A method of remote fault diagnosis based on multilayer SOM

This paper establishes the model of remote fault detect system, using typical fault data of a type of planepsilas undercarriage braking system to implement the diagnosis. The result shows that he Ensemble Method of Remote Fault Diagnosis based on PCA and ICA and Multi-layer SOM, integrates the advantage of PCA in low noise and ICA in high noise. Multilayer SOM can classify the input data rapidly with associative memory and parallel processing. The computing results accord with the actual situation and indicate that the ensemble PCA-ICA-MSOM method has large potential in recognizing remote patterns.