摘要
针对形态学滤波在处理含有大量噪声的机械振动信号中结构元素长度不能自适应调整的问题,提出了一种可自适应调整结构元素长度的形态学滤波新方法。结构元素长度的不同会导致对信号特征提取效果的不同。该方法通过以峭度值为指标,找寻出使峭度值最大的一系列结构元素长度。然后通过计算不同长度滤波后信号的冲击特征比值,找寻出使故障特征最突出、最明显的结构元素长度。以此长度为最优长度对信号进行滤波,能够较好地提取出滚动轴承的故障特征,找到故障特征频率。通过实验台信号验证了所述方法的有效性和优点。
Aiming at resolving the problem that the structure element length can not be adjusted adaptively when morphological filter is processing mechanical vibration signal with too much noise,a new method of morphological filter which can adjust structure element length adaptively is presented.The difference of the length of the structure element leads to the diversity of the effect of the feature extraction signal.By using kurtosis as the index,the method identifies a series of structural element length which makes the value maximum.Through calculating the impact characteristic ratio of signals filtered in different length,the length is found,which makes the fault feature most prominent and obvious.Taking this length as the optimal length of the signal filtering,the fault feature can be extracted,and the fault feature frequency can be obtained.The availability and advantages of the method are verified through signals at the experimental station.
出处
《石家庄铁道大学学报（自然科学版）》
2018年第1期100-104,共5页
Journal of Shijiazhuang Tiedao University(Natural Science Edition)
基金
国家自然科学基金(11227201
11372199
11572206)
河北省自然科学基金(A2014210142)
关键词
形态学滤波
峭度
冲击特征比值
故障特征频率
morphological filter
kurtosis
impact characteristic ratio
fault feature frequency
