目的 针对印刷设备由机械型号大、运转速度高所致的轴承振动信号噪声大、故障特征微弱、难以精准诊断等问题,开展基于参数迁移的变工况故障诊断算法研究,以提高复杂工况下轴承故障诊断的准确性和可靠性。方法 构建DSCDSN（Deep separable convolutional dense network）智能故障诊断模型,通过数据预处理模块降噪;引入深度可分离卷积层,重构稠密神经网络卷积层,并确定其结构和超参数。基于微调迁移学习思想,构建Fine-Tuning参数迁移学习模型,利用源域数据集预先训练网络,转移参数,冻结部分结构,微调最后的稠密块卷积层超参数,并进行优化。结果 采用美国凯斯西储大学的轴承数据集和实际工况下印刷机轴承数据集进行实验,验证了Fine-Tuning模型在变工况下识别轴承故障的有效性,具备良好的诊断性能。结论 该算法将DSCDSN与Fine-Tuning模型结合,发挥了迁移学习优势,实现了复杂工况下轴承故障的有效诊断,可为实际生产设备故障诊断提供有效方案。
Abstract
The work aims to conduct research on a variable-condition fault diagnosis algorithm based on parameter transfer to improve the accuracy and reliability of bearing fault diagnosis under complex working conditions, so as to address the challenge that the large mechanical specifications and high operating speeds of printing equipment lead to significant noise in bearing vibration signals, weak fault characteristics, and difficulties in precise diagnosis. A DSCDSN (Deep Separable Convolutional Dense Network) intelligent fault diagnosis model was constructed, and the noise was reduced through a data preprocessing module. A depthwise separable convolution layer was introduced to reconstruct the convolutional layer of the dense neural network, with its structure and hyperparameters determined. Based on the concept of fine-tuning transfer learning, a Fine-Tuning transfer learning model was built. The source-domain dataset was used to pre-train the network, transfer parameters, freeze part of the structure, and optimize by fine-tuning the hyperparameters of the last dense block's convolutional layer. Validation on the Case Western Reserve University dataset showed that the Fine-Tuning model could effectively identify bearing faults under variable conditions, demonstrating good diagnostic performance. In conclusion, this algorithm combines DSCDSN with the Fine-Tuning model, giving full play to the advantages of transfer learning, achieving effective diagnosis of bearing faults under complex conditions, meeting the research objectives, and providing an effective solution for fault diagnosis of actual production equipment.
关键词
变工况 /
轴承故障诊断 /
深度迁移学习 /
参数迁移 /
稠密神经网络
Key words
variable conditions /
bearing fault diagnosis /
deep transfer learning /
parameter migration /
dense neural network
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] SHAO H D, JIANG H K, LIN Y, et al.A Novel Method for Intelligent Fault Diagnosis of Rolling Bearings Using Ensemble Deep Auto-Encoders[J]. Mechanical Systems and Signal Processing, 2018, 102: 278-297.
[2] XIA M, LI T, XU L, et al.Fault Diagnosis for Rotating Machinery Using Multiple Sensors and Convolutional Neural Networks[J]. IEEE/ASME Transactions on Mechatronics, 2018, 23(1): 101-110.
[3] GAO S Z, LI T C, ZHANG Y M.Rolling Bearing Fault Diagnosis of PSO-LSSVM Based on CEEMD Entropy Fusion[J]. Transactions of the Canadian Society for Mechanical Engineering, 2020, 44(3): 405-418.
[4] VERSTRAETE D, FERRADA A, DROGUETT E L, et al.Deep Learning Enabled Fault Diagnosis Using Time-Frequency Image Analysis of Rolling Element Bearings[J]. Shock and Vibration, 2017, 2017(1): 5067651.
[5] 张训杰, 张敏, 李贤均. 基于二维图像和CNN-BiGRU网络的滚动轴承故障模式识别[J]. 振动与冲击, 2021, 40(23): 194-201.
ZHANG X J, ZHANG M, LI X J.Rolling Bearing Fault Mode Recognition Based on 2D Image and CNN-BiGRU[J]. Journal of Vibration and Shock, 2021, 40(23): 194-201.
[6] 张弘斌, 袁奇, 赵柄锡, 等. 采用多通道样本和深度卷积神经网络的轴承故障诊断方法[J]. 西安交通大学学报, 2020, 54(8): 58-66.
ZHANG H B, YUAN Q, ZHAO B X, et al.Bearing Fault Diagnosis with Multi-Channel Sample and Deep Convolutional Neural Network[J]. Journal of Xi'an Jiaotong University, 2020, 54(8): 58-66.
[7] CHEN Z Q, LI C, SANCHEZ R V.Gearbox Fault Identification and Classification with Convolutional Neural Networks[J]. Shock and Vibration, 2015, 2015(1): 390134.
[8] MANJURUL ISLAM M M, KIM J M. Reliable Multiple Combined Fault Diagnosis of Bearings Using Heterogeneous Feature Models and Multiclass Support Vector Machines[J]. Reliability Engineering & System Safety, 2019, 184: 55-66.
[9] 刘永志. 基于深度迁移学习的滚动轴承故障诊断方法研究[D]. 成都: 西南交通大学, 2022: 47-66.
LIU Y Z.Research on Rolling Bearing Fault Diagnosis Method Based on Deep Transfer Learning[D]. Chengdu: Southwest Jiaotong University, 2022: 47-66.
[10] 唐波, 陈慎慎, 郭必奔, 等. 基于特征参数迁移的滚动轴承故障诊断[J]. 计量学报, 2022, 43(3): 386-391.
TANG B, CHEN S S, GUO B B, et al.Fault Diagnosis of Rolling Bearings Based on Migration of Characteristic Parameters[J]. Acta Metrologica Sinica, 2022, 43(3): 386-391.
[11] 贾宁. 基于深度学习和迁移学习的旋转部件跨域故障诊断技术研究[D]. 重庆: 重庆理工大学, 2023: 37-69.
JIA N.Research on Cross-Domain Fault Diagnosis Technology of Rotating Parts Based on Deep Learning and Migration Learning[D]. Chongqing: Chongqing University of Technology, 2023: 37-69.
[12] 邵文琦, 赵梦奇. 基于参数迁移的摇臂传动系统跨工况故障诊断算法[J]. 煤矿机械, 2025, 46(1): 186-189.
SHAO W Q, ZHAO M Q.Cross-Condition Fault Diagnosis Algorithm for Rocker Arm Transmission System Based on Parameter Transfer[J]. Coal Mine Machinery, 2025, 46(1): 186-189.
[13] 李继光, 苏燕辰, 任继炜, 等. 迁移学习在轴承保持架故障诊断中的研究[J]. 机械设计与制造, 2023(11): 17-20.
LI J G, SU Y C, REN J W, et al.Research on Transfer Learning in Fault Diagnosis of Bearing Cage[J]. Machinery Design & Manufacture, 2023(11): 17-20.
[14] TANG G, LUO G G, ZHANG W H, et al.Underdetermined Blind Source Separation with Variational Mode Decomposition for Compound Roller Bearing Fault Signals[J]. Sensors, 2016, 16(6): 897.
[15] GANIN Y, USTINOVA E, AJAKAN H, et al.Domain-Adversarial Training of Neural Networks[J]. Journal of Machine Learning Research, 2016, 17(59): 1-35.
[16] GUO W, TSE P W, DJORDJEVICH A.Faulty Bearing Signal Recovery from Large Noise Using a Hybrid Method Based on Spectral Kurtosis and Ensemble Empirical Mode Decomposition[J]. Measurement, 2012, 45(5): 1308-1322.
[17] ZHAO M H, ZHONG S S, FU X Y, et al.Deep Residual Shrinkage Networks for Fault Diagnosis[J]. IEEE Transactions on Industrial Informatics, 2019, 16(7): 4681-4690.
基金
国家自然科学基金（52205127）; 陕西省教育厅重点科学研究计划（23JY062）
PDF(9553 KB)
渝公网安备 50010702501716号 渝ICP备15012534号-2
