基于能量熵和CL-LSTM的故障诊断模型  

Fault Diagnosis Model Based on Energy Entropy and CL-LSTM

作　　者：侯鑫烨 董增寿[1] 刘鑫[1] 段敏霞 HOU Xinye;DONG Zengshou;LIU Xin;DUAN Minxia(School of Electronic Information Engineering,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024,China)

机构地区：[1]太原科技大学电子信息工程学院,山西太原030024

出　　处：《机床与液压》2021年第16期180-184,203,共6页Machine Tool & Hydraulics

基　　金：国家留学基金资助项目;山西省留学归国人员择优资助项目(201802);山西省重点研发计划项目(201903D321012);山西省研究生教育创新项目(2019SY487)。

摘　　要：针对长短时记忆网络(Long Short Term Memory,LSTM)处理大数据集时运行时间长、存在维数灾难的问题,提出基于能量熵和CL-LSTM(Long Short Term Memory Network with Center Loss)的智能故障诊断模型。利用自适应白噪声的完整集合经验模态分解对原始信号进行分解;结合相关系数筛选IMF分量并计算其能量熵作为新样本输入到LSTM中,增强了样本间的差异性,减小了数据维度。将中心损失引入Softmax损失中,使类内距离更小,进一步提高分类精度。利用西储大学轴承数据集进行实验,验证了所提方法在识别滚动轴承故障状态时准确率高、稳定性好。The LSTM(long short-term memory)has the problems of long running time and dimension disaster in processing large data sets.In order to solve the problems,an intelligent fault diagnosis model based on energy entropy and CL-LSTM(long short term memory network with center loss)was proposed.The original signal was decomposed into several components with complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)adaptively;IMF components were selected through correlation coefficients,and its energy entropy was calculated as a new input into the LSTM,which enhanced the sample diversity and reduced the data dimension.Center loss was introduced into the Softmax loss to make the intra-class distance smaller and further improve the classification accuracy.The bearing data set of Western Reserve University was used for experiment.It is verified that the proposed method has high accuracy and good stability in the identification of rolling bearing fault states.

关 键 词：能量熵 滚动轴承 故障诊断 损失函数 长短时记忆网络 

分 类 号：TH133.33[机械工程—机械制造及自动化]