Özet:

This paper focuses on the prediction of the ultimate compressive capacity of axially loaded concrete-filled steel (CFST) tube section columns using machine learning (ML) techniques. The use of CFST columns in the construction industry has been popular due to their superior structural performance both under axial loads as well as under lateral seismic loads. Studies carried out on the contribution of CFST members on lateral seismic resistance have revealed that the ideal combination of stiffness and ductility inherent in concrete and steel, respectively, results in superior performance under lateral loads. The evaluation of the load-carrying capacity of structural members ML-based predictive techniques has been popular among researchers. In this study, the case of CFST columns under axial loading is studied. The dataset needed for the prediction was acquired from the related existing research which provided the results of 215 experimental studies. The ML techniques that were used included two prevalent techniques namely Artificial Neural Networks (ANN) and Random Forest (RF). In this study, the axial compressive capacity was predicted using these techniques and finally, the performances of the techniques were compared. Overall, the RF prediction technique was found to be in very close agreement with the experimental results acquired from the literature.

Anahtar Kelimeler: