Abstract:
In the last few decades the construction of small dwellings, particularly government subsidy houses, has increased markedly in South Africa. Large numbers of these light structures have suffered severe damage and many have been demolished after being occupied for only a small fraction of their expected life-span. The research detailed in this thesis deals with a search for reasons for these widespread failures and investigations into economically feasible solutions. The investigation began with an assessment of the current methods of identification and assessment of potentially problematic expansive soils. This revealed serious shortcomings in a number of the standard testing procedures. These shortcomings may have consequences for a wider range of engineering projects than just housing development. Progress towards more reliable testing methods has led to an appreciation of the importance of variability of soil properties and the necessity of taking this into account in engineering design. Investigation of clay fraction determination by the hydrometer has led to the realization that the unreliability of this procedure has been a cause of failure in subsidy housing projects as well as other light structure developments. Assessment of expansive potential cannot safely rely on hydrometer determination of clay fraction, which is a principal parameter for a number of popular design procedures. It was also established that attractive alternatives in the form of automatic particle sizing equipment (designed for the analysis of industrial powders) are not suitable for natural soils containing the clays commonly found in South African soils. Alternative procedures, particularly small-scale suction testing, have been found to provide more reliable insights into design for expansive clays. The investigations suggest that some form of Reliability Based Design, which takes account of specific variability in soil properties, may be the only way to achieve rational design for foundations on
expansive clay.
