Date of Award

3-2020

Document Type

Thesis

Degree Name

Master of Science

Department

Department of Engineering Physics

First Advisor

Nicholas Herr, PhD

Abstract

A process was developed to identify potential defects in previous layers of Selective Laser Melting (SLM) Powder Bed Fusion (PBF) 3D printed metal parts using a mid-IR thermal camera to track infrared 3.8-4 m band emission over time as the part cooled to ambient temperature. Efforts focused on identifying anomalies in thermal conduction. To simplify the approach and reduce the need for significant computation, no attempts were made to calibrate measured intensity, extract surface temperature, apply machine learning, or compare measured cool-down behavior to computer model predictions. Raw intensity cool-down curves were fit to a simplified functional form designed to approximate the behavior of radiative, conductive, and convective heat transfer, yielding a conductivity parameter. This parameter was able to identify minor changes, less than twenty five percent, in the thickness of a single layer of the walls of single-pass, thin-wall structures and the presence of underlying unfused powder. Small voids were not produced in the test prints analyzed here. The thermal signature of underlying fusion defects, as measured by the conductivity parameter, is clearly present for many layers. This method was shown to be effective in detecting changes in conductivity of the material which represents a large portion of the defects commonly found in SLM additive manufacturing.

AFIT Designator

AFIT-ENP-MS-20-M-080

DTIC Accession Number

AD1106235

Download

Share

COinS