基于AP-FBSOMP组合算法的轴承微弱故障智能诊断分析  

Intelligent diagnosis and analysis of bearing weak faults based on AP-FBSOMP combined algorithm

作　　者：于春霞[1] 张建国[2] 李明 YU Chunxia;ZHANG Jianguo;LI Ming(Department of Computer Science,Yellow River Institute of Science and Technology,Zhengzhou 450063,Henan,China;School of Mechanical Engineering,Henan Polytechnic University,Zhengzhou 451460,Henan,China;Technical Development Department,Henan Litian Tool Co.,Ltd.,Zhengzhou 450053,Henan,China)

机构地区：[1]黄河科技学院计算机系,河南郑州450063 [2]河南理工大学机械工程学院,河南郑州451460 [3]河南力天刀具有限公司技术开发部,河南郑州450053

出　　处：《中国工程机械学报》2022年第2期178-183,共6页Chinese Journal of Construction Machinery

基　　金：河南省民办高等学校品牌专业建设资助项目(ZLG201903)。

摘　　要：为增强轴承微弱故障信号分块压缩感知(CS)重构性能,综合运用信号分块自适应(AP)以及K奇异值分解(KSVD)优化正交匹配追踪(OMP)算法,设计一种AP-FBSOMP组合算法,并开展故障智能诊断实测分析。研究结果表明:随着重构支撑集的有效支撑集原子数增加后,获得了跟初始信号误差更低的重构信号,重构信号达到了更低误差,显著优化了信号重构效果。相对OMP、SP、SOMP的重构时间明显缩短,FBSOMP算法都具备更优的重构结果。利用不同算法处理相对误差及其重构时间测试结果得到,以FBSOMP算法重构相对误差最小,能够更加快速达到重构的效果。信号重构可以达到准确重现初始信号故障特征的状态,有效实现了理想的匹配效果。FBSOMP算法则可以实现高效运行的性能,由此大幅提高重构信号精度。In order to enhance the reconstruction performance of weak bearing fault signals by block compression sensing(CS),an AP-FBSOMP algorithm was designed by using the AP and KSVD optimized OMP algorithm.The results show that with the increasing of the atomic number of effective support set,the reconstructed signal with lower error than the original signal is obtained,and the reconstructed signal achieves lower error,which significantly optimizes the signal reconstruction effect.Compared with OMP,SP and SOMP,FBSOMP algorithm has better reconstruction results.By using different algorithms to deal with the relative error and reconstruction time test results,FBSOMP algorithm reconstruction relative error is the smallest,which can achieve the reconstruction effect more quickly.The signal reconstruction can accurately reproduce the initial signal fault features and effectively achieve the ideal matching effect.FBSOMP algorithm can realize efficient operation performance,thus greatly improving the reconstructed signal accuracy.

关 键 词：压缩感知(CS) 自适应分块 重构算法 稀疏字典 

分 类 号：TG156[金属学及工艺—热处理]