Personnalisation des propriétés mécaniques des tissus mous du fessier humain par méthodes d’éléments finis et expérimentations in vivo

The purpose of this master degree project was to personalize the mechanical properties of the diverse soft tissue layers and to evaluate the necessity of the approach when the buttock lies on a rigid support and a foam cushion.

The first objective of the project was to develop a biomechanical FE model of the human buttock that integrates hyperelastic material formulations for the soft tissues with material properties taken from the literature. A magnetic resonance imaging protocol was realized to acquire geometric data of a healthy human male buttock in non-weight-bearing (MRI #1) and weight-bearing (MRI #2) conditions. An axial slice from MRI #1 was then used to reconstruct the 2D contours of the pelvis and buttock soft tissues while vertical sagging of the muscle and adipose tissues were measured by comparing correspondent slices from MRI #1 and #2. After the MRI acquisition, interfaces pressures were acquired experimentally in weight-bearing condition and were associated to the corresponding vertical sagging. These vertical sagging and interface pressure were necessary to realize the second objective, the mechanical properties personalization. After the experimental acquisition, a detailed FE model of the buttock was realized by creating a 5 mm extrusion of the 2D reconstructed profile of the buttock and by meshing the resulting volume using the ANSYS software (ANSYS Inc., Canonsburg, PA). Soft tissues were divided in muscles and adipose tissue layers, both meshed using a total of 7689 linear brick elements oriented toward the principal loading direction. The skin layer was meshed with 699 linear shell elements. Since the pelvic had a largely superior stiffness when compared to the soft tissues, it was modeled as a rigid body using equation-constraints on the contour of the pelvic-muscle interface. Contact elements were added to represent the buttock-support mechanical interaction and to simulate the friction at the interface.

The second objective of the project was to personalize the mechanical properties of the FE model and to evaluate their impact on the mechanical behaviour of the model in contact with a rigid support and a foam cushion. The optimisation was realized with a first order fitting approach that integrates the gradient method from the ANSYS’ design of experimentation module. The objective function to minimize was based on the difference between the volumes of the soft tissue layers measured experimentally and predicted by the FE model. The dependent variables were the hyperelastic material parameters (C₁₀, C₀₁ and ν) of the muscle and adipose tissues. Once the optimisation problem was established, dependent variable were modified through the process until the objective function was minimized and the states variables satisfied. The states variables were defined by the interface pressures (±10%) and the vertical sagging (±3 mm) measured experimentally. Ranges of tolerance were based on the acquisition process (measurement errors on the MRI images and the force sensing array matrix).

Once the personalization process was established on a rigid support, the relevance of such process was evaluated with the buttock in contact with a flat foam cushion normally used in pressure sores prevention. Hence, a FE model of a flat foam cushion was integrated to the FE model of the human buttock using 800 linear brick elements. A multi-linear curve was used to represent the hyperelastic behaviour of the cushion’s material. Results showed a variation of the interface and internal stresses and strains when using personalized mechanical properties. More precisely, the interface pressure increased of 11% while muscle’s Von Mises stresses increased by 20%. Moreover, a decrease of 19% was observed for the muscle’s peak internal strain. Finally, the personalization process resulted in a higher difference between the Von Mises stresses of the muscles and adipose tissues, thus confirming the premise for the development of deep tissue injury. Without this precise personalization, it will be impossible to represent the real internal phenomena of the buttock, such as the stresses augmentation from the buttock-support interface to the pelvis-muscle interface. These results clearly show the importance of the mechanical properties personalization to properly understand the interaction between the buttock and his support, as well as to evaluate a cushion’s capacity to distribute the internal stresses. (Abstract shortened by UMI.)