NUMERICAL ANALYSIS AND PREDICTION OF THE VELOCITY FIELD IN CURVED OPEN CHANNEL USING ARTIFICIAL NEURAL NETWORK AND GENETIC ALGORITHM

This paper presents numerical analysis and prediction of flow field in a 90 degrees bend using Artificial Neural Networks (ANN) and Genetic Algorithm (GA). Firstly, a 3D Computational Fluid Dynamics (CFD) model is used to investigate the flow patterns and velocity profiles. Numerical simulation in two phases is done using the ANSYS-CFX software and k-epsilon turbulence model is used to solve turbulence equations. The results show secondary flow and centrifugal force influenced flow pattern and have good agreement with experimental data. Then two similar ANNs are trained based on GA and Back-Error Propagation (BEP) technique for velocity prediction in different sections of bend and their test results are compared with each other and with actual data. Since obtaining experimental data in every point of channel is not easy, ANN is used to obtain the velocity in some sections where experimental data are not available, and the results are compared with CFX's result.