Title: Application of Fracture Mechanics to Steel-Concrete Bond Analysis

Author(s): S. L. McCabe, D. Darwin, O. C. Choi, and H. Hadje-Ghaffari

Publication: Symposium Paper

Volume: 134

Issue:

Appears on pages(s): 101-114

Keywords: bond (concrete to reinforcement); bond stress; cover; failure; cracking (fracturing); epoxy resins; finite element method; models; fracture properties; tension; Design

DOI: 10.14359/3098

Date: 9/1/1992

Abstract:
The recent introduction of epoxy coating to reinforcing steel has made the study of bond, and the effect of this coating, all the more important. A recent large scale study of bond performance of epoxy-coated and uncoated reinforcement conducted at the University of Kansas using modified cantilever beam-end specimens has shown the effects of various parameters on bond. These specimens placed the bar and surrounding concrete in tension, simulating the situation in actual members. A prescribed bond test region, the bonded length, was placed at a discrete distance, the lead length, from the front of the specimen to prevent surface effects from interfering with the test region. The experimental work has provided ample evidence of the fundamental fracture mechanics aspects of bond failure and the subsequent specimen failure. Splitting failure of the beam-end specimens was observed consistently in all tests where a fracture plane formed above the bond test region and propagated quickly through the tension region of the specimen. Tests indicated that the bonded length of the bar, the value of the lead length, and the amount of cover were all important parameters. The paper presents the results of an analytical evaluation of the bond process and specimen fracture that was observed in the laboratory, using nonlinear finite element analysis to study the effects of interface properties on the fracture behavior and failure load. The majority of the beam-end specimen was modeled using linear elastic elements representing one-half of the symmetric experimental specimen. The actual bar-concrete interface was modeled using link elements and a Mohr-Coulomb failure model. Rod elements joined the specimen to the specified crack plane located at the center line of the specimen. The fracture process was modeled using Hillerborg's fictitious crack model. The behavior observed in the laboratory for coated and uncoated bars has been accurately predicted using this procedure. The fracture process and resulting overall bond performance has been studied analytically to assess the effects of interface properties on the fracture behavior. The analytical studies confirmed that coating reduced the relative bond strength with respect to that of an uncoated bar, while the absolute bond strength was found to increase with additional cover and lead length.