A5100610894- Machine Fault Diagnosis Techniques
- Energy Load and Power Forecasting
- Solar Radiation and Photovoltaics
- Magnetic Field Sensors Techniques
- Power Systems and Technologies
- Power System Reliability and Maintenance
- Reinforcement Learning in Robotics
- Photovoltaic System Optimization Techniques
- Fault Detection and Control Systems
- Non-Destructive Testing Techniques
- Advanced Sensor and Control Systems
- Wind Energy Research and Development
- Power Systems and Renewable Energy
- Gear and Bearing Dynamics Analysis
- Magnetic Properties and Applications
- Sensor Technology and Measurement Systems
- Advanced Decision-Making Techniques
- Power Transformer Diagnostics and Insulation
- Electric Power System Optimization
- High voltage insulation and dielectric phenomena
- Smart Grid and Power Systems
- Structural Health Monitoring Techniques
- Advanced Algorithms and Applications
- Magneto-Optical Properties and Applications
- Geoscience and Mining Technology
Beihang University
2015-2024
PowerChina (China)
2022-2024
Zhejiang University of Science and Technology
2018-2024
Nanjing Agricultural University
2022-2024
Henan Normal University
2022-2023
South China Agricultural University
2023
Tongji University
1991-2023
State Key Laboratory of Oral Diseases
2023
Sichuan Agricultural University
2023
Guizhou Minzu University
2022-2023
Since wind at the earth's surface has an intrinsically complex and stochastic nature, accurate power forecasts are necessary for safe economic use of energy. In this paper, we investigated a combination numeric probabilistic models: Gaussian process (GP) combined with numerical weather prediction (NWP) model was applied to wind-power forecasting up one day ahead. First, wind-speed data from NWP corrected by GP, then, as there is always defined limit on generated in turbine due controlling...
Condition monitoring and early fault detection of wind turbine faults can reduce maintenance costs prevent cascaded failures. This article proposes a new Normal Behavior Modeling (NBM) method to predict electric pitch system failures using supervisory control data acquisition (SCADA) information. The proposed is particularly effective for online applications at reasonable computational complexity. Briefly, in the preprocessing stage method, order remove interferential information improve...
Unsupervised domain adaptation methods aim to alleviate performance degradation caused by domain-shift learning domain-invariant representations. Existing deep focus on holistic feature alignment matching source and target distributions, without considering local features their multi-mode statistics. We show that the learned patterns are more generic transferable a further distribution enables fine-grained alignment. In this paper, we present method for jointly aligning Comparisons...
In deep reinforcement learning, off-policy data help reduce on-policy interaction with the environment, and trust region policy optimization (TRPO) method is efficient to stabilize procedure. this article, we propose an TRPO method, TRPO, which exploits both on- guarantees monotonic improvement of policies. A surrogate objective function developed use keep We then optimize by approximately solving a constrained problem under arbitrary parameterization finite samples. conduct experiments on...
Camellia oleifera Abel. is an economically important woody oil plant native to China. To explore the genetic diversity of wild C. phenotypic traits and effectively protect these germplasm resources, this study provides a thorough evaluation variability cluster 143 resources. A total 41 characters, including leaves, flowers, fruits, seeds, quality were investigated based on quantization physical chemical descriptors digital image analysis. The findings revealed significant variations among...
Fault diagnosis of wind turbine (WT) gearboxes can reduce unexpected downtime and maintenance costs. In this paper, a new fault framework is proposed based on deep bi-directional Long Short-term Memory (DB-LSTM). Even though learning has been used in rotating machines, models with the input raw time-series or frequency data face computational challenges. Additionally, deviation between datasets be triggered easily by operating condition variation, which will highly performance models....
We propose a fast and accurate determination method for transverse relaxation of the spin-exchange-relaxation-free (SERF) magnetometer. This is based on measurement magnetic resonance linewidth via chirped field excitation amplitude spectrum analysis. Compared with frequency sweeping separate sinusoidal excitation, our can realize within only few seconds meanwhile obtain good resolution. Therefore, it avoid drift error in long term improve accuracy determination. As SERF magnetometer very...
Accurate wind speed forecasts are necessary for the safety and economy of renewable energy utilization. The can be obtained by statistical model based on historical data. In this paper, a novel W-GP (wavelet decomposition Gaussian process learning paradigm) is proposed short-term forecasting. nonstationary nonlinear original series first decomposed into set better-behaved constitutive subseries wavelet decomposition. Then these sub-series forecasted respectively GP method, forecast results...
Wind turbine yaw control plays an important role in increasing the wind production and also protecting turbine. Accurate measurement of angle is basis effective controller. The accuracy affected significantly by problem zero-point shifting. Hence, it essential to evaluate shifting error on turbines on-line order improve reliability real time. Particularly, qualitative evaluation could be useful for farm operators realize prompt cost-effective maintenance sensors. In aim qualitatively...
The deep Q-network (DQN) and return-based reinforcement learning are two promising algorithms proposed in recent years. DQN brings advances to complex sequential decision problems, while have advantages making use of sample trajectories. In this paper, we propose a general framework combine most the algorithms, named R-DQN. We show performance traditional can be improved effectively by introducing learning. order further improve R-DQN, design strategy with measurements which qualitatively...
Improving sample efficiency has been a longstanding goal in reinforcement learning. This paper proposes VRMPO algorithm: efficient policy gradient method with stochastic mirror descent. In VRMPO, novel variance-reduced estimator is presented to improve efficiency. We prove that the proposed needs only O(ε−3) trajectories achieve an ε-approximate first-order stationary point, which matches best complexity for optimization. Extensive empirical results demonstrate VRMP outperforms...
Based on the combination of gas engine cogeneration heat and power (GECHP) technology temperature humidity independent control air-conditioning technology, an system combined with desiccant wheel driven by GECHP was proposed to decrease cooling energy consumption utilize waste engine. In this study, Dymola software adopted model hybrid conduct hourly simulation throughout heating season for two types large public buildings: supermarket building office building. Results show that applied...