|
چکیده
|
To avoid a brittle failure of reinforced concrete (RC) frame, the ductility of beams has an important role during earthquake. The calculation of the accurate values of ductility of confined RC beams is complicated, and therefore, a direct and accurate approach is necessary. This study presents two approaches to predict the curvature ductility of RC beams using results of a new method. In the new method, a procedure was developed that curvature ductility of 250 confined and unconfined RC beams were calculated based on actual characteristics of a confined and unconfined compression concrete. In the first approach, an equation was proposed to predict curvature ductility of RC beam by using nonlinear regression based on result of 250 beams. In the second approach, an ANN model was trained to predict curvature ductility of RC beam based on result of 250 beams. An experimental database containing 51 RC beams as flexural test was used to validate the proposed equation and ANN model. A statistical analysis was then performed to evaluate and compare the proposed equation and ANN model for tested RC beams. The statistical analysis results show that ANN model can predict the curvature ductility of RC beam with slightly less scatter in comparison with the proposed equation. Overall, both approaches can predict curvature ductility of RC beam by a proper accuracy.
|