Session: 16-01-01: Government Agency Student Poster Competition

Paper Number: 150742

150742 - Cadmap: Creating Mapped Solid Models of Deformed As-Manufactured Geometries That Link to an Original Reference Design 

The accurate, precise, and comprehensive representation of the shape details of manufactured components is critical for the design, simulation, control, and optimization of next-generation cyber-based manufacturing systems. However, in many cases, the design intent encoded in the computer aided design (CAD) model cannot be applied directly to the as-manufactured part because the physical part never perfectly matches the geometry of the designed part. Therefore, a systematic bi-directional mapping framework between as-designed models and as-manufactured geometries is needed to connect design and production. For example, we can optimize an additive manufacturing process with the mapping that shows how the design has been distorted during manufacturing. Similarly, we can create digital twins that fuse manufacturing and inspection data from different sources and serial numbers across the product lifecycle to optimize maintenance and value to the end user.

We propose to address these challenges by breaking the problem into two sub-problems: 1) matching landmarks, key points, and/or coarse shapes between the original and as-manufactured geometries and 2) retaining the CAD topology such as continuity and connectivity, especially for relatively complex geometries and boundary representations of topologically connected trimmed non-uniform rational B-splines (NURBS) surfaces, in the fitted representation. Solving these sub-problems will enable the definition of mappings between the original design and the as-manufactured part.

This project aligns with two of the NSF's ten big ideas: “Harnessing the Data Revolution” and “Growing Convergence Research” and will facilitate the digital integration of design and manufacturing within larger product lifecycle ecosystems. By building a closer connection between the as-designed and as-manufactured geometries, the proposed research has potential implications that include data fusion and integration from different resources and modalities, geometric distortion and shrinkage quantification in a manufacturing process to enable process optimization, and engineering analysis of the as-manufactured CAD model, to develop a digital twin of the manufactured part. In addition, our work will facilitate centralizing all kinds of information about a part, including multiple design revisions, finite element results, etc. For example, in current technology, selecting boundary regions for finite element boundary conditions is a tricky manual process. Our persistent mappings offer the potential to identify such regions once and reuse that definition on an updated design. Our work will transform manufacturing by making it possible to routinely connect all sources of information from design office to the factory floor or maintenance hangar.  

We test our framework with a simple bar geometry as proof-of-concept and extend it to arbitrarily trimmed NURBS models with a teapot example. We obtain the point cloud of the physical part using XCT/3D laser scanner and fit the authority CAD model to the actual part shape with consistent CAD topology and seamless surface connection. In the experiments results, we demonstrated that our framework effectively converts the as-designed CAD model to the actual shape without topology change and harmful gaps.

Presenting Author: Lijie Liu Iowa State University

Presenting Author Biography: Lijie Liu is currently a fourth-year Ph.D. student at Iowa State University majoring in Industrial Engineering. He obtained a master's degree in Statistics at the University of Wisconsin-Madison in 2019. He have a solid background in statistics, operational research, machine learning, and mathematics, with strong programming skills in Python, R, and Matlab. His main research work at ISU is mapping the inconsistent geometry between the product design and final manufactured parts, which was funded by the Center for Nondestructive Evaluation (CNDE) at ISU and recently got new funding from the National Science Foundation (NSF) for further research. Besides his main work, he also collaborated with other departments like Agricultural and Biosystems Engineering and Psychology to do data analysis in interdisciplinary fields, leveraging his solid knowledge in statistics, machine learning, and programming. He received 2022 and 2023 R.B. Thompson Graduate Fellowship for his outstanding research outcome in CNDE. He actively presented his work and built professional connections at IISE annual conference, ASNT research symposium, and IMSE research symposium several times. He also contributed significantly to the ideas, preliminary work, and writing of successful NSF proposals and internal proposals. He has an internship at MxD to bring his research to benefit the industry. Besides excellent record in research and courses, he mentored three undergraduate students in two semesters and showed strong leadership and collaboration skills.

Authors:

Lijie Liu Iowa State University
Stephen Holland Iowa State University
Adarsh Krishnamurthy Iowa State University
Qing Li Iowa State University