Fault diagnosis of rolling bearing based on empirical mode decomposition and improved manhattan distance in symmetrized dot pattern image

Abstract

The vibration signal of rolling bearing is typical non-stationary signal, it is difficult for traditional time–frequency analysis technology to apply. Therefore, a novel fault diagnosis approach based on improved manhattan distance in Symmetrized Dot Pattern (SDP) image is proposed. In this method, the sample vibration signal data is divided as 10 equal parts on average, and empirical mode decomposition (EMD) is used to decompose the equally-divided signals into several Intrinsic Mode Function (IMF) components. The first five IMF components are adopted to transform into five symmetrical snowflake image in polar coordinates, after denoising corresponding feature matrices are calculated, and the mean feature matrix and the maximal characteristic root may be obtained. In this way the improved manhattan distance between the local matrix of each IMF components and corresponding mean matrix is extracted. Different vibration signal of rolling bearing is classified according to this improved manhattan distance. Finally, an experimental experiment is carried out to verify the feasibility and effectiveness of this novel fault diagnosis approach.