Speaker：Natasa Trisovic,教授，贝尔格莱德大学

邀请人：李伟

报告时间：2022年5月13日-2022年5月15日

腾讯会议ID：502 7594 2895

报告人简介：塞尔维亚贝尔格莱德大学机械工程学院教授，CEEPUS(Central European Exchange Program for University Studies)计划成员，捷克工程大学力学工程学院客座教授（2009），波西尼亚巴尼亚卢卡大学力学工程学院客座教授（2010，2012），斯洛伐克科技大学力学工程学院客座教授，美国莱斯大学MEMS学院客座教授（2012-2014），EUREKA计划和ESPRIT项目成员，塞尔维亚结构完整性与生命协会成员，塞尔维亚力学学会秘书长，研究方向为理论与应用力学，主要集中于对结构系统振动疲劳可靠性的数值研究。发表学术论文60多篇，参加项目10余项。

报告题目1：CONFIDENCE INTERVALS FOR DIFFERENCE OF POPULATION VARIANCES

Abstract：We consider confidence intervals for the population variance and the difference in variances of two populations based on the ordinaryt-statistics combined with the bootstrap method. Theoretical and practical aspects of the suggested techniques are presented, as well as their comparison with existing methods (methods based on Chi-square statistics andF-statistics). In addition, application of presented methods in domain of insurance property data set is described and analyzed. For data from exponential distribution confidence intervals, which are calculated using described methods (based on transformation of thet-statistics and bootstrap technique), give consistent and best coverage in comparisons with other methods.

报告题目2：Conceptual design of a plant for anaerobic biological treatment of sludge from the process of sanitary wastewater treatment for a city of 50000 ES

Abstract：Reducing greenhouse gas emissions, and thus preventing climate change, is one of the greatest challenges of our time. Consequently, a significant reduction in the use of fossil fuels and a greater use of renewable energy sources are essential. One of these sources is biogas, the production of which I will study in more detail in this paper.

报告题目3：INFLUENCE OF DIFERENT METAMATERIAL GEOMETRY TO MECHANISM’S SINTESYS

Abstract：This paper investigates the influence of different geometrical structures on new concepts for the formation of technical systems. The ability that some geometric structures are able to withstand a certain level of deformation, was used to replace joints in certain assemblies of technical systems. Now all movements are accomplished with the deformation of geometrical structures. Obtained results from simulations, define the level of deformation that structures can withstand. Designing of 3D models and simulations was conducted in SOLIDWORKS 2016. Several different structures of metamaterials will be examined. 64 simulations were conducted by changing the internal structure, thickness, and orientation of metamaterials. For each simulation results were presented as stresses and displacements. Additionally, three models of pliers have been simulated, and their results were compared with results from previous simulations.

报告题目4：MACHINE LEARNING BASED ARM SWING MODEL DURING HUMAN WALKING

Abstract：Wrist-worn energy harvesters have great potential to power wearable devices for Internet of Things sustainably. Considerable attentions have been paid to improve their energy conversion efficiency through structural designs and smart materials. However, little research about the model of arm swing movements was conducted in previous studies, which significantly influenced the power generation of wrist-worn energy harvesters. To fill the gap of energy harvesting in arm swing, firstly,we proposethe dynamic model of accelerations on wrist during human walking. It reveals that the body accelerations and arm swing movements have positive effects on the power generation. Then to accurately describe arm swing movements, the walking experimental data of 300 participants (150 female and 150 male) were tested by different machine learning methods, thus showing that the Random Forest regression (RFR) method has better performance for arm swing predictions. Based on it, we construct the arm swing (shoulder-angle and elbow-angle) models trained by the experimental dataset. With the statistical analysis for the derived models, it is clarified thatthe sequence of feature importance influencing arm swing movements is walking speed, height, weight, body mass index (BMI) and age. Comparing prediction results between the proposed arm swing model and previous linear regression model, the RFR model could more successfully describe the arm swing movements of the experimental dataset. These findings provide theoretical guidance for further design and optimization of wrist-worn energy harvesters in the future