An experimental study of the size effect on core shrinkage behaviour of reactive soils

This study experimentally investigates the effect of specimen size on the measured axial and radial strains in core shrinkage tests, which are generally used in designing lightweight structures on reactive soils. Remoulded core specimens were prepared in the laboratory using a reactive soil of newer volcanic basaltic origin obtained from near Melbourne, Australia. One group of core specimens was prepared with 15 mm, 38 mm, 50 mm, 63.5 mm and 90 mm diameters and different lengths to produce a length-to-diameter (L/D) ratio of 1.75 for each case of diameter. The other group of specimens was 50 mm in diameter, and the length was varied to yield L/D ratios of 1, 1.5, 1.75, 2, 2.5 and 3. Both axial and volumetric strains showed a marked dependency on specimen diameter, with larger specimens showing lower strains. This behaviour was attributed to the effect of the shrinkage cracks developed in specimens that were more abundant in larger specimens. In contrast, neither systematic nor significant variations of axial or volumetric strain were observed for varying L/D ratios of specimens. This was understood to be an effect of shrinkage cracks, which were smaller in size and consistent across different L/D ratios. It is recommended to keep the drying rate as low as possible in core shrinkage tests to control moisture gradients within the specimen and minimise surface cracking, as observed in the 90 mm diameter specimens.