9 » IC Electronic information » Category: A

IC Electronic information

A design of electronic equipment failure Library

In Electronic Infomation Category: A | on April 28,2011

Fault diagnosis expert system is a compelling diagnosis of the development direction of the field is one of the most studied, most widely used diagnostic technique for a class of intelligent, mainly for those who are not difficult to build accurate mathematical models or mathematical models of complex systems . Has gone through two stages, namely, shallow knowledge-based fault diagnosis of the first generation of expert systems and MC7805CTG datasheet and knowledge-based second generation of deep fault diagnosis expert system. Based on shallow knowledge (the experience of human expert knowledge) of the fault diagnosis system is based on experts in the field and MC7805CTG price and experience of the operators heuristic knowledge of the core, or generated by deductive reasoning based reasoning to obtain diagnostic results. Based on deep knowledge (diagnostic object model knowledge) diagnosis of the fault diagnosis system requirements for every aspect of the object with a clear relationship between input and MC7805CTG suppliers and output expression, diagnosis, diagnosis of the first object by the actual output and expected outputs of the inconsistency caused by these generation * reasons for the discrepancy, and then in the field of diagnosis of the first law of object knowledge (with a clear scientific basis of knowledge) and their relationship within the particular constraints, the use of certain algorithms, to identify possible fault source. In the fault diagnosis expert system, expert system fault library is the core, and embodies the level of expert systems, but the failure or malfunction of the construction of the library access to knowledge has become the bottleneck of building expert systems.

1 fault library construction steps

Fault database is designed to establish an expert system is the most important and most difficult task. The initial problem of knowledge, including knowledge of design, knowledge conceptualization, the concept of formal, formal rules and regulations of legalization. Problem knowledge, that is, research to identify the nature of the problem; knowledge conceptualization, that is, general knowledge of key concepts that need their relations; the concept of formal, that is, to determine the data structure used to organize the form of knowledge; formal rules, that is the preparation of rules to transform the grounds of formalized knowledge representation for computer programming language statements and procedures for implementation; rules legalized, you acknowledge that the rules of the rationality of knowledge, testing the validity of the rules.

Failure to establish the first library by knowledge engineers from experts in the field test equipment get in the diagnostic process or by obtained from the user knowledge, knowledge acquisition module and then sent to the knowledge repository.

Main difficulty is to obtain knowledge on how to properly grasp the concepts used by experts in the field, relationships and problem solving methods. According to expert knowledge and experience to provide the characteristics of test equipment, by way of direct access, interactive way to get access to valuable diagnostic information. Fault Database for

is a very complex process, it is necessary to follow certain methods and procedures. Usually the steps shown in Figure 1 in accordance with the existing design information from the extract qualitative knowledge about principles and rules of this knowledge into the form of induction. Below is shown in Figure 1 to illustrate the various steps are:

Figure 1 Fault Database for steps

(1) function, structure, level of decomposition. As the modern design of the modular design, so the electronic products in the functional and structural obvious hierarchy, so to extract the knowledge of fault diagnosis of electronic devices, the system decomposition is necessary.

(2) to establish the normal function of components described. For each of the above by parts, it is necessary to work a qualitative description logic.

(3) determine the components of the test point set. Enter the measuring point is divided into parts and components, measuring points out measuring point, a component of P in the diagnosis, only all of its input signals are normal, it can output to determine whether the component failure.

(4) to determine possible fault set of components. By the above modules are failure analysis, for the parts, we must first determine a possible fault types, component P determine the possible fault set can be based on experience, FMEA analysis and collection, and many other devices Zi Liao .

(5) to establish the relationships of qualitative network. For the components P, make sure each of the output caused by failure characterized by e, that the output of a combination of data measurement points. Then, on the part of all measuring points P data classification, a class of its own failures of that component test data, recorded as PSO; non-self-induced failure (caused by the input exception) go to another type of failure, denoted by PU O . W PSO when the test data, the fault diagnosis can be sure that part P; when the test data W PUO, the need to test parts of the input P, the part P of the input is the output of other components, therefore, can by level reasoning, until the fault location will be in one or several components.

(6) the failure of induction into the rules of the form of relationship networks. The above causal Fenxi get the Jieguozhuanhua Cheng two kinds Guize Xing Shi :

IF W PSO T HEN P fault;

IF W PUO TH EN input of the P test;

(7) failure to write the rule library, through the general knowledge base editing system makes it easy to write the rules of the above fault knowledge base.

2 Fault Database Design

2. 1 Data collection and knowledge acquisition

Data collection in the fault diagnosis system to occupy an important role. For the diagnostic system, the more information collected, the more easy to locate and determine the fault, this system will increase the necessary hardware design, as well as reduce the servo system hardware design complexity, according to the characteristics of servo systems and expert the lessons learned, the use of servo system itself necessary hardware connection, a slight increase in hardware design, component diagnostic system for data collection.

System knowledge by production (Pro duct ion Rules) *, also known as rules *. Production is often used to represent causal knowledge, the basic form of P Q, or IF PT HEN Q. Where, P representative of conditions, such as the premise, state and causes; Q representative of the results, such as conclusions, action, consequences. The implication is: if the precondition P is satisfied, you can launch the implementation of the provisions of Q conclusions or actions. Put together a group of production, so that they complement each other, synergies, the conclusions of a production can be generated for use as a prerequisite to another production, obtained in this way, problem solving, such a system is called the generating systems, also known as rule-based system.

Failure for multiple factors, the use of fault tree to represent the fault tree model reflects the level of fault propagation and the son of a causal relationship between the parent node, fault tree, the parent of a child node is the source of failure, so use of fault relationship between the tree nodes, according to the fault phenomena to identify the problem.

2. 2 data table design

Order to play a role in the computer knowledge to generate problem-solving ability, must be formalized by an entity that after the kinds of knowledge into the computers internal form, but also to establish the necessary mechanisms and the interpretation of a good user interface. The system uses Access 2003 database to create a knowledge base built by A ccess 2003 system knowledge base, knowledge base can make the search more efficient and management more convenient.

Knowledge Base consists of four tables, which are the main failure mode table, fact table, the test point table and the rules table.

Main failure mode table test equipment used to store the main fault type. The structure of the table Module (FaultN ame, Child), which, FaultName the name for the fault type, Child signs of this fault type the corresponding name of the table. Shown in Table 1.

Table 1, the main failure mode table


fact and conclusions of the table, including table, two tables of the same structure. The structure of the table Fact (Name, ID), which, Name for the fault symptoms (Conclusion) name, ID for the fault symptoms (conclusions) of the code. Table 2.

Table 2 fact sheet

Test point table used to store the test channel of the test points. The structure of the table T est _ Point (Po intName, PCBName, X, Y, ID, ToneName, Solution), which, PointName name for the test points, PCBName PCB plans to test the channel files, X for the test point of the abscissa , Y coordinates for the vertical test points, ID code for the test points, ToneName the name of the test channel, Solut io n to take measures to resolve the fault. Shown in Table 3.

Table 3 test points table

Rules table including shallow knowledge and deep knowledge of the rules rule table table, two tables of the same structure. Rule table to store rule antecedent and the consequent parts of the code.

Inference engines internal working process is the use of the fact that parts of the match prior to encoding, using a comprehensive database on the reasoning process and after the first pieces of recorded cases. The facts in the reasoning process and the final results of the encoding codes can be found in the fact table corresponding name. The structure of the table (ID, Co ndition_number, Conclut ion_number, sig1, sig2,!, Sig10, con1, co n2,!, Con10), which, ID number for the rule, Co ndition_number number of preconditions for the rule, Co nclut ion_number number of conclusions for the rules, sig 1, sig2,!, sig10 used to store the premise of the rule, co n1, con2,!, co n10 used to store the rules of the conclusions. As shown in Table 4.

Table 4 rule table

3 inference implementation

Fault diagnosis in the test is most important is the inference engine, for the certainty of knowledge, the use of forward chaining expert system approach to reasoning. For the uncertain, ambiguous information, neural network fuzzy inference. Inference engine to achieve through the establishment of class EsReaso n, the following is defined in the class the main pointers, arrays, variables and functions, such structure is as follows:

Class EsReason


pr otected:

Majo rTable * M ajor;

SigalTable * Sig al;

Co nclut ionTable * Conclutio n;

Po int Table * point;

RuleTable * Rule;

BOOL Connectio nDb ();

BOOL LoadT able (CStr ing name);

CStr ing Var iantToCStr ing (co nst _ var iant_t & var);

CStr ing Curr entConditio ns [100];

CStr ing Curr entConclusio ns [100];

CStr ing pr og ram;

unsig ned int To tal_Cur rentCo nditions;

unsig ned int To tal_Cur rentCo nclusio ns;

BOOL Reasoning ();

BOOL NEWff ();


Which, * Majo r, * Sigal, * Co nclut ion, * point, and * Rule definition of the table point to the main failure mode, Symptoms, conclusions table, test points and rules of the structure of the table pointer in the database to facilitate the table data records and interviews. Members through the establishment of BOOL type function Connect io nDb and Lo adTable to realize the database connection and load. Loading the database, by setting the SQL language to open the preparation of each table, and then call the recordset Open member function pointers to open each individual table records. Move through the record set pointer record set records each table number, and then move the record set pointer record to the table to the application stored in dynamic memory.

Comprehensive database through an array of Cur rent Condit io ns [100] and Cur rent Conclusio ns [100] implementation of a comprehensive database used to record the reasoning process of the initial conditions, the intermediate results and final conclusions of the code to facilitate the inference engine by Comprehensive database of the contents of forward reasoning, through CStr ing prog ram record variable of type inference rules used in the process to achieve the reasoning process of interpretation.

In the class of functions is the core EsReason Reasoning and NEWf f, Reasoning function works as follows: the fact that the integrated database as the initial conditions, and the rule base to match the rule antecedent; when the rule is activated, activated by these rules, reasoning functions into the integrated database in the conclusions of reasoning continues until no other rule with the antecedent of the fact that the consolidated database to match up. Used in the reasoning process is a very important member variables of type BOOL Rule _Used records used rules to avoid death occurred during the reasoning cycle. NEWf f the work function of the reasoning process can select the appropriate decision logic reasoning to achieve the automation of fault diagnosis.

4 Concluding Remarks

Because of modern electronic devices such as their characteristics and environmental factors in complex conditions, the traditional fault diagnosis methods can not meet the diagnostic requirements. Library-based fault diagnosis method for intelligent devices rely on the principle of qualitative analysis, the full use of existing design data, without the need to quantify the actual problem of over-simplification, the underlying data than relying on a more quantitative analysis close to practical engineering. Proposed for the electronic equipment electronic equipment failure, a failure to achieve the program library to be effective in fault location and make repairs guidance, effectively improving the general maintenance staff, fault diagnosis and repair efficiency.

MC7805CTG datasheetMC7805CTG suppliersMC7805CTG Price

All right © 2010-2016 Certificate