Evaluating computational performances of hyperelastic models on supraspinatus tendon uniaxial tensile test data

Accurate modelling of the mechanical behaviour of tendon tissues is vital due to their essential role in the facilitation of joint mobility in humans and animals. This study focuses on the modelling of the supraspinatus tendon which helps to maintain dynamic stability at the glenohumeral joint in humans. It is observed that in sporting activities or careers that involve frequent arm abduction, injuries to this tendon are a common cause of discomfort. Therefore, this paper evaluates the relative modelling capabilities of three hyperelastic models, namely the Yeoh, Ogden and Martins material models on the tensile behaviour of three tendon specimens. We compare their fitting accuracies, convergence rates during optimisation, and the different forms of sensitivities to data-related features and initial parameter estimates. We find that the Martins model outperforms the other models in fitting accuracies; the Yeoh model has the most stable performance across all initial parameter estimates (with correlations above 99 %) and has the fastest convergence rates (above 20 and 8 times as fast as the Ogden and Martins models’ rates, respectively); and that the Ogden model does not depend on differences in the topological features of the test data. The material parameters of relevant constitutive model may be used for further development of computational models.