Multi‐Objective Collaborative Process Optimization Based on Deep Learning of Laser In Situ Forging Additive Manufacturing Ti‐6Al‐4V

Aiming at the problems of insufficient forming quality and poor surface performance in the preparation of Ti‐6Al‐4V alloy by laser powder bed fusion (LPBF), this study proposed a laser in situ forging additive manufacturing (LIF‐AM) technology, which integrates LPBF with femtosecond laser in situ shock forging and realizes defect control and surface modification via plasma oscillation and thermo‐mechanical coupling effect induced by ultrafast laser. To solve the bottleneck of difficult parameter optimization caused by numerous adjustable parameters and strong coupling nonlinear correlation in the LIF‐AM process, a deep learning‐driven collaborative optimization method for process parameters was developed. First, orthogonal experiments were carried out to collect the roughness and hardness data of Ti‐6Al‐4V specimens, and a convolutional neural network (CNN)‐transformer prediction model was constructed based on the previous research to accurately establish the mapping relationship between LIF‐AM process parameters and forming quality. Next, the CNN‐transformer model combined with the reference‐point based non‐dominated sorting genetic algorithm (RNSGA) III algorithm is used to conduct multi‐objective collaborative optimization with the goals of minimizing surface roughness and maximizing microhardness. Experimental verification results show that under the optimized process parameters, the internal pore defect volume of Ti‐6Al‐4V specimens is reduced by about 90% compared with the traditional LPBF process; the microhardness is increased by 15.8%–389 HV 0.1 , and the high‐cycle fatigue strength is significantly improved by 32.3%–514.7 ± 24.7 MPa. This method breaks through the limitations of traditional parameter optimization methods for LIF‐AM, providing an accurate and efficient process parameter optimization scheme for forming quality and mechanical properties improvement of Ti‐6Al‐4V alloy prepared by additive manufacturing.