Computational Concepts in Simulation of Welding Processes

Modelling and simulation of welding processes is a powerful engineering tool that is gaining importance in industrial applications. If implemented correctly, it allows to precisely predict how a welding process will take place and what will be its final result, determining its geometry, resultant microstructure and even the mechanical performance. However, the simulation of welding processes is still not as widespread as is the case for other production processes. For example, the use of modelling in forming processes is much more common, where the relative simplicity of the process contrasts with the complex nature of welding. In general, welding modelling requires more data, more variables and thus significantly more computational time. Furthermore, in most industrial applications, welding is now part of a large process chain and optimizing all these processes is a very time-consuming activity. Therefore, several assumptions and simplifications must be applied to reduce simulation time and ensure the practicality of a simulation-based approach.

A key element of a successful welding simulation is the right selection of the model used as the heat source. The characteristics of the heat source, such as geometry, dimensions and heat value, must be selected according to the welding process and the weld geometries. This chapter describes various models that have been used so far to represent different welding processes and geometries. Arc welding, resistance welding, beam welding and friction stir welding (FSW) processes as well as common welding, narrow gap welding and key-hole welding are discussed. The validation methods of the models used for the heat source are also explained. In this way, a reliable thermal modelling of the welding processes can be achieved by considering the thermal boundary conditions and various mechanisms of heat dissipation such as conduction, convection, and radiation. The output of the thermal modelling is then suitable to be used in mechanical modelling in a subsequent stage.

Phase transformations in structural steels determine both thermal and mechanical properties of steel. Phase transformations depend on chemical composition, initial microstructure, maximum temperature and the cooling rate. During welding a gradient of temperature, occurring both temporally and spatially, may cause a variety of phases to co-exist and whose percentages are locally variable. Continuous-cooling-transformation (CCT) diagrams are reliable tools to calculate the phases percentages. The thermal and mechanical properties of steel for every location and every time can be determined based on the percentage of each phase. With the implementation of variable material properties in both thermal and mechanical analyses, a reliable and accurate prediction of joint performance is feasible during welding simulation. This chapter describes how material properties can be calculated and implemented in the simulation of a welded process and resultant welded joint.

Mechanical analysis of welded joint is performed to calculate the residual stress and distortion generated during the welding process and cooling. Thermal elastic–plastic FEM analysis for doing so is usually performed using static implicit or quasi-static explicit methods. The strategies which can be employed to reduce the calculation time are described in this chapter. It is also explained how a mechanical analysis can be performed in a large-scale structure using approaches such as the inherent strain method. The effect of phase transformation in mechanical analysis is also discussed. Considering plasticization behavior of material and phenomena such as creep and annealing on mechanical analysis can increase the accuracy of simulation and in this chapter it is explained how one can implement these considerations in mechanical analysis. Finally, the effect of these phenomena on residual stress during welding are also addressed.

Although the use of computational methods has significantly reduced the cost of weld design and optimization processes, eliminating the number of experiments needed, there are still important limitations on the capabilities of these processes. In many cases, multiple modelling runs may be needed, greatly increasing the length of the design process. However, multiple strategies have been developed to reduce the number of runs needed, greatly increasing the effectiveness of using models in welded joint design workflows. Calibration procedures are also fundamental for modelling welded joints, being essential to ensure that the model correlates well with experimental data and produces valid results. Optimization strategies can be used also to calibrate the simulation parameters as well and some of the most important of these approaches are described in detail in this chapter.