Date of Award

Spring 5-2023

Degree Type

Thesis

Degree Name

MS Geological Engineering

Department

Geological Engineering

Committee Chair

Mary MacLaughlin

First Advisor

Lorne Arnold

Second Advisor

Nathan Huft

Abstract

The ability to monitor and predict the evolution of rock damage and fracture is key for creating reliable and sustainable civil infrastructure. Rock crack propagation is a complex process that is dependent on material properties such as tensile strength, compressive strength, density, elastic modulus, Poisson's ratio, friction coefficient, cohesion, and specimen conditions such as preexisting microcracks, particle size distribution, and mineral distribution. In conjunction with the 56th US Rock Mechanics/Geomechanics Symposium, researchers primarily at Purdue University organized the Damage Mechanics Challenge to determine the current state and future direction of the rock modeling community’s ability to simulate rock crack formation. In this thesis, the ability of the Particle Flow Code 3D (PFC3D) distinct element method software and the Irazu3D hybrid finite-discrete element method software to accurately recreate cracks propagated in rock specimens was analyzed. Simulations were constructed based on the laboratory data from three-point bending, unconfined compressive strength, and Brazilian tests conducted on 3D-printed blocks provided by Purdue University. The constraints of using computing power that is comparable to an average engineering industry setting, limiting the use of complex coding, and restricting the implementation of additional post processing software were applied to gain more targeted insight into the current state of the modeling community. “Average computing power” was defined as a commercially available desktop containing a single graphics card with no cloud computing or supercomputing capabilities. The two software packages chosen for this study (PFC3D and Irazu3D) were selected based on their ability to model fracture propagation; the usability and computational performance were compared in addition to quantitative model behavior as evaluated using the calibration laboratory datasets. The time spent constructing, refining, and interpreting models was recorded III for comparison, as well. Simulations of three models (unconfined compressive strength, Brazilian, and three-point bending tests) were built and calibrated using the two software packages to predict the fracturing behavior of a fourth challenge geometry. Two types of models were developed using PFC3D: traditional ball models and rigid body polygon models constructed using the new “rblock” feature. All the PFC3D and Irazu3D models were successfully calibrated. The challenge simulation was very similar for these three models. The specimens all exhibited failure due to the formation of a single fracture oriented from the notch location to the center rod. The simulation’s force-displacement curves had good agreement with the physical experiment. The PFC3D models exhibited the exact behavior and were within the range of values of the results. The Irazu3D model showed less agreement with the results because it had a lower peak strength, stronger residual strength, and represented a more ductile behavior. The results were all within the same order of magnitude and could be used to approximately predict the challenge block’s behavior. Based on the force-displacement results, both Irazu3D and PFC3D were determined to be effective methods for simulating fracturing in rock blocks. The fracture texturing of the rock blocks will need further evaluation to better determine the accuracy of the simulation’s predictions.

Comments

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Geological Engineering

Download

Share

COinS