An improved needle arrangement design method of low-damage structures for needled composites based on machine learning

  • Yunchao Qi*
  • , Zihui Hao
  • , Guodong Fang
  • , Jun Liang
  • , Guizhe Zhao
  • *Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

The composites prepared by needled preforms as reinforced skeleton have been widely used in aerospace structures. The in-plane long fibers are deflected to the thickness direction in the needled area due to the extrusion of the needle, so the interlaminar mechanical properties will be enhanced. However, the long fibers are damaged by the needle, so the in-plane mechanical properties of the material will be severely degraded. It is an urgent problem to be solved for reducing the damage and fracture of long fibers in the needling processes. The needle arrangements have an important effect on the distribution of the needled areas and the damage of the long fibers. The needled areas in the composites can be divided into fiber bending areas and fiber breaking areas. In order to reduce the needled area of long fiber fracture, an improved needle arrangement design method based on the representative volume cell model and machine learning model is proposed. The low-damage material structure and the needle arrangements are designed to ensure that the number of the fiber breaking areas is minimized. The feasibility and validity of this method are verified by in-plane and interlaminar shear experiments. The results show that the composites designed by this method have more excellent in-plane shear strength properties. The in-plane shear strength of the composite prepared by the designed process has increased by 15.50 % and 28.03 % compared with the material strength of the contrast other processes. Meanwhile the interlaminar shear strength of the composite will not significantly reduce. Therefore, the needle arrangement designed by this approach can effectively reduce the damage degree of long fibers during the production process of needled composites. This method also ensures the in-plane and interlaminar mechanical properties of materials simultaneously. This study can effectively reduce the design cost of needled composite materials and promote the application of fiber-reinforced composites in the aerospace industry.

Original languageEnglish
Article number113787
JournalMaterials Today Communications
Volume49
DOIs
Publication statusPublished - Dec 2025
Externally publishedYes

Keywords

  • Low-damage structure
  • Machine learning model
  • Needle arrangement
  • Needled composites
  • Shear strength properties

Fingerprint

Dive into the research topics of 'An improved needle arrangement design method of low-damage structures for needled composites based on machine learning'. Together they form a unique fingerprint.

Cite this