Bayesian optimization to infer parameters in viscoelasticity

2025/02/26 by Isaac Y. Miranda-Valdez, Miranda-Valdez, Isaac Y., Tero Mäkinen +5 · 1 citation
Chemical Engineering · Engineering · Physics and Astronomy · #Elasticity and Material Modeling #FOS: Physical sciences #Model Reduction and Neural Networks #Rheology and Fluid Dynamics Studies #Soft Condensed Matter (cond-mat.soft)

paper · pdf · doi:10.48550/arxiv.2502.19132

openalex publication_date 2025/02/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Inferring viscoelasticity parameters is a key challenge that often leads to non-unique solutions when fitting rheological data. In this context, we propose a machine learning approach that utilizes Bayesian optimization for parameter inference during curve-fitting processes. To fit a viscoelastic model to rheological data, the Bayesian optimization maps the parameter values to a given error function. It then exploits the mapped space to identify parameter combinations that minimize the error. We compare the Bayesian optimization results to traditional fitting routines and demonstrate that our approach finds the fitting parameters in a less or similar number of iterations. Furthermore, it also creates a "white-box" and supervised framework for parameter estimation in linear viscoelasticity modeling.

Cited by

• Analysing gelation transition through fractional viscoelasticity and Mittag-Leffler-Prabhakar function

Related

• Machine learning for viscoelastic constitutive model identification and parameterisation using Large Amplitude Oscillatory Shear
• Inverse Modeling of Viscoelasticity Materials using Physics Constrained Learning
• Deep Q-learning of global optimizer of multiply model parameters for viscoelastic imaging
• Learning rheological parameters of non-Newtonian fluids from velocimetry data
• Learning constitutive models and rheology from partial flow measurements