关键词: 神经网络, L-M算法, 专家系统, 故障诊断

Abstract: An improved neural network based on L-M algorithm was applied to fault diagnosis expert system against to the slow convergence rate of conventional BP neural network .The expert system structure based on neural network was introduced, and a fault diagnosis model was designed combining with 7216 tapered bearings experiment .Simulation results indicated this model remarkably reduced the training time, with its relatively high accuracy, surpassed the conventional BP neural network model. It was feasible for mechanical fault diagnosis.

Key words: neural network, L-M algorithm, expert system, fault diagnosis