Liquefaction Modelling using the PM4Sand Soil Constitutive Model in PLAXIS 2D

Abstract

The phenomenon of earthquake-induced soil liquefaction has been a major issue among geotechnical engineers mainly due to the dramatic consequences this could have on civil structures. Basically, earthquakes propagate shear waves generating excess pore pressures and thus weakening the effective shear resistance of granular soils to their minimum until they liquefy behaving like a viscous fluid. This occurs because the soil experiences undrained behaviour against these rapid cyclic loads. In practice, the cyclic shear resistance of the soil against earthquakes is assessed by means of empirical correlations from in-situ penetration tests and cyclic laboratory tests. However, the liquefaction phenomena analysis is still a topic under research in which soil constitutive models play an important role.
The PM4Sand is an advanced soil constitutive model that has been developed to simulate soil liquefaction behaviour of granular soils by defining mainly three model parameters being an easy calibration model and therefore very attractive for the industry. The current project aims, firstly, to verify the PM4Sand model response at soil element level and secondly, to validate its use for quay wall structures design using the finite element methods software, PLAXIS 2D.
During the first phase of the project, a comparison between the PM4Sand model response and documented cyclic DSS tests documented by Sriskandakumar (2004) is performed. In this, a parametric assessment identifies the influence of the model parameters on the model response, allowing also to evaluate the original calibration methodology proposed by Boulanger and Ziotopoulou (2017). Consequently, initial state conditions are evaluated. It was observed that a proper calibration of the PM4Sand model provides satisfactory response both in terms of stress paths and generation of excess pore pressure, even though the model tends to overestimate the cyclic resistance of the soil at higher cyclic stress levels and to underestimate this at lower levels with respect to a target ‘CRR vs Nc’ relation. Moreover, static shear stress effect is not well captured by the model but this is still under discussion as this effect is not fully understood yet.
In the next phase, modelling of the case study is developed based on research carried out by Iai and Kameoka (1993). The PM4Sand model is calibrated based on the representative SPT tests at the site to then be implemented on the upper liquefiable soil layers. The dynamic analysis of the collapsed quay wall was applied using different approaches: dynamic analysis with and with consolidation effect, and using free-field and tied-degree of freedom lateral boundaries. The results showed that the PM4Sand model is able to properly simulate the onset of liquefaction even though the displacements obtained were much lower than those documented.
Finally, an initial evaluation of the post liquefaction effect of the model was performed that could be considered as a starting point for future research.