中文摘要：

噪声的干扰导致滚动轴承早期故障信号不易被发现,有效滤除噪声信号是实现设备早期故障诊断的关键。对基于Sigmoid函数的变步长自适应算法（G-SVSLMS）进行相关性改进,并将该方法应用于振动信号的初始降噪。根据冗余提升原理对一次降噪后信号进行再降噪,从而实现设备振动信号的有效提取。实验及工程数据分析表明该复合方法的应用效果明显高于单一滤波方法。

英文摘要：

The early fault signal of rolling bearing is not easy to be discovered,effective filtering of the noise signal is key to early fault diagnosis. The correlation of variable step size adaptive algorithm is improved based on Sigmoid function and the algorithm is used to filter the original signal. According to the principle of redundant lifting to reduce the noise signal again,realize the effective extraction of early fault signals.The results of experiments and engineering data show that this method is more effective than single filtering method.