|About this Abstract
||2019 TMS Annual Meeting & Exhibition
||Computational Approaches for Big Data, Artificial Intelligence and Uncertainty Quantification in Computational Materials Science
||Comprehensive Quality Assurance of Additive Manufacturing Ti-6Al-4V by Learning from Prior Studies
||Sen Liu, Branden Kappes, Aaron Stebner, Xiaoli Zhang
|On-Site Speaker (Planned)
In the past decades, efforts in metal additive manufacturing (AM) have produced mounts of experiments or simulation data. Because the extent of human intuition and limited number of publications a human can read, it is hard for researchers to follow up to have a comprehensive learning, which easily causes repeated studies with time/resource costs. In this study, an automated framework was established for systematically gather prior publication knowledge and effectively extract information to accelerate the multi-properties optimization process. The collected database contains objects (e.g. machine, Ti-6Al-4V), feature space (e.g. laser power, scan speed, hatch spacing etc.) and properties (e.g. density, hardness etc.). A Na´ve Bayesian Network is adopted to explore the complex correlations. A multi-properties recommendation method based on Genetic Algorithm was applied to learned models for comprehensive quality assurance. A set of experiments were performed to evaluate its feasibility and reliability in acceleration multi-properties processing optimization of metal AM.
||Planned: Supplemental Proceedings volume