The Use of Engineering Index Correlations with the Shrink-Swell Test

Expansive soils are a very common in many parts of the world and are often a direct cause of damage to housing, through their tendency to undergo significant volumetric change in reaction to changes in moisture content. The shrink-swell test (AS1289.7.1.1-2003) has become a widely adopted laboratory test in Australia for estimating the degree of reactivity of an in-situ soil due to its simplicity, low cost and ability to generate relatively reliable data which is independent of soil suction and initial water content. Anecdotal reports indicate that it has become increasingly common to use engineering index properties such as the Liquid Limit (LL), Plasticity Index (PI) and Linear Shrinkage (LS) in place of a shrink-swell test, with the results being correlated to a Shrink-Swell Index (Iss) value, using a variety of published correlations. This is typically occurring both in response to cost and time pressures placed on the geotechnical consultant, as well as in certain situations where an undisturbed sample (required for the shrink-swell test) is not able to be obtained in the field. However, the use of such correlations without regard for the context in which they were developed, together with conservative interpretation, can have a significant effect on the resulting site classification and construction cost. This paper examines and discusses the use of these correlations in Australian geotechnical consulting practice within the context of site classification to AS2870-2011. It is demonstrated that using a non- geologically specific correlation between Iss and engineering index in the calculation of ground surface movement and subsequent site classification is not able to be achieved without an unacceptable likelihood of significant error and adverse cost implications.