A5101453831- Cyclone Separators and Fluid Dynamics
- Fluid Dynamics and Heat Transfer
- Plant Surface Properties and Treatments
- Combustion and flame dynamics
- Electrohydrodynamics and Fluid Dynamics
- Remote-Sensing Image Classification
- Remote Sensing and Land Use
- Machine Fault Diagnosis Techniques
- Gear and Bearing Dynamics Analysis
- Turbomachinery Performance and Optimization
- Image Retrieval and Classification Techniques
- Mechanical Failure Analysis and Simulation
- Coal Combustion and Slurry Processing
- Spectroscopy and Chemometric Analyses
- Wind and Air Flow Studies
- Fluid Dynamics Simulations and Interactions
- Particle Dynamics in Fluid Flows
- Mechanical stress and fatigue analysis
- Heat Transfer Mechanisms
- Fluid Dynamics and Mixing
University of Science and Technology of China
2024-2025
Yanshan University
2018-2020
Recently, deep learning has been used for hyperspectral image classification (HSIC) due to its powerful feature and ability. In this letter, a novel learning-based framework based on DeepLab is proposed HSIC. Inspired by the excellent performance of in semantic segmentation, applies excavate spatial features (HSI) pixel pixel. It breaks through limitation patch-wise most existing methods More importantly, it can extract at multiple scales effectively avoid reduction resolution. Furthermore,...
Gearboxes are critical components in wind turbines, and their fault diagnosis has gained increasing considerable attention. Compared to traditional vibration-based methods, current-based significant advantages terms of cost, implementation, reliability. However, it is quite challenging extract informative fault-related features from raw current signals due the presence dominant fundamental component harmonic as well electrical noise. In order address this challenge, article presents a novel...
Abstract To study the coherent structure in a combustion chamber equipped with double-stage swirler, Unsteady Reynolds Averaged Navier-Stokes (URANS) of Realizable k-ε and renormalization Group (RNG) , along Large Eddy Simulation (LES) are employed. The time-averaged results indicate that three turbulence models exhibit certain level consistency experimental data. Analysis instantaneous streamlines Q criterion iso-surface reveal unsteady demonstrates poor predictive capablity. In contrast,...
This paper proposes a new intelligent fault diagnosis approach based on multimodal deep learning to fuse vibration and current signals diagnose wind turbine gearbox faults. The proposed method typically consists of modality-specific feature network fusion network, specifically popular model named belief networks (DBNs). First, two individual DBNs are designed learn fault-related features directly from raw signals, respectively. Then, the learned vibration-based current-based further fused by...
The investigation of droplet impingement plays a crucial role in understanding the dynamics fuel impact on walls inside engine combustion chambers. To study and heat transfer characteristics two-component mixed droplets impacting upon an inclined stainless steel heated wall, effect wall tilt angle (0°–40°), Weber numbers (We, ranging from 50 to 210), temperatures (Tw, between 25 350 °C) models spreading diameters is conducted. Through quantitative analysis high-speed images captured during...
In this paper, the isothermal swirling flow in a combustion chamber equipped with double-stage swirler is studied by combination of experiments and numerical simulations at Reynolds number (Re) ranging from 2712 to 43 396. The swirl numbers inside outside entrances are 0.81 0.89, respectively. effect Re on mean field, oscillation evolution characteristics instantaneous vortex structures such as breakdown precessing core (PVC) systematically analyzed. It found that there significant...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
Aiming at the problem of image information loss, dilated convolution is introduced and a novel multi-scale convolutional neural network (MDCNN) proposed. Dilated can polymerize without reducing resolution. The first layer used spectral step to reduce dimensionality. Then aggregation extracted features through applying shortcut connection. which represent properties data were fed Softmax predict samples. MDCNN achieved overall accuracy 99.58% 99.92% on two public datasets, Indian Pines Pavia...