the location of the porosity has been shown to be important for fatigue life, with shorter lives associated with cracks originating from pores located near the test piece surface, for ti-6al-4v samples produced by bothebm 2, 8 and selective laser melting (slm) 18, 19 . murakami 20 has suggested that pores generate a stress intensity factor, which is dependent on the size of the defects and is greater for defects at the surface, while finite element (fe) modelling has confirmed that the stress concentration generated by pores is greater when they are within one diameter of a free surface 21 . inclusions can also result in a much shorter fatigue life when they are located close to a sample surface, in comparison to the bulk 22 . previously, for aluminium fatigue samples containing pores, it has been shown that it is possible to predict where a fatigue crack will initiate using criteria based on the location of maximum stress and plastic strain concentration 23 . in standard engineering components, where the number density of defects is accurately known, it is possible to predict the likelihood of a defect lying in a location where it will initiate a crack, and so predict the probable lives of components due to defect initiated fatigue 22 . however, for ebm am, both the pore number density and pores effect on fatigue have not been fully characterised.
in this paper we have demonstrated the use of a novel time-lapse x-ray computed tomography (ct) method to improve the understanding of the effect of pores on the fatigue life of ebm components. ct allows the complete defect (pore) distribution in fatigue samples to be identified prior to testing. only by such a method can the pores that do not lead to fatigue crack initiation be identified and contrasted to the initiating pores on the fracture surface. we have then examined the extent to which we can identify the pores most likely to nucleate a fatigue crack using ranking strategies that include the pore size, aspect ratio, proximity to the surface and to other pores, and the propensity for local plastic strain. this has allowed us to examine the extent to which these factors control which defect is likely to become the initiating one. furthermore, by combining this knowledge with that regarding the non-uniform spatial distribution of pores in ebm components, general recommendations to optimise the fatigue life of components can be found. 3d9ccd7d82