Data-Driven Flexural Stiffness Model of FRP-Reinforced Concrete Slender Columns

Abstract

Predicting the nominal axial capacity of slender fiber-reinforced polymer (FRP) reinforced concrete (RC) columns is dependent majorly on their flexural stiffness. However, the current design provisions do not incorporate design equations to estimate the flexural stiffness of slender FRP-RC columns yet due to the limited research work on this aspect. Although limited research studies proposed flexural stiffness models for slender FRP-RC columns, these models show inaccurate results with large discrepancies. This study, therefore, compiles and analyzes a surveyed database of 53 tested slender FRP-RC columns found in the literature to construct a simplified and accurate model to predict the flexural stiffness of the slender FRP-RC columns. In this approach, the experimental-based flexural stiffness values of the tested specimens were used to build a nonlinear regression flexural stiffness model and examine the influence of the critical design parameters affecting this value. As a result, the proposed model showed a strong agreement with the experimental flexural stiffness values evidenced by having the least root mean square error (RMSE) compared to the other proposed models in the literature. Moreover, the proposed model was theoretically evaluated accounting for the second-moment order effect by which a data set of 36,000 cases were generated and compared to the results of the proposed model to increase its creditability and repeatability. Moreover, a design example was presented to quantify the difference in predicting the flexural stiffness of the slender FRP-RC columns between the proposed and available models. Accordingly, the proposed model revealed better representation of the flexural stiffness with higher accuracy compared to the available models which will help the engineers to accurately design the FRP-RC columns.