GAMES Webinar 2017-12期 | 陈德赛（麻省理工学院计算机科学与人工智能实验室）

【GAMES Webinar 2017-12期】

报告嘉宾：陈德赛，麻省理工学院计算机科学与人工智能实验室

报告时间：2017年9月7日（星期四）晚20:00-21:30（北京时间）

主持人：吕琳，山东大学（个人主页：http://irc.cs.sdu.edu.cn/~lulin）

报告题目：Multiscale Methods for Fabrication Design

报告摘要：

Modern manufacturing technologies such as 3D printing enable the design and fabrication of objects with extraordinary complexity. By arranging materials to form functional structures, a much wider range of physical properties can be achieved compared to the constituent materials. Computational design algorithms have previously been developed to automatically design objects with specified physical properties. However, many types of physical properties are still very challenging to optimize because predictive and efficient simulations are not available for problems such as high-resolution non-linear elasticity or dynamics with friction and impact. Even with problems such as linear elasticity where accurate simulation is available, the simulation resolutions is still orders of magnitudes below available printing resolutions. We propose to speed up simulation and inverse design process of fabricable objects by using multiscale methods. First, we construct a library of microstructures and their bulk material properties. Each bulk material property is stored using a small number of parameters. At runtime, fine element structures are replaced by coarse elements with precomputed material properties to speed up simulation. For validation, we performed design optimization and real-world fabrication for static and dynamic elastic objects. For 3D cubic-symmetric microstructures with linear material models, we also explored the space of achievable bulk parameters that we call the material property gamut. The gamut is represented as a level set. The level set representation allows us to use discrete and continuous sampling methods to push the structures to the limits of achievable material properties. It is also used in topology optimization algorithms to design functional objects using microstructures. Finally, the gamut also enables automatic discovery of microstructures families with extremal physical properties.

讲者简介：

陈德赛是麻省理工学院第六年博士生。2011年本科毕业于卡内基梅隆大学计算机系。2013年于麻省理工获得计算机系硕士学位。研究兴趣主要是固体的有限元模拟，使用模拟和优化算法来设计实际物体，提高实际部件的特性。尤其注重于在3D打印方面的应用。主页：http://people.csail.mit.edu/desaic/

直播链接：http://webinar.games-cn.org