GOALI: Engineering-Driven Modeling of Multi-Resolution Data for Surface Variation Control

GOALI：用于表面变化控制的多分辨率数据的工程驱动建模

• 批准号: 1434411
• 负责人: Hui Wang
• 金额: $ 27.74万
• 依托单位: Florida State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2014
• 资助国家: 美国
• 起止时间: 2014-03-01 至 2017-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1434411&HistoricalAwards=false
• 关键词: GOALI; Engineering; Driven; Modeling; Multi

This GOALI project develops methodologies and algorithms to reduce surface shape variation, enabling cost-effective high-precision machining by integrating process physics with multi-resolution surface data for quality diagnosis and monitoring. The integration of the process physics model and multi-resolution data will improve surface data modeling, variation prediction accuracy, and interpretability. Based on the improved data model, a cost-effective surface measurement strategy and monitoring scheme will be developed by which abnormal process variations can be detected and identified with reduced false alarm rates. The model will also improve surface variation diagnosis. By combining the modeling and monitoring approaches, a two-level surface variation control methodology is established to improve surface quality for multistage machining processes. The PI and co-PIs will collaborate with major US automakers and/or powertrain system suppliers for data acquisition, experiments, algorithm verification and validation, and technology transfer.The outcome of this research provides new insights into surface variations in multistage machining processes and will transform in-plant quality control practice from dimensional variation reduction to surface shape variation control. If successful, this project will enhance US manufacturing competitiveness by improving powertrain machining precision and quality. The surface modeling and control framework established in this research focuses on face milling and can potentially benefit other surface machining applications. In addition, the methodology of surface variation control established in this research will contribute to development of a novel optical surface measurement system that can intelligently select metrology resolution and determine defective surface regions. The surface variation monitoring scheme can also be extended to quality control for micro manufacturing processes by reducing the measurement time of high-definition metrology such as atomic force microscopy and 3D profilometry.

该守门员项目开发了方法和算法，以减少表面形状变化，从而通过将过程物理与多分辨率的表面数据整合在一起，从而实现了具有成本效益的高精度加工，以进行质量诊断和监测。过程物理模型和多分辨率数据的集成将改善表面数据建模，变化预测准确性和解释性。基于改进的数据模型，将开发出具有成本效益的表面测量策略和监测方案，通过该策略，可以通过降低的错误警报率检测并确定异常过程变化。该模型还将改善表面变化诊断。通过结合建模和监测方法，建立了两级表面变化控制方法，以提高多阶段加工过程的表面质量。 PI和CO-PIS将与美国主要汽车制造商和/或动力总成系统供应商合作，以进行数据获取，实验，算法验证和验证以及技术转移。这项研究的结果提供了对多阶段加工过程中表面变化的新见解，并将从植入质量质量控制实践中从减小的变性转化表面形成对表面变量的变量。如果成功，该项目将通过提高加工精度和质量来提高美国制造竞争力。这项研究中建立的表面建模和控制框架集中在面部铣削上，并有可能使其他表面加工应用受益。此外，本研究中建立的表面变异控制方法将有助于开发新型的光学表面测量系统，该系统可以智能地选择计量学分辨率并确定缺陷的表面区域。表面变化监测方案也可以通过减少高清计量学的测量时间（例如原子力显微镜和3D修复计）来扩展到微型制造过程的质量控制。

项目成果

Surface Variation Modeling by Fusing Multiresolution Spatially Nonstationary Data Under a Transfer Learning Framework

• DOI: 10.1115/1.4041425
• 发表时间: 2018-10
• 期刊: Journal of Manufacturing Science and Engineering
• 作者: Jie Ren;Hui Wang

Iterative multi-task learning for time-series modeling of solar panel PV outputs

• DOI: 10.1016/j.apenergy.2017.12.058
• 发表时间: 2018-02
• 期刊: Applied Energy
• 影响因子: 11.2
• 作者: Tahasin Shireen;Chenhui Shao;Hui Wang;Jingjing Li;Xi Zhang;Mingyang Li