CAREER: Simulation of Geometrically Flexible Materials with Applications to Computer Graphics and Computational Science

职业：几何柔性材料的模拟及其在计算机图形学和计算科学中的应用

• 批准号: 2153851
• 负责人: Chenfanfu Jiang
• 金额: $ 52.43万
• 依托单位: University of California-Los Angeles
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-10-01 至 2025-05-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2153851&HistoricalAwards=false
• 关键词: CAREER; Simulation; Geometrically; Flexible; Materials

High-fidelity physics-based simulation of 3D materials and natural phenomena has become essential in many computational science domains such as structural engineering and vehicle/aircraft design, as well as a critical component in motion pictures, visual effects (VFX), animation, and video games. Numerical simulations have further found an increasing breadth of new applications such as real-time VFX previews, virtual reality games, interactive surgical training, predictive soft robotics, and computational fabrication. While theoretical computation capacity is now less of an impediment, a timely opportunity emerges for innovations in designing new numerical algorithms that mathematically resolve complex geometry and multi-physics with high accuracy and can best utilize new computational platforms with plausible scalability. While contributing towards this direction, the project will also directly promote modern interdisciplinary studies and education in scientific computing, mechanical engineering, and human-robot interaction. The application to simulating virtual humans will enable clinical training software, which not only improves patient care but also eliminates animal experiments. The support for large-scale geophysical simulation saves lives by improving the prediction of disasters like avalanches and landslides. The innovation of a versatile multi-physics system facilitates advances in climate sciences by modeling Arctic sea ice. This project will produce highly useful software systems for non-simulation experts and educational tools for STEM students. It will also strongly encourage the involvement of undergraduate students, underrepresented minorities, and women through a versatile set of educational events, exchange programs, and outreach activities.This project will develop innovative computational algorithms, including flexible treatment of thin structures with the Material Point Method, a unified multi-material multi-physics framework to capture versatile phenomena, along with novel approaches harnessing the power of next-generation multi-GPU platforms. Co-dimensional geometries (metallic shells, fluid sheets, filaments, biological membranes, fibrous composites, threaded alloys, etc.) will be a primary focus. The project will build innovative geometric representations that are robust for heterogeneous materials, and numerical algorithms that naturally capture multi-physics. The innovative treatment of thin structures will enable new applications such as fiber-level wood crack prediction and fibrous food design/processing. The unified framework will create an exciting opportunity to improve clinical planning and training by enabling high-fidelity biomechanical simulation directly from tomographic imaging, while investigations into numerical stability and computational scalability will advance synergistic domains in computer graphics and computational science at large.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

3D材料和自然现象的高保真物理模拟在许多计算科学领域（如结构工程和车辆/飞机设计）中已经变得至关重要，并且是电影，视觉效果（VFX），动画和视频游戏的关键组成部分。 数值模拟进一步发现了越来越广泛的新应用，如实时VFX预览，虚拟现实游戏，交互式手术培训，预测软机器人和计算制造。虽然理论计算能力现在不那么成为障碍，但在设计新的数值算法方面出现了及时的创新机会，这些算法可以高精度地在数学上解决复杂的几何和多物理问题，并且可以最好地利用具有合理可扩展性的新计算平台。在朝着这个方向做出贡献的同时，该项目还将直接促进科学计算、机械工程和人机交互方面的现代跨学科研究和教育。模拟虚拟人的应用将使临床培训软件成为可能，这不仅改善了患者护理，而且消除了动物实验。对大规模地球物理模拟的支持通过改善对雪崩和山体滑坡等灾害的预测来拯救生命。多功能多物理场系统的创新通过模拟北极海冰促进了气候科学的进步。该项目将为非模拟专家提供非常有用的软件系统，并为STEM学生提供教育工具。该项目还将通过一系列多样化的教育活动、交流项目和外联活动，大力鼓励本科生、代表性不足的少数民族和妇女的参与。该项目将开发创新的计算算法，包括用材料点方法灵活处理薄结构，这是一个统一的多材料多物理框架，可以捕捉多种现象，沿着利用下一代多GPU平台的新方法。共维几何形状（金属壳、流体片、细丝、生物膜、纤维复合材料、螺纹合金等）将成为主要焦点。该项目将建立创新的几何表示，对异质材料和自然捕捉多物理的数值算法是强大的。薄结构的创新处理将使新的应用，如纤维级木材裂纹预测和纤维食品设计/加工。统一的框架将创造一个令人兴奋的机会，以改善临床规划和培训，使高保真生物力学模拟直接从断层成像，该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的知识产权进行评估来支持。优点和更广泛的影响审查标准。

项目成果

Second-order Stencil Descent for Interior-point Hyperelasticity

• DOI: 10.1145/3592104
• 发表时间: 2023-07
• 期刊: ACM Transactions on Graphics (TOG)
• 作者: L. Lan;Minchen Li;Chenfanfu Jiang;Huamin Wang;Yin Yang

A novel discretization and numerical solver for non-fourier diffusion

• DOI: 10.1145/3414685.3417863
• 发表时间: 2020-11
• 期刊: ACM Transactions on Graphics (TOG)
• 作者: Tao Xue;Haozhe Su;Chengguizi Han;Chenfanfu Jiang;Mridul Aanjaneya

HoD-Net: High-Order Differentiable Deep Neural Networks and Applications

• DOI: 10.1609/aaai.v36i8.20799
• 发表时间: 2022-06
• 作者: Siyuan Shen;Tianjia Shao;Kun Zhou;Chenfanfu Jiang;Feng Luo;Yin Yang

The power particle-in-cell method

• DOI: 10.1145/3528223.3530066
• 发表时间: 2022-07
• 期刊: ACM Transactions on Graphics (TOG)
• 作者: Ziyin Qu;Minchen Li;F. D. Goes;Chenfanfu Jiang

High-order differentiable autoencoder for nonlinear model reduction

用于非线性模型简化的高阶可微自动编码器

• DOI: 10.1145/3450626.3459754
• 发表时间: 2021
• 期刊: ACM Transactions on Graphics
• 影响因子: 6.2
• 作者: Shen, Siyuan;Yang, Yin;Shao, Tianjia;Wang, He;Jiang, Chenfanfu;Lan, Lei;Zhou, Kun
• 通讯作者: Zhou, Kun