Enhanced dynamic homogenization of hexagonally packed granular materials with elastic interfaces

TL;DR

This paper introduces an advanced dynamic homogenization method for hexagonally packed granular materials, improving the accuracy of continuum models in capturing both static and dynamic behaviors, including optical and acoustic branches.

Contribution

It develops a high-frequency dynamic homogenization scheme using enhanced continualization, resulting in non-local micropolar continua that better simulate granular materials' behavior.

Findings

Accurately models optical branches of the discrete system.

Addresses instability issues in classical models.

Shows convergence of continuum response to discrete system with higher order models.

Abstract

It is well known that the classical energetically consistent micropolar model has limits in simulating the frequency band structure of packed granular materials (see Merkel et al., 2011). It is here shown that if a standard continualization of the difference equation of motion of the discrete model is carried out, an equivalent micropolar continuum is obtained which is able to accurately simulate the optical branches of the discrete model. Nevertheless, this homogenized continuum presents non-positive defined elastic potential energy, a deficiency that limits the reliability of the model and implies instability phenomena (destabilizing effects) in the acoustic branches. This drawback is circumvented here through an high-frequency dynamic homogenization scheme which is based on an enhanced continualization of the discrete governing equations into pseudo-differential equations. Through a formal Taylor expansion of the pseudo-differential operators a higher order differential equation corresponding to the governing equation of a non-local continuum thermodynamically consistent are obtained. The resulting approach allows obtaining an equivalent micropolar continuum characterized by inertial non-locality. Moreover, higher order continua with non-local constitutive and inertial terms may be derived. The proposed continuum models are proved to be able to accurately describe both the static and dynamic behavior of the discrete granular model. Finally, the convergence to the response of the discrete system is shown when increasing the order of the higher order continuum.