Vibration Characteristics and Damage Detection of Composite Structures with Anisotropic Damage Using Unified Particle Swarm Optimization Technique

Abstract

Composite structures may undergo several types of damages such as delamination, fiber breakage, etc. due to the defects in the manufacturing process or fatigue loading during service which in turn affects the effective performance of the structure. The damage in the composite is anisotropic in nature whose extent is distributed in several orthogonal directions and the extent of damage in any orthogonal direction is independent of the other. The anisotropic damage is introduced into the composite using concept of reduction in stiffness due to reduction in effective load bearing area which is represented by damage parameters or variables. Vibration based damage detection is one of an effective scheme to detect damages due to its simplicity in implementation and ability of acquiring both the global and the local damage information of the structure. From the literature, it has been revealed that the frequency response function (FRF) based damage detection is more accurate compared to frequency based damage detection since it provide much more information on damage for a particular frequency range. A numerical procedure is developed in MATLAB to detect and quantify damages in composite structures using vibration response parameters with the help of unified particle swarm optimization (UPSO) algorithm. From the numerical and experimental results it is inferred that the presence of damage in composite structure affect the vibration characteristics such as frequency, mode shapes and FRF curves as well. Moreover, the damage parameters show strong orthogonality on vibration characteristics. It has been observed that, damage in the direction of fiber results in steeper deterioration of natural frequencies. The outcome shows that as the intensity of damage or the area of damage increases, the natural frequency deteriorates irrespective of ply orientation. The proposed methodology is able to detect and quantify damages accurately for various single and multiple damage scenarios.