Finite Elements Model of a composite human femur and validation through Digital Image Correlation

In the field of osteosynthesis, modelling of medical devices in silico is essential to predict and simulate their post-implantation behavior. The purpose of these procedures is to develop and facilitate an increasingly patient-specific approach that prevents the occurrence of complications due to errors in the placement of the implant. The term osteosynthesis refers to a reconstructive surgical technique whose purpose is to join two or more bone fragments after their realignment, using screws, plates and other devices, involving materials with both mechanical and biological properties appropriate to the implant. In this regard, it is of utmost importance to adopt a patient-specific approach, as implant failure or success depends on both technical choices and patient-related factors. Into this scenario, there are several studies that propose finite element modeling of the human femur, with the aim of obtaining fracture risk prediction and assessment. The study tries to define an approach focused on proper bone modeling of the femur that takes into account its mechanical and physiological characteristics. The load was simplified and considered as a uniaxial compression force concentrated on the femoral head, hypothetically representing the load of the coxo-femoral joint. It was chosen to use the finite element approach, discretizing bone geometry with tetrahedral elements and assigning the mechanical properties of Sawbone, the synthetic material chosen to mimic the femur. The model obtained was loaded uniaxially and bound in the region of the condyles using an encastre, simplifying the real case. Since the purpose was the validation of the finite element model, a comparison of displacements and strains obtained with Finite Element Analysis (FEA) was made with those obtained through an experimental procedure. This involved the use of Digital Image Correlation (DIC) which allowed the recording of the displacements and strains of a Sawbone femur bound to the test machine. The comparison procedure was automated using a MATLAB script, calculating the errors and accuracy of the Finite Element Model (FEM) estimate compared to the real experimental case with DIC. The results obtained showed a satisfactory prediction of displacements along the longitudinal axis and the transverse axis. The deformations, however, did not show a good degree of prediction, resulting in high errors and poor accuracy. The reasons for this could be the lack of further experimental evidence and acquisitions, together with the need to perform new FE simulations. In the latter case, changing the test parameters and the physical-mechanical characteristics of the bone could lead to improvements in the predicted results. Knowing the limitations of the previous test set-up, a new structure is presented in the second part of the work, designed considering the modularity of the applications and the anatomical position of the femur as accurately as possible. The new set-up consists of six elements, three of which are made of steel and three to be made of ABS with 3D printing to mimic the shape of the femur. The new experimental set-up needs to be tested using the DIC technique to validate a FE model to be built from scratch based on the geometry created. The optimization of the model could make up for the shortcomings highlighted by the preliminary analysis, reproducing more accurately the mechanical properties of the femur.