VMD和CICA在齿轮箱故障诊断中的应用

Application of VMD and CICA in fault diagnosis of gearbox

针对齿轮箱故障信号微弱且易受周围噪声影响的问题,提出了一种基于变分模态分解(variational mode decomposition,VMD)的约束独立分量分析(constrained independent component analysis,CICA)算法。利用单通道加速度传感器采集齿轮箱的混合故障信号,通过VMD算法将混合信号分解为多个不同的本征模态函数(intrinsic mode function,IMF),然后依据峭度和互相关系数选取有效的IMF分量进行重构信号,对于重构信号利用CICA进行降噪处理,根据CICA降噪后得到齿轮和轴承的故障特征,对齿轮和轴承混合故障进行仿真及实验研究,结果表明,VMD-CICA算法可以很好地提取齿轮和轴承的故障特征频率,同时与经验模态分解-约束独立分量分析(EMD-CICA)和集成经验模态分解-约束独立分量分析(EEMD-CICA)算法相比得到的故障特征频率更明显。

A constrained independent component analysis(CICA)algorithm based on variational mode decomposition(VMD)is proposed to solve the problem of weak fault signals and susceptible to ambient noise.A single-channel acceleration sensor was used to collect the mixed fault signal of gearbox,and the mixed signal was decomposed into multiple intrinsic mode functions(IMF)by VMD algorithm.Then,the effective IMF components were selected according to kurtosis and correlation coefficient to reconstruct the signal.For the reconstructed signal,the CICA was used to de-noise,and the fault characteristics of gears and bearings were obtained according to the CICA after de-noising.Through the simulation and experimental study of the mixed fault of gear and bearing,the results show that the VMD-CICA algorithm can extract the fault characteristic frequency of gear and bearing well.Compared with the empirical mode decomposition-constrained independent component analysis(EMD-CICA)and the integrated empirical mode decomposition-constrained independent component analysis(EEMD-CICA)algorithm,the characteristic frequency of barrier is more obvious.