Soil Liquefaction
Loosely packed soils tend to a decrease of soil volume during shearing. In case the soil volume is water saturated, i.e. the voids in the soil skeleton are filled with water, the water has to drain so that the soil volume can decrease. However, if the loading is faster than the drainage time of the water or if it is prevented from flowing out, a pore water pressure is generated. The pore water pressure causes the grains to be pressed apart and thus, the contact surfaces between the soil grains and the contact forces reduce. As the pore water pressure increases, the contacts between the soil grains diminish. In the following, the soil behaves like a fluid – the soil liquefies.
Soil liquefaction can occur e.g. during an earthquake. In case the subsoil consists of a coarse-grained material, water can escape more quickly and soil liquefaction can be prevented. At the Institute of Geotechnical Engineering, the usage of ballast columns as a measure to prevent soil liquefaction is investigated.
The mechanism of soil liquefaction can be illustrated by means of a model test consisting of a water-saturated sand column. The basic principle of the experiment is to induce a dynamic impulse into a sand volume trapped in a tank.
At the beginning of the experiment, dry sand is trickled into a tank made of Plexiglas and then flooded with water from the bottom to the top. This way, for the sand volume a low density and high water saturation are generated. For a low density and a high degree of water saturation, the dynamic impulse leads to swinging of the tank and thus to cyclic shear loading. This causes densification of the sand. If the impulse entry takes place very quickly, so that an outflow of pore water is hardly possible during this time, short-term undrained (no change in volume) conditions can be assumed. This leads to a generation of pore water pressures which reduce the grain-to-grain stresses. When the grain-to-grain stresses become very small, the rigidity of the soil is nearly zero and the soil behaves like a liquid. Large settlements at the surface are the consequences.
During the model test the generated pore water pressures, at different depths of the test stand, can be determined by the deflections of water columns.
The soil liquefaction is significantly influenced by the granulometric properties of the soil. The highest tendency regarding liquefaction can be observed for sandy soils. For these soils at full water saturation a compression can take place only by expressing the pore water. Depending on the water permeability an excess pore water pressure can occur which reduces the effective stresses at constant total stresses. If the excess pore water pressure is equal to the effective stress the grain skeleton disintegrates and the soil behaves like a liquid.
As part of this project, the liquefaction tendency of various coarse-grained soils is assessed using a self-developed identification test.
For this purpose, a numerical model using an advanced hypoplastic constitutive model is also created.
Earthquakes cause enormous economic and humanitarian damage, especially when the earthquakes cause soil liquefaction. Soil liquefaction is usually referred to as the vanishing of the effective stresses due to a generation of pore water pressures. This is manifested by a loss of soil stiffness and shear strength. Sand boiling, settlement differences, flow slides and a loss of bearing capacity can be the result.
At the Institute of Geotechnical Engineering, centrifuge experiments on a flooded earth dam were numerically recalculated. For this purpose, an advanced hypoplastic constitutive model was used. On the one hand, the aim was to analyse the behaviour of the earth dam under seismic action. On the other hand, the reliability of the constitutive model used in dynamic boundary value problems was verified. The simulation results displayed that the calculated and measured displacements corresponded very well. Both for the settlements with time during the earthquake and the final amount of settlements, the calculated and measured values were very close. However, there were differences in the measured and calculated developments of pore water pressure in proximity to the slope.
The consistent simulation results are based on a laborious calibration of the material parameters. Here, cyclic undrained triaxial compression tests were recalculated and both the decrease of the shear modulus with increasing shear strain as well as the decrease of the effective mean stress with the number of cycles were evaluated.
The usage of stone columns to prevent soil liquefaction has increased in recent years. However, studies on their operating principles are hardly available. This can be explained by the fact that both numerical calculations with simple constitutive models and small-scale laboratory tests are problematic.
At the Institute of Geotechnical Engineering, the application of soil improvement columns for the prevention of soil liquefaction was numerically investigated. For this purpose, a hypoplastic constitutive model was used. The analyses focused on the influence of stiffness and water permeability of the columns.
Furthermore, the influence of column production on the surrounding soil is investigated. Centrifuge experiments were carried out and recalculated numerically in order to validate the numerical calculations. The calculation results show that the soil improvement columns can prevent soil liquefaction due to their high stiffness as well as their high water permeability.