摘要
针对滚动轴承信号的非线性、非平稳特性,致使轴承状态难识别的问题,提出分层极限学习机(HELM)故障诊断模型。首先采用集合经验模式分解(EEMD)将轴承信号分解为若干个本征模式分量(IMFs),并提取其能量熵值构建特征向量;其次利用自动编码器(AE)对极限学习机的隐含层进行分层,且使隐含层节点的输入权值和阈值满足正交条件;最后将构建的特征向量作为H-ELM算法的输入,通过训练建立H-ELM滚动轴承故障分类模型。实验结果表明:H-ELM滚动轴承故障分类模型比SVM、ELM故障分类模型具有更高的精度、更强的稳定性。
According to the nonlinear and non - stationary characteristics of the rolling bearing signals, cause the bearing condition difficult to identify, and the hierarchical extreme learning machine fault diagnosis model is proposed. Firstly the ensemble empirical mode decomposition （EEMD） bearing signal decomposition for multiple intrinsic mode of IMFS and extract the energy entropy composition feature vector; secondly, the hidden layer of the extreme learning machine is layered by using automatic encoder, and the input value and threshold value of the hidden layer nodes are satisfied ; finally, the combined feature vector is used as the input of the algorithm, and the fault classification model of the rolling bearing of the H - ELM is established. The experimental results show that the H - ELM roiling bearing fault classification model has higher accuracy and stronger stability than ELM and SVM fault classification model.
出处
《制造技术与机床》
北大核心
2017年第4期73-76,81,共5页
Manufacturing Technology & Machine Tool
基金
国家自然科学基金(51565046)
内蒙古自然科学基金(2015MS0512)
包头市科技计划发展项目(2015X2011)
关键词
滚动轴承
故障诊断
自动编码器
极限学习机
rolling bearing
fault diagnosis
automatic encoder
extreme learning machine
