Innovative Models for Capacity Estimation of Reinforced Concrete Elements in Terms of Soft Computing Techniques

Abstract

In reinforced concrete (RC) moment frames, the performance of the whole structure to provide a suitable strength against lateral loads depends on the behavior of each structural element. Therefore, the capacity of such members plays a vital role in the final response of RC buildings. However, the nonlinear nature of complex materials like concrete and its combination with other parameters makes the calculation of capacity more difficult. An estimation of the strength is a critical challenge in civil engineering with many applications, especially for damage detection, retrofitting, and strengthening programs. In this article, new computational frameworks are evaluated to introduce innovative approaches with the aim of determining the structural capacity of beams, columns, and joints in RC frames using soft computing (SC) techniques. Neuro-fuzzy models, group method of data handling (GMDH), and, also, three different structures of neural networks are presented and discussed. To this end, three collections of experimental datasets are gathered to assess the models. Moreover, mathematical equations are extracted from the considered systems for practical usages. Finally, the best predictive model for each structural element is proposed in detail. A comparison study between the results obtained from the presented methods and the laboratory values indicates that the proposed models can estimate the capacity of the elements with high accuracy and can be used as a powerful tool for evaluating the capacity of the structural members of RC frames.