Optimization of Power Spinning Process Parameters for D506A New-Generation Ultra-High Strength Steel Based on the Improved BOA-BP Algorithm
Year: 2026
Type: conference-paper
Abstract: This research is dedicated to optimizing the power spinning process parameters of the new-generation ultra-high-strength steel D506A. Three key process variables were selected, namely the thinning ratio, feed ratio, and first-pass allocation ratio, with the wall-thickness springback of the workpiece adopted as the evaluation indicator. A regression model for predicting the wall-thickness springback of spun components was established using the Box-Behnken response surface methodology. In addition, an improved BOA-BP algorithm was developed by integrating the advantages of the Butterfly Optimization Algorithm (BOA) and Back Propagation (BP) neural network, which is characterized by rapid convergence, high prediction accuracy, and strong adaptability. This hybrid algorithm was utilized to achieve high-precision springback prediction for D506A steel power-spun parts. The results indicate that the optimal combination of process parameters is determined as follows: a thinning rate of 82.4%, a feed ratio of 0.8 mm·r⁻¹, and a first-pass allocation ratio of 50%, which results in a minimum springback of 0.90 mm. Experimental verification of the improved BOA-BP algorithm shows that all relative prediction errors are controlled within 6%, which proves the high reliability of the model and its considerable guiding significance for the industrial production of D506A ultra-high-strength steel components.
Authors: Feng Yang, Changlin Du
Subfield: Mechanical Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7204833760
FWCI:
Cites:
Cited by:
Related to: