A Novel Approach for Optimization of Soft Material Constitutive Model Parameters Based on a Genetic Algorithm and Drucker's Stability Criterion

The growing interest in soft materials to develop flexible devices involves the need to create accurate methodologies to determine parameter values of constitutive models to improve their modeling. In this work, a novel approach for the optimization of constitutive model parameters is presented, which consists of using a genetic algorithm (GA) to obtain a set of solutions from data of uniaxial tensile tests, which are later used to simulate the mechanical test using finite element analysis (FEA) software to find an optimal solution considering Drucker's stability criterion. This approach was applied to the elastomer Ecoflex 00-30 considering the Warner and Yeoh models and Rivlin's phenomenological theory. The correlation between the experimental and the predicted data by the models was determined using the root mean squared error (RMSE), where the found parameter sets provided a close fit to the experimental data with RMSE values of 0.022 (ANSYS) and 0.024 (ABAQUS) for Warner's model, while for Yeoh's model were 0.014 (ANSYS) and 0.012 (ABAQUS). It was found that the best parameter values accurately follow the experimental material behavior using FEA. The proposed GA not only optimizes the material parameters but also has a high reproducibility level with average RMSE values of 0.024 for Warner's model and 0.009 for Yeoh's model, fulfilling Drucker's stability criterion.

Automated summary

In this study, a novel approach using a genetic algorithm (GA) is presented to optimize the parameters of constitutive models for soft materials. The GA is applied to obtain a set of solutions from uniaxial tensile test data, which are then used to simulate mechanical tests using finite element analysis (FEA), resulting in an optimal solution that satisfies Drucker's stability criterion. The proposed GA not only accurately predicts experimental data but also demonstrates high reproducibility.