Non-linear multivariable model for predicting the steel to concrete bond after high temperature exposure. (20th July 2020)

• Record Type: Journal Article
• Authors: Varona, F.B.; Baeza, F.J.; Bru, D.; Ivorra, S.
• Is Part Of: Construction & building materials. Volume 249(2020)
• Publication Date: 2020-07-20
• Subjects: Bond strength -- Pull-out strength -- Fibre reinforced concrete -- High performance concrete -- High temperature; Building materials -- Periodicals; 624.18
• DOI: 10.1016/j.conbuildmat.2020.118713
• Languages: English
• ISSNs: 0950-0618

Abstract

Highlights: An exhaustive review on steel–concrete bond at high temperature is included. Results of 17 experimental studies are systematically collected in a database. Bond at high temperature is predicted through multiple variable regression models. Ratio of bond strength to compressive strength at elevated temperatures is studied. The Model Code 2010 equation for peak bond strength is adapted for high temperature. Abstract: The bond mechanism between steel and concrete can be compromised during a fire and is one of the least investigated phenomena in concrete research. In this work we present a thorough review of the experimental data available on this topic. The results from the tests reported by a number of researchers have been systematically collected in a database. This work also reports the results obtained in the bond strength tests carried out on four batches of normal and high strength concretes exposed to temperatures up to 825 °C. The database provides the source for a multiple regression analysis which is performed in order to define a model aimed at predicting the bond strength as a function of several variables: the exposure temperature, the concrete compressive strength at ambient temperature, the type of fibre addition, the fibre volume fraction, the age at testing, the bond length and the concrete cover of the steel bar. Based on different error measurements, our model is favourably compared to the set of experimental results reported here and also other prediction models reported in the literature.