Work

Configurational Energy as a Microstructural Descriptor of Failure Precursors in 2D Frictional Granular Materials

Year: 2026

Type: article

Abstract: ABSTRACT Localized deformation in dense granular materials, often culminating in the formation of shear bands, is a key failure mechanism in geotechnical and material systems. However, predicting the onset of such localization remains a fundamental challenge due to the system's inherent inelasticity and microstructural complexity. In this study, we propose that the evolution of internal configuration—characterized by changes in contact topology and stored potential energy—governs the collective mechanical response and encodes precursors to material failure. To quantify this evolving internal state, we introduce the notion of configurational energy, defined as the change in contact‐scale elastic potential energy resulting from a controlled loading—unloading probe. This metric is first formulated at the contact level and subsequently analyzed at the specimen scale using Discrete Element Method (DEM) simulations of biaxial compression. Our results demonstrate that configurational energy captures the system's sensitivity to perturbations and reflects local instability: both positive and negative values emerge at the contact level, with large magnitudes concentrated near regions of active rearrangement. Despite this local variability, the specimen‐scale configurational response remains strictly negative, and its magnitude increases systematically as the material approaches failure. Notably, spatial localization of configurational energy precedes the formation of macroscopic shear bands with an evolving internal length scale, offering a mesoscale energetic signature of incipient failure. These findings establish configurational energy as a physically grounded descriptor of microstructural evolution and a promising tool for anticipating failure in frictional granular systems.

Source: International Journal for Numerical and Analytical Methods in Geomechanics

DOI: 10.1002/nag.70209

Authors: Zhennan Zhu, Francois Nicot, Antoine Wautier, Klaus Regenauer‐lieb, Manman Hu