Structural health monitoring of aircraft through prediction of delamination using machine learning

Rajeswari D; Osamah Ibrahim Khalaf; Srinivasan R; Pushpalatha M; Habib Hamam

doi:10.7717/peerj-cs.1955

Structural health monitoring of aircraft through prediction of delamination using machine learning

PeerJ Comput Sci. 2024 Mar 27:10:e1955. doi: 10.7717/peerj-cs.1955. eCollection 2024.

Authors

Rajeswari D¹, Osamah Ibrahim Khalaf², Srinivasan R³, Pushpalatha M³, Habib Hamam⁴

Affiliations

¹ Department of Data Science and Business Systems, School of Computing, College of Engineering and Technology, SRM Institute of Science and Technology, Kattankulathur, Tamilnadu, India.
² Department of Solar, Al-Nahrain Research Center for Renewable Energy, Al-Nahrain University, Jadriya, Baghdad, Iraq.
³ Department of Computing Technologies, School of Computing, College of Engineering and Technology, SRM Institute of Science and Technology, Kattankulathur, Tamil Nadu, India.
⁴ Université de Moncton, Moncton, Canada.

Abstract

Background: Structural health monitoring (SHM) is a regular procedure of monitoring and recognizing changes in the material and geometric qualities of aircraft structures, bridges, buildings, and so on. The structural health of an airplane is more important in aerospace manufacturing and design. Inadequate structural health monitoring causes catastrophic breakdowns, and the resulting damage is costly. There is a need for an automated SHM technique that monitors and reports structural health effectively. The dataset utilized in our suggested study achieved a 0.95 R2 score earlier.

Methods: The suggested work employs support vector machine (SVM) + extra tree + gradient boost + AdaBoost + decision tree approaches in an effort to improve performance in the delamination prediction process in aircraft construction.

Results: The stacking ensemble method outperformed all the technique with 0.975 R2 and 0.023 RMSE for old coupon and 0.928 R2 and 0.053 RMSE for new coupon. It shown the increase in R2 and decrease in root mean square error (RMSE).

Keywords: Delamination; Machine learning; Prediction; Stack ensemble; Structural health monitoring.

Grants and funding

The Natural Sciences and Engineering Research Council of Canada (NSERC) and New Brunswick Innovation Foundation (NBIF) provided financial support for the global project. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.