This study utilizes artificial intelligence to predict the optimum lateral confinement ratio for concrete columns through an innovative dual-phase approach. First, a comprehensive artificial neural network (ANN) model was developed using analytically generated data encompassing diverse parameters: section geometry, dimensions, concrete unconfined compressive strength, transverse steel yield strength, longitudinal reinforcement ratio, vertical spacing between ties, and confinement configurations. This model was carefully trained and optimized to ensure reliable performance across a wide range of conditions. Secondly, the model was validated against actual experimental results collected from various studies available in the literature, achieving an impressive ability to accurately predict the confinement effectiveness coefficient, with a mean square error of 0.01 and a mean absolute error of 0.07. This methodology effectively bridges the gap between theoretical models and experimental findings while providing a practical and efficient tool for estimating confinement effects, eliminating the need for extensive laboratory testing, complex mathematical formulations, or time-consuming procedures. |
