Laser Beam Welding Parametric Optimization for AZ31B and 6061-T6 Alloys: Residual Stress and Temperature Analysis Using a CCD, GA and ANN

Abstract

Laser beam welding (LBW) is a technology that efficiently joins materials using a concentrated laser beam, providing advantages such as a small heat-affected zone (HAZ), high precision, adaptability to various materials, and seamless integration into automated processes. Precision in LBW is limited by challenges in selecting appropriate process parameters, performing expensive and risky physical experiments, and dealing with materials with high thermal conductivity and low melting temperatures, such as AZ31B and AA6061. This study aims to predict laser welding process parameters, encompassing beam power (4000W-6000W), spot size (0.4mm-0.8mm), welding speed (30mm/s-50mm/s), and segment number (5-15) utilizing the Response Surface Methodology (RSM), Genetic Algorithm (GA) and backpropagation neural networks. Beam power significantly affects peak temperature, leading to deeper laser penetration and a higher maximum temperature, while the segment number plays a crucial role in residual stress by potentially causing uneven cooling and distinct thermal gradients. The ANN results confirm the model's accuracy and the GA-predicted process parameters include a laser power of 4000W, welding speed of 50mm/s, spot size of 0.6mm, segment number of 15, peak temperature range of 888.9°C to 1020.5°C, and a preferred equivalent stress of 1396.62 MPa.