线上学术报告通知
时间：2020年12月11日上午
会议室: 腾讯会议909 491 034
主持人：陈宇翔博士（加拿大阿尔伯塔大学）
时间安排：
1) 9:00-9:45 “Cooling and heating the buildings by using energy deep underground and deep space”
Dr. Yongqiang Luo, Assistant Professor, School of Environmental Science and Engineering, Huazhong University of Science and Technology
Abstract: Energy data from China, UN or IEA are all pointing out that building is the biggest energy consumer by taking about 30% of the total contribution. Since the conventional energy is limited and our request for energy utilization is increasing, the renewable, clean and sustainable energy sources should be the way out. In our recent study, we tend to believe that we can use the energy far away from us, to provide an highly efficient way of cooling and heating of buildings. First, we showed that the deep geothermal energy is very suitable for building heating. And we developed the first analytical model to describe the complex heat transfer process. Secondly, we showed that the cooling energy from deep space is very suitable for building cooling. Some simulations as well as experiments are providing encouraging results. In conclusion, we think the energy deep underground and deep space has the potential to better serve for building energy saving and facilitate the promise made by President Xi about “Carbon Neutral” before 2060.
2) 9:45-10:30 “Ergonomic risk assessment and productivity improvement in industrialized construction”
Dr. Xinming Li, Assistant Professor, Department of Mechanical Engineering, University of Alberta
Abstract: The construction manufactory industry has a disproportionately high number of lost-time injuries due to the high physical demand of labour-intensive tasks despite the involvement of machinery. These conditions, in turn, delay the production due to the lack of immediate worker replacement when injuries occur. Our research team seeks improvements in construction by identifying ergonomic risks and investigating corresponding corrective measures to secure the health and safety of workers and enhance workplace productivity. Our areas of focus include physical demand analysis, physiological measurement, 3D visualization ergonomic analysis, virtual reality, motion capture, lean manufacturing and workspace design. The outcomes of our research help the industry partner to reduce workplace injuries and claims, develop a strong return-to-work program, reduce workers’ compensation premium rates, and improve productivity.
3) 10:30-11:15 “Data-driven models for thermal dynamic analysis of buildings”
Dr. Yuxiang Chen, Assistant Professor, Department of Civil and Environmental Engineering, University of Alberta
Abstract: Data-driven modelling has been widely applied in building operation optimization, energy management, ongoing commissioning, and so on. This study focuses on data-driven modelling for the analysis of building thermal dynamics – parameter estimation and thermal response prediction. First, three types of data-driven models, namely, transfer-function based models (TF models), resistor-capacitor based models (RC models), and artificial-intelligence-based models (AI models) are compared for their interpretability of physical meanings, and prediction accuracy. Then further study is conducted on the development of RC models. A backward selection approach involving a genetic algorithm and asymptotic confidence intervals is used to identify the model structures with identifiable parameters. A Bayesian estimation approach with unscented Kalman filtering is used to improve the estimation of the parameter based on offline data. Then the Bayesian approach is enhanced with joint state-parameter estimation for real-time thermal dynamic analysis. A three-storey house with in-situ measured data is used for validating and demonstrating the methodologies.
4) 11:15-12:00 “Computer vision for behavior-based safety in construction”
Dr. Weili Fang, Research Scientist, School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology
Abstract: People unsafe behaviour is a significant contributor to accidents in construction. A behavior-based safety (BBS) approach has been proven to be an effective approach to address this problem. However, it should be acknowledged that BBS approaches have been widely criticized for workers not reporting unsafe behaviors and neglecting the root cause of unsafe behavior, ignoring issues regarding values and attitudes, and hiding management commitment and inadequacies. Against this contextual backdrop, our research develops a robust and proactive approach for BBS by utilizing the techniques of computer vision, and deep learning approaches, including: (1) automated identification of people unsafe behaviour; (2) knowledge extraction and recommendation for behaviour modification; and (3) automated prediction and tracking of people unsafe behaviour.
