MR23A-4332:
An Experimental Investigation into Failure and Localization Phenomena in the Extension to Shear Fracture Transition in Rock

Tuesday, 16 December 2014
Robert Charles Choens II1, Frederick M Chester2, Stephen J Bauer3 and Gregory M Flint3, (1)Texas A&M University, HOUSTON, TX, United States, (2)Texas A & M University, Geology & Geophysics, College Station, TX, United States, (3)Sandia National Laboratories, Albuquerque, NM, United States
Abstract:
Fluid-pressure assisted fracturing can produce mesh and other large, interconnected and complex networks consisting of both extension and shear fractures in various metamorphic, magmatic and tectonic systems. Presently, rock failure criteria for tensile and low-mean compressive stress conditions is poorly defined, although there is accumulating evidence that the transition from extension to shear fracture with increasing mean stress is continuous. We report on the results of experiments designed to document failure criteria, fracture mode, and localization phenomena for several rock types (sandstone, limestone, chalk and marble). Experiments were conducted in triaxial extension using a necked (dogbone) geometry to achieve mixed tension and compression stress states with local component-strain measurements in the failure region.

The failure envelope for all rock types is similar, but are poorly described using Griffith or modified Griffith (Coulomb or other) failure criteria. Notably, the mode of fracture changes systematically from pure extension to shear with increase in compressive mean stress and display a continuous change in fracture orientation with respect to principal stress axes. Differential stress and inelastic strain show a systematic increase with increasing mean stress, whereas the axial stress decreases before increasing with increasing mean stress. The stress and strain data are used to analyze elastic and plastic strains leading to failure and compare the experimental results to predictions for localization using constitutive models incorporating on bifurcation theory. Although models are able to describe the stability behavior and onset of localization qualitatively, the models are unable to predict fracture type or orientation. Constitutive models using single or multiple yield surfaces are unable to predict the experimental results, reflecting the difficulty in capturing the changing micromechanisms from extension to shear failure.

Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Deopartment of Energy’s National Security Administration under contract DE-AC04-94AL85000.

SAND2014-16578A