Experimental study and constitutive modelling of the stress-strain behaviour of Hong Kong soils

Abstract

The findings about stress-strain behaviour of granitic saprolite as reported in published literature were not only limited but also inconsistent. The inconsistency might be due to (i) the complexity and variability in behaviour of granitic saprolite and/or (ii) the fragile characteristics of granitic saprolite and the lack of a standardised triaxial-test specimen formation methodology that can effectively minimise disturbance induced during specimen formation. The inconsistent findings also impede a systematic study and further investigation such as constitutive modelling. In this study, an essentially operator independent methodology that can effectively minimise the disturbance during specimen formation is developed for forming a triaxial specimen of Hong Kong granitic saprolite from a Mazier tube sample. Experimental studies on Hong Kong Completely Decomposed Granite were conducted based on a series of triaxial drained tests using intact Mazier samples obtained from different sites. The isotropic compression responses and stress-strain behaviour including small-strain stiffness were investigated in detail. Triaxial tests were also conducted on reconstituted specimens for comparison. The reconstituted specimens were formed and tested with the same dry densities and testing conditions of the corresponding intact specimens. Differences were consistently observed between the behaviour of the intact and reconstituted specimens. One of the differences is that dilatancy behaviour was found consistently in the intact specimens. In particular, the dilatancy behaviour was still observed in one conventional triaxial drained test with the applied confining stress of four times more than the experienced in-situ stress. The new and consistent observation not only unambiguously demonstrates the existence of structure preserved in the intact specimens, but also reflects the success of the proposed triaxial-specimen formation methodology in minimising the disturbance during the formation process. The consistent experimental results obtained in this study allow further study about constitutive modelling of granitic saprolite. Hardening Soil model with small-strain stiffness (HSS model) was evaluated. It is found that this soil model is able to reproduce most of the essential characteristics of Hong Kong granitic saprolite CDG. The model parameters can be objectively determined through triaxial tests only. A method is proposed to infer the maximum shear modulus without the need of direct laboratory measurements such as bender-element or resonant-column measurements. Good quality reproduction performance is observed through the comparison between the simulation results of HSS model without calibration and the experimental results. Finite-element (FE) analyses with the applications of Hardening Soil model (HS model) and HSS model are presented to show the significance of incorporating the small-strain stiffness and cap-yield surface in a constitutive model for FE settlement analysis of a footing on Hong Kong granitic saprolite. A series of triaxial tests was also conducted for investigating the small-strain stiffness of Hong Kong Alluvium, and intact Mazier samples were used. The samples were obtained from the same site, and the alluvium strata were complex and with high variability that at least three difference types of alluvium were found. The specimen preparation methodology developed for Hong Kong granitic saprolite is found to be applicable to Hong Kong Alluvium. The method developed for granitic saprolite for inferring the maximum shear modulus is also shown to be applicable to Hong Kong Alluvium as well.