Fault classification on vibration data with wavelet based feature selection scheme

Fault classification based upon vibration data is an essential building block of a sophisticated conditional based monitoring (CBM) system. Multiple sensor channels are called for to assure the redundancy and to achieve the desired reliability and accuracy. The shortcoming of using multiple sensor input channels is the need to deal with high dimensional features set, a computational expensive task in classification. It is vital to reduce the feature dimension via an effective feature extraction and feature selection scheme. A simple wavelet based feature selection scheme is proposed. This scheme overcomes the disadvantages faced by existing feature selection methods, by producing a general feature set, reducing the dimensionality of features and requiring no prior information of the problem domain. The proposed feature selection scheme is based on the strategy of "divide and conquer" that significantly reduce the computation time without compromising the classification performance. The simulation results show the proposed feature selection scheme provides at least 65% reduction of the total number of features without compromising the classification performance.