The University of Queensland

Research Projects for RHD Students

Research projects for RHD students are available here.

Computational Geomechanics -- Research

Geomechanics

Geomechanics encompasses the disciplines of Soil Mechanics, Rock Mechanics and Engineering Geology required for a rational approach to the mechanics of earth structures including man-modified and natural landscapes. Computational Geomechanics allows the realisation of a rational approach to real practical problems. Geomechanics deals with surface and underground structures. Applications include building foundations, dams, embankments, earthen slopes, and road pavements. Some critical problems related to disasters such as landslides and debris flows due to tsunamis can rationally be analysed using Geomechanics principles.

Methods

The geomaterials under consideration are a highly complex mixture of solid particles and fluids. This mixture cannot be treated simply as a solid or as a fluid! Therefore, innovative approaches, employing complex mathematics and algorithms, must be developed. Large-scale computer resources are required to solve and predict these problems. The concept of parallel computing, where hundreds or even thousands of processors are connected and employed at the same time for calculations, will be adopted. Leading edge technologies such as graphics cards multi-processor computing are also being considered. We will put effort onto solving very large-scale problems using large-scale computer clusters and powerful computing techniques.

Aims

In addition to conventional methods (e.g. finite element methods, FEM), our research group extends and develops new methods for the prediction of the mechanical behaviour of (porous) solids and granular materials, including (where desired) multi-component fluid flow. Continuum and discrete approaches are considered. The computer software, named MechSys (Open Library for Mechanical Systems http://mechsys.nongnu.org) serves as a basis for the development of the models and solvers. MechSys includes also discrete element methods, smoothed-particle hydrodynamics (SPH) and the lattice Boltzmann method (LBM). MechSys is under constant improvement in order to deal with disaster-like scenarios.

Our group is also engaged on Engineering Optimisation, in particular employing advanced new optimisation techniques such as Genetic Algorithms. In this front, the computer library, named SGA (Simple Genetic Algorithms http://sga.nongnu.org) serves as a basis for the development of optimisation tools

Due to the complexity and generality of our methods, many other important problems in Engineering can also be solved. For instance, the following can potentially be simulated by our methods: impact of comets or projectiles on the ground or underground bunkers, armour behaviour against projectile penetration, cracking and fracturing of solids and structures, mining materials processing, and many others. Accuracy is one of our main concerns, in light of the fact that reliability is a key constraint posed by Civil Engineering.

Some examples are given below:

Fig. 1 - MechSys/DEM: discrete element simulations of particles of quasi-general (complex) shapes.

Quasi-general particles, including particles of non-convex shapes

Solid (cohesive) block of 'Voronoi' grains

Granular packing of quasi-general grains

Fracture of a solid disk using DEM

Fig. 2 - MechSys/LBM: study of the microscopic aspects of flow through porous media. Colours show velocity intensity: red meaning higher intensity. It is interesting to note how tortuosity may rule the flow showing that conventional methods, such as Darcy's equation, may not apply to such heterogeneous packing.

Fig. 3 - MechSys/LBM: microscopic simulation of viscous liquid flow under gravity with an obstacle.

Fig. 4 - MechSys/LBM: microscopic simulation of multi-component (liquid, its own vapour, and gas) flow. The initial bubble (blue) expands, moves to the surface, and then pops up, while the liquid (red) settles downward.