PFI:BIC Next Generation Real-Time Distributed Manufacturing Service Systems Using Digital Process Planning and GPU-Accelerated Parallel Computing

PFI：BIC 使用数字流程规划和 GPU 加速并行计算的下一代实时分布式制造服务系统

• 批准号: 1631803
• 负责人: Thomas Kurfess
• 金额: $ 100万
• 依托单位: Georgia Tech Research Corporation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2016
• 资助国家: 美国
• 起止时间: 2016-09-01 至 2020-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1631803&HistoricalAwards=false
• 关键词: PFI; BIC; Next; Generation; Real

The research project will enable scaling of novel manufacturing analysis applications in a smart, human-centered manufacturing service system. This manufacturing service system will enable highly functioning relationships between a range of users, including service consumers (e.g., designers, entrepreneurs, makers) and service producers (e.g., manufacturers). Design organizations and new entrepreneurs have limited access to the advanced manufacturing processes needed for innovation due to longstanding barriers associated with: (1) cost and availability of advanced manufacturing platforms and (2) technical skill and experience needed to design manufacturable products in a correct-by-construction framework. While direct-to-print paradigms, such as 3D printing, have increased the capabilities for users to make parts directly from digital product designs, the capital equipment for manufacturing functional components are still housed in mostly private enterprise facilities. Additionally, challenges to accurately specify product manufacturing requirements remain as barriers toward facilitating large-scale, consumer-driven manufacturing services. Smart manufacturing service systems that address these persistent challenges will dramatically accelerate innovation across multiple industrial sectors by removing technological barriers for functional manufacture. The consumerization of manufacturing services will also enable manufacturers with excess production capacity to access a significantly broader customer base, facilitating stabilization of production capacity through decentralization of service demand. Democratization of access to manufacturing capabilities and the associated sharing economy benefits will lead to transformative advances in product realization for the US economy.To facilitate these collaborations, this research project leverages prior discovery for rapid, digital process planning based on novel hybrid dynamic tree representations and graphics processing unit (GPU) accelerated parallel computing. To implement this novel discovery in a distributed service system, the digital process planning method will provide product and process intelligence for each user class. Human subjects experiments and user requirements modeling will inform system design for user interface functionality, service consumer cognitive processing, service provider informational needs, and consumer-to-producer interaction requirements. The service system architecture will be designed to interface with users, facilitate user transactions, receive and shape input data, and enable archival functionality. To realize the foundational digital process planning discovery in a novel human-centered service system, new information models for manufacturing-related data and service provider production capabilities will be established. Advanced computational algorithms will also be derived for scaling GPU-accelerated process analyses in a many-user, cloud-based environment. Comprehensive system testing will inform efficacy of the underlying user requirement models, service system architecture, manufacturing information models and GPU-accelerated computing for generalized human-centered manufacturing service systems.The research project assembles a collaborative team of experts in the fields of manufacturing, design, computing and psychology and a diverse range of industrial partners and nonprofit agencies to realize a next-generation manufacturing service system. The lead institution is Georgia Tech with faculty from mechanical engineering, computing and psychology. The industrial partners, Tucker Innovations (small business), Mazak (large business), Morris South (large business), represent manufacturing technology developers and platform builders with interest in facilitating broad access of enterprise-level technologies to consumer end users. The partner industrial consortia, National Center for Manufacturing Sciences (non-profit), National Center for Defense Manufacturing and Machining (non-profit) and Digital Manufacturing And Design Innovation Institute (non-profit), provide access to a strong membership base spanning many industrial sectors. Collaborations at Georgia Tech with The Center for the Enhancement of Teaching and Learning, Vertically Integrated Projects Program and VentureLab will facilitate connections of research outcomes to educational outcomes for the broad community of makers.

该研究项目将能够在以人为本的智能制造服务系统中扩展新型制造分析应用。该制造服务系统将在一系列用户之间建立高效的关系，包括服务消费者（例如设计师、企业家、制造商）和服务生产者（例如制造商）。由于长期存在的障碍，设计组织和新企业家获得创新所需的先进制造工艺的机会有限：(1) 先进制造平台的成本和可用性；(2) 以正确的方式设计可制造产品所需的技术技能和经验。通过构建框架。虽然 3D 打印等直接打印范例提高了用户直接根据数字产品设计制造零件的能力，但用于制造功能组件的资本设备仍然大多位于私营企业设施中。此外，准确指定产品制造要求的挑战仍然是促进大规模、消费者驱动的制造服务的障碍。解决这些持续挑战的智能制造服务系统将通过消除功能制造的技术障碍，极大地加速多个工业领域的创新。制造服务消费化也将使产能过剩的制造商获得更广泛的客户群，通过分散服务需求促进产能稳定。制造能力的民主化和相关的共享经济利益将带来美国经济产品实现的变革性进步。为了促进这些合作，该研究项目利用先前的发现，基于新颖的混合动态树表示和图形处理单元 (GPU) 加速并行计算。为了在分布式服务系统中实现这一新颖的发现，数字流程规划方法将为每个用户类别提供产品和流程智能。人体实验和用户需求建模将为用户界面功能、服务消费者认知处理、服务提供商信息需求和消费者与生产者交互需求的系统设计提供信息。服务系统架构将被设计为与用户交互、促进用户交易、接收和塑造输入数据以及启用归档功能。为了在以人为本的新型服务系统中实现基础数字流程规划发现，将建立制造相关数据和服务提供商生产能力的新信息模型。还将衍生出先进的计算算法，用于在多用户、基于云的环境中扩展 GPU 加速的过程分析。全面的系统测试将揭示底层用户需求模型、服务系统架构、制造信息模型和 GPU 加速计算的有效性，以实现以人为本的广义制造服务系统。该研究项目聚集了制造、设计领域的专家协作团队、计算和心理学以及各种工业合作伙伴和非营利机构，以实现下一代制造服务系统。主导机构是佐治亚理工学院，拥有机械工程、计算机和心理学教师。工业合作伙伴 Tucker Innovations（小型企业）、Mazak（大型企业）、Morris South（大型企业）代表制造技术开发商和平台构建者，他们有兴趣促进企业级技术向消费者最终用户的广泛获取。合作伙伴工业联盟，国家制造科学中心（非营利性）、国家国防制造和加工中心（非营利性）以及数字制造和设计创新研究所（非营利性），提供了覆盖多个领域的强大会员基础。工业部门。佐治亚理工学院与教学与学习增强中心、垂直整合项目计划和 VentureLab 的合作将促进研究成果与广大创客社区的教育成果的联系。

项目成果

Model-free subtractive manufacturing from computed tomography data

• DOI: 10.1016/j.mfglet.2017.06.004
• 发表时间: 2017-08
• 期刊: Manufacturing letters
• 影响因子: 3.9
• 作者: Jing Yu;Roby Lynn;Thomas M. Tucker;C. Saldana;T. Kurfess

Effects of spatial energy distribution-induced porosity on mechanical properties of laser powder bed fusion 316L stainless steel

• DOI: 10.1016/j.addma.2021.101875
• 发表时间: 2021-03-01
• 期刊: ADDITIVE MANUFACTURING
• 影响因子: 11
• 作者: Jost, Elliott W.;Miers, John C.;Saldana, Christopher
• 通讯作者: Saldana, Christopher

Thin wall deposition of IN625 using directed energy deposition

• DOI: 10.1016/j.jmapro.2020.04.032
• 发表时间: 2020-08-01
• 期刊: JOURNAL OF MANUFACTURING PROCESSES
• 影响因子: 6.2
• 作者: Kim, Myong Joon;Saldana, Christopher
• 通讯作者: Saldana, Christopher

Multi-Axis Voxel-Based CNC Machining of Centrifugal Compressor Assemblies

基于多轴体素的离心压缩机组件 CNC 加工

• 发表时间: 2018
• 期刊: American Helicopter Society Forum 74
• 作者: Kurfess, T.R.

A Deep Ensemble Classifier for Surface Defect Detection in Aircraft Visual Inspection

用于飞机视觉检测中表面缺陷检测的深度集成分类器

• DOI: 10.1520/ssms20200031
• 发表时间: 2020
• 期刊: Smart and Sustainable Manufacturing Systems
• 影响因子: 1
• 作者: Ren, Ivan;Zahiri, Feraidoon;Sutton, Gregory;Kurfess, Thomas;Saldana, Christopher
• 通讯作者: Saldana, Christopher