Uncertainty Quantification With Sparsely Characterized Parameters: An Example Applied to Femoral Stem Mechanics

Abstract

Abstract The advent of state-of-the-art additive manufacturing (AM) processes has facilitated the manufacturing of complex orthopedic metallic implants such as femoral stems with porous portions based on lattice structures. These struts often have rough and not smooth textured surfaces, for which the irregularities may influence mechanical properties. To make robust predictions about the behavior of this kind of system, the variability effect of its parameters on the stem stiffness must be considered in the processes of modeling and design of porous femoral stems. Also, to improve the credibility of computational models used for hip implant analysis, which involves numerous uncertainties, there is a need for rigorous uncertainty quantification (UQ) framework for proper model assessment following a credible-modeling standard. This work proposes a UQ framework in the presence of sparsely characterized input parameters using the maximum entropy principle for analyzing a femoral stem implant model and thus to clarify how uncertainties impact the key properties of a porous femoral stem. In this study, uncertainties in the strut thickness, pore size, Young's modulus, and external forcing are considered. The UQ framework is validated using experimental results available from literature, following the guidelines set in an ASME standard.