摘要
深度学习作为一种实用的大数据处理工具,在机械智能故障诊断领域也受到广泛关注,许多研究者已经成功地将深度学习模型应用于故障诊断领域。但这些研究往往忽略了两个重要的问题:(1)当原始训练数据集不足时,模型训练过程不理想;(2)网络模型的学习内容不明确。为了克服上述不足,提出一种新的数据增强的堆叠自编码器(DESAE)框架,该框架由数据增强模块和故障分类模块组成。在数据增强模块中,采用SAE生成模拟信号,对不充足的训练数据进行增强。在故障分类模块中,利用增强的数据集训练另一个SAE模型并进行故障样本类型识别。同时,利用轴承数据集验证该方法的有效性。此外,为了更便于理解DESAE工作过程,对其各层学习特性进行可视化分析。
As a practical tool for large data processing,deep learning has been widely concerned in the field of mechanical intelligent fault diagnosis.Many researchers have successfully applied the deep learning model to the field of fault diagnosis.However,these studies always neglect two important points as follows:(1)the model training process will not be ideal when the original training dataset is insufficient;(2)the learning content of the network model is not clear.In order to surmount above deficiencies,a novel framework named data-enhanced stacked autoencoders(DESAE),which consists of a data enhancement module and a fault classification module,is proposed.In the data enhancement module,SAE is adopted to generate simulation signals to strengthen the insufficient training data.In the fault classification module,the enhanced datasets are used to train another SAE model for fault type recognition.Meanwhile,the bearing datasets are employed to validate the proposed method.In addition,the visual analysis of the learning characteristics in each hidden layer is presented to understand the working process of DESAE.
作者
王晓玉
刘桂芳
韩宝坤
王金瑞
石兆婷
WANG Xiaoyu;LIU Guifang;HAN Baokun;WANG Jinrui;SHI Zhaoting(School of Mechanical and Electronic Engineering,Shandong University of Science and Technology Qingdao 266590,Shandong,China)
出处
《噪声与振动控制》
CSCD
北大核心
2021年第2期100-104,110,共6页
Noise and Vibration Control
基金
国家自然科学基金资助项目(52005303)
山东省自然科学基金资助项目(ZR202020QE157)
中国博士后科学基金面上资助项目(2019M662399)。
关键词
智能故障诊断
深度学习
数据增强的堆叠自编码器
仿真信号
intelligent fault diagnosis
deep learning
data-enhanced stacked autoencoder(DESAE)
simulation signals
