Understanding and Modeling the Effects of Surface Texture on Fatigue Behavior of Additively Manufactured Parts

Abstract

Additive manufacturing, considered as a future manufacturing technique, provides an unique advantage of fabricating near-net-shape parts without any additional processes. While there are extensive efforts to improve mechanical properties of additively manufactured materials, their fatigue strength, particularly those with an unmachined surface condition, still remains significantly low. It delays the adoption of additive manufacturing in practical applications and weakens its primary advantage of direct usage after fabrication as the low fatigue strength requires post-surface treatments. This dissertation aims to investigate the effects of surface texture on fatigue behavior of additively manufactured materials. Firstly, the surface texture of unmachined additively manufactured parts is examined to understand its effect on fatigue behavior under force-controlled and strain-controlled cyclic loadings. In addition to unmachined surface condition, the fatigue behavior of additively manufactured parts after different post-surface treatments is investigated. Based on the experimental findings, several objectives have been identified and addressed. These include the development of a hybrid surface roughness parameter tailored to represent the surface texture of additively manufactured parts. Additionally, surface features that are critical to fatigue strength of additively manufactured material have been investigated through various surface measurement techniques. Using the surface texture analyses and fatigue test outcomes, a non-destructive surface texture based fatigue prediction model has been proposed. The prediction model has been validated by comparing predicted fatigue lives with empirical data. It has further been employed to estimate fatigue lives under multiaxial loading conditions. The developed hybrid surface roughness parameter effectively correlates the surface texture of additively manufactured parts with various surface conditions to their fatigue strength. Finally, most of the fatigue lives predicted by the surface texture based fatigue prediction model fall within scatter bands of 3 compared to empirical data.