论文题目 |
基于多变换域特征提取和机器学习的滚动轴承故障诊断方法 |
答辩人 |
NGUYEN VIET HUNG |
指导教师 |
程军圣 |
答辩委员会 主席 |
于德介 |
学科专业 |
机械工程 |
学院 |
威廉希尔 |
答辩地点 |
威廉希尔516 |
答辩时间 |
2018年5月27日 下午5：00 |
学位论文简介
The proposed system would give practically variable solution. We intend to introduce following innovations in our solution set:
Vibration feature extraction based on IMFSC-SVD is proposed by combination EMD with SVD methods which is to extract and select the features. These features are fed into the optimal ACROSVKF classifier model to identifying the faults. The experimental results shown that IMFSC-SVD can represent the fault features of roller bearing in the low dimension, and the roller bearing conditions is identified by the ACROSVKF model.
By using the multi-aspect feature method is to express the composite features of roller bearing and use the GDA method for reducing the feature dimension, and finally input to CRSVM model. The expert MAF-GDA-CRSVM fault diagnosis technique is forward proposed for bearing fault diagnosis under different working conditions. The experimental results show that the method can effectively identify the fault of rolling bearing under different working conditions.
Aiming at the problem of fault diagnosis of roller bearing with multi-level fault. An EEMD-DLN diagnosis technique is proposed. The high-dimensionality feature set is extracted by EEMD. Through the autoencoder of the DLN the low dimensionality feature set is generated meaningful. The experimental results verify the effectiveness of the proposed technique.
Based on deep learning stacked autoencoder network and optimal LSSVM-PSO classifier, a diagnosis technique is proposed which is used to identify the multi-level fault of roller bearing. The experimental results verify the effectiveness of vibration features extracted by deep learning network in the proposed technique.
主要学术成果
Nguyen V H, Cheng J S, Thai V T. “An integrated generalized discriminant analysis method and chemical reaction support vector machine model (GDA-CRSVM) for bearing fault diagnosis”. Advances in Production Engineering & Management, 2017,12(4): 321-336. (SCI Q3).
VietHung Nguyen, JunSheng Cheng, VanTrong Thai and XuanChung Nguyen. “Identification of Bearing Fault Signal based on Adaptive Feature Extraction and Optimal ACROSVKF Classification Model”. International Journal of Engineering Science and Computing, 2017,7(12): 11. (Ei compendex).
V.Hung Nguyen, J.Sheng Cheng, Yang Yu and V.Trong Thai. “An Architecture of deep learning network based on ensemble empirical mode decomposition in precise identification of bearing vibration signal”. Journal of Mechanical Science and Technology (Under review).
VietHung Nguyen, JunSheng Cheng, Yang Yu and TienDung Hoang. “Deep learning stacked autoencoder network-based feature representation and optimal LSSVM-PSO classifier model in bearing fault diagnosis”.
VanTrong Thai, JunSheng Cheng,VietHung Nguyen, DucHieu Le. “Application EEMD energy entropy technique and BSA-SVM method for gear fault diagnosis”. Journal of Vibroengineering (Revised).
VanTrong Thai, JunSheng Cheng,VietHung Nguyen, PhuongAnh DaoThi. “Optimizing SVM’s parameters based on Backtracking search optimization algorithm for gear fault diagnosis”.International Journal of Computer Integrated Manufacturing (Under review)
