Optimization of V-Bending for Grade 4A Titanium: A Combined Experimental and Artificial Intelligence Approach

¹ Sousse Higher Institute of Applied Sciences and Technology (ISSATSo), University of Sousse, Cité Taffala, Ibn Khaldoun 8, 4003 Sousse, Tunisia
² LMS, National School of Engineers of Sousse (ENISO), University of Sousse, Route de Ceinture Sahloul, Cité Hammam Maarouf, 4054 Sousse, Tunisia
³ LMMP, National High School of Engineering of Tunis (ENSIT), 5 Avenue Taha Hussein, Montfleury 1008 Tunis, Tunisia

Abstract

Cold bending of Grade 4A titanium bone plates is a crucial process requiring optimization to prevent crack formation. This study integrates experimental analysis, numerical simulations, and artificial intelligence modeling. Tensile tests were performed to determine the mechanical properties of titanium, validating a finite element model for V-bending simulation. A design of experiments was conducted to assess the influence of bending parameters on key responses: equivalent plastic strain (PEEQ) and springback. Results highlight the predominant role of material thickness and die shoulder distance. To enhance prediction accuracy and process optimization, neural networks using the Bayesian regularization algorithm were applied. This approach contributes to improving the design and manufacturing of orthopedic implants.