摘要
针对直升机自动倾斜器滚动轴承转速低、钢球多,故障特征频率难提取的问题,利用VMD-SE和DEELM对直升机滚动轴承进行故障诊断。首先,对振动信号进行变分模态分解,利用中心频率法选择模态数,通过样本熵提取模态分量的特征;其次,利用差分进化算法对ELM进行优化;最后,利用课题组测得的真实数据进行故障诊断试验。结果表明:相比ELM和BA-ELM方法,DE-ELM在隐含层节点数较少情况下对轴承故障识别率达到99%以上,验证了该方法用于直升机自动倾斜器滚动轴承故障诊断的有效性。
Aiming at difficulty in characteristic frequency extraction of fault for helicopter swashplate rolling bearings with a number of steel balls under low speed,the fault diagnosis forthe bearings is carried out based on VMD-SE and DE-ELM. Firstly,the variation mode decomposition is used to decompose vibration signal,the mode number is selected by center frequency method,and the features of mode components are extracted by sample entropy. Then,the differential evolution algorithm is used to optimize ELM. Finally,the real data are measured by research group to carry out fault diagnosis experiments. The results show that compared with ELM and BA-ELM methods,the bearing fault recognition rate obtained by DE-ELM method is more than 99% in the case of fewer hidden nodes,and the validity of method is verified for fault diagnosis of helicopter swashplate rolling bearings.
出处
《轴承》
北大核心
2017年第8期53-57,共5页
Bearing
基金
航空科学基金项目(2016ZD56008
2010ZD56009)
江西省教育厅科学技术研究项目(GJJ14519)
