Inertial Navigation System Fault Diagnosis Expert System

On the satellite measure and control ship, the inertial navigation system has complex structure and requires high precision, so that the special fault diagnosis expert system is one of the effective approaches to improve the security and reliability of it.In this article, firstly, the basic principles, characteristics, realizing methods and applicable domain of the Fault Tree Analysis, Case-based Reasoning and Rule-Based Reasoning expert system are analyzed and summarized. Secondly, based on the features and actual situation of the inertial navigation system, the system is analyzed and the fault model is built, further the fault diagnosis expert system constructed by FAT and CBR methods is presented. The deep reasoning is realized by FAT based on system's structure, action and function, and the shallow reasoning is carried out by CBR, so that the deep knowledge and shallow knowledge are combined to realize the fault diagnosis expert system. In this system, production rules approach and frames approach are applied to represent and storage the knowledge, the forward chaining reasoning and case-based searches reasoning are applied to the inference engine. These methods effectively improve the system reasoning efficiency and accurateness. To the knowledge bottleneck of expert system, the learning mechanism of CBR can resolve it. By this means the CBR method can make up the lack of the rule-based reasoning and enhance the intelligence of the fault diagnosis system. Finally, according to this design idea, a fault diagnosis expert system is programmed and realized, and its validity and feasibility are proved. Some technical problems such as realizing platform, embedding programming about expert system, and communication between foreground with background are given.