有向无环图决策支持向量机和经验模式分解在轴承故障诊断中的应用

当旋转机械发生故障时,其振动信号常常表现出较为复杂的调制形式,经验模式分解能根据信号的真实物理意义完成自适应分解。支持向量机由于其出色的学习性能和良好的推广能力,使其在包括故障诊断在内的众多领域得到较为广泛的应用。利用经验模式分解结果提取频带能量特征向量,采用有向无环图决策支持向量机实现对轴承状态的判别,并基于留一法优化支持向量机的模型参数。最终的应用结果表明,基于EMD和有向无环图决策支持向量机方法可以有效实现对轴承的状态判别。

Application of EMD and DDAGSVM in Bearing Fault Diagnosis

When fault occurs,vibration signal of rotation machine behaves in complex form of modulation.The method of EMD can adaptively decompose signals based on the physical meaning of signal.As SVM has excellent learning performance and favorable generalization performance,SVM has been used in many fields including fault diagnosis.In this paper,energy eigenvector of frequency band is extracted through EMD.State judgment for bearing is realized by DDAGSVM.The most excellent model parameter is supported based on LOO...