Selection of Driven Pile Design Parameters for the I-15 Reconstruction Project

Abstract

The I-15 Reconstruction Project consisted of the complete reconstruction of nearly 17 miles of urban interstate highway through Salt Lake City, Utah and the metropolitan area south of the city. The $1.6 billion design/build project included construction of 142 new bridges, each supported on driven pile foundations. Subsurface conditions along the alignment ranged from deep, lightly over-consolidated Lake Bonneville clays to dense alluvial sands above the groundwater table. Piles derived support from shaft friction or a combination of shaft friction and end bearing, depending on the strata present. Design pile capacities were evaluated using the effective stress, or beta, method. The UNIPILE computer program was selected, as it was preferred by the Utah Department of Transportation (UDOT) and because of previous successful application in Lake Bonneville deposits. The basic design parameters are side shear (beta) and end bearing (Nt) factors. The design-build contract required full-scale load tests to assist in selecting the design parameters. However, because of the construction schedule, about half the bridges were designed before the pile load tests were completed. Therefore, initial beta and Nt parameters were selected based on geotechnical evaluation of the subsurface data, combined with past experience in Lake Bonneville deposits. Nine full-scale load tests were performed at locations representative of subsurface conditions along the alignment. Static compression, uplift and lateral load tests (not discussed in this paper) were conducted, with each test pile carried to failure in compression. Load distributions in the test piles were evaluated using internal strain measuring devices that allowed the test loads to be separated into side shear and end bearing, and the corresponding beta and Nt values to be determined. Many of the test pile locations were adjacent to structures that would be designed and constructed. Therefore, correlations were made between the measured beta and Nt values and the design pile capacities calculated/predicted before the load tests. With one exception, the ratios of measured to predicted pile capacities were within 0.94 to 1.3. The single exception was a location where the test pile was driven into dense sands above the water table, and the Nt value assigned to the dense sand layer was too great. Adjustments in the design pile capacities were made for this subsurface condition as a result of the load test program, and pile design was completed using the measured design values.