Weak fault detection of rail vehicle suspension system based on MPCA

Along with the rapid development of urban rail transit, the security and reliability of rail transit vehicle system draws more and more attention. Suspension system plays a key role in supporting the car body, bogies and wheels, therefore it is vitally important to implement real-time performance monitoring and weak fault detection. This paper introduces a multilinear principle component analysis (MPCA) framework for weak fault feature extraction by using tensor data. The aim of the proposed framework is to determine a multilinear projection that captures most of the variation on the original tensor input. Experimental results show that MPCA discussed in this paper achieves better performance for weak fault detection and isolation compared with classical principle component analysis (PCA) and its extension algorithm dynamic principle component analysis (DPCA).