Computational Modelling of Friction Spinning for Hot Forming of Metals
Please login to view abstract download link
Similar to friction welding, friction spinning is a solid-state process. But friction spinning is designed for hot metal forming to shape a workpiece by rapidly rotating it against a blunt tool. The high velocity at the contact interface generates frictional heat, which conducts into the workpiece and causes thermal softening. This softening facilitates easier and more permissive deformation, which enables to manufacture a wide range of shapes. Friction spinning can also be employed to produce rivets, known as friction-spun joint connectors (FSJC), and to join components (https://doi.org/10.1007/s11740-021-01094-8). A computational model is presented based on (http://hdl.handle.net/11427/36777, https://doi.org/10.1007/s00466-023-02303-0), using the finite element method to simulate forming and joining via FSJC. The model is implemented using the C++ software library deal.II. Several challenges arise due to the characteristics of the process. Frictional heating induces significant thermal softening, allowing for extreme deformations. At the same time, steep temperature gradients at the contact interface lead to highly localized material flow. These effects are captured through an arbitrary Lagrangian-Eulerian (ALE) formulation, which enables controlled smoothing of the underlying Lagrangian mesh. The rapid spinning motion — reaching approximately 10000 rpm — must also be modelled efficiently. This is achieved by decoupling the rotational motion from the mesh deformation/motion. The latter is modelled in 2D using axisymmetry, whilst preserving torsional effects. The frictional heat generation and temperature-induced material softening interact to bring the process temperature close to the material’s melting point. Consequently, the workpiece experiences a wide range of strain rates and temperatures (approximately 20 °C to 1500 °C for steel), necessitating suitable constitutive models. These models incorporate material properties computed using JMatPro (based on the CALPHAD method) to accurately represent this broad process window.
