JPH01267799A - Abnormality diagnosing device - Google Patents

Abstract

PURPOSE:To support a proper operation and a proper maintenance by rearranging plural alarms in a cause-and effect generation sequence and fixing and displaying a first alarm corresponding to a first cause to cause the alarm. CONSTITUTION:When plural alarms are generated by an alarm detector 21, an alarm rearranging device 22 rearranges plural alarms in the cause-and-effect generation sequence based on data stored in a knowledge data base 25, decides the first alarm, and outputs the information to an alarm cause reader 23. The alarm cause reader 23 reads the candidate of the generation cause of the first alarm stored in a knowledge data base 26 beforehand, and outputs the information to a confident degree calculation . operation guidance reader 24. Thus, the contents of a generated abnormality can be properly decided, and the operation of the operator or the maintenance of a CE can be supported.

Description

[Detailed Description of the Invention] (Industrial Application Field) The present invention relates to an abnormality diagnosis device, and is particularly suitable for accurately and quickly identifying the cause when a stage number alarm is issued in a power generation plant or the like. The present invention relates to an abnormality diagnosis device.

(Prior Art) The following methods are available for inferring or identifying causes of abnormalities in various systems, plants, etc. using a so-called knowledge base.

(1) As described in Japanese Patent Application Laid-Open No. 57-707,
When an abnormality occurs and its cause and situation are unknown, the operating methods and procedures for correcting the abnormality are determined by computer processing based on knowledge base, improving safety and economic efficiency when an abnormality occurs. something that attempts to achieve

(2) As described in Japanese Unexamined Patent Publication No. 14303/1983, a group of cause-effect rules, a group of effect-cause rules, and a group of cause estimation rules are prepared in advance as a knowledge base, and when an abnormality occurs. Second, it attempts to perform system failure diagnosis based on each of the above-mentioned rule groups and data indicating the operating status of the system.

(3) As described in Japanese Patent Application Laid-Open No. 62-49408, when an abnormality occurs due to multiple causes, the accuracy of the combination of multiple causes that can cause the abnormality (
The reliability of the system is calculated from the average value of the reliability of each cause, and system failure diagnosis is performed based on the reliability of the combination of the plurality of causes.

In the conventional technologies described above, when multiple alarms are issued in a short period of time, the certainty of the cause of each alarm is independently given, and failure diagnosis is performed based on the average value or combination of these. However, the cause-and-effect relationship between each warning was not considered at all.

(Problem to be solved by the invention) Generally, when a certain abnormality occurs in a plant etc. and an alarm is issued, multiple alarms may be issued due to one abnormality. The exact order of occurrence is not necessarily suitable for understanding the true cause.

In other words, an alarm is generally generated by converting a continuous quantity into an on/off signal using a threshold value, and is generated when the continuous quantity exceeds the threshold value, but the order in which the threshold value is exceeded does not necessarily indicate the occurrence of an abnormality. Because it does not match the order or cause of the cause, it is not possible to accurately determine the cause unless the causal order is known.

For example, in a case where an alarm is issued saying "vibration hole" in the target plant, and then an alarm saying "exhaust chamber temperature high" is also issued within a short time, "Due to the increase in vibration, the exhaust chamber It is not always the case that the temperature in the exhaust chamber has increased, but a situation may also occur in which the vibrations have increased because the temperature in the exhaust chamber has increased.

Therefore, even if the certainty factor of the cause of the "vibration hole" is calculated in such a case, accurate failure diagnosis cannot be performed.

In other words, when an alarm is issued, whether or not the first alarm that was issued in terms of time is the first cause that triggered the occurrence of the abnormality depends on other alarms that were issued along with that alarm. depends on the content.

Therefore, when an abnormality occurs due to the number of stages, first determine the order in which alarms occur as a cause-and-effect relationship, and
As a result, from a causal standpoint, the cause of the abnormality occurrence and its certainty were determined based on the alarm that was recognized to be the first cause that triggered the abnormality occurrence (hereinafter referred to as the first alarm). It is necessary to understand.

The purpose of the present invention is to solve the above-mentioned problems, understand the first alarm as a cause-and-effect relationship, investigate the cause of the first alarm, and support operations by operators or maintenance by maintenance personnel. It is in.

(Means for Solving the Problems) In order to solve the above-mentioned problems, the present invention provides a system based on the causal relationship of the warnings stored in the database when multiple warnings are issued simultaneously or within a short period of time. Then, the alarms for each ship are rearranged into an order of alarm occurrence based on causal relationships (hereinafter abbreviated as causal order of occurrence), and the first alarm corresponding to the first cause that triggered the plurality of alarms is generated. The feature is that it is equipped with a means for sorting alarms to identify and display them.

Furthermore, it is equipped with a certainty calculation means that requests the plant for the data necessary to calculate the certainty of the cause of the first alarm, and automatically calculates the certainty by taking in the data from the plant. The points are distinctive.

(Operation) According to the above configuration, when a plurality of alarms are issued, the first alarm can be identified by rearranging the plurality of alarms in a causal order of occurrence.

Therefore, it becomes possible to accurately determine the nature of the abnormality that has occurred.

Furthermore, since data regarding the cause of the alarm can be automatically obtained manually from the plant, it becomes possible to easily and quickly calculate the confidence level regarding the cause of the first alarm.

(Example) The present invention will be explained below with reference to the drawings.

FIG. 1 is a block diagram of a wage diagnosis device which is an embodiment of the present invention.

In the figure, the alarm detection device 21 detects the threshold value when some abnormality occurs in the target plant 20 and as a result, the value of the signal detected by the detector (not shown) exceeds the threshold value. A warning corresponding to the signal exceeding the threshold is issued, and the information is output to the warning rearranging device 22.

When the alarm detection device 21 issues a plurality of alarms, the alarm rearranging device 22 rearranges the plurality of alarms into a causal order of occurrence based on the data stored in the knowledge database 25 and determines the first alarm. Then, the information is output to the alarm cause reading device 23.

In other words, the order in which the alarms are issued by the alarm detection device 21 is the order in which the signal exceeds the threshold, that is, the order in which the alarms are issued based on time (hereinafter abbreviated as "temporal order of occurrence"). However, the order in which the alarms are issued in time is not necessarily This is not the order in which the abnormality occurs. Therefore, when multiple alarms are issued, the alarm sorting device RL 122 selects the first alarm, which is identified as the main alarm from a causal standpoint, based on data stored in the knowledge database 25 in advance. decide.

The alarm cause reading device 23 uses the knowledge database 26 in advance.
reading candidates for the cause of the first alarm stored in the
The information is calculated by the reliability calculation/operation guidance reading device 24.
Output to.

When candidates for the cause of the first alarm are read out, the operator inputs the data necessary to calculate the degree of certainty that the candidate can be the cause of the first alarm using the data input device 29. It is sent to the calculation/operation guidance reading device @24.

Furthermore, when a candidate for the cause of occurrence of the i-th alarm is read out, the certainty factor calculation/operation guidance reading device I\24 automatically sends all of the data necessary to calculate the certainty factor of the cause of occurrence to the plant 20. The confidence level may be automatically calculated based on the input data for both the table and the table requested.

In addition, only data that cannot be entered automatically due to the inability to be automatically converted into electrical signals by sensors, etc., or data that shows subtle signs of abnormality that only the operator can judge, must be input by the operator. Input device 29
Other data may be automatically supplied from the plant 20 to the reliability calculation/operation guidance reading device 24.

When the data necessary to calculate the certainty factor of the cause is collected as described above, the certainty factor calculation/operation guidance reading device 24 calculates the certainty factor based on the certainty factor calculation method stored in advance in the knowledge database 26. Do calculations.

When the certainty level of each cause is determined, the data editing device 2
In step 7, the above-mentioned first alarm, cause, confidence level of the cause, etc. are edited, and the information is output to the display bag g128.

Next, the functions and operations of each part will be explained in more detail while following the flow of information.

In the target plant 20, for example, when the vibration of a certain part exceeds a predetermined threshold, the alarm detection device 21 detects this and issues an alarm of "large vibration". Furthermore, within a short period of time, an alarm will be issued stating that the exhaust chamber temperature is 1°.

These two alarms, "high vibration" and "high exhaust chamber temperature", are sent to the alarm rearranging device 22, where they are rearranged into a causal order of occurrence based on data stored in the knowledge database 25.

The knowledge base stores procedures and methods for knowledge processing as software based on the knowledge of experts in the field and empirical know-how (Know-1 (OW). In other words, this knowledge The base is the knowledge of abnormality prediction/detection and abnormality treatment methods obtained through the past experience of the expert, If..., th
It is organized according to the format of en... and registered as rules.

Taking the case where alarms A to C are registered as alarms as an example, rearranging alarms using such a knowledge base is as follows.
For example, this is performed by a method as described below.

(1) The causal order of occurrence is alarm A, alarm B, alarm C.
How to set it in fixed order.

In other words, if alarm A and alarm B are issued, alarm A
is identified as the first alarm, when alarm B and alarm C are issued, alarm B is identified as the first alarm, and when alarm A and alarm C are issued, alarm A is identified as the first 'D alarm. A method of setting the causal order of occurrence of each alarm in a fixed manner, such as identifying the alarm.

(2) A method of relatively setting the causal order of occurrence.

That is, alarms A and II%! If alarm B was issued, alarm A was identified as the first alarm, and if alarm B and alarm C were issued, alarm B was identified as the first alarm, and alarm 1111A and alarm C were issued. In this case, identify alarm C as the first alarm,
If Alarm A, Alarm B, and Alarm C are all issued, Alarm B is identified as the first alarm. How to set the order of occurrence.

(3) A method in which the order (time) in which the alarm is issued is added to the method shown in (2) above.

In other words, even if Alarm A, Alarm B, and QJ Alarm C are all issued, if Alarm A is issued first in terms of time, Alarm B is identified as the first alarm, and the When alarm B is issued, alarm C is identified as the first alarm, and so on, the order in which each alarm was issued is taken into consideration for combinations of alarms that were issued simultaneously or within a short period of time. How to set the causal order of occurrence.

In this embodiment, based on the knowledge in the knowledge database 25, it is determined that the vibration has increased because the salinity of the exhaust chamber has increased. That is, "high exhaust chamber temperature" is identified as the first alarm.

When the first alarm is identified, the information is sent to the alarm cause reading device 23, and the alarm cause reading device 23 reads out the cause of the first alarm, “exhaust chamber temperature high”, from the knowledge database 26. .

FIG. 2 is a schematic diagram for explaining the method of calculating the reliability after the cause of the first alarm "exhaust chamber temperature high" is read out by the alarm cause reading device 23 as described above.

In this embodiment, three causes of the first alarm, ie, "exhaust chamber temperature high", are read out: strainer clogged, spray valve inoperable, and condenser tube clogged.

When the cause of the first alarm is read in this way, the certainty calculation/operation guidance reading device 24 calculates the certainty based on the function stored in the knowledge database 26. requests the plant 20 for data necessary for calculation.

In other words, if "strainer clogging" is read as the cause, the certainty calculation/operation guidance reading device i ¥24
In order to calculate the certainty that the strainer is clogged, data regarding the strainer differential pressure is requested from the plant 20.

When the data regarding the strainer differential pressure is obtained, the reliability calculation/operation guidance reading device 24 applies the data to the data table 8-1 that shows the relationship between the strainer differential pressure and the reliability stored in the knowledge database 26. Calculate the confidence level.

Similarly, in order to calculate the confidence that condenser tubes are clogged, the plant 20 is requested to know the temperature difference at the condenser inlet and outlet, and the obtained data is used to calculate the relationship between the temperature difference at the condenser inlet and outlet and the confidence. The confidence level is calculated by applying the data table 8-2 shown in FIG.

If the discharge pressure of the spray valve cannot be determined automatically, the data is determined using an appropriate means such as a monitor, the value is input by the operator from the data manual device 29, and the same is repeated thereafter. “Calculate the confidence that the spray valve is inoperable.

In this example, the strainer pressure difference is ΔP1, the temperature difference between the condenser inlet and outlet is ΔT1, the spray valve discharge pressure is ΔP2,
After applying each data to the data table, the confidence level of each cause was 0.9.0.0.0.2.

When the certainty of each cause is determined in this way, the above-mentioned first alarm, the cause of the first alarm, the certainty of the cause, etc. are output as an alarm analysis result to a display device 28 such as a display or a typewriter. Ru.

Further, information regarding the cause and the operation guidance according to the degree of certainty of the cause may be stored in the knowledge database 26 in advance, and the operation guidance may be displayed together with the cause.

FIG. 3 is a diagram showing the display contents of the display, in which the first alarm "exhaust chamber altitude", the cause of the first alarm, and its certainty are displayed.

Furthermore, as mentioned above, "Strainer clogged.

For the cause, the operation guidance ``Replace the strainer'' is displayed, and for the cause ``Spray valve not working'', the operation guidance ``Check the spray valve 2'' is displayed.

Further, if the degree of certainty of strainer clogging 2 is not very high in this embodiment, for example, about 0.2, the operation guidance "Check the strainer" is displayed.

In this way, in this embodiment, not only is the operation guidance displayed for each cause, but the content of the operation guidance changes depending on the degree of certainty of the cause.

It also displays the plant system diagram on the display.
As mentioned above, if parts requiring replacement or checking are clearly indicated by high brightness, Tτ/X different color display, flashing display, etc., abnormal locations can be easily understood.

FIG. 4 is a block diagram of another embodiment of the present invention, and FIG. 5 shows the contents displayed on the display in this embodiment. In this figure, the same reference numerals as in FIG. 1 represent the same or equivalent parts.

In this embodiment, the storage device 30 is configured to store various measurement data necessary for calculating the certainty factor of the cause measured in the plant 20 together with time (as a function of time).
It is distinctive in that it can be memorized.

That is, in the embodiment shown in FIG. 1, the reliability calculation is performed based on data at the time when the first alarm is identified and the reliability calculation/operation guidance reading device 24 requests data from the plant 20. However, in this embodiment, the data before the alarm is issued can also be called up, and the confidence calculation/operation guidance reading device 24 calculates the confidence based on the time course of the data and the data at that point shown in FIG. Since the degree is calculated, it becomes possible to further improve the accuracy of the confidence degree.

Furthermore, since a lot of valuable information is included in the graph showing the temporal transition of data, according to this embodiment,
Even in cases where the symptoms of an abnormality are ambiguous and the degree of certainty cannot be calculated accurately using only the information stored in the knowledge database 26, accurate failure can be determined by checking the time trends of the data by an expert. Be able to make a diagnosis.

Furthermore, regardless of whether or not the data was read out as the cause of the first alarm, if the time course of any data can be called up on the display upon request from the operator, it will be possible to diagnose faults when an alarm occurs. It will also be possible to support daily plant maintenance and inspections.

(Effects of the Invention) As is clear from the above explanation, according to the present invention, when an alarm indicating an abnormality in a plant is issued, detailed analysis results of the cause that triggered the alarm can be provided to the operator. In addition, data for calculating the certainty of the cause of the alarm is automatically supplied from the target plant, making it easy to In addition, it becomes possible to perform failure diagnosis quickly.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of one embodiment of the present invention. FIG. 2 is a schematic diagram for explaining a method of calculating certainty. FIGS. 3 and 5 are diagrams showing the contents displayed on the display. FIG. 4 is a block diagram of another embodiment of the invention. 20...Plant, 21...Alarm detection device, 22.
... Alarm sorting device, 23... Alarm cause reading device, 24... Confidence calculation/operation guidance reading device,
25.26...Knowledge database, 27...Data editing device, 28...Display device, 29...Data input device, 30...Storage device

Claims (8)

[Claims] (1) When an abnormality occurs in the target plant, an abnormality diagnosis device that detects the abnormality and issues an alarm simultaneously or for a short period of time with the first database that stores the cause-and-effect relationship of the alarm. when a plurality of alarms are issued, the plurality of alarms are rearranged in a causal order of occurrence based on the causality of the alarms stored in the first database, and the trigger for the plurality of alarms is An abnormality diagnosis device characterized by comprising: an alarm sorting means for identifying and displaying a first alarm corresponding to a first cause. (2) When a plurality of alarms are issued at the same time or within a short period of time, the alarm rearranging means is configured to sort the plurality of alarms into causal relations based on the causal relations among the alarms stored in the first database. 2. The abnormality diagnosis apparatus according to claim 1, wherein the abnormality diagnosis apparatus displays the abnormality in the order in which the abnormality occurs. (3) A patent characterized in that the invention further comprises: a second database storing the cause of the alarm; and an alarm cause reading means for reading out the cause of the first alarm from the second database and displaying it. An abnormality diagnosis device according to claim 1 or 2. (4) The feature further comprises confidence calculation means for collecting data necessary for calculating the confidence of the cause of the first alarm, calculating and displaying the confidence based on the data. An abnormality diagnosis device according to any one of claims 1 to 3. (5) The abnormality diagnosis device according to claim 4, wherein the data collection is performed by an operator inputting data into a certainty calculation means. (6) The collection of the data is automatically performed by the certainty calculation means requesting necessary data from the plant and importing the data from the plant. Abnormality diagnosis device described in section. (7) The abnormality diagnosing device according to any one of claims 3 to 6, further comprising storage means for storing data regarding the cause together with time. (8) Claims 3 to 3 further include operation guidance reading means for reading out and displaying operation guidance from the second database in accordance with the cause of the occurrence of the alarm and the degree of certainty thereof. The abnormality diagnosis device according to any of Item 7.