Neural Network-Based Multi-Objective Optimization of Adjustable Drawbead Movement for Deep Drawing of Tailor-Welded...

Area of Science:

• Materials Science and Engineering
• Manufacturing Processes
• Computational Mechanics

Background:

• Deep drawing of tailor-welded blanks presents challenges in formability due to material variations.
• Controlling defects such as fracture and centerline deviation is crucial for successful manufacturing.

Purpose of the Study:

• To enhance the formability of tailor-welded blanks in deep drawing applications.
• To optimize the movement of an adjustable drawbead to simultaneously minimize fracture and centerline deviation.

Main Methods:

• Utilized finite element simulations to model deep drawing processes.
• Employed multi-objective optimization with artificial neural networks (ANN) and response surface method (RSM) for prediction.
• Applied the non-dominated sorting-based genetic algorithm II (NSGA-II) for optimization.

Main Results:

• Artificial neural networks demonstrated superior prediction accuracy compared to the response surface method.
• The non-dominated sorting-based genetic algorithm II effectively identified optimal parameters.
• The optimized adjustable drawbead significantly improved formability.

Conclusions:

• An adjustable drawbead, optimized through advanced computational methods, offers an effective solution for enhancing deep drawing formability.
• The integration of ANNs and NSGA-II provides a robust framework for optimizing complex manufacturing processes.