Affiliation:
1. University of Connecticut, Storrs, Connecticut, United States
2. Amazon.com, Tempe, Arizona, United States
Abstract
Abstract
The capability of metal additive manufacturing (AM) to produce parts with complex geometries manifests its potential to revolutionize manufacturing. However, the presence of heterogeneous defects even under optimized part design and processing conditions imposes critical barriers towards scaling metal AM to production environments. The recent advancement in imaging technology leads to a data-rich environment in AM and provides a unique opportunity to enhance understanding of design-quality interactions. However, the state-of-the-art image-guided methodologies focus on the characterization of the individual defect and are less concerned about transferring knowledge across heterogeneous defects to improve modeling and prediction ability. This study introduces a novel data-driven methodology to simultaneously quantify interactions between design parameters (e.g., orientation, thickness, and height) and heterogeneous defects (e.g., geometry distortion, discontinuity, and porosity) in the metallic AM build. A real-world case study on complex thin wall structures is evaluated with the single-response characterization models. Experimental results show average improvement of 64.8%, 70.4%, 73.3% in RMSE in comparison with single response model with inducing percentage u = 0.25, 0.5, and 0.75, respectively. The proposed model enables providing practitioners with a guideline to understand multimodal defects interrelation and fabrication of thin walls with minimal defects.
Publisher
American Society of Mechanical Engineers
Cited by
3 articles.
订阅此论文施引文献
订阅此论文施引文献,注册后可以免费订阅5篇论文的施引文献,订阅后可以查看论文全部施引文献
1. Brain-inspired computing for in-process melt pool characterization in additive manufacturing;CIRP Journal of Manufacturing Science and Technology;2023-04
2. Feature Engineering in Additive Manufacturing;Engineering of Additive Manufacturing Features for Data-Driven Solutions;2023
3. Applications in Data-Driven Additive Manufacturing;Engineering of Additive Manufacturing Features for Data-Driven Solutions;2023