Automatic Model Creation Method of Hierarchical Parallel Network Model for Transfer Diagnosis of Rotating Machinery

doi:10.3901/JME.2022.22.115

Abstract

Abstract: In view of the current deep learning-based fault diagnosis methods of rotating machinery relying on manual modeling experience, requiring manual parameter adjustment and continuous trial-and-error iteration, re-creation of diagnosis models for different diagnostic tasks, an automatic creation method of hierarchical parallel network model for transfer diagnosis of rotating machinery is proposed, which can quickly and automatically search high precision diagnosis models with heterogeneous transfer performance according to different diagnosis tasks. Based on neural architecture search(NAS) and modular design ideas, two types of foundation blocks of parallel structure containing multiple layers are designed, which is different from the traditional NAS method to search layer by layer, but searches based on the foundation blocks. The controller outputs decision sequence to determine the foundation blocks' structure and stacks them to form a hierarchical parallel candidate model. Then according to the verification results of the candidate model on the diagnosis task, the controller is optimized using the strategy gradient algorithm, and the diagnosis accuracy of the candidate model is continuously improved by iterating the above process. The hierarchical parallel structure of the candidate model supports its good heterogeneous transfer performance. In addition, in order to solve the time-consuming bottleneck problem of NAS method, the weight sharing mechanism is set in the candidate model training process to improve the efficiency of automatic modeling. The proposed method is used to conduct automatic modeling and heterogeneous transfer diagnosis experiments for four different rotating machinery fault datasets, and the results show that the proposed method can efficiently create 100% accuracy diagnosis models for four different diagnosis tasks, consuming 313s to 1 601 s, and the candidate model can achieve a transfer diagnosis accuracy of more than 95% for the target diagnosis task with only 10% of the target domain data and 100s of fine-tuning.

Key words: rotating machinery, hierarchical parallel network, automatic model creation, weight sharing, transfer diagnosis

CLC Number:

TH165
TG506

ZHOU Jian, ZHENG Lian-yu, WANG Yi-wei, WANG Yi-chuan. Automatic Model Creation Method of Hierarchical Parallel Network Model for Transfer Diagnosis of Rotating Machinery[J]. Journal of Mechanical Engineering, 2022, 58(22): 115-128.

References

[1] 刘东东,程卫东,万广通.基于故障特征趋势线模板的滚动轴承故障诊断[J].机械工程学报,2017,53(9):83-91.LIU Dongdong,CHENG Weidong,WAN Guangtong.Bearing fault diagnosis based on fault characteristic trend template[J]. Journal of Mechanical Engineering,2017,53(9):83-91.
[2] DHAMANDE L S, CHAUDHARI M B. Compound gear-bearing fault feature extraction using statistical features based on time-frequency method[J].Measurement,2018,125:63-77.
[3] YAN Ruqiang,GAO R X. Energy-based feature extraction for defect diagnosis in rotary machines[J]. IEEE Trans.Instrum. Meas.,2019,58(9):3130-3139.
[4] 黄包裕,张永祥,赵磊.基于布谷鸟搜索算法和最大二阶循环平稳盲解卷积的滚动轴承故障诊断方法[J].机械工程学报,2021,57(9):99-107.HUANG Baoyu, ZAHNG Yongxiang, ZHAO Lei.Research on fault diagnosis method of rolling bearings based on cuckoo search algorithm and maximum second order cyclostationary blind deconvolution[J]. Journal of Mechanical Engineering,2021,57(9):99-107.
[5] 姚德臣,杨建伟,程晓卿,等.基于多尺度本征模态排列熵和SA-SVM的轴承故障诊断研究[J].机械工程学报,2018,54(9):168-176.YAO Dechen,YANG Jianwei,CHENG Xiaoqing,et al.Railway rolling bearing fault diagnosis based on multi-scale IMF permutation entropy and SA-SVM classifier[J]. Journal of Mechanical Engineering,2018,54(9):168-176.
[6] 胡茑庆,陈微鹏,程哲,等.基于经验模态分解和深度卷积神经网络的行星齿轮箱故障诊断方法[J].机械工程学报,2019,55(7):9-18.HU Niaoqing,CHEN Weipeng,CHENG Zhen,et al. Fault diagnosis for planetary gearbox based on EMD and deep convolutional neural networks[J]. Journal of Mechanical Engineering,2019,55(7):9-18.
[7] 周兴康,余建波.基于深度一维残差卷积自编码网络的齿轮箱故障诊断[J].机械工程学报,2020,56(7):96-108.ZHOU Xingkang,YU Jianbo. Gearbox fault diagnosis based on one-dimension residual convolutional auto-encoder[J]. Journal of Mechanical Engineering,2020,56(7):96-108.
[8] 董绍江,裴雪武,吴文亮,等.基于多层降噪技术及改进卷积神经网络的滚动轴承故障诊断方法[J].机械工程学报,2021,57(1):148-156.DONG Shaojiang,PEI Xuewu,WU Wenliang,et al.Rolling bearing fault diagnosis method based on multilayer noise reduction technology and improved convolutional neural network[J]. Journal of Mechanical Engineering,2021,57(1):148-156.
[9] WANG Yiwei, ZHOU Jian, ZHENG Lianyu. An end-to-end fault diagnostics method based on convolutional neural network for rotating machinery with multiple case studies[J]. Journal of Intelligent Manufacturing,2020,33:809-830.
[10] 沈长青,汤盛浩,江星星,等.独立自适应学习率优化深度信念网络在轴承故障诊断中的应用研究[J].机械工程学报,2019,55(7):81-88.SHEN Changqing, TANG Shenghao, JIANG Xingxing,et al. Bearings fault diagnosis based on improved deep belief network by self-individual adaptive learning rate[J]. Journal of Mechanical Engineering,2019,55(7):81-88.
[11] LI Xiang,ZHANG Wei,DING Qian. A robust intelligent fault diagnosis method for rolling element bearings based on deep distance metric learning[J]. Neurocomputing,2018,310:77-95.
[12] WEN Long,GAO Liang,LI Xinyu. A new deep transfer learning based on sparse auto-encoder for fault diagnosis[J]. IEEE Transactions on Systems,Man,and Cybernetics:Systems,2019,49(1):136-144.
[13] ZHOU Tianyi,PAN Jialin,TSANG W I,et al. A deep learning framework for hybrid heterogeneous transfer learning[J]. Artificial Intelligence,2019,275:310-328.
[14] GUO Liang,LEI Yaguo,XING Saibo,et al. Deep convolutional transfer learning network:a new method for intelligent fault diagnosis of machines with unlabeled data[J]. IEEE Transactions on Industrial Electronics,2019,66(9):7316-7325.
[15] ZHOU Jian,ZHENG Lianyu,WANG Yiwei. A multistage deep transfer learning method for machinery fault diagnostics across diverse working conditions and devices[J]. IEEE Access,2020,8:80879-80898.
[16] BAKER B,GUPTA O,NAIK N,et al. Designing neural network architecture using reinforcement learning[C/CD]//Proc. Int. Conf. Learn. Representations,2017.
[17] REAL E,MOORE S,SELLE A,et al. Large-scale evolution of image classifiers[C]//Proceedings of the 34th International Conference on Machine Learning,PMLR,2017,70:2902-2911.
[18] WANG Ruixin,JIANG Hongkai,LI Xingqiu,et al. A reinforcement neural architecture search method for rolling bearing fault diagnosis[J]. Measurement,2020,154:107417.
[19] JIE Cao,MA Jialin,HUANG Dailin,et. al. Finding the optimal multilayer network structure through reinforcement learning in fault diagnosis[J].Measurement,2022,188:110377.
[20] ZHOU Zheng,LI Tianfu,ZHANG Zilong,et al. Bayesian differentiable architecture search for efficient domain matching fault diagnosis[J]. IEEE Transactions on Instrumentation and Measurement,2021,70:1-11.
[21] ZHANG Kaiyu,CHEN Jinglong,HE Shuilong,et. al.Differentiable neural architecture search augmented with pruning and multi-objective optimization for time-efficient intelligent fault diagnosis of machinery[J].Mechanical Systems and Signal Processing,2021,158:107773.
[22] ZHONG Zhao,YAN Junjie,WU Wei,et al. Practical block-wise neural network architecture generation[C/CD]//Proc. IEEE Conf. Comput. Vis. Pattern Recognit.,2018:2423-2432.
[23] LIU Hanxiao,SIMONYAN K,VINYALS O,et al.Hierarchical representations for efficient architecture search[C/OL]//ICLR 2018,arXiv:1711.00436.
[24] SZEGEDY C,LIU Wei,JIA Yangqing,et al. Going deeper with convolutions[C/CD]//2015 IEEE Conference on Computer Vision and Pattern Recognition,7-12,June,2015,Boston,MA.
[25] SZEGEDY C, LOFFE S, VANHOUCKE V, et al.Inception-v4,Inception-ResNet and the impact of residual connections on learning[C/CD]//2016 IEEE Conference on Computer Vision and Pattern Recognition,27-30 June,2016,Las Vegas,NV,USA.
[26] SZEGEDY C, VANHOUCKE V, LOFFE S, et al.Rethinking the inception architecture for computer vision[C/CD]//2016 IEEE Conference on Computer Vision and Pattern Recognition,27-30 June,2016,Las Vegas,NV,USA.