3D Spatial Reconstruction of Thermal Characteristics in Directed Energy Deposition through Optical Thermal Imaging

Open Access
Author:
Kriczky, Dennis Andrew
Graduate Program:
Mechanical Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
May 12, 2014
Committee Members:
  • Panagiotis Michaleris, Thesis Advisor
  • Edward William Reutzel, Thesis Advisor
Keywords:
  • thermal metrics
  • thermal imaging
  • directed energy deposition
  • 3D spatial reconstruction
  • Ti-6Al-4V
Abstract:
An application to visualize thermal metrics extracted from coaxial thermal images in three dimensions during directed energy deposition is presented in this thesis. The extraction of thermal metrics is useful for correlation of microstructure for process monitoring and control of additive manufacturing. The thermal metrics attained from the coaxial images include the thermal gradient at the solidus to liquidus region, the maximum temperature in the melt pool, the melt pool pixel area, and the length-to-width ratio of the melt pool. The current procedure for part qualification in additive manufacturing is through destructive methods. The use of thermal metrics in a 3D spatial reconstruction allows for a non-destructive means to distinguish material microstructure. For this reconstruction, two Ti-6Al-4V L-shaped parts were deposited with a 1-bead wide deposition on one leg of the build and 3-bead wide deposition on the second leg of the build. A filtering scheme of the coaxial thermal images is utilized to produce melt pools with distinguishable solidus to liquidus regions. The acquisition of laser location during deposition is used to create a three dimensional representation of the calculated thermal metrics. Differences in thermal metric values between separate legs of the L-shaped parts express changes in the thermal history and hence the microstructure development for the transient and steady state regions of melt pool movement. For process monitoring, the cross sectional cuts of the three dimensional representation of thermal metrics can correlate to variations in material microstructure from the cross sectional cuts of actual L-shaped builds.