摘要
为准确地评估滚动轴承的性能退化状态,提出了一种基于经验模态分解(empirical mode decomposition,EMD)和逻辑回归的评估方法。首先,提取轴承振动信号的本征模函数(intrinsic mode function,IMF)能量作为特征向量;其次,以轴承正常状态数据和失效状态的特征向量建立逻辑回归模型,获取回归参数;最后计算轴承信号全寿命周期的评估指数(confidential value,CV)。评估结果表明,该方法能及时发现早期故障,也能很好地描述轴承性能退化的各个阶段。
A method based on empirical mode decomposition (EMD) and logistic regression is proposed to accurately assess the degenerate state of roling bearing. Firstly, the intrinsic mode function (IMF) energy of a beating's vibration signal was extracted to be eigenvector. According to the eigenvectors under the normal data and failure data of the bearing, a logistic regression model was then established to obtain regression parameters. The state and degree of the bearing degeneration were finally decided based on the calculation of the confidential value(CV) of bearing signal's full life cycle. Evaluation results show that the proposed method is able to detect bearing fault in its early stage and can provide a reasonable interpretation to the evaluated bearing health condition.
出处
《机械设计与研究》
CSCD
北大核心
2016年第5期72-75,79,共5页
Machine Design And Research
基金
国家自然科学基金资助项目(51205130)
江西省科协重点活动项目(赣科协字[2014]88号
关键词
滚动轴承
性能退化
IMF能量
逻辑回归
rolling bearing
performance degeneration
IMF energy
logistic regression