Improving welding joints with fluid dynamics simulations - Dr. Marcin Serdeczny

Creating strong and repeatable joints at high speed with laser welding is a task of utmost importance in today’s engineering world. Ever-increasing requirements from the automotive industry push this technology to its maximum. Laser welding is a key step in manufacturing hairpin motors, which have several advantages over conventional winding motors when used in electric vehicles (EV). Hairpin technology relies on winding the electric motor with rectangular wires rather than traditional round wires. This enables denser and more deterministic winding which translates to higher power density, as well as easier automation of the production process. All in all, it brings us closer to creating less expensive and more powerful electric vehicles.

The challenge in creating hairpin motors is reliable melting and joining of hairpins with lasers. From the moment when laser irradiation is absorbed at the copper bar surface, the welding process begins. The two adjacent hairpins are heated above the melting point and start to flow, creating a melt pool that is very sensitive to the amount of absorbed laser radiation, and which solidifies to create a joint after the laser shuts off. All of this happens in fractions of milliseconds and is repeated thousands of times for a single electric motor.

Computational fluid dynamics (CFD) simulations with FLOW-3D WELD let engineers look at this process at high spatial and temporal resolution, thanks to its high fidelity multiphysics solver. FLOW-3D WELD numerically solves the conservation equations of fluid flow, allowing the user to predict the shape and size of the melt pool during and after welding. This virtual test bench of the process enables optimization of the welding parameters and identification of the sources of defects and variations in the joint.

Hairpin welding is only one of many examples where FLOW-3D WELD can be used to improve the laser welding process. Another prominent example is an optimization of the laser heat input during the welding of EV battery cells and cases, where joints are required to have minimal electrical resistivity and good mechanical strength, while too much laser power can lead to over-penetration and damaging of the cell. In fact, many challenges in laser welding boil down to fluid dynamics of the melt pool that can be predicted with FLOW-3D WELD. These are keyhole formation, collapse, and trapping of pores, instability and spatter, or surface roughness formation. Thanks to advances in computer science, parameters of laser welding no longer need to be guessed at but can be engineered using physics-driven simulations.

During this year’s MCWASP conference, we will be focusing on another interesting application of FLOW-3D WELD, which is prediction of alloys mixing during dissimilar materials welding. Laser welding of dissimilar materials is widely applied across the industry and is challenging due to differences in thermophysical properties of alloys. Moreover, the history and intensity of mixing alloys drives the formation of brittle intermetallic phases that degrade the quality of the joint. With FLOW-3D WELD, we can optimize the process parameters to minimize the dissimilar materials mixing, while ensuring good weld penetration.

By Dr. Marcin Serdeczny