JPS61228501A - Method for deciding treatment of plant abnormality - Google Patents

Abstract

PURPOSE:To improve the safety and the working efficiency of a plant by utilizing the knowledge of the interlocking method of an automatic safety protection system for an atomic or thermoelectric power plant, etc. and the knowledge of the physical causal relation of an abnormal phenomenon to produce a cause- effect chart of the abnormality and studying the cause of the abnormality and its present state and then estimating its future state. CONSTITUTION:The monitor data obtained from a plant 1 consisting of a large- scale system is supplied to an abnormality process deciding device 2 as well as an input device 7. The device 2 contains an alarm table 3, a knowledge base 4, a reasoning part 5 and a display device 6. The alarm signal, the signal showing the start/stop of devices and the output signal of the device 7 which monitors the analog quantity and detects the abnormality are supplied to the table 3. While the base 4 stores the knowledge showing the name and the importance degree of the alarm and the knowledge of the physical causal relation of the fault. The part 5 studies the cause of the fault and its present state and also estimates the future state of the abnormality to display them on the device 6.

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a method for deciding what to do in the event of an abnormality in a large-scale system such as a nuclear power plant or a thermal power plant.

[Technical background of the invention and its problems] In large-scale systems such as nuclear power plants or thermal power plants, many alarms are generated when an abnormality occurs, and many signals exhibit behavior different from normal times. Therefore, there was a problem in that the operator himself was confused and could not immediately determine the cause of the abnormality and appropriate measures to take in response to it.

The conventional solution to such an event is the third method.
Assuming all the abnormal events that may occur in the plant as shown in the figure, starting from the cause of each event, and combining the behavior of the plant data when the abnormality progresses, we calculate the cause-effect of the abnormality. Tree-like data connecting (Cause
Consequence T ree below COT
It has been proposed that a system is prepared in advance, and when an abnormality occurs, the cause can be identified by tracing OCT using plant data at that time.

However, with the conventional solution methods described above, it is difficult to list all possible causes of abnormalities, and if an abnormality that has not been taken into account occurs, it may be impossible to identify the cause; As the number of events increases, OCT to describe them becomes extremely complex and extremely difficult to create.

Moreover, the capacity to store this data has become enormous, and OCT
There have been problems such as difficulty in correcting errors in the system, changing OCT when there is a change in the plant configuration, adding OCT when a new abnormality is discovered, etc.

[Purpose of the Invention] The present invention was made in view of the above circumstances, and its purpose is to automatically provide a system equivalent to OCT using plant data when an abnormality occurs in a large-scale system such as a power generation plant. To provide a method for deciding what to do in the event of an abnormality in a plant, which improves plant safety and operation rate by estimating the cause, current status, and future of the abnormality and displaying the corresponding measures to the operator. There is a particular thing.

[Summary of the Invention] In order to achieve the above object, the present invention utilizes knowledge regarding the interlock of automatic safety protection systems in large-scale systems such as nuclear power plants or thermal power plants, and knowledge regarding the physical cause-and-effect relationship of abnormal phenomena. A diagram connecting the cause and effect of an abnormality is created from the plant data at the time of an abnormality using , and this diagram is used to estimate the cause of the abnormality, the current state of the abnormality, and the future state of the abnormality, and determine the corresponding measures. The present invention relates to a method for deciding what to do in the event of an abnormality in a plant.

The status of large-scale plants is detected by a large number of analog signals and digital signals such as alarms. In the present invention, for analog signals, we use an abnormality detection method that matches the characteristics of each signal based on safety analysis during plant design and operational experience (for example, predicting the current behavior from a physical model or statistical properties of the signal's past behavior). By determining the threshold value for determining normality/abnormality (comparison with the AR model, which is a model for The state of the plant is understood as a combination of signals expressed in binary values, using installed alarms and signals indicating the operation/stopping of equipment. (Hereinafter, all of these signals are referred to as alarms in a broad sense.) If an abnormality occurs, collect only the alarms that are "ON" and use knowledge about interlocking of automatic safety protection systems (Iji) to collect them. For the child Kaaka plant, knowledge such as ``If the main pump trips, the reactor will automatically scram'') and knowledge about the physical cause and effect of abnormal phenomena (``If the pump trips, the discharge pressure will drop.'' The most upstream alarm indicates the cause of the abnormality, the most downstream indicates the current situation, and the next possibility is that it will become "ON". A certain state is determined to be the future state.

As a processing guide to give to operators, an accident processing manual should be prepared that describes the cause of turning on each of the above-mentioned alarms and the corresponding processing. If the item is ranked at a level where it is possible to restore the plant to a normal state by performing the following actions, a manual for the alarm indicating the cause will be displayed.

Furthermore, if there is an alarm ranked in the importance level of the "ON" alarms at a stage where recovery from the abnormality is impossible and the plant needs to be safely stopped, this is displayed with priority. Furthermore, it predicts the next alarm that should be turned on by the automatic function, and checks whether the automatic function is operating normally. If there are multiple alarms indicating the cause, it is incomplete to determine whether it is a multiple abnormal event or the knowledge base, so this will be displayed and left to the operator's discretion, but later analysis will show that the knowledge base If any deficiencies are found, they will be corrected and correct inferences will be made regarding the same event the next time.

As described above, when an abnormality occurs in a plant, by providing appropriate guidance according to the degree of abnormality, it is possible to improve the safety and operation rate of the plant.

[Embodiments of the invention] Embodiments of the present invention will be described with reference to the drawings.

FIG. 1 shows a functional configuration diagram of an embodiment of the present invention.
Monitoring data obtained from the plant 1 is input to the abnormality action determining device 2 and other monitoring functions 7. This abnormality action determining device 2 has a warning table 3. Knowledge base 4. It is composed of an inference section 5 and a display device 6.

In alarm table 3, the original alarm signals provided on the control panel, signals indicating equipment operation/stop, and output signals of other monitoring systems that monitor analog quantities and detect abnormalities are summarized as 0N10FF. Create a table that describes the plant status using a collection of binary signals.

The knowledge base 4 stores knowledge about attributes such as the name and importance of each alarm, knowledge about logical connections between alarms by interlocking of an automatic safety protection system, and knowledge about physical causal relationships of abnormal phenomena. This knowledge base 4
The content of ``can extract the same information as the conventional OCT shown in Figure 3, but instead of describing the behavior of each signal in a fixed state connected to the cause as in the conventional OCT, it extracts the same information as the conventional OCT. Because it is structured as a collection of knowledge based on the conditions under which alarms are connected to other alarms, it is easy to create, and it is also easy to modify, change, and add after completion. .

The inference unit 5 reads the plant status from the alarm table 3, uses the knowledge base 4 to connect the alarms that are "ON", and immediately creates something equivalent to OCT every time an abnormality occurs to determine the cause and current situation. , estimate the future. Furthermore, by displaying something equivalent to this OCT, the reason for the inference is displayed, increasing the reliability of the operator.

Next, the algorithm according to the method of the present invention will be explained with reference to the flowchart of FIG.

First, the plant status is periodically fetched from the alarm table 3 in the first step 11, and in the second step 12, it is compared with the previously input alarm pattern to determine whether the automatic function is normal or not, and if it is normal, the next In the third step 13, it is determined whether a new alarm has occurred, and if no new alarm has occurred, the process returns to the first step 11 and the above-described procedure is repeated. Furthermore, if a new alarm occurs, the following inferences are made regarding it. In the fourth step 14, the most upstream and @downstream alarms are estimated using knowledge about interlocks for new alarms, and in the fifth step 15, the most upstream and most downstream alarms are estimated using knowledge about physical causal relationships. Then, in a seventh step 17, a treatment manual for an alarm indicating the cause or a treatment manual for safely shutting down the plant is displayed depending on the degree of abnormality.

Further, in the sixth step 16, the next alarm to be generated is predicted and stored. If there is a request for something equivalent to an automatically created COT as a result of the rough inference of an abnormality, it can be displayed as a reason for determination in the seventh step 17. Further, if the automatic function is not normal in the second step 12, an automatic function abnormality alarm is generated in the eighth step 18.

[Effects of the Invention] As explained above, according to the method for deciding what to do when an abnormality occurs in a plant according to the present invention, when an abnormality occurs in a plant, knowledge about the interlock of the automatic safety protection system and the physical cause and effect relationship of the abnormal phenomenon can be used. Search the knowledge base that is summarized as knowledge that can be easily created, modified, changed, and added using the knowledge related to OCT, automatically create something equivalent to OCT, and understand the cause of the abnormality and the current situation.

Since the future can be estimated and the corresponding measures can be displayed to the operator, it is possible to support the operator in dealing with abnormalities and to improve the safety and operation rate of the plant.

[Brief explanation of the drawing]

FIG. 1 is a functional configuration diagram of an embodiment of the present invention, FIG. 2 is a flowchart showing an algorithm according to the method of the present invention, and FIG.
The figure is a diagram showing tree-like data connecting causes and results of abnormalities according to a conventional method. 1... Plant, 2... Abnormal treatment determining device 3... Alarm table, 4... Knowledge base 5.
... Reasoning section, 6... Display section (8733)
Agent Patent attorney Yoshiaki Inomata (one idiot) Figure 1 Figure 2

Claims (1)

[Claims] Diagrams that connect the causes and effects of abnormalities from plant data during abnormalities using knowledge about the interlock of automatic safety protection systems in large-scale systems such as nuclear power plants or thermal power plants and knowledge about the physical cause and effect of abnormal phenomena. A method for deciding what to do in the event of an abnormality in a plant, characterized in that the cause of the abnormality, the current state of the abnormality, and the future state of the abnormality are estimated using this chart, and corresponding measures are determined.