FMRG: Cyber: Cyber-Coordinated Analytical Framework for Multi-stage Distributed Future Manufacturing Systems

FMRG：网络：多阶段分布式未来制造系统的网络协调分析框架

• 批准号: 2134409
• 负责人: Hongyue Sun
• 金额: $ 233.31万
• 依托单位: SUNY at Buffalo
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-11-15 至 2024-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2134409&HistoricalAwards=false
• 关键词: FMRG; Cyber; Coordinated; Analytical; Framework

A manufactured product is the end of a long chain of interwoven manufacturing steps that may span geography, industries, and manufacturing processes. Each step may be optimized, yet for a step there may be uncertainty about the larger context of its contribution that if known during production would allow for greater efficiencies, better quality, and improved productivity. If a global system analysis and optimization could, in addition, be conducted cooperatively with each step during production, the end product and overall production would greatly benefit. Advances in hybrid distributed control and optimization, in highly networked cyber-physical systems, and in artificial intelligence and machine learning have opened new opportunities in cyber-enabled manufacturing. This Future Manufacturing Research Grant (FMRG) CyberManufacturing project connects and coordinates the ensemble of various manufacturing processes, operating at different manufacturing stages under a cyber-coordinated close-knit system, defined by an analytical framework, known as STREAM, a multi-stage distributed future manufacturing system. This project provides a public online repository that hosts the STREAM data, models, simulators, controllers, analytics, and empirical studies, and that facilitates sharing research experiences about STREAM with the research and industry communities. The project creates new outreach and workforce development activities for K-12, undergraduate and graduate students, as well as working professionals in the field of manufacturing. These research results will be included in the university curriculum for advanced manufacturing, architecture design, machine learning, simulation, and system control and optimization. This project is an interdisciplinary research project with three interconnected research tasks: (1) STREAM architecture design and implementation. There is a novel bridging middleware architecture to enable seamless machine interoperability, and software- and hardware-based accelerators will be built to enable efficient communication and computing in cyber-manufacturing systems. (2) STREAM modeling and quality control. There is a novel multi-layer functional graphical model to address the high-dimensional functional dependencies over STREAM machines and a spatial-temporal iterative learning control method to achieve an accurate and efficient process quality control. (3) STREAM simulation and production control. To handle the multi-level dynamic process interactions in the system that collectively meet process quality, cost, and resource constraints, there are coordinated, high-fidelity multi-resolution simulators, assisting with the hybrid control of STREAM production. These research tasks are validated in an open-source additive manufacturing testbed for the manufacture of high energy/power density supercapacitors. The modeling and analysis methodologies form a rigorous basis for continuous improvement of quality, manufacturability, and productivity of future multi-stage and distributed manufacturing systems.This project is jointly funded by the Division of Computer and Network Systems (CNS) in the Directorate for Computer and Information Science and Engineering (CISE) and the Divisions of Civil. Mechanical and Manufacturing Innovation (CMMI) and Engineering Education and Centers (EEC) in the Directorate for Engineering (ENG).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.

制造出来的产品是一长串交织在一起的制造步骤的末端，这些步骤可能跨越地理、行业和制造过程。每个步骤都可以进行优化，但是对于一个步骤，如果在生产过程中知道其贡献的更大背景，则可能存在不确定性，该背景将允许更高的效率、更好的质量和改进的生产率。如果在生产的各个环节都能协同进行全局的系统分析和优化，最终产品和整体生产都会受益匪浅。混合分布式控制与优化、高度网络化的信息物理系统、人工智能和机器学习等技术的进步，为网络化制造带来了新的机遇。这个未来制造研究基金（FMRG）网络制造项目连接和协调各种制造过程的集合，在一个网络协调的紧密结合的系统下，在不同的制造阶段运行，由一个被称为STREAM的分析框架定义，一个多阶段分布式未来制造系统。该项目提供了一个公共在线存储库，其中包含STREAM数据、模型、模拟器、控制器、分析和实证研究，并促进了与研究和行业社区分享有关STREAM的研究经验。该项目为K-12，本科生和研究生以及制造业领域的专业人士创造了新的推广和劳动力发展活动。这些研究成果将被纳入大学先进制造、建筑设计、机器学习、仿真、系统控制与优化等课程。本项目是一个跨学科的研究项目，有三个相互关联的研究任务：(1)STREAM架构的设计和实现。有一种新颖的桥接中间件体系结构，可以实现无缝的机器互操作性，基于软件和硬件的加速器将被构建，以实现网络制造系统中的高效通信和计算。(2) STREAM建模和质量控制。本文提出了一种新的多层功能图形模型来解决流机器的高维功能依赖，并提出了一种时空迭代学习控制方法来实现准确有效的过程质量控制。(3)流模拟与生产控制。为了处理系统中共同满足过程质量、成本和资源约束的多层次动态过程交互，有协调的、高保真的多分辨率模拟器，协助流生产的混合控制。这些研究任务在一个用于制造高能量/功率密度超级电容器的开源增材制造试验台上进行了验证。建模和分析方法为未来多阶段和分布式制造系统的质量、可制造性和生产率的持续改进奠定了严格的基础。该项目由计算机和信息科学与工程理事会（CISE）的计算机和网络系统司（CNS）以及土木司共同资助。机械和制造创新（CMMI）和工程教育和中心（EEC）的工程理事会（ENG）。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

A Campus Prototype of Interactive Digital Twin in Cyber Manufacturing

网络制造中交互式数字孪生的校园原型

• DOI: 10.1145/3560905.3568049
• 发表时间: 2022
• 期刊: ACM
• 作者: Rubino, Matthew;Weng, Michelle;Chen, Jiasheng;Saptarshi, Shardul;Francisco, Marcus;Francisco, Alex;Zhou, Chi;Sun, Hongyue;Xu, Wenyao
• 通讯作者: Xu, Wenyao

DTFab: A Digital Twin based Approach for Optimal Reticle Management in Semiconductor Photolithography

• DOI: 10.1007/s11518-023-5564-x
• 发表时间: 2023-05
• 期刊: Journal of Systems Science and Systems Engineering
• 影响因子: 1.2
• 作者: Chandrasekhar Komaralingam Sivasubramanian;Robert W. Dodge;Aditya S. Ramani;David Bayba;M. Janakiram

Multiclass Reinforced Active Learning for Droplet Pinch-Off Behaviors Identification in Inkjet Printing

喷墨打印中液滴夹断行为识别的多类强化主动学习

• DOI: 10.1115/1.4057002
• 发表时间: 2023
• 期刊: Journal of Manufacturing Science and Engineering
• 作者: Li, Zebin;Segura, Luis Javier;Li, Yifu;Zhou, Chi;Sun, Hongyue
• 通讯作者: Sun, Hongyue

Analytical Study and Experimental Verification of Shear-Thinning Ink Flow in Direct Ink Writing Process

直写墨水剪切稀化流动的分析研究与实验验证

• DOI: 10.1115/1.4056926
• 发表时间: 2023
• 期刊: Journal of Manufacturing Science and Engineering
• 作者: Guo, Zipeng;Fei, Fan;Song, Xuan;Zhou, Chi
• 通讯作者: Zhou, Chi

A two-stage data-driven approach to remaining useful life prediction via long short-term memory networks

• DOI: 10.1016/j.ress.2023.109332
• 发表时间: 2023-05
• 期刊: Reliab. Eng. Syst. Saf.
• 作者: Huixin Zhang;Xiaopeng Xi;Rong Pan