The Sacramento-San Joaquin Delta is a 700,000 acre estuary at the confluence of the Sacramento and San Joaquin rivers, which consists of below sea-level islands surrounded by levees. Many of theses levees are not engineered structures, consisting of uncompacted sands, silts, clays, and organics often founded upon natural levees and in some cases on peaty organic soil. Many hazards threaten the Delta, but the seismic hazard is exceptional because of the potential for multiple simultaneous breaches inundating many islands within the Delta. The objectives of this research are to improve fundamental understanding of cyclic and post-cyclic behavior of peat that could affect levee performance, and develop analysis tools to predict this behavior.
Laboratory testing is performed on 22 undisturbed Shelby tube samples gathered from Sherman Island at depths ranging from approximately 1 to 6 m. The shear testing in this research is carried out using the UCLA bi-directional broadband simple shear device, which is a digitally-controlled device with capabilities for chamber pressure control and multidirectional excitation. The apparatus was improved as part of this research program to add capabilities for testing under constant-height and stress-controlled conditions.
Oedometer tests are carried out on Sherman Island peat to evaluate its compressibility properties. To facilitate accurate determination of the end of primary consolidation, a new consolidometer was fabricated that provides single drainage through the top of the specimen, while pore pressure is measured at the bottom. For specimens with high organic content (OC > 28%), values of coefficient of consolidation (c_v) for normally-consolidated load stages are observed to decrease as vertical effective stress increases, often by more than two orders of magnitude over the stress range tested in the consolidometer. For the Sherman Island peat c_v is as high as 400〖?10〗^(-4) cm2/s at σ_p^'. The trend of compression index (Cc), and recompression index (Cr) confirms that they are positively correlated with in situ water content (w0) and OC. The values of C_α⁄C_c (where C_αis secondary compression index) for the Sherman Island peat ranges from 0.05 to as high as 0.12 with an average of 0.08. Hydraulic conductivity k depends on void ratio, and decreases with the decrease in void ratio. The slope of e versus logk (i.e., C_k) increases with initial void ratio (e0), and the best fit for the data is C_k=0.20e_0.
Monotonic test results show that lightly overconsolidated peat with OCR < 2 shows contractive behavior, while higher OCR’s result in dilative behavior. The Normal Consolidation Line (NCL) and Failure State Line (FSL) are approximately parallel for OC ≤ 35%. Normalized shear strength (the normalization is with respect to pre-shear vertical effective stress) has been evaluated as a function of OCR. The soil behavior generally supports the concept of normalization, with the strength ratio being higher for high OC (80-85%) than for low OC (≤35%).
Cyclic strain-controlled tests show that for low shear strain amplitude, γ_c (< 0.7%), although hysteretic loops form, cyclic degradation of stiffness does not occur and pore pressures do not accumulate. Accordingly, stress paths are similar to those for a drained test. For higher 〖 γ〗_c, r_ur increases from cycle-to-cycle and reaches to around 0.1 after 15 cycles at 4% shear strain. The soil stiffness degrades slightly to achieve the uniform strain amplitude.
One of the principle contributions in this dissertation is to demonstrate that evaluation of the rate of secondary compression following primary consolidation is related to the vertical distance in void ratio – effective stress space between soil state and a secondary compression reference line (SCRL). The traditional approach takes this rate as logarithmically decaying with time following load application, but is shown to not be generally applicable. For example, it fails for the case of a small load increment applied to a soil element. I develop a conceptual and analytical framework to compute the rate of void ratio change as a function of soil state given this framework, which predicts slower secondary compression as OCR (overconsolidation ratio) increases. The proposed framework provides a much improved match to observations for conditions differing from those in traditional consolidation tests in which the ratio of load increase to initial stress is approximately unity.
When peaty organic soils are cyclically loaded, they can experience an increase of secondary compression rate relative what would have been present without dynamic loading. This increased rate of volume change occurs even for shear strain amplitudes that do not induce a pore pressure response. For a given soil, the rate change increases with strain amplitude and number of cycles, which can be viewed as a partial resetting the secondary compression clock without change in total stress. However, rather than modeling this in a time-based framework, I demonstrate that this reset behavior can be captured by a vertical shift of the SCRL, which can be quantified by a reset index (I_R). The value of I_R varies between 0 (no reset) to 1 (fully reset). An empirical model is developed for I_R as a function of cyclic shear strain amplitude, number of uniform loading cycles, organic content, over-consolidation ratio, initial overburden pressure, and amount of static shear stress.
A simplified procedure is developed to estimate post-earthquake settlement of organic soils in consideration of post-earthquake pore pressure dissipation and accelerated rates of secondary compression. The procedure uses 1-D site response analysis to find representative profiles of peak shear strain and its phasing in time. The irregular shear strain time series is then converted to uniform shear strain cycles at some specified amplitude as a fraction of the peak. Reset index and pore pressure ratio at the conclusion of shaking are then calculated using predictions equations conditioned on strain amplitude, number of cycles, over-consolidation ratio, organic content, and pre-earthquake stress conditions. The resulting values of reset index and pore pressure can be used in a non-linear consolidation code to calculate post-earthquake settlement versus time.