FMSG: Cyber: Process Monitoring Methods for the Product Quality Improvement of Electron Beam Powder Bed Fusion Additive Manufacturing Processes

增材制造是一种计算机控制的过程，通过逐层沉积材料来制造具有复杂结构的三维物体。在众多增材制造工艺中，电子束粉末床熔融（EB-PBF）在金属零件的生产中显示出了巨大的前景。阻碍EB-PBF和许多其他增材制造工艺被广泛采用的最重要的障碍之一是缺乏有效的产品质量控制方法。因此，EB-PBF机器制造质量不符合规格的次品的概率非常高。为了应对这一挑战，该项目专注于研究，以促进建立EB-PBF增材制造过程的原位质量控制框架的可能性。该框架使用实时传感数据监控产品质量，并在必要时自适应调整过程控制参数以提高质量。本研究探索了EB-PBF的新能力，以显著提高制造零件和制造过程的质量。该项目还包括各种教育和推广组成部分，为吸引和培训下一代制造业专业人员奠定基础。本项目旨在为EB-PBF增材制造工艺的现场质量控制提供依据。主要有三个研究方向。首先是一种计算效率高的运行时缺陷检测方法，该方法使用一种新的统计学习模型，该模型建立在实时电子数据基础上，用于检测制造过程中可能对产品质量产生不利影响的缺陷。第二种是保护隐私的联邦模型，用于制造社区参与根本原因缺陷诊断；它建立在一个新的基于正则化的统计学习模型上，特别是EB-PBF，以识别导致产品缺陷的关键过程控制参数。第三部分着重于设计和实施实验来验证这些模型的有效性。自动缺陷检测和根本原因诊断方法不仅是EB-PBF现场质量控制的基础，而且显著丰富了统计学习、数据分析和机器学习社区的知识库。该未来制造奖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）以及工程理事会（ENG）的土木、机械和制造创新司（CMMI）支持。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。

Additive manufacturing is a computer-controlled process that creates three-dimensional objects with complex structures by depositing materials layer-by-layer. Among many additive manufacturing processes, Electron Beam Powder Bed Fusion (EB-PBF) has shown its great promise in the production of metallic parts. One of the most significant barriers that prevents the EB-PBF and many other additive manufacturing processes from becoming more widely adopted is the lack of effective product quality control methods. Consequently, EB-PBF machines have a very high probability of fabricating defective products whose quality fails to meet specifications. To address this challenge, this project focuses on research to facilitate the possibility of establishing an in-situ quality control framework for EB-PBF additive manufacturing processes. The framework uses real-time sensing data to monitor the quality of products and adaptively adjusts process control parameters for quality improvement whenever necessary. This research explores new capabilities for EB-PBF to significantly improve the quality of manufactured parts and manufacturing processes. The project also has a variety of educational and outreach components that form a basis for attracting and training the next generation of manufacturing professionals.The objective of this project is to provide a basis for in-situ quality control of EB-PBF additive manufacturing processes. There are three primary research directions. The first is a computationally-efficient run-time defect detection method using a novel statistical learning model that builds on real-time electron-based data for detecting defects during manufacturing that would adversely affect product quality. The second is a privacy-preserving federated model to engage the manufacturing community on root-cause defect diagnostics; it is constructed on a new regularization-based statistical learning model, particularly for EB-PBF, to identify crucial process control parameters responsible for product defects. The third focuses on designing and conducting experiments to validate the effectiveness of these models. The automatic defect detection and root-cause diagnostics methods not only serve as the fundamental basis of the in-situ quality control of EB-PBF, but also significantly enrich the knowledge base of the statistical learning, data analytics, and machine learning communities.This Future Manufacturing award is supported by the Division of Computer and Network Systems (CNS) of the Directorate for Computer and Information Science and Engineering (CISE), and by the Division of Civil, Mechanical and Manufacturing Innovation (CMMI) of 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.