Inclined gabion drops improve hydraulic performance relative to conventional USBR stilling basins under variable flow conditions

This study experimentally and numerically investigates the hydraulic performance of the Inclined Gabion Drop (IGD) as a sustainable energy dissipator compared with conventional USBR stilling basins. Laboratory experiments were conducted under varying flow conditions, drop heights, and angles to evaluate energy dissipation and hydraulic jump behavior. The results demonstrate that the porous gabion mattress substantially enhances energy dissipation through throughflow, interstitial friction, turbulence generation, and flow partitioning. Compared with USBR basins, the IGD reduced relative residual energy (E d /E u ) by 30–36% and downstream relative depth (y d /h) by 37–43%. The downstream Froude number decreased from 4.49 to 8.35 to 1.28–2.64, corresponding to an average reduction of approximately 64%, resulting in significantly shorter hydraulic jump lengths and improved flow stabilization. To complement the experimental analysis, three soft computing models (ANN-MLP, ICA-ANN, and ACO-ANN) were developed to predict yd/h and Ed/Eu. The ACO-ANN model exhibited superior predictive performance, achieving R² values of 0.9862 and 0.9918 for E d /E u and y d /h, respectively, outperforming benchmark models. SHAP-based interpretability and signed sensitivity analysis identified downstream Froude number (Fr d ) as the most influential parameter with a strong negative contribution, followed by relative critical depth (y c /h), while drop angle (β) showed negligible influence. These results confirm strong physical consistency between hydraulic behavior and machine learning predictions. Overall, the IGD configuration provides a robust, economical, and hydraulically efficient alternative to conventional energy dissipation structures. The proposed ACO-ANN framework offers a reliable surrogate model for rapid prediction and design optimization of IGD systems in hydraulic engineering applications.