Machine learning based very-high-cycle fatigue life prediction of AlSi10Mg alloy fabricated by selective laser melting

Abstract

Few machine learning models are applied to investigate the influence of defect features on very-high-cycle fatigue performance of additively manufactured alloys and these models usually suffer from data scarcity. Interpolation methods are run to enlarge dataset size and machine learning models are established to investigate the synergic influence of layer thickness, stress ratio, stress amplitude, defect size, shape and location on fatigue life of selective laser melted AlSi10Mg. Results show that the increases in defect distance to surface, circularity, and layer thickness favor higher fatigue life; however, the increases in stress amplitude, stress ratio, and defect size decrease fatigue life.