JPS58182711A - Factor predicting device for fault of plant - Google Patents

Abstract

PURPOSE:To set automatically a hypothesis for the factor of a fault and at the same time to perform an inspection by an operator, by combining the factor predicting function for the factor of the fault with the function which extracts the data out of the record of the plant faults and feeds them in order of time. CONSTITUTION:The record of four types of data including a fault detecting data of a data base 1 is fed to a factor predicting device 3 consisting of a computer containing a specific software. At the same time, the operating state of a plant is recorded to a fault recorder 4 through a data fetching device 13 and then fed to a fault record extracting device 6 consisting of a computer containing a specific software together with the output of a set value recorder 5 which records the set value of the process quantity. Then an estimated plant data is delivered and fed to a factor estimating device 3. The factor is inspected for an estimated fault, and this result of inspection is displayed on a man-machine interface device with which an operator can operate.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an apparatus for estimating the cause of an abnormality in a plant, and in particular, when an abnormality occurs in a plant such as a thermal power plant, a nuclear power plant, or a large-scale chemical plant, the present invention estimates the cause of the abnormality and helps the plant operator. (Plan to assist in judgment and operation) 74 Pertaining to an ordinary cause estimation device.

Conventionally, the cause-effect relationship tree (Cause-C
A device that utilizes onsequenceTree (CCT) has been proposed. This device is a CC'l' that connects the cause-and-effect relationship of events that occur in the plant on the tree.
The system instantly analyzes the data captured online and estimates the cause of the abnormality, allowing operators to determine the cause of the abnormality.

However, with conventional equipment, even if you notice a possible cause of an abnormality other than what is displayed, you have to enter it into the equipment and estimate the cause of the equipment error in an interactive manner. I couldn't.

In addition, there are devices that use knowledge engineering in the medical field to estimate causes in an interactive manner. A device that automatically performs verification using data has not yet been realized.

The purpose of the present invention is to overcome the above-mentioned drawbacks of the prior art.
In addition to automatically setting a hypothesis for the cause of an abnormality when an abnormality occurs in a plant, operators can also set a hypothesis, and the hypothesis is continuously tested using *, t+ data. An object of the present invention is to provide a device for estimating the cause of plant abnormality.

Furthermore, in the process of setting and testing the above-mentioned hypothesis, another purpose is to
A plant where you can check the operating status of equipment!
I! 4. To provide a constant cooling factor estimation device.

In order to achieve the above-mentioned object, the present invention combines a function of estimating the cause of an abnormality and a function of extracting data from plant accident records and providing it in chronological order. The main point is that a hypothesis about the cause of R-loss is established and tested, and that setting this hypothesis makes it easier for operators to understand.

Hereinafter, embodiments of the present invention will be explained based on the details.

FIG. 1 shows a configuration diagram of a plant abnormality cause estimation device according to the present invention. In Fig. 1, numeral 1 is a database that contains data for outputting causes that are different from abnormality detection data, data for reasoning about phenomena that have occurred, and data that should actually occur but should actually occur. This is to record 4mm data which is enough to enumerate the phenomena that did not occur. This can be constructed by storing the data of the 4th era in the storage device of the computer.

Reference numeral 2 is a cause estimation device which inputs data regarding the operating status of the plant, searches the database 1, and determines the probable causes of abnormalities, the reasons for the phenomena that have occurred, and the causes that should have occurred but actually did not occur. This is a device to test for phenomena that did not occur and for inconsistencies in input data. It can consist of a total of n machines with specific software.

3 is a data capture device for providing a means to capture data regarding the operating status of the plant, and ¥f
It can be configured with a computer input device.

4 is an accident recording device that records data on the operating status of the plant, hypotheses on the presumed causes of abnormalities, reasons for phenomena that have occurred, phenomena that should have occurred but did not actually occur, and discrepancies between hypotheses. This is a device to record the presence/absence. This can be configured in the computer's storage device.

Reference numeral 5 denotes a setting value recording device for recording the setting values of the process, and this can be constituted by a storage device of a storage bag.

6 is an accident record extraction device, which is an accident record device IT4.
From the records of the set value recording device, find the time when the process quantity changed beyond the set value in [5] or the time when the operating state of the equipment changed, and calculate the operating state of the plant if each hypothesis at that time was correct. Assumed plant, calculated by 1 in chronological order,
This is a device that cannot be used to output data. This can be configured using a computer with specific software.

7 is an accident record entry device for entering the output of the cause estimation device 2 into the accident record device 4, and can be composed of a total of 3I machines having specific software.

8 displays the output of the cause estimating device 2 using the man-machine interface device 11, inputs a hypothesis regarding the cause of the abnormality inferred by the operator, and uses the accident record entry device 17 to input the output in the accident record device [4]. There is a device for this purpose. This can be configured with a computer terminal device.

Table 1 shows an example of the contents of the setting value i recording device 5, and Table 2 shows a part of the example of the contents of the database 2 in this embodiment.

Table 1 Table 2 In this example, the data of the four types are concentrated in one place so that they can be used in common, and the stored data is based on comparing data regarding the operating status of the plant with its condition section. ,
The format is such that if a match occurs, the probable cause of the abnormality and the phenomenon that should occur can be searched from the conclusion section. Next, in explaining the operation of the embodiment according to the present invention described above,
Explain the phenomenon that is assumed to have occurred. Figure 2 shows a conceptual diagram of input data to the plant abnormality cause estimation device. First, the plant low pressure condensate pump (LPCP) fails, the feed water flow rate decreases, the reactor water level falls, and the reactor water level L3 A scram occurred and the steam isolation valve (
Assume that the xis factor estimator is activated after MSIV) is closed. At this time, it is assumed that the LPCP failure is not input to the &4 wealth cause estimating device.

Here, LPCP failure is data that is input to the usual cause estimator if instrument failure is also desired to occur, but in order to explain the process of setting and testing assumptions according to the present invention, we will explain the actual Although the probability of occurrence (9) is low, we dare to assume multiple failures as described above. In addition, when an LPCP failure is input to the abnormality cause estimation device, the processing is performed as follows:
This process is the same as the process described below, except that loss of coolant accident J (LOCA) and "water supply system abnormality" are not set. In addition, changes in reactor pressure due to MSIV closure and activation of safety systems such as the high-pressure core staghorn system due to reactor water level L2 will be ignored in the explanation since they would be complicated to explain. In addition, regardless of the occurrence of an abnormality, the data acquisition device 3 is in an operating state, and no abnormality is detected due to operator instructions or warnings.
At the time the cause estimation device is activated, the accident recording device i[
Data as shown in Figure 4KM2 has been accumulated.

First, according to the flowchart of the processing algorithm of the accident record extraction device 6 shown in FIG.
works. That is, since there is no hypothesis being tested or a hypothesis to be estimated, and only unprocessed data regarding the operating status of the plant is stored in the accident recording device 4, the reactor water level is initially recorded in the set value recording device ( l
O) Search the accident recording device [4] for time type ◎ data that has changed beyond the set value L5 recorded in 5, and transmit the results to the cause estimation device i12. The contents of the transmitted data are shown in Figure 3-18.

Next, the cause estimating device 2 calibrates the contents of the transmitted data using the data in the data pace l.

In the cause estimating device, processing is performed according to the flowchart of the processing algorithm shown in FIG.

Here, the data conditions of Table 2 JI61 are satisfied,
Estimated cause! d2 outputs the presumed cause of the difference.

At this time, it can be determined that an abnormality has occurred since the condition part of the second clothing store 1 is satisfied. At the same time, by searching for the knotted part, you can find the cause ii5! can be output. That is, the data of the second part 1 is used for two purposes: to detect an abnormality and to output a hypothesis of the cause.

At this time, since the data of the second fiA1 is data for setting a hypothesis, it is marked so as not to be reused. FIG. 3-1b shows the output of the cause estimation device 2 at this point. This output is entered into the accident recorder 4 by the accident recorder (11) device 7.

Next, the accident record extraction device f, 6 is restarted, and although there is no hypothesis being tested this time, there are unprocessed hypotheses estimated at time to, so one of them is selected.
The data is transmitted again to the cause estimation device 2 as data at time t0. The contents transmitted at this time are shown in FIG. 3-2a. This time, the data in Table 2 I61 is not searched because it has already been applied, and only the data in Table 2 42 is searched.Since the conclusion part is included in the input data, the reason for the conclusion part is the condition part. will now be output. At this time, Table 2 A2
The data is used to explain the phenomenon that occurred. FIG. 3-2b shows the output contents at this time.

Next, when the accident record extraction device #6 is activated again, this time it is assumed that the structure being verified is correct, and the time 1 when the reactor water level has changed beyond L4. data is extracted and transmitted to the cause estimate f&2. The contents of the input and output to the cause estimation device 2 at this time are shown in 3rd LV-3a and 3rd LV-3b.

(12) The above 1st degree is time 1. ．． 1. It is also repeated in sequence. Figure 3-4a.

Figure 3-4b is for time t2, Figure 3-5a, Figure 3-
5b shows the contents of the input and output of the cause estimation device 20 for the case of time 13, respectively.

Time 1! t/(j? is the non-rich gas treatment system (SG'
The startup of l'81 is expected based on the data of the second clothes store 4 as a search result of Data Pace, but since it has not started according to the data collected from the plant, it is output as a failure. Here, the data in Table 2 A4 is used to enumerate phenomena that should have occurred but did not occur in the actual VAK. Similar processing is also performed at time 14.

When the 111th hypothesis has been tested up to the current time in the process described above, the second hypothesis is then tested, but this process is the same as that for the first hypothesis. Regarding the input and output to the cause estimation device, the case of time 1o is shown in Fig. 3-6a and Fig. 3-6b, and the case of time t1 is shown in Fig. 3-6.
Figure 7a, Figure 3-7bK, Figure 3 for the case of time tl.
8a and (13) are shown in Figure 3-8b, respectively. At time t2, the data of the second clothing store 3 is searched, and the dry well blur is predicted. According to the input data, the dry well pressure is not high, and unlike machines such as the 5GTS, there is no failure, so it is determined that there is a contradiction in the input data, and "inconsistency" is output as an accuracy evaluation of the input data. The output of the cause estimation device at this time is as shown in FIG. 3-8b.

Since the 21st Nth hypothesis was denied midway through, the accident record extraction value &6 determines that the testing of the second hypothesis has ended, and the third hypothesis is then tested. This process is exactly the same as in the case of the first hypothesis, so the explanation will be omitted.

When the above-described processing is completed, a display screen as shown in FIG. 4-1 is displayed to the operator through the man-machine interface device 8.

Furthermore, the operator who looked at the display deduced from his own experience that the cause was that one LPCP tripped, and determined that the cause of the accident at time 1o (14) was detected through the man-machine interface device 8 and the event recording device 7. As the 44th hypothesis of the cause, one LPCP trip can be suggested. The accident zero error extraction device 16 produces an output as shown in FIG. 3-9a for the unprocessed hypothesis, and is processed in the same way as other hypotheses, such as LPCP.
Comments such as the reactor water level falling too rapidly for a single trip are displayed to the operator. FIG. 4-2 shows the content displayed in this case.

As is clear from the above embodiments, according to the present invention, when an abnormality occurs in printing, a hypothesis regarding the cause of the abnormality can be displayed first. Second, it is possible to test that hypothesis and the hypothesis regarding the cause of the abnormality estimated by the operator. At this time, it is also possible to check the operating state of the +lj& device when each hypothesis is correct. Furthermore, in this embodiment, since the hypotheses are tested one by one in order, even if JIII of the rule for setting hypotheses is adjusted, the most probable hypothesis is tested first, so information regarding the cause of the abnormality that is most doubtful is checked first. can be displayed. This percentage value is determined by the cause estimation device 2 using a slow computer (15
) Especially when it comes to realization.

In addition, because it is an independent data base for cause estimation, setting values, etc. prepared in advance, it is easier to make corrections when changing plant production, and at the same time, one data can be used for multiple It can be used for the purpose. In this embodiment, as an evaluation of the accuracy of input data, intermittent evaluation such as "presence or absence of contradiction" is used, but continuous evaluation such as "probability that input data is correct" may be used. At this time, the accident record extraction device 6 may determine that the test is complete if the probability is less than a certain value based on a certain hypothesis.

Further, in this embodiment, data for abnormality detection, data for cause estimation, etc. are omitted, but they may be prepared separately. Also, these data are kept independent as a database l, for example, a program word Ii'OR'I')
rlF (conditional) conclusion using LA N's IF statement.
It would be better to embed it in the software for realizing the cause estimation device 12, as in the example shown in FIG.

Similarly, data regarding set values is also stored in the accident recording device (16
) It may be embedded in the software for realizing the configuration 6. Also, the abnormality cause extraction device may be realized using one Irt's machine, and a data machine for cause estimation and a data machine for extracting the accident reporter! ! 'tJlIso etc. may be realized using dedicated computers for each function. At this time, 11't3 for cause estimation
I made multiple units of I machine and 11 machines for extracting accident reporters,
Each hypothesis may be tested in parallel.

If the hypothesis is correct again, fcJj! When data regarding the operation of the plant at time 1 is sent from the accident record 4 extraction device 6 to the cause estimation device 12, data regarding each hypothesis may be sent completely separately at time III as in this embodiment, or
Some additional data may be provided so that the cause estimation device 2 can distinguish which hypothesis each data relates to, and may be sent at the same time.

1 As mentioned above, according to the present invention, the nuclear power plant 11
When an abnormality occurs in a plant such as a plant! Display the hypothesis regarding the cause of J4 wealth, and estimate the hypothesis and the operator! ! It is possible to provide a plant abnormality cause estimation device that tests the hypotheses regarding the four causes of wealth and checks the operating state of the machine when each hypothesis (17) is correct. This has the effect that operators can always clearly understand the cause of J4 wealth and the status of the plant and operate the plant safely.

[Brief explanation of drawings]

Figure 1 is a configuration diagram of the plant abnormality cause estimation device, the second cause is a conceptual diagram of input data to the plant abnormality cause estimation device, and the third
The figure is a conceptual diagram of input data and output data to the cause estimation device, FIG. 4 is a display screen diagram, and FIG. 5 is a flowchart of the processing algorithm of the accident record extraction device. FIG. 6 is a flowchart 1r8W of the processing algorithm of the cause estimation device. l...Data base, 2...Cause estimation device, 13
...Data import device, 4...Accident recording device, 5
...Setting value recording device, 6...Accident record extraction device, 7
...According to the accident record, X [W, a...Man-machine interface device. (18) Figure 1 Number 2 Figure 30-/L:L Figure 5-/4V, 3[
21-2 Arrow 73 Figure-2 Each Figure 3-E
Fig. 3 - 3'6'' 1 & 3 Fig. - Male Calculation 3 - 4 each 13 Fig. - 5 Good Stem 3 Fig. - μ m 3 Fig. - C Good Pill 3 Fig. - μ Fig. 3 - 7 Good Kaya 3 Mi - 74 No. Figure 3-80-Pole 30-3 Each Figure 3-Rotation ¥3m-94 Figure 5¥4 Kasumi

Claims (1)

[Scope of Claims] 1. means for recording plant data regarding the operating state of the plant; means for recording a hypothesis regarding the cause of abnormality in the plant; and data for outputting the hypothesis; an accident recording means having a means for calculating and outputting assumed plant data based on the plant data if it is correct; and =1+ outputted by the accident recording means.
The assumed plant data for each hypothesis is inputted for each hypothesis, and it is determined whether or not there is a contradiction in the content of the assumed plant data for each hypothesis. If there is a contradiction, the data for outputting the hypothesis regarding the normal cause of the 1viI period and each of the above hypotheses is determined. and a cause estimation means for outputting a hypothesis regarding the cause of the abnormality using the assumed plant data for each, the hypothesis of the cause of the abnormality outputted from the cause estimation means is recorded in the accident recording means, and the hypothesis of the cause of the abnormality is recorded in the accident recording means, Repeatedly search for a consistent hypothesis of the cause of the abnormality! +! A device for estimating the cause of plant abnormality. 2. Plant abnormality cause estimation according to claim 1, wherein the cause estimation means calculates the certainty of the plant data, tests a hypothesis regarding the abnormality cause, and outputs a test result. Dressing W. 3. The cause estimating means outputs the reason for the phenomenon that has occurred using data prepared in advance for explaining the phenomenon that has occurred. Plant abnormality cause estimation device. 4. The cause estimation means uses data prepared in advance to enumerate phenomena that should have occurred but did not actually occur. The plant abnormality cause estimating device according to item 1 of the scope of special PIFM search, characterized in that it outputs phenomena that have not occurred. 5. The accident recording means uses a pre-prepared set value for the process lid, and records (1) at the time when the recorded process lid exceeds the set value or when the equipment changes its operating state. Plan) 14. The ordinary cause estimation device according to claim 1, which calculates and outputs assumed graft data when the recorded hypothesis is correct. 6. The cause estimating means displays the output of the cause estimating means to the operator, inputs a hypothesis regarding the cause of the abnormality inferred by the operator, and outputs it to the accident means. - The plant abnormality cause estimation device according to claim 1, characterized in that it has a manno interface means.