|
|
FAULT DIAGNOSIS MODEL BASED ON LEAST SQUARE SUPPORT VECTOR MACHINE OPTIMIZED BY LEAVE-ONE-OUT CROSS-VALIDATION |
LI Feng;TANG Bao-Ping;ZHANG Guo-Wen
|
(The State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing 400030, China) |
|
|
Abstract A new fault diagnosis model is proposed based on Least square support vector machine(LS-SVM) optimized by Leave-one-out cross-validation: firstly, fault vibration signals are decomposed into several stationary IMFs, then the instantaneous amplitude L2-norms of the IMFs which have fault modulation characteristics are computed and regarded as the input characteristic vector of lin-kernel LS-SVM optimized by Leave-one-out cross-validation for fault classification. EMD decomposition adaptively isolates fault modulation signals from original signals. The differences among instantaneous amplitude L2-norm vectors reflect different fault type. Adopting Leave-one-out cross-validation to optimize punish parameter can not only enhance fault prediction accuracy of LS-SVM which adopts lin-kernel, but also improve adaptive diagnosis capacity of LS-SVM. The deep groove ball bearings fault diagnosis example proves the effectivity of this new model.
|
Received: 11 June 2009
|
|
|
|
|
|
|