Élaboration et validation d'un modèle numérique de genou

The authors report the methodology of the construction of a multibody model of the knee and the validation of the kinematics of the modelled knee. The construction of the model includes: the rigid bodies represented by osseous components (femur, tibia, fibula, patella), the ligamentous structures (collateral ligaments, patellar ligament, cruciates ligaments), the muscular part represented by the quadriceps. Morphological data were acquired through 3D CT scans for the bones and a biometrical study of the ligaments (insertions, orientation, length, section). Ligament biomechanics was modelled as bilinear springs (in compression the tightness is null; in traction it is a function of length, section and Young modulus of elasticity). The quadriceps was modelled as a sliding channel with a translatory servocommand. Contacts at the interfaces (femur/patella; femur/tibia) were evaluated according to the index of penetration (distance D) between two bodies where effort was: D × 105 N/mm2). The model was tested simulating a symmetrical kneeling (800 N body weight) and required a ground link modelled as a ball and socket joint. The model was developed under ADAMS software. The validation of the kinematics of the modelled knee was provided according to the data of Wilson et al. who have shown that (i) in normal knees, internal/external rotation, abduction/adduction and all three components of translation are coupled to flexion angle both in passive flexion and extension; (ii) the tibia rotates internally as the knee is flexed. The consistency of the coupled motions support the model's premise that passive knee motion is guided by isometric fascicles in anterior and posterior cruciates, by the medial collateral ligament and by articular contact in the medial and lateral compartments. The main curves (internal/external rotations; posterior/anterior translation) of the model conforms with the framework of Wilson.