Thickness Optimization and Experimental Validation of Incremental Collar Forming of Explosive Welded Al/Cu Bimetal Sheet with an Obround Hole Using Finite Element, Whale Algorithm and Neural Network

Bimetal sheets are prominently expanded in industries due to their unique properties including high strength, low density, and corrosion resistance compared to single-layer sheets. On the other hand, incremental forming is found to be one of the most suitable methods to form complex parts. In this regard, a numerical and experimental study was conducted to consider the behavior of explosively welded aluminum–copper sheets. A nonlinear explicit solver was employed to simulate the incremental collar forming process. Experimental tests were performed to validate thickness variation, dimension and tool average force. A coupled artificial neural network and finite element method are utilized to generate an objective function of thickness reduction. The new metaheuristic whale optimization algorithm using a trained neural network (ANN–WOA) is utilized to improve sheet thickness distribution by finding optimum parameters. The developed artificial neural network could reliably predict thickness reduction in Al/Cu bimetal sheet during incremental collar forming while minimum thickness reduction occurs in the circular area of obround shape with less than 4% deviation from ANN results. It is also shown that optimum incremental forming parameters of obround shapes with length to diameter ratio between 2 and 3 are very close.