Case Studies, Metals and Glass
building icon

The Research Group

Navy STTR
#N6833518C0706 PI: Aaron Stebner (Mines)

NAVSEA
Colorado School of MINES
ADAPT
question mark icon

The Problem

  • Accelerate qualification of additively manufactured components
  • Create an ML model that can screen out components based on build parameters, as likely to have large pores
gear process icon

The Process

  • Analyze porosity in additively manufactured (AM)-printed IN718 components produced with varying build parameters (including laser orientation and power density.
  • The team identified a bimodal distribution in this data, with one cluster of components more likely than the other to fail max pore size qualification standards.
bullseye

The Outcome

  • A machine learning (ML) model that can predict likelihood of falling on pore size for a given set of build parameters.
log max pore diameter v median pore diameter
Download this case study as a PDF