Mean Flow Velocity Prediction of Lateral Intakes by Using Computational Fluid Dynamic , Artificial Neural Network and the Flowmeter Measurements

Lateral intakes are one of the most common structures of dividing the flow in irrigation and drainage systems. Due to complexity of the velocity profile in divide zone, measurement of mean flow velocity is become very difficult. In this paper the velocity profile of lateral intakes were calculated whit high accuracy by using of artificial neural network. To do this, the following steps have been taken: (1) Computational fluid dynamic model of lateral intakes in various wide ratios were modeled and validated with a published experimental study. The results shown that the numerical model has high accuracy in modeling the flow of lateral intakes. (2) By using the computational fluid dynamic model, the velocity that measured with a hypothetical flowmeter that placed at the middle of the cross section were extracted. (3) A multilayer perceptron model were designed to predicting the mean flow velocity by using of the flowmeter measured velocity, width ratio and longitudinal coordinate. The results shown that using of combination of flowmeter measurement and artificial neural network could predict the accurate mean flow velocity in lateral intakes.