CHS: Small: Collaborative Research: Computational Acoustic Design for Digital Manufacturing

CHS：小型：协作研究：数字制造的计算声学设计

• 批准号: 1815372
• 负责人: Wojciech Matusik
• 金额: $ 25万
• 依托单位: Massachusetts Institute of Technology
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-09-01 至 2021-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1815372&HistoricalAwards=false
• 关键词: CHS; Small; Collaborative; Research; Computational

The acoustic characteristics of objects are among the most important properties that influence or define their function. Thus, design for fabrication of objects with high-fidelity acoustic specifications is a widely desired engineering task. Currently, this design process remains slow and expensive, commonly requiring many trial-and-error iterations; as a consequence, an object's acoustic properties are often not optimized at all, as to do so would drastically increase the design's cost and time. The long and suboptimal design cycle further limits the use of new materials and the exploration of shape variations. This research aims to address this fundamental challenge by developing new numerical models that simulate and optimize acoustics. Project outcomes will include a suite of computational tools that design the geometric shape and structure of an object so as to realize its acoustic properties using novel fabrication methods such as additive manufacturing. These tools have the potential to improve the function of many of today's products, thereby shifting the current paradigm of visual design into computational audiovisual design. The new tools will be made available to a large population of expert designers, engineers, hobbyists, and students, thereby facilitating the design, manufacture and customization of musical instruments, mechanical structures, consumer products, and novel materials with unique acoustic properties. Additional broad impact will derive because students trained by this project will become skilled in developing, extending, and harnessing numerical algorithms and interactive design tools in a wide array of engineering design tasks. The investigators will seek participation from underrepresented groups at the doctoral, masters, and undergraduate levels.This project advances a complete pipeline of computational design for manufacturing objects with high-fidelity acoustic specifications. When provided with the desired acoustic properties of an object such as sound spectrum and frequency scattering profile, the pipeline will computationally design the object's geometric shape and structure to realize its acoustic properties using digital fabrication. The proposed methods will bridge the longstanding gap between parametric computer-aided design (CAD) models and discrete finite-element meshes, and will embrace automatic computational optimization of designed structures with the user in the loop. Thus, they will reshape today's acoustic design, shifting from its current "design then analysis" process to a "design with analysis" process. To this end, the research will address a number of fundamental research questions including: (1) accurate, efficient design-oriented simulation methods to predict sound properties of a given design; (2) automated optimization algorithms that translate desired acoustic properties to shape and material specifications; and (3) efficient computer aided design systems with intuitive, interactive user interfaces incorporating simulation codes. The approach will address all these aspects in an integrated interdisciplinary research plan, leveraging the expertise of both investigators in their respective fields.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.

物体的声学特性是影响或定义其功能的最重要的特性之一。因此，设计制造具有高保真声学规格的物体是一项广泛需要的工程任务。目前，这种设计过程仍然缓慢且昂贵，通常需要多次反复试验；因此，一个物体的声学特性往往根本没有得到优化，因为这样做会大大增加设计的成本和时间。长而次优的设计周期进一步限制了新材料的使用和形状变化的探索。本研究旨在通过开发模拟和优化声学的新数值模型来解决这一基本挑战。项目成果将包括一套计算工具，用于设计物体的几何形状和结构，以便使用新型制造方法（如增材制造）实现其声学特性。这些工具有可能改善当今许多产品的功能，从而将当前的视觉设计范式转变为计算视听设计。这些新工具将提供给大量的专家设计师、工程师、爱好者和学生，从而促进乐器、机械结构、消费产品和具有独特声学特性的新材料的设计、制造和定制。其他广泛的影响将会产生，因为这个项目训练的学生将在开发、扩展和利用数值算法和交互设计工具在广泛的工程设计任务的技能。研究人员将寻求在博士、硕士和本科阶段未被充分代表的群体的参与。该项目推进了一个完整的计算设计管道，用于制造具有高保真声学规格的物体。当提供了物体所需的声学特性，如声谱和频率散射剖面时，管道将计算设计物体的几何形状和结构，以使用数字制造实现其声学特性。所提出的方法将弥合参数化计算机辅助设计（CAD）模型和离散有限元网格之间长期存在的差距，并将在用户参与的情况下实现设计结构的自动计算优化。因此，他们将重塑今天的声学设计，从目前的“先设计后分析”的过程转变为“先设计后分析”的过程。为此，本研究将解决一些基础研究问题，包括：(1)准确、高效的以设计为导向的仿真方法来预测给定设计的声音特性；(2)将所需声学特性转化为形状和材料规格的自动优化算法；(3)高效的计算机辅助设计系统，具有包含仿真代码的直观、交互式用户界面。该方法将在一个综合的跨学科研究计划中解决所有这些方面，利用两位研究者在各自领域的专业知识。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Diversity-Guided Multi-Objective Bayesian Optimization With Batch Evaluations

• 发表时间: 2020
• 作者: Mina Konakovic-Lukovic;Yunsheng Tian;W. Matusik