Abstract

The characterisation of the properties of porous materials is of great importance in the effective management of natural and manmade systems. A sophisticated network model, 'Pore-Cor', of some of these properties has been previously developed. The present study has significantly extended the scope of the model's predictive capabilities. Flow and transport behaviour was examined in laboratory sand columns of various depths. These experiments examined unsaturated flow of water and conservative solute tracer transport through homogeneous sand samples. Flow through these was not homogeneous or repeatable. Experimental observations found that this may have been due to subtle random variations in packing, and the network model was shown to be able to simulate these. Solute transport of bromide was studied, applied both uniformly and from a point source. Both scenarios were modelled using a convection-dispersion equation, and it was demonstrated that the lateral component of such transport was highly significant. It was shown how convection-dispersion equation predictions of uniformly applied tracer transport might be improved by the application of the network model and a method for improving predicted lateral solute transport was outlined. It has been shown that levels of correlation in the distribution of differently sized voids within porous material may be responsible for large variations in permeability. This can make accurate modelling of permeability very difficult. A technique was developed for assessing the degree and nature of such correlations. The new method was tested on a variety of artificial and real samples and demonstrated to provide a quantitative assessment of such correlations. A method by which this could be used to improve network model simulations of materials possessing such correlation was described.

Document Type

Thesis

Publication Date

1999

Download

Share

COinS