NEURAL NETWORK APPLICATION FOR FINE-BLANKED EDGE QUALITY IN ROLLED STEEL SHEETS

Abstract

The backing plate plays a crucial role in drum brakes, pushing the brake shoe against the drum for effective vehicle braking and road safety. Manufactured from steel sheets, it exhibits properties akin to a composite material, with characteristics varying based on directionality. The backing plate's profile features a continuous series of arcs, making the fine-blanked edge quality contingent on the arc radius. This research delved into fineblanked manufacturing through a combination of experimentation and simulation to discern the impact of the arc radius on edge quality. Employing a scanning electron microscope for edge quality investigation and benchmarking the finite element model against it, the study ensured a comprehensive understanding. The wellaligned finite element model served as input data for training an artificial neural network (ANN), specifically engineered to accurately estimate backing plate edge quality. This ANN is anticipated to be instrumental in designing future steel plate profiles boasting multiple arcs, offering precision in the manufacturing process for enhanced edge quality.