JPH0390892A - Accident and fault relapse prevention assisting device - Google Patents

Abstract

PURPOSE:To make a systematic and inclusive judgement without depending upon experience by analyzing the cause of an accident or fault and determining a counter measure for prevention against a relapse according to information from an accident and fault data base and a planiprocess(PP) signal. CONSTITUTION:When the PP signal at the time of the occurrence of an accident or fault to a plant is inputted to a cause analysis part 4 through a user interface (UI) 2, the analysis part 4 retrieves a fault tree which describes the cause-effect relation of events in the accident and fault data base(DB) 3 to search for a chain of the same events with the PP signal and when there is the same event, the cause is inquired into; and a countermeasure determination part 5 selects a countermeasure corresponding to the cause stored in the DB 3 and offers it to a user through the UI2. If the chain of the same events with the PP signal can not be found, a similar event is retrieved and when the analysis part 4 judges that there is a similar event example, a regarding relapse prevention countermeasure is taken, but when it is judged that there is no similar event example, information on a relative event sequence, etc., is offered and an expert is requested to register the analysis result in the DB 3 through an event registration part.

Description

[Detailed Description of the Invention] [Objective of the Invention] (Field of Industrial Application) The present invention is designed to analyze the cause and take measures to prevent recurrence when an accident or failure occurs in a process plant such as a power plant or a chemical plant. Regarding the accident/failure recurrence prevention support equipment we provide.

(Conventional technology) Generally, in process plants, such as nuclear power plants, safety and reliability are taken into consideration from the design stage and necessary measures are taken, and sufficient operation management is carried out on a daily basis even during the operation stage. .

However, abnormal events such as equipment failures occur between various measures that are considered to be perfect. For such abnormal events, it is necessary to investigate the cause and take measures to prevent recurrence of accidents and failures, including similar events. In the case of nuclear power plants, which are growing as Japan's central energy source, preventing the recurrence of accidents and breakdowns is an extremely important issue when considering their social impact.

Traditionally, when an accident or breakdown occurs in a plant, experts in the relevant field would collect related plant data, investigate the cause, and decide to replace mechanical parts to prevent recurrence, all based on past experience. It's here.

(Problems to be Solved by the Invention) However, as plants have become larger and more complex, the number of devices that make up the plants has increased, and failure cases have become more diverse. For this reason, it has become increasingly difficult to make appropriate judgments based solely on the judgments that experts have traditionally made based on their experience.

The present invention has been made in view of the above circumstances, and its purpose is to analyze the causes of abnormal events that occur in plants, and to provide an accident/failure recurrence prevention support system that is equipped with measures to prevent the recurrence of accidents/failures. It is about providing.

[Configuration of the Invention (Means and Actions for Solving the Problem) In order to achieve the above object, the accident/failure recurrence prevention support device of the present invention has an accident/failure database storing accident/failure data and a process plant. a cause analysis unit that inputs plant signals from the plant and information from the accident/failure database and analyzes the causes of accidents/failures occurring in the plant; and causes of accidents/failures from the cause analysis unit and the accident/failure database. a countermeasure determining unit that inputs information from the computer and determines countermeasures to prevent recurrence; an event registration unit that registers new accident/failure cases in the accident/failure database; the cause analysis unit; and the countermeasure determining unit. and a user interface that is common to the event registration section and that inputs and outputs data with the user.

Therefore, according to the accident/failure recurrence prevention support device of the present invention, the causes of abnormal events occurring in the plant are analyzed,
It is possible to provide measures to prevent recurrence of accidents and breakdowns.

(Example) Embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a functional configuration diagram of an embodiment of the present invention, which will be explained below by taking a nuclear power plant as an example.

As shown in Figure 1, instructions are received from a user 1 via a user interface 2 (e.g. CRT/keyboard), and process signals (e.g. parameter trends, equipment operation) from the nuclear power plant at the time of an accident or failure are received. (changes in status, etc.) are input to the cause analysis section 4. The cause analysis unit 4 analyzes the causes of accidents and failures based on the data stored in the accident and failure database 3. The accident/failure database 3 includes, for example, a fault tree that logically expresses the causal relationship of events caused by accidents/failures,
It stores information such as countermeasure methods and procedures in the event of an accident or failure, the importance of the event, changes in plant processes related to the event (trends and event sequences), plant system diagrams and control circuit block diagrams related to the event, etc. Faultry is created based on accidents and failure cases that have occurred in the past, but in order to realize adaptation to events that have never been experienced before, in addition to experience-based events, it also uses plant design information (e.g. equipment configuration, etc.). It also includes "design-based events" obtained from connection relationships, interlocks, etc.).

The cause analysis unit 4 searches for a fault tree that describes the causal relationship of events in the accident/failure database 3, and searches for a chain of events that is the same as the plant process signal. If the same chain of events as the input plant process signal is searched, the cause of the accident/failure has been found, and in this case, the countermeasure determining unit 5
A countermeasure (for example, replacement of related parts, improvement of maintenance procedures, etc., including prevention of recurrence) corresponding to the cause stored in the user interface 2 is selected and provided to the user 1 via the user interface 2.

If the same chain of events as the plant process signal cannot be found in the accident/failure database 3, a search for similar cases is performed. The purpose of searching for similar cases is to search the accident/failure database 3 for events that are very similar to the chain of plant process signals, and is performed, for example, by the method described below. In other words, the fault tree stored in the accident/failure database 3 expresses the causal relationship of various events leading to the occurrence of the top-level event (for example, reactor scram, main pump stoppage, etc.) using logical elements such as ANDloR. This is what I did. Therefore, the event corresponding to the lowest level on the fault tree (the highest level in terms of the flow of events) can be said to be the causative event that causes the highest level event to occur. Generally, there are multiple event sequences that cause the highest event on a fault tree, and each event sequence is composed of multiple events based on causal relationships.

By the way, the event sequence that may cause the top event (T) is Sji (1:1.N, where N:T
(the total number of event sequences included in St), and the events that make up St are Et+7. That is, 5ji=E eye 1 umbrella E
ti2...ENm...E jiM (however, the total number of events forming the Mis mouth) Each Etii has a score G11m for calculating the similarity. Furthermore, consider a logical variable LIiI11 that indicates whether Elim has occurred. That is, Llim = 1 (if Ejim occurs) = O
(If Elim does not occur) Based on the above preparations, the degree of similarity is calculated as follows.

All Slis in the database are compared with a given plant process signal sequence, and each Sli is
The degree of similarity pt+ with is calculated as follows.

P [1=ΣGjii umbrella Lfim/ΣG t im
(m=l, M) Based on the above calculations, it is determined that the event sequence (1) that gives the maximum value of Pti is the most similar. As is clear from the above calculation formula, O≦PI+≦1
Therefore, when P ti = 1, there is a complete match.

Since the realistic event sequence varies depending on the plant situation, generally Q<Pji<1,
A threshold value ε is set for determining similarity, and a case is determined to be similar when the maximum value of the Pth case is ≧ε.

If the cause analysis unit 4 determines that there is a similar case, it provides related countermeasures to prevent recurrence. If it is determined that there are no similar cases, information such as the event sequence related to the user 1 is provided via the user interface 2, the analysis is entrusted to an expert in the field, and the analysis results are sent to the event registration unit 6 to report the accident. ◆Register in the failure database 3.

The event registration unit 6 registers the event sequence analyzed by the expert in the database 3 in a fault-free format as a new accident/failure case.

Next, the flow of processing according to the present invention will be explained with reference to FIG.

When accident/failure information (process signals) is input to the cause analysis unit 4 when an accident/failure occurs in a plant (first step 11), the cause analysis unit 4 describes the cause-and-effect relationship of events in the accident/failure database 3. A search is made for a chain of events that is the same as the plant process signal (second step 12). If the same chain of events as the input plant process signal is searched (third step 13
), the cause of the accident/failure has been determined, and the countermeasure determining unit 5 selects a countermeasure corresponding to the cause stored in the accident/failure database 3 and provides it to the user 1 via the user interface 2. (6th step 16)
. The same chain of events as a plant process signal can lead to accidents and
If it cannot be found in the failure database 3, a similar case is searched (fourth step 14). If the cause analysis unit 4 determines that there is a similar case (fifth step 15)
), and provide related measures to prevent recurrence (Sixth Step 16). If it is determined that there are no similar cases, provide information such as the event sequence related to the user 1 via the user interface 2. Seventh step 17) Entrust the analysis to an expert in the field and register the analysis results as an event. Part 6
is registered in the accident/failure database 3 in a fault-free format (eighth step 18).

[Effects of the Invention] As explained above, the present invention has the advantage that experts can systematically and comprehensively analyze the causes of accidents and failures and determine measures to prevent recurrence based on their experience. ,
Effective in preventing recurrence of accidents and breakdowns in process plants.

[Brief explanation of drawings]

FIG. 1 is a functional configuration diagram of an embodiment of the present invention, and FIG. 2 is a diagram showing a processing procedure of the present invention. 1... User 2... User interface 3... Accident/failure database 4... Cause analysis section 5... Countermeasure determination section 6... Event registration section (8733) Agent Patent attorney Sho Inomata Akira (and others)
1 person) Figure

Claims (1)

[Claims] an accident/failure database storing accident/failure data; a cause analysis unit that inputs plant signals from the process plant and information from the accident/failure database to analyze causes of accidents/failures occurring in the plant; A countermeasure determination unit inputs accident/failure causes from the cause analysis department and information from the accident/failure database to determine measures to prevent recurrence, and registers new accident/failure cases in the accident/failure database. and a user interface that is common to the cause analysis section, the countermeasure determination section, and the event registration section and that inputs and outputs data with a user. Prevention support device