针对水轮机导轴承工作环境复杂,测得信号掺杂噪声给信号处理带来困难的问题,提出一种基于独立分量分析的水轮机振动信号的信噪分离方法,并通过信号仿真验证了该方法分离两源混合信号(原始信号和噪声)的准确性。在此基础上引入经验模态分解算法对采集振动信号的噪声分离,经过仿真信号与实测信号分离的对比,验证了所提方法在水力机械故障诊断方面的工程应用价值。 In view of the complicated working environment of turbine guide bearing and the difficulty of signal processing caused by signal noise measurement, a method of signal-noise separation of turbine vibration signal based on independent component analysis is proposed. The signal simulation shows that the method separates two Accuracy of source mixed signal (original signal and noise). On the basis of this, the empirical mode decomposition algorithm is introduced to separate the noise of the vibration signal. The comparison between the simulation signal and the measured signal shows that the proposed method is applicable to hydraulic machinery fault diagnosis.