High-throughput Screening and Quality Assessment Testing and an AI-based In-Situ Process Monitoring and Quality Control Framework for Additive Manufacturing New Builds and Repaired Parts
Abstract
This research aims to advance the state-of-the-art in-situ process monitoring and quality control technology by developing a high-throughput screening and quality assessment testing framework for new and repaired builds and parts (powder bed) pertinent to turbine engine components. Recent interest in AM technology is continually growing in many industries such as power generation, aerospace, automotive, and bio-medical, but components often result in highly variable performance. Determining optimal process parameters for the build and repair process can be an extensive and costly endeavor due to a limited knowledge base, proprietary restrictions, heterogeneity of energy sources, and scanning strategies, among other variables affecting the quality of the builds. Quantifying the significant variability of the material properties and performance in AM components is challenging as there are many underlying causes, including machine-to-machine differences, recycles of powder, balling effects, poor weldability, and erroneous process parameter selection. Therefore, there have been significant advancements in AI-based AM process modeling and predictive tools for online monitoring and controlling material properties and microstructural defects. Many relevant training data sets are critically needed to accomplish the high-fidelity characteristics of these modeling tools. These data can be obtained via conventional material evaluations and assessment tests such as tensile and fatigue. Some of these tests, for example, fatigue, are time-consuming to test for the desired composition and performance requirements. Furthermore, these AM components must be qualified at a Òbuild specificÓ level, necessitating rapid testing and high-fidelity models for design qualifications. For that, in this project, an integrated high-throughput screening and quality assessment testing framework will be developed for determining the optimal process parameters using the specimens made from additive manufacturing builds and repaired parts. Furthermore, the data characteristics and quality features accumulated from the assessment testing will be used to construct novel deep-learning driven in-situ process monitoring and quality prediction tools named Predictive Digital Twins. The outcomes of this work will engender a new comprehensive regulatory approval adaptive AM process monitoring and control framework to produce quality parts consistently with expected performances for new and repaired AM parts in DoD applications.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 26, 2023
- Source ID
- W911NF2310148
Entities
People
- Mo-how Shen
Organizations
- Army Contracting Command
- Office of the Secretary of Defense
- University of North Texas