Development of a stochastic approach for fatigue life prediction of AlSi12 alloy processed by selective laser meltingTuesday (27.09.2016) 17:45 - 18:00 Part of:
Selective laser melting (SLM) is an additive manufacturing process for metals with high design freedom and short lead time. The processed alloys have a unique microstructure and mechanical properties which can satisfy the freedom needs for automotive and aerospace industries. During experimental testing of fatigue performance of AlSi12, high scatter of fatigue values was observed as well as a sensitivity of fatigue life to residual structural defects and as-built surface roughness.
To address the effects of these limitations, an approach was developed to quantify the effect of these process-specific material parameters on the fatigue life as well as the corresponding scatter. The approach is based on crack propagation modelling, the maximum likelihood method, Weibull probability density function and the weakest link theory, and implemented in Matlab. In-process base plate heating at 200 °C was used to cause a lower cooling rate during consolidation. This, in turn, leads to an enhancement of stability of the melt pool, thus residual gas porosity accumulation is avoided. Although, with such remnant porosity, a density higher than 99% can be obtained, the residual pores serve as preferred crack initiation sites. On the other hand, if the samples were tested in the as-built condition, surface roughness causes a notch like effect promoting crack initiation form the surface and reducing fatigue life significantly. So along with experimental constant amplitude testing (CAT), analytical modeling will help to understand the failure mechanism in presence of different defect types and states. A statistical inference based on Weibull distribution of experimental values will capture the highest possible fatigue life from which an estimation of the fatigue scatter can be made.
According to the prediction model, the reduction of residual porosity of base plate heated samples leads to a decrease of fatigue scatter and a slight decrease of fatigue endurance in high cycle fatigue (HCF) regime, which is in conformance with the experimental results. The reduction of endurance was more pronounced in very high cycle fatigue (VHCF) regime. One more highlight of the model is role of micro-porosity in crack bluntation which acts as a cause of crack inhibition.