JPH03145846A - Fault diagnostic method - Google Patents

Abstract

PURPOSE:To easily correct and add a fault diagnostic rule by forming a description form of a fault state storage table and a description form of a fault status table in a same format and comparing the table from a head area sequentially when the tables are compared. CONSTITUTION:A fault diagnostic rule 7a, a fault status table 7b, and a fault cause candidate table 7c are stored in a file capable of read/write to form a fault diagnostic intelligent database (DB) 7. When the fault diagnostic operation is started, a fault diagnostic section (deduction engine) 6 extracts a fault status data 104 from a fault state storage table 5, the fault diagnostic rule 7a from the fault diagnostic intelligent database 7, and diagnostic data 105 from the fault status table 7b and the fault cause candidate table 7c and applies fault diagnosis by the method where the fault state storage table 5 and the fault status table 7b are compared according to the diagnostic rule 7a.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to the configuration and processing method of a diagnostic system in an information network fault diagnosis system, the configuration of a diagnostic knowledge database, the structure of rules, a guidance method for operators, etc. .

[Conventional technology]

Fault diagnosis systems in information network management systems have conventionally only collected and stored information on faults occurring in networks, leaving maintenance personnel, operators, and the like to grasp fault conditions and diagnose faults. In such cases, it is difficult for beginners to understand the fault situation and isolate the fault, and even for experienced users, when a large amount of fault information appears due to the occurrence of a fault, it is difficult to understand and isolate the fault situation.
In some cases, it was difficult to isolate the fault.

Therefore, systems are being developed that allow even beginners to easily diagnose faults. First, as a first step, a method was adopted in which the order of fault isolation as described in the maintenance manual was programmed, and fault diagnosis was performed according to the fault messages that occurred. In this method, diagnosis is performed according to the order determined by the fault that has occurred, which leads to a fixed diagnosis in which the same judgment is always made depending on the fault, and is inflexible in that it cannot respond to changes in the network configuration.

For this reason, attempts are being made to apply knowledge engineering to fault diagnosis systems for information networks and incorporate the experiential knowledge of skilled operators and maintainers into the fault diagnosis systems. An example of this type of device is the literature, J.R. Fox, Concept Prototyping a Real-Time Switch Maintenance Expert System, International Conference on Communications 88, 1988.
Year, (J, R, Fax, “Concept Prot
otyping a RealTime 5w1tc
h Maintenance Expert S
system”.

International Conference
on Communications'88. p,p,
1450-1454. (1988)), the Be1lcore online fault diagnosis system for switching equipment
-RTES is known.

Further, as a method for diagnosing faults in systems other than information networks, for example, Japanese Patent Application Laid-open No. 175060/1983.

JP-A-62-212764, JP-A-62-22512
There is a description in Publication No. 7, etc.

[Problem to be solved by the invention]

The above conventional technology does not take into account the growth and development of the system, so when the configuration of the diagnosis target changes, the diagnosis rules must be revised, and it is not possible to immediately respond to this, such as modifying or adding failure diagnosis rules. Since no consideration is given to so-called knowledge learning or knowledge acquisition, it is not easy to modify or add knowledge, and no consideration is given to operator errors, etc., so information on the occurrence of failures arrives at the diagnostic system in chronological order. Therefore, there were problems such as the risk of erroneous diagnosis results depending on when the diagnosis was started.

An object of the present invention is to provide a method for easily modifying or adding fault diagnosis rules in a fault diagnosis system for an information network, and also to semi-automatically add fault diagnosis rules as the system grows and develops. It consists in realizing a method of acquiring knowledge.

Another object of the present invention is to realize a method that can semi-automatically modify fault diagnosis rules as the system grows and develops, in other words, a knowledge learning function that updates conventional knowledge. It is in.

Another object of the present invention is to advise the operator on when to give an instruction to start diagnosis, and to go even further, to automatically start the diagnosis. The goal is to make it possible to finally ascertain the

[Means to solve the problem]

In order to achieve the above object, the present invention uses the following means.

A fault status table that stores fault status such as fault information and test results of fault detection parts, a fault cause candidate table that stores candidates for fault causes that cause the phenomenon, and a combination of these fault status tables and fault cause candidate tables. A fault diagnosis knowledge database is used in which a plurality of diagnosis rules, each of which is considered in combination as a pair of diagnosis rules, is stored in a readable/writable file. The fault diagnosis system stores the fault status of a fault detection part in a fault status storage table from time to time, and performs fault diagnosis by comparing the fault storage status table and the fault status table.

In such a fault diagnosis system, a fault status storage table when a fault occurs due to an artificially generated fault at an arbitrary location is converted into a fault status table,
Expand your knowledge base by storing. In addition, when the correction processing function of the fault diagnosis knowledge database is filled in and the fault diagnosis result is judged, the fault cause candidate table at the time of fault occurrence is automatically corrected if necessary, and the knowledge base is updated by storing it. Fix it.

In addition, by focusing on the connection form of the devices that make up the network, and classifying the devices, we use a fault status table that represents a set of classified devices, so that the amount of memory does not need to be increased even when changing the network configuration. method.

Furthermore, evaluation indicators such as the number of uses for the entire diagnostic rule, the number of uses for the fault cause candidate table, and the number of diagnosis matches (confidence) for the fault cause candidate elements in the fault cause candidate table are provided, and when the rule is used, these Make corrections and additions to
The method was to search for rules and present candidate causes based on the index.

Furthermore, an evaluation index for the degree of agreement between the fault status storage table and the fault status table belonging to each diagnosis rule is provided for each rule and table, and system operators are provided with guidance on when to start fault diagnosis and This is a method to deal with misdiagnosis due to automatic execution or errors in the timing of the system operator's failure diagnosis start operation.

[Operation] In the fault diagnosis system according to the present invention, the description format of the fault status storage table and the description format of the fault status table are the same format, and both tables can be modified by simply rewriting the description framework of these formats. When comparing tables in , a method is used in which comparisons are made sequentially starting from the first area, so there is no possibility of malfunction.

In addition, when classifying devices by focusing on their connection configuration, we create a correspondence table between the failure situation storage table and the failure status table, and use this as an intermediary to compare the tables, which prevents malfunctions. There is no.

Evaluation indicators such as the number of uses in the diagnosis rule, the number of uses in the failure cause candidate table, and the number of times diagnosis matches (confidence) are set for failure cause candidate elements in the failure cause candidate table, and corrections and
Since a processing unit that performs additions is provided to enable corrections and additions, and the search for rules and presentation of cause candidates are performed based on the index, malfunctions will not occur.

〔Example〕

The present invention will be explained below using examples.

FIG. 1 shows a configuration example of a fault diagnosis system using the information network fault diagnosis method according to the present invention.

First, a fault diagnosis system 1 provided in an information network management device is notified of a fault in a network component device or the like. These are collectively shown as failure information source 2. The notified fault information 101 is received by the fault information receiving section 3 and passed to the fault information analyzing section 4 as a fault message 102.

The failure information analysis unit 4 analyzes the failure message 102,
The analysis data 103 is stored in the failure status storage table 5 from time to time.

FIG. 2 shows an example of the operation from receiving and analyzing fault information to storing the fault status in the fault status storage table 5 in the sequence described above. Fault information 101 notified from the fault information source 2 is received by the fault information receiving section 3 and passed to the fault information analyzing section 4 as a fault message 102 . The failure message 102 includes a common information area where the failure information analysis unit 4 analyzes data for storage in the failure status storage table 5, and a unique information area that is not used directly for analysis but is used only for mere display or local analysis. It is divided into information areas. The failure information analysis unit 4 analyzes the failure message 102 and stores the analysis data 103 in the failure situation storage table 5. As shown in FIG. 2, only the information in the common information area is developed in the table.

Returning to Figure 1 again.

On the other hand, the diagnostic rule 7a for fault diagnosis includes a fault status table 7b that holds the fault state at the time of fault occurrence;
This table is associated with the failure cause candidate table 7c.

This fault status table 7b is configured to have a fault cause candidate table 7c corresponding to the combination of network component devices in which a fault has occurred.

The failure cause candidate table 7c includes first failure cause candidates 7c1. Second failure cause candidate 7c2...
Data such as this is stored.

The above-mentioned fault diagnosis rules 7a, fault status table 7b, and fault cause candidate table 7c are stored in a readable/writable file to create a fault diagnosis knowledge database (
DB) 7.

When you start the fault diagnosis, the fault diagnosis section (inference engine) 6
is the failure status data 104 from the failure status storage table 5.
and the fault diagnosis rule 7 from the fault diagnosis knowledge database 7.
Fault diagnosis is performed by taking out the fault status table 7b and fault cause candidate table 7c as diagnostic data 105, and comparing the fault status storage table 5 and fault status table 7b according to the diagnostic rule 7a.

The results of the fault diagnosis are sent to the diagnosis result display section 8 as diagnosis result data 106 and displayed on the display 9. In fault diagnosis, when an adjustment or a test request 108 to the operator 11 is required, the fault diagnosis unit 6 requests the adjustment processing unit 10 107, and the operator 11 takes measures to respond to the request and responds. 109. Failure status storage table 5 via auxiliary information input interface 12
Correct the data. Using the fault information in the fault situation storage table 5 that has been corrected in this way.

The fault diagnosis unit diagnoses the fault again.

FIG. 3 shows an example of the content of data included in the fault diagnosis knowledge database 7 and the content of data in the fault situation storage table 5. The fault diagnosis section 6 uses the diagnosis rule 7.
According to a, the failure status storage table 5 and the failure status table 7b are compared, and the elements of the failure cause candidates are extracted from the failure cause candidate table 7C described in the diagnosis rule with which the tables match, and are arranged in the order of their failure cause certainty. indicate.

Here, the description format of the failure status storage table 5 and the description format of the failure status table 7b are made the same format as shown in the figure, and both tables can be modified by simply rewriting the description framework of these formats. It is possible to do so. Also, by doing this,
It is sufficient to compare the tables in the rules sequentially starting from the top areas of both tables.

Next, the present invention will be explained using a second embodiment.

FIG. 4 shows a second configuration example of a fault diagnosis system using the information network fault diagnosis method according to the present invention. As shown in the first embodiment, the process is the same until diagnosis of a fault is made. However, in this example, a new feature is diagnostic knowledge! iD
A B update processing section 13 is provided.

The fault diagnosis unit 6 diagnoses the fault and identifies fault cause candidates 7cl.
, 7c2... is displayed on display 9,
The fault diagnosis unit 6 uses the adjustment processing unit 10 to request the operator 11 to judge the fault diagnosis result. The operator 11 judges the results and instructs the fault diagnosis unit 6 115.
I do. As a result, if the data in the failure cause candidate table 7c needs to be corrected, the diagnostic knowledge DB update processing unit 13
Failure cause candidate element 7c1 in failure cause candidate table 7c
7c2, etc. For example, the failure cause certainty is corrected.

Next, the present invention will be explained using a third embodiment.

FIG. 5 shows a third configuration example of a fault diagnosis system using the information network fault diagnosis method according to the present invention. As shown in the first embodiment, the process is the same until diagnosis of a fault is made. However, in this embodiment, a diagnostic knowledge DB update processing section 13 is newly provided as in the second embodiment.

The fault diagnosis unit 6 diagnoses the fault and determines the fault cause candidate 7cc.
After displaying 1, 7c2, . The operator 11 makes a judgment on the first diagnosis result and issues an instruction 115 to the fault diagnosis section 6. As a result, if it is necessary to add or modify the diagnostic rule 7a, the diagnostic knowledge 11DB update processing unit 13 performs this. By this method, knowledge (diagnosis rules) can be added to and modified in the fault diagnosis knowledge database 7.

Next, the present invention will be explained using a fourth embodiment.

FIG. 6 shows a fourth configuration example of a fault diagnosis system using the information network fault diagnosis method according to the present invention. As shown in the first embodiment, the method of diagnosing a fault is common. However, in this embodiment, the diagnostic knowledge ff1 is newly added.
DB update processing unit 13. Knowledge acquisition processing interface 14
has been established. Fault information receiving unit 3. The failure information analysis unit 4 constantly collects failure information 101 and stores the failure status data in the failure status storage table 5.

Here, if an artificial failure occurs at any location in the information network, the failure status is stored in the failure status storage table 5. At this time, the operator 11 instructs the knowledge acquisition interface 14 to acquire knowledge, and at the same time inputs the cause of the artificial failure at that time 113. Then, the knowledge acquisition interface 14 updates the diagnostic knowledge DB update processing unit 13.
0 diagnosis knowledge fiDB which performs database modification instruction 114 to
The update processing unit 13 performs necessary additions and corrections to the fault diagnosis rules 7a, the fault status table 7b, or the fault cause candidate table 7c in the fault diagnosis knowledge database 7 (112). With this method, even if there is no diagnosis knowledge at all, new knowledge can be acquired in the fault diagnosis knowledge database 7, and fault cause candidates can be added to the contents of known knowledge ff1 (diagnosis rules) even after system operation.

Next, a fifth embodiment will be explained.

In the first to fourth embodiments, the failure status storage table 5 increases proportionally as the number of network components increases. The following methods can be used to solve this memory capacity constraint. That is, in this example,
Focusing on the connection form of the devices that make up the network, we classify the devices and use a failure status table that represents a set of devices classified into one category.

FIG. 7 shows an example of a network configuration using a high-speed multiplexing device. The network fault diagnosis device 21 is present in a device that manages a network constituted by high-speed multiplexing bags Wt22, 23, 24, 25°26, 27, and 28. The high-speed multiplexing devices are connected using high-speed digital lines as relay lines. Network fault diagnosis device 2
1 uses a portion of the high-speed digital line to collect network management information from each high-speed multiplexing device, analyzes this information, and uses it for fault diagnosis.

Here, we will focus on the connection form of high-speed multiplexers that constitute such a network, and classify high-speed multiplexers according to their connection characteristics. In this example, it is classified into four groups.

Connection characteristic 1 means that if a failure occurs in a high-speed multiplexer with this characteristic, the entire network cannot be managed. This is the conversion device 22.

Connection characteristic 2 means that if a failure occurs in a high-speed multiplexer with this characteristic, it becomes impossible to manage some devices in the network. It is.

Connection characteristic 3 means that even if a failure occurs in a high-speed multiplexing device that has this characteristic, there is no problem in managing other devices in the network, but it becomes impossible to manage only the device in which the failure has occurred. , and this is a case where there is only one trunk line. This is the high speed multiplexer 28.

Furthermore, connection characteristic 4 means that even if a failure occurs in a high-speed multiplexing device having this characteristic, there is no problem in managing other devices in the network, and only the device in which the failure has occurred cannot be managed. Moreover, this applies when there are two or more relay lines.

This is the case with high speed multiplexers 24,26.

The reason for this classification based on the number of trunk lines is that there is only one trunk line.
In the case of a book, if a failure occurs in this high-speed multiplexer 28, there will be no response to the network fault diagnosis device 21, but in the case of two or more, the high-speed multiplexer 24, 26 uses a different route. It is possible to notify the network failure diagnosis device 21 of failure information.
This is because the phenomena when a failure occurs are different.

Focusing on the connection form of the high-speed multiplexing devices constituting the network in this way, the results of classifying the high-speed multiplexing devices based on their connection characteristics are shown in a table (FIG. 7(b)).

The classification according to the connection type shown here is not unique to the fault diagnosis system using this diagnosis method, but is a common problem in any system that uses any diagnosis method for boat fishing, and it is suitable for any fault diagnosis. It can also be applied to systems.

FIG. 8 shows a configuration example of a fault diagnosis system using the present fault diagnosis method when high-speed multiplexing devices are classified according to the connection characteristics of network component devices as shown in FIG.

Fault information 101 is notified to the fault diagnosis system 1 from a diagnostic information source 2 such as an information network management device or network component equipment. The notified fault information 101 is received by the fault information receiving section 3 and passed to the fault information analyzing section 4 as a fault message 102. The failure information analysis unit 4 analyzes the failure message 102 and stores the analysis data 103 in the failure situation storage table 5 from time to time.

On the other hand, the diagnostic rule 7a for fault diagnosis includes a fault status table 7b that holds the fault state at the time of fault occurrence;
This table is associated with the failure cause candidate table 7c. This fault status table 7b is configured to have a fault cause candidate table 7c corresponding to the combination of network component devices in which a fault has occurred. The failure cause candidate table 7c includes the first failure cause candidate 7c for each chip.
Data such as 1° Second failure cause candidate 7c2... is stored. In addition, using the network configuration information 13, the device characteristic analysis unit 14 uses the seventh
A connection characteristic analysis as shown in the figure is performed, and the results are stored as a connection characteristic correspondence table 7d. The above fault diagnosis rule 7a and fault status table 7b
, the failure cause candidate table 7c, and the connection characteristic correspondence table 7d are stored in a readable and writable file, which is called a failure diagnosis knowledge database 7.

When you start the fault diagnosis, the fault diagnosis section (inference engine) 6
is the failure status data 104 from the failure status storage table 5.
and the fault diagnosis rule 7 from the fault diagnosis knowledge database 7.
a, a failure status table b, a failure cause candidate table 7c, and a connection characteristic correspondence table 7d as diagnostic data 10.
5, fault diagnosis is performed by comparing the fault status storage table 5 and the fault status table 7b according to the extraction 1 diagnosis rule 7a.

The results of the fault diagnosis are sent to the diagnosis result display section 8 as diagnosis result data 106 and displayed on the display 9. When an adjustment or a test request 108 to the operator 11 is required in fault diagnosis, the fault diagnosis section 6 requests the arrangement processing section 1o 107, and the operator 11 takes measures to deal with this and responds to the request 109. , the data in the fault situation storage table 5 is corrected via the auxiliary information input interface 12, and the fault diagnosis unit diagnoses the fault again using the fault information in the fault situation storage table 5 corrected in this way. .

The fifth embodiment shown here focuses on the connection form of devices constituting a network, classifies the devices, and uses a failure status table representing a set of classified devices. This is shown as a modification of the embodiment shown in the figure.

Therefore, this method can be directly applied to FIGS. 4, 5, and 6.

FIG. 9 shows the contents of data included in the fault diagnosis knowledge database 7 and the data in the fault status storage table S when high-speed multiplexing devices are classified according to the connection characteristics of network component devices as shown in FIG. This is an example of the content. Unlike the diagnosis rule 7a, the failure diagnosis unit 6 compares the failure status storage table 5 and the failure status table 7b, and selects elements of failure cause candidates from the failure cause candidate table 7c described in the diagnosis rule with which the tables match. and display them in order of failure cause certainty. However, when comparing the fault status table 7b and the fault situation storage table 5, the fault diagnosis section 6 refers to the fault connection characteristic correspondence table 7d and reads out, for example, the address of the fault detection device from the fault situation storage table 5. , the process of creating the failure status table 7b is performed at any time by performing processes such as changing information as a connection characteristic and adding a detection information address as auxiliary information to the failure status table 7b.

Next, a sixth embodiment will be described.

In this embodiment, as shown in FIG. 10, the number of uses is 781 for the diagnostic rule 7a. The number of uses 701. is displayed in the failure cause candidate table 7c.
Number of times the diagnosis matches the failure cause candidate element 702 (confidence level)
This figure shows an example of the content of data included in the fault diagnosis knowledge database 7 when evaluation indicators such as the following are provided.

An actual operational example using the number of uses 7al in the diagnostic rule 7a will be described with reference to FIG. In the fault diagnosis system, it was previously shown that when the fault diagnosis section 6 executes a diagnosis, it first compares the fault status storage table 5 and the fault status table 7b according to the diagnosis rule 7a.

For this reason.

As the amount of information in the diagnostic rule 7a increases, the load for table comparison processing increases. Therefore, as shown in FIG. 10, the fault diagnosis unit 6 performs a process of accumulating the number of times the diagnosis rule 7a has been used for diagnosis after the diagnosis is completed. The number of times this is used is 7al.

When the fault diagnosis unit 6 compares the fault status storage table 5 and the fault status table 7b, the number of uses 7a
Comparison processing is performed in the order of l.

Further, an actual operation example in which the failure cause candidate table 7c is provided with the number of uses 7cl, and the failure cause candidate elements are provided with evaluation indicators such as the number of diagnosis matches 702 (confidence) will be explained using FIG.

In the fault diagnosis system 1, the fault diagnosis unit 6 uses the adjustment processing unit 10 to issue a fault diagnosis result determination request 108 to the operator 11. The operator 11 makes a determination and issues an instruction 115 to the failure diagnosis unit 6. As a result, if the data in the failure cause candidate table 7c needs to be corrected, the diagnostic knowledge DB update processing unit 13 uses the failure cause candidate table 7c. Number of times 7cA, failure cause candidate elements 7cl, 7c2...
Correct the diagnosis match count of 7 cB. The number of times of use 7cA and the number of times of diagnosis matching 7cB are used as indicators of the presentation type of failure cause candidate elements to be presented as failure cause candidates after diagnosis is performed in the failure diagnosis section 6 according to the diagnosis rule 7a. In other words, the more times 7cB of diagnosis matches the failure cause candidate based on one rule, the more likely it is to be the cause of the failure.

Next, in the information network fault diagnosis system,
Furthermore, an evaluation index for the degree of matching between the fault status storage table and the fault status table belonging to each diagnosis rule is provided for each rule and table, and the operator can be provided with instructions on when to start fault diagnosis, and An example of a method for dealing with misdiagnosis due to an error in the timing of starting a fault diagnosis operation will be described. Misdiagnosis caused by an error in the operator's failure diagnosis start operation is not unique to the failure diagnosis system using this diagnosis method, but is a common problem in any system that uses any diagnosis method for boat fishing.

Next, a seventh embodiment will be described.

FIG. 11 shows, as an example, the number of comparison tables 782. which is an index indicating the number of tables to be compared included in one diagnosis rule 7a. Fault status table 7
This figure shows an example of the content of data included in the fault diagnosis knowledge database 7 when an evaluation index such as a diagnostic coefficient 7b2 as auxiliary information is provided in b.

By providing such an evaluation index, the following diagnostic processing becomes possible.

First, it is possible to deal with misdiagnosis caused by an error in the timing of the operator's fault diagnosis start operation. An embodiment of this method will be described using FIG. 4.

When a fault occurs in the network, the fault diagnosis system 1
Failure information 101 is reported. The fault information 101 does not all occur at the same time, and the information is often delayed depending on the cycle and frequency of polling (delivering fault information to each device). Furthermore, failure information messages to the failure diagnosis system 1 also arrive in chronological order.

Therefore, the fault diagnosis must be started only after all the fault information 101 has been reported to the fault diagnosis system 1. Here, in the system in which the operator instructs the fault diagnosis system 1 to start the fault diagnosis, the fault diagnosis starts. If the timing of the instruction is incorrect, diagnosis will be performed based on insufficient fault information 101, resulting in incorrect judgment.

Therefore, 1 diagnosis rule 7 of fault diagnosis knowledge database 7
The number of comparison tables in a is 782. Fault status table 7
b is provided with an evaluation index called a diagnostic coefficient 7b2 as auxiliary information. The processing up to failure diagnosis is as already explained. The fault diagnosis unit 6 may have misjudged the timing of the start instruction and performed the diagnosis based on insufficient fault information 101. Therefore, after performing the fault diagnosis as described above, the fault diagnosis unit 6 may have performed the diagnosis based on the diagnosis rule 7a as a diagnosis result. Among other diagnostic rules whose evaluation index is larger than the comparison table number 7a2, which is the evaluation index included in the selected compatible diagnosis rule, the fault status that is the same as the fault status table 7 trigram used for diagnosis is included in the if part of the rule. Find the diagnostic rule that includes table 7b. As a result, if there is a diagnostic rule that seems to be related, the fault diagnosis section 6 will provide the missing information in the fault status table 7b to the operator 11. If there is any other fault information, input the fault information (instruction 115), and the fault diagnosis unit 6 will perform the diagnosis again. Furthermore, if there is no failure, the process ends as is.

The diagnostic coefficient, which is auxiliary information of the fault status table 7b, provides information on the frequency of occurrence of the fault status table 7b. In the above example, when a diagnostic rule is found, it is used as a measure of which diagnostic rule should be prioritized.

Next, an eighth embodiment will be described.

The evaluation index can also be used as guidance for instructing the operator when to start fault diagnosis, or for automatically starting diagnosis. An embodiment of this method will be explained using FIGS. 4 and 11.

When a fault occurs in the network, the fault diagnosis system 1
Failure information 101 is reported. As already explained, the fault information messages to the fault diagnosis system 1 arrive in chronological order. therefore.

If the fault diagnosis is started at the wrong timing, the diagnosis will be performed based on insufficient fault information 101 and an incorrect judgment will be made. The fault diagnosis unit 6 prevents operational errors by the operator 11, instructs when to start fault diagnosis, and automatically executes the diagnosis.

The contents of the fault diagnosis knowledge database 7 have already been explained.

When the fault information 101 is reported, the fault diagnosis section 6 performs a preliminary fault diagnosis using only the currently available information. That is, the fault information analysis unit 4 writes fault information to the fault situation table 5 every moment (action 103). therefore,
The fault diagnosis unit 6 constantly monitors the fault situation table 5 and
As soon as the failure status table 5 is changed, a search is made to see if there is a diagnostic rule 7a that matches the current failure status. As a result of the search, if there is no matching diagnostic rule 7a, the process is temporarily stopped until the next failure information occurs.

If there is a matching diagnosis rule 7a, the fault diagnosis unit 6
presents the fault diagnosis result to the operator 11 (process 10
6). Such a method allows automatic execution of the diagnosis.

The method of using the evaluation index is the same as in the seventh embodiment.

Furthermore, the following is an example of a method for providing guidance to the operator to start fault diagnosis.

The fault diagnosis unit 6 searches to see if there is a diagnostic rule 7a that matches the current fault situation.

As a result, even if there is a matching diagnostic rule 7a, for example, if the number of uses 7a1 in the diagnostic rule 7a is small, it is determined that the probability of occurrence is low), or a compatible diagnostic rule 7a.
If the diagnostic coefficient of the failure status table 7b corresponding to the failure status table 7b (information on the frequency of occurrence of the failure status table 7b is given) is small, the operator is given guidance to wait for the start.

mentioned here. The seventh embodiment and the eighth embodiment are just examples, and any evaluation index can be used.

〔Effect of the invention〕

Since the present invention is configured as described above, the following effects can be expected.

The fault diagnosis system can perform fault diagnosis through simple memory area comparison processing between the fault storage status table and the fault status table, so that processing speed can be increased. Furthermore, since the configuration of the diagnostic rules is simple, it is easy to create new diagnostic rules or modify them.

Furthermore, in such a fault diagnosis system, the knowledge base can be easily added by converting the fault status storage table when a fault occurs due to an artificially generated fault at any location into a fault status table and storing it.
Can be expanded.

In addition, a correction processing function for the fault diagnosis knowledge database is provided, and when the fault diagnosis result is determined, the fault cause candidate table at the time of fault occurrence is automatically corrected if necessary, and the knowledge base is corrected by storing it. This makes it easy to add up your knowledge.

In addition, evaluation indicators such as the number of times a diagnosis rule is used, the number of times it is used in the failure cause candidate table, and the number of diagnosis results (confidence) for failure cause candidate elements in the failure cause candidate table are set up, and corrections and additions are made. Since the system allows searching for rules and presenting cause candidates based on the index, it is possible to apply knowledge acquisition and correction functions to diagnostic rules.

Furthermore, by focusing on the connection form of the devices that make up the network, and classifying the devices, it becomes possible to respond to frequent changes in the network configuration by using a fault status table that represents a set of classified devices. .

Furthermore, an evaluation index for the degree of agreement between the failure status storage table and the failure status table belonging to each diagnosis rule is provided for each rule and table, so that it can be used in case the system operator makes a mistake in starting the failure diagnosis operation. It is possible to deal with misdiagnosis, to guide the system operator when to start fault diagnosis, or to automatically start fault diagnosis, making the system more efficient and easy to operate. It can be used in many ways.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a fault diagnosis system using a fault diagnosis method for an information network according to an embodiment of the present invention;
The figure is an operation sequence diagram from receiving fault information to storing it in the fault status storage table. Figure 3 shows an example of the content of data included in the fault diagnosis knowledge database and the content of data in the fault status storage table. Explanatory drawings, FIGS. 4 and 5. 6 and 8 are configuration diagrams of a fault diagnosis system using a fault diagnosis method for an information network according to another embodiment of the present invention, and FIG. 7 is a block diagram showing an example of a network configuration using a high-speed multiplexing device. Figures 9, 10, and 11 showing the classification are explanatory diagrams of the fault diagnosis knowledge database when high-speed multiplexing devices are classified according to the connection characteristics of the network components shown in Figure 7. . 1...Fault diagnosis system, 2...Fault information source, 3.
... Fault information receiving section, 4... Fault information analysis section, 5...
- Fault status storage table, 6... Fault diagnosis unit (inference engine), 7... Fault diagnosis knowledge database, 7a.
...Diagnosis rule, 7a1...Number of times the diagnosis rule is used,
7a2...Number of comparison tables for diagnostic rules, 7b...
Fault status table, 7bl... Auxiliary information of the fault status table, 7b2 Diagnosis coefficient of the fault status table, 7c... Fault cause candidate table, 7cl.
...Second failure source candidate, 7c2...Second failure source candidate, 7
cA... Number of uses of failure cause candidate element, 7cB...
Diagnosis matching number of failure cause candidate elements, 7d... Connection characteristic correspondence table, 8... Diagnosis result display section, 9... Display, 10... Time adjustment processing section, 11... Operator, 12.・・Auxiliary information input interface, 13・
...Diagnostic knowledge DB update processing unit, 14...Knowledge acquisition interface, 21...Network fault diagnosis device, 2
2゜23. 24, 25, 26, 27, 28... High-speed multiplexing device, 29... Table classifying high-speed multiplexing devices by connection characteristics, 101... Fault information, 102... Fault message , 103... Analysis data, 104... Failure status data, 105... Diagnosis data, 106... Diagnosis result data, 107... Request to the adjustment processing unit, 10
8... Test request, 109... Response to timing, 112... Addition/modification of fault diagnosis knowledge database, 113... Input of fault cause, 114... Database correction instruction, 115... Database correction instructions given by the operator. Tomi 1 Fig. Tomi Z Fig. Tomi 3 Fig. Fig. Fig. Fig. Tomi 7 Fig.

Claims (1)

[Claims] 1. In a fault diagnosis system for an information network management system, there is provided a fault status table for storing fault information on fault detection parts and fault conditions such as test results, and candidates for fault causes that cause these phenomena. A fault diagnosis knowledge database is used that stores a fault cause candidate table for storage, and a plurality of diagnostic rules in which the combination of these fault status tables and fault cause candidate tables is regarded as a pair of diagnostic rules, in a read/write file. , the failure status of the failure detection part is stored in a failure status storage table, which is a table for momentarily storing the failure status, and the failure diagnosis is performed by comparing the failure status storage table and the failure status table. Fault diagnosis method. 2. In the information network fault diagnosis system according to claim 1, the description format of the fault status storage table and the description format of the fault status table are the same, and both tables can be written by simply rewriting the description framework of this format. A fault diagnosis method that allows the format to be modified and is characterized in that when comparing tables in order to search for compatible rules in a rule, the comparison is performed sequentially from the top area. 3. In the information network failure diagnosis system according to claim 2, the failure status of the failure detection part can be stored in the failure status storage table from time to time, so that failures due to artificially caused failures can be detected at any location. A fault diagnosis method characterized by expanding a knowledge base by converting a fault status storage table at the time of occurrence into a fault status table and storing it, and creating a fault cause candidate table at that time. 4. In the information network fault diagnosis system according to claim 2, a correction processing function of the fault diagnosis knowledge database is provided, and when a fault diagnosis result is determined, it is necessary to automatically correct the fault cause candidate table when a fault occurs. A fault diagnosis method characterized in that the knowledge base is modified by modifying and storing the knowledge base. 5. In the information network fault diagnosis system according to claims 2, 3, and 4, by further focusing on the connection form of devices constituting the network and classifying the devices, a set of classified devices is represented. A fault diagnosis method characterized by using a fault status table. 6. In the information network fault diagnosis system according to claims 2, 3, 4, and 5, the number of uses in the diagnosis rule, the number of uses in the fault cause candidate table, and the diagnosis match with the fault cause candidate element in the fault cause candidate table. A fault diagnosis method characterized by providing an evaluation index such as the number of times (confidence), making it possible to modify or add the index, and searching for rules and presenting cause candidates based on the index. 7. In the information network fault diagnosis system according to claim 6, an evaluation index such as the degree of coincidence between the fault status storage table and the fault status table belonging to each diagnostic rule is further provided for each rule and the table. , To deal with misdiagnosis when a system operator makes a mistake in the timing to start a fault diagnosis, or to provide guidance to a system operator about the timing to start a fault diagnosis, or to automatically start a fault diagnosis. A fault diagnosis method characterized in that: