F. Sikström, M. Ericsson, University West, Trollhattan, Sweden
This paper describes a promising approach where finite element analysis (FEA) is combined with computer aided robotics (CAR) in off-line programming (OLP) of advanced robotized welding. FEA are used to find an optimized weld power signal based on weld trajectories obtained from CAR-simulations that in turn are used to find an optimized operating sequence. The weld power signal is calculated by applying simulated feedback control during the FE weld simulation. The optimization ensures a full penetration weld while the total specific thermal energy input is minimized in order to mitigate unwanted residual stress and distortion.
This simulation approach is especially interesting in the case of complex geometries where thermal boundary conditions vary due to abrupt changes in the geometry or when heat sinks like clamps are in the vicinity of the weld path. The objective is to support the design of robotized welding and significantly reduce the number of costly trials and errors in physical implementations.
The approach has been successfully verified by experiments where a fabricated jet engine object of austenitic steel was welded using a robotized Nd:Yag laser. This investigation clearly highlights the need for a varying input power signal in order to compensate for changes in thermal boundary conditions. The experiments showed good agreement between weld simulation and real welding. Assumptions behind the modeling and the experiments together with an evaluation of the conformity between them are discussed. The relevance of this paper is a useful method for off-line optimization of robot trajectories and varying process parameters.
Summary: Finite element analysis (FEA) is combined with computer aided robotics (CAR) in off-line programming (OLP) of advanced robotized welding. This is a useful method for off-line optimization of robot trajectories and varying process parameters.
