This project application devoted to the CFD simulation of metal fusion in welding and additive manufacturing with electric arc and laser is in the continuation of the SNIC projects 2021/5-584 (Medium Compute) and 2021/6-337 (Medium Storage) that addressed metal fusion with a gas metal arc (GMA).
In earlier SNIC projects, we developed and applied a model for free surface thermal flow including solid, liquid, and gas phases, thermocapillary force, electromagnetic force (EMF), and metal transfer. We improved it step by step. In the projects 2020/5-674 and 2020/6-274, different assumptions made for the EMF model were relaxed. This work is now published in open access (DOI 10.1016/j.ijheatmasstransfer.2022.123068). As the resultant EMF model still assumed a frozen-free surface, we removed this simplification and investigated the consequences during the projects run in 2022. In addition, we extended the work to pulsated arcs and investigated different pulsation models. Two journal manuscripts are currently in preparation to publish these results. Jointly with the two articles published in earlier SNIC projects, they will contribute to a Ph.D. thesis. Furthermore, a preliminary study aiming at extending the CFD model to powder bed fusion AM with a laser was also conducted.
The first objective of the proposed project is to assess the improved model for GMA metal fusion in the context of a real welding application since most of the test cases made in 2021-222 were so-called academic. In that intent, GMA welding for a V-groove joint at various substrate orientations (uphill, downhill, horizontal, and sideways) will be studied. Computations made with both the initial model (from 2020) and the improved model will be comparatively analyzed to check whether weld defects can now be better predicted. Validation by comparison to experimental measurements already made at our facilities will also be performed, and a journal manuscript will be prepared.
The second objective is to investigate the thermal flow field produced by an Adjustable Core & Ring Mode laser. This is a beam-shaped laser that was observed to produce fewer defects than a standard laser. However, selecting the power density distribution is challenging. Numerical simulation will provide process understanding that complements experimental observation and supports process control and development. We thus aim to perform simulations with five different beam power distributions in the core and ring and present the results at a conference with a peer-reviewed publication.
The proposed study will also include preliminary parametric and mesh studies to prepare for the above-mentioned GMA and laser test cases. Based on the former projects, it is evaluated that to conduct this study, in total, a minimum computational time of 80 000 core hours per month and storage for at least 8250 GiB and 6 000 000 files for a duration of 1-year will be needed.