Complexity as a measure for machine fault detection and diagnosis

Abstract

Conventional techniques used for the deteciion and diagnosis of machine defects, such as spectral analysis and timefrequency analysis, are based on the assumption that a physical sysrem possesses linear transfer functions. However, these techniques cannot truthfully identify fault features when the actual behavior of the physical system is far from linear due to the change of its operating conditions, and involves nonlineari@. This paper presents a nonlinear dynamics method called complexi@, which has been investigated to extract feature parameters from raw vibration signals measured from a bearing system. The result3 demonstrated that complexi@presents a good measure for detecting machine defects.