JP6884078B2 - Plant abnormality monitoring system - Google Patents

Description

The present disclosure relates to a plant abnormality monitoring system for monitoring plant abnormalities.

In the plant, abnormality monitoring is performed to prevent the occurrence of failures. For example, the measured values (sensor values) of a plurality of sensors installed on various devices (equipment) constituting the plant are monitored. However, a monitoring method is known in which an alarm is issued to the driver and the manager when the sensor value exceeds a certain threshold value. Among the abnormalities notified in this way, for example, an alarm is issued due to a water supply leak (leakage) from damage caused by high temperature corrosion or wear on a heat transfer tube such as a furnace wall tube of a boiler. In some cases, a failure occurs, and when a failure occurs, it is necessary to take measures such as stopping the equipment (plant) in order to prevent its expansion.

Therefore, a method of detecting a sign of an abnormality before it is detected has been proposed (Patent Documents 1 and 2). For example, Patent Document 1 describes an abnormality sign monitoring system that monitors signs (signs) of plant abnormalities based on the operation history of the monitored plant, and an abnormality sign monitoring system that monitors the results of abnormality signs detected by the abnormality sign monitoring system. A plant operation system including an abnormality diagnosis system for performing abnormality diagnosis of a plant is disclosed. It is disclosed that the cause of the abnormality is identified by this abnormality diagnosis using a Bayesian network (statistical model).

Further, in Patent Document 2, in the plant monitoring and diagnosis method, the operating state of the plant, the operating state of the equipment, and the environmental state are detected, the detection data is accumulated, the inspection data of the plant is accumulated, and the above-mentioned accumulation is performed. It is disclosed to diagnose the current state of the plant based on the history information of the plant consisting of the detection data and the inspection data. As a result, even if the current state is sound, for example, if an abnormality sign or the like is detected in the past from the history information, a monitoring diagnosis is performed, and when an abnormality is detected, the abnormality location and the abnormality content are identified. However, it is said to predict anomalous ripple events.

International Publication No. 2017/051576 Japanese Unexamined Patent Publication No. 6-331507

Detecting signs of abnormality using the operation history of the plant as in Patent Documents 1 and 2 and predicting the abnormal event before the occurrence of the abnormality is effective as a method for preventing failure of the plant. Then, the inventors of the present invention have invented a new method for predicting (identifying) what kind of abnormal event the sign of abnormality develops.

In view of the above circumstances, at least one embodiment of the present invention aims to provide a plant abnormality monitoring system capable of predicting an abnormality event occurring in a plant before it occurs.

(1) The plant abnormality monitoring system according to at least one embodiment of the present invention is
A plant anomaly monitoring system that monitors anomalies based on multiple operational data that indicate the operating status of the plant.
An abnormality sign determination unit that determines the presence or absence of an abnormality sign in the plant based on the monitoring abnormality degree, which is an abnormality degree calculated based on the monitoring operation data including the plurality of operation data.
When the abnormality sign determination unit determines that the abnormality sign is present, the monitoring operation data is compared with at least one reference operation data including the plurality of operation data indicating the abnormality sign acquired in the past. It is equipped with an abnormality diagnosis unit that performs abnormality diagnosis of the plant based on the above.
The abnormality diagnosis unit
The contribution degree to the monitoring abnormality degree is calculated for each of the plurality of operation data constituting the monitoring operation data, and among the monitoring operation data, the top N items having a large monitoring contribution degree, which is the calculated contribution degree ( N represents an integer of 2 or more.) A diagnosis target data group extraction unit that extracts a diagnosis target data group composed of the above-mentioned operation data, and a diagnosis target data group extraction unit.
The top M items having a large reference contribution, which is the contribution to the abnormality calculated based on the reference operation data for each of the diagnosis target data group and the plurality of operation data constituting the reference operation data ( M indicates an integer of 2 or more.) A matching index calculation unit that calculates a matching index with the diagnostic reference data group composed of the operation data, and a matching index calculation unit.
It has an abnormality prediction unit that predicts the abnormality of the plant based on the reference operation data in which the match index is equal to or higher than the match determination threshold value.

According to the configuration of (1) above, the plant abnormality monitoring system monitors operation data (monitoring operation data) obtained through measurements by a plurality of sensors installed in the plant, and is calculated based on the monitoring operation data. When an abnormality sign is detected based on the degree of abnormality (monitoring abnormality degree, for example, Mahalanobis distance (MD)), the abnormality indicated by this abnormality sign is diagnosed. Specifically, the contribution (monitoring contribution, for example, the SN ratio when the abnormality is MD) of a plurality of operation data (monitoring operation data) used for calculating the monitoring abnormality is calculated. Next, the contributions calculated by the same method for the top N (diagnosis target data group) with a large degree of monitoring abnormality and at least one reference operation data accumulated as indicating an abnormality sign of an abnormality that occurred in the past. The matching index with the top M (diagnosis reference data group) of the degree (reference contribution. Past monitoring operation data) is calculated. Then, the reference operation data in which the monitoring operation data and the match index are equal to or higher than the match determination threshold value is searched, and the occurrence of an abnormality corresponding to the abnormality sign indicated by the obtained reference operation data is predicted.

In other words, matching the currently monitored monitoring operation data (diagnosis target data group) with the reference operation data (diagnosis reference data group) showing abnormal signs of abnormalities (trouble) experienced in the past contributes to each degree of abnormality. It is performed from the viewpoint of degree, and the abnormality is predicted assuming that there is a high possibility that the abnormal sign indicated by the matched reference operation data has occurred. As a result, it is possible to predict an abnormal event (content) before an abnormality that has occurred in the past occurs, and it is possible to prevent a plant failure by dealing with the predicted abnormality.

(2) In some embodiments, in the configuration of (1) above,
The matching index refers to i operations belonging to both the diagnosis target data group and the diagnosis reference data group with respect to the total monitoring contribution of the operation data constituting the diagnosis target data group or the diagnosis reference data group. Includes the total percentage of the monitoring contributions for the data (where i ≦ N, M).
According to the configuration of (2) above, by using the contribution to the calculation of the match index, the match index between the diagnosis target data group extracted from the monitoring operation data and the diagnosis reference data group extracted from the reference operation data. Can be evaluated appropriately.

(3) In some embodiments, in the configuration of (1) above,
The match index is the ratio of the number of i operation data (however, i ≦ N or M) belonging to both the diagnosis target data group and the diagnosis reference data group to the number of the N or M. Including.
According to the configuration of (3) above, the matching index is calculated as the ratio of the types of operation data common to the diagnosis target data group extracted from the monitoring operation data and the diagnosis reference data group extracted from the reference operation data. Therefore, the matching index can be evaluated appropriately.

(4) In some embodiments, in the configuration of (1) above,
When the matching index is given a point value that increases as the reference contribution or the monitoring contribution increases with respect to the operation data constituting the diagnosis reference data group or the diagnosis target data group, the matching index is described. It includes the sum of the point values for i of the operation data (provided that i ≦ N or M) belonging to both the diagnosis target data group and the diagnosis reference data group.
According to the configuration of (4) above, by using the point value for the calculation of the match index, the match index between the diagnosis target data group extracted from the monitoring operation data and the diagnosis reference data group extracted from the reference operation data. Can be evaluated appropriately. Since this point value increases as the contribution of the operation data constituting the diagnosis reference data group increases, it contributes more than calculating the matching index based on the type of operation data common to the diagnosis target data group and the diagnosis reference data group. It is possible to calculate a matching index that reflects the order of magnitude of degree.

(5) In some embodiments, in the configurations (1) to (4) above,
The degree of abnormality is the Mahalanobis distance.
The contribution is the SN ratio of the operation data.
According to the configuration of (5) above, the abnormal sign can be appropriately detected based on the Mahalanobis distance, and the diagnosis target data group and the diagnosis reference data group can be appropriately matched based on the SN ratio. Can be done.

(6) In some embodiments, in the configurations (1) to (5) above,
An operation data storage unit for storing the reference operation data is further provided.
According to the configuration of (6) above, the reference operation data can be reliably stored as information indicating an abnormality sign of an abnormality that has occurred in the past.

(7) In some embodiments, in the configurations (1) to (6) above,
The reference operation data is data indicating the operation state of another plant different from the plant from which the monitoring operation data was acquired.
According to the configuration of (7) above, the diagnosis is performed using the reference operation data indicating the abnormal sign of the abnormality that has occurred in another plant. That is, the reference operation data is shared between the plants, and it is possible to predict an abnormality that has already occurred in another plant but has not occurred in the own plant before the abnormality occurs.

(8) The plant abnormality monitoring method according to at least one embodiment of the present invention is
A plant abnormality monitoring method that monitors anomalies based on multiple operation data that indicate the operating status of the plant.
An abnormality sign determination step for determining the presence or absence of an abnormality sign in the plant based on the monitoring abnormality degree, which is an abnormality degree calculated based on the monitoring operation data including the plurality of operation data.
When the abnormality sign determination step determines that the abnormality sign is present, the monitoring operation data is compared with at least one reference operation data including the plurality of operation data indicating the abnormality sign acquired in the past. It is provided with an abnormality diagnosis step of performing an abnormality diagnosis of the plant based on the above.
The abnormality diagnosis step is
The contribution degree to the monitoring abnormality degree is calculated for each of the plurality of operation data constituting the monitoring operation data, and among the monitoring operation data, the top N items having a large monitoring contribution degree, which is the calculated contribution degree ( N represents an integer of 2 or more.) A diagnosis target data group extraction step for extracting a diagnosis target data group composed of the above-mentioned operation data, and a diagnosis target data group extraction step.
The top M items having a large reference contribution, which is the contribution to the abnormality calculated based on the reference operation data for each of the diagnosis target data group and the plurality of operation data constituting the reference operation data ( M represents an integer of 2 or more.) A matching index calculation step for calculating a matching index with the diagnostic reference data group composed of the operation data, and a matching index calculation step.
It has an abnormality prediction step of predicting the abnormality of the plant based on the reference operation data in which the match index is equal to or higher than the match determination threshold value.

According to the configuration of the above (8), the same effect as the above (1) is obtained.

According to at least one embodiment of the present invention, there is provided a plant abnormality monitoring system capable of predicting an abnormality event occurring in a plant before it occurs.

It is a figure which shows roughly the structure of the plant abnormality monitoring system which concerns on one Embodiment of this invention. It is a figure for demonstrating the reference operation data which concerns on one Embodiment of this invention. It is a figure for demonstrating the calculation method of the matching index which concerns on one Embodiment of this invention. It is a figure which shows the plant abnormality monitoring method which concerns on one Embodiment of this invention.

Hereinafter, some embodiments of the present invention will be described with reference to the accompanying drawings. However, the dimensions, materials, shapes, relative arrangements, etc. of the components described as embodiments or shown in the drawings are not intended to limit the scope of the present invention to this, and are merely explanatory examples. Absent.
For example, expressions that represent relative or absolute arrangements such as "in a certain direction", "along a certain direction", "parallel", "orthogonal", "center", "concentric" or "coaxial" are exact. Not only does it represent such an arrangement, but it also represents a state of relative displacement with tolerances or angles and distances to the extent that the same function can be obtained.
For example, expressions such as "same", "equal", and "homogeneous" that indicate that things are in the same state not only represent exactly the same state, but also have tolerances or differences to the extent that the same function can be obtained. It shall also represent the existing state.
For example, an expression representing a shape such as a quadrangular shape or a cylindrical shape not only represents a shape such as a quadrangular shape or a cylindrical shape in a geometrically strict sense, but also a concavo-convex portion or chamfering within a range in which the same effect can be obtained. The shape including the part and the like shall also be represented.
On the other hand, the expressions "equipped", "equipped", "equipped", "included", or "have" one component are not exclusive expressions that exclude the existence of other components.

FIG. 1 is a diagram schematically showing a configuration of a plant abnormality monitoring system 1 according to an embodiment of the present invention. FIG. 2 is a diagram for explaining reference operation data Dr according to an embodiment of the present invention. Further, FIG. 3 is a diagram for explaining a method of calculating the matching index Rm according to the embodiment of the present invention.

As shown in FIG. 1, the plant abnormality monitoring system 1 is a system that monitors an abnormality based on a plurality of operation data D indicating the operation state of the plant 9. For example, in a power plant such as a thermal power plant or a nuclear power plant having a commercial boiler, or in a plant 9 such as a waste incinerator or a chemical plant, a plurality of devices (equipment) constituting the plant 9 are used for operation control and abnormality monitoring. ) Is measured by installing multiple sensors, and the status of various devices including valves is monitored. The plant abnormality monitoring system 1 is a distributed control system (DCS) that controls the operation of various devices in addition to the measured values (hereinafter, appropriately sensor values) of these plurality of sensors and the above-mentioned on / off states. ), Control information such as the operation content of the operation panel is periodically acquired as operation data D, and abnormality monitoring is performed based on these plurality of operation data D.

For example, multiple devices that make up a thermal power plant include boilers, steam turbines (high-pressure turbines, low-pressure turbines, etc.), superheaters, reheaters, heat exchangers such as economizers, and steam sent to steam turbines. There are various devices (denitration device, electrostatic collector, etc.) for treating the exhaust gas discharged from the boiler, valves and dampers that adjust the flow rate and pressure of the turbine. The operation data D in this case is the temperature, pressure, and flow rate of the steam generated by the boiler, the vibration of the steam turbine, the number of revolutions, the opening degree of the valve and the damper, the temperature of the superheater and the reheater, and the like. ..

In the embodiment shown in FIGS. 1 to 3, as shown in FIG. 1, the plant 9 has an operation history creation function realized by a computer, and various types as described above acquired at each periodic timing. The operation history H is created for each set of the operation data D of the above. Further, the operation history H created in this way is sent to the plant abnormality monitoring system 1 after a predetermined number of operation histories H such as one set are generated or at predetermined time intervals. The operation history H may be received through the plant abnormality monitoring system 1 requesting the plant 9 to transmit the operation history H to the plant 9 on a regular basis. Further, the plant abnormality monitoring system 1 and the plant 9 may be connected via a wide area network (WAN), and the above-mentioned operation history H may be connected via a virtual dedicated line using IP-VPN technology such as MPLS-VPN. You may send and receive.

Next, the plant abnormality monitoring system 1 will be described in detail.
As shown in FIG. 1, the plant abnormality monitoring system 1 includes an abnormality sign determination unit 3 and an abnormality diagnosis unit 4. The plant abnormality monitoring system 1 is composed of a computer, and includes a CPU (processor) (not shown), a memory such as ROM and RAM, and a storage device 14 (external storage device). Then, the CPU operates (data calculation, etc.) according to the instruction of the program loaded in the memory (main storage device) to realize each of the above-mentioned functional units included in the plant abnormality monitoring system 1. Each of the above functional parts will be described.

In the following description, the degree of abnormality T and the degree of contribution C, which will be described later, will be described as the Mahalanobis distance (MD) and the SN ratio used in the pattern recognition technique called the MT method (Mahalanobis Taguchi method), respectively. In this MT method, a normal group is defined as a unit space based on normal multivariate data (here, a plurality of operation data D), and from the unit space of the target data (here, monitoring operation data Dt). Determine the abnormality by finding the distance (Mahalanobis distance). This makes it possible to comprehensively diagnose the plant 9 using only one index, the Mahalanobis distance. In addition, the MT method can detect anomalies earlier than the method of diagnosing whether or not various operating parameters are within the control value (within the abnormality determination threshold value La) before the damage to the equipment progresses. It is known as a method that can prevent or minimize damage to equipment. However, the present invention is not limited to the present embodiment. The degree of abnormality T may be calculated by another method based on multivariate data such as the K-nearest neighbor method as long as the abnormality sign S or the abnormality can be detected.

The abnormality sign determination unit 3 is the plant 9 based on the monitoring abnormality degree Tt which is the abnormality degree T (Mahalanobis distance in the present embodiment; the same applies hereinafter) calculated based on the monitoring operation data Dt including a plurality of operation data Ds. The presence or absence of the abnormal sign S in the above is determined. In the embodiment shown in FIGS. 1 to 3, as shown in FIG. 1, the plant abnormality monitoring system 1 monitors including a plurality of operation data D (operation history H) sent from the plant 9 side as described above. The operation data collecting unit 12 for acquiring the operation data Dt and the abnormality degree calculation unit 21 for calculating the monitoring abnormality degree Tt based on the monitoring operation data Dt acquired by the operation data collection unit 12 are provided. In the present embodiment, the operation history H generated by the plant 9 is transmitted to the plant abnormality monitoring system 1 every time it is created, and when the operation data collection unit 12 receives it from the plant 9, it receives the latest operation history H. The operation history H is transmitted to the interconnected abnormality degree calculation unit 21. The abnormality degree calculation unit 21 calculates the monitoring abnormality degree Tt by setting a plurality of operation data D included in the latest operation history H input from the operation data collection unit 12 as monitoring targets (monitoring operation data Dt). Then, the abnormality sign determination unit 3 is connected to the above-mentioned abnormality degree calculation unit 21, and acquires the monitoring abnormality degree Tt calculated by the abnormality degree calculation unit 21.

Further, in the present embodiment shown in FIGS. 1 to 3, the plant abnormality monitoring system 1 includes an abnormality detection unit 22, and the above-mentioned abnormality degree calculation unit 21, abnormality detection unit 22, and abnormality sign determination unit 3 monitor the abnormality. It constitutes a part 2. The abnormality sign determination unit 3 and the abnormality detection unit 22 are each connected to the abnormality degree calculation unit 21, and acquire the monitoring abnormality degree Tt calculated by the abnormality degree calculation unit 21. The difference between the two is that the abnormality sign determination unit 3 has the presence or absence of the abnormality sign S based on the monitoring abnormality degree Tt, such as comparison between the monitoring abnormality degree Tt and the threshold value for detecting the abnormality sign S (abnormality sign determination threshold value Ls). On the other hand, the abnormality detection unit 22 determines whether or not an abnormality has occurred due to the progress of the event corresponding to the abnormality sign S, the monitoring abnormality degree Tt and the threshold value for detecting the abnormality (abnormality determination threshold value La). ), Etc., which is a point to be judged based on the monitoring abnormality degree Tt. When an abnormality occurs, the abnormality degree T greatly deviates from the normal value (normal range) (see time t2 in FIG. 2A and FIG. 3A), but the abnormality determination threshold La has an abnormality degree T from the normal value. The deviant event is set to a detectable value, and the abnormality sign determination threshold Ls is determined by analyzing the transition of the abnormality degree T in the past abnormality cases (trouble cases) and determining the event (change) that can be said to be the abnormality sign S. Above, the event is set to a detectable value. Therefore, the absolute value of the abnormality sign determination threshold value Ls is smaller than the absolute value of the abnormality determination threshold value (| Ls | <| La |).

Then, as shown in FIG. 1, the abnormality detection unit 22 (abnormality monitoring unit 2) notifies the driver, the manager, or the like that an abnormality has occurred, such as issuing an alarm when the abnormality is determined. .. On the other hand, when the abnormality sign determination unit 3 (abnormality monitoring unit 2) determines that the abnormality sign S is present, the abnormality diagnosis unit 4 to be described next is notified. The abnormal sign determination threshold value Ls used for detecting the abnormal sign S is a value common to a plurality of types of abnormal sign S, or is set to detect a plurality of types of abnormal sign S collectively. In many cases, it is not clear what kind of abnormality the abnormality sign S detected by the part 3 is based on, and this is to predict (identify) the abnormality indicated by the abnormality sign S.

When the abnormality diagnosis unit 4 determines that the abnormality sign S is present by the abnormality sign determination unit 3 described above, the abnormality diagnosis unit 4 includes at least the monitoring operation data Dt and a plurality of operation data D indicating the abnormality sign S acquired in the past. The abnormality diagnosis of the plant 9 is performed based on the comparison with one reference operation data Dr. The above-mentioned reference operation data Dr corresponds to the abnormality sign S obtained through the analysis of the abnormality that has occurred in the past. When an abnormality occurs, for example, as shown in FIG. 2A, the degree of abnormality T (MD value) increases as the abnormality determination threshold La exceeds (FIGS. 2A and 3B). t2) Through the analysis of the abnormality, it is determined at which stage (time zone) of the abnormality degree T that changes with the passage of time, and the abnormality degree T indicating the abnormality sign S is calculated. The plurality of operation data Ds that have been used are stored as reference operation data Dr. In the embodiment shown in FIG. 2, an event occurs in which the MD value changes at a time t1 before the time t2 when the abnormality determination is made. Then, as a result of determining that the abnormality degree T between the time t1 and the time t2 indicates the abnormality sign S, the abnormality degree T at that time is calculated based on Ma operation data D in total. However, from among them, Mb (Ma ≧ Mb ≧ 2) operation data D capable of determining the abnormality sign S with higher sensitivity are extracted and stored as reference operation data Dr.

Then, the abnormality diagnosis unit 4 predicts what kind of abnormality event the detected abnormality sign S is through the abnormality diagnosis based on the above-mentioned reference operation data Dr. Specifically, the abnormality diagnosis unit 4 performs the above-mentioned abnormality diagnosis (abnormality prediction) based on the contribution C to each abnormality degree T of the plurality of operation data D used for calculating the abnormality degree T. This contribution degree C is an index showing which of the plurality of operation data Ds contributes to changing the abnormality degree T until the abnormality sign determination threshold Ls is exceeded, and contributes to the abnormality degree T. It is shown that the higher the degree C, the higher the operation data D, the more the detected abnormality sign S corresponds to the main factor. Therefore, if the contribution C (monitoring contribution Ct) related to the monitoring operation data Dt and the contribution C (reference contribution Cr) related to the reference operation data Dr match, the main factors of the abnormality sign S match. Become. Based on such findings, the present invention predicts the abnormality indicated by the detected abnormality sign S. In the embodiment shown in FIGS. 1 to 3, as shown in FIGS. 2 to 3, the SN ratio obtained by using an orthogonal array with respect to the MD value having the abnormality degree T is calculated as the contribution degree C. (See FIGS. 2 to 3).

More specifically, as a configuration for performing the above-mentioned abnormality diagnosis, the abnormality diagnosis unit 4 includes a diagnosis target data group extraction unit 5, a match index calculation unit 6, and an abnormality prediction unit 7.
In the embodiment shown in FIGS. 1 to 3, as shown in FIG. 1, the reference operation data Dr is stored in the storage device 14 included in the plant abnormality monitoring system 1 as an abnormality sign example. Further, the contribution degree C (reference contribution degree Cr) of each of the plurality of past operation data (reference operation data Dr) constituting the reference operation data Dr is calculated in advance when the reference operation data Dr is stored in the storage device 14. By doing so, it is stored in the storage device 14 together with the reference operation data Dr. In some other embodiments, only the reference operation data Dr may be stored in the storage device 14.

As shown in FIG. 3, the diagnosis target data group extraction unit 5 calculates the contribution C to the monitoring abnormality degree Tt for each of the plurality of operation data Ds constituting the monitoring operation data Dt, and among the monitoring operation data Dt. , The diagnosis target data group Gt composed of the top N operation data D having a large monitoring contribution Ct, which is the calculated contribution C, is extracted. The above N is an integer satisfying Na ≧ N> 2 when the number of operation data D constituting the monitoring operation data Dt is Na, and it is desirable to set the value so as to appropriately include the characteristics of the abnormality sign S. ..

The match index calculation unit 6 calculates the match index Rm between the diagnosis target data group Gt extracted by the diagnosis target data group extraction unit 5 and the diagnosis reference data group Gr. The diagnostic reference data group Gr is a higher-order M having a large reference contribution Cr, which is a contribution C to the abnormality degree T calculated based on the reference operation data Dr for each of the plurality of operation data D constituting the reference operation data Dr. It is composed of a piece of operation data D. The matching index Rm is an index that quantitatively indicates how much the diagnostic target data group Gt and the diagnostic reference data group Gr match. For example, the larger the value, the larger the diagnostic target data group Gt and the diagnostic reference data group. It is an index showing that Gr matches (there are many common items). The specific calculation method will be described later.

In the embodiment shown in FIGS. 1 to 3, as shown in FIG. 1, the match index calculation unit 6 is connected to the diagnosis target data group extraction unit 5 and the storage device 14, respectively, and the diagnosis target data group extraction unit 5 It is configured to calculate the matching index Rm between the diagnosis target data group Gt input from the storage device 14 and the diagnosis target data group Gt input from the storage device 14. Further, M is Mb, which is the total number of reference operation data Drs, and is the same as N, which is the number of operation data constituting the diagnosis target data group Gt (N = M = Mb). ).

The abnormality prediction unit 7 predicts an abnormality in the plant 9 based on the reference operation data Dr in which the match index Rm is equal to or higher than the match determination threshold value V. For example, the reference operation data Dr is associated with an abnormal event (trouble case) such as a water supply leak from the furnace wall pipe and stored in the storage device 14, and the match index Rm is equal to or higher than the match determination threshold V. The occurrence of an abnormal event in the operation data Dr may be output as a prediction result. In the embodiment shown in FIGS. 1 to 3, information on an abnormal event of one or more reference operation data Drs in which the match index Rm is equal to or higher than the match determination threshold value V is output from the abnormality prediction unit 7 together with the information on the match index Rm. Is displayed on a display device 16 such as a display.

Next, a plant abnormality monitoring method corresponding to the processing by the plant abnormality monitoring system 1 described above will be described with reference to FIG. FIG. 4 is a diagram showing a plant abnormality monitoring method according to an embodiment of the present invention. The plant abnormality monitoring method is a method of monitoring an abnormality based on a plurality of operation data D indicating the operation state of the plant 9, and as shown in FIG. 4, the abnormality sign determination steps (S1 to S3) and the diagnosis target data. It includes a group extraction step (S5), a match index calculation step (S6), and an abnormality diagnosis step (S4 to S7) having an abnormality prediction step (S7).
The plant abnormality monitoring method will be described according to the flow shown in FIG. The flow of FIG. 4 shall be performed every time the operation history H is acquired from the plant 9.

The abnormality sign determination step is performed in steps S1 to S3 of FIG. In the abnormality sign determination step, the abnormality sign S in the plant 9 is based on the monitoring abnormality degree Tt, which is the abnormality degree T (Mahalanobis distance in this embodiment) calculated based on the monitoring operation data Dt including a plurality of operation data Ds. This is a step of determining the presence or absence. In the embodiment shown in FIG. 4, the monitoring operation data Dt (operation history H) is collected in step S1, the monitoring abnormality degree Tt is calculated in step S2, and the presence or absence of the abnormality sign S is determined in step S3. Then, when it is determined in step S4 that there is an abnormality sign S, the abnormality diagnosis steps (S4 to S7) described below are executed, and when it is determined that there is no abnormality sign S, the flow of FIG. 4 is performed. To finish. The flow of FIG. 4 may be terminated after displaying the fact that it is normal on the display device 16. Since the abnormality sign determination step is the same as the process performed by the abnormality sign determination unit 3 described above, detailed description thereof will be omitted.

The abnormality diagnosis step is performed in steps S5 to S8 of FIG. The abnormality diagnosis step includes at least one monitoring operation data Dt and a plurality of operation data D indicating the abnormality sign S acquired in the past when the abnormality sign determination unit 3 determines that the abnormality sign S is present. This is a step of performing an abnormality diagnosis of the plant 9 based on a comparison with one reference operation data Dr. More specifically, the diagnosis target data group extraction step is executed in step S5, the match index calculation step is executed in step S6, and the abnormality prediction step is executed in step S7. The diagnosis target data group extraction step (S5) is the same as the process performed by the diagnosis target data group extraction unit 5 described above, and the match index calculation step (S6) is the same as the process performed by the match index calculation unit 6 described above. Yes, the abnormality prediction step (S7) is the same as the process performed by the abnormality prediction unit 7 described above. Therefore, the details of these steps will be omitted.

Then, in step S8, the diagnosis result of the abnormality diagnosis performed in the abnormality diagnosis steps (S4 to S7) is output and displayed on a display device 16 such as a display (see FIG. 1). As a result, the abnormality indicated by the abnormality sign S is notified to the operator and the manager of the plant 9.

According to the above configuration, the plant abnormality monitoring system 1 monitors the operation data D (monitoring operation data Dt) obtained through measurement by a plurality of sensors installed in the plant 9, and calculates based on the monitoring operation data Dt. When the abnormality sign S is detected based on the abnormality degree T (monitoring abnormality degree Tt, for example, Mahalanobis distance (MD)), the abnormality indicated by the abnormality sign S is diagnosed. Specifically, the contribution of a plurality of operation data D (monitoring operation data Dt) used for calculating the monitoring abnormality degree Tt to the monitoring abnormality degree Tt (monitoring contribution degree Ct. For example, when the abnormality degree T is MD, the SN ratio. .) Is calculated. Next, the top N (diagnosis target data group) having a large monitoring abnormality degree Tt and at least one reference operation data Dr accumulated as indicating the abnormality sign S of the abnormality that occurred in the past are calculated by the same method. The matching index Rm with the upper M (diagnosis reference data group Gr) of the obtained contribution C (reference contribution Cr. Past monitoring operation data Dt) is calculated. Then, the reference operation data Dr in which the monitoring operation data Dt and the match index Rm are equal to or higher than the match determination threshold value V is searched, and the occurrence of an abnormality corresponding to the abnormality sign S indicated by the obtained reference operation data Dr is predicted.

That is, matching between the currently monitored monitoring operation data Dt (diagnosis target data group Gt) and the reference operation data Dr (diagnosis reference data group Gr) indicating the abnormality sign S of the abnormality (trouble) experienced in the past, respectively. The abnormality is predicted from the viewpoint of the degree of contribution C to the degree of abnormality T, assuming that there is a high possibility that the abnormality sign S indicated by the matched reference operation data Dr has occurred. As a result, it is possible to predict an abnormal event (content) before an abnormality that has occurred in the past occurs, and it is possible to prevent a failure of the plant 9 by dealing with the predicted abnormality.

Next, some embodiments relating to the above-described method for calculating the match index Rm will be described.
In some embodiments, as shown in FIG. 3B, the match index Rm is the sum of the monitoring contributions Ct of the operation data D constituting the diagnosis target data group Gt or the diagnosis reference data group Gr (FIG. 3 (FIG. 3). In b), the total monitoring contribution Ct for i operation data D (however, i ≦ N, M) belonging to both the diagnosis target data group Gt and the diagnosis reference data group Gr with respect to Sa) (FIG. 3 (b). ) Includes the proportion of _Sp). In the embodiment shown in FIG. 3, as shown in FIG. 3 (b), the sum of the monitoring contribution Ct operating data D that constitutes the diagnostic object data group Gt as Sa, the matching index Rm, Rm = S _p ÷ Calculated in Sa. In some other embodiments, the match index Rm may be calculated by the same calculation formula, where Sa is the total of the reference contribution Cr of the operation data D constituting the diagnostic reference data group Gr. In some other embodiments, the match index Rm may be calculated by another calculation formula including the above calculation formula.

According to the above configuration, by using the contribution degree C to calculate the match index Rm, the diagnosis target data group Gt extracted from the monitoring operation data Dt and the diagnosis reference data group Gr extracted from the reference operation data Dr. The matching index Rm can be appropriately evaluated.

In some other embodiments, the match index Rm comprises the ratio of the number of i operating data Ds belonging to both the diagnostic target data group Gt and the diagnostic reference data group Gr to the number of N or M. .. That is, the match index Rm is calculated by an arithmetic expression including Rm = i ÷ N or Rm = i ÷ M. Each of the plurality of operation data D is a factor of the abnormality sign S, and in the present embodiment, the ratio of the factor of the abnormality sign S common to the diagnosis target data group Gt and the diagnosis reference data group Gr is obtained.

According to the above configuration, the match index Rm is used as the ratio of the types of operation data D common to the diagnosis target data group Gt extracted from the monitoring operation data Dt and the diagnosis reference data group Gr extracted from the reference operation data Dr. By calculating, the matching index Rm can be appropriately evaluated.

In some other embodiments, the match index Rm becomes a larger value as the reference contribution Cr or the monitoring contribution Ct is larger with respect to the operation data D constituting the diagnosis reference data group Gr or the diagnosis target data group Gt. When the point value is given, the total of the point values for i operation data D (provided that i ≦ N or M) belonging to both the diagnosis target data group Gt and the diagnosis reference data group Gr is included. The above point value is, for example, a ranking value of the SN ratio in which contribution degrees C are arranged in ascending order and numbers are assigned from 1 to the arrangement as shown in the table of FIG. 2 (b) or FIG. 3 (b). It may be the reciprocal.

In the embodiment shown in FIG. 3, as shown in FIG. 3B, the match index Rm is the higher M of the reference operation data Dr in the list of the SN ratio ranking of the monitoring operation data Dt constituting the diagnosis target data group Gt. It is calculated as the sum of all the reciprocals (point values) of the SN ratio ranking for each monitoring operation data Dt that has a column for the presence or absence of up to the number. More specifically, in FIG. 3B, at least the 1st, 2nd, 5th, and 7th rankings belong to both the diagnosis target data group Gt and the diagnosis reference data group Gr, so that the matching index Rm is , Rm = 1/1 + 1/2 + 1/5 + 1/7 + ...

According to the above configuration, by using the point value for calculating the match index Rm, the match index Rm between the diagnosis target data group extracted from the monitoring operation data and the diagnosis reference data group extracted from the reference operation data can be obtained. Can be evaluated appropriately. Since this point value increases as the contribution of the operation data constituting the diagnosis reference data group increases, the matching index Rm is calculated based on the type of operation data common to the diagnosis target data group and the diagnosis reference data group. The matching index Rm that reflects the rank of the degree of contribution can be calculated.

Next, an embodiment relating to other configurations relating to the plant abnormality monitoring system 1 will be described.
In some embodiments, as shown in FIG. 1, the plant anomaly monitoring system 1 further comprises an operation data storage unit 15 that stores reference operation data Dr. In the embodiment shown in FIG. 1, the operation data storage unit 15 is formed in the storage device 14. As a result, the reference operation data Dr can be reliably stored as information indicating the abnormality sign S of the abnormality that has occurred in the past.

However, the present invention is not limited to the present embodiment. In some other embodiments, another external system or the like manages the reference operation data Dr, and the plant abnormality monitoring system 1 (match index calculation unit 6 or the like) is referred to by another external system or the like. Data Dr or diagnostic reference data group Gr may be acquired.

Further, in some embodiments, the reference operation data Dr is of another plant (for example, plant B of FIG. 1) different from the plant 9 (for example, plant A of FIG. 1) from which the monitoring operation data Dt was acquired. It may be data indicating an operating state. In the embodiment shown in FIG. 1, the plant abnormality monitoring system 1 obtained the abnormality diagnosis of the abnormality sign S of the monitoring operation data Dt collected from the plant A through the analysis of the abnormalities that have occurred in the past in the plant A and the plant B. The reference operation data Dr is registered in the storage device 14 (operation data storage unit 15) and used for abnormality diagnosis. That is, the reference operation data Dr (trouble case) is shared by the plurality of plants 9.

According to the above configuration, the diagnosis is performed using the reference operation data Dr indicating the abnormality sign S of the abnormality occurring in the other plant 9. In other words, the reference operation data Dr (trouble case) is shared between the plants 9, and even if the abnormality has already occurred in another plant but has not occurred in the own plant, the abnormality is predicted before the abnormality occurs. can do.

The present invention is not limited to the above-described embodiment, and includes a modified form of the above-described embodiment and a combination of these embodiments as appropriate.
For example, in the embodiment shown in FIG. 1, the plant abnormality monitoring system 1 is commonly provided for a plurality of plants 9, but in some other embodiments, the plant abnormality monitoring system 1 is provided for each plant 9. May be provided respectively.

1 Plant abnormality monitoring system 12 Operation data collection unit 14 Storage device 15 Operation data storage unit 16 Display device 2 Abnormality monitoring unit 21 Abnormality calculation unit 22 Abnormality detection unit 3 Abnormality sign determination unit 4 Abnormality diagnosis unit 5 Diagnosis target data group extraction unit 6 Match index calculation unit 7 Abnormality prediction unit 9 Plant

H Operation history D Operation data Dr Reference operation data Dt Monitoring operation data S Abnormality sign T Abnormality Tt Monitoring abnormality La Abnormality judgment threshold Ls Abnormality sign judgment threshold C Contribution Cr Reference contribution Ct Monitoring contribution Gr Diagnosis reference data group Gt Diagnosis target data group Rm Match index V Match judgment threshold

Claims (8)

A plant anomaly monitoring system that monitors anomalies based on multiple operational data that indicate the operating status of the plant.
An abnormality sign determination unit that determines the presence or absence of an abnormality sign in the plant based on the monitoring abnormality degree, which is an abnormality degree calculated based on the monitoring operation data including the plurality of operation data.
When the abnormality sign determination unit determines that the abnormality sign is present, the monitoring operation data is compared with at least one reference operation data including the plurality of operation data indicating the abnormality sign acquired in the past. It is equipped with an abnormality diagnosis unit that performs abnormality diagnosis of the plant based on the above.
The abnormality diagnosis unit
The contribution degree to the monitoring abnormality degree is calculated for each of the plurality of operation data constituting the monitoring operation data, and among the monitoring operation data, the top N pieces having a large monitoring contribution degree, which is the calculated contribution degree ( N represents an integer of 2 or more.) A diagnosis target data group extraction unit that extracts a diagnosis target data group composed of the above-mentioned operation data, and a diagnosis target data group extraction unit.
The top M items having a large reference contribution, which is the contribution to the abnormality calculated based on the reference operation data for each of the diagnosis target data group and the plurality of operation data constituting the reference operation data ( M indicates an integer of 2 or more.) A matching index calculation unit that calculates a matching index with the diagnostic reference data group composed of the operation data, and a matching index calculation unit.
A plant abnormality monitoring system comprising an abnormality prediction unit that predicts the abnormality of the plant based on the reference operation data in which the match index is equal to or higher than the match determination threshold value. The matching index refers to i operations belonging to both the diagnosis target data group and the diagnosis reference data group with respect to the total monitoring contribution of the operation data constituting the diagnosis target data group or the diagnosis reference data group. The plant abnormality monitoring system according to claim 1, further comprising a ratio of the total of the monitoring contributions for data (where i ≦ N, M). The match index is the ratio of the number of i operation data (provided that i ≦ N or M) belonging to both the diagnosis target data group and the diagnosis reference data group to the number of the N or M. The plant abnormality monitoring system according to claim 1, further comprising. When the matching index is given a point value that increases as the reference contribution or the monitoring contribution increases with respect to the operation data constituting the diagnosis reference data group or the diagnosis target data group, the matching index is described. The plant according to claim 1, wherein the plant includes the sum of the point values for i of the operation data (provided that i ≦ N or M) belonging to both the diagnosis target data group and the diagnosis reference data group. Abnormality monitoring system. The degree of abnormality is the Mahalanobis distance.
The plant abnormality monitoring system according to any one of claims 1 to 4, wherein the contribution is the SN ratio of the operation data. The plant abnormality monitoring system according to any one of claims 1 to 5, further comprising an operation data storage unit for storing the reference operation data. The plant according to any one of claims 1 to 6, wherein the reference operation data is data indicating the operation state of another plant different from the plant from which the monitoring operation data was acquired. Abnormality monitoring system. A plant abnormality monitoring method that monitors anomalies based on multiple operation data that indicate the operating status of the plant.
An abnormality sign determination step for determining the presence or absence of an abnormality sign in the plant based on the monitoring abnormality degree, which is an abnormality degree calculated based on the monitoring operation data including the plurality of operation data.
When the abnormality sign determination step determines that the abnormality sign is present, the monitoring operation data is compared with at least one reference operation data including the plurality of operation data indicating the abnormality sign acquired in the past. It is provided with an abnormality diagnosis step of performing an abnormality diagnosis of the plant based on the above.
The abnormality diagnosis step is
The contribution degree to the monitoring abnormality degree is calculated for each of the plurality of operation data constituting the monitoring operation data, and among the monitoring operation data, the top N items having a large monitoring contribution degree, which is the calculated contribution degree ( N represents an integer of 2 or more.) A diagnosis target data group extraction step for extracting a diagnosis target data group composed of the above-mentioned operation data, and a diagnosis target data group extraction step.
The top M items having a large reference contribution, which is the contribution to the abnormality calculated based on the reference operation data for each of the diagnosis target data group and the plurality of operation data constituting the reference operation data ( M represents an integer of 2 or more.) A matching index calculation step for calculating a matching index with the diagnostic reference data group composed of the operation data, and a matching index calculation step.
A plant abnormality monitoring method comprising: an abnormality prediction step of predicting the abnormality of the plant based on the reference operation data in which the matching index is equal to or higher than the matching determination threshold value.

Images