针对齿轮箱振动信号的非平稳性和非线性, 提出一种多重分形和支持向量机相结合的故障

诊断方法。运用多重分形理论方法对齿轮振动信号进行分析, 通过分析发现多重分形谱和广义维数作

为故障特征能够很好地反映齿轮箱的工作状态 对支持向量机的参数利用粒子群优化算法进行优化, 并

将齿轮箱振动信号的多重分形特征量作为支持向量机的输入参数以识别齿轮的故障类型。实验结果表明, 该方法在样本较小的情况下能够准确对齿轮箱的故障类型进行分类-Gearbox vibration signal of non-stationary and non-linear, the combination of a multifractal support vector machine fault diagnosis method. The use of multiple fractal theory methods to analyze gear vibration signals, through the analysis found multifractal spectrum and broad-dimensional number as the fault feature able to well reflect gearbox work state　support vector machine parameters utilize particle swarm optimization algorithm to optimize and the multifractal characteristic quantities of the gearbox vibration signal as the input parameters of support vector machine to identify the type of gear failure. Experimental results show that the method in the case of a smaller sample can be accurately classify the type of failure of the gear box