Simulation of non-conservative transport using particle tracking methods with an application to soils contaminated with heavy metals.

Abstract

This thesis focuses on the development and application of a discrete time random walk particle tracking model to the simulation of non-conservative transport in porous media. The model includes the simulation of solute transport, reversible bimolecular reactions, and sorption. The functionality of the discrete time random walk method is expanded to allow for the simulation of more complicated chemical systems than previously achieved. The bimolecular reaction simulation is based on a colocation probability function method. This reaction simulation method is analysed to investigate the effects of the controlling parameters on its behaviour. This knowledge is then used to inform a discussion of its application to the simulation of mixing limited reactive transport and comparison with other approaches. The reaction simulation method developed in the thesis possesses a greater flexibility than previously developed methods for the simulation of reactions using particle tracking.

The developed model is also applied, in combination with a chemical speciation model, to enable the production of a reduced complexity model to simulate effects of an amendment
scheme on soils contaminated with heavy metals. The effect of the soil amendment scheme on the partitioning of Pb between solution, soil surfaces, and dissolved organic matter is approximated by rules fitted as functions of concentrations of single components within the soil amendment. This allows for the simulation of complicated chemical systems using particle tracking methods.

As well as expanding the functionality of particle tracking methods the issue of the computational expense is also addressed. A scheme for the optimization of the reaction simulation is presented and its effectiveness investigated. Together with the use of graphics processing units for code acceleration, the computational and temporal expense of the solution is reduced. The combination of the expansion in functionality and reduction in run time makes particle tracking a more attractive simulation method.