Can Grains of Sand & Rock Predict Fault Line Instability?

Predicting how granular media — rocks and soils — move in fault lines could be crucial for everything from geophysics to construction engineering. 

Existing models for predicting these interactions don't explicitly account for particle-scale interactions. This can lead to inaccurate predictions of shear strength and instability, making it difficult to predict when a fault gouge will yield and other aspects of fault line behavior. 

Recent machine learning models have attempted to predict fault line behavior, but they aren't ideal as the many steps and processes they use internally remain opaque to outside observers — their "black box" nature obscures how they reach certain conclusions. In a new paper by Exponent mechanical engineer Adyota Gupta and collaborators from Johns Hopkins University, the team developed a new two-dimensional, physics-based model that explicitly accounts for the forces between grains. 

The authors used this analytical, dynamic force chain model to understand how force chain stress states and kinematics give rise to instabilities, leading to drops in stress and the flow of energy between force networks. They examined the flow of energy at contacts between the strong and weak force networks of a granular material to understand how this flow of energy corresponds with macroscopic stress fluctuations. Based on their results, the authors offer insights into how their new model can help predict fault line behavior.

From the publication: "This work is the first to emphasize that one can analytically show, in the context of the model presented in this paper, that stress fluctuations stem from the competition between (1) the rate of change in lateral support of force chains and (2) the rate of lateral deformation of force chain segments."