Structural damage identification with uncertain modelling error and measurement noise by clustering based tree seeds algorithm. (15th April 2019)

• Record Type: Journal Article
• Authors: Ding, Zhenghao; Li, Jun; Hao, Hong; Lu, Zhong-Rong
• Is Part Of: Engineering structures. Volume 185(2019)
• Page Start: 301
• Page End: 314
• Publication Date: 2019-04-15
• Subjects: Tree seeds algorithm -- Clustering -- Structural damage identification -- Uncertainty -- Modal data -- Noise; Structural engineering -- Periodicals; Structural analysis (Engineering) -- Periodicals; Construction, Technique de la -- Périodiques; Génie parasismique -- Périodiques; Pression du vent -- Périodiques; Earthquake engineering; Structural engineering; Wind-pressure; Periodicals; 624.105
• DOI: 10.1016/j.engstruct.2019.01.118
• Languages: English
• ISSNs: 0141-0296

Abstract

Highlights: This paper proposes a clustering based tree seeds algorithm for damage identification. Uncertain modelling errors and measurement noises are considered. Benchmark studies are conducted to demonstrate the improvement of the approach. Numerical and experimental verifications are conducted to demonstrate the performance. The accuracy and robustness of the proposed approach are compared with existing methods. Abstract: This paper proposes a novel structural damage identification approach by using the clustering based Tree Seeds Algorithm, termed as C-TSA, taking into account of both the finite element modeling errors and measurement noise. In order to make the standard TSA more powerful and robust, K-means cluster technique is introduced into the standard TSA before starting the seeds search, which is beneficial to enhance the algorithm's global optimization performance. The objective function based on the modal data is formulated for structural damage identification. Numerical studies on benchmark functions and a 61-bar truss structure are conducted to investigate the accuracy and robustness of the proposed approach. The finite element modelling errors and noises in the measurement data are considered. Experimental verifications on a laboratory steel frame structure model is conducted to further validate the accuracy of the proposed approach. The results from the numerical and experimental studies are compared with those obtained from several latest evolutionary algorithms. The identification results demonstrate that the proposed approach is more competitive and robust for structural damage identification even considering the modelling errors and measurement noises.