Machine Learning and Optimization-Driven Analysis of Build Orientation Effects on the Microstructure and Mechanical Behavior of Ti-6al-4v Fabricated by Electron Beam Melting Additive Manufacturing


Sicakdemir B., ÖZSOY K.

Journal of Materials Engineering and Performance, 2026 (SCI-Expanded, Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1007/s11665-026-13415-y
  • Dergi Adı: Journal of Materials Engineering and Performance
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Chemical Abstracts Core, Compendex, INSPEC
  • Anahtar Kelimeler: additive manufacturing, ANOVA, build orientation, electron beam melting, machine learning, Ti-6Al-4 V
  • Isparta Uygulamalı Bilimler Üniversitesi Adresli: Evet

Özet

The study examined the effects of manufacturing on the mechanical and microstructural properties of Ti-6Al-4 V specimens. This study aimed to take advantage of the unique design flexibility enabled by electron beam melting (EBM) technology to overcome existing production constraints by altering part build orientation and to reveal the effect of orientation on performance through mechanical and microstructural analyses of specimens produced at different angles. In this study, the specimens designed with different build orientations were manufactured using the EBM process with Ti-6Al-4 V Grade 5 metal powder provided by ARCAM©. Flexural and hardness tests, as well as dimensional accuracy and surface roughness measurements, were conducted to determine the mechanical and physical properties of the specimens. The microstructure and fractured surfaces of the specimens were analyzed using a scanning electron microscope (SEM). The study revealed that the mechanical characteristics of Ti-6Al-4 V specimens with different build orientations are highly dependent on build orientation and thermal effects. While the highest flexural strength and stiffness were obtained in specimens built at 90°, the specimens produced at 0° exhibited superior ductility, indicating that build orientation significantly affects both strength and deformation behavior. It was shown that build orientation significantly influenced the microstructure and mechanical properties of Ti-6Al-4 V specimens produced by EBM, with 0°, 60°, and 90° orientations resulting in distinct α/β phase morphologies. Flexural strength was affected by orientation at a rate of 92.68%, with an F-value of 190.00 confirmed by ANOVA. According to the classification results obtained using machine learning algorithms, the random forest model consistently outperformed the others in predicting dimensional accuracy, surface roughness, and flexural strength, achieving the highest accuracy rates of 92%, 96%, and 94%, respectively. These results demonstrate its superior accuracy, stability, and generalization performance across all evaluation metrics and build orientations.