A5101208893- Power Systems and Renewable Energy
- Industrial Technology and Control Systems
- Machine Fault Diagnosis Techniques
- Smart Grid and Power Systems
- Indoor and Outdoor Localization Technologies
- Multimodal Machine Learning Applications
- Digital Transformation in Industry
- Nuclear Engineering Thermal-Hydraulics
- Natural Language Processing Techniques
- Advanced Manufacturing and Logistics Optimization
- Advanced Adaptive Filtering Techniques
- Thermal Analysis in Power Transmission
- Technology and Security Systems
- Power Systems and Technologies
- Advanced Neural Network Applications
- Fault Detection and Control Systems
- Robot Manipulation and Learning
- Advanced Decision-Making Techniques
- Blockchain Technology Applications and Security
- Evaluation Methods in Various Fields
- Advanced Image and Video Retrieval Techniques
- Speech and Audio Processing
- Magneto-Optical Properties and Applications
- Assembly Line Balancing Optimization
- Gear and Bearing Dynamics Analysis
University of Helsinki
2024-2025
Xijing University
2023-2024
Xinjiang University
2024
Kunming University of Science and Technology
2024
Henan University of Science and Technology
2023
China Academy of Information and Communications Technology
2023
Inner Mongolia Electric Power (China)
2022
Guizhou Electric Power Design and Research Institute
2014
Human–object interaction (HOI) detection aims to localize and recognize the relationship between humans objects, which helps computers understand high-level semantics. In HOI detection, two-stage one-stage methods have distinct advantages disadvantages. The can obtain high-quality human–object pair features based on object but lack contextual information. transformer-based model good global cannot benefit from detection. ideal should of both methods. Therefore, we propose Pairwise...
Abstract To address the challenges of accuracy and interpretability in mechanical fault detection models, this study proposes a shaft orbit generative adversarial network (SOGAN) applies it to detect blade fouling faults. Variational autoencoder (VAE) is used as foundational architecture for extracting high-dimensional latent features from images. Concurrently, invariant moments images are extracted embedded bypass within generator, thereby enhancing detection. Two sets real-world data...
Abstract Bearing fault diagnosis is of great significance to the normal operation machinery, and its performance life span directly affect operational efficiency safety whole equipment. For existing image coding methods which detecting bearing with a large number training samples complex neural networks achieve desired detection performance, Multiscale Permutation Entropy Gray Image Coding (MPEGIC) method proposed. In order fully extract feature information time series signal, this paper...
Generative retrieval generates identifiers of relevant documents in an end-to-end manner using a sequence-to-sequence architecture for given query. The relation between generative and other methods, especially those based on matching within dense models, is not yet fully comprehended. Prior work has demonstrated that with atomic equivalent to single-vector retrieval. Accordingly, exhibits behavior analogous hierarchical search tree index when semantic identifiers. However, prior focuses...
This paper analyzes the influence of high permeability new energy on frequency characteristics system, and proposes a multi-objective optimization model for under load shedding (UFLS), which takes proportions each round as variables. A is established, in mean deviation amount are considered objectives. And system dynamic set constraints. An example 10 machines 39-nodes shows that proposed scheme can make have better stability reduce shedding.
Abstract A new method to measure the voltage in a grid has been proposed based on studies of novel electronic transducer for high-voltage power transmission equipment using charge induction principle and electromagnetic computing method. The measuring improved, real-time electric field position orientation near lines measured calculated by inverse problem. According simulation with test data, data compensation calibration calculation were carried out. With rapid engineering voltage,...
In recent years, due to the rapid development of pet industry, many people began raise cats and dogs other pets. However, accelerated pace life young abundance travel or spare time activities elderly, there are more "empty nest pets". Especially during new coronavirus epidemic, a large number pets facing crisis phenomenon water food shortage. How take care their remotely, conveniently with high quality is immediate need majority owners, so feeding problem period solitude has become major...
To address the trust problem between transaction subjects in current cloud manufacturing system, this paper combines blockchain technology with proposes a system architecture incorporating technology, designs resource matching algorithm within smart contract, constructs an optimal scheduling model minimum time, cost, highest qualification rate and trust, uses NSGA-II to solving. Finally, efficiency of under environment based on while maintaining absolute is demonstrated experimentally.
In order to overcome the subjective limitation in existing evaluation methods, based on determination of indicator weight by using rough set theory, a new fuzzy synthesized method e-business website is proposed. It reduces influence and increases objectivity result. An example shows effectiveness practicality this method.
A two-speed dual clutch transmission (DCT) model was built, while a Clutch-To-Clutch shift control algorithm based on the closed loop of input speed designed following slipping torque theory.On mat lab/Simulink platform, models were completed for vehicle, DCT gearbox chain and software.A simulation carried out with blade electric SUV to verify effectiveness power logic theoretical feasibility closed-loop reveal relationship between time friction work.The results prove that in this paper can...