چكيده به لاتين
Due to occurrence of structural damages arising from sudden and special loadings such as earthquake, undesirable changes are created. In this situation it is necessary to use such methods for damage detection in order to reparation and rehabilitation of damage location before more serious damages or demolition. One of the most noticeable subjects in civil engineering, mechanic engineering and aerospace science from past to now is detection of location and severity of structural damage in early stages and their health monitoring. Nowadays researches on damage detection methods based on dynamic parameters have developed much, because these methods can be used for continuous and online check of structures and prevent happening more damages. The main idea of these methods is that each structure can be considered as a dynamic system with unique mass, stiffness and damping changing when structure damages. So by evaluating differences of modal parameters between model of intact structure and damaged structure, damages locations and severities can be found. Model updating technique is one of the dynamic methods.
In this thesis, a new method is investigated wherein using generalized flexibility matrix in the objective function as an optimization problem for damage detection. Damage is defined in the form of stiffness reduction of element in this research. Damage location and severity are specified based on change in generalized flexibility matrix. In generalized flexibility matrix, the effect of higher modes decreases in comparison with other modal parameters and even flexibility matrix, and lower modes are used. Also in order to minimize objective function, Flower pollination algorithm is employed that is one of the metaheuristic optimization algorithm. For evaluating proposed method, three numerical examples are considered include a 21 elements 2D truss, a 21 elements 2D frame and a shear frame with 15 stories that on which one of those, three different damage scenarios include presence of single or multi damaged elements at noisy and without noise conditions are applied. The amount of noise that is considered for truss and frame is 5 percent and 3 and 8 percent for shear frame. Results indicate that proposed method is satisfactory, precise and fast for detection of damage location and severity in structures.
