The effect of irregular seismic loading on the validity of the simplified liquefaction procedures

Abstract

Soil liquefaction has been one of the major hazards for civil engineering projects relating to earthquakes. The simplified liquefaction procedure which is used to assess liquefaction susceptibility in practice is still based on semi-empirical methods. These rely on the assumption that irregular seismic motions can be represented fully by an equivalent number of cycles of uniform stress amplitude, which is based on the peak acceleration measured at ground surface. Most methodologies used to calculate the equivalent number of cycles are based on Miner's damage concept developed for the fatigue analysis of metals. Several researchers have questioned the validity of this concept, as soils have a highly non-linear response. The present work investigates numerically the concept of the equivalent uniform amplitude cycles. Effective stress-based non-linear finite element analyses are performed with a modified bounding surface plasticity model that allows to realistically simulate liquefaction, reproducing the cyclic strength of sands accurately. The seismic response of a 15 m deep uniform level-ground sand deposit is simulated with full hydro-mechanical coupling to establish the benchmark extent of liquefaction zone. In parallel, the analyses are repeated assuming drained conditions to compute the irregular time-histories, which are then converted to an equivalent number of uniform amplitude cycles. The constant amplitude series are then applied in single element simple shear test simulations, with initial conditions those corresponding to the 7 m depth in the deposit. The results in terms of the predicted triggering of liquefaction are contrasted to the predictions of the fully coupled benchmark analyses at the corresponding depth to assess the validity of the Seed et al. (1975) methodology, based on Miner's cumulative damage concept. (English) [ABSTRACT FROM AUTHOR]