Interpretation of the loading-wetting behaviour of compacted soils within the 'MPK' framework. Part II: Dynamic compaction

摘要

Dynamic compaction is commonly used to construct structural fills for various geo-infrastructures. Current practice is to specify a minimum dry density and moisture content criterion to be used in the field on the basis of Proctor compaction carried out in the laboratory. However, there are still no practical methods for predicting compacted clay behaviour under expected mechanical and environmental loadings. Current theories are difficult to apply in practice due to difficulties in determining the necessary parameters. In this paper, the recently developed "void ratio - moisture ratio (volume of water / volume of solids) - net stress space" (MPK) framework is extended to cover dynamically compacted soils, with significant supporting experimental evidence. Two types of soils are used: lightly reactive kaolin and reactive Merri Creek clay. As the compaction stress was unknown for dynamic compaction, recompression of soil specimens from compacted soil was used to establish the "loading wetting state boundary surface" (LWSBS). Independent tests show that the framework can predict well the behaviour of compacted soils under loading-unloading and yielding, collapse during wetting, change of loading yield stress after wetting, and swelling pressure development during constrained wetting. The value of the approach is that the testing methods are straightforward, do not require specialized equipment, and testing times are much shorter. In addition, the uncertainty that laboratory dynamic compaction may not relate directly to field roller compaction can be addressed with the developed framework. Soil specimens obtained from field soil pads compacted by actual rollers can be used to establish the LWSBS. This information will allow prediction of the likely behaviour of field-compacted fills under expected environmental and mechanical loadings subject to one-dimensional conditions. Extension to triaxial conditions would require further experimental work and theoretical modelling.