Application research of Fault Diagnosis Expert System for Photoelectric Tracking Device Based on BP NN

In order to overcome the deficiencies of poor adaptive capacity, lack of inspiration and narrow domain knowledge of expert system and fundamentally improve the diagnostic efficiency, an intelligent fault diagnosis expert system for photoelectric tracking devices, based on BP neural network, is put forward. Firstly, in this paper, some key basic concepts and principles of intelligent fault diagnosis technology are proposed. Secondly, according to the difficulty of multiple and coupling fault diagnosis, after making a comparative analysis of the related BP neural network algorithms, such as the quasi-Newton method, the stretch BP method and the conjugate gradient method, a neural network fault diagnosis reasoning method based on the Levenberg-Marquardt is designed, which combined the implementation of the diagnosis expert system. Finally, several interrelated essential implementation issues, such as the architecture of the system and the VR technology based on OpenGL, are also discussed. Practical experiments and applications demonstrate that the proposed approach is effective, robust and universal.