Gear fault feature extraction and diagnosis method under different load excitation based on EMD, PSO-SVM and fractal box dimension

Abstract

Aiming at the problem of gear fault feature extraction and fault classification under different load excitation, we present a new fault diagnosis method that combines three methods, including empirical mode decomposition (EMD), particle swarm optimization support vector machine (PSO-SVM) and fractal box dimension. First, the non-stationary original vibration signal of gear fault is decomposed into several intrinsic mode functions (IMF) by EMD method. Then, the time, frequency, energy characteristic parameters and box dimension are calculated separately from the time domain, frequency domain, energy domain and fractal domain. And then the gear fault characteristics under different load excitation are obtained. Finally, the extracted feature parameters are input into the PSO-SVM model for gear fault classification. The experimental results show that the proposed method can effectively identify gear failure types under different load excitation.