JP2000214924A - Plant abnormality monitoring device and abnormality occurrence location identification method - Google Patents

Abstract

(57) [Summary] [Problem] To compare qualitative changes of each process state value of a plant with a knowledge database on its propagation relations, and to easily identify an abnormal occurrence location even for abnormal events that gradually change over a long period of time. . SOLUTION: The plant abnormality monitoring device is a plant state measuring device for measuring a process value of each device in the plant.
And a data collection device 12 that converts a process state signal into a numerical signal and saves it in a time series as plant state data.
And an abnormality diagnosis device 13 for fixing an abnormality occurrence location based on a database storing knowledge of qualitative characteristics of each process state value over a long period of time, and a display device 14 for displaying the obtained diagnosis results in the form of illustrations and charts And

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a plant abnormality for diagnosing the location of an abnormality and the cause of the abnormality by inference based on a knowledge database in a large-scale plant including a plurality of devices and equipment such as a power plant and a chemical plant. The present invention relates to a monitoring device and a method for identifying an abnormality occurrence location.

[0002]

2. Description of the Related Art In a large-scale plant such as a power plant or a chemical plant, it is necessary to detect a local equipment abnormality at an early stage and to take appropriate measures to suppress the spread of the abnormality to the entire plant and to stop the entire plant. This is a very important issue in avoiding the situation. In the past, methods such as displaying process state quantities measured for monitoring equipment abnormalities on instruments and collecting them as measurement data were used.However, all of the huge process state quantities were operated. It is virtually impossible for personnel to constantly monitor.

For this reason, computer-dependent data collection has recently been adopted as a method that does not rely on humans. In a system mainly composed of this computer, a process state signal is given to the computer, and it is possible to configure a plant monitoring apparatus in which everything including data processing is automated. Furthermore, many plant abnormality monitoring apparatuses have been proposed today that diagnose the state of the plant at that time based on a process signal measured by the plant and support monitoring and diagnosis.

[0004] A typical example assumes an abnormality that may occur in a plant, and stores a typical transient change pattern of the state quantity of the plant in that case, and compares the pattern with the change pattern when an abnormality occurs. Thus, what kind of abnormality has occurred in the plant is identified. This result is displayed on the display of the display device, and it is possible to immediately start a countermeasure by clarifying the location where the abnormality has occurred.

[0005]

However, such a plant abnormality monitoring device has the following problems.
That is, when a change exceeding the set value of the alarm set individually for each process signal is detected, it is necessary to individually check the alarm for each signal, so that a great deal of labor is required to respond quickly. Need.

Further, for an event in which the process state value is sufficiently faster than the response speed of the control system and appears as a transient change exceeding the alarm set value, the change in the signal before the abnormality propagates to the whole. By examining the pattern, it was possible to identify the location where the abnormality occurred by the conventional abnormality diagnosis method.However, it was controlled within the response speed range of the control system, and did not exceed the set value of the conventional alarm. It is extremely difficult to identify the location where an abnormality has occurred for an abnormal event that changes to.

An object of the present invention is to collate a qualitative change in each process state value of a plant with a knowledge database relating to the propagation relation thereof, and to easily identify an abnormal occurrence location even for an abnormal event that gradually changes over a long period of time. An object of the present invention is to provide a plant abnormality monitoring device and an abnormality occurrence location identification method that can be performed.

[0008]

According to the present invention, there is provided a plant condition measuring device for measuring a process condition value of each device in a plant, and a given process condition signal is converted into a numerical signal.
A data collection device that stores time series as plant state data, an abnormality diagnosis device that identifies the location of an abnormality based on a database that stores knowledge of long-term qualitative characteristics of each process state value, and obtained diagnosis results And a display device for displaying an illustration and a chart in a table.

In the plant abnormality monitoring apparatus having the above configuration, the qualitative change of each process state value of the plant is collated with the knowledge about the propagation relation, and even in the case of an abnormal event that gradually changes over a long period, the location where the abnormality occurs can be easily determined. Can be identified.

Preferably, in the present invention, the abnormality diagnosis apparatus further includes a database in which knowledge of the cause of the abnormality and the causal relationship between the events is further stored.

In the plant abnormality monitoring apparatus having the above configuration, it is possible to estimate the cause of the abnormality from the information on the device stored in the database with respect to the candidate of the location where the abnormality has occurred.

Further, in the method for identifying an abnormality occurrence location according to the present invention, when identifying an abnormality that has occurred in an apparatus, the apparatus constituting the target system is divided into units in which the fluid flow rate and the differential pressure are measured and grouped. It is determined whether or not the flow balance is maintained for each of the grouped systems, and when the flow balance is maintained, a qualitative change pattern and a parameter of the abnormality diagnosis knowledge database are further obtained for each parameter. Is determined by comparing the qualitative propagation relationship between the two and the presence or absence of inconsistency. If the flow balance is not maintained or there is inconsistency in the propagation relationship, the system is identified as abnormal.

In such an abnormal location identification method, if an abnormality occurs in a target system, it is compared with a knowledge database relating to propagation relations, and an abnormal occurrence location that is gradually changed over a long period can be easily identified. It becomes possible to do.

Preferably, the method of the present invention compares the absolute values of the flow rates at the inlet and outlet of each group when checking the flow balance for each group.

Further, in the method of the present invention, when confirming the qualitative propagation relation for each group, it is desirable to connect the propagation relation between the process values in the normal state set in advance and the flow direction of the actual observation value. The qualitative change patterns are sequentially compared in order of to find contradictions.

Preferably, in the method of the present invention, when confirming a qualitative propagation relationship for each group, a propagation relationship between process values in a normal state and actual observation values set in advance for each diagnosis group are determined. Compare the qualitative change pattern of each observed value with the change pattern of the outlet flow as a reference to find contradictions.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (First Embodiment) First Embodiment of the Invention
An embodiment will be described with reference to the drawings. In FIG. 1, a plant abnormality monitoring device is a plant state measurement device 1.
1, a data collection device 12, an abnormality diagnosis device 13, and a display device 14.

The plant state measuring device 11 is a plant 10
Measure the process status value of the equipment inside. The data collection device 12 converts the transmitted process state signal into a numerical signal and stores it as time series data as plant state data. The abnormality diagnosis device 13 identifies an abnormality occurrence location from the qualitative change of each process state value, and estimates a candidate for the cause of the abnormality. The display device 14 displays the obtained diagnosis results in a chart.

FIG. 2 shows the abnormality diagnosis device 13 and the display device 1.
4 is shown in detail. The abnormality diagnosis device 13 includes an abnormality diagnosis processing unit 15, an abnormality diagnosis knowledge database 16, and an abnormality factor database 17. The abnormality diagnosis processing unit 15 identifies an abnormality occurrence location by comparing a change pattern of the process state signal observed by the abnormality diagnosis knowledge database 16 with a qualitative propagation relationship between the signals. FIG. 3 shows a configuration example of the abnormality diagnosis knowledge database 16.

Further, the abnormality diagnosis processing unit 15 estimates the abnormality factor by searching the observed signal behavior from the abnormality factor database 17 concerning the fixed abnormality occurrence location. FIG. 4 shows a configuration example of the abnormality cause database. The display device 14 includes a display processing unit 18 and a display 19.

The display processing unit 18 performs processing for illustration and charting of the numerical signal so that the diagnosis result transmitted from the abnormality diagnosis processing unit 15 can be visually discriminated. The display 19 displays the processing result in a chart or the like.

FIG. 5 is a flow chart showing the contents of processing in the abnormality diagnosis processing section 15. The system to be diagnosed is divided into groups in which the flow rate F and the differential pressure dP (= Pj-Pi) are measured in advance, grouped, and stored as a part of the abnormality diagnosis knowledge database 16. FIG. 6 shows the concept of grouping based on the parameter observation status of the system to be diagnosed.

The processing in the abnormality diagnosis processing unit 15 is as follows.
The flow rate balance (conservation of mass) is confirmed by comparing the absolute values of the inlet and outlet of each group (steps 1 to 3). Next, when the flow rate balance is maintained, a qualitative change pattern observed for each parameter in the diagnosis group and a qualitative propagation relationship between the parameters stored in the abnormality diagnosis knowledge database 16 are described. Collation is performed (step 4). FIG. 7 shows the concept of the propagation relation check.

Here, the observation values in the diagnosis group are sequentially confirmed from the upstream side to the downstream side of the flow in the order of connection on the system. For example, if the observed value of the pressure on the upstream side increases (decreases), and if the observed value of the flow rate on the downstream side decreases (increases), it is determined that there is a contradiction, and the observed qualitative change pattern is If there is no inconsistency with the propagation relation defined for the target system, the process proceeds to the confirmation of the next group (step 5).

If it is determined in the flow rate balance check (step 2) that the balance is not maintained, or if it is determined in the propagation relation check (step 4) that there is an inconsistency, the corresponding diagnosis group is used. Alternatively, the vicinity of the observation parameter is determined to be a candidate for an abnormal occurrence location and registered in a candidate list (step 6).

The above processing is sequentially performed from the upstream to the downstream of the target system, and when the processing has been completed for all the observation points (step 1), the abnormality factor is determined for all the candidates of the abnormality occurrence location registered in step 6. Database 17
Then, a corresponding signal behavior is searched from the information on the device stored in the device, and an abnormal factor is estimated (step 7).

Through the above steps, it is possible to identify the location where the abnormality has occurred and to estimate the cause of the abnormality.

According to the present embodiment, the qualitative change of each process state value of the plant is collated with the knowledge about the propagation relation, and the location where the abnormality occurs can be easily determined even for abnormal events that gradually change over a long period of time. Can be identified.

(Second Embodiment) Further, a second embodiment of the present invention will be described. FIG. 8 shows the concept of the propagation relation check of the present embodiment. Here, the propagation relation is confirmed based on the qualitative change pattern of the outlet flow rate of the diagnostic group with respect to the observation value in the diagnostic group. For example, when the observed value of the flow rate at the outlet of the diagnostic group increases (decreases), if the differential pressure within the diagnostic group increases (decreases), it is determined that there is no inconsistency, and the flow rate at the outlet of the diagnostic group is observed. When the value increases (decreases), if the differential pressure in the diagnosis group decreases (increases), it is determined that there is a contradiction. The processing other than the confirmation of the propagation relation is the first processing described above.
The processing is performed in the same procedure as that described in the embodiment.

According to the present embodiment, each process of the plant compares the qualitative change of the state value with the knowledge about the propagation relation, and easily determines the location of the occurrence of an abnormality even if the abnormality changes gradually over a long period of time. Can be identified.

[0031]

According to the present invention, a qualitative change in each process state value of a plant is collated with knowledge on its propagation relation, and the location of an abnormal event that is gradually changed over a long period can be easily fixed. be able to.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a plant abnormality monitoring device according to the present invention.

FIG. 2 is a block diagram showing details of an abnormality diagnosis device and a display device of FIG. 1;

FIG. 3 is a diagram showing an example of an abnormality diagnosis knowledge database according to the present invention.

FIG. 4 is a diagram showing an example of an abnormal cause database according to the present invention.

FIG. 5 is a flowchart showing processing contents of an abnormality diagnosis processing unit of the present invention.

FIG. 6 is a diagram illustrating a concept of grouping based on parameter observation states of a system to be diagnosed.

FIG. 7 is a view showing the concept of propagation relation check.

FIG. 8 is a diagram showing a concept of a propagation relation check according to the second embodiment of the present invention.

[Explanation of symbols]

 REFERENCE SIGNS LIST 11 Plant state measurement device 12 Data collection device 13 Abnormality diagnostic device 14 Display device 15 Abnormality diagnostic processing unit 16 Abnormality diagnostic knowledge database 17 Abnormality factor database 18 Display processing unit 19 Display

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Yoshihiko Uhara 8th Shinsugita-cho, Isogo-ku, Yokohama-shi, Kanagawa Prefecture Inside the Toshiba Yokohama Office (72) Inventor Hiroki Yamamoto 8th Shinsugita-cho, Isogo-ku, Yokohama-shi, Kanagawa A-term in the Toshiba Yokohama office of the company (reference) 2G075 AA01 BA03 CA02 DA04 DA05 EA05 FB07 FB09 FB13 FB16 FC05 FD01 FD09 GA34 5H223 AA02 BB01 EE06 FF06 9A001 BB06 HH03 JJ01 KK37 KK54 LL05 LL09

Claims (6)

[Claims] 1. A plant condition measuring device for measuring a process condition value of each device in a plant, a data collecting device for converting a given process condition signal into a numerical signal and storing the same in a time series as plant condition data; An abnormality diagnosis device for identifying an abnormality occurrence location based on a database storing knowledge of qualitative characteristics of a process state value over a long period of time, and a display device for displaying the obtained diagnosis results in illustrations and charts Plant abnormality monitoring device. 2. The abnormality diagnosis apparatus according to claim 1, further comprising a database storing knowledge of a cause of abnormality and a causal relationship between events in the device causing the abnormality, and estimating a cause of abnormality of the device from the knowledge of the database. The plant abnormality monitoring device according to claim 1, wherein: 3. In identifying an abnormality occurring in a device,
The equipment constituting the target system is divided into groups in which the fluid flow rate and the differential pressure are measured and grouped, and it is determined whether the flow balance is maintained for each of the grouped systems and whether or not the flow rate is different. When the balance is maintained, the qualitative change pattern of each parameter and the qualitative propagation relationship between the parameters of the abnormality diagnosis knowledge database are compared to determine whether there is any inconsistency, and the flow balance is maintained. An abnormal location identification method that identifies the system as abnormal when it is not set or there is an inconsistency in the propagation relationship. 4. The method according to claim 3, wherein comparing the absolute value of the flow rate at the inlet and the absolute value of the flow rate at the outlet of each group when checking the flow balance for each group. 5. Confirming a qualitative propagation relationship for each group, a qualitative change pattern is sequentially set in the order of connection in the direction of flow of a preset propagation relationship between normal process values and a flow of actual observed values. 5. The method according to claim 3, wherein a comparison is made to find a contradiction. 6. In confirming the qualitative propagation relation for each group, the pattern of change in the flow rate at the group outlet for each diagnosis group is determined for each diagnosis group with respect to the propagation relation between the normal process values and the actual observation value set in advance. 5. The method according to claim 3, wherein a comparison is made between the qualitative change pattern of each observation value and a contradiction.