Article
Chemistry, Analytical
Jiajun He, Ping Wu, Yizhi Tong, Xujie Zhang, Meizhen Lei, Jinfeng Gao
Summary: This paper proposes a novel bearing fault diagnosis method using an improved multi-scale convolutional neural network (IMSCNN) to extract multi-scale features and mitigate the effect of noise in vibration signals. Experimental results show the superiority of the proposed method compared to other related methods.
Article
Engineering, Electrical & Electronic
Linshan Jia, Tommy W. S. Chow, Yu Wang, Yixuan Yuan
Summary: A novel fault diagnosis framework called MRA-CNN is proposed in this article to learn discriminative multiscale features from vibrational signals and reduce noises. Experimental results show that the proposed method achieves higher accuracy in highly noisy environments compared to state-of-the-art methods.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Engineering, Electrical & Electronic
Junbin Chen, Ruyi Huang, Kun Zhao, Wei Wang, Longcan Liu, Weihua Li
Summary: The MSCNN-FA method proposes a feature alignment module to enhance the shift-invariance of convolutional neural networks for bearing fault diagnosis. It utilizes a multiscale convolution strategy to extract robust features and constructs a classifier with fully connected layers, outperforming other CNN-based methods in terms of diagnosis accuracy and feature robustness in bearing experiments.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Engineering, Electrical & Electronic
Xinglong Pei, Xiaoyang Zheng, Jinliang Wu
Summary: The paper introduces a novel Transformer convolution network (TCN) based on transfer learning, which has achieved highly accurate fault diagnosis. Experimental results demonstrate the robustness and effectiveness of the proposed method.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2021)
Article
Computer Science, Information Systems
Long Qian, Binbin Li, Lijuan Chen
Summary: This paper presents a method for diagnosing motor faults based on the multi-feature fusion of convolutional neural network (CNN), which improves the accuracy and stability of motor fault diagnosis through various processing methods such as multi-scale feature extraction and time series fusion.
Article
Chemistry, Analytical
Xiaorui Shao, Chang-Soo Kim
Summary: This article proposes a domain adaptive and lightweight framework for fault diagnosis based on 1D-CNN, which can extract features with robustness and domain invariance through CORAL processing to minimize domain shifts, effectively improving FD performance.
Article
Computer Science, Artificial Intelligence
Shuyang Luo, Xufeng Huang, Yanzhi Wang, Rongmin Luo, Qi Zhou
Summary: In this paper, an improved stacked autoencoder based on convolutional shortcuts and domain fusion strategy is proposed for fault diagnosis of rolling bearing. The feasibility of the proposed method is validated on two publicly available bearing datasets and a custom-built experiment device, and the results show its superior performance in different working conditions, cross-domain, and limited labeled data situations.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Physics, Multidisciplinary
Ming Xu, Jinfeng Gao, Zhong Zhang, Heshan Wang
Summary: Deep learning bearing-fault diagnosis has shown strong vitality in recent years. A novel intelligent bearing-fault diagnosis method based on signal-to-RGB image mapping (STRIM) and multichannel multiscale CNN (MCMS-CNN) is proposed to overcome the limitations of traditional fault-diagnosis methods and existing intelligent bearing-fault diagnosis approaches. The proposed method achieves feature fusion by converting signals from multiple sensors into RGB images and effectively extracts RGB image features using MCMS-CNN. Experimental results demonstrate the superior performance of the proposed method in fault diagnosis compared to other approaches.
Article
Engineering, Electrical & Electronic
Jianbin Xiong, Minghui Liu, Chunlin Li, Jian Cen, Qinghua Zhang, Qiongqing Liu
Summary: Under nonlinear and nonstationary dynamic conditions, traditional multidimensional dimensionless indicators (MDIs) fail to effectively diagnose faults in petrochemical units. To address this, this article proposes a new dimensionless indicator named complementary ensemble multidimensionless indicators (CEMDIs), by combining complementary ensemble empirical mode decomposition (CEEMD) and MDI. The CEMDI processed data is then converted into Gramian angular fields (GAFs) using the sequential mapping method, and convolutional neural networks (CNNs) are used to identify different fault types in sparse data. The proposed method is validated using three datasets, showcasing its effectiveness and superiority over traditional methods.
IEEE SENSORS JOURNAL
(2023)
Article
Engineering, Mechanical
Defeng Lv, Huawei Wang, Changchang Che
Summary: The study proposes a fault diagnosis method for rolling bearings based on MCNN and decision fusion, which can extract deep features of vibration signals and achieve robust fault diagnosis results. The model can accurately diagnose faults in long time series of vibration signals with noise.
INDUSTRIAL LUBRICATION AND TRIBOLOGY
(2021)
Article
Engineering, Electrical & Electronic
Shaopu Yang, Zhaoyang Cui, Xiaohui Gu
Summary: This article introduces a novel unsupervised domain adaptation network called BDTN for data-driven bearing fault diagnosis. BDTN maps data with distinct marginal and conditional distributions onto the same feature subspace and considers the relative importance of marginal distribution adaptation and conditional distribution adaptation.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Electrical & Electronic
Peizhe Yin, Jie Nie, Xinyue Liang, Shusong Yu, Chenglong Wang, Weizhi Nie, Xiangqian Ding
Summary: Fault diagnosis for rolling bearings is a significant engineering problem, and engineers use vibration signals to analyze the features and detect damage. Deep-learning technology has gained research interest for fault diagnosis, but most existing methods have limitations in mining the relationship between signals.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Multidisciplinary
Changchang Che, Huawei Wang, Xiaomei Ni, Ruiguan Lin
Summary: This method extracts time-domain features of vibration signals from rolling bearings and uses convolutional neural networks and deep belief networks to process grayscale images and time series samples respectively. By combining multiple deep learning models, comprehensive fault prediction results are obtained, demonstrating higher fault diagnosis accuracy compared to individual deep learning models and traditional methods.
Article
Chemistry, Multidisciplinary
Cheng-Jian Lin, Chun-Hui Lin, Frank Lin
Summary: A vector-based convolutional fuzzy neural network (vector-CFNN) was developed in this study to diagnose faults from signals. The fusion layer played a crucial role in combining characteristics and improving the model's performance. Experimental results showed that vector-CFNN outperformed other neural networks in terms of accuracy and parameter efficiency, making it feasible for online spindle vibration monitoring.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Multidisciplinary
Tian Han, Longwen Zhang, Zhongjun Yin, Andy C. C. Tan
Summary: This paper combines CNN and SVM for bearing fault diagnosis, improving the model's generalization ability and accuracy. Experimental results show the system has advantages of less time-consuming, high accuracy, and strong generalization ability.
Article
Chemistry, Analytical
Duy Tang Hoang, Xuan Toa Tran, Mien Van, Hee Jun Kang
Summary: This paper introduces a novel method for bearing fault diagnosis by fusing information from multiple sensor systems using a convolutional neural network. Experimental results show that the proposed method outperforms other techniques in feature extraction and fusion processes.
Article
Chemistry, Analytical
Anh Tuan Vo, Thanh Nguyen Truong, Hee-Jun Kang
Summary: This article presents an advanced prescribed performance-tracking control system for uncertain robotic manipulators, with finite-time convergence stability. The proposed controller exhibits improved properties, including estimated convergence speed and predefined upper and lower limits for maximum overshoot during transient responses. Furthermore, it can provide a smooth control torque without losing its robustness, and the control errors inevitably converge to zero within a finite time.
Article
Chemistry, Multidisciplinary
Phu-Nguyen Le, Hee-Jun Kang
Summary: The study introduces a manipulator calibration algorithm aimed at reducing positional errors of an industrial robotic manipulator by selecting optimal measurement poses using a genetic algorithm. It utilizes conventional kinematic calibration and a radial basis function neural network to compensate for compliance errors, showing effectiveness through experimental calibration and validation processes.
APPLIED SCIENCES-BASEL
(2022)
Article
Automation & Control Systems
Thanh Nguyen Truong, Anh Tuan Vo, Hee-Jun Kang
Summary: A neural network-based non-singular fast terminal sliding mode control method is proposed for path tracking control of uncertain nonlinear systems, achieving faster convergence time, higher tracking accuracy, and less chattering with strong control performance for the entire closed-loop control system.
INTERNATIONAL JOURNAL OF CONTROL AUTOMATION AND SYSTEMS
(2022)
Article
Engineering, Mechanical
Quang Dan Le, Hee-Jun Kang
Summary: In this paper, an active fault-tolerant control for robot manipulators is proposed, which combines a novel finite-time synchronous fast terminal sliding mode control and extended state observer. The proposed controller can suppress the effects of faults and guarantee the acceptable performance of robot manipulators when faults occur. By integrating the observer compensation, sliding mode, and synchronization technique, the proposed controller can effectively deal with uncertainties and disturbances in normal operation mode and reduce the effects of faults, especially loss-effective faults.
Article
Chemistry, Analytical
Anh Tuan Vo, Thanh Nguyen Truong, Hee-Jun Kang
Summary: This paper investigates the problem of an APPTMC for manipulators and proposes several improved methods, including modified PPF, modified NISMS, fixed-time USOSMO, and ASTwCL. An observer-based control solution is synthesized from these methods to achieve PCP in the sense of finite-time Lyapunov stability, significantly improving precision, robustness, and chattering reduction.
Article
Engineering, Mechanical
Van-Cuong Nguyen, Xuan-Toa Tran, Hee-Jun Kang
Summary: This paper proposes a novel fault-tolerant control tactic for robot manipulator systems using only position measurements. The tactic combines a nonsingular fast terminal sliding mode control (NFTSMC) and a novel high-speed third-order sliding mode observer (TOSMO). The proposed strategy provides excellent features, such as fast convergence time, high tracking precision, chattering phenomenon reduction, robustness against unknown input, and elimination of velocity requirement. The proposed observer improves the convergence speed of the estimated signals and increases the system's dynamic response.
Article
Chemistry, Analytical
Thanh Nguyen Truong, Anh Tuan Vo, Hee-Jun Kang
Summary: This research proposes an improved control approach based on real-time Prescribed Performance Control (PPC) to address the dynamic uncertainty and exterior perturbations in magnetic levitation systems. By introducing a modified function and an improved sliding mode observer, the controlled errors quickly converge to the equilibrium point within a prescribed performance range. The combination of the proposed control methods ensures stable positioning of the controlled ball and achieves impressive performance in terms of tracking accuracy, fast convergence, stabilization, and chattering reduction.
Article
Mathematics
Anh Tuan Vo, Thanh Nguyen Truong, Hee-Jun Kang
Summary: This paper proposes a fixed-time neural network-based prescribed performance control method (FNN-PPCM) for robot manipulators. It suggests alternative approaches to address the limitations of a fixed-time sliding mode controller (SMC) designed with its strengths and weaknesses in mind. The method integrates a radial basis function neural network (RBFNN) for precise uncertainty estimation, unconstrained dynamics, and a fixed-time convergence sliding surface based on transformed errors, achieving fixed-time prescribed performance while effectively addressing chattering and requiring only a partial dynamics model of the robot.
Article
Engineering, Electrical & Electronic
Anh Tuan Vo, Thanh Nguyen Truong, Quang Dan Le, Hee-Jun Kang
Summary: This work proposes a hybrid trajectory tracking control algorithm (HTCA) for robot manipulators with uncertain dynamics and external disturbances. The control algorithm includes a uniform second-order sliding mode disturbance observer (USOSMDO) for active disturbance rejection and a fixed-time singularity-free terminal sliding surface (FxSTSS) for fixed-time convergence of the tracking control error. The HTCA is formed based on the FxSTSS and the fixed-time power rate reaching law (FxPRRL) using information from the USOSMDO. Numerical simulations show the effectiveness and advantages of the proposed HTCA applied to a FARA robot.
Review
Computer Science, Artificial Intelligence
Thanh Nguyen Truong, Anh Tuan Vo, Hee-Jun Kang
Summary: This paper investigates the application of sliding mode control based on neural networks in robot manipulators. Firstly, the advantages, disadvantages, and applications of sliding mode control and its variants are assessed. Secondly, recent advancements in control systems and the use of neural networks as an alternative approach are introduced. Finally, the advantages and limitations of these combined approaches are evaluated based on previous studies and future development directions.
Proceedings Paper
Computer Science, Artificial Intelligence
Thanh Nguyen Truong, Anh Tuan Vo, Hee-Jun Kang, Tien Dung Le
Summary: In this paper, an observer-based fixed-time sliding mode controller is proposed for a class of second-order nonlinear systems with matched uncertainties and disturbances. By utilizing a designed fixed-time disturbance observer and a fixed-time sliding mode method, the controller effectively reduces chattering, improves tracking performance, and ensures a guaranteed closed-loop convergence time.
INTELLIGENT COMPUTING METHODOLOGIES, PT III
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Anh Tuan Vo, Thanh Nguyen Truong, Hee-Jun Kang, Tien Dung Le
Summary: Scientists have always been interested in robot manipulators due to their important role in practical applications. In this paper, an advanced terminal sliding mode control method is developed to address position tracking problems of robot manipulators, resulting in significant performance improvements.
INTELLIGENT COMPUTING METHODOLOGIES, PT III
(2022)
Article
Computer Science, Information Systems
A. N. H. T. U. A. N. VO, T. H. A. N. H. N. G. U. Y. E. N. TRUONG, HEE-JUN KANG
Summary: This paper investigates complex modified function projective lag synchronization with fixed-time stability guarantees for hyperchaotic systems. The proposed method ensures the synchronization and desired performance of complex hyperchaotic systems within a bounded time, without requiring the initial conditions from the systems. The fixed-time stability guarantees are achieved using the Lyapunov principle, and the settling time is calculated by solving differential equations.
Article
Engineering, Electrical & Electronic
Van-Cuong Nguyen, Duy-Tang Hoang, Xuan-Toa Tran, Mien Van, Hee-Jun Kang
Summary: The article introduces a novel method for bearing fault diagnosis using DNN, which improves feature extraction with multiple-domain image representation data and a multi-branch structure DNN, leading to better fault diagnosis performance.