针对传统的机械故障源分离方法限于非高斯、平稳和相互独立的源信号,且传感器的观测数多于源数,机械故障源信号通常不易满足这些假设的局限性,提出了一种基于经验模态分解的机械故障欠定盲源分离方法.该方法对混合观测信号进行经验模态分解,然后重组本征模函数分量作为新的观测信号进行盲源分离,仿真和实验结果表明,该方法是有效的. In view of the traditional method of mechanical fault source separation is limited to non-Gaussian, stationary and independent source signals, and the number of sensors is more than the number of sources, the mechanical fault source signal is usually not easy to meet the limitations of these assumptions, State decomposition of mechanical fault underdetermined blind source separation method. The method of empirical observation of the mixed signal decomposition, and then recombined eigenmodel components as a new observation signal blind source separation, simulation and experimental results show that the method is Effective.