Application of Adaptive Local Iterative Filtering Decomposition Energy Distribution and LSSVM in Gear Fault Recognition

Abstract

Focusing on the gear measured signal can not accurately reflect the fault characteristics due to noise interference, an adaptive local iterative filtering (ALIF) decomposition combining energy distribution with Least Squares Support Vector Machine (LSSVM) is introduced. ALIF can decompose the non-stationary signal of gear into finite stationary intrinsic mode functions. By calculating the energy distribution of each intrinsic mode function (IMF), it could be found that the energy distribution of the first several IMF can represent the characteristics of different fault types. By calculating the normalized energy distribution of the IMF of vibration signal decomposition under different working conditions, LSSVM distinguishesdifferent gear fault type. Results show the proposed method provides a new way for gear fault recognition.