A genetic algorithm-calibrated heat source model for reconstructing thermal history and precipitate evolution in additive friction stir multilayer deposition

Abstract

Additive friction stir deposition (AFSD) has shown significant potential for additive manufacturing of large structures, which involves unique thermal cycling effects in multilayer deposition. The present endeavor firstly attempts to develop a universal moving heat source model (the double-hump model) for finite element multilayer thermal simulation, which mathematically simplifies the non-uniform distribution of macroscopic heat flux densities in the deposition zone. By adopting a planar heat source formulation, the model consequently neglects the radial contribution of plastic dissipation. The main objective is to seek a convenient simulation means to reconstruct the thermal history of the multilayer deposition and to get its impact on the precipitate evolution and mechanical properties. The heat source model parameters were determined by a genetic algorithm based on the principle of monarch selection and uniform crossover strategy to match the experimental data. The temperatures predicted by the finite element model for the preheating and multilayer deposition stages agreed well with those obtained experimentally. The location of the maximum heat flux density in the deposition zone was evaluated for the first time. The simulation data and characterization results show that the simulation of the multilayer AFSD process based on the heat source model can effectively reflect the second phase evolution in the thermal cycle and the influence on the mechanical properties. Furthermore, this study presents a novel thermal analysis of distinctive phenomena occurring in the AFSD process, showing the potential of heat source modelling in the thermal simulation of multilayer deposition.