| About this Abstract |
| Meeting |
MS&T24: Materials Science & Technology
|
| Symposium
|
Computational Materials for Qualification and Certification
|
| Presentation Title |
Fast, Cheap & In Control: Application of Surrogate Models to Explore Microstructure-Properties Relationships for AM-Based Materials |
| Author(s) |
Vanessa Oklejas, Ross Gregoriev, Kyle Rosenow, Scott Cochran |
| On-Site Speaker (Planned) |
Vanessa Oklejas |
| Abstract Scope |
The advantages attributed to additive manufacturing (flexibility and portability) are also accompanied by substantial drawbacks (decreased reliability and reproducibility) that, so far, have precluded widespread adoption of AM technology. It’s widely believed that the development of methods to understand process-structure-properties relationships, resulting from the highly variable processing conditions and complex multi-physics, will enable standardization (characterized by the twin hallmarks qualification and certification) of AM technology.
This work will present the development of computationally inexpensive surrogate models and their application toward unraveling structure-properties relationships in Nickel-based superalloys. This work will examine the efficiency (i.e., amount of required training data) with which robust surrogate models can be developed, as well as the uncertainty associated with surrogate model predictions. Further application of surrogate models toward the development of process-structure relationships, as well as methods for chaining multiple surrogate models together, will be discussed.
Copyright 2024 Lockheed Martin Corporation |