Post weld heat treatment, or PWHT, is often used to improve the mechanical properties of materials that have been welded. Several publications have investigated the effects of the PWHT process using experimental designs. However, the modeling and optimization using the integration of machine learning (ML) and metaheuristics have yet to be reported, which are fundamental steps toward intelligent manufacturing applications. This research proposes a novel approach using ML techniques and metaheuristics to optimize PWHT process parameters. The goal is to determine the optimal PWHT parameters for both single and multiple objective perspectives. In this research, support vector regression (SVR), K-nearest neighbors (KNN), decision tree (DT), and random forest (RF) were ML techniques employed to obtain a relationship model between PWHT parameters and mechanical properties: ultimate tensile strength (UTS) and elongation percentage (EL). The results show that the SVR demonstrated superior performance among ML techniques for both UTS and EL models. Then, SVR is used with metaheuristics such as differential evolution (DE), particle swarm optimization (PSO), and genetic algorithms (GA). SVR-PSO shows the fastest convergence among other combinations. The final solutions of single-objective and Pareto solutions were also suggested in this research.
This study investigates the effects of post-weld heat treatment (PWHT) on the microstructures and mechanical properties of plasma arc-welded 316 stainless steel. The experimental parameters included the solid solution temperatures of 650 °C and 1050 °C, solid solution durations of 1 h and 4 h, and quenching media of water and air. The mechanical properties were evaluated using Vickers hardness testing, tensile testing, scanning electron microscopy (SEM), and optical microscopy (OM). The highest ultimate tensile strength (UTS) of 693.93 MPa and Vickers hardness of 196.4 in the welded zone were achieved by heat-treating at 650 °C for one hour, quenching in water, and aging at 500 °C for 24 h. Heat-treating at 650 °C for one hour, followed by quenching in water and aging at 500 °C for 24 h results in larger dendritic δ grains and contains more σ phase compared to the other conditions, resulting in increased strength and hardness. Additionally, it shows wider and shallower dimple structures, which account for its reduced impact toughness.
This study investigates the effects of a back plate preheating assistance system and deep rolling (DR) on axial force and tunnel defects during friction stir welding (FSW). Different preheating configurations—advancing side (AS), retreating side (RS), and both sides—were examined to evaluate their impact on axial force reduction, temperature distribution, and defect minimization. Axial force measurements were taken using a dynamometer, and temperature histories were recorded with a thermal camera. The results demonstrate that a preheating temperature of 200 °C is optimal, reducing axial force by 30.24% and enhancing material flow. This temperature also facilitated deeper tool penetration, especially when preheating was applied to both sides. Preheating on the AS resulted in the smallest tunnel defects, reducing defect size by 80.15% on the RS and 96.91% on the AS compared to the non-preheated condition. While DR further reduced tunnel defects, its effectiveness was limited by the proximity of defects to the surface. These findings offer significant insights for improving the FSW process.
This study investigates the effects of deep rolling parameters, pressure, speed, and offset, on the residual stress distribution and material deformation in UIC 860 Grade 900A railway rails. We will model deep rolling to simulate the process and predict the residual stress profile in railway rails. Subsequently, we will rigorously compare and analyze the FEM simulation results with experimental data to optimize deep rolling parameters for improved residual stress distribution. Using both experimental methods and finite element analysis via ANSYS 2023 R1, the study varied deep rolling parameters. Experimental deep rolling pressure was set at 150 bar, speed at 1800 mm/min, and offset at 0.1 mm, while FEA simulations predicted corresponding pressures of 157 bar and speed of 1796.52 mm/min. These parameter settings were chosen to induce significant surface compressive stresses that could enhance the material’s mechanical performance. The experimental results showed an average compressive residual stress of 498.9 MPa, closely aligning with the FEA-predicted value of 502.5 MPa. A paired t-test revealed no statistically significant difference between the two results, with a T-value of −0.22 and a p-value of 0.833, validating the reliability of the FEA model. The consistent deformation observed in both experimental and FEA simulations, especially with a 0.1 mm offset, confirmed that the rolling parameters were effective in producing uniform stress distribution, albeit with a slightly extended processing time due to the small offset. Overall, the findings confirm that optimizing the deep rolling parameters of pressure, speed, and offset leads to favorable residual stress distributions and improved material properties. The results indicate that FEA is a reliable tool for predicting the outcomes of deep rolling, and this study provides a strong foundation for further refinement of the process to enhance performance in practical applications, such as railway rail treatments.