Improving axial load-carrying capacity prediction of concrete columns reinforced with longitudinal FRP bars using hybrid GA-ANN model

This study aims to develop a hybrid machine learning model, so-called Genetic algorithm–Artificial neural network (GA-ANN), for efficiently predicting the axial load-carrying capacity (ALC) of concrete columns reinforced with fiber reinforced polymer (FRP) bars. For that, a set of 280 experimental test data is collected to develop the GA-ANN model. Seven code-based and empirical-based formulas, which were proposed by various design codes and published studies, are also included in comparison with the developed machine learning model. The performance results of GA-ANN are compared with those of seven previous equations. Statistical properties including goodness of fit (R2), root mean squared error (RMSE), and a20 - index are calculated to evaluate the accuracy of those predictive models. The comparisons demonstrate that GA-ANN outperforms other models with very high R2 and a20 - index values (i.e., 0.993 and 0.89, respectively), and a small RMSE (148 kN). Moreover, the influence of input parameters on the predicted ALC is assessed. Finally, an efficient graphical user interface tool is developed to simplify the practical design process of FRP-concrete columns. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.