Optimization of V-Bending of Grade 4 Titanium Bone Plates: A Combined Experimental, Numerical, and Artificial Intelligence Approach

The cold V-bending of Grade 4 titanium bone plates at room temperature is a critical forming operation that must be optimized to control strain localization and springback and to reduce the risk of surface cracking. This study proposes a combined experimental, numerical, and artificial intelligence-based approach for the analysis and optimization of this process. Tensile tests were first performed to characterize the mechanical behavior of the material and to calibrate the constitutive law used in the finite element model. The numerical model was then validated through comparison with experimental V-die bending results. A design of experiments was subsequently applied to investigate the effects of sheet thickness, die shoulder distance, punch radius, and punch displacement on two key responses: equivalent plastic strain (PEEQ) and spring back. The results show that sheet thickness and die shoulder distance are the most influential parameters. In addition, artificial neural network models were developed to predict process responses, and Bayesian regularization showed the best overall predictive performance among the tested ANN training algorithms, namely Levenberg–Marquardt, Bayesian regularization, and scaled conjugate gradient. The proposed framework provides a basis for optimizing the forming of titanium orthopedic implants.