Backfilling Mine and Tunnel Structures with Glass Waste Slurries

Project description

A typical mining process (in particular for coal mining) involves the removal and recovery of economically valuable minerals from the earth’s crust. The resulting excavation is then revived by a process referred to as backfilling. In this process, a cementitious slurry/grout is pumped through cased drill holes directly into underground cavities to fill them, thereby stabilising the area from collapsing. Grout/slurry backfilling has also been successfully used to (i) stabilise collapsing in abandoned mines, (ii) fill the annular gap between the outside of a segmental lining and the excavated surface of the ground in tunnel boring (TBM) operations, and (iii) prevent surface subsidence in high-use areas. Therefore, the backfill can perform both as a support system and a working platform with the requirement of its mechanical properties governed by the different roles mentioned earlier.

Backfilling grouts and slurries consist principally of water and high proportions of sand. In recent years, backfilling supply has therefore suffered from a significant rise in the overall cost mainly due to steep hikes in the price of sand. This price rise can be attributed to the widespread use of sand in today’s blended aggregate industries (concrete, shotcrete, pavement) that has made sand the second-most consumed resource on earth (after freshwater) and the most important solid substance on the planet. In addition to the increased cost of backfilling operations, such unsustainable exhaustion of sand has damaged aquatic habitats, caused beach erosion, made coastal communities vulnerable to floods, and has impacted biodiversity on land. Recycling local wastes to replace the rapidly reducing natural and quarried sand is therefore becoming of urgent interest worldwide. Among multiple types of waste, glass is a perfect cost-effective candidate to replace sand as it is theoretically 100% re-useable, does not degrade, can be recycled over and over again without any reduction in quality, is predominantly composed of silica (SiO2), a key mineral found in natural sand, and it shares similar chemical composition to natural sand. However, it also has different thermal conductivity, abrasivity, permeability, and hydro-mechanical parameters to that of sand and rock. Hence, it is expected that if used in backfilling production, its different mechanical parameters could affect the cracking/fracturing behaviour, modulus, deformation resistance, fatigue, and ultimate fracture strength in backfill grouts and slurries.

While experiments can provide a detailed and reliable characterization of mechanical properties, performing a large number of experiments is prohibitive. To this extent, computational modelling can be used as a surrogate to accelerate the design of the slurry system. Modelling the slurry system using computational methods presents its own challenges. Due to the complex nature and the thermodynamics associated with the dissolution-precipitation process of cement setting, the constitutive model of the system varies with time. The cementitious slurry has a continuously evolving constitutive relationship, ranging from viscoplastic and viscoelastic in the early stages of hydration to brittle elastic in the later stages after the setting of the slurry. While several methods, including finite element and discrete element methods, have been used to study these systems, calibrating the model requires many experimental measurements. To address these challenges, this study aims to also develop a physics-informed machine learning approach to directly learn the constitutive relationship of the slurry as a function of time from the experimental data. The model will be based on a neural operator, which will be trained to learn the constitutive relationship while respecting the fundamental laws, including conservation of energy and linear momentum.

Through high-fidelity experiments, involving high-stress slurry consolidometer and true triaxial testing, coupled with and machine-learning modelling, the ambitious goal of this study is to understand the key processes controlling the behaviour of crushed waste glass used in backfill slurries, from macro to granular scale a topic which has been largely understudied in the literature.

Outcomes

The proposed project consists of the following overlapping stages:

i. Mineralogical and image-based shape analyses, X-ray fluorescence spectroscopy, optical microscopy, and geotechnical testing (including sieve analysis, specific gravity, minimum & maximum dry density, permeability, abrasion loss, and direct shear strength testing) of natural aggregates commonly used in backfilling grout/slurries as well as glass waste products available in the Australian and Indian markets;

ii. Experimental (physical modelling) of backfilling slurries with waste glass aggregates as a viscous, cohesive, non-Newtonian radial slurry using the world-class true triaxial soil and rock testing facilities at UQ Civil;

iii. A physics-informed machine learning model for modelling the mechanics of cementitious slurry with glass aggregates for accelerated discovery of optimal cementitious grout;

iv. Publishing of the ideas and outcomes of the project in high-impact journals and conferences.

Essential capabilities

Knowledge of Geotechnical and/or Structural and/or Mechanical Engineering with academic achievement.

Desireable capabilities

Experience in experimental testing and machine learning basics. In case the student does not have the relevant experience, the student will be encouraged to undertake additional studies in this field at UQ and/or IIT Delhi.

Expected qualifications (Course/Degrees etc.)

Master’s Degree in Geotechnical, Civil, or Computer Science or a related discipline.

Candidate Discipline

Tunnel and Mine Backfilling Crushed Waste Glass Artificial Intelligent & Machine Learning Fracture Mechanics.