Behavior and modeling of RC columns confined with large-rupture strain FRP composites

Abstract

A very popular application of FRP composites is as external jacket for the seismic strengthening/retrofit of RC columns. It has been well established that FRP confinement can enhance both the strength and ductility of RC columns. The three common FRP composites used in column strengthening/retrofit include carbon FRP (CFRP), glass FRP (GFRP) and aramid FRP (AFRP) composites, which are referred to as conventional FRPs in this thesis. These FRP composites have a linear elastic stress-strain response with brittle rupture failure at a small rupture strain (typically around or less than 1.5% for CFRP, around or less than 2.5% for GFRP and around or less than 3% for AFRP). For seismic retrofit applications, an FRP jacket with a larger hoop rupture strain is generally beneficial as it leads to more ductile behavior and greater energy absorption with the same degree of strength enhancement. In recent years, a new category of FRP composites has emerged as an alternative to conventional FRP composites for the seismic retrofit of RC columns. These FRPs are made of new materials such as Polythylene Naphthalate (PEN) or Polyethylene Terephthalate (PET) fibers which possess a large rupture strain (LRS) (usually larger than 5%) but a relatively low elastic modulus. Compared to conventional FRPs, these LRS FRPs are cheaper and more environmentally friendly because they are usually made from recycled materials (e.g., waste plastic bottles). Therefore, the use of LRS FRPs is expected to lead to a more economical and ductile solution for seismic retrofit applications. The low elastic modulus of these new FRP composites is generally not a problem because it can be compensated by the use of a thicker FRP jacket. Against this background, this thesis is concerned with the investigation of application of the LRS FRP in seismic strengthening / retrofit of RC columns. This thesis presents a systematic study covering the behavior and modeling of RC columns confined with LRS FRP. This thesis is composed of three parts: (1) behavior and modeling of LRS FRP-confined plain concrete; (2) interaction between the external FRP jackets and internal longitudinal reinforcing bars in FRP-confined RC columns; (3) behavior and modeling of FRP-confined RC columns subjected to combined constant axial load and cyclic lateral load. In the first part of this thesis, an experimental study on LRS FRP-confined plain concrete cylinders subjected to monotonic axial compression was conducted. Results from tensile coupon tests for LRS FRP indicated that LRS FRP fibers possess an approximately bilinear tensile stress-strain relationship, which has a significant effect on the axial compressive stress-strain behavior of FRP-confined concrete. A recent confinement model for conventional FRPs was compared with the present test results, indicating that the model significantly overestimates the ultimate axial strain. A modified version of the model was then presented to provide more accurate predictions for the test results. Furthermore, tests on LRS FRP-confined plain concrete columns under cyclic axial compression were carried out. A cyclic compressive stress-strain model for LRS FRP-confined concrete was developed by combining the proposed monotonic model for the envelope curve and a recent model for the cyclic unloading/reloading rules. In the second part of this thesis, interaction between the external FRP jacket and the internal longitudinal reinforcing bars in FRP-confined RC columns was investigated. It is well known that in FRP-confined RC columns, the lateral deflection of longitudinal reinforcing bars influences the performance of the external FRP jacket (i.e., mainly the rupture strain efficiency and non-uniform strain distribution along the column section circumference), and vice visa. In this study, the buckling behavior of longitudinal reinforcing bars, which is believed to be the most representative phenomenon for the interaction mechanism and the major reason for the rupture strain efficiency and non-uniform strain distribution of FRP jacket along the column section circumference, was studied through both experimental and numerical methods. At the beginning of the second part, experimental studies on FRP-confined RC columns subjected to monotonic and cyclic axial compression were presented. For the FRP-confined RC columns under monotonic axial loading, the contribution of longitudinal reinforcing bars could be separated by subtracting the total axial load by the contribution of the corresponding FRP-confined plain concrete columns. A curved beam model was utilized to evaluate the confining stiffness provided by the external FRP jackets on the longitudinal reinforcing bars. Then a "beam with springs" model was proposed to calculate the monotonic compressive stress-strain response of longitudinal reinforcing bars in FRP-confined RC columns. Favorable agreement between the predicted stress-strain curves and test results validated the accuracy and applicability of the proposed model. Furthermore, parametric studies based on the proposed "beam with springs" model were conducted to probe into the influence of the three main factors on the compressive stress-strain curves of longitudinal reinforcing bars. An empirical model was proposed to describe the monotonic compressive stress-strain response of longitudinal reinforcing bars in FRP-confined RC columns through the least-squares regression method. The proposed monotonic compressive stress-strain model, together with a monotonic tensile stress-strain model and the Menegotto-Pinto model accounting for the cyclic loops, comprises a complete stress-strain model for reinforcing bars under cyclic loading. The proposed stress-strain model for laterally supported reinforcing bars and a simplified version for the proposed model for FRP-confined plain concrete under cyclic compression were implemented into the OpenSees software platform. Comparisons between test results and results predicted by OpenSees verified the accuracy of the proposed reinforcing steel model. Results from OpenSees simulations also demonstrated that the plastic strains of plain concrete, reinforcing bars, and the whole RC column are inconsistent with each other: when the whole RC column is unloaded to zero, the FRP-confined plain concrete tends to be in compression, while the reinforcing bars tend to be in tension. The third part of this thesis presented an experimental study on seismic behavior of RC columns confined with FRP composites. Three RC columns strengthened with PEN FRP composites, were tested under combined constant axial loading and cyclic lateral loading. The results were compared with those of three conventional CFRP-confined RC columns and one control RC column. Hysteretic responses and failure mode of these columns, as well as the strain evolution histories of longitudinal reinforcing bars and external FRP jackets, were discussed in detail. Numerical simulations based on a fiber element model were conducted to reproduce the test results, and close predictions could be obtained by numerical simulations.