Open AccessMachine Learning as a Tool for Estimating Discharge Coefficient for Rectangular Weir with Multiple Slots
Author(s) -
Mona A. Hagras
Publication year - 2020
Publication title -
international journal of engineering and advanced technology
Language(s) - English
Resource type - Journals
ISSN - 2249-8958
DOI - 10.35940/ijeat.d7348.049420
Subject(s) - weir , discharge coefficient , artificial neural network , correlation coefficient , coefficient of determination , support vector machine , decision tree , mathematics , computer science , set (abstract data type) , algorithm , statistics , artificial intelligence , engineering , nozzle , mechanical engineering , cartography , geography , programming language
In terms of higher water demand due to increases of agriculture lands which depend on irrigation channels, there is a need to increase the efficiency of weirs already constructed on those channels. The most reasonable solution is to modify the existing weirs instead of replacing them by new ones to avoid costly solutions. In this study, a potential of using different Machine Learning regression algorithms has been investigated to estimate the coefficient of discharge Cd for rectangular weir with multiple circular slots. Using experimental data set, three Machine Learning regression models; Decision Tree Regressor (DTR), Artificial Neural Networks (ANN) and Support Vector Machine Regressor (SVR) were developed and compared to find most suitable algorithm. Based on the simulation results of the three developed models, it was found that the ANNs algorithm is the superior one which can be used to estimate discharge coefficient Cd for rectangular weirs with multiple circular slots. It gives the highest matching between measured and predicted values with correlation coefficient (R2 ) value of 0.759, minimum MAE with value 0.001 and minimum MSE with value 0.022. Finally, an equation using ANNs is presented to estimate the discharge coefficient.