Strength and Stiffness of Undisturbed and Reconstituted Fluvial Sand: The Role of Initial Fabric

While it has been known for decades that the initial fabric of soils affects their mechanical behavior, this fact is often overlooked in engineering practice. Samples of sands are often reconstituted in the laboratory without rigorous consideration of the history and depositional environment of the in situ deposit. In this research, high-quality undisturbed samples of a fluvial sand were obtained by ground freezing to assure that their initial fabric remains intact. In the laboratory, undisturbed specimens were carefully prepared and sheared monotonically under drained and undrained conditions in a triaxial apparatus. The same undisturbed specimens were then reconstituted at identical void ratios using different reconstitution methods: dry deposition, moist tamping, water spooning, and slurry deposition. Each group of undisturbed and reconstituted specimens was tested under identical testing conditions. Four distinct groups of undisturbed and reconstituted specimens produced a consistent data set that enabled objective assessment of which the reconstitution method best replicated the triaxial response of the undisturbed fluvial sand. Results from this unique experimental study indicate that slurry deposition is the method that captures most closely the response of the undisturbed sand, and, conversely, moist tamping shows the greatest discrepancy. This highlights the importance of replicating as well as possible the depositional environment of an element of soil when reconstituting specimens in the laboratory. The presented results have major implications for testing and analysis of liquefaction of fluvial sand and provide insights into the behavior of offshore and tailings sands.