摘要
采用提升多小波的方法,以峭度为优化目标,遗传算法为优化算法,针对信号特征进行多小波的自适应构造.以构造的多小波作为谱峭度的滤波器,针对多小波特点改进了峭度图,提出了多小波谱峭度方法.该方法建立在传统谱峭度方法的基础上,不仅克服了原方法中滤波器变化有限的劣势,而且提高了谱分析的分辨率.将该方法应用于滚动轴承的故障诊断中,以试验台与电力机车的滚动轴承故障诊断为例进行验证,结果表明,该方法不仅提高了频带选择的准确性与滤出信号的信噪比,而且获得了更好的诊断效果.
A new multiwavelet is constructed according to the signal feature,which utilizes mulitiwavelet lifting,taking kurtosis as the optimal object and genetic algorithm as the optimal strategy.The new multiwavelet is substituted for the filter in spectral kurtosis（SK） and the kurtogram is improved accordingly.The multiwavelet spectral kurtosis is proposed based on the conventional SK,which makes up the SK in filter limitation and improves the resolution of spectral analysis.The examples for rolling bearing in test rig and locomotive result indicate the improved accuracy of frequency-band-selection and the signal-to-noise-ratio of signal filtered.
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2010年第3期77-81,共5页
Journal of Xi'an Jiaotong University
基金
国家高技术研究发展计划资助项目(2006AA04Z430)
教育部高等学校博士学科点专向科研基金资助项目(200806980011)
关键词
多小波
自适应构造
谱峭度
滚动轴承
故障诊断
multiwavelet
adaptive construction
spectral kurtosis
rolling bearing
fault diagnosis
