Predicting the Peak Strength of FRP Confined Rectangular Concrete Columns Using Multi-Layered Perceptron Artificial Neural Networks

Due to various reasons including an increase in the applied load level due to changes in the occupancy of the structures during their service life, inaccurate estimation of the applied loads, or damages to structural components caused by different reasons, Reinforced Concrete (RC) structures may require rehabilitation and/or strengthening. Nowadays, the application of Fiber Reinforced Polymers (FRP) for confining RC columns is a common method that provided desirable results in increasing the strength and ductility of the structural members. It is proved by previous studies that the effectivity of FRP confinement in rectangular columns is not comparable with circular sections, due to the non-uniform distribution of confinement pressure in rectangular sections. Many of the previous studies on FRP confined RC columns were focused on the proposition of relation to predicting confined concrete strength, and due to the above-mentioned issue, most of these relations were proposed for columns with circular sections, and many of the existing models for rectangular columns do not provide sufficient accuracy. Therefore, the present study is an attempt to develop a predictive model based Artificial Neural Network (ANN) for estimating the peak confined strength of FRP-confined rectangular RC columns using a large database of experimental results collected from the literature. A comparison of the performance of the developed model against some of the other existing predictive relations demonstrates the higher accuracy of the proposed model.

Keywords:

Confined concrete , Column , Rectangular section , FRP , Artificial Neural Network.