Pore pressure response of liquefiable soil treated with prefabricated vertical drains : experimental observations and numerical predictions

Prefabricated vertical drains represent a soil improvement technique that achieves liquefaction mitigation by decreasing the drainage path length and hence expediting the dissipation of excess pore pressures. When evaluating the required spacing between vertical drains to achieve the desired reduction in pore pressure response, simplified design charts or more sophisticated finite element analyses are used to predict the pore pressure response. These charts and programs have not been evaluated in terms of their accuracy because there exists little data with which to compare the numerical predictions. More recently, the effectiveness of prefabricated vertical drains for liquefaction mitigation has been evaluated via small – scale centrifuge testing performed on untreated soil deposits and on soil deposits treated with vertical drains. In particular, the performance of the soil deposits subjected to sinusoidal motions and actual earthquake recordings was tested. The main goal of this research is to compare the experimental observations of pore pressure response from the centrifuge experiments with the numerical predictions. The comparison focuses on the average excess pore pressure ratio (r_(u,avg)) that was developed in the location of a vertical pore pressure array in both the untreated and drain – treated sides of the models. In parallel, a parametric study is performed for the numerical predictions in order to study the effect of each input parameter that influences the pore pressure prediction, namely the effect of soil properties, ground motion characteristics and drain parameters. The numerical predictions are found to provide reliable predictions of the pore pressure response despite the simplicity of the constitutive model employed. The numerical predictions of r_(u,avg) time – histories are generally in good agreement with the recorded values in the centrifuge experiments. In most of the cases, the numerical model managed to predict the same maximum average excess pore pressure ratio, which is the parameter that is used in drain design. To incorporate any uncertainty on the soil properties or on the characteristics of shaking, the use of a smaller pore pressure threshold for drain design is recommended.