基于本征模式分量符号化分析的滚动轴承故障诊断方法研究

RESEARCH ON FAULT DIAGNOSIS METHOD OF ROLLING BEAR BASED ON SYMBOLIC ANALYSIS OF INTRINSIC MODE FUNCTION

研究在经验模式分解的基础上提出了一种基于本征模式分量符号化分析的故障诊断方法:首先,对故障信号进行拓延,利用经验模式分解获取本征模式分量,实现原始信号在不同尺度下的分离;其次,依据符号动力学理论对本征模式分量进行符号化序列研究;最后,计算符号化后的本征模式分量序列的符号熵,构建特征向量表征故障特性,结合模式分类方法实现诊断。研究通过实验证明了所提出方法对于典型轴承故障有着很好的识别效果,并针对印刷机故障轴承进行实验,成功检测出了不同转速下的轴承故障。所提出诊断方法仅通过时域分量特征提取就获得了较好的诊断效果,相对减少了同类研究中频域特征提取的计算量,具备一定工程应用前景。

A fault diagnosis method of the intrinsic mode components symbolic analysis is proposed on the basis of empirical mode decomposition. Firstly, the extension fault signal is decomposed into several intrinsic mode functions under different time scales with empirical mode decomposition. Secondly, analyze the symbolization of intrinsic mode functions in order to find relationship between different faults hidden in them. At last, calculate entropy of symbolized sequence to establish a characteristic vector which can express fault in some way. And then realize the fault diagnosis with the help of some kind of pattern recognize method. The results show that there is a good recognition effect for typical bearing fault and printing machine fault. Bearing failures in offset printing under different rotational speed are detected successfully. The method has a good effect at features extraction in time domain and reduces computation in frequency domain compare with similar research. The symbolic entropy has good application prospects in engineering.