Gabion weirs are environment-friendly structures widely used for irrigation and drainage network purposes. These structures' hydraulic performance is fundamentally different from solid weirs' due to their porosity and the existence of a through-flow discharge. This paper investigates the reliability and suitability of a number of Machine learning models for estimation of hydraulic performance of gabion weirs. Generally, three different Boosting ensemble models, including Gradient Boosting, XGBoost, and CatBoost, are compared to the well-known Random Forest and a Stacked Regression model, with respect to their accuracy in prediction of the discharge coefficient and through-flow discharge ratio of gabion weirs in free flow conditions. The Bayesian optimization approach is used to fine-tune model hyper-parameters automatically. Recursive feature elimination analysis is also performed to find optimum combination of features for each model. Results indicate that the CatBoost model has outperformed other models in terms of estimating the through flow discharge ratio (Q_in/Q_t) with R² = 0.982, while both XGBoost and CatBoost models have shown close performance in terms of estimating the discharge coefficient (C_d) with R² of CatBoost equal to 0.994 and R² of XGBoost equal to 0.992. Weakest results were also produced by Decision tree regressor with R² = 0.821 and 0.865 for estimation of C_d and Qin/Qt values.

Various regions are becoming increasingly vulnerable to the increased frequency of floods due to the recent changes in climate and precipitation patterns throughout the world. As a result, specific infrastructures, notably bridges, would experience significant flooding for which they were not intended and would be submerged. The flow field and shear stress distribution around tandem bridge piers under pressurized flow conditions for various bridge deck widths are examined using a series of three-dimensional (3D) simulations. It is indicated that scenarios with a deck width to pier diameter (Ld/p) ratio of 3 experience the highest levels of turbulent disturbance. In addition, maximum velocity and shear stresses occur in cases with Ld/p equal to 6. Results indicate that increasing the number of piers from 1 to 2 and 3 results in the increase of bed shear stress by 24% and 20% respectively. Finally, five machine learning algorithms, including Decision Trees (DT), Feed Forward Neural Networks (FFNN), and three Ensemble models, are implemented to estimate the flow field and the turbulent structure. Results indicated that the highest accuracy for estimation of U, and W, were obtained using AdaBoost ensemble with R2 = 0.946 and 0.951, respectively. Besides, the Random Forest algorithm outperformed AdaBoost slightly in the estimation of V and turbulent kinetic energy (TKE) with R2 = 0.894 and 0.951, respectively.