IFD与KELM结合的滚动轴承故障诊断方法

针对滚动轴承振动信号非线性、非平稳的特点,提出基于迭代滤波分解(iterative filtering decomposition, IFD)提取各分量特征,结合核极限学习机(kernel extreme learning machine, KELM)的故障诊断方法.通过对原始信号进行IFD分解,得到一组本征模态函数(intrinsic mode functions, IMF).计算包含主要故障信息在内的IMF分量能量与排列熵组成的故障特征向量,将特征向量作为KELM输入识别轴承的故障类型.实验分析结果表明,以IFD作为预处理器的特征融合方法比经验模态分解(empirical mode decomposition, EMD)为预处理器的特征融合方法有更高的故障识别率,并且该方法在少量样本情况下仍能有效识别故障类型.

Fault diagnosis method of rolling bearing based on IFD and KELM

Aiming at the nonlinear and non-stationary characteristics of the rolling bearing vibration signal, a fault diagnosis method based on iterative filtering decomposition (IFD) is proposed to extract the features of each component and combined with the kernel extreme learning machine (KELM). By IFD decomposition the original signal into a set of intrinsic mode functions (IMF), the energy and permutation entropy of the IMF components that containing the main fault information are calculated to form a fusion feature vector, the feature vector is intruduced as the input of the KELM to identify the fault type of the bearing. The experimental results show that the feature fusion method with IFD as the preprocessor has higher fault recognition rate than the empirical mode decomposition (EMD) as the preconditioner''s feature fusion method, and the method can still effectively identify the fault type in the case of a small number of samples.