Characterization and mesoscale modeling of an enhanced UHMWPE fabric treated with bis-diazirine: multicriteria crosslinker selection and surrogate-based inverse parameter estimation

TL;DR

This study improves UHMWPE fabric for armor by selecting the best crosslinker and using a model to predict its mechanical behavior.

Contribution

A novel approach combining multicriteria decision-making and surrogate-based inverse modeling for UHMWPE fabric characterization.

Findings

The optimal crosslinker increased yarn interactions by over 100%.

A mesoscale model accurately predicted tensile and shear responses of treated and untreated fabrics.

Inverse analysis captured difficult-to-measure properties like yarn modulus and friction.

Abstract

Ultra-high molecular weight polyethylene (UHMWPE) woven fabrics are commonly used in armor applications due to their superior biaxial mechanical and physical properties. In this study, three different diazirine-based crosslinker options were initially considered as the chemical treatment applied to a dry UHMWPE plain weave to improve a range of its properties. The optimum crosslinker was then selected using a VICOR multicriteria decision-making model. Specifically, through the bias-extension and yarn pull-out tests, it was observed that the optimum crosslinker significantly enhanced (> 100%) the crossover interactions between the warp/weft yarns. Subsequently, a mesoscale finite element model was developed to predict both the tensile and shear responses of the untreated and treated fabrics. In developing this model, an inverse analysis was employed to capture the effect of yarn transverse elastic modulus and the friction at the crossovers—two properties that are known to be difficult to measure directly in the weave form of yarns. These parameters were sampled using a design of computational experiments and then optimized via a surrogate-based model. Finally, challenges presented by the crosslinking at the single yarn level during characterization are discussed and resolved numerically. For both the treated and untreated fabrics, the mesoscale model is shown to predict the material behavior accurately.

Figures