Application of Finite Element Analysis for Thermal Cycle Prediction in Butt Welded Plates

The knowledge of the temperature history during the welding process is the crucial first step for understanding the metallurgical and mechanical effects of welding. The present study focused on finite element method for prediction of the thermal cycles induced during three-pass arc welding of butt plates. Two dimensional numerical analysis was performed by taking into account plane at the mid-section of the welded plate. Considering observed welded plates are symmetrical, half of the model was discretized into mesh of rectangular elements with four nodes. The nodal heat quantities were determined according to the real welding parameters and heat inputs were introduced in numerical analysis for defined finite element nodes. The heat flow in the thickness direction was neglected. The finite element model and heat inputs were defined through ANSYS Parametric Design Language code. The thermal analysis boundary condition was defined as convection which was specified using a convection coefficient of 5 center dot 10(-6) W/mm(2)K and reference temperature of 25 degrees C at all nodes. The analysis was carried out for three different finite element models in order to define optimal mesh model. The main difference between observed finite element models is element size and total number of elements. In order to determine the most appropriate finite element model, numerical results of temperature distribution were compared with real experimental values that were measured in transverse direction, during welding process by using thermocouples. Six locations in total, each at varying distances from the weld line, were used as observed points for temperature measuring. The practical application of the used approach was confirmed by establishing that, in the case of all FE models at each observed point, the values of the experimental measurement nearly match the FEA results. The optimal FE model was defined according to cost and time reduction.