2021 International Conference on Sensing, Measurement and Data Analytics in the era of Artificial Intelligence (ICSMD2021)

111 / 2021-07-30 18:43:36

An Improved Symplectic Geometry Mode Decomposition Method for Rolling Bearing Fault Diagnosis under Variable Speed Conditions

Special Session > SS4 - Weak Signal Processing and Incipient Fault Detection

Fault diagnosis of rolling bearings under variable speed conditions is of vital importance and also quite challenging in industrial areas. Tacholess order tracking (TLOT) methods, which are based on successive time-frequency analysis, ridge detection and components extraction, would be probably interfered by some tough problems, such as error accumulation and inapplicable computational burden. In this paper, an improved symplectic geometry mode decomposition (ISGMD) method is proposed. Based on the symplectic geometry similarity transformation and Jensen-Shannon divergence (JSD), Harmonic components are extracted without the limitation of prior knowledge. With the harmonic relationship identification, the TLOT is conducted to detect the local defect of the rolling bearing utilizing the order spectrum. The vibration signal collected from a SpectraQuest test-rig is used for validation. The experimental result exhibits that the proposed method is more accurate and flexible in vibration signal analysis under variable speed conditions.