Soft Computing Techniques for Predicting Aeration Efficiency of Gabion Stepped Weir

Abstract

The paper investigates the modeling performance of aeration efficiency (E20) of gabion stepped weir with varying steps (no step, one step and two steps) by using soft computing techniques. The output values of aeration efficiency are computed using Random Forest (RF) and Artificial Neural Network (ANN). The actual aeration efficiency is taken by conducting experiments in laboratory flume and taking mean size, porosity, discharge, drop height and Reynolds number as input parameters. For comparing the results by these soft computing techniques standard statistical parameters such as coefficient of correlation (CC) and root mean square error (RMSE) are utilized. RF obtained the coefficient of correlation and root mean square error value of 0.8738 and 0.0649 compared to the values of 0.9357 and 0.0550 respectively attained by ANN. The findings of this paper will help in selecting better modeling technique for predicting the aeration efficiency of gabion stepped weir.