Prediction of the aerodynamic behavior of a rounded corner square cylinder at zero incidence using ANN

The aerodynamic behavior of a square cylinder with rounded corner edges in steady flow regime in the range of Reynolds number (Re) 5–45; is predicted by Artificial Neural Network (ANN) using MATLAB. The ANN has trained by back propagation algorithm. The ANN requires input and output data to train the network, which is obtained from the commercial Computational Fluid Dynamics (CFD) software FLUENT in the present study. In FLUENT, all the governing equations are discretized by the finite volume method. Results from numerical simulation and back propagation based ANN have been compared. It has been discovered that the ANN predicts the aerodynamic behavior correctly within the given range of the training data. It is additionally observed that back propagation based ANN is an effective tool to forecast the aerodynamic behavior than simulation, that has very much longer computational time.