Prediction model for compressive arch action capacity of RC frame structures under column removal scenario using gene expression programming

• Published in: Structures (Oxford) Vol. 25; pp. 212 - 228
• Main Authors: Azim, Iftikhar, Yang, Jian, Javed, Muhammad Faisal, Iqbal, Muhammad Farjad, Mahmood, Zafar, Wang, Feiliang, Liu, Qing-feng
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.06.2020
• Subjects: Compressive arch action; Gene expression programming; Prediction model; Progressive collapse; RC beam-column substructure; Progressive collapse; Compressive arch action; RC beam-column substructure; Prediction model; Gene expression programming
• ISSN: 2352-0124; 2352-0124
• DOI: 10.1016/j.istruc.2020.02.028

•GEP is used for predicting the CAA capacity of RC beam-column substructures.•L/d is the most influential while ρ is the least influential parameter.•The GEP based model is simple, accurate and applicable to a wide range of data.•Sensitivity and parametric analyses are also carried out. Compressive arch action (CAA) is one of the important resistance mechanisms against progressive collapse in reinforced concrete (RC) frame buildings at small deformations. In this study, the distinguishing features of Gene Expression Programming (GEP) are utilized in order to establish a prediction model for the CAA capacity of RC beam-column substructures. The proposed equation correlates CAA capacity with six input parameters, i.e., compressive strength of concrete (fꞌc), double-beam span-to-depth ratio (Ld), width of beam (b), top and bottom longitudinal reinforcement ratio at the joints (ρt &ρb, respectively), and the ratio of axial restraints stiffness to the axial stiffness of beam (α). The data for the above-mentioned variables is collected from different studies to develop the proposed model and subsequently implemented for the verification purpose. Different statistical and external validations criteria recommended in literature are carried out to assess the performance of the developed model. Furthermore, sensitivity and parametric analyses are also performed using criteria available in literature. The performance of the genetically developed GEP model was checked and compared with models developed by using regression techniques. The GEP model demonstrated superior performance to linear and non-linear regression methods. The results show that the developed model is successfully capable of evaluating the CAA capacity of RC beam-column substructures and is incredibly imperative for prediction rationales.