A hierarchical neural network for identification of multiple damage using modal parameters

• Main Authors: S.J.S., Hakim, J.M., Irwan, M.H.W, Ibrahim, S., Shahidan, S.S., Ayop, N., Anting, T.N.T, Chik
• Format: Conference or Workshop Item
• Language: English
• Published: 2023
• Subjects: TA329-348 Engineering mathematics. Engineering analysis
• Online Access: http://eprints.uthm.edu.my/11643/1/P16682_74e0e3ab01b75e48402da271f894d8d4%209.pdf; http://eprints.uthm.edu.my/11643/
• Institution: Universiti Tun Hussein Onn Malaysia

Artificial neural networks have been applied extensively in recent years due to their excellent performance in pattern recognition, which is useful for detecting damage in structural elements. The application of multiple damage cases by an ensemble neural network using dynamic parameters of structure is very limited. Therefore, in this paper, an ensemble neural network based on damage identification techniques was developed and applied for damage localization and severity identification of quad-point damage cases in I-beam structure. Experimental modal analysis and finite element simulation were carried out for I-beam with four-point damage cases to generate the modal parameters of the structure. Based on the results, it is found that the ensemble neural networks achieve a high detecting accuracy and good robustness of quad-point damage cases in I-beam structures