Integrated Neural Network Fault Diagnosis Technique and Applications

The features and existed problems of single sub-NN fault diagnosis are analyzed for advancing the veracity of NN fault diagnosis extremely, INN based on signal types and various eigenvector combinations is introduced for information fusion. A three-layer-series INN is constructed by building information distributed network, sub-NN as well decision-making information fusion network. Diagnosing from all aspects, the effect is outstanding with input signals assembled.