Aiming at that the gearbox vibration signals are nonlinear and non-stationary

a fault diagnosis method based on the theory of multi-fractal and support vector machine（SVM） is proposed.First the multi-fractal theory are applied to analyze gearbox vibration signals

the analysis results show that multi-fractal spectrum and general dimension give a good presentation for gearbox working condition.Then particle swarm optimization is applied to optimize the parameter of support vector machine.The multi-fractal characteristic parameters of gearbox vibration signals are regarded as the fault characteristic vectors and served as input parameters of SVM classifier to classify the fault types of the gearbox.The experimental results show that in the case of small number of samples

this method can classify the different fault types of gearbox accurately.