NEURAL NETWORK APPLICATION FOR STRUCTURE DESIGN OPTIMIZATION OF THIN-WALL STRUCTURES

Neural networks are trained to predict the response of a thin wall tube under dynamic impact loading then they are integrated with an optimization algorithm to improve the crashworthiness design of the thin wall tube. LS-DYNA is used to simulate the tube's response under dynamic impact while MATLAB is used to train the neural networks and the optimization algorithm. The results show that the suggested approach succeeded in improving the thin wall tube design at an affordable computational cost. It is suggested that the approach can be expanded for the design improvement of more complex structures.