LASER WELDING
OPEN-SOURCE SIMULATION OF ADVANCED LASER WELDING
SEAN ROCHE-WATSON ET AL*, UNIVERSITY OF MANCHESTER
Figure 1: Path of a single ray into a capillary caused during laser welding, the ray can be seen to reduce in energy as it reflects off the alpha phase surface geometry , as it changes from red to blue.
FROM ROSENTHAL TO REALITY
Mathematical modelling of welding processes has evolved greatly since Daniel Rosenthal predicted thermal history in the 1930s by assuming a moving point heat source on a semi- infinite plate [1]. By the late 1970s Mazumder and Steen had shown a numerical finite difference method for a 3D heat transfer model of a gaussian laser heat source [2].
Today, numerical simulation has matured into two primary methods. The first, based on thermo- metallurgical and mechanical studies, often relies on the standard Goldak double ellipsoid heat source dominated by conduction (used in e.g. Ansys Mechanical, SYSWELD & Simufact). The second uses studies from integrating energy transfer and fluid dynamics in a multiphysics environment. Ray tracing is integrated to express laser energy absorption by a material, the vaporisation that creates the keyhole, and the hydrodynamics of liquid and vapor flows (used in e.g. Ansys Fluent, COMSOL Multiphysics and bespoke software FLOW-3D WELD) [3]. The latter are the modern answer to Rosenthal’s limitations.
26 | LASER USER 119 MARCH 2026
Figure 2: ‘laserbeamFoam’ solver simulation and X-ray imaging of laser welding in Ti–6Al–4V. (a–f) Simulated melt evolution of transition from conduction to keyhole mode and deepening of the vapour depression (scale bar 100 µm) [6] . (A–H) Corresponding experimental frames from Cunningham et al. [8].
LIGHTING THE WAY
Multiphysics packages make hidden phenomena visible, such as internal melt pool velocity patterns and pressure distributions that are practically impossible to capture experimentally. These predictive models serve as the foundation for digital twins, enabling virtual production planning that replaces costly, time-consuming experimental methods with streamlined, data- driven optimisation [4]. Commercial packages can be costly for researchers and use proprietary software, meaning users cannot inspect the underlying mathematical models, making them less flexible for high-level research.
OPEN-SOURCE ALTERNATIVE: LASERBEAMFOAM
As an alternative, LaserbeamFoam is an open-source collaboration between University of Manchester, University College Dublin, SINTEF Industry and Universität Stuttgart
to simulate laser welding, laser drilling, selective laser melting and diode array AM. It offers additional flexibility for users to manipulate underlying parameters, compared with commonly available software. Beyond this transparency, the solver offers uncapped scalability; parallel processing is limited by available hardware rather than restrictive licence fees, meaning users can increase simulation scale and reduce run times at no extra software cost. The platform can be customised and integrated with the wider
OpenFOAM.com ecosystem for advanced meshing and post- processing.
The ‘laser melting suite’ includes two main OpenFOAM solvers, ‘laserbeamFoam’ and ‘compressiblelaserbeamFoam’. The original ‘laserbeamFoam’ solver uses a Volume of Fluid method; it treats the metallic substrate as incompressible and fully describes the fusion and melting state. The topology of the
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32
