JP7175786B2 - Abnormality detection device, simulator, plant monitoring system, abnormality detection method and program - Google Patents

Description

The present invention relates to an abnormality detection device, a simulator, a plant monitoring system, an abnormality detection method, and a program.

In some cases, the operating state of actual plant equipment is monitored through various sensors to determine the current state of the plant or predict the future state of the plant. For example, there is known a driving support device using a predictive simulator that predicts future driving conditions based on monitoring data of actual equipment.

In addition, in Patent Document 1, an abnormality occurrence time estimation method is determined based on the identification result of an abnormal event, and a conduit for estimating the occurrence time of the identified abnormal event based on the compatibility and the abnormality occurrence time estimation method An anomaly detection device is disclosed.

JP 2017-002554 A

A simulator is known that faithfully simulates the operation of an actual plant by computer calculation. It is desired to apply such a simulator to accurately and quickly detect an abnormality in an actual machine.

The present invention provides an abnormality detection device, a simulator, a plant monitoring system, an abnormality detection method, and a program capable of accurately and early detection of an abnormality in an actual plant using a simulator of the actual plant.

According to the first aspect of the present invention, an abnormality detection device includes an acquisition unit that acquires an actual machine measurement value from an actual plant machine, acquires a control command from a control device of the plant, and indicates the state of the actual machine of the plant. a simulation unit that calculates a simulator calculated value based on the control command; and when the degree of divergence of the actual machine measured value from the calculated simulator calculated value exceeds a predetermined judgment threshold, it is determined that an abnormality has occurred in the actual machine. an anomaly determination unit for determining, an estimation unit for estimating the time when the anomaly occurred based on the time series of the actual machine measurement values up to the time when the anomaly was determined to occur, wherein, when it is determined that the abnormality has occurred, the estimation unit notifies the simulation unit of an accident signal indicating that an abnormality has occurred, and the simulation unit acquires the accident signal, Calculation of the simulator-calculated value based on the control command is stopped, the simulator-calculated value is traced back to the value of the simulator-calculated value at the time when the abnormality occurred, and the simulator calculation reflects the abnormality after that time. Calculate the value.

According to the second aspect of the present invention, the estimating section further estimates the type of abnormality based on the condition satisfied by the actual machine measurement value at the time when it is determined that the abnormality has occurred in the actual machine.

According to the third aspect of the present invention, the estimation unit further reflects the occurrence of the abnormality of the estimated type in the simulation unit, and simulates from the time when the abnormality occurred to the current time. Based on the time series of simulator calculated values obtained from operation, the type and scale of the abnormality are estimated.

According to the fourth aspect of the present invention, the estimating unit includes a time series of the actual machine measurement values acquired by a time when a predetermined time has passed after at least the time when it is determined that the abnormality has occurred, and , the type and scale of the anomaly are estimated based on the time-series similarity of the simulator-calculated values.

According to the fifth aspect of the present invention, the estimation unit estimates the time when the abnormality occurred based on the time-series rate of change of the actual machine measurement value.

According to the sixth aspect of the present invention, the simulator includes an acquisition unit that acquires an actual machine measurement value from an actual plant machine, acquires a control command from a control device of the plant, and a simulator calculation that indicates the state of the actual machine of the plant. a simulation unit that calculates a value based on the control command; and an abnormality judgment that determines that an abnormality has occurred in the actual machine when the degree of divergence of the actual machine measured value from the simulator calculated value exceeds a predetermined judgment threshold. and an estimating unit that, when it is determined that the abnormality has occurred, estimates the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the determination was made , The estimating unit notifies the simulation unit of an accident signal indicating that the abnormality has occurred when it is determined that the abnormality has occurred, and the simulation unit acquires the accident signal and performs the control operation based on the control command. Calculation of the simulator calculated value is stopped, the simulator calculated value is traced back to the value of the simulator calculated value at the time when the abnormality occurred, and the simulator calculated value reflecting the abnormality after that time is calculated. .

According to a seventh aspect of the present invention, a plant monitoring system includes the control device and the abnormality detection device described above.

According to the eighth aspect of the present invention, an abnormality detection method comprises the steps of acquiring an actual machine measurement value from an actual plant machine and acquiring a control command from a control device of the plant; a step of calculating a simulator calculated value shown based on the control command; and when the degree of divergence of the actual machine measured value from the calculated simulator calculated value exceeds a predetermined determination threshold, an abnormality has occurred in the actual machine. when it is determined that the abnormality has occurred, the step of estimating the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the determination was made; a step of notifying the simulator of an accident signal indicating that an abnormality has occurred when it is determined that a , and causing the simulator-calculated value to trace back to the value of the simulator-calculated value at the time when the abnormality occurred, and calculating the simulator-calculated value reflecting the abnormality after that time .

According to the ninth aspect of the present invention, the program acquires, in the computer of the abnormality detection device, an actual machine measurement value measured from an actual machine of the plant, and acquires a control command from the control device of the plant; calculating a simulator-calculated value indicating the state of the actual plant based on the control command; a step of determining that an abnormality has occurred in the actual machine; and, if it is determined that the abnormality has occurred, estimating the time at which the abnormality has occurred based on the time series of the actual machine measurement values up to the time when the determination is made. the step of notifying the function for calculating the simulator-calculated value of an accident signal indicating that the abnormality has occurred when it is determined that the abnormality has occurred; and the function of calculating the simulator-calculated value, When the accident signal is acquired, the calculation of the simulator-calculated value based on the control command is stopped, the simulator-calculated value is traced back to the value of the simulator-calculated value at the time when the abnormality occurred, and the simulator-calculated value after that time. and a step of calculating the simulator calculated value reflecting the abnormality .

According to the abnormality detection device, the simulator, the plant monitoring system, the abnormality detection method, and the program described above, an abnormality in the actual plant can be accurately and quickly detected using the simulator of the actual plant.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows the whole plant monitoring system structure which concerns on 1st Embodiment. 3 is a diagram showing the functional configuration of a simulator according to the first embodiment; FIG. It is a figure which shows the processing flow of the simulator based on 1st Embodiment. FIG. 4 is a diagram for explaining in detail the processing of the simulator according to the first embodiment; FIG. FIG. 4 is a diagram for explaining in detail the processing of the simulator according to the first embodiment; FIG. FIG. 4 is a diagram for explaining in detail the processing of the simulator according to the first embodiment; FIG. FIG. 4 is a diagram for explaining in detail the processing of the simulator according to the first embodiment; FIG. FIG. 4 is a diagram for explaining in detail the processing of the simulator according to the first embodiment; FIG.

<First Embodiment>
The plant monitoring system according to the first embodiment will be described in detail below with reference to FIGS. 1 to 8. FIG.

(Overall configuration of plant monitoring system)
The plant monitoring system 1 is, for example, a system that monitors the operation of an actual PLR in a nuclear power plant. The plant monitoring system 1 includes a control device 10 and a simulator PLV.

The control device 10 outputs a control command for controlling the operation of the actual PLR of the plant. Further, the control device 10 also outputs to the simulator PLV the same signal as the control command output to the actual machine PLR. The control device 10 acquires a measured value (hereinafter also referred to as “actual device measured value”) from the actual PLR. Similarly, the simulator PLV also acquires actual machine measurement values. The simulator PLV acquires a calculated value (hereinafter, also referred to as “simulator calculated value”) indicating the operating state of the actual machine PLR calculated by itself. That is, the simulator PLV acquires the actual machine measured value and the simulator calculated value from each of the actual machine PLR and the simulator PLV that operate based on the same control command. The simulator PLV also functions as an anomaly detection device that detects an anomaly of the actual PLR by comparing the acquired actual machine measured values and the simulator calculated values.

The actual PLR is a facility composed of a turbine, a boiler, and various types of piping. Various sensors are provided at various locations in the actual machine PLR, and the control device 10 acquires actual machine measurement values via these sensors. The various sensors include various instruments capable of measuring pipe temperature, pressure, flow rate, etc., tank water storage amount detection sensors, valve open/close detection sensors, and the like. Actual machine measurement values acquired through a plurality of sensors are sequentially acquired by the control device 10 and the simulator PLV.

The simulator PLV is a so-called digital twin of the actual PLR, and is actually an arithmetic unit (computer). That is, the simulator PLV is a real machine simulation model that simulates the operation of the real machine PLR. The simulator PLV is constructed by a group of functions representing physical phenomena that occur in the actual PLR during operation. The simulator PLV incorporates the control commands received from the control device 10, performs physical calculations, and simulates the operating state of the actual machine PLR. The simulator PLV is, for example, developed for monitoring of the actual PLR, operation training, etc., and is preferably constructed with a simulation model that has a proven track record of consistency with the actual PLR.

In this embodiment, the actual PLR is assumed to be a device, equipment, etc. installed in a nuclear power plant, but in other embodiments, it may be a power plant other than nuclear power (thermal power plant, etc.). , devices, facilities, etc. installed in plants other than power plants (for example, chemical plants, waste treatment plants, etc.).

(Simulator functional configuration)
FIG. 2 is a diagram showing the functional configuration of the simulator according to the first embodiment.
As shown in FIG. 2, the simulator PLV includes a CPU 100, a memory 101, a communication interface 102, a monitor 103, an input device 104, and a storage 105.

The memory 101 is a so-called main storage device, and expands commands and data for the CPU 100 to operate based on programs.

The communication interface 102 is an interface device for communicating with the outside of the simulator PLV (actual PLR, control device 10, etc.).

The monitor 103 is a display device that visually displays information, and may be, for example, a liquid crystal display or an organic EL display.

The input device 104 is an input device that receives the operation of the user of the simulator PLV, and may be, for example, a general mouse, keyboard, touch sensor, or the like.

The storage 105 is a so-called auxiliary storage device, and may be, for example, an HDD (Hard Disk Drive), an SSD (Solid State Drive), or the like. The storage 105 records an abnormality estimation table DB (described later).

The CPU 100 is a processor that controls the overall operation of the simulator PLV. The CPU 100 according to this embodiment functions as an acquisition unit 1000, an abnormality determination unit 1001, a notification processing unit 1002, an estimation unit 1003, and a simulation unit 1004, as shown in FIG. Each function of the CPU 100 will be described below.

Acquisition unit 1000 acquires real machine measurement values from various sensors provided in real machine PLR. Acquisition unit 1000 also acquires a simulator calculated value from simulation unit 1004 . The acquisition unit 1000 also acquires the control command output by the control device 10 to the actual PLR.

The abnormality determination unit 1001 determines whether there is an abnormality in the operation of the actual PLR. Specifically, the abnormality determination unit 1001 determines that an abnormality has occurred in the actual PLR when the degree of divergence of the actual machine measured value from the simulator calculated value acquired by the acquisition unit 1000 exceeds a predetermined determination threshold. .

When the abnormality determination unit 1001 determines that an abnormality has occurred in the operation of the actual PLR, the notification processing unit 1002 notifies the operator of the occurrence of the abnormality through the monitor 103 or the like.

When the abnormality determination unit 1001 determines that an abnormality has occurred in the operation of the actual PLR, the estimation unit 1003 estimates various information regarding the abnormality. Specifically, when it is determined that an abnormality has occurred in the operation of the actual PLR, the estimation unit 1003 estimates the time when the abnormality occurred, the type of abnormality, and the scale of the abnormality. A specific aspect of the processing in which the estimation unit 1003 estimates information on these abnormalities will be described later.

The simulation unit 1004 reflects the actual machine measurement values and control commands acquired by the acquisition unit 1000 in a function group representing physical phenomena in the plant, performs physical calculations, and calculates simulator calculation values. For example, when a part of the control command output by the control device 10 is input, the function group outputs the same value as the actual machine measured value measured by the actual machine PLR that received the same control command. The simulation unit 1004 acquires the latest control command through the acquisition unit 1000 and continuously calculates the simulator calculated value using the value. The simulation unit 1004 may use actual machine measurement values to calculate the simulator calculation values. As a result, the simulator PLV functions as a digital twin that simulates the operating state of the actual PLR in real time. The values calculated by the simulator include physical quantities such as temperature and pressure to be monitored at locations where actual sensors are not installed. By referring to the simulator-calculated values, the operator can obtain various information that cannot be obtained from the actual machine measurement values alone, which can be useful for plant monitoring. Therefore, when the simulator calculated values output by the simulation unit 1004 are used for plant monitoring, the simulator calculated values are required to maintain high simulation accuracy.

(Simulator processing flow)
FIG. 3 is a diagram illustrating a processing flow of the simulator according to the first embodiment;
4 to 8 are diagrams for explaining in detail the processing of the simulator according to the first embodiment.

The processing flow shown in FIG. 3 is continuously and repeatedly executed during operation of the actual PLR.

First, the acquiring unit 1000 of the simulator PLV acquires actual machine measured values through various sensors of the actual machine PLR, and also acquires simulator calculated values from the simulation unit 1004 (step S30). The simulator calculated value is the physical quantity α calculated by the simulation unit 1004 using the control command or the like acquired by the acquisition unit 1000 . This physical quantity α is also included in the actual machine measurement value acquired by the acquisition unit 1000 .

Next, the abnormality determination unit 1001 of the simulator PLV calculates the degree of deviation between the actual machine measured value and the simulator calculated value acquired in step S30, and determines whether or not the degree of deviation exceeds a predetermined determination threshold. (step S31). For example, the abnormality determination unit 1001 compares the actual machine measured value of the physical quantity α with the simulator calculated value of the physical quantity α. If the degree of divergence does not exceed the predetermined determination threshold (step S31; NO), the abnormality determination unit 1001 determines that the actual PLR has no abnormality. In this case, the CPU 100 returns to the process of step S30.

If the degree of divergence exceeds a predetermined determination threshold (step S31; YES), it is determined that an abnormality has occurred in the actual PLR. In this case, the notification processing unit 1002 of the simulator PLV notifies the operator of the occurrence of the abnormality through the monitor 103 or the like (step S32).

The processing of steps S31 to S32 described above will be described in detail with reference to FIG.
The graph shown in FIG. 4 shows the time series DR of actual machine measured values and the time series DV of simulator calculated values obtained from a certain reference time t0 to the current time (time ta). FIG. 4 shows a comparison between the actual machine measured value for one parameter (for example, the pressure of a certain pipe A) obtained from the actual machine PLR and the simulator calculated value for the same parameter (the pressure of the pipe A). there is In the present embodiment, actual machine measured values and simulator calculated values are actually compared for each of a plurality of types of parameters.

As shown in FIG. 4, it is assumed that the degree of divergence g (g=|DRa−DVa|) between the actual machine measured value DRa and the simulator calculated value DVa obtained at a certain time ta exceeds the judgment threshold gth (step S31 ; YES). In this case, the abnormality determination unit 1001 determines that an abnormality has occurred in the actual PLR. Accordingly, the notification processing unit 1002 notifies the occurrence of an abnormality at time ta (step S32). That is, at time ta, the occurrence of an abnormality is detected by the simulator PLV (abnormality detection device).

Returning to FIG. 3, subsequently, the estimating unit 1003 of the simulator PLV executes the following processing in order to estimate various types of information regarding the abnormality that has occurred.

The estimation unit 1003 estimates the abnormality occurrence time (step S33). The abnormality occurrence time is the time when the abnormality actually occurred in the actual PLR, and is the time before the time when the abnormality was detected (time ta shown in FIG. 4).

The processing of step S33 described above will be described in detail with reference to FIG.
Similar to FIG. 4, the graph 5a shown in FIG. 5 shows the time series DR of actual machine measured values and the time series DV of simulator calculated values obtained from a certain reference time t0 to the current time (time ta). A graph 5b shown in FIG. 5 shows a rate of change ΔDR that is a time differential of the time-series DR of the actual machine measurement values.

In this embodiment, the estimating unit 1003 determines whether an abnormality has occurred in the actual PLR based on the time series DR of the actual measurement values from the predetermined reference time t0 to the time (time ta) when it is determined that there is an abnormality. Estimate the time. Specifically, estimating section 1003 calculates the rate of change ΔDR of the time-series DR of the actual device measured values up to time ta, and calculates the time (time tb) when the rate of change ΔDR exceeds a predetermined determination threshold value ΔDRth. Identify. Here, the time tb at which the rate of change ΔDR exceeds the determination threshold ΔDRth can be regarded as the change point of the trend of the time-series DR of the actual machine measured values. Therefore, the estimation unit 1003 estimates the time tb specified as described above as the time when the abnormality occurred. The estimation unit 1003 notifies the simulation unit 1004 of an accident signal indicating the occurrence of an abnormality. Upon acquiring the accident signal, the simulation unit 1004 temporarily stops calculating the simulator calculated value. Furthermore, estimation section 1003 instructs simulation section 1004 to return the simulated state of the actual PLR to the operating state of the actual PLR at time tb. The simulation unit 1004 reproduces the simulator calculated value at time tb. For example, the simulation unit 1004 records the simulator calculated value, the actual machine measured value, and the control command for each time in the storage 105 in association with the time. Upon receiving an instruction from the estimating unit 1003 to return the simulated state to time tb, the simulating unit 1004 reads the simulator calculated values and the like at time tb from the storage 105 and reproduces the operating state of the actual PLR at time tb.

Returning to FIG. 3, next, the estimation unit 1003 determines the type of abnormality that has occurred in the actual PLR (step S34). The type of abnormality is information specifying the mode, nature, occurrence location, etc. of the abnormality.

The processing of step S34 described above will be described in detail with reference to FIG.
The information table shown in FIG. 6 is an example of the abnormality estimation table DB stored in the storage 105 in advance. As shown in FIG. 6, the abnormality estimation table DB records combinations of conditions satisfied by a plurality of types of actual machine measurement values for each "type of abnormality" defined in advance. Each condition T1, T2, . . . , P1, P2, . . . , F1, F2, . below), or the range of variation of each actual machine measurement value DR_T, DR_P, DR_F, . . .

In step S34 described above, the estimating unit 1003 refers to the abnormality estimation table DB, and the time series DR of the actual machine measurement values acquired up to the current time (time ta) is defined for each "type of abnormality". It is determined whether or not the combination of conditions described above is satisfied. For example, when the actual machine measured value DR_T satisfies the condition T1, the actual machine measured value DR_P satisfies the condition P1, and the real machine measured value DR_F satisfies the condition F1, the estimating unit 1003 determines that the accident event X1 It is presumed that (pipe A breakage) occurred.

Depending on the time-series combination of actual device measurement values, it is conceivable that a combination of multiple types of conditions defined in the abnormality estimation table DB may apply at the same time. For example, it is conceivable that a time-series combination of a certain actual machine measurement value satisfies both the conditions of the accident event Xa and the conditions of the accident event Xb. In this case, the estimation unit 1003 estimates that "either or both of the accident events Xa and Xb have occurred" in the actual PLR.

Returning to FIG. 3, next, the estimating unit 1003 performs the processing of steps S35 to S37 described below in estimating the scale of the abnormality that has occurred in the actual PLR. The processing of step S35 will be described in detail with reference to FIG.

In the initial state, the operation of the actual PLR that the simulator PLV (simulation unit 1004) simulates by physical calculation is calculated on the premise that no abnormality has occurred in the actual PLR. Therefore, when an abnormality occurs in the actual PLR, there is a discrepancy in the operating conditions between the actual PLR in which the abnormality has occurred and the simulator PLV that performs calculations on the assumption that no abnormality has occurred. Therefore, the estimating unit 1003 causes the simulator PLV, which is traced back to the time when the abnormality occurred (time tb identified in step S33), to reflect the type of abnormality identified in step S34 (input malfunction). Then, the simulated operation is redone (step S35).
Here, "to reflect the abnormality in the simulator PLV" means that, as shown in FIG. 7, the type of abnormality (for example, "pipe A rupture ”) is incorporated into the computation of the simulation unit 1004 . As a result, the simulator PLV becomes capable of simulating operation of the actual machine PLR in the state where the abnormality (pipe A breakage) has occurred.
In addition, in the processing of the first step S35, the estimation unit 1003 randomly determines the “scale of abnormality” to be reflected in the simulator PLV (for example, the fracture area: 1.0 cm ² , etc.).

As described above, in step S35, the estimation unit 1003 reflects the type of abnormality (for example, pipe A fracture) and the scale of the abnormality (for example, fracture area: 1.0 cm ² ) in the simulator PLV. , a simulated operation is executed from the time when the abnormality occurred (time tb) to the present time. Next, the estimating unit 1003 changes the scale of the abnormality reflected last time (for example, changes the fracture area: 1.0 cm ² → 2.0 cm ² ), and again changes the abnormality occurrence time (time tb) to the current time. Perform a simulated operation up to This is repeated for variations in the scale of anomalies set in advance.

Next, the estimating unit 1003 obtains the time series of the actual machine measured values up to the present time, and the simulator calculated values of all variations obtained by the simulated operation of the simulator PLV reflecting the occurrence of the abnormality from the abnormality occurrence time (time tb). The similarity with the time series of is calculated (step S36).
The magnitude of the anomaly whose degree of similarity between the two is most consistent with the predetermined determination threshold is estimated as the magnitude of the anomaly occurring in the actual PLR (step S37).

The processing of steps S36 and S37 described above will be described in detail with reference to FIG.
The estimating unit 1003 executes the processing of steps S35 and S36, and as shown in FIG. Get time series of multiple simulator calculation values. For example, in the example shown in FIG. 8, the estimating unit 1003 calculates the time series DV1 of the simulator calculation values as a result of the simulated operation with the fracture area of "1.0 cm ² " and the simulated operation with the fracture area of "2.0 cm ² ". A time series DV2 of the simulator calculated values obtained as a result of the operation and a time series DV3 of the simulator calculated values obtained as a result of the simulated operation with the fracture area of "1.5 cm ² " are obtained. In this case, the estimation unit 1003 makes the determination in step S36 based on the high similarity between the time series DV3 of the simulator calculated values and the time series DR of the actual machine measured values. Then, in step S37, the estimating unit 1003 acquires an estimation result of "fracture area: 1.5 cm ² " for the scale of the abnormality that occurred in the actual PLR.

In the example shown in FIG. 8, the estimating unit 1003 calculates the actual machine measured value and the simulator calculated value acquired until the time (time tc) at which a predetermined time has passed after the time ta at which it is determined that an abnormality has occurred. are included in the similarity calculation. By doing so, the similarity determination is made including the trend of the driving state after the time when the abnormality was detected, so it is possible to improve the accuracy of estimating the scale of the abnormality. However, other embodiments are not limited to this aspect. You can calculate degrees. In other words, the similarity determination may be performed including the trend of the driving state before the time when the abnormality was detected.

Next, the notification processing unit 1002 notifies various types of information about the abnormality (step S38). For example, the notification processing unit 1002 notifies the operator of information such as the occurrence time of the abnormality, the detection time of the abnormality, the type of the abnormality, the scale of the abnormality, etc. through the monitor 103 or the like (step S38).

(action, effect)
As described above, the simulator PLV (abnormality detection device) according to the first embodiment acquires the actual machine measured value from the actual plant PLR, and acquires the simulator calculated value from the simulator PLV that simulates the operation of the actual plant PLR. an acquisition unit 1000, an abnormality determination unit 1001 that determines that an abnormality has occurred in the actual PLR when the degree of divergence of the actual machine measured value from the acquired simulator calculated value exceeds a predetermined determination threshold, and an abnormality occurrence and a notification processing unit 1002 that notifies that an abnormality has occurred when it is determined that an abnormality has occurred.
In this way, when the simulator PLV deviates from the operating state calculated on the assumption that no abnormality has occurred, the actual machine PLR is determined to be abnormal, so that the occurrence of abnormality can be detected quickly and accurately. can.

Further, the simulator PLV according to the first embodiment further includes an estimating unit 1003 for estimating information (abnormality occurrence time, type of abnormality, scale of abnormality) regarding the occurrence of an abnormality when the occurrence of an abnormality is detected. It has
By doing so, the operator can grasp the information about the abnormality that has occurred, and can take prompt and appropriate measures. As a result, early convergence of the abnormality can be achieved.

The estimating unit 1003 also has a function of estimating the time of occurrence of an abnormality and retroactively tracing the operating state of the actual PLR simulated by the simulator PLV to the time of the time of occurrence of the abnormality. The abnormality occurrence time is just before the difference between the simulator calculated value and the actual machine measured value occurs, and the simulator calculated value calculated by the simulation unit 1004 at the abnormality occurrence time reproduces the operating state of the actual machine PLR with high accuracy. The estimating unit 1003 performs malfunction input to the operating state simulated with high accuracy, and estimates the type and scale of the abnormality, so that the type and scale of the abnormality occurring in the actual PLR can be accurately estimated. be able to.

In addition, the simulation unit 1004 calculates simulator calculation values based on the actual machine measurement values and control commands. The operating state of the actual PLR is simulated so as to match the . Then, at least part of the simulator calculated values gradually deviate from the actual operating state of the actual PLR, and the simulation accuracy decreases. For example, when using the simulator PLV for the purpose of monitoring, the operator will refer to an erroneous simulator calculated value for monitoring. On the other hand, according to the estimation unit 1003, it is possible to specify the time (abnormal occurrence time) when the simulated operating state starts to deviate from the operating state of the actual PLR, and notify the simulation unit 1004 of an accident signal. As a result, continuation of an erroneous simulation by the simulation unit 1004 can be stopped. Also, the period before the abnormality occurrence time can be determined as a period in which the simulator calculated value is correct. The simulator-calculated values for the period in which the values are determined to be correct can be used, for example, later for analysis of the occurrence of anomalies.

The process of each process in the simulator PLV (abnormality detection device) described above is stored in a computer-readable recording medium in the form of a program. will be Here, the computer-readable recording medium refers to magnetic disks, magneto-optical disks, CD-ROMs, DVD-ROMs, semiconductor memories, and the like. Alternatively, the computer program may be distributed to a computer via a communication line, and the computer receiving the distribution may execute the program.

Further, the program may be for realizing part of the functions described above. Further, it may be a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in the computer system.
Also, the simulator PLV (abnormality detection device) may be composed of one computer, or may be composed of a plurality of computers communicably connected.

In addition, it is possible to appropriately replace the components in the above-described embodiments with well-known components without departing from the scope of the present invention. Moreover, the technical scope of the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.
For example, the functions of the acquisition unit 1000, the abnormality determination unit 1001, the notification processing unit 1002, and the estimation unit 1003 may be provided in the control device 10. FIG.

1 plant monitoring system 10 control device 100 CPU
1000 acquisition unit 1001 abnormality determination unit 1002 notification processing unit 1003 estimation unit 1004 simulation unit 101 memory 102 communication interface 103 monitor 104 input device 105 storage PLR actual device PLV simulator (abnormality detection device)
DB Abnormality estimation table

Claims (9)

an acquisition unit that acquires an actual machine measurement value from an actual machine of a plant and acquires a control command from a control device of the plant;
a simulation unit that calculates a simulator calculated value indicating the state of the actual plant based on the control command;
an abnormality determination unit that determines that an abnormality has occurred in the actual machine when the degree of divergence of the actual machine measured value from the calculated simulator calculated value exceeds a predetermined judgment threshold;
an estimating unit that, when it is determined that the abnormality has occurred, estimates the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the abnormality was determined;
with
The estimating unit notifies the simulation unit of an accident signal indicating that the abnormality has occurred when it is determined that the abnormality has occurred;
When the simulation unit acquires the accident signal, the simulation unit stops calculating the simulator-calculated value based on the control command, and causes the simulator-calculated value to trace back to the value of the simulator-calculated value at the time when the abnormality occurred; calculating the simulator calculated value that reflects the abnormality after the time;
Anomaly detector. The estimation unit further
The anomaly detection device according to claim 1 , wherein the type of the anomaly is estimated based on a condition satisfied by the actual machine measurement value at a time when it is determined that the anomaly has occurred in the actual machine. The estimation unit further
After reflecting the occurrence of the estimated type of abnormality in the simulation unit, the abnormality is determined based on the time series of the simulator calculated values obtained from the simulated operation from the time when the abnormality occurred to the current time. The anomaly detection device according to claim 2 , which estimates a scale. The estimation unit
Based on the degree of similarity between the time series of the actual machine measured values and the time series of the simulator calculated values acquired by the time when a predetermined time has elapsed after at least the time when it is determined that the abnormality has occurred, The anomaly detection device according to claim 3 , which estimates the scale of the anomaly. The estimating unit estimates the time when the abnormality occurred based on the time-series change rate of the actual machine measurement value.
The abnormality detection device according to any one of claims 1 to 4 . an acquisition unit that acquires an actual machine measurement value from an actual machine of a plant and acquires a control command from a control device of the plant;
a simulation unit that calculates a simulator calculated value indicating the state of the actual plant based on the control command;
an abnormality determination unit that determines that an abnormality has occurred in the actual machine when the degree of divergence of the actual machine measured value from the simulator calculated value exceeds a predetermined judgment threshold;
an estimating unit that, when it is determined that the abnormality has occurred, estimates the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the abnormality was determined;
with
The estimating unit notifies the simulation unit of an accident signal indicating that the abnormality has occurred when it is determined that the abnormality has occurred;
When the simulation unit acquires the accident signal, the simulation unit stops calculating the simulator-calculated value based on the control command, and causes the simulator-calculated value to trace back to the value of the simulator-calculated value at the time when the abnormality occurred, calculating the simulator calculated value that reflects the abnormality after the time;
simulator. The abnormality detection device according to any one of claims 1 to 5 ;
the control device;
plant monitoring system. a step of acquiring an actual machine measurement value from an actual machine of a plant and acquiring a control command from a control device of the plant;
a step in which a simulator calculates, based on the control command, a simulator-calculated value indicating the state of the actual plant;
a step of determining that an abnormality has occurred in the actual machine when the degree of divergence of the actual machine measured value from the calculated simulator calculated value exceeds a predetermined judgment threshold;
when it is determined that the abnormality has occurred, estimating the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the abnormality was determined;
a step of notifying the simulator of an accident signal indicating that an abnormality has occurred when it is determined that the abnormality has occurred;
When the simulator acquires the accident signal, it stops calculating the simulator-calculated value based on the control command, causes the simulator-calculated value to trace back to the value of the simulator-calculated value at the time when the abnormality occurred, and a step of calculating the simulator calculated value reflecting the abnormality after the time;
An anomaly detection method comprising: In the computer of the anomaly detection device,
a step of acquiring an actual machine measurement value measured from an actual machine of a plant and acquiring a control command from a control device of the plant;
a step of calculating a simulator calculated value indicating the state of the actual plant based on the control command;
a step of determining that an abnormality has occurred in the actual machine when the degree of divergence of the actual machine measured value from the calculated simulator calculated value exceeds a predetermined judgment threshold;
when it is determined that the abnormality has occurred, estimating the time when the abnormality occurred based on the time series of the actual machine measurement values up to the time when the abnormality was determined;
a step of notifying the function for calculating the simulator calculated value of an accident signal indicating that the abnormality has occurred when it is determined that the abnormality has occurred;
When the function for calculating the simulator-calculated value acquires the accident signal, it stops calculating the simulator-calculated value based on the control command, and converts the simulator-calculated value to the simulator-calculated value at the time when the abnormality occurred. a step of calculating the simulator-calculated value that reflects the abnormality after the time by going back to the value;
program to run.

Images