Three-layer information fusion for braking system fault diagnosis

In this paper, information fusion fault diagnosis technology was applied to a hoist system, and three-layer information fusion fault diagnosis was proposed. Compared with the principal component analysis of fault diagnosis and the two-layer information fusion, the results indicate that Elman neural network has ability with higher accuracy classification and better stability than Error Back Propagation (RBF) Neural Network in small training sample. If more evidence exist, Dempster Shafer (DS) fusion method will be more practical than Yager fusion method. Experiment verified the feasibility of this method. It will improve the accuracy of diagnosis system, and provide greater reliability for coal mine safety production.