AbstractThis thesis examines two available approaches for predicting the swelling behaviour of unsaturated expansive clays:
(i) simplified empirical relationships for rapidly identifying swelling soils and
(ii) a more accurate prediction method suitable for design purposes and based upon experimentally measured swell moduli.
The work was prompted by the absence of a universally accepted procedure for use with these soils.
The initial (and shorter) phase of the work uses published swell data to assess available empirical swell prediction methods and also determine the relative significance of soil parameters used for this purpose.
The work suggests that potentially expansive soils can be identified and classified, and the national or regional soil classification systems may be easily amended to allow for this.
The results indicated:
(i) the swell prediction methods are suitable for indicating the degree of swell behaviour only, and methods for actually quantifying volume change are usually limited to localised application
(ii) in situ suction correlated closely with swell for available data, therefore it is suggested that the suction-swell relationship be derived for numerous soils in order to assess the possibility of using it as a stand-alone indicator.
The second and most significant phase of the work involved an assessment of Fredlund's constitutive model approach for unsaturated soils when applied to the swell prediction of expansive clays. This required considerable equipment development for measuring volume changes and controlling the stresses of the soil, air and water phases. In addition, two conventional consolidometers (oedometers) were linked to a chart recorder to facilitate continuous data acquisition.
An experimental programme of three test series was then instigated to provide data by which Fredlund's theory could be assessed; these included Null tests to assess the validity of the chosen stress state variables, Volume change (uniqueness tests)to evaluate Fredlund's constitutive equations and finally, Consolidometer swell tests to further evaluate the constitutive equations and assess an established soils laboratory test for swell prediction purposes.
The results showed:
(i) in general, Fredlund’s constitutive model approach is suitable for predicting volume changes in unsaturated expansive clays
of a liquid limit range typically found in the field (up to 142%).
(ii) despite equipment limitations, equal changes of the stress components (cr, Ua and Uw) produced negligible volume changes, thereby confirming the stress state variables.
(iii) both the stress control and consolidometer tests indicated reasonable agreement between measured and predicted swell values, thus confirming the suitability of Fredlund’s constitutive model approach to expansive unsaturated clays.
(iv) the theory makes no provision for (a) discontinuous phases in soils found at saturation levels below 25% and above 80-85% and
(b) the non elastic behaviour of unsaturated soils following a reversal in stress change.
Fredlund’s theory should therefore be further studied for soils composed of discontinuous phases and the effects of stress reversals upon volume change. The development of a full triaxial testing system is recommended.
|Date of Award||Apr 1988|