摘要
针对传统形态学结构元素选择的不确定性和广义形态学结构元素间相互影响的缺点,提出一种根据局部极值步长确定形态学结构元素尺寸的方法,充分利用信号的局部信息和局部极值步长,达到自适应选取最优结构元素尺寸的效果,解决了形态学结构元素选取时存在的不足。针对数学形态学在强噪声下滤波效果不理想这一不足,构造了广义形态学差值滤波器,将其与传统形态学滤波器进行仿真对比,结果显示广义差值滤波器的降噪和故障特征提取的效果明显优于传统形态滤波器,并将其应用到滚动轴承故障信号的特征提取中,结果表明该方法能够有效的抑制噪声,明显的提取滚动轴承的故障信息特征,实现滚动轴承的故障诊断。
The traditional morphological structure element size selection is blindness, generalized morphological structure elements have influence on each other, which makes the structure element selection difficulty, so it put forward the local extreme point step length to determine the morphological structure elements size, making full use of local information and signals local extreme step to select the optimal adaptive structure element size, it effectively solve the shortness. Aimed at that mathematical morphology under strong noise processing effect is not ideal. It builds generalized morphological difference filter, to denoise the fault signal and to extract fanlt feature in strong noise environment. Comparing generalized morphological difference filter simulation with the traditional morphology filter simulation, the simulation results show that the generalized difference filter processing effect is better than the traditional morphological filters. It is applied to rolling bearing fault signal feature extraction, and the results show that this method can effectively suppress noise and extract fault information characteristic of rolling bearing, realizing the fault diagnosis of rolling bearing.
出处
《机械设计与制造》
北大核心
2015年第6期25-29,共5页
Machinery Design & Manufacture
基金
国家自然科学基金资助项目(51475339)
关键词
广义数学形态学
局部极值
故障特征频率
轴承故障
Generalized Mathematical Morphology
Local Extreme Point
Fault Characteristic Frequency
BearingFault
