Finite Element Calibration of Muscle for Human Body Models

Abstract

Accurate modelling of muscle tissue is critical for realistic human body simulations in biomechanics and vehicle safety applications. This project focuses on calibrating and validating visco-hyperelastic material models for muscle tissue, specifically comparing the Ogden model and a General Hyper-elastic rubber model, implemented in LS-DYNA. The study leverages uniaxial compression test data at strain rates ranging from quasistatic (0.01/s) to dynamic (90/s) to fit material parameters. Analytical derivations and numerical optimizations (using LS-OPT) were employed to calibrate the models, followed by validation via Single-Element Tests and Unit Cell Tests to assess stability and accuracy. Results demonstrate that the Ogden model effectively captures quasi-static and low strain-rate behaviour but exhibits discrepancies in visco-elastic regimes, overestimating stresses at higher strains and strain-rates. The General Hyper-elastic model provided a comparable fit but required higher-order terms for accuracy. Both models achieved numerical stability close to 70% compressive strain, benchmarked against a material implementation known to be numerically stable. Full-scale impact simulations using the SAFER Human Body Model revealed close alignment with experimental data for humerus plate impacts, though deviations occurred in bar impact scenarios. Key challenges included interpreting LS-DYNA’s visco-elastic implementation. Future work should address viscous behaviour modelling, expand experimental datasets, and refine geometry-specific calibrations. This study advances the fidelity of muscle tissue representation in HBMs, supporting safer automotive design and injury prediction.