A5036221608- Building Energy and Comfort Optimization
- Energy Load and Power Forecasting
- Urban Heat Island Mitigation
- Integrated Energy Systems Optimization
- Air Quality Monitoring and Forecasting
- Wind and Air Flow Studies
- Structural Health Monitoring Techniques
- Water Quality Monitoring and Analysis
- High Altitude and Hypoxia
- Control Systems and Identification
- History and advancements in chemistry
- Textile materials and evaluations
- Power Systems and Renewable Energy
- Advanced Physical and Chemical Molecular Interactions
- Image and Video Quality Assessment
- Infection Control and Ventilation
- Thermoregulation and physiological responses
- Mineral Processing and Grinding
- Advanced Chemical Physics Studies
- Adipose Tissue and Metabolism
- Advanced Algorithms and Applications
- Neural Networks and Applications
- Solar Radiation and Photovoltaics
- Hydrological Forecasting Using AI
- Infrared Thermography in Medicine
Tianjin University of Technology
2022-2023
National Engineering Research Center of Electromagnetic Radiation Control Materials
2023
Tsinghua University
2002-2013
Universiti Sains Malaysia
2013
Academy of Military Medical Sciences
1991
The drastic growth in the consumption of electrical household appliances becomes a concerned issue recent years. It is difficult for traditional methods to consider both subject's comfort and energy savings. Moreover, different people have preferences requirements. In this paper, novel integrated control method blind lights proposed it can provide an intelligent comfortable living environment office building. presented has ability intelligently learn subjects' by analyzing their behaviors....
Pollutant forecasting is an important problem in the environmental sciences. Data mining approach to discover knowledge from large data. This paper tries use data methods forecast ?PM?_(2.5) concentration level, which air pollutant. There are several tree-based classification algorithms available mining, such as CART, C4.5, Random Forest (RF) and C5.0. RF C5.0 popular ensemble methods, are, builds on CART with Bagging C4.5 Boosting, respectively. level predictive models based by using R...
Intelligent building system attracts more and attention in both academic industrial communities. Learning human comfort requirements incorporating it into control is one of the important issues. In traditional HVAC system, thermal acoustic are often conflicted we lack a scheme to trade off them well. this paper, propose unified framework based on reinforcement learning balance multiple dimension comforts, including comforts. We utilize user's complaints sensations as feedback combine current...
Learning the model of user's thermal complaints in daily office environment and apply it to control Heating, Ventilating, Air Conditioning (HVAC) is more user-friendly intelligent. But modeling complaint behavior challenging because stochastic time, unbalanced sample sets, individual differences uncertainties. Most existing human comfort models are not feasible under those conditions. In this paper, we propose a novel one-class (complaint only) classifier complaints. The method extracts...
When the output observation noise is output-dependent, identifying unknown system parameters becomes challenging. Traditional methods based on Mean Square Error, even ones with corrections still have biased estimations in this case. Many practical cases such as bounded sensor, uncertainty of expression human involved identification, and physiological or biological model identification actually problem. In paper, some algorithms were proposed to obtain unbiased estimation for...
A facile approach to control the cooling rate of high temperature charging was first suggested. The furnace tube surrounded with electrical wires put into Dewar flask whose degree vacuum controlled. Switching off power and changing when heated desired value. Since Dewar's thermal conducting coefficient mainly depends on vacuum, so controllable a great extent. corresponding experiment carried out prove its feasibility.
Accurate heat load prediction is a prerequisite for feed-forward control and on-demand supply in district heating system. However, considering that the experimental data used to train model are often not optimal or most energy efficient, accurate difficult achieve effective energy-saving. This paper proposes hybrid energy-saving combines similar sample selection approach (SSA) deep neural network. A new weighted Euclidean norm (WEN) select suitable datasets, novel strategy proposed reduce...