Inverse modelling of fusion welding processes

The technique of solving the inverse heat conduction problem to find the steady state welding conditions by observable or desired weld features is outlined. The objective function is an augmented sum of weighted squares of differences in value between calculated and measured temperature-related functions, constraints and regularisation of unknown parameters. The algorithm for finding the minimum of the objective function is based on the Gauss-Newton method. As examples, the heat input distributions during laser and electron beam welding of steel and aluminium alloy thin sheets and thick plates are found. The weld interface, meltpool boundary and weld texture orientation have been used as the temperature-related functions. The results of simulation are directly compared with the experimental evidence. The suggested technique allows one to evaluate the heat efficiency of tire welding process. It can be applied to an optimisation problem.