一种脊线提取方法在轴承故障诊断中的应用  

Applications of a Ridgeline Extraction Method in Bearing Fault Diagnosis

作　　者：陈剑[1,2] 杨斌[1,2] 黄凯旋 蔡坤奇 刘圆圆 刘幸福 CHEN Jian;YANG Bin;HUANG Kaixuan;CAI Kunqi;LIU Yuanyuan;LIU Xingfu(Institute of Sound and Vibration Research,Hefei University of Technology,Hefei,230009;Automotive NVH Engineering&Technology Research Center Anhui Province,Hefei,230009)

机构地区：[1]合肥工业大学噪声振动研究所,合肥230009 [2]安徽省汽车NVH工程技术研究中心,合肥230009

出　　处：《中国机械工程》2021年第10期1157-1163,共7页China Mechanical Engineering

基　　金：国家自然科学基金(11604070);安徽省科技重大专项(17030901049)。

摘　　要：针对Crazy Climber算法脊线识别能力低、提取脊线与实际脊线有差距等问题,通过对该算法中Climber的移动规则和局部最优峰值提取方法进行改进,提出了一种改进的Crazy Climber算法。首先对时域信号进行短时傅里叶变换获取时频矩阵,再用改进的Crazy Climber算法对时频矩阵进行脊线提取,最终获得可以反映频率随时间变化的时频脊线。仿真信号分析结果证明了该算法优于原始的Crazy Climber算法。将该算法用于轴承的故障诊断中,可准确提取故障阶次。To solve the problems of low ridgeline identification ability and differences between the extracted ridge and the actual ridge in Crazy Climber algorithm,the Climber's movement rule and local optimal peak extraction method in the algorithm were improved.An improved Crazy Climber algorithm was proposed.First,STFT was applied to the time-domain signals to obtain the time-frequency matrix,and then the improved Crazy Climber algorithm was used to extract the time-frequency matrix to obtain the time-frequency ridge that might reflect the frequency changes with time.The simulation signal analyses prove that this method is superior to the original Crazy Climber algorithm.The method is used in the fault diagnosis of bearings,and the fault order is accurately extracted.

关 键 词：时频分析 短时傅里叶变换 Crazy Climber算法 脊线提取 故障诊断 

分 类 号：TH17[机械工程—机械制造及自动化]