Register now After registration you will be able to apply for this opportunity online.
Parallel computing: MPI-Based Parallelization of Laser Fusion Simulation with SPH Framework
The Advanced Manufacturing Lab (am|z) is excited to announce a thesis opportunity focusing on the development of a highly parallelizable modeling framework for additive manufacturing (AM) processes, particularly laser powder bed fusion (LPBF). Our research primarily delves into advancing manufacturing techniques, with a special emphasis on additive manufacturing. We have developed a robust numerical simulation framework called iMFREE utilizing Smoothed Particle Hydrodynamics (SPH) for multi-physics applications like LPBF. However, there is a need to enhance computational efficiency, specifically through parallelization via Message Passing Interface (MPI). This project offers an excellent chance for students to deepen their knowledge in parallel computation while working hands-on with a mature computational framework.
LPBF is a promising AM process for fabricating complex shapes, where layers of material powder are selectively melted to build up a part. Our in-house simulation framework, iMFREE, originally designed for machining processes, has been adapted for additive manufacturing, particularly LPBF. While particle simulations are ideal for material transport and phase change problems like LPBF, they suffer from poor computational efficiency. Parallelization, especially through MPI, offers a solution to this challenge by distributing computations across multiple processing units.
LPBF is a promising AM process for fabricating complex shapes, where layers of material powder are selectively melted to build up a part. Our in-house simulation framework, iMFREE, originally designed for machining processes, has been adapted for additive manufacturing, particularly LPBF. While particle simulations are ideal for material transport and phase change problems like LPBF, they suffer from poor computational efficiency. Parallelization, especially through MPI, offers a solution to this challenge by distributing computations across multiple processing units.
This project aims to develop an MPI implementation of the existing SPH simulation framework with the following objectives:
• Development of an MPI framework for SPH
• Testing, optimization, and validation of the parallel implementation
• Investigation of runtime and speed-up factors under different settings
This project aims to develop an MPI implementation of the existing SPH simulation framework with the following objectives: • Development of an MPI framework for SPH • Testing, optimization, and validation of the parallel implementation • Investigation of runtime and speed-up factors under different settings