Development of a stochastic approach for fatigue life prediction of AlSi12 alloy processed by selective laser melting

Parts manufactured by selective laser melting (SLM) process possess unique features in terms of surface roughness, microstructure, residual stresses as well as defect distribution. These defects are responsible for failure of the parts in functional applications. When fatigue loading is applied, these defects are the dominant cause of crack initiation, resulting in scatter of fatigue properties. This scatter occurs due to interacting phenomena like defect size, location as well as the magnitude and type of load. For the purpose of investigating the effect of defects on fatigue life performance of AlSi12 manufactured by selective laser melting, a procedure was developed based on the weakest-link theory and Weibull's probability density function. Using various destructive and non-destructive techniques, defects, including remnant porosity and surface roughness, have been characterized in amount, size and location. Therefore fatigue life prediction, relying on equations constituted from crack propagation properties, was carried out. Predicted fatigue life and Weibull's statistical parameters were used to compare the effect of both defect types on fatigue reliability of AlSi12 produced by SLM. The most probable fatigue life for a sample was interpreted based on Weibull probability density function with respect to maximum probability of occurrence. The prediction of numerous possible values enabled an estimation of fatigue scatter to be made. Thus, the findings of this novel approach enabled conclusions about strength and reliability of different SLM AlSi12 configurations and gave a prelude towards application-oriented design of SLM components. (C) 2017 Elsevier Ltd. All rights reserved.