Numerical Analysis and Perdiction of The Mean Velocity in The Intake and Rivers Using Artificial Neural Networks (ANN) andANSYS-CFX

The simplest method of deviating water in irrigation systems, agriculture is using dividing flow channels. Intake is considered a type of dividing flow channel. One of the essential hydraulic parameters in intake efficiency increase is measuring the mean velocity in intake. The mean velocity in an intake is predicted using ANNMLP neural network model for the ratio of different widths (w) in this study. To accomplish that the flow field was first threedimensionallysimulated through using ANSYS-CFX software in an intake with 90 degrees of diversion and then the flow mean velocity is predicted through using numerical model and the artificial neural network ANNMLP. The artificial neural network used includes 4 inputs, (y*oordinates, the ratio of the branch channel width to the main channel (w) and linear mean velocity which has been measured by numerical modelsin two different places of the intake channel (v*line 1and v*line 2). The results form comparing the numerical results and the experimental results indicated the proper accuracy of the numerical model in predicting the flow field specification in the intake and comparing the results from the predictions made by ANN-MLP model with the experimental results shows the acceptable accuracy of the artificial neural network in predicting mean velocity of flow in the intakes and for different width ratios

Keywords:

dividing flow channels , mean velocity , ANSYSCFX software , ANN-MLP model