摘要
针对从滚动轴承振动信号中所提取的故障信息精度低的问题,提出一种基于鲸鱼优化算法(WOA)-变分模态分解(VMD)能量熵的特征提取方法,并采用改进鲸鱼优化算法(WOA)-支持向量机(SVM)进行故障诊断。首先,利用鲸鱼优化算法对变分模态分解模态个数K和惩罚参数α寻优,然后根据VMD处理信号得到若干模态分量,筛选后进一步提取能量熵作为特征向量。最后,针对WOA种群迭代机制易陷入局部极值等缺点,引入随机变异策略进行改进,根据改进WOA-SVM对轴承信号进行故障诊断。实验表明,该方法能够准确提取故障信息,提高轴承数据故障识别率,准确率高达99.2%。
Aiming at the problem of low precision in extracting fault information from vibration signals of rolling bearings,a feature extraction method based on whale optimization algorithm and variational modal decomposition(WOAVMD)energy entropy is proposed with improved WOA-SVM used for fault diagnosis.First of all,the WOA is used to optimize the number of modal decomposition modes K and the penalty parameterα.Then the signal is processed by VMD to obtain several modal components,and the energy entropy is further extracted as a feature vector after the screening.Finally,in view of the shortcoming that the WOA population iteration mechanism may easily fall into local extremes,a random mutation strategy is introduced,and the fault diagnosis of bearing signals is realized by the improved WOA-SVM.Experiments show that this method can accurately extract fault information,improve the fault recognition rate of bearing data.The accuracy rate is up to 99.2%.
作者
张萍
张文海
赵新贺
吴显腾
刘宁
ZHANG Ping;ZHANG Wenhai;ZHAO Xinhe;WU Xianteng;LIU Ning(School of Artificial Intelligence and Data Science,Hebei University of Technology,Tianjin 300130,China)
出处
《噪声与振动控制》
CSCD
北大核心
2021年第4期86-93,275,共9页
Noise and Vibration Control
基金
国家自然科学基金资助项目(51207042)
河北省教育厅青年基金资助项目(Q2012103)。
关键词
故障诊断
变分模态分解
能量熵
鲸鱼优化算法
支持向量机
fault diagnosis
variational mode decomposition(VMD)
energy entropy
whale optimization algorithm(WOA)
support vector machine(SVM)