Impact of defects on high cycle fatigue life in wire-arc additive manufactured TC17 alloy

Abstract

Engineering applications for additively manufactured (AM) titanium alloy components are often constrained by suboptimal fatigue properties and high variability in fatigue data due to defects. This study aims to address these limitations by developing a fatigue life prediction model that incorporates the influence of defects in wire arc additive manufacturing (WAAM) TC17 alloy. The microstructure and mechanical properties of WAAM-TC17 were thoroughly characterized. Results revealed that the average α-grains length and width in WAAM-TC17 was significantly smaller, approximately one-twelfth and one-seventeenth of that in Forged-TC17, respectively. The yield strength of the WAAM-TC17 horizontal and vertical specimens was approximately 93% of the Forged-TC17. However, the high-cycle fatigue (HCF) performance of WAAM-TC17 specimens was inferior due to crack initiation dominated by porosity and lack of fusion (LOF) defects. To enhance fatigue life prediction accuracy for defective WAAM-TC17 specimens, a novel parameter K*, derived from the stress concentration factor (K_t) using support vector regressor (SVR) in machine learning (ML), was introduced. The K*-N mean curve demonstrated high predictive accuracy for the HCF life of defective WAAM-TC17 specimens, with a standard deviation (STD) of 0.33.