小波阈值函数旋转机械振动信号去噪方法研究

从强背景噪声中提取出微弱的旋转机械振动故障特征信号一直是技术性难题。针对传统全局阈值函数去噪在阈值处不连续和存在恒定偏差的问题,提出一种改进的小波阈值函数分层去噪方法。首先对旋转机械故障信号去噪中的小波参数进行了筛选,然后采用改进的阈值函数,利用最优小波参数对振动信号进行分层阈值降噪处理。理论仿真和实测结果表明,对比传统阈值去噪方法,该方法能有效去除背景噪声,保留振动信号原貌特征信息,提高信噪比和减小均方根误差,适合非平稳振动信号去噪,为旋转机械故障诊断奠定了信号预处理的基础。

Research on Rotating Machinery Vibration Signal Denoising with Wavelet Threshold Function

Extracting weak fault characteristics of rotating machinery vibration signal from strong noisy background has been a technical difficulty.To solve the problems of discontinuity at threshold point and constant deviation in traditional global function,an improved wavelet threshold fuction layered denoising approach is proposed.Firstly,the wavelet parameters are selected from the rotating machinery fault signal denosing,then the vibration signal is de-noised under the improved threshold fuction and optimal wavelet parameters.Theoretical simulation and measured results demonstrate that this approach performs better in suppressing the background noise,keeping original vibration signal characteristic information,enhancing the signal-to-noise ratio,and reducing mean-square-error than traditional threshold denoising method.Therefore,the proposed approach works well in non-stationary vibration signal denoising,and thus lays a foundation for the rotating machinery fault diagnosis and signal processing.