A novel compound neural network for fault diagnosis
WeiDong Jiao,ShiXi Yang,GongBiao Yan
DOI: https://doi.org/10.1109/MESA.2006.297004
2007-01-01
Abstract:Independent Component Analysis (ICA) is a powerful tool for redundancy reduction and nongaussian data analysis. And, Artificial Neural Network (ANN), especially the Self-Organizing Map (SOM) based on unsupervised learning is a kind of excellent method for pattern clustering and recognition. By combining ICA with ANN, we proposed a novel compound neural network for pattern classification. First, two neural ICA algorithms were applied to fusion of multi-channel measurements by sensors. Moreover, a unit for further feature extraction was used to capture statistical features higher than second order, which embedded into the measurements. Second, certain a typical neural classifier such as Multi-Layer Perceptron (MLP), Radial Basis Function (RBF) or SOM was trained for the final pattern classification. The results from contrast experiments in fault diagnosis show that the proposed compound neural network with ICA based feature extraction can classify various fault patterns at considerable accuracy, and be constructed in simpler way, both of which imply its great potential in fault diagnosis.