EMD算法和循环平稳理论在机械故障诊断中的应用

The Integrated Application of EMD Algorithm and Cyclostationary Theory in Signal Extraction

针对非平稳非线性的循环平稳信号提出了基于Hilbert变换的EMD分解重构消噪法;用带有强噪音和高斯噪声的典型循环平稳信号进行了仿真,并结合二阶循环平稳分析法对信号进行了信息提取.在实验部分,该文试用基于Hilbert变换的EMD分解重构和循环平稳理论结合的方法对运转中的故障轴承信号进行了分析,实验结果表明,这种应用能有效提取出信号中的有用信息,成功提取旋转机械运转产生的信号中的故障信息.

In view of the cycle of non-stationary non-linear signals,we present a new reconstruction method basing on the EMD and Hilbert transform to remove the noise,and then we make a simulation with a typical cycle stationary signal which is disturbed by strong random and Gaussian noises.In the experiment,combining the cycle stationary with the method based on the EMD Hilbert transform decomposition and reconstruction,we analyze the fault of the bearing signal.Results show the method can extract effectively the useful information;moreover the fault signals due to the rotating machinery also can be extracted effectively.