CN104376033A - Fault diagnosis method based on fault tree and database technology - Google Patents

Abstract

A fault diagnosis method based on fault tree and database technology, which manages the knowledge base and reasoning base in the fault diagnosis process through the database, and completes the automatic configuration and overall control of the reasoning process and diagnosis strategy, including the following steps: a. Receive users Input, including fault type and fault phenomenon keyword; b. According to the fault type and fault phenomenon keyword retrieval, judge whether the fault is a bottom fault, if it is a bottom fault, display the content of the corresponding "maintenance suggestion" field, save the diagnosis result, This diagnosis process is over; if it is not the underlying fault, search the field "node code" in the fault phenomenon table corresponding to the fault, get all fault nodes contained in the next layer, and automatically deduce the fault node corresponding to the diagnosis conclusion. Its purpose is to provide a fault diagnosis method based on fault tree and database technology that can quickly perform fault diagnosis and improve fault diagnosis efficiency when a complex system fails.

Description

A Fault Diagnosis Method Based on Fault Tree and Database Technology

technical field

The invention relates to a fault diagnosis method based on fault tree and database technology. the

Background technique

At present, the commonly used fault diagnosis methods mainly include acoustic diagnosis method, vibration diagnosis method, temperature diagnosis method, strength diagnosis method, pollutant diagnosis method, pressure flow diagnosis method and electrical parameter diagnosis method, etc. These methods consider that system failures will lead to changes in the operating parameters of the system. Under normal conditions, the system parameters should be in a nominal state, and when the system fails, these parameters will deviate from their nominal state. Therefore, fault diagnosis When using these nominal states as reference quantities, it is judged whether the system is faulty or not according to the estimated residual characteristics of the system output or state variables. the

Based on this, the traditional fault diagnosis method lists and summarizes the causes of all faults, and processes them in a standardized and organized manner. Once the system fails, it needs to be checked in sequence according to the established procedures and eliminated one by one. However, this method has some disadvantages, such as failure to classify the faults, which cannot reflect the impact of each fault on the operation of the system, and the troubleshooting process is time-consuming and inefficient. Especially when users need to know the functional logic relationship and degree of correlation between faults at different levels of the system, and hope to find and eliminate system faults accurately and efficiently at the fastest speed and at the lowest cost, this method cannot meet the requirements. Require. the

Contents of the invention

The purpose of the present invention is to provide a fault diagnosis method based on fault tree and database technology that can quickly perform fault diagnosis and improve fault diagnosis efficiency when a complex system fails. the

A fault diagnosis method based on fault tree and database technology of the present invention manages the knowledge base and reasoning base in the fault diagnosis process through the database, and completes the automatic configuration and comprehensive control of the reasoning process and diagnosis strategy, including the following steps:

a. receiving user input, the user input includes a fault phenomenon keyword;

b. Retrieve the fault diagnosis information table in the fault diagnosis database according to the fault type and fault phenomenon keywords, and judge whether the fault is a bottom fault. The content of the "Maintenance Suggestion" field of the fault, and save the diagnosis result, the diagnosis process is over; if it is not the underlying fault, search the field "node code" in the fault phenomenon table corresponding to the fault, and get the next fault node in the fault tree All the faulty nodes contained in one layer, and then according to the diagnosis results, automatically infer the faulty nodes corresponding to the diagnostic conclusions among all the faulty nodes contained in the next layer;

Wherein, if it is not the bottom fault as described in step b, find the field "node code" in the fault phenomenon table, obtain all fault nodes contained in the next layer of the fault node in the fault tree, and then automatically list them under it according to the diagnosis result Among all the faulty nodes included in the first layer, the faulty node corresponding to the diagnostic conclusion is deduced. The specific reasoning method is:

According to the retrieved field "node code" in the fault phenomenon table corresponding to the fault, retrieve the fault diagnosis reasoning information table in the fault diagnosis database, call out its corresponding "diagnostic content" field, and display the "diagnostic content" to the user , allowing the user to operate according to the prompts of the diagnosis content, and at the same time, according to the "node type" in the corresponding fault diagnosis information table, the fault diagnosis process type is divided into the detection amount of the acquisition card and the manual observation amount;

If it is the detection quantity of the acquisition card, call out the "item code" in the corresponding fault diagnosis information table, and search the evaluation plan configuration table according to the "item code", and display the detection method and detection steps of the item, the position of the detection point, and the installation of the sensor method, set the test conditions, prepare for the test, and then pass the corresponding parameters to the test program according to the measurement point channel mapping table, prompt the user to prepare the test conditions, and then automatically complete the test; retrieve all faulty nodes in the next layer, and test The result is matched with the feature quantity of each faulty node, and the successful match is the faulty node corresponding to the next diagnosis;

If it is a manual observation quantity, the user inputs the observation quantity according to the prompt of the diagnosis content; retrieves all the faulty nodes in the next layer, and matches the observation quantity result with the feature quantity of each faulty node. faulty node;

c. According to the faulty node, obtain its fault type and fault symptom keywords, and then proceed to step b. the

A fault diagnosis method based on fault tree and database technology of the present invention, when each step is diagnosed, the "receipt number", "diagnosis serial number", "node code", "node name", "node position" of the step diagnosis will be ", "Node Type", "Diagnostic Content", "Item Code", "Number of Measuring Points", "Length of Single Measuring Point Data", "Measuring Point Data", "Number of Indicators", "Technical Indicator Data", The "maintenance suggestion" information is stored in the fault diagnosis process information table. When the final diagnosis result is output, the information of each single-step diagnosis in the fault diagnosis process information table in the fault diagnosis database is called out and output in the form of a report. the

The present invention is a fault diagnosis method based on fault tree and database technology. According to the response relationship between fault events and fault causes, layered reasoning is performed; the knowledge information and detection process information of each fault node are defined and stored in the fault In the diagnosis database, manual detection or automatic detection is implemented for each fault node in the fault tree, and the detection resource allocation is effectively carried out. When a complex system fails, it can quickly diagnose the fault and improve the efficiency of fault diagnosis. the

A fault diagnosis method based on fault tree and database technology of the present invention will be further described in detail below in conjunction with the accompanying drawings. the

Description of drawings

Figure 1 is a schematic diagram of the fault tree expanded with fault A as the top fault;

Fig. 2 is a kind of fault diagnosis method flowchart based on fault tree and database technology of the present invention;

Fig. 3 is an example of electrical fault tree when the diesel engine is started based on the present invention. the

Detailed ways

Before setting forth the specific embodiments of the present invention, the meanings of some names, characters and graphical symbols involved will be explained. the

First of all, it is necessary to obtain as much system fault data as possible, layer the fault data according to expert experience, and carry out logical processing to construct a fault tree. The fault tree is composed of top-level faults, bottom-level faults and middle-level faults. Figure 1 is a schematic diagram of the fault tree expanded with fault A as the top fault. the

Top-level fault: The initial fault that the system least expects to occur or is specified by the user for logical analysis is taken as the top-level fault in the fault diagnosis process, and placed at the top of the fault diagnosis process, indicated by "□". the

Bottom faults: all the most basic faults that lead to the occurrence of fault phenomena are regarded as bottom faults, indicated by "○". the

Middle-level faults: from the top-level faults to the bottom-level faults, the faults obtained by intermediate reasoning are middle-level faults, which are located between the top-level faults and the bottom-level faults, and are represented by "◇". the

Node code: Each fault diagnosis node corresponds to a record in the fault diagnosis information table of the database, which is usually identified by a node code, which is used to form a pointing relationship between faults, and there are corresponding The node code, which is the key to traction and control the smooth execution of the entire diagnostic process. the

Referring to Fig. 2, a fault diagnosis method based on fault tree and database technology of the present invention manages the knowledge base and reasoning base in the fault diagnosis process through the database, and completes the automatic configuration and comprehensive control of the reasoning process and diagnosis strategy, including the following steps:

a. Receive user input, the user input includes the keyword of the fault phenomenon;

b. Retrieve the fault diagnosis information table in the fault diagnosis database according to the fault phenomenon keywords, and judge whether the fault is a bottom fault. If it is a bottom fault, it means that the diagnosis result is already at the bottom of the fault tree, and then display the corresponding "maintenance suggestion" The content of the field, and save the diagnosis result, the diagnosis process is over; if it is not the bottom fault, search the field "node code" in the fault phenomenon table corresponding to the fault, and get the node code contained in the next layer of the fault node in the fault tree All faulty nodes, and then according to the diagnosis results, automatically infer the faulty nodes corresponding to the diagnostic conclusions among all the faulty nodes contained in the next layer;

If it is not the bottom fault in step b, find the field "node code" in the fault symptom table, get all the fault nodes contained in the next layer of the fault node in the fault tree, and then according to the diagnosis result, automatically in the next layer Among all the fault nodes included, infer the fault node corresponding to the diagnostic conclusion, the specific reasoning method is:

According to the retrieved field "node code" in the fault phenomenon table corresponding to the fault, retrieve the fault diagnosis reasoning information table in the fault diagnosis database, call out its corresponding "diagnostic content" field, and display the "diagnostic content" to the user , allowing the user to operate according to the prompts of the diagnosis content, and at the same time, according to the "node type" in the corresponding fault diagnosis information table, the fault diagnosis process type is divided into the detection amount of the acquisition card and the manual observation amount;

If it is the detection quantity of the acquisition card, call out the "item code" in the corresponding fault diagnosis information table, and search the evaluation plan configuration table according to the "item code", and display the detection method and detection steps of the item, the position of the detection point, and the installation of the sensor method, set the test conditions, prepare for the test, and then pass the corresponding parameters to the test program according to the measurement point channel mapping table, prompt the user to prepare the test conditions, and then automatically complete the test; retrieve all faulty nodes in the next layer, and test The result is matched with the feature quantity of each faulty node, and the successful match is the faulty node corresponding to the next diagnosis;

If it is a manual observation quantity, the user inputs the observation quantity according to the prompt of the diagnosis content; retrieves all the faulty nodes in the next layer, and matches the observation quantity result with the feature quantity of each faulty node. failure node;

c. According to the faulty node, obtain its fault type and fault symptom keywords, and then proceed to step b. the

In each step of diagnosis, the "document number", "diagnosis serial number", "node code", "node name", "node position", "node type", "diagnosis content", "item code" of the step diagnosis , "Number of measuring points", "Length of single measuring point data", "Measuring point data", "Number of indicators", "Technical indicator data", and "Maintenance suggestion" information are stored in the fault diagnosis process information table, and the final diagnosis When the results are output, the information of each single-step diagnosis in the fault diagnosis process information table in the fault diagnosis database is called out, and output in the form of a report. the

Referring to FIG. 3 , based on the example of the electrical fault tree when the diesel engine is started in the present invention, the example of the fault diagnosis reasoning process can be demonstrated with table reasoning, as shown in Table 1 below. the

Table 1 Reasoning method for electrical faults when starting diesel engine based on the present invention

According to the response relationship between the fault event and the fault cause, the present invention performs layered reasoning, defines the detection process information of each fault node and stores it in the fault diagnosis database, and implements manual detection or automatic detection of each fault node in the fault tree. Detection can configure detection resources more effectively and realize comprehensive management of the fault diagnosis process. the

In terms of fault data storage, establish an overall fault diagnosis database structure, and define database information for each fault tree node on the basis of various fault data summaries, system databases, and system database relationship structure construction according to database design requirements and data processing procedures , generate a fault phenomenon table, a fault diagnosis information table, a fault diagnosis reasoning information table, and store the diagnosis process information in the fault diagnosis process information table. the

Claims (2)

1. the method for diagnosing faults based on fault tree and database technology, it is characterized in that being managed the knowledge base in failure diagnostic process and reasoning storehouse by database, complete the automatic configuration of reasoning process and Diagnostic Strategy and comprehensive management and control, comprise the following steps:

A. receive user's input, described user's input comprises phenomenon of the failure key word;

B. according to the failure diagnosis information table in described phenomenon of the failure key search Fault Diagnosis Database, judge whether this fault is bottom fault, if bottom fault, illustrate that this diagnostic result has been positioned at the bottom of fault tree, then show the content of corresponding " maintenance suggestion " field, and preserve diagnostic result, this diagnostic procedure terminates; If not bottom fault, retrieve the field " station code " in phenomenon of the failure table corresponding to this fault, obtain all malfunctioning nodes that in fault tree, lower one deck of this malfunctioning node comprises, then according to diagnostic result, automatically, in all malfunctioning nodes that one deck comprises under it, the malfunctioning node corresponding to diagnosis is inferred;

Wherein described in step b if not bottom fault, find the field " station code " in phenomenon of the failure table, obtain all malfunctioning nodes that in fault tree, lower one deck of this malfunctioning node comprises, then according to diagnostic result, automatically in all malfunctioning nodes that one deck comprises under it, infer the malfunctioning node corresponding to diagnosis, concrete inference method is:

According to the field retrieved in phenomenon of the failure table corresponding to this fault " station code ", fault diagnosis reasoning information table in retrieval Fault Diagnosis Database, recall " diagnosis content " field of its correspondence, and " diagnosis content " is shown to user, user is allowed to operate according to diagnosis content prompting, simultaneously according to " node type " in the failure diagnosis information table of its correspondence, failure diagnostic process type is divided into capture card detection limit and manual observation amount;

If be capture card detection limit, then recall " project code " in corresponding failure diagnosis information table, according to " project code ", retrieval Te st grogram allocation list, show detection method and the detecting step of this project, check point position, sensor installation method, setting test condition, setup test, then according to measuring point channel map, relevant parameter is passed to test procedure, after prompting user gets out test condition, automatically complete test; Retrieve its lower all malfunctioning node of one deck, and test result mated with the characteristic quantity of each malfunctioning node, what the match is successful is malfunctioning node corresponding to next step diagnosis;

If be artificial observed quantity, user points out according to diagnosis content, input observed quantity; Retrieve its lower all malfunctioning node of one deck, and observed quantity result mated with the characteristic quantity of each malfunctioning node, what the match is successful is malfunctioning node corresponding to next step diagnosis;

C. according to malfunctioning node, obtain its fault type and phenomenon of the failure key word, then continue to perform step b.

2. a kind of method for diagnosing faults based on fault tree and database technology according to claim 1, it is characterized in that: when each step diagnosis, " document number " that this step is diagnosed by capital, " diagnosis sequence number ", " nodes encoding ", " nodename ", " node location ", " node type ", " diagnosis content ", " project code ", " measuring point number ", " single measuring point data length ", " measuring point data ", " index number ", " technology index data ", " maintenance suggestion " information is stored in failure diagnostic process information table, when last diagnostic result exports, the information of each single step diagnosis in failure diagnostic process information table in Fault Diagnosis Database is recalled, and export according to the form of form.