Dynamic Shake Testing of a Model Levee on Peaty Organic Soil in the Sacramento-San Joaquin Delta
- Author(s): Reinert, Edward Thomas
- Advisor(s): Brandenberg, Scott J
- et al.
The Sacramento-San Joaquin Delta is the hub of California's water distribution system, which consists of below sea-level islands surrounded by levees. Delta levees are constructed of local fill, have typically been unengineered and are notorious for breaching, causing flooding of the islands inside. One major concern is the seismic performance of Delta levees, which have not experienced a significant seismic event in over a century. Many of these levees are founded on local peaty organic soils, and the seismic performance of these levees is poorly understood.
As part of a collaborative research investigation to study to study the seismic performance of Delta levees, a series of dynamic field tests were performed on a model levee constructed on very soft and compressible peaty organic soils on Sherman Island. This first-of-its-kind test applied dynamic loads to the model levee-peat system using the large NEES@UCLA MK-15 eccentric mass shaker mounted on the levee crest. Two sets of tests were performed in 2011 and 2012.
Geotechnical and geophysical investigations performed at the site found a 11m thick peat deposit rests atop permeable Pleistocene sand. The peaty soils consist of 9m of soft saturated peat with a Vs of 30 m/ and a 2m stiff desiccated crust with a Vs of 60 m/s lying atop the soft peat. Artesian pressures exist in the soft saturated peat due to hydraulic connection with the nearby Sacramento River, with a zero effective stress condition existing at the peat-sand interface. Remote data monitoring measured settlement and pore pressure dissipation of the levee using embedded piezometers and a slope inclinometer. The remote monitoring found fast dissipation of pore pressures underneath the levee and continued settlement of the levee due to a high rate of secondary compression. Prior to the 2012 tests, a berm was constructed around the levee and the ground was flooded, to create more realistic soil conditions in the unsaturated crust.
Dynamic base shear-displacement and moment-rotation relations were made for the levee. The model levee translated and rotated visibly during testing, demonstrating a response that differs from the one-dimensional wave propagation assumption used to analyze seismic levee response. High radiation damping was observed, and translation of the levee was found to go out-of-phase at peak shaker frequencies. Complex-valued stiffness of the levee-peat system was analyzed and compared to analytical solutions for a rigid foundation on an elastic halfspace. Little agreement was found between the field test results and the analytic solution, suggesting that the levee-peat foundation is flexible.
Dynamic shear strains measured underneath the levee crest and toe measured an average value of shear strains at the bottom of the stiff crust and top of the soft peat. Peak shear strains measured during testing went up to 0.4%, with higher shear strains occurring underneath the levee toe, due to the rocking behavior. Comparison of residual pore pressure ratios generated during testing show a trend in increasing residual pore pressure with increasing shear strain. Comparison of field test results with dynamic laboratory testing showed very little increase in residual pore pressures from field tests, suggesting that pore pressures underneath the levee dissipated quickly due to high horizontal permeability.
A series of finite element simulations were performed with elastic isotropic materials to compare different hypothetical soil conditions and loading scenarios. Good agreement in shear strains between the field test and the finite element simulations were found. Higher shear strains were found to exist beneath the levee for softer soils and uniform base excitation. A study investigated the development of shear stresses within the levee fill, and found an increase in peak shear stresses compared to shear stresses calculated for a simple shear case. This has implications for liquefaction triggering analysis, and the finite element simulations suggest that the current methodology used in evaluating seismic demand may be underestimating shear stresses within the levee fill.