Dr. Prahalada Rao
Dr. Kevin Cole
Dr. Carl Nelson
Date of this Version
Kobir, M.H., 2021, “Thermomechanical Modeling in Laser Powder Bed Fusion Additive Manufacturing using Graph Theory: Application to Prediction of Recoater Crash,” M.S. thesis, University of Nebraska-Lincoln.
This work pertains to the laser powder bed fusion (LPBF) additive manufacturing process. The objective of this thesis is to predict a frequently occurring type of thermal-induced process failure in LPBF called recoater crash. To ascertain the likelihood of a recoater crash before the part is printed, we develop and apply a computationally efficient thermomechanical modeling approach based on graph theory.
Despite its demonstrated ability to overcome the design and processing constraints of conventional subtractive and formative manufacturing, the production-level scaleup of LPBF is hindered by frequent build failures. For example, the part often deforms as it is being printed due to uneven heating and cooling. This thermal-induced deformation of the LPBF part during processing causes it to interfere with the deposition mechanism (recoater) leading to a common build failure called recoater crash. A recoater crash not only destroys the part involved but also causes an entire build to be abandoned resulting in considerable time and material losses.
In this context, fast and accurate thermomechanical simulations are valuable for practitioners to identify and correct problems in the part design and processing conditions that can lead to a recoater crash before the part is even printed. Herein, we propose a novel thermomechanical modeling approach to predict recoater crashes which is based on two sequential steps. First, the temperature distribution of the part during printing is predicted using a meshfree graph theory-based computational thermal model. Second, the temperature distribution is used as an input into a finite element model to predict recoater crashes. The accuracy and computational efficiency of this graph theory-based approach is demonstrated in comparison with both non-proprietary thermomechanical finite element analysis (Abaqus), and a proprietary LPBF simulation software (Netfabb). Based on numerical (verification) and experimental (validation) studies, the proposed approach is 5 to 6 times faster than the non-proprietary finite element modeling and has the same order of speed as Netfabb. This physics-based approach to prevent recoater crashes can engender substantial savings by supplanting existing build-and-test optimizations of part design and parameters.
Advisor: Prahalada Rao