Browsing by Subject "temperature field"

Now showing 1 - 3 of 3

Numerical Simulation of Temperature Fields in Powder Bed Fusion Process by Using Hybrid Heat Source Model
(University of Texas at Austin, 2017) Luo, Zhibo; Zhao, Yaoyao Fiona
Powder bed fusion (PBF) process is capable of producing a complex geometrical part with less material and energy consumption compared with conventional manufacturing methods. The performance of PBF processed part is mainly controlled by many process parameters such as scanning speed, scanning pattern, scanning strategy, and layer thickness. Usually, these parameters are optimized through detailed experiments which are time-consuming and costly. Therefore, numerical methods have been widely adopted to investigate the effects of these process parameters on temperature fields and thermal stress fields. As the laser/electron beam introduces huge temperature gradients within the irradiated region, which will result in the distortion even delamination of solidified layers, the study of the history of temperature distribution is the basic and crucial step in the modeling of PBF process. Most of the current research utilizes moving Gaussian point heat source as heat input to model the temperature distribution of a part. However, due to the small diameter of laser/electron beam, a small enough time step size is required to accurately model the real heat input, which will lead to significant computational burden. In this research, a hybrid of moving Gaussian point and line heat source model is developed, which makes the modeling of PBF process efficient without losing too much accuracy. In addition, an adaptive mesh scheme, which is capable of dynamically refining the mesh near the beam spot and coarsening the mesh far away from the beam spot, is adopted to accelerate the simulation process. Specifically, moving Gaussian point heat source is applied to the region of interest where accuracy is more concerned such as the temperature field within overhang feature. While the line heat source is applied to the region of interest where efficiency is more concerned such as temperature field within the inner region of a square. The simulation result shows that the temperature fields by using hybrid source model are comparable to the temperature fields by using the moving Gaussian point heat source model, and much less central processing unit time is required when the hybrid heat source is applied.
Numerical Thermal Analysis in Electron Beam Additive Manufacturing with Preheating Effects
(University of Texas at Austin, 2012-08-16) Shen, Ninggang; Chou, Kevin
In an early study, a thermal model has been developed, using finite element simulations, to study the temperature field and response in the electron beam additive manufacturing (EBAM) process, with an ability to simulate single pass scanning only. In this study, an investigation was focused on the initial thermal conditions, redesigned to analyze a critical substrate thickness, above which the preheating temperature penetration will not be affected. Extended studies are also conducted on more complex process configurations, such as multi-layer raster scanning, which are close to actual operations, for more accurate representations of the transient thermal phenomenon.
Predicting Temperature Field for Metal Additive Manufacturing using PINN
(University of Texas at Austin, 2023) Peng, B.; Panesar, A.
Machine-learning-based methods are gaining traction as an alternative to numerical methods in many engineering applications. Physics-informed neural network (PINN), a self-supervised method, is particularly attractive with its unique capability of guiding the training with physical laws written in the forms of partial differential equations. Thermomechanical simulation for additive manufacturing (AM), a multi-scale, multi-physics problem could potentially benefit from the use of PINN, as demonstrated in some successful attempts in the literature. In this work, PINN is applied to different metal AM processes and several challenges that limit the robustness of PINN are observed. This paper aims to provide a summary of the observations and a preliminary attempt to account for such observations in order to pave the path for future work that aims to unleash the full promise of PINN in AM-related applications.