Multi-objective optimization of steel fusion welding
Steel fusion welding is characterized by phase transformations influencing the final mechanical properties. Such properties and modifications are strongly related to welding parameters such as speed, current, voltage and heat input. In the present paper hardness, residual stresses, phase transformations, tensile, fatigue and impact properties of different steel welds have been related to the material composition, geometry and the welding conditions by employing a multi-objective optimization software (modeFRONTIER). As a matter of fact the weight of the different parameters influence have been evaluated through such kind of study. An optimization analysis has been performed in order to identify the best welding condition for each kind of steel taking as final goal the fatigue and impact strength of the joints. Very few information are available on the microstructure-fracture-fatigue properties of fusion welded joints in the open literature. The practical application of any steel on a larger scale is critically dependent on its weldability for fabrication.
In all arc-welding processes, the high heat source produced by the arc and the associated local heating and cooling result in a number of consequences in material behaviour and several metallurgical phase changes occur in different zones of a weldment. The microstructure and stress characteristics of welded joints differ from those of the base material and the performance of the welded structure is usually limited by the initiation of failure within the Heat Affected Zone (HAZ) of the base material. Therefore, to ensure the reliability of large-scale structures which will be subjected to dynamic impact loading conditions, it is essential to evaluate the mechanical properties of their structural materials, including their weld metals. This paper presents the results of a broad experimental campaign performed on different steel joints obtained with different processing parameters with a special focus on the resulting microstructural properties and consequently mechanical properties. The data were employed to build a predictive database through modeFRONTIER for industrial welding procedures.