Static axial pile foundation response using seismic piezocone data
Niazi, Fawad Sulaman
MetadataShow full item record
Ever since the use of cone penetration testing (CPT) in geotechnical site investigations, efforts have been made to correlate its readings with the components of static axial pile capacity: unit base resistance (qb) and unit shaft resistance (fp). Broadly, the pile capacity analysis from CPT data can be accomplished via two main approaches: rational (or indirect) methods, and direct methods. The rational methods require a two-step approach, whereby CPT data are first used to provide assessments of geoparameters that are further utilized as input values within a selected analytical framework to enable the evaluation of the pile capacity components. In contrast, direct CPT methods use the measured penetrometer readings by scaling relationships or algorithms in a single-step process to obtain fp and qb for full-size piling foundations. The evolution of the CPT from mechanical to electrical to electronic versions and single-channel readings (i.e., measured tip resistance, qc) to the piezocone penetration test (CPTu), that provides three readings of point stress (qt), sleeve friction (fs), and porewater pressure (u1 or u2), has resulted in the concurrent development of multiple CPT-based geotechnical pile design methods. It is noted, however, that current CPT-based methods focus only on an estimate of "axial pile capacity", corresponding to a limiting load or force at full mobilization. A more comprehensive approach is sought herein utilizing the CPT readings towards producing a complete nonlinear load-displacement-capacity (Q-w-Qcap) on axial pile response. In particular, the seismic cone penetration test (SCPT) provides the profile of shear wave velocity (Vs) that determines the fundamental small-strain shear modulus: Gmax = gt?Vs2, where gt = total mass density of soil. With the penetrometer readings useful in assessing foundation capacity, the stiffness Gmax finds application within elastic continuum solutions towards evaluating the load-displacement (Q-w) response. In this study, a concise review of the deep foundation systems is presented, including pile types and characteristics, various arrangements of axial pile load testing in static mode, and interpretations of the load test data. In addition a comprehensive state-of-the-art review of CPT-based rational and direct methods of pile capacity evaluations is compiled. It is recognized that the direct methods offer more convenience in their straightforward approach in estimation of the pile capacity. The piezocone-based UniCone direct method proposed by Eslami and Fellenius (1997) is selected for further refinements, as it utilizes all three CPT readings in its design formulations. Concerning the analysis of pile deformations under axial loading, a brief review covers designs employing empirical formulations, analytical solutions, load-transfer (t-z) methods, numerical simulations, variational approaches, and those using hybrid methods. Specifically, the analytical elastic solution by Randolph and Wroth (1978; 1979) is covered in more detail since it is simple and convenient in application with extended applications to uplift and bidirectional O-cell types of loadings. This elastic approach also serves well in modeling a stacked pile solution for layered soil profiles. The last part of the review covers various shear modulus reduction schemes, since evaluation of the applicable stiffnesses is considered to be the most delicate phase in the nonlinear Q-w response analysis of axially loaded piles. It is identified that the most appropriate scheme applicable to static axial loading of pile foundations is the one that can be derived from the back-analyses of actual load tests within the framework of analytical elastic solution. In order to conduct a comprehensive research study on the axial Q-w-Qcap response of deep foundations from CPT readings, a large database is compiled. This includes 330 case records of pile load tests at 70 sites from 5 continents and 19 different countries of the world, where pile foundations were load tested under top-down compression or top-applied uplift (tension) loading, or both, or by bi-directional Osterberg cell setups. All test sites had been investigated using CPT soundings; in most cases by the preferred SCPTu that provides all four readings from the same sounding: qt, fs, u2, and Vs. In a few cases, sites were subjected to CPT or CPTu and the profiles of shear wave velocities were obtained by other field geophysical techniques, otherwise by empirical estimations. Results of the new correlation efforts are offered to derive coefficients Cse for shaft component and Cte for base component of the axial pile capacity from CPTu data. The UniCone type of soil classification chart is refined by delineating 11 soil sub-zones along with their respective Cse, in contrast to the 5 zones originally proposed. The CPT material index, Ic (Robertson, 2009) is then used to establish direct correlations linking Cse vs. Ic and Cte vs. Ic. Statistical relationships offer continuous functions for estimating the coefficients over a wide range of Ic values, thereby eliminating the need for use of the soil classification chart as well as improving the reliability in the evaluations of fp and qb. The effects of the pile loading direction (compression vs. uplift) and loading rate are also incorporated in the proposed design formulations. New sets of shear stiffness reduction curves are developed from the back-analysis of pile load tests and Gmax profiles obtained from the SCPT data. Alternative functions formats are provided in terms of hyperbolic tangent expressions or exponential curves, developed as normalized shear stiffness (G/Gmax) vs. logarithm of percent pseudo-strain (gp = w/d, where w = pile displacement and d = pile diameter). These charts offer convenience in the axial Q-w analysis of different pile categories within the framework of analytical elastic solution. The results also account for the plasticity characteristics of the soil formations within the database. A stacked pile model for Q-w analysis is presented in which certain adaptations are proposed in the elastic continuum solution. These adaptations enable plotting of separate modulus reduction curves (G/Gmax vs. gp) as function of depth for each layer, and treating pile as a stack of smaller pile segments embedded in a multi-layered soil media. The solution can be used to address the question of progressive failure with depth in a multi-layer soil media that exhibits nonlinear soil stiffness response. Finally, the closed-from analytical elastic pile solution for predicting the Q-w response is decoupled and modified to account for different setup cases and multi-stage loading of bi-directional O-cell tests. The decoupling accounts for separate assessments of the response to axial loading for different segments of pile shaft and different stages of loading, while the modifications include: (1) reduced maximum radius of influence for the upward displacements of the upper shaft segment, and (2) modeling the non-linear ground stiffness from the back-analysis of a well-documented dataset of O-cell load tests.