Seismic response prediction of FRC rectangular columns using intelligent fuzzy-based hybrid metaheuristic techniques

This research study focused on the dynamic response and mechanical performance of fiber-reinforced concrete columns using hybrid numerical algorithms. Whereas test data has non-linearity, an artificial intelligence (AI) algorithm has been incorporated with different metaheuristic algorithms. About 317 datasets have been applied from the real test results to detect the promising factor of strength subjected to the seismic loads. Adaptive neuro-fuzzy inference system (ANFIS) was carried out as an AI beside the combination of particle swarm optimization (PSO) and genetic algorithm (GA). Extreme Machine Learning (ELM) was also performed in order to approve the obtained results. According to the findings, it is demonstrated that ANFIS-PSO predicts the lateral load with promising evaluation indexes [R-2 (test) = 0.86, R-2 (train) = 0.90]. Mechanical performance prediction was also carried out in this study, and the results showed that ELM predicts the compressive strength with promising evaluation indexes [R-2 (test) = 0.66, R-2 (train) = 0.86]. Finally, both ANFIS-GA and ANFIS-PSO techniques illustrated a reliable performance for prediction, which encourage scholars to replace costly and time-consuming experimental tests with predicting utilities.

Automated summary

This research study focused on the dynamic response and mechanical performance of fiber-reinforced concrete columns using hybrid numerical algorithms. By incorporating an artificial intelligence (AI) algorithm with different metaheuristic algorithms, the study analyzed non-linear test data. The findings showed that ANFIS-PSO predicts lateral load accurately, while ELM effectively predicts compressive strength. Both ANFIS-GA and ANFIS-PSO techniques demonstrated reliable performance for prediction.