Diagnosing corroded reinforced concrete structures based on hybrid artificial intelligence model

Abstract

ABSTRACT

One of the main causes of deterioration in structural strength of reinforced concrete structures is due to corrosion of reinforcing bars. Prediction of bearing capacity of reinforced concrete beams corroded has been investigated from experimental and theoretical perspectives. Most of the research work has been done using empirical formulas and single prediction models. This study uses a hybrid model between the least squares vector support machine and the differential evolution algorithm on the computing environment of Matlab software. The model is built and tested on a collected dataset in Ho Chi Minh City, Vietnam. The comparison results show that the hybrid model has the highest predictive performance in estimating the residual strength of corroded reinforced concrete beams compared with the individual models. This study demonstrates an effective predictive application of artificial intelligence for structural strength estimation early in building maintenance planning

Keywords: Artificial intelligence; corrosion; hybrid modeling; data mining