CAREER: Multiobjective Learning Control Strategies for Additive Manufacturing

职业：增材制造的多目标学习控制策略

• 批准号: 1254313
• 负责人: Sandipan Mishra
• 金额: $ 40万
• 依托单位: Rensselaer Polytechnic Institute
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-06-01 至 2019-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1254313&HistoricalAwards=false
• 关键词: CAREER; Multiobjective; Learning; Control; Strategies

The research objective of this Faculty Early Career Development (CAREER) award is to create system-theoretic design and analysis tools for high-throughput and high-precision control of additive manufacturing. Despite their tremendous potential, additive manufacturing have not yet delivered functional parts at a production scale due to poor process reliability. A key reason for the low reliability is the lack of accurate process models coupled with open-loop process control. To counter this challenge, the unique layer-by-layer material deposition mechanism of additive manufacturing will be exploited by learning algorithms that rely on iterative refinement based on sensor data in conjunction with partial model information. Inspired by multi-objective optimization, the proposed layer-to-layer learning control algorithm will simultaneously address multiple conflicting objectives to deliver acceptable part geometry and mechanical properties, in the presence of modeling errors and uncertainties in operating conditions. Deliverable of this project include (1) design and analysis tools for multi-objective learning control algorithms tailored towards jet-based printing and selective metal melting processes, (2) verification of proposed algorithms on experimental test beds, (3) development of hands-on experimentation and interactive course modules on additive manufacturing and automation at the K-12 level, and (4) training of undergraduate and graduate students for careers in automation, manufacturing, and control. The successful completion of this project will make a strong positive impact on the rapidly expanding billion-dollar additive manufacturing industry by increasing reliability, repeatability, and production rate. Specifically, enhancing reliability and performance of polymer-based additive manufacturing process can have a significant impact on the fabrication of one-off mass-customized parts including biomedical implants, opto-mechanical components, and tissue engineering. On the other hand, reliable metal-based additive manufacturing processes can produce parts for safety-critical applications such as in aviation industry. Through the proposed outreach and education plan, this project will spread awareness about the importance and opportunities in advanced manufacturing, train students in critically needed skills, and inspire talented young engineers to pursue careers in automation, control, and advanced manufacturing.

该学院早期职业发展(Career)奖的研究目标是为添加剂制造的高通量和高精度控制创造系统论设计和分析工具。尽管具有巨大的潜力，但由于工艺可靠性较差，添加剂制造尚未实现规模化生产。可靠性低的一个关键原因是缺乏准确的过程模型和开环过程控制。为了应对这一挑战，添加剂制造独特的逐层材料沉积机制将被依赖于基于传感器数据和部分模型信息的迭代精化的学习算法所利用。受多目标优化的启发，所提出的逐层学习控制算法将同时解决多个相互冲突的目标，以在存在建模误差和操作条件的不确定性的情况下提供可接受的零件几何和机械性能。该项目的成果包括(1)针对喷墨印刷和选择性金属熔化过程量身定做的多目标学习控制算法的设计和分析工具；(2)在实验试验台上验证所提出的算法；(3)在K-12水平上开发关于添加剂制造和自动化的动手实验和互动课程模块；以及(4)为自动化、制造和控制领域的本科生和研究生提供培训。该项目的成功完成将通过提高可靠性、重复性和生产率，对快速扩张的十亿美元添加剂制造业产生强大的积极影响。具体地说，提高基于聚合物的添加剂制造工艺的可靠性和性能可以对一次性大规模定制部件的制造产生重大影响，包括生物医学植入物、光机械部件和组织工程。另一方面，可靠的基于金属的添加剂制造工艺可以生产安全关键应用的部件，例如航空工业。通过拟议的外展和教育计划，该项目将传播对先进制造业的重要性和机会的认识，培训学生迫切需要的技能，并激励有才华的年轻工程师追求自动化、控制和先进制造业的职业生涯。