Open Access

ARTICLE

The Flow Behavior Investigation of 5754 Aluminum Alloy Based on ACO-BP-ANN

Fengjuan Ding¹, Lu Suo², Tengjiao Hong^1,3,*, Fulong Dong¹, Dong Huang¹

1 College of Intelligent Manufacturing, Anhui Science and Technology University, Chuzhou, 233100, China
2 College of Information and Data Science, King Mongkut’s University of Technology North Bangkok (KMUTNB), Bangkok, 10250, Thailand
3 Graduate School, Stamford International University, Bangkok, 10250, Thailand

* Corresponding Author: Tengjiao Hong. Email:

(This article belongs to the Special Issue: Applications of Neural Networks in Materials)

Computers, Materials & Continua 2025, 85(3), 4551-4570. https://doi.org/10.32604/cmc.2025.069565

Received 26 June 2025; Accepted 09 September 2025; Issue published 23 October 2025

Abstract

The complex phenomena that occur during the plastic deformation process of aluminum alloys, such as strain rate hardening, dynamic recovery, recrystallization, and damage evolution, can significantly affect the properties of these alloys and limit their applications. Therefore, studying the high-temperature flow stress characteristics of these materials and developing accurate constitutive models has significant scientific research value. In this study, quasi-static tensile tests were conducted on 5754 aluminum alloy using an electronic testing machine combined with a high-temperature environmental chamber to explore its plastic flow behavior under main deformation parameters (such as deformation temperatures, strain rates, and strain). On the basis of true strain-stress data, a BP neural network constitutive model of the alloy was built, aiming to reveal the influence laws of main deformation parameters on flow stress. To further improve the model performance, the ant colony optimization algorithm is introduced to optimize the BP neural network constitutive model, and the relationship between the prediction stability of the model and the parameter settings is explored. Furthermore, the predictability of the two models was evaluated by the statistical indicators, including the correlation coefficient (R²), RMSE, MAE, and confidence intervals. The research results indicate that the prediction accuracy, stability, and generalization ability of the optimized BP neural network constitutive model have been significantly enhanced.

---

Keywords

---