NONLINEAR DYNAMICAL PRINCIPLE

OF MECHANICAL FAULT DIAGNOSIS

CHEN Yushu

(School of Astronautics, Harbin Institute of Technology, Harbin 150001)

Abstract: The dynamical principle of fault diagnosis is addressed, and the initial concept of nonlinear dynamical principle of mechanical fault diagnosis is introduced. The mechanical devices can be divided into two classes of systems, namely: possible-of-modeling systems and impossible-of-modeling systems. For possible-of-modeling systems, the analysis of fault mechanism based on the bifurcation theory may provide an instruction to mechanism, control and prediction of vibrating faults which are difficult to diagnose. The modeling method is described using the rotating machinery as an example. For impossible-of-modeling systems, using the chaotic dynamical theory, the phase space can be reconstructed in terms of the measuring vibration data, and the sources of generating faults can be justified through the form of singular spectrum which describes the energy distribution of the system. The studies on the nonlinear dynamical principle of fault diagnosis may significantly enhance the ability of our manufacturing industry of complicated devices and also make great contribution to the nonlinear dynamical theory.

Key words: Nonlinear dynamics Principle of fault diagnosis Possible-of-modeling systems Impossible-of-modeling systems Singular spectrum analysis

CLC No: TH113

2006年中国机械工程学会年会推荐论文. Received 20061205,  received in revised form 20061215