Finite element modeling to predict bulk mechanical properties of 3D printed metal foams

Abstract

Worldwide yearly 800.000-1.000.000 people receive a total artificial hip. 8-9% Of all patients requires a second surgery to correct the primary arthroplasty. In more than 70% of cases, aseptic loosening is the cause for implant failure. Possible pathways leading to aseptic loosening are stress shielding and micro-motions. Stress shielding can be reduced by less stiff implant material in the susceptible areas and micro-motions can be minimized by a stable fixation in the bone, e.g. due to bone ingrowth into the implant. Three dimensional (3D) printing techniques provide the possibility to combine solid metal with metal foam in one implant. The apparent density (AD) of the foam is in direct relation with the Young’s modulus, so the stiffness of the foam can be tailored with the AD. The printing does entail imperfections in the foam, such as irregular cross-sections of the struts and porosity within the struts. Analytical models are based on the perfectly regular situation and therefore do not satisfy in the prediction of the foam stiffness. A custom-made finite element (FE) modeling tool was developed to generate models of metal foams that do include these irregularities. The struts were all composed of several beam elements to which different cross-section sizes were assigned based on a Gaussian distribution. In addition, porosity within the struts was modelled by assigning a void percentage to the matrix material, which was also Gaussian distributed. In this study, the predicted bulk mechanical properties of models generated with the modeling tool were compared to analytical models and to experimental results in order to validate the FE results. It was shown that FE modeling is a promising method to predict the stiffness of 3D printed metal foams. Especially the stiffness of foams with a low AD was properly predicted by the FE model. Further development is required to optimize the accuracy of the outcome. It is recommended to include non axial alignment of the beam elements. In the future, FE modeling can be used to optimize the geometrical and mechanical properties of patient specific implants.