Multiple linear regression model for the assessment of bond strength in corroded and non-corroded steel bars in structural concrete

Abstract

With the growth and ageing of the stock of existing structures, structural assessment and retrofitting are fast acquiring a significant role in the construction industry. The benefits of upgrading existing reinforced concrete (RC) structures or extending their service life and of ensuring greater durability in designs for de novo construction have led to a need to include deterioration as a factor in structural safety models. Bond between reinforcing steel and concrete is of cardinal importance in this respect. The present paper proposes a unified formulation for assessing bond strength in corroded and non-corroded steel bars, and an associated model to accommodate the effect of transverse pressure where appropriate. The formulation is the result of applying multiple linear regression analysis to a database built from the findings of over 650 bond tests on corroded and non-corroded reinforcing steel reported in the literature. The data collected include a wide range of variables affecting bond strength, such as bar diameter, concrete compressive strength, concrete cover, anchorage length, confinement ratio and cross-sectional loss. A number of statistical criteria are used to compare the proposed formulation to the other bond strength assessment models, including the fib Model Code 2010 proposal for corroded steel bars. Further to the statistical tests conducted, the model proposed can be usefully applied to assess the structural safety of corroded RC members.