Excellence in Research: A Cyber-Physical System Framework for In-process Quality Assurance of Inkjet-based Additive Manufacturing

这项历史悠久的黑人学院和大学-卓越研究(HBCU-EIR)赠款支持为添加剂制造提供与质量保证相关的新知识的研究，促进科学进步和国家繁荣。喷墨打印是一种典型的添加剂制造工艺，它是基于热或声波形成液滴并通过喷嘴喷射出来的。它的巨大前景已在电子、能源、医疗保健和生物医药行业得到展示。然而，喷墨打印对环境、材料、机械和电子等因素非常敏感，工艺容易偏离理想的工作状态，导致零件出现缺陷。这往往会导致材料和能源的浪费，并影响许多重要工程系统的结构健康和功能完整性。该奖项支持基础性研究，为开发包括喷墨打印质量保证的神经网络在内的整体框架提供所需的知识。该项目有望显著提高喷墨添加剂制造工艺的生产率、质量和材料效率，从而造福美国经济和社会。这项研究综合运用了制造、计算机视觉、控制理论和机器学习等多学科的方法，帮助代表不足的群体更广泛地参与研究并促进工程教育。本项目的目标是建立一个无缝集成过程中基于视频的监控与闭环控制和补偿的综合框架，以有效检测并随后纠正高质量喷墨打印的过程漂移和异常。该框架由三个基于神经网络的协同数字双胞胎组成，神经网络是一种模仿人工智能领域人脑操作的技术。第一个数字双胞胎旨在对微滴的运动学和形态状态进行闭环控制。第二个数字双胞胎侧重于对印刷图案的几何和形态状态进行闭环控制。第三个数字双胞胎确定并实施对有缺陷的图案的补偿策略。具体目标是1)确定创建和集成数字双胞胎的方法，以保持理想的液滴状态，获得所需的图案并实施有效的补偿；2)得出在质量保证中使用神经网络的实用指南，包括输入选择和准备、网络设计和优化、输出选择和使用以及可转移性和适应性；3)从神经网络的角度了解喷墨打印中材料特性、控制变量、工艺参数和打印结果之间的关系。该项目有望为添加剂制造中的计算机物理系统的设计、开发和实施提供基本的理解。开发的框架可以适用于其他宏观和微观规模的添加剂制造工艺。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This Historically Black Colleges and Universities - Excellence in Research (HBCU-EiR) grant supports research that contributes new knowledge related to quality assurance for additive manufacturing, promoting both the progress of science and the advancement of national prosperity. Inkjet printing is one representative additive manufacturing process based on thermal or acoustic formation and ejection of liquid droplets through a nozzle. Its great promise has been demonstrated in electronics, energy, healthcare and biomedical industries. However, inkjet printing is sensitive to environmental, material, mechanical and electronical factors, and the process can easily deviate from the desirable working status, resulting in defective parts. This tends to lead to material and energy waste and affects the structural health and functional integrity of many important engineering systems. This award supports fundamental research to provide needed knowledge for the development of a holistic framework involving neural networks for quality assurance in inkjet printing. This project holds the potential to significantly improve productivity, quality and material efficiency for inkjet-based additive manufacturing processes, thus benefiting the U.S. economy and society. Using a multi-disciplinary approach involving manufacturing, computer vision, control theory, and machine learning, this research helps broaden participation of underrepresented groups in research and promotes engineering education.The goal of this project is to establish a comprehensive framework that seamlessly integrates in-process video-based monitoring with closed-loop control and compensation to effectively detect and subsequently correct the process drift and anomalies toward high-quality inkjet printing. The framework consists of three synergic digital twins based on neural networks, a technique that mimics the operations of a human brain in the artificial intelligence field. The first digital twin aims at closed-loop control of the kinematic and morphological status of the micro droplets. The second digital twin focuses on closed-loop control of the geometrical and morphological status of the printed patterns. The third digital twin determines and implements compensation strategies for defective patterns. Specific objectives are to 1) identify methodology for creation and integration of digital twins to maintain desirable droplet status, obtain required patterns and implement effective compensation, 2) derive practical guidelines of using neural network in quality assurance, including input selection and preparation, network design and optimization, output selection and usage, and transferability and adaptability, and 3) understand the relationship between material properties, control variables, in-process parameters and print outcome in inkjet printing from the perspectives of neural network. This project is expected to provide fundamental understanding of the design, development, and implementation of cyber-physical systems in additive manufacturing. The developed framework can be adapted to other macro- and micro-scale additive manufacturing processes.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.