As a novel polishing technology, polishing by laser beam radiation can be used to improve the sample surface finish without causing material losses. In order to study the effect of laser polishing on the surface morphology of S136D die steel, an L₁₆(4⁴) orthogonal experiment was designed to describe the variation trend of surface roughness with energy density. The two-dimensional transient model of laser polishing was established to simulate the evolution process of material surface morphology during laser polishing by combining numerical simulation with the experiment. The model uses a moving laser heat source to study the effects of capillary pressure and thermocapillary pressure in the laser polishing process. The experimental results show that the minimum roughness can be reduced to 0.764 μm, and the error between the actual molten pool depth and the simulated molten pool depth is 5.3%.

As a surface finishing technique for rapid remelting and re-solidification, laser polishing can effectively eliminate the asperities so as to approach the feature size. Nevertheless, the polished surface quality is significantly sensitive to the processing parameters, especially with respect to melt hydrodynamics. In this paper, a transient two-dimensional model was developed to demonstrate the molten flow behavior for different surface morphologies of the Ti6Al4V alloy. It is illustrated that the complex evolution of the melt hydrodynamics involving heat conduction, thermal convection, thermal radiation, melting and solidification during laser polishing. Results show that the uniformity of the distribution of surface peaks and valleys can improve the molten flow stability and obtain better smoothing effect. The high cooling rate of the molten pool resulting in a shortening of the molten lifetime, which prevents the peaks from being removed by capillary and thermocapillary forces. It is revealed that the mechanism of secondary roughness formation on polished surface. Moreover, the double spiral nest Marangoni convection extrudes the molten to the outsides. It results in the formation of expansion and depression, corresponding to nearby the starting position and at the edges of the polished surface. It is further found that the difference between the simulation and experimental depression depths is only about 2 μm. Correspondingly, the errors are approximately 8.3%, 14.3% and 13.3%, corresponding to Models 1, 2 and 3, respectively. The aforementioned results illustrated that the predicted surface profiles agree reasonably well with the experimentally measured surface height data.