New predictive models via gene expression programming and multiple nonlinear regression for SFRC beams
Yasser Sharifi, Adel Moghbeli
Abstract
A soft computing based study aimed to predict the shear capacity of steel fiber RC (SFRC) beams is qualified and new formulae based on gene expression programming (GEP) and nonlinear multiple regression (NMR) are suggested. To extend the proposed formulae, extensive experimental data collected from existing studies in the literature were considered. The data utilized in the GEP and NMR models are ordered in a format of six input such as concrete compressive strength, effective depth, shear span to the effective depth ratio, the ratio of the longitudinal reinforcement, length of fiber to the diameter of itself, and fiber content by volume. The accuracy of the developed formulae is verified versus the experimental data and the rates of efficiency and performance are compared with those provided by suggested equations presented by some of the researchers. The developed models that investigated in present study demonstrated that the GEP and NMR methodologies displayed strong potential for estimating the shear capacity of SFRC beams.