摘要
针对轴承在运行过程中状态难以预测的问题,提出一种基于最小二乘支持向量机(LSSVM)结合分形盒维数的方法预测滚动轴承状态。首先,分析滚动轴承在正常和三种故障状态下的盒维数和峭度值;然后,分析盒维数和峭度值对轴承运行状态的描述;最后,借助改进LSSVM预测轴承信号的盒维数和峭度值。实验结果表明,分形盒维数能对正常状态、内外圈及滚动体故障进行区分,结合峭度值能提高分类识别效果,改进的LSSVM方法能准确地预测特征参数,从而实现对轴承状态的预测。
Aiming at the difficulty to predict the condition of a rolling bearing during its operation,the improved LSSVM with fractal box dimension method is proposed.Firstly,the box dimension and kurtosis values of rolling bearings under normal and three fault conditions are analyzed respectively.Then,on this basis,the description of the operating state of the bearing by the box dimension and kurtosis values is analyzed.Finally,the box dimension and kurtosis values of bearing signals are predicted by using the improved LSSVM.The experimental results show that the improved LSSVM method combined with fractal box dimension can effectively distinguish the normal from rolling bearing faults and the inner and outer ring faults.By combining the kurtosis value the classification and recognition effect can be improved.At the same time,the method can predict the characteristic parameters more accurately,so as to realize the prediction of the bearing state.
作者
刘天顺
谷晓娇
李时雨
LIU Tianshun;GU Xiaojiao;LI Shiyu(Shenyang Ligong University,Shenyang 110159,China)
出处
《沈阳理工大学学报》
CAS
2023年第3期82-87,共6页
Journal of Shenyang Ligong University
基金
国家自然科学基金项目(52004165,51905357)
沈阳理工大学博士科研启动基金项目(1010147000818)。
关键词
滚动轴承
盒维数
特征参数
最小二乘支持向量机
rolling bearing
box dimension
characteristic parameter
least square support vector machine
作者简介
刘天顺(1995-),男,硕士研究生;通信作者:谷晓娇(1989-),女,副教授,博士,研究方向为设备故障诊断、运行状态监测、安全性预测。
