چكيده به لاتين
Nowadays, timely identification of damage in structures in preventing future damage and improving the performance of structures to increase the safety of residents and the useful life of structures, has a very important and significant role. In recent years, many efforts have been made to reach this goal in order to provide diverse and effective methods in this regard. Among the methods presented, model updating methods based on modal parameters have a large contribution to the studies due to their proper functioning. In most of the proposed methods of identifying damage, the basis of the work is to compare the structure and modal parameters related to it before and after the damage. The lack of precise structural information, limited structural modal information, and the existence of errors in measured data are among the limitations in these methods. Therefore, in all of the proposed methods, taking into account the constraints mentioned is essential and inevitable.
In this study, four methods for identifying damage to structures have been proposed. In each of the proposed methods, an objective function based on the modal parameters is set. Pseudo Modal strain energy, Mode shape, static displacement and generalized flexibility matrix were used as modal parameters in proposed methods. In order to optimize proposed methods, four meta-heuristic optimization algorithms which are Moth Flame Optimization (MFO), Grasshopper Optimization Algorithm (GOA), Whale Optimization Algorithm (WOA) and Moth Swarm Algorithm (MSA) have been used. In order to evaluate the performance of proposed methods, limited modal data and noise effects on modal parameters obtained for similar the actual condition of the vibration of the structure was considered.
The function of proposed method along with optimization algorithms presented on ten numerical examples, including two-dimensional and three-dimensional trusses, two-dimensional moment frames and shear frames were evaluated. Also, in a laboratory setting, a six-story shear frame was constructed and investigated for evaluating the performance of proposed methods. All the results obtained in different methods of detecting the extent and severity of damage in numerical and experimental examples had an error of less than 2.5%. The results obtained in this study indicate the exact function of the proposed methods and the appropriate convergence of the proposed optimization algorithms to detect the location and extent of damage.
