摘要
为提取被噪声干扰的有效轴承故障特征信息,提出一种SVD归一化强度降噪方法,并对有效奇异值进行软阈值处理,降低噪声的干扰。轴承信号降噪结果表明,在不同的噪声强度干扰下,该降噪方法均能够保留源信号中的主要频率信息,大幅提高信噪比;该方法较SVD差分谱方法能够有效提取故障特征频率信息,避免信号的过降噪。SVD归一化强度软阈值降噪方法降噪原理清晰、计算简便,能够为轴承故障的精密诊断提供可靠的数据基础。
In order to extract bearing fault characteristic information disturbed by noise, a SVD denoising method based on normalized intensity was proposed. The soft threshold processing was conducted to the effective singular value in order to restrain the noise. The noise reduction results of bearing signal showed that, this method can retain the main frequency characteristic information in the source signal and greatly improve the signal-to-noise ratio of the simulation signals under different noise intensity interference, can effectively extract fault characteristic frequency information than SVD difference spectrum method, and can avoid excessive signal noise reduction. The proposed method has clear principle and simple calculation, and can provide reliable data basis for further accurate diagnosis of bearing faults.
作者
陈涛
王立勇
唐长亮
徐小力
CHEN Tao;WANG Li-yong;TANG Chang-liang;XU Xiao-li(The Ministry of Education Key Laboratory of Modern Measurement and Control Technology, Beijing Information Science and Technology University, Beijing 100192, China)
出处
《组合机床与自动化加工技术》
北大核心
2019年第8期71-75,共5页
Modular Machine Tool & Automatic Manufacturing Technique
基金
国家自然科学基金资助项目51605035
北京市教委科技计划项目(KZ201611232032,KZ201611232004)
关键词
归一化强度
有效降噪阶次
软阈值处理
轴承故障诊断
normalized intensity
effective denoising order rank
soft threshold processing
bearing fault diagnosis
