基于IMF希尔伯特解调的风电齿轮箱复合故障识别方法

摘要

为克服风电齿轮箱部件故障各调制载波边带重叠的影响和传统滤波方法造成的信号相位变化,提高故障诊断的精度,引入基于IMF希尔伯特解调的复合故障识别方法.该方法首先通过经验模态分解得到若干个对应不同的调制频率族的内禀模态函数,然后采用希尔伯特解调分析提取调制信号对应的内禀模态分量的故障信息,以达到精确识别故障的目的.结合实际案例分析,验证了该方法可以有效地提取非线性、非稳定性和多调制混杂复合信号中的故障信息,有效地提高了风电齿轮箱故障识别的精度.%In order to improve the accuracy of fault diagnosis and overcome the effect of the each modulated carrier sideband overlapping of the gearbox failure and signal phase changes caused by the traditional filtering methods, a composite fault method based on Hiibert envelope analysis of IMF is proposed in this paper. At first several IMF components corresponding to different modulation frequency clusters is obtained by EMD, and then the Hiibert envelope analysis of the amplitude modulation signal corresponding to the IMF component is used to comply the fault information extraction that can achieve accurate fault identification. Actual case studies verify the method can effectively extract the fault information in the hybrid composite signal of the non-linear, non-stability and multi—modulation and it improves the accuracy of the wind turbine gearbox fault diagnosis effectively.