A5085781660- Reliability and Maintenance Optimization
- Machine Fault Diagnosis Techniques
- Fault Detection and Control Systems
- Gear and Bearing Dynamics Analysis
- Risk and Safety Analysis
- Advanced Computational Techniques and Applications
- Quality and Safety in Healthcare
- Evaluation and Optimization Models
- Mechanical Failure Analysis and Simulation
- Embedded Systems and FPGA Design
- Metallurgy and Material Forming
- Software Reliability and Analysis Research
- Advanced Multi-Objective Optimization Algorithms
- Metaheuristic Optimization Algorithms Research
- Statistical Distribution Estimation and Applications
- Engineering Diagnostics and Reliability
- Structural Integrity and Reliability Analysis
- Advanced Decision-Making Techniques
- Fatigue and fracture mechanics
- Probabilistic and Robust Engineering Design
- Tribology and Lubrication Engineering
- Wireless Sensor Networks and IoT
- Life Cycle Costing Analysis
- EFL/ESL Teaching and Learning
- Time Series Analysis and Forecasting
Taiyuan University of Science and Technology
2012-2025
Shihezi University
2024
Wuhan University of Technology
2013-2023
University of California, San Diego
2020-2021
University of Huddersfield
2021
Shandong Normal University
2019
Shanghai Electric (China)
2019
West Virginia University
2018
Northwestern Polytechnical University
2009-2018
Henan Polytechnic University
2013-2014
Abstract Support vector machines (SVMs) have good processing performance for small sample datasets. The giant search space kernel parameters and the tendency of parameter optimization to fall into local optima are two essential factors that affect generalization ability SVM models and, thus, accuracy fault diagnosis results. Propose using fast inter-class distance (FICDF) in feature reduce function then use differential mutation particle swarm (DMPSO) optimize improve classification model....
Abstract Existing multi-objective evolutionary algorithms (MOEAs) have demonstrated excellent efficiency when tackling tasks. However, its use in computationally expensive issues is hindered by the large number of reliable evaluations needed to find Pareto-optimal solutions. This paper employs semi-supervised learning technique model training aid for addressing issues, resulting assisted algorithm (SLTA-MOEA). In SLTA-MOEA, value every objective function determined as a weighted mean values...
Traditional fault diagnosis methods for rolling bearings rely on nemerous labeled samples, which are difficult to obtain in engineering applications. Moreover, when unseen categories appear the test set, these models fail achieve accurate diagnoses, as not represented training data. To address challenges, a zero-shot model is proposed, realizes knowledge transfer from seen by constructing attribute information, thereby reducing dependence samples. First, an method Discrete Label Embedding...
With improvements in industrial automation, the reliability of gearbox, a key transmission device, has become increasingly crucial for stable operation an entire operating system. However, predicting remaining useful life gearbox is challenging because complex working environments and dynamic load changes. Several existing methods assume inaccurate model structure parameter estimation during prediction, owing to limited availability similar fault sample data. In this study, we analyse...
Abstract The use of multi-sensor information fusion techniques is essential for condition monitoring and prediction in large complex systems. In this paper, a new distributed model method proposed to predict the remaining useful life (RUL) nonlinear Wiener process. First, state–space process established, based on monitoring, Kalman filtering algorithm used filter fuse measurement data received from multiple sensors. Next, parameters degradation states are estimated updated online real time...
Low-contrast or uneven illumination in real-world images will cause a loss of details and increase the difficulty pattern recognition. An automatic image perception adaptive correction algorithm, termed as GLAGC, is proposed this paper. Based on Retinex theory, an extracted through discrete wavelet transform. Two features that characterize illuminance are creatively designed. The first feature spatial luminance distribution feature, which applied to gamma local lighting. other global...
For the remaining useful life (RUL) prediction of complex systems, some data in a large amount multisensor monitoring do not effectively characterize degradation while there is redundancy between sensor data, which leads to low accuracy results. Therefore, high-dimensional kernel density estimation (KDE) RUL method was proposed with an adaptive relative window width based on multisource information fusion, contains multiindicator evaluation algorithm entropy weight and maximum...