Aeration of square jets in an open channel: experimental analysis and modeling

ABSTRACT A high concentration of dissolved oxygen is essential for the maintenance of healthy aquatic ecosystems. Aeration studies have been conducted in both closed systems and open channels utilizing conventional hydraulic structures. However, the feasibility of aeration through screens with square openings has not yet been explored. This study aimed to evaluate the aeration efficiency (E20) in an open channel system using screens with square jets. Five input parameters were analyzed: angle of inclination, number of square jets, discharge, cumulative hydraulic radius, and Froude number. The results showed that Nj, α, and Q significantly influence E20. The highest E20 recorded was 37% on level terrain with moderate discharge in the open channel. This method has the potential to enhance oxygen concentrations in rural regions where skilled labor and mechanised systems may be challenging. The study also focused on identifying suitable soft computing models for predicting E20. Five machine learning approaches were employed: artificial neural network, random forest with Bagging, Gaussian process utilizing the Pearson VII function kernel (GP_PUK), support vector machine with the PUK kernel function (SVM_PUK), and a radial basis function (SVM_RBF). The GP model employing the PUK kernel function demonstrated superior performance.