A5014860174- Energy Load and Power Forecasting
- Solar Radiation and Photovoltaics
- Electric Power System Optimization
- Photovoltaic System Optimization Techniques
- Advanced Steganography and Watermarking Techniques
- Video Coding and Compression Technologies
- Software System Performance and Reliability
- Digital Media Forensic Detection
- Network Security and Intrusion Detection
- 3D Shape Modeling and Analysis
- Integrated Energy Systems Optimization
- Power System Reliability and Maintenance
- Machine Fault Diagnosis Techniques
- Complex Network Analysis Techniques
- Smart Grid Energy Management
- Bone Tissue Engineering Materials
- Optimal Power Flow Distribution
- Blockchain Technology Applications and Security
- Anomaly Detection Techniques and Applications
- Image Processing and 3D Reconstruction
- Dental Implant Techniques and Outcomes
- Smart Grid Security and Resilience
- Power System Optimization and Stability
- Advanced Data Compression Techniques
- Medical Image Segmentation Techniques
Sichuan University
2020-2025
Huawei Technologies (China)
2023-2024
National Renewable Energy Laboratory
2022-2024
Jilin University
2021-2023
Chinese Academy of Agricultural Sciences
2022-2023
China Medical University
2023
Zhengzhou Normal University
2014-2023
Agricultural Genomics Institute at Shenzhen
2022-2023
Harbin University of Science and Technology
2023
Guizhou Minzu University
2023
Renewable energy forecasting is crucial for integrating variable sources into the grid. It allows power systems to address intermittency of supply at different spatiotemporal scales. To anticipate future impact cloud displacements on generated by solar facilities, conventional modeling methods rely numerical weather prediction or physical models, which have difficulties in assimilating information and learning systematic biases. Augmenting computer vision with machine overcomes some these...
Both deterministic and probabilistic load forecasting (DLF PLF) are of critical importance to reliable economical power system operations. However, most the widely used statistical machine learning (ML) models trained by optimizing global performance, without considering local behaviour. This paper develops a two-step short-term (STLF) model with Q-learning based dynamic selection (QMS), which provides reinforced forecasts (DLFs PLFs). First, pool (DMP) (PMP) built on 10 state-of-the-art ML...
Solar forecasting accuracy is highly affected by weather conditions, therefore, awareness models are expected to improve the performance. However, it may not be available or reliable classify different tasks only using predefined meteorological categorization. In this paper, an unsupervised clustering-based (UC-based) solar method developed for short-term (1-h-ahead) global horizontal irradiance (GHI) forecasting. This UC-based consists of three parts: GHI time series clustering, pattern...
Building occupancy patterns facilitate successful development of the smart grid by enhancing building-to-grid integration efficiencies. Current detection is limited lack widely deployed non-intrusive sensors and insufficient learning power shallow machine algorithms. This paper seeks to detect real-time building from Advanced Metering Infrastructure (AMI) data based on a deep architecture. The developed model consists convolutional neural network (CNN) long short-term memory (LSTM) network....
Wind power ramps are significantly impacting the balance of system operations. Understanding statistical characteristics ramping features would help operators better manage these extreme events. Toward this end, paper develops an analytical generalized Gaussian mixture model (GGMM) to fit probability distributions different features. The nonlinear least-squares method with trust-region algorithm is adopted optimize tunable parameters components. optimal number components adaptively solved by...
Abstract Motivation Protein–protein interaction (PPI), as a relative property, is determined by two binding proteins, which brings great challenge to design an expert model with unbiased learning architecture and superior generalization performance. Additionally, few efforts have been made allow PPI predictors discriminate between properties intrinsic properties. Results We present sequence-based approach, DeepTrio, for prediction using mask multiple parallel convolutional neural networks....
The evolution of fissures and permeability associated with mining the upper protective layer coal seam is crucial for pressure relief gas drainage underlying seam. To understand influence on within seam, this study utilized M16 M18 seams in Qinglong Coal Mine Guizhou. Theoretical analysis, discrete element numerical simulation, field tests were used to characterize fractures effects protected results show that mining-related stress changes controlled development fractures, altering values...
Articular cartilage is refractory to self-healing due the absence of vascular, nervous, and lymphatic systems, its repair remains a clinical challenge. Tissue regeneration through in situ recruitment stem cells via cell-free scaffolds promising alternative strategy. Herein, kind functional injectable hydrogel system (Col-Apt@KGN MPs), which collagen-based microsphere-embedded scaffold, was designed achieve spatiotemporal regulation endogenous mesenchymal (MSCs) their chondrogenic...
Abstract Based on the moving least‐squares (MLS) approximation, we propose a new approximation method—the complex variable (CVMLS) approximation. With CVMLS trial function of two‐dimensional problem is formed with one‐dimensional basis function. The number unknown coefficients in less than MLS and can thus select fewer nodes meshless method that from are required no loss precision. derived also has greater computational efficiency. From for elasticity problems—the (CVMM)—and formulae CVMM...