- Machine Fault Diagnosis Techniques
- Gear and Bearing Dynamics Analysis
- Manufacturing Process and Optimization
- Time Series Analysis and Forecasting
- Design Education and Practice
- Mechanical Failure Analysis and Simulation
- Occupational Health and Safety Research
- Simulation and Modeling Applications
- Sustainable Supply Chain Management
- Product Development and Customization
- Reliability and Maintenance Optimization
- Advanced Measurement and Metrology Techniques
- Neural Networks and Applications
- Teleoperation and Haptic Systems
- Railway Engineering and Dynamics
- Hand Gesture Recognition Systems
- Fault Detection and Control Systems
- Advanced machining processes and optimization
- Digital Transformation in Industry
- Advanced Manufacturing and Logistics Optimization
- Engineering Diagnostics and Reliability
- Advanced Surface Polishing Techniques
- Industrial Vision Systems and Defect Detection
- Transport and Logistics Innovations
- Technology Assessment and Management
Southwest Jiaotong University
2016-2025
Zhejiang Chinese Medical University
2024
China Institute of Atomic Energy
2023
Science and Technology Department of Sichuan Province
2023
Huazhong University of Science and Technology
2006-2022
Union Hospital
2022
Duke University
2003
The effectiveness of cluster-based distributed sensor networks depends to a large extent on the coverage provided by deployment. We propose virtual force algorithm (VFA) as deployment strategy enhance after an initial random placement sensors. For given number sensors, VFA attempts maximize field coverage. A judicious combination attractive and repulsive forces is used determine motion paths rate movement for randomly-placed Once effective positions are identified, one-time with energy...
Formulae display:?Mathematical formulae have been encoded as MathML and are displayed in this HTML version using MathJax order to improve their display. Uncheck the box turn off. This feature requires Javascript. Click on a formula zoom.
For data-driven remaining useful life (RUL) prediction of rolling bearing, deep learning methods usually do not consider the difference in distribution due to different operating conditions, which adversely affects results. Recently, transfer has been a research hotspot, and it can mitigate above problem effectively. However, for bearings under same working condition, data changes location time failure. To solve these problems, model based on dynamic benchmark(DB) proposed predict RUL this...
With the advent of Industry 4.0, design a smart product to work in factory, or at home city over its lifecycle has consider intelligent interaction with external environment (physical, human and cyberspace environments). Thus, interaction-driven products becomes an important research field, facing huge challenge integrating crossing all life phases. In this paper, high-profile framework is proposed for guiding model-driven through-life design. It three core elements: (1) generic model...
Abstract The success of rotating machines’ data-driven remaining useful life (RUL) prognosis approaches depends heavily on the abundance entire cycle data. However, it is not easy to obtain sufficient run-to-failure data in industrial practice. Data generation technology a promising solution for enriching but fails address intrinsic complexity nonlinear stage degradation and time correlation long-term This research proposes an RUL approach improved by trend feature variational autoencoder....
The intelligent diagnosis of rolling bearing (RB) faults under different working conditions has attracted significant attention. two main limitations existing domain-adaptation-based fault methods for RBs are as follows. One is that the source domain transfer features contain a large amount redundant information interfering with adaptation. other discrepancies in distribution between same class samples lead to low accuracy. Aiming at overcoming these limitations, this study, cross-domain...
In the existing bearing remaining useful life (RUL)-prediction model based on deep learning, advantages and disadvantages of extracted features are evaluated by prediction accuracy; thus, analytical ability is poor. At same time, change working conditions has a great influence accuracy. To overcome these limitations, method RUL feature evaluation transfer learning proposed. The proposed can solve above problems: (1) selection for trend consistency index was designed. (2) this study, domain...
Designing a complex mechatronic product involves multiple design variables, objectives, constraints, and evaluation criteria as well their nonlinearly coupled relationships. The space can be very big consisting of many functional parameters, structural behavioral (or running performances) parameters. Given inexplicit relations among them, how to optimally in an optimization process is challenging research problem. In this paper, we propose systematic method based on reduction surrogate...
Intelligent mechanical fault diagnosis has developed very fast in recent years due to the advancement and application of deep learning technologies. Thus, there are many network models that have been explored classification diagnosis. However, still limitations research on relationship between location, type, severity. In this paper, a novel method for bearing using hierarchical multitask convolution neural networks (HMCNNs) is proposed, taking into account mentioned relationships. The HMCNN...
The transfer learning (TL) method represented by domain adaptation (DA) can effectively improve the prediction accuracy of rolling bearings' remaining useful life (RUL) under different working conditions. However, difference in bearing degradation process same conditions limits reliability and generalization RUL model. Owing to aforementioned problems, this study proposed an predicting for bearings based on conditions' common benchmark. An attention mechanism autoencoder (AMAE) is extract...
The workpiece contour errors from previous process affect current process. In order to improve allowance distribution, this article presents a registration and localization adjustment method with contact inspection. According the analysis of machining deviation, rotational translational matrixes multi-tolerance surfaces are obtained by surface characteristic vectors. (position orientation) can be realized transforming coordinate systems 5-axis machine tool. Through adaptive iterative...
A railway vehicle is a complicated mechatronic system. Its mechanical structure and connections are key influences on its performance. When runs at very high speeds dynamic performance becomes of considerable interest. Therefore, the identification design factors that significantly impact vehicle’s running behavior, such as safety, stability, comfort reliability, step towards optimization. The optimal difficult to perform using existing methods due complexity inherent in rail/track...
Transfer across bearings produces a greater domain shift than transfer working conditions (WCs). Because different may have differences in structural parameters, measurement environments, and WCs, direct significantly lower the diagnostic accuracy of target bearing. A novel fault diagnosis method based on pseudo-label transitive adaptation networks (PLTDANs) is proposed to address this problem. First, empirical selection criteria are offered ensure that appropriate intermediate domains...
A GPGPU-based collision detection algorithm is proposed. Firstly, the information of OBB hierarchy tree and triangles tested objects are mapped into some data textures designed for calculation, such as triangle vertex textures, bounding box size texture, node relationship etc., then these downloaded to GPU complete preparation. Secondly, whole executed on GPU, in which three key contents fulfilled: reading necessary from related correctly by order coordinate method index method, detecting...