Abstract
Machine learning techniques are potent tools to predict structural failures and forecast future failure patterns. They can provide insights that are not immediately recognizable to human operators, and hence, enhance the maintenance decision-making process, leading to better safety and longevity of structures. This research employed a genetic algorithm based on an identification technique to estimate the location and depth of cracks in circular arches. The arches were made of homogeneous material and had a rectangular cross-section. The study adopted two different procedures for detecting cracks in the circular arches. The first method used natural frequency contours represented in a three-dimensional plot to identify the location and depth of cracks by intersecting a set of natural frequency contours. The second procedure utilized the regression method and an optimization technique to minimize the cost function objective function and determine the crack parameters. The influences of crack location and depth on the vibrational behavior of cracked arches were presented and discussed. Similarly, the results indicate a high level of consistency between the two techniques in predicting both crack size and location.
Acknowledgment
The authors are grateful to the reviewers for their valuable comments and suggestions that helped us to improve the quality of our article.
Authors’ contribution
All authors have contributed equally to the planning, programming, and writing of this work.
Omar 0utassafte is the corresponding author of this work, he planned, programmed, and wrote the article.
Ahmed Adri: he participated in the supervision of this work.
Yassine El Khouddar: he participated in the programming and writing of this article.
Issam El Hanati: participated in the writing and correction of this article.
Rhali Benamar: conception and analysis of the method of resolution of nonlinear dynamic equations.
Disclosure statement
The authors have no financial or non-financial interest to disclose.
Ethical statement
The manuscript has not been submitted or published anywhere and will not be submitted anywhere until the editorial process is completed.