摘要
传统智能故障诊断算法需要依赖人工特征提取和专家知识,而旋转机械设备复杂的工作环境和工况使得传统算法在实际应用中缺乏良好的自适应性和泛化性.针对以上问题,提出基于卷积神经网络(Convolutional neural network, CNN)的层级化故障诊断算法(CNN based hierarchical fault diagnosis, CNN-HFD).首先,将原始振动信号进行分段预处理,以实现数据扩容;然后,分别根据故障类型和故障程度设计多个卷积神经网络,并将原始振动数据以某一时间步进行分割,作为卷积神经网络的输入进行训练;最后,将待识别信号送入CNN-HFD模型,经过分层故障诊断,在...
学科与关键词
关键词
卷积神经网络深度学习层级化故障诊断滚动轴承振动信号 convolutional neural networkdeep learninghierarchical fault diagnosisrolling bearingvibration signal
- 基于卷积神经网络的滚动轴承故障诊断研究高统林; 朱坚民; 黄之文农业装备与车辆工程, 2019, (12): 1-9.
- Intelligent Fault Diagnosis for Rotary Machinery Using Transferable Convolutional Neural NetworkChen, Zhuyun; Gryllias, Konstantinos; Li, Weihua*IEEE Transactions on Industrial Informatics, 2020, 16(1): 339-349.被引 185
- Parallel multi-fusion convolutional neural networks based fault diagnosis of rotating machinery under noisy environmentsLi, Guoqiang; Wu, Jun*; Deng, Chao*; Chen, ZuoyiISA TRANSACTIONS, 2022, 128: 545-555.被引 33
