Stress‐strain curves for polyurethane elastomers: A statistical assessment of constitutive models

Abstract Polyurethane elastomers (PUEs) are broadening applications attributed to their highly tunable mechanical properties, and the stress–strain curve is one of the most important characters. Theoretically, dozens of constitutive models have been developed to interpret the stress–strain curve for rubber‐like materials but their applicability on PUEs is still suspicious, which results in the selection of models and the initialization of model parameters need expertise experience. We performed a statistical assessment of 25 constitutive models (10 physical and 15 phenomenological) based on a comprehensive dataset with 529 stress–strain curves for PUEs. The average coefficients of determination for the whole curve can be improved from 0.676 for the phantom model (1 parameter) to 0.990 for the Bechir model (6 parameters), and the percentage of well‐fitted curves increases from 6% to 92%. Constitutive models with both the first‐ and the second‐invariant, in logarithmic or exponential expression for strain perform better.