Model for predicting heat generation and temperature in friction stir welding from the material properties

Abstract

This paper describes a simple numerical model for predicting the heat generation in friction stir welding (FSW) from the material hot deformation and thermal properties, the process parameters, and the tool and plate dimensions. The model idealises the deformation zone as a two-dimensional axisymmetric problem, but allowance is made for the effect of translation by averaging the three-dimensional temperature distribution around the tool in the real weld. The model successfully predicts the weld temperature field and has been applied with minimal recalibration to aerospace aluminium alloys 2024, 7449 and 6013, which span a wide range of strength. The conditions under the tool are presented as novel maps of flow stress against temperature and strain rate, giving insight into the relationship between material properties and optimum welding conditions. This highlights the need in FSW for experimental high strain rate tests close to the solidus temperature. The model is used to illustrate the optimisation of process conditions such as rotation speed in a given alloy and to demonstrate the sensitivity to key parameters such as contact radius under the shoulder, and the choice of stick or slip conditions. The aim of the model is to provide a predictive capability for FSW temperature fields directly from the material properties and weld conditions, without recourse to complex computational fluid dynamics (CFD) software. This will enable simpler integration with models for prediction of, for example, the weld microstructure and properties.