ORIGINAL PAPER Extension of a Constitutive Model for Sands to Account for the Sample Preparation Method William Fuentes 1 • Carlos Lascarro 1 Received: 15 August 2017 / Accepted: 2 June 2018 / Published online: 18 June 2018 Ó Indian Geotechnical Society 2018 Abstract In this article, an existing constitutive model for sands is extended to account for the fabric effect arising from the sample preparation method. The ISA constitutive model, proposed by Fuentes and Triantafyllidis (Int J Geomech 39:1235–1254, 2015), serves as reference model to be extended. The proposed extension modifies the for- mulation of the characteristic void ratios, namely the maximum and critical void ratios. It features the conser- vation of a unique critical state line under large deviator strain amplitudes in order to be consistent with other works. The model performance is evaluated with some element test simulations of samples with different prepa- ration methods. It also includes the simulation of a scaled foundation test on a sand deposited by aerial discharge. The simulations showed that many effects related with the material fabric are captured with the proposed formulation. Keywords Intergranular  Strain  Anisotropy  Hypoplasticity  Preparation method Introduction It is well known that the deposition of a granular soil induces a natural fabric affecting its strength and stiffness characteris- tics. A different deposition method may result in different particle contact distributions which influences the overall mechanical behavior of the sand. Beside this, the compression of the material under a gravitational field creates a horizontal bedding plane perpendicular to the gravity direction respon- sible of the fabric anisotropy [3, 4, 29, 31, 33, 45, 47]. These effects explain why some samples exhibit a very different compressive-dilational behavior for each preparation method [4, 14, 19, 23, 32, 32, 38]. To mention some examples, it is well known that moist tamped samples are less contractive and show a higher anisotropy than air pluviated samples [25]. Hence, many authors have proposed to represent the material fabric through some enhanced formulations [16, 29, 30, 32] aiming to be accounted by constitutive models and other applications, see e.g. [4, 12, 21, 39]. Of course, the mathe- matics behind these formulations may be complicated and their calibration are time consuming. The literature have shown different ways to capture the material fabric. When its mechanical behavior presents a considerable variation depending on the sample prepara- tion method, independent of the loading direction and the bedding plane, then a scalar function representing its iso- tropic fabric would be the most appropriated method [24, 26]. This method has proved to perform well on some sands whereby the effect of the bedding plane is small compared to the isotropic fabric arising from each prepa- ration method. The common belief that the isotropic component of the fabric is directly related to the void ratio is debatable, considering the experience of some existing models intending to capture this behavior. To give some examples, the model of Papadimitriou et al. [34] consid- ered a scalar variable in their hardening modulus which must be calibrated for a specific preparation method. The model of Yang et al. [45] defined a scalar parameter adjusted to each preparation method as well. Additional examples of models which had already included the void ratio dependency, and required an additional variable to capture the effect of the sample preparation method can be & William Fuentes fuentesw@uninorte.edu.co 1 Department of Civil and Environmental Engineering, Universidad del Norte, Km.5 Va Puerto Colombia, Barranquilla, Colombia 123 Indian Geotech J (August 2019) 49(4):381–397 https://doi.org/10.1007/s40098-018-0315-5 Journal: Indian Geotechnical Journal Year: 2019