Bond between Smooth Prestressing Wires and Concrete: Finite Element Model and Transfer Length Analysis for Pretensioned Concrete Crossties

Abstract

Pretensioned concrete ties are increasingly employed in railroad high speed and heavy haul applications. The bond between prestressing wires or strands and concrete plays an important role in determining the transfer length of pretensioned concrete members, but little research was done to characterize the transfer length in terms of steel reinforcement and concrete factors for railroad concrete ties. The Federal Railroad Administration is sponsoring a comprehensive test program at Kansas State University (KSU) aimed at quantitatively correlating prestressing steel and concrete variables with the transfer length of pretensioned concrete crossties. The Volpe Center has been applying the data obtained in the KSU test program to develop bond models that can be used in transfer length prediction and failure analysis of concrete ties. This paper describes finite element (FE) model development related to the smooth prestressing wire whose dominant bonding mechanisms with concrete are chemical adhesion and friction. The commercial FE software, Abaqus, is employed, and the steel-concrete interface is discretized with cohesive elements. A user bond model is developed within the elastoplastic framework and implemented for axisymmetric and 3D cohesive elements. The bond model defines constitutive relations in terms of normal and shear stresses vs. interfacial dilation and slips. The bond behavior is initially linear elastic, followed by adhesion and friction that are governed by a yield function and a plastic flow rule specific for the smooth wire-concrete interface. The main bond material parameters are normal and shear elastic stiffness, initial adhesive strength, plastic slip at which adhesion first breaks completely, and coefficient of friction. Except for the coefficient of friction, which is determined with reference to the open literature, the bond parameters are calibrated from untensioned pullout tests and pretensioned prism tests conducted at KSU. The calibrated bond parameters exhibit a dependence on the nominal compressive strength of concrete at the time of pretension release. Because considerable concrete creeping has been observed in the periods between pretension release and concrete strain measurement in the test program, an additional concrete material parameter, basic creep compliance, can be calculated and applied to adjust the concrete surface strain data. The user bond model is then validated with transfer length data measured on actual concrete crossties made with a smooth prestressing wire in a tie manufacturing plant.