共振解调是滚动轴承故障诊断中最常用的方法之一，然而其带通滤波器参数的选取通常比较困难。谱峭度法能根据峭度最大化原则自动确定带通滤波器参数，取得了一定的诊断效果，但由于滚动轴承的早期故障信号中含有强烈的背景噪声，诊断效果有时也不够明显。为此，本文提出一种基于EMD降噪和谱峭度法的滚动轴承早期故障诊断新方法，首先采用基于互相关系数和峭度准则的EMD降噪对采样信号进行预处理，突出高频共振成分，再利用谱峭度法选取最佳带通滤波器参数，最后使用带通滤波和包络解调进行故障诊断，并通过实际工程信号进行了验证。

It is difficult to select the band-pass filter parameters of resonance demodulation which is one of common methods for diagnosing rolling element bearing, but it can be automatically determined by spectral kurtosis based on the principle of max- kurtosis. Though some diagnosis effect has been achieved by this method, it is not obvious enough sometimes due to the strong background noise involved in the fault early signal of rolling element bearing. Therefore, a new fault early diagnosis method of rolling element bearing based on EMD denoising and spectral kurtosis is proposed in this paper. Firstly, the sampling signal is preprocessed by EMD denoising which is combined with the criterions of cross-correlation coefficient and kurtosis, which can highlight high-frequency resonance components. Secondly, the optimal band-pass filter parameters are selected by spectral kurtosis. Finally, fault diagnosis is implemented by band-pass filter and envelope demodulation. The application in a practical engineering signal verifies its capability.