Hierarchical Bayesian constitutive model selection for high-strain-rate soft material characterization

Area of Science:

• Materials Science
• Biophysics
• Mechanical Engineering

Background:

• Characterizing soft, tissue-like materials under ultra-high-strain-rate conditions is vital for engineering and medicine.
• Existing microcavitation techniques face challenges with measurement noise and parameter estimation uncertainty.
• Accurate material property determination requires robust methods for soft materials under extreme conditions.

Purpose of the Study:

• To address limitations in microcavitation techniques for soft material characterization.
• To develop a hierarchical Bayesian model selection method for Inertial Microcavitation Rheometry (IMR).
• To accurately determine constitutive models and material parameters for soft viscoelastic materials.

Main Methods:

• Employed a hierarchical Bayesian model selection framework with Inertial Microcavitation Rheometry (IMR).
• Utilized a weighted Gaussian likelihood with a hierarchical noise scale for uncertainty quantification.
• Incorporated physically informed priors to penalize complex models and ensure model parsimony.

Main Results:

• The method successfully explored constitutive model parameter spaces for laser-induced microcavitation bubble oscillations.
• Probabilistic model selection provided initial estimates for maximum a posteriori (MAP) material parameters.
• Synthetic tests and experimental data from gelatin, fibrin, polyacrylamide, and agarose showed accurate model reproduction and consistent cross-institutional results.

Conclusions:

• The developed Bayesian method enhances the fidelity of soft material characterization at ultra-high strain rates.
• It effectively quantifies uncertainty and selects the most credible constitutive models for viscoelastic hydrogels.
• This approach offers a robust and consistent framework for material property determination across different institutions.