Muthanna Journal of Engineering and Technology
Volume (11), Issue (2), Year (30 December 2023), Pages (09-29)
DOI:10.52113/3/eng/mjet/2023-11-02/09-29
Research Article By:
Ali Abdullah Hassooni and Hussein Yousif Aziz
Corresponding author E-mail: ali.abdullah1230@gmail.com
ABSTRACT
In recent decades, Concrete filled steel tube (CFST) columns have been widely utilized in construction due to their high strength, ductility, energy absorption, fire resistance, and cost reduction due to the absence of formwork. Estimating the Axial compressive capacity (ACC) of short rectangular CFST columns has been the subject of numerous experiments. In this study, artificial neural network was used to make predictions regarding the ACC of CFST columns. Multi fed forward back propagation and Adaptive neuro fuzzy inference system were used. 512 experimental tests were collected from the literature. One thousand models for multi fed forward back propagation and one hundred for adaptive neuro fuzzy inference system were trained and tested. artificial neural network models are evaluated by using statistical analysis to validate and test the prediction models. The best models were selected by using the Root mean square error (RMSE) and Coefficient of determination (R2). Using the RMSE, R2, and Mean absolute percentage error (MAPE), the best models were compared to design code formulas. As a result, the best model performed much better in every performance measure. The best model for multi fed forward back propagation has better performance in comparison with the best model for adaptive neuro fuzzy inference system. For interested users and researchers, a graphical user interface was created using the best model for multi fed forward back.
Keywords: Artificial neural network, Multi fed forward back propagation, Adaptive neuro fuzzy inference system, Concrete filled steel tube columns, Axial compressive capacity.