【作者】 文莉； 刘正士；

【Author】 Wen Li;Liu Zhengshi;

【机构】 合肥工业大学动态测试中心；

【摘要】 平移不变量小波去噪方法是在Donoho的阈值法基础上的改进。在信号的不连续点邻域,用阈值法去噪会产生伪吉布斯现象。仿真实验可以看出,该方法和阈值法相比,不仅能有效的抑制伪吉布斯现象,而且能减小原始信号和估计信号之间的均方根误差,提高信噪比。将这种方法用于变速箱齿轮故障信号的去噪处理,同时与阈值法去噪的结果进行了比较,结果表明,该方法可以有效地去除强噪声的干扰,提取齿轮故障特征信息,具有很好的工程实用性。更多还原

【Abstract】 Translation Invariant Wavelet Denoising is an improvement upon Donoh’s thresholding denoising method.In the neighborhood of discontinuities of the signal,the thresholding denoising may cause pseudo-Gibbs phenomena.Simulated experiments show that as compared to the thresholding denoising,the translation invariant denoising not only can suppress pseudo-Gibbs phenomena,but also will reduce RSME between the original signal and estimated signal.At the same time,SNR of the estimated signal is also improved.We apply this method to denoise the gearbox gear fault signal and compare this method with the thresholding denoising.The results show that the translation invarinat denoising can greatly remove noise and extract the gear fault characteristic imformation effectively.It also has practical engineering applications.更多还原