Study On Fault Diagnosis Expert System For The Commander Panoramic Sight System

The commander panoramic sight system in Type-99 tanks is characterized by multiple functions, complicated structure, and numerous components. Compared to the current systems in infantry fighting vehicles and Type-96 tanks, it is more advanced with more complicated structure. As a result, malfunctions may occur in many places of the system, thus resulting in great difficulties in maintenance. The inspection system in service has failed to meet the current needs for a long time. Thus, it is of great urgency to develop a fault diagnosis expert system for the commander panoramic sight system in Type-99 tanks. This thesis designed and established a fault diagnosis expert system for the commander panoramic sight system in Type-99 tanks, and the specific research contents are described as follows:(1) The knowledge acquisition module of the fault diagnosis expert system for the commander panoramic sight system is researched. Fault tree analysis and BAM neural network were used in this study for knowledge acquisition, and the bottleneck problem of knowledge acquisition in the fault diagnosis expert system for the commander panoramic sight system was solved. The fault tree of each subsystem is firstly established based on expertise, professional materials consultation and a great deal of experimental data; then, item events on the fault trees are determined according to fault manifestations; finally, fault tree analysis is used to determine the minimum cutset and fault cause. If the results obtained from fault trees do not match the facts, fault tree analysis could be used to process the fault manifestations and generate the data that can be identified by the BAM diagnosis module; then, BAM network module is started to perform associative recall, decompose the system, and generate training samples of BAM neural network, devised a new method of weighting matrix, overcome the BAM neural network of the original data samples redundancy problem, improve the BAM neural network training efficiency and convergence speed, finally, specific fault location and cause could be obtained.(2) Knowledge base of the fault diagnosis expert system for the commander panoramic sight system is built. The establishment of system knowledge base is the hardcore of the fault diagnosis expert system for the commander panoramic sight system. This study mainly used the knowledge representation method of production rules to analyze and manage the rule base, conclusion base, explanation base, and condition base. The database software, Access 2003, was used to establish and store the system knowledge base and Visual C++6.0 were used to modify, maintain, and create the knowledge base.(3) Inference engine of the fault diagnosis expert system for the commander panoramic sight system is built. Based on the analysis of the system, facts or data provided by users, knowledge in the knowledge base, and comparison and research of several inference methods, an improved forward reasoning method was established for inference through the designed inference steps, and then fault location, cause, and repair advice could be obtained to further improve the efficiency of the inference engine and shorten fault diagnosis time.(4) Software part of the fault diagnosis expert system for the commander panoramic sight system is designed. According to the major constituent parts of the expert system, the functional modules of the fault diagnosis expert system for the commander panoramic sight system, including login system, knowledge acquisition system, knowledge base management and maintenance system, were designed respectively. Visual C++6.0 was used to write the program of the system software, and Access 2003 was used to create and manage all the databases, so as to give full play to the functions of all the functional software modules of the fault diagnosis expert system for the commander panoramic sight system, and make the system perform well in human-computer interaction.