Alexander Ilin

Computer Aided Development of the Crack-Free Laser Welding Processes

The present paper describes the new methodology used for the development of the crackfree welding processes. The presented approach is based on the accurate experimental observations on binary Al-Si alloys, which clearly demonstrate that the crack initiation is a result of the accumulation of macroscopic tensile strain in a microscopic intergranular liquid film of segregates at the final stage of the weld metal solidification. The numerical model takes into account the effects of strain accumulation as well as the influence of thermo-dynamical and thermo-mechanical properties of the welded material. The new approach provides a clear phenomenological interrelation between the cracking susceptibility, parameters of the welding process and properties of the base and filler material. It is successfully applied for development of technological means for elimination of the solidification cracking during welding of aluminium alloys AA6056, such as a multi-beam welding.

Nonlinear Optimization Methods for the Determination of Heat Source Model Parameters

In this paper the technique of parameter identification is investigated to reconstruct the 3D transient temperature field for the simulation of laser beam welding. The reconstruction bases on volume heat source models and makes use of experimental data. The parameter identification leads to an inverse heat conduction problem which cannot be solved exactly but in terms of an optimal alignment of the simulation and experimental data. To solve the inverse problem, methods of nonlinear optimization are applied to minimize a problem dependent objective function.In particular the objective function is generated based on the Response Surface Model (RSM) technique. Sampling points on the RSM are determined by means of Finite-Element-Analysis (FEA). The scope of this research paper is the evaluation and comparison of gradient based and stochastic optimization algorithms. The proposed parameter identification makes it possible to determine the heat source model parameters in an automated way. The methodology is applied on welds of dissimilar material joints.