Evaluation of Laterally Loaded Pile Group at Roosevelt Bridge

Abstract

An isolated single pile and a large-scale test group of 16 prestressed concrete piles spaced at 3 diameters were subjected to a static lateral loading using a fixed-headed production group for reaction. The foundation consisted of sand overlying a partially cemented sand at the Roosevelt Bridge replacement, Stuart, Fla. Ten piles of the test group, six piles of the reaction group, and a single pile were instrumented with strain gauges and inclinometers. The piles were 76 cm², and approximately 16.5 m long. Standard penetration tests (SPT), cone penetration tests (CPT), DMT, and PMT in-situ tests were used to establish the soil profile and p-y curves. Subsequent to testing, the strain gauges and inclinometer data were reduced to “measured”p-y curves. The p-y curves developed from SPT correlations and PMT results provided an accurate soil representation. The single pile was subjected to 320 kN, while the pile groups were loaded to about 4,800 kN. The testing results show that the nonlinear characteristics of cracked prestressed concrete piles dominate analyses and data reduction. Consequently, the FLPIER program, with its nonlinear concrete capabilities, could predict properly the “postcracking” response. The group interaction was accurately modeled by p-y (actually p, only) multipliers, which were determined as 0.8, 0.7, 0.3, and 0.3 for the leading, middle leading, middle trailing, and trailing rows, respectively, with the overall group p-y multiplier being 0.55.