Degree Name

Doctor of Philosophy


Department of Civil, Mining and Environmental Engineering


In Australia and worldwide, highly erodible and dispersive soils have been reported as a major problem initiating failure of earth structures such as embankment dams, rail/road embankments, canal banks and foundations due to surface and internal erosion (piping). Adopting a suitable ground improvement technique to control soil erosion is necessary to avoid property damage and high maintenance costs caused by the erodible and/or dispersive soils. To eliminate or reduce erosion and internal piping, chemical stabilisation has been proven to be an appropriate and cost effective technique worldwide. However, the traditional soil stabilisers such as cement, lime, fly ash, slag and gypsum have been identified to cause serious environmental problems (e.g. changing the pH of soil and ground water, thus negative impact on agriculture and aquaculture etc.). also these soils tend to exhibit excessive brittleness (post-peak) that has clear implications on the stability of infrastructure, especially during cyclic and impact loading conditions prevailing in high speed rail and aircraft runways. In this context, lignin-based chemicals such as lignosulfonate (LS) have shown promising potential in stabilising erodible and dispersive soils.

In the recent past, experimental investigations have been carried out to investigate the effectiveness of LS treatment to verify that LS can enhance the erosion resistance of various soils. However, currently there is no rational (theoretical) erosion model which correlates the erosion rate of lignosulfonate treated soils with commonly adopted soil properties that can be readily determined in the laboratory. Furthermore, it is important to capture the shear strength and volume change behaviour of lignosulfonate treated soil through an advanced constitutive model.

In this research, a novel theoretical model was developed to calculate the rate of internal erosion of both lignosulfonate treated and untreated soil. The model is based on the law of conservation of energy, and the effects lignosulfonate stabilisation were captured using the increased strain energy required to break the inter-particle bonds. The model correlates the erosion resistance of LS treated soil with the basic soil properties such as the mean particle diameter and specific gravity, the shear strength of soil, and the flow characteristics of the eroding fluid.

An important phase of this research included laboratory investigations required to validate the proposed erosion model. A series of direct shear tests were conducted on a highly erodible silty sand to obtain the stress-strain-volumetric responses of both LS-treated and untreated soil. Effective normal stresses varying from 5kPa - 42kPa and lignosulfonate dosages varying from 0.2% - 1.2% by dry soil weight were considered. The laboratory shear tests indicated that the peak and ultimate shear strength, as well as the angle of internal friction increased with the increasing amount of lignosulfonate. The volume change behaviour showed a dilative response (i.e. less compressive) after the lignosulfonate treatment. However, the enhancement in the elastic deformation (secant) modulus was not significant, and also the change in soil ductility caused by lignosulfonate treatment was marginal compared to the brittleness attributed to cement treatment.

Laboratory results were used to validate the proposed erosion model, whereby the model parameters were empirically determined from the experimental data. An independent set of erosion test data was used to compare with the model predictions. These comparisons proved that the proposed erosion model could accurately capture the erosion behaviour of lignosulfonate treated and untreated silty sand.

As the final phase of the study, in order to predict the stress-strain and volume change behaviour of the LS treated soil, a constitutive model was formulated based on the Disturbed State Concept (DSC). The relative intact (RI) responses of both treated and untreated soil were modelled incorporating the δ0 version of the HiSS plasticity models, whereby the non-associated yielding was considered through the disturbance function. The response of lignosulfonate bonds was model separately using the same concept (DSC) and that response was used as the fully adjusted (FA) response in modelling the treated soil behaviour. The laboratory shear test results corresponding to the effective normal stresses of 10kPa and 22kPa were adopted for calibrating the material parameters representing acceptable accuracy of this DSCbased model.