JPH06123642A - Method and apparatus for analyzing abnormality of plant - Google Patents

Abstract

PURPOSE:To provide a method and an apparatus for diagnosing the abnormality of a plant to investigate the. fundamental cause of at least one failure generated in the plant system with the use of a failure mechanism model. CONSTITUTION:A plant abnormality diagnosing apparatus 1 is constituted of a data input/output device 2 which inputs/outputs data necessary for the plant abnormality diagnosing method and manipulation data by a user including the diagnosing result of abnormality, a database 3 for recording the data, if necessary, a data managing device 4 for managing data including registration, deletion and retrieval of data to the database 3, and a data operating/processing device 5 for executing the plant abnormality diagnosing method.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormality diagnosing method and an abnormality diagnosing apparatus for various components constituting a large-scale plant system.

[0002]

2. Description of the Related Art Conventional techniques for diagnosing abnormalities and maintenance of a plant include a technique for diagnosing deterioration of a specific device using a failure mechanism model, and reliability of the final transfer mode of degradation or failure by FTA (Failure Tree Analysis). There was a technology to evaluate the sex and a technology to estimate the cause of the failure phenomenon using a causal matrix.

A technique for diagnosing deterioration of specific equipment of a plant and formulating an appropriate inspection / maintenance plan by using the above failure mechanism / model is disclosed in the prior application of the applicant of the present application (Japanese Patent Laid-Open No. Hei 4-
285117). According to this technique, it is possible to predict a change in aged reliability of the entire device by modeling a deterioration / fault mechanism of a specific device. By predicting the reliability of this device over time,
It is possible to grasp changes in the reliability of equipment that depended on the inspection of equipment and parts and the replacement thereof, and it is possible to exert a great effect on the diagnosis of equipment deterioration and the formulation of appropriate inspection and maintenance plans.

Further, the reliability evaluation technique by FTA described above models a failure mechanism of the entire system composed of various devices by a failure tree analysis and its similar method, and a failure rate and a failure rate are assigned to each block corresponding to a failure mode. It gives the unreliability function. With this method, the reliability evaluation technique by FTA can evaluate the total reliability of the top event (final transmission mode of deterioration or failure) of each block of the component equipment by Boolean algebra or the like.

Further, the above-mentioned causal matrix modeling technique classifies the failure phenomenon mode and the cause mode of a specific device in advance, and gives the degree of their relevance as numerical values to the elements of the causal matrix created by them. Is. According to this method, the cause candidate can be estimated using the observed failure phenomenon mode and the model, and is applied to various abnormality diagnosis devices.

[0006]

Although the above-mentioned conventional plant abnormality diagnosing techniques exist, the occurrence of a failure phenomenon involves a probabilistic problem, and not only the failure phenomenon occurs due to a single failure mode. It has a complicated mechanism that various failure modes affect each other and various failure phenomena appear. Therefore, in order to investigate the cause of the abnormality, it is necessary to model the failure mechanism with extremely high accuracy and consider the influence between failure modes.

The abnormality diagnosis method for diagnosing the deterioration of the specific equipment of the plant using the failure mechanism model is effective for investigating the cause of the abnormality of the specific equipment. For example, in the case of rotating equipment such as a pump, as the wear of the bearing progresses, vibration due to eccentricity of the shaft increases, which in turn causes contact and galling between the impeller and the casing, and as a result, the casing or the impeller. Events such as chipping, cracking and breakage occur and progress. According to the above failure mechanism / model method, it was possible to accurately model the phenomenon of scraping, cracking, or breakage of the casing or impeller, and to accurately identify the cause of the abnormality. Further, if the object is piping, for example, if the flange bolts are loosened and this causes the piping to vibrate, resulting in wear and cracks of the seal ring, then according to the method of the failure mechanism model described above, Wear of seal ring
The complex effect such as cracking, that is, the effect transmission (interference effect) of the deterioration / failure mode could be accurately modeled, and the cause of the abnormality could be accurately identified. However, the occurrence of abnormality in the system of the plant is not always due to the single failure mode in the specific device as described above. For example, the piping is vibrated due to insufficient mounting of the piping support, the vibration of the piping causes damage to the components of the valve installed on the piping, and fragments of the components flow into the downstream pump, resulting in The seal part of the pump bearing may be damaged and fluid may leak. In this example, from the phenomenon of pump leakage by the method of the failure mechanism model described above, even if the damage of the bearing seal part can be evaluated as the cause event, the root cause, that is, the cause event of poor support of piping Cannot be identified.

On the other hand, the FTA method is
You can model the failure mechanism of the entire device,
A fault tree model can be used to analyze the above cases. However, since the FTA method is a method of developing the influence transmission based on one predetermined top event, the created model of the interference effect represents only the partial failure mode of the system / equipment. There is a case. In this case, in order to complete a failure model for the entire system / equipment without omission, it is necessary to set several top events and create a failure model.

Further, in the method based on the causal matrix, it is necessary to classify the failure phenomenon and the failure cause of a specific device in advance, but in general, the failure mechanism of the device is very complicated, and the failure mode is simply defined as the phenomenon and the cause. Since it is difficult to classify, there is a problem that accuracy is insufficient even if a model is constructed.

As described above, the current technology has not proposed a diagnostic method and a diagnostic device for investigating the root cause of an abnormal phenomenon occurring in a plant system and its components.
The investigation of the causes of abnormal events that occurred in plant systems and equipment relied solely on the experience and intuition of specialist engineers.

Therefore, an object of the present invention is to extend the abnormality diagnosis method based on the failure mechanism model for the specific equipment to the level of the plant system and the group of constituent equipments, and use the failure mechanism model to generate in the plant system. Another object of the present invention is to provide a plant abnormality diagnosing method and a plant abnormality diagnosing apparatus for investigating the root cause of at least one failure phenomenon.

[0012]

In order to achieve the above-mentioned object, a method of diagnosing a plant abnormality according to the present invention is a failure mode created by two-dimensionally arranging the failure modes of various components constituting a system of a plant. Substituting a numerical value indicating the directional influence transfer between the failure modes and the degree thereof into each element of the interference matrix, and using matrix calculation, a failure phenomenon from a plurality of failure modes that cause the failure of the constituent device is determined. The present invention is characterized by constructing a failure mechanism model showing a hierarchical influence transfer path leading to a plurality of failure modes that appear, and graphically representing the failure mechanism model in a hierarchical block diagram using a directed graph. .

Further, in the plant abnormality diagnosing method according to the present invention, the degree of relevance of the failure mode that causes the failure of the failure phenomenon is input by inputting or setting the failure mode that becomes the failure phenomenon using the failure mechanism model. Of the above, the certainty of the cause of the failure, and the temporal transition in which the failure mode appears as a failure phenomenon are analyzed, the cause of the failure phenomenon that occurred and the support information for maintenance measures are provided. is there.

The plant abnormality diagnosing device according to the present invention includes an information input / output device for inputting / outputting information necessary for the plant abnormality diagnosing method, operation information by a user, and an abnormality diagnosing result, and the information. A database which is recorded as required, an information management device which manages information including registration, deletion, and retrieval in the database, and an information processing device which carries out the plant abnormality diagnosis method. Is.

[0015]

According to the plant abnormality diagnosing method of the present invention, the failure modes of the components constituting the plant system are two-dimensionally arranged to create a failure mode interference matrix, and a failure occurs for each element of the failure mode interference matrix. Substitute a numerical value that indicates the directional influence transmission between modes and its degree. Next, the failure mode interference matrix is directly calculated between each failure mode by matrix calculation.
Convert to a reachable matrix that includes indirect transfer effects as elements.
Furthermore, the reachability matrix is hierarchized in the order of failure modes (elements) that are only affected, and the hierarchized reachability matrix is diagrammed in a hierarchical block diagram by a directed graph to obtain a visually recognizable failure mechanism model.

Since this failure mechanism model includes the direct and indirect influence transmission paths between failure modes of various constituent devices constituting the plant system and its degree as elements, the failure phenomenon of a specific device occurring in the plant system is dealt with. Thus, it is possible to estimate the failure mode of another component that causes the failure, and to visually recognize the failure cause and its degree of association with the hierarchical block diagram.

Further, the failure interference matrix can set a plurality of failure causes for a plurality of failure phenomena, and when a plurality of failure phenomena are set, a failure mode which is a common failure cause for these failure phenomena is weighted. This makes it possible to accurately identify the root cause of the plant abnormality.

In order to carry out the above-described plant abnormality diagnosis method, the plant abnormality diagnosis apparatus of the present invention records the information and the input / output apparatus for inputting / outputting information necessary for diagnosis, user operation information, diagnosis result and the like. Since it has a database, an information management device that manages the data in the database, and an information processing device that executes the abnormality diagnosis method, the user can easily set the numerical values of the elements of the fault interference matrix from the database and The failure mechanism model can be constructed by various modifications, the diagnosis result can be visualized by the hierarchical blocks to easily estimate the cause of the failure, and the diagnosis result can be registered in the database as necessary.

Further, since the plant abnormality diagnosing device according to the present invention has the input / output device for inputting the operation information of the user, the user analyzes the diagnosis result through the input / output device and determines the cause of failure. If necessary, it is possible to obtain support information useful for plant maintenance measures, such as the degree of association of failure modes, the certainty of failure causes, and the temporal transition in which failure modes appear as failure phenomena.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the plant abnormality diagnosing method and plant abnormality diagnosing apparatus according to the present invention will be described below. FIG. 1 shows an example of the configuration of a plant abnormality diagnosing apparatus for carrying out the plant abnormality diagnosing method of the present invention. The plant abnormality diagnosing device 1 is composed of a computer in order to deal with a huge amount of information processing of the abnormality diagnosing method of the present invention. Information input / output device 2 for inputting / outputting information, database 3 for recording information as necessary, information management device 4 for registering, deleting, and searching information in database 3, and plant abnormality diagnosis method of the present invention And an information processing unit 5 for executing The information input / output by the information input / output device 2 is
Equipment information, parts information, FMEA (failure mode influence analysis)
There are information, element data of the failure mode interference matrix, diagnostic conditions, diagnostic results, and the like. These pieces of information are recorded in the database 3 as needed. The information input / output device 2 has a color / graphics monitor device 6, a keyboard device 7, a mouse device 8, a floppy disk drive device 9, a kanji / graphics printer device 10, etc. for handling these information.

Next, the process flow of the plant abnormality diagnosis method of the present invention will be described. FIG. 2 shows a processing flow of the plant abnormality diagnosis method. Step 100 shows a step of starting the creation of a failure mechanism model for a plant system or equipment group for which a plant abnormality diagnosis is performed. In step 110, necessary information (numerical value of elements of failure mode interference matrix) is retrieved from the database 3 and set. Next, in step 120, step 1
It is examined whether the information obtained in 10 is appropriate as information used for diagnosis. If the information is not correct, step 13
At 0, necessary information is registered in the database 3 and the above process is repeated.

If the information obtained in step 110 is appropriate,
In the next step 140, the failure mode interference matrix is set. Next, in step 150, consider whether the elements of the failure mode interference matrix are appropriate. If the elements of the failure mode interference matrix are not suitable, then at step 160 the elements of the failure mode interference matrix are modified by the user. If the elements of the failure mode interference matrix are appropriate, matrix computation is performed by the information processing unit 5 in step 170 to create a failure mechanism model. Next, in step 180, it is confirmed to the user whether or not the failure mechanism model needs to be graphed. When graphing, a directed graph is created and displayed in step 190. Next, in step 200, it is confirmed whether or not the failure mechanism / model needs to be registered, and in the case of registration, it is registered in the database 3 via the information management device 4 in step 210. If the failure mechanism model is inappropriate and does not need to be registered, then in step 160 the elements of the failure mode interference matrix are modified and the above process is repeated.

Once the failure mechanism / model has been constructed in this way, the root cause investigation of the abnormalities in the plant system / device group in step 210 and thereafter is performed. Step 220
At, the plant failure mode is input and set in the created failure mechanism model. In step 230 for this abnormal mode, the failure mode associated with the path associated with the abnormal mode is searched for and extracted. Next, in step 240, the rank of the degree of association and the certainty factor of the failure mode that causes the failure are calculated. In this case, the diagnostic condition can be changed. Next, in step 250, it is confirmed whether or not it is necessary to list the failure modes of the failure causes, and in the case of listing, in step 260, the failure cause candidate list and the priority of the detailed observation event are displayed. Next, in step 270, it is confirmed whether or not to continue the observation. When the observation is to be terminated, the diagnosis result is registered in the database 3 in step 280, and it is confirmed whether or not the abnormality diagnosis is terminated in the next step 290. To do. If the observation is not ended, in step 300 and step 310, the information about the device on the cause candidate list and the device information by the patrol, the monitoring device robot, etc. are collected, and step 2
The processing from 20 is repeated. If the confirmation to end the diagnosis is obtained in step 290, the end processing of the diagnosis is performed (step 320). If not, in step 330, the diagnosis target, the information from the database 3, the elements of the failure mode interference matrix, Change the calculation conditions etc. and go to Step 100.
Alternatively, the processing from step 210 is repeated.

Next, an example in which the plant abnormality diagnosis method of the present invention is applied to a concrete plant system will be described with reference to FIGS.
Explained by. FIG. 3 shows a motor-operated valve 11, a pipe 12,
1 shows an example of a simple plant system having a pump 13 and transmission of its failure cause. In FIG. 3, the motor operated valve 11
Is a device that is the root cause of the failure, the pipe 12 is a device that interferes with transmission of the cause of the failure, and the pump 13 is a device in which a failure event is observed. In the example of failure cause transmission shown in FIG. 3, when a failure mode called seat leak occurs in the motor-operated valve 11, thermal deformation occurs in the pipe 12 due to this effect, and further thermal deformation of the pipe 12 causes poor support of the pipe 12 and the pump. Leakage of the shaft seal portion 13 occurs. Further, the poor support of the pipe 12 affects the seat leak of the electric valve 11, and the interference transmission of the above influence is repeated.

When the cause of the fault phenomenon of the plant system shown in FIG. 3 is analyzed by the plant abnormality diagnosis method of the present invention, first, the information input / output device 2 provides information such as the name of the system to be diagnosed and component equipment. Is searched from the database 3 and set, and then information for creating a failure mechanism model, such as FMEA information and failure mode interference matrix, is searched and set from the database 3. When the information retrieved and set from the database 3 is not appropriate, the user can input or correct the data via the information input / output device 2.

By the above operation, for example, the failure mode interference matrix shown in FIG. 4 is set and displayed on the color graphic monitor device 6 of the information input / output device 2. The row of the failure mode interference matrix in FIG. 4 shows the failure mode that causes the failure, and the column shows the failure mode of the failure phenomenon. Each element of the failure mode interference matrix indicates the degree of failure transfer (hereinafter referred to as failure transfer rate).

The numerical values of the respective elements of the failure mode interference matrix of FIG. 4 are examples of the failure transfer rates between the failure modes of the failure transfer of the plant system of FIG. Specifically, ID. N
O. The electric valve seat leak of 01-01-002 is I
D. NO. 01-01-001 motor operated valve malfunction (failure transmission rate 0.8), ID. NO. It affects the pipe deformation of 01-02-003 (fault transmission rate 0.7). Also,
ID. NO. The pipe deformation of 01-02-003 is ID.
NO. 01-02-005 piping support failure (failure transmissibility 0.8), ID. NO. It affects the pump leakage of 01-03-006 (failure transfer rate 0.7). Also,
ID. NO. The defective piping support of 01-02-005 is ID. NO. 01-01-002 motor-operated valve seat leak (failure transmissibility 0.7) and ID. NO. 01-02
-Transfect the pipe deformation of 003 (fault transmission rate 0.8).

The numerical values of the elements of the above failure mode interference matrix are
If there is already data in the database, it will be searched automatically.
Is displayed. In this case, the failure mode interference matrix can be obtained by performing only the correction / addition. In addition, if the retrieved / displayed failure mode interference matrix data is inadequate or does not exist, the user is free to input and correct it.

The numerical values (fault transferability) input / set to the elements of the failure mode interference matrix differ depending on the matrix calculation applied. The four methods of setting the numerical values of the elements of the failure mode interference matrix are outlined below.

(1) Method of Inputting Value of 0 or 1 In this case, matrix calculation is performed by ISM (Interpretiv).
e Structural Model) method is applied (see “Participatory Systems Approach”; edited by Yoshikazu Sasaki and Kazuhiko Kawamura; published by Nikkan Kogyo Shimbun). Here, element 0 indicates that there is no effect transfer between failure modes, and element 1 indicates that there is effect transfer between failure modes.

(2) Method of inputting any value between 0 and 1 The matrix calculation in this case is FSM (Fuzzy Struc
The “Tual Modeling” method is applied (“Structuring of social system by fuzzy theory”; Eiichiro Tasaki; separate volume “Mathematical Science: Road to Fuzzy Theory”; see Science Publishing). Here, the size of the numerical value of the element corresponds to the degree of influence transmission, and the value close to 1 indicates that the influence transmission is large. The example of FIG. 4 is based on this method.

(3) Method of setting degree of influence transmission by group questionnaire In this case, the presence or absence of influence transmission between two failure modes is 0 by a specialist in design and maintenance of the target plant system and equipment group. Do one questionnaire. Next, the average value of this questionnaire result can be calculated and the FSM method of (2) above can be applied. Alternatively, the average value of the above-mentioned questionnaire result is rounded to 4 and converted into 0 or 1,
The ISM method of (1) above can be applied.

(4) Method of setting degree of influence transmission with certainty factor According to this method, in addition to the element numerical values obtained by any of the above methods (1) to (3), the certainty factor Are also input and set. For example, a certainty factor of 0.3 is added to the failure transfer rate [0.8], and the failure transfer rate [0.8, 0.
3] is set. In this case, the effect transmission between the failure modes is relatively large (corresponding to 0.8), but the effect actually occurs and the possibility of transmission is relatively low (part corresponding to 0.3). It means that.

Next, regarding the example of FIGS. 3 and 4, the above
The application of matrix calculation by the ISM method in (1) will be described. In this matrix calculation, the values of the elements shown in FIG. 4 are rounded to 4 and converted into values of 0 or 1. The failure mode interference matrix shown in FIG. 4 shows only a direct relationship between two failure modes. That is, the motor-operated valve 11 and the pipe 1
2 and the direct relationship between the pipe 12 and the pump 13 are shown, but the indirect relationship between the motor-operated valve 11 and the pump 13 is not shown. In the ISM method, a failure mode interference matrix including only such a direct relationship is called an adjacency matrix. If this adjacency matrix is A and the identity matrix is I, then (A + I) ^r-1 ≠ (A + I) ^r = (A + I) ^{r + 1} = T under the Boolean algebra operation.
The reachability matrix T can be obtained by (reachability matrix). The reachability matrix T includes all the direct relationships of the adjacency matrix A and also includes the indirect relationships at the same time.

By setting the hierarchical structure of each element (fault mode) from this reachable matrix T, the hierarchical structure of the failure mechanism model (intermediate model) can be obtained. Furthermore, by calculating the adjacency matrix (skeleton matrix) of the reachability matrix T, it is possible to simplify the influence transfer path and construct a failure mechanism model. The process of simplifying the influence transmission path is performed by the route in which the failure mode A transmits the influence to the failure mode C via the failure mode B and the failure mode A.
Corresponds to a process of omitting the influence transmission route of the latter when there is a route transmitting the influence to the direct failure mode C at the same time. However, in this process, since the influence transmission path is omitted even when the actual physical phenomenon has transitivity, in the plant abnormality diagnosis method and the diagnosis device of the present embodiment, the effect exists in the original failure mode interference matrix. It is possible to revive the influence transmission path. By this operation, the obtained failure mechanism model can properly reflect the actual physical phenomenon.

The failure mode interference matrix and the failure mechanism model in each processing stage as described above are visually represented by a directed graph. Here, the directed graph includes a failure mode interference matrix and a failure mechanism on the color graphic monitor device 6 by blocks and arrows.
This is a model. The plant abnormality diagnosing method and the diagnosing apparatus of the present embodiment are configured so as to be able to create three types of directed graphs of a failure mode interference matrix, a failure mechanism model hierarchical structure (intermediate model), and a failure mechanism model. . FIG. 5 shows a directed graph of the failure mechanism model obtained from the failure mode interference matrix shown in FIG. FIG. 5 shows the components (motorized valve 1
1, the influence transmission direction and the influence transmission direction of the failure mode of the pipe 12, the pump 13) are shown for easy understanding.

In the plant abnormality diagnosing method and apparatus according to this embodiment, the failure mechanism model of FIG. 5 is automatically created from the failure mode interference matrix shown in FIG. 4 through the user's operation shown in FIG. Further, the constructed failure mechanism model is registered and stored in the database 3 as needed.

Next, a method for investigating the cause of failure using the above failure mechanism model will be described below. In the plant abnormality diagnosing method and the diagnosing device according to the present embodiment, for example, when "leakage from the shaft seal part of the pump" is observed by a patrol or inspection robot by a field maintenance worker or the like, it is displayed on the color graphic monitor device 6. Figure 5
A block indicating "pump leakage" on the failure mechanism / model is designated by using the mouse device 8 or the like. By this operation, the block display is reversed in black and white or colored. Further, by performing this operation a plurality of times, it is possible to set a plurality of failure observation events (fault modes).

For the failure phenomenon set in this way,
The information processing unit 5 lists the failure modes having a high degree of association and outputs them to the color / graphics monitor 6 and the Kanji / graphics printer 10. This list of failure modes has three types of output formats depending on the numerical values input to each element of the failure mode interference matrix. The three types of output formats will be described below.

FIG. 6 shows a failure mechanism model (FIG. 6) for a failure mode interference matrix consisting of 0 or 1 elements.
(A)) and a failure cause list (Fig. (B)) are shown.
In FIG. 6A, blocks A to J indicate failure modes, the failure phenomenon is shown in the upper part, and the cause of the failure is shown in the lower part. When there is influence transmission between failure modes, the element of the failure mode interference matrix is 1, and the blocks of the failure mechanism model of FIG. 6A are connected by lines.
On the other hand, when there is no effect transmission between failure modes, the element of the failure mode interference matrix is 0, and the blocks of the failure mechanism model of FIG. 6A are not connected by lines.
This failure mechanism model does not take into account the degree of influence transmission between blocks. In the failure mechanism model shown in FIG. 6A, when failure mode D is observed, failure modes A and B which are failure phenomena and failure modes F, G, I and J which are failure causes are shown in FIG. They are listed in parallel as shown in (b).

FIG. 7 shows a failure mechanism model (FIG. 7A) and a failure cause list (FIG. 7) for a failure mode interference matrix consisting of elements having any value between 0 and 1.
(B)) is shown. In FIG. 7A, blocks A to J indicate failure modes, the failure phenomenon is shown in the upper part, and the cause of the failure is shown in the lower part. If there is influence transmission between failure modes, the elements of the failure mode interference matrix are 0 to 1
7A, the blocks of the failure mechanism model in FIG. 7A are connected by a line, and the line indicates the degree of influence transmission. On the other hand, when there is no effect transmission between failure modes, the element of the failure mode interference matrix is 0, and the blocks of the failure mechanism model of FIG. 7A are not connected by lines.

In the failure mechanism model shown in FIG. 7A, when failure mode D is observed, failure modes A and B which are failure phenomena and failure modes F, G, I and J which are failure causes. 7 according to the order of their degree of association.
It is output as shown in (b). In this case, for example, the degree of association of the failure mode J is calculated as 0.5 × 0.7 = 0.35 and arranged in a predetermined order of the degree of association.

FIG. 8 shows a failure mechanism for a failure mode interference matrix whose elements are numerical values consisting of the degree of association and the certainty factor.
Model (Fig. 8 (a)) and failure cause list (Fig. 8 (b))
Is shown. In FIG. 8A, blocks A to J indicate failure modes, the failure phenomenon is shown in the upper part, and the cause of the failure is shown in the lower part. When there is an influence transfer between failure modes, blocks in failure modes are connected by a line, and a relevance of a value between 0 and 1 and a reliability of a value between 0 and 1 are attached to this line. Has been done.

In the failure mechanism model shown in FIG. 8A, when failure mode D is observed, failure modes A and B which are failure phenomena and failure modes F, G, I and J which are failure causes are Output is performed as shown in FIG. 8B according to the product of the degree of association and the degree of reliability. in this case,
For example, the order of the failure mode F is arranged in a predetermined order according to the size of the product 0.21 of the degree of association 0.3 and the degree of reliability 0.7.

Based on the failure cause list obtained in this way, additional inspections by the maintenance personnel of the plant and inspections by the inspection robot can be performed, and observation data of new equipment conditions can be collected. When a device status is newly observed, the cause of the failure can be narrowed down more accurately by inputting this observation event into the failure mechanism model.

In the example of FIG. 5, for example, when "pump leakage" is observed, "pump foreign matter mixing" is inferred as the first candidate of the failure cause, and "pipe deformation" is inferred as the second candidate. As a result of inspecting both failure modes based on the reasoning of the failure cause, if "foreign matter contamination of the pump" is not observed and "deformation of the piping" is observed, the failure mechanism model in Fig. Input "Transformation". Based on this new failure phenomenon, the plant abnormality diagnosing method and the diagnosing apparatus of the present embodiment can infer "the seat leak of the motor-operated valve" as the first candidate of the failure cause. That is, according to the plant abnormality diagnosing method and the diagnosing device of the present embodiment, it is possible to infer the indirect root cause, "seat leak of the motor-operated valve" from the failure phenomenon of "pump leakage". Thus, according to the plant abnormality diagnosis method and the diagnosis apparatus of the present embodiment, it is possible to construct a failure mechanism model for investigating the root cause of the abnormality of the plant system or equipment group, and by using this failure mechanism model, It is possible to investigate the cause of failure and formulate appropriate repair measures, improve the efficiency of plant inspection, and improve the ability to prevent plant abnormalities.

[0047]

As is apparent from the above description, according to the plant abnormality diagnosing method and the plant abnormality diagnosing device of the present invention, the degree of influence transmission between the failure modes of the components of the plant is quantified, and the failure mode interference is quantified. By creating a matrix and performing matrix calculation on this failure mode interference matrix, it is possible to construct a failure mechanism model necessary for investigating the cause of failure of an abnormal event in the plant system and component equipment group. The root cause event of the event can be evaluated. As a result, when a failure occurs in the plant system and equipment group, it is possible to accurately identify the cause of the failure and prevent the expansion of the abnormal event, and it is also possible to formulate an appropriate repair / maintenance plan. As a result, it is possible to improve the operating rate of the plant, optimize the maintenance plan, and improve the reliability of the entire plant.

[Brief description of drawings]

FIG. 1 is a block diagram showing a device configuration of an embodiment of a plant abnormality diagnosis device according to the present invention.

FIG. 2 is a block diagram showing a processing flow of a plant abnormality diagnosis method according to the present invention.

FIG. 3 is a block diagram illustrating transmission of failure modes of a plant system and its constituent devices.

4 is a diagram exemplifying a failure mode interference matrix for the plant system of FIG.

5 is a directed graph showing a failure mechanism model obtained from the failure mode interference matrix of FIG. 4;

FIG. 6 is a diagram showing a failure mechanism model obtained from a failure mode interference matrix having 0 or 1 elements and an output format of the failure cause list.

FIG. 7 is a diagram showing a failure mechanism model obtained from a failure mode interference matrix whose elements are values between 0 and 1 and an output format of the failure cause list.

FIG. 8 is a diagram showing a failure mechanism model obtained from a failure mode interference matrix whose elements are a set of a transmission influence degree and reliability, and an output format of the failure cause list.

[Explanation of symbols]

 1 Plant abnormality diagnosis device 2 Information input / output device 3 Database 4 Information management device 5 Information arithmetic processing device

Front page continuation (72) Inventor Akira Abe 8 Shinsita-cho, Isogo-ku, Yokohama-shi, Kanagawa Stock company Toshiba Yokohama office

Claims (3)

[Claims] 1. A directional influence transfer between the failure modes and its extent to each element of a failure mode interference matrix created by two-dimensionally arranging the failure modes of various components constituting the system of the plant. Substituting the numerical value indicating
While constructing a failure mechanism model showing a hierarchical effect transfer path from a plurality of failure modes that cause a failure of the component to a plurality of failure modes that appear as failure phenomena, the failure mechanism model is used by using a directed graph. A plant abnormality diagnosing method which is characterized in that it is diagrammed in a hierarchical block diagram. 2. By using the analyzed failure mechanism model to input or set a failure mode that is a failure phenomenon, the order of relevance of the failure modes that cause the failure of the failure phenomenon and the failure cause A method for diagnosing a plant abnormality, characterized by analyzing the certainty factor and the temporal transition in which a failure mode appears as a failure phenomenon, and investigating the cause of the occurring failure phenomenon and providing support information for maintenance measures. 3. An information input / output device for inputting / outputting information necessary for the plant abnormality diagnosis method, operation information by a user, and an abnormality diagnosis result, and a database for recording the information as necessary. An information management device that manages information including registration, deletion, and search in the database,
An apparatus for diagnosing plant abnormality, comprising an information processing unit for implementing the method for diagnosing plant abnormality.