generated to describe the operations performed by operators on the real factory, are executed for each real factory. The decision-making unit determines the reliability of the prediction results for each of the multiple operating modes. as well as The display control unit executes the display control of each alarm based on the reliability of the prediction results. Each of the multiple operating modes is configured with load information representing the load of the actual factory when each operating mode is executed. When the prediction unit performs predictions for the multiple operating modes, it uses interpolation or extrapolation of each model generated using different load information to perform predictions corresponding to the load information of each of the multiple operating modes. The decision unit determines the reliability of the prediction result based on the interpolation status of the interpolation or extrapolation.

2. The information processing device according to claim 1, characterized in that, The more different the generation conditions of each model are from the conditions under which each alarm is predicted, the lower the reliability of the prediction result determined by the decision-making unit.

3. The information processing apparatus according to claim 1 or 2, characterized in that, When making predictions for the multiple operating modes, the model corresponding to the load information set for each of the multiple operating modes is used from the models generated using different load information.

4. The information processing apparatus according to claim 1 or 2, characterized in that, The display control unit generates display screens for the predicted alarms in a time sequence according to the occurrence order for each of the multiple operating modes. The display control unit suppresses the display of alarms whose prediction results have a reliability lower than a threshold among the multiple alarms displayed on the display screen.

5. An alarm prediction method, characterized in that, The computer will perform the following processing: Using models generated under different conditions, predict alarms that will occur when multiple operating modes, which are virtually generated to describe the operations performed by operators on the real factory, are executed for the real factory. The reliability of the prediction results is determined for each of the multiple operating modes; as well as The display control of each alarm is executed based on the reliability of the prediction results. Each of the multiple operating modes is configured with load information representing the load of the actual factory when each operating mode is executed. When making predictions for the multiple operating modes, interpolation or extrapolation of each model generated using different load information is used to perform predictions corresponding to the load information of each of the multiple operating modes. The reliability of the prediction result is determined based on the interpolation status of the interpolation or extrapolation.

6. A computer-readable storage medium, characterized in that, An alarm prediction program is stored, which causes the computer to perform the following processes: Using models generated under different conditions, predict alarms that will occur when multiple operating modes, which are virtually generated to describe the operations performed by operators on the real factory, are executed for the real factory. The reliability of the prediction results is determined for each of the multiple operating modes; as well as The display control of each alarm is executed based on the reliability of the prediction results. Each of the multiple operating modes is configured with load information representing the load of the actual factory when each operating mode is executed. When making predictions for the multiple operating modes, interpolation or extrapolation of each model generated using different load information is used to perform predictions corresponding to the load information of each of the multiple operating modes. The reliability of the prediction result is determined based on the interpolation status of the interpolation or extrapolation.

Images