| Title |
Prediction of Shear Strength of Steel Fiber Reinforced Concrete Beams Using Artificial Neural Networks |
| Keywords |
Steel fiber ; Shear strength ; Neural network ; Back-propagation ; Bayesian regularization |
| Abstract |
Many investigators identified share force characteristics through experiments with use of materials, shapes, mixture ratios, and steel ratios so as to identify the share force characteristics of steel fiber enforced concrete. However, the experiments to identify the characteristics of steel fiber enforced concrete require large expenses and long time for equipment arrangement in general, the characteristics of steel fiber enforced concrete have been used with limitation. Therefore, for various applications, an easier approaching method was required to expect the share force characteristics of unsaturated soils. In consideration of such situation, a method to expect the share force characteristics of steel fiber enforced concrete was suggested and applied in this study as neural network theory. Back-propagation algorithm was applied as learning algorithm of neural network and learning was performed so as to be converged within the range of 0.001. In addition, nonlinear function was uses as objective function and the problem of overfitting was resolved with more generalized method by adopting Bayesian regularization technique as generalization process. |