Statistical quantitative evaluation of bending strength of corroded RC beams via SMFL technique. (15th April 2020)

• Record Type: Journal Article
• Authors: Qiu, Junli; Zhou, Jianting; Zhao, Siyu; Zhang, Hong; Liao, Leng
• Is Part Of: Engineering structures. Volume 209(2020)
• Publication Date: 2020-04-15
• Subjects: Corroded reinforced concrete (RC) beams -- Statistical evaluation -- Bending strength -- Self-magnetic flux leakage (SMFL) -- Bayesian method; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2020.110168
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: Bending strength of corroded RC beams evaluated via innovative NDT technique SMFL. SMFL index K for quantitative evaluation was proposed theoretically. Quantitative correlation between index K and bending strength loss rate β was established. Bending strength of corroded RC beams was quantitatively evaluated via K based on Bayesian model. Abstract: The self-magnetic flux leakage (SMFL), a non-destructive testing (NDT) technique, was applied innovatively to the quantitative statistical evaluation of the bending strengths of corroded reinforced concrete (RC) beams. In this study, the variations of the SMFL and bending strength caused by steel bar corrosion were analyzed theoretically and experimentally. Based on theoretical analysis, a SMFL index K that is not affected by the initial magnetization and magnetization history was proposed. The bending strength loss rate β of the corroded RC beams could be evaluated quantitatively using index K . 1775 K data from 18 corroded RC beams with different corrosion degrees were extracted, and the experimental correlation between K and β was established based on the experimental data and a theoretical correction. The results indicated that the magnitudes of K from different beams were similar, which proved that K eliminated the difference between the initial magnetization and magnetization history between different individuals. K and β were approximately exponentially related, and their correlation was not significantly affected by the stirrups. In view of the discreteness of the experimental sample K, β was statistically evaluated via K based on a Bayesian model. The posterior probability distribution of β in different value intervals when K was in a certain value interval was given. Thus, the quantitative classification of β was realized. This work is beneficial for the engineering evaluation of the bending strengths of corroded RC beams and provides an innovative research idea.