loading page

Artificial intelligence (AI) assisted fatigue fracture recognition based on morphing and Fully Convolutional Networks
  • +5
  • Yetao Lyu,
  • Zi Yang,
  • Hao Liang,
  • Beini Zhang,
  • Ming Ge,
  • Rui Liu,
  • Zhefeng Zhang,
  • Haokun Yang
Yetao Lyu
Hong Kong Productivity Council

Corresponding Author:[email protected]

Author Profile
Zi Yang
Hong Kong Productivity Council
Author Profile
Hao Liang
Hong Kong Productivity Council
Author Profile
Beini Zhang
The Hong Kong University of Science and Technology
Author Profile
Ming Ge
Hong Kong Industrial Artificial Intelligence and Robotics Centre (FLAIR) Hong Kong People’s Republic of China
Author Profile
Rui Liu
Institute of Metal Research Chinese Academy of Sciences
Author Profile
Zhefeng Zhang
Institute of Metal Research Chinese Academy of Sciences
Author Profile
Haokun Yang
Hong Kong Productivity Council
Author Profile

Abstract

Fatigue fracture is one of the most common metallic component failure cases in manufacturing industries. The observation on fractography can provide the direct evidence for failure analysis. In this study, an image segmentation method based on Fully Convolutional Networks (FCNs) was proposed to figure out the boundary between fatigue crack propagation and fast fracture regions from optical microscope (OM) fractography images. Furthermore, novel morphing-based data augmentation method was adopted to enable few-shot learning of sample images. The proposed framework can successfully segment two categories, namely the crack propagation and fast fracture regions, thus differentiating the boundary of two regions in one image. The artificial intelligence (AI) assisted fatigue analysis architecture can complete the failure analysis procedure in 0.5 second and prove the feasibility of fatigue failure analysis. The segmentation accuracy of developed network achieves 95.4% for the fatigue crack propagation region, as well as 97.2% for the fast fracture region, which possesses comparable accuracy to the segmentation results using Mask R-CNN Regional Convolutional Neural Network (Mask R-CNN), one state-of-the-art deep learning networks.
13 Sep 2021Submitted to Fatigue & Fracture of Engineering Materials & Structures
13 Sep 2021Submission Checks Completed
13 Sep 2021Assigned to Editor
18 Sep 2021Reviewer(s) Assigned
06 Oct 2021Review(s) Completed, Editorial Evaluation Pending
29 Oct 2021Editorial Decision: Revise Major
27 Dec 20211st Revision Received
27 Dec 2021Submission Checks Completed
27 Dec 2021Assigned to Editor
02 Jan 2022Reviewer(s) Assigned
10 Jan 2022Review(s) Completed, Editorial Evaluation Pending
10 Jan 2022Editorial Decision: Revise Major
26 Feb 20222nd Revision Received
26 Feb 2022Submission Checks Completed
26 Feb 2022Assigned to Editor
27 Feb 2022Reviewer(s) Assigned
16 Mar 2022Review(s) Completed, Editorial Evaluation Pending
16 Mar 2022Editorial Decision: Accept
Jun 2022Published in Fatigue & Fracture of Engineering Materials & Structures volume 45 issue 6 on pages 1690-1702. 10.1111/ffe.13693