This report addresses the critical challenge of characterizing the residual strength in p-y curves for rock-socketed pile shafts, particularly the Strong Rock (SR) p-y curves commonly used in practical applications. The brittle behavior exhibited by the original SR p-y curves, with a sudden drop to nearly zero residual strength, has practical implications for the design and construction of rock-socketed pile foundations. The absence of empirical evidence supporting this drastic eduction in residual strength necessitates a reevaluation of these curves.
The objective of this study is to modify the post-peak residual strength of SR p-y curves. The selection of SR p-y curves for this study is based on their minimal input parameter requirements, rendering them a potential alternative to more intricate p-y relationships during sensitivity analysis. The principal assumption in this study is that once rock fractures occur at the rock-pile interface due to lateral pressure, the lateral resistance reduces to that of very poor quality, heavily fractured rock with an RQD of 25%. To characterize the reduction of lateral resistance, a three-dimensional (3D) model of a laterally loaded pile was developed in FLAC3D, employing the Hoek-Brown constitutive model (Hoek and Brown, 1997) to simulate the response of rock layers with varying Rock Quality Designation (RQD). First, the 3D model was validated against a lateral load test involving a 6-foot-diameter reinforced concrete (RC) rock-socketed shaft (Schnabel, 2014). Subsequently, the validated numerical model was employed in a parametric study to characterize the p-y response of rocks with different RQDs, which were subsequently used to characterize the residual strength of rock p-y curves. The determined residual strength was then incorporated into the SR p-y curves within the LPILE software, thereby facilitating a comparison of the predicted lateral response of the shaft using both the original and modified SR curves.
Figure E-1 shows the original and modified SR p-y curves for three different rock qualities, including RQD values of 25%, 60%, and 100%, serving as illustrative examples. The relationships for Rpult and Rres are provided in Equations E-1 and E-2 respectively and can be employed to develop p-y curves for rocks with differing RQD values. The proposed adjustments to the original SR p-y curves include two main aspects:
- The ultimate lateral resistance (pult) is defined for varying RQDs by employing a reduction factor (Rpult), as outlined in Equation E-1. Notably, the original SR curves, as implemented in LPILE, lack explicit guidance on reducing pult or Qu for RQD values below 100%. In this study, a reduction factor (Rpult) is introduced to modify pult for different RQD values. This approach essentially mirrors the reduction of Qu, given that pult within SR p-y relationships exhibits linear proportionality to Qu.
- Within the modified SR p-y curves, the residual strength is characterized using Equation E-2. In the original SR p-y curves, the lateral resistance reduces to a nearly zero value as deflections surpass 0.0024 times the pile diameter. In the modified SR p-y curves, the residual strength is defined by a non-zero value calculated using Equation E-2. It is important to note throughout the formulation of both the original and revised SR p-y curves, the unconfined compressive strength (Qu) corresponds to an intact rock specimen.