Pore-water Pressure Accumulation Under Dynamic Loadings
Advisor: Professor Timothy D. Stark
Abstract
The primary objective of this study was to study the impact of closely spaced low-magnitude earthquakes on the stability of upstream raised tailings dam. For this purpose, a series of numerical analyses were conducted and the level of excess pore-water pressure accumulated by earthquakes was investigated. Results of laboratory triaxial tests, drained and undrained, and cyclic direct simple shear tests on loose sandy tailings were used to calibrate the soil constitutive model under low applied vertical stress conditions. The seismically induced pore-water pressures calculated from the numerical analysis are compared with previous analytical methods so that empirical correlations can be used for estimating seismically induced pore-water pressures when performing a complex seismic numerical analysis is not feasible.
Flow failure stability analysis was also performed with the calculated values of seismic-induced pore-water pressures. This analysis was needed because localized liquefied strength may not result in a flow failure of the entire dam. The inverse analysis results showed the subject dam was stable before the earthquakes. However, the closely-spaced earthquakes, high static piezometric level, which saturated most of the tailings deposit, and loose sand conditions resulted in positive pore-
water pressures being generated during each earthquake. The decrease in the factor of safety against flow failure (FoSFlow) due to pore-water pressure accumulation under closely spaced earthquakes indicated this failure mechanism should be considered for future design and/or periodic stability assessments.
Triggering of liquefaction due to accumulation of positive pore-water pressure, however, could not be explained using previous flow failure assessment procedures. Therefore, a new procedure was developed to assess the susceptibility of sandy soils to undrained shear strength loss that could result in mobilizing a liquefied strength under static or dynamic conditions in dams and embankments. This procedure is based on the critical state soil mechanics framework in the effective stress domain and no longer uses a state parameter or an assessment of contractive/dilative shear behavior. Instead, the effective stress ratio (p’0/p’cs) for segments along the potential failure surface are estimated from cone penetration test measurements and used to classify the soil behavior and estimate the amount of excess pore-water pressure needed to mobilize the liquefied strength. This enables identification of zones that could mobilize a liquefied strength during static or low-level shaking/disturbances as well as high-level earthquake.