A simple method to estimate liquefaction- induced deformation was developed to estimate the settlement of river levees. In the analyses the author assumed that the residual deformation would occur in liquefied ground due to the reduction of shear modulus. Selected levees were damaged and non-damaged levees during past two earthquakes. Analyzed settlements agreed fairly well the actual settlements.
Stability of river levees during earthquakes is usually evaluated by the safety factor of the slope. However, in the seismic design of the levees under very intense seismic motion, it is necessary to evaluate not only the stability but also the functionally or serviceability of the levee. The functionally of levees can be evaluated by the amount of settlement. The authors have developed a simplified procedure for the analysis of liquefaction induced residual deformation, based on the tests on stress-strain relationships of liquefied sands (Yasuda et al., 1999). The simplified method was applied to estimate the settlement of river levees. Analyzed results were compared with the actual settlements which were induced during the 1993 Hokkaido-nansei-oki and 1995 Kobe earthquakes.
Figure 1 shows the concept of the stress-strain curves which are used in the proposed method. Line l denotes a back bone curve at the beginning of the earthquake. Point A is supposed to be initial state of a soil element in the ground. When excess pore water pressure generates, material properties such as shear strength and elastic moduli change. Suppose that the back bone curve moves from l to m due to liquefaction, then strain should increases in order to hold the driving stress. Torsional shear test apparatuses were used to obtain the back bone curves of liquefied soils (Yasuda et a., 1998). Specimens were saturated, applied back pressure and consolidated.