Damage identification of structures using incomplete mode shape and improved TLBO-PSO with self-controlled multi-stage strategy

• Published in: Structures (Oxford) Vol. 35; pp. 1101 - 1124
• Main Authors: Das, Subhajit, Dhang, Nirjhar
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2022
• Subjects: Damage identification; Hybrid ITLBO-PSO; Inverse problem; Metaheuristics; Structural health monitoring; Hybrid ITLBO-PSO; Damage identification; Structural health monitoring; Metaheuristics; Inverse problem
• ISSN: 2352-0124; 2352-0124
• DOI: 10.1016/j.istruc.2021.07.089

The present work presents an efficient multi-stage optimization method-based damage detection method for truss and frame structures equipped with a limited number of sensors. In this approach, a Finite Element (FE) model is developed to simulate the response of the actual structure. The limited sensor condition for this FE model is achieved by the modal reduction method. A comparison study among three well-established modal reduction methods has been performed, and Iterated Improved Reduction System (IIRS) approach has been selected for the present study. Next, the damage identification problem is defined as an unconstrained optimization problem. The objective function of the optimization problem is formulated using the weighted linear combination of the frequencies and mode shapes obtained from the actual damaged structure and reduced FE model. This objective function is minimized by the improved version of hybrid Teaching Learning Based Optimization - Particle Swarm Optimization (ITLBO-PSO) utilizing a self-controlled multi-stage (SCMS) strategy. In this method, the SCMS strategy automatically reduces the search dimension of the optimization problem in each stage. Four examples with different damage scenarios from the relevant literature are considered in the present study to demonstrate the efficacy of the proposed method. The proposed method results for both with noise and without noise are compared with existing literature and nine other well-established algorithms. The results show that the proposed ITLBO-PSO with SCMS strategy identifies damages with adequate precision and outperforms the other algorithms regarding the accuracy and computational cost.