Diagnosis of Damages in Beam Structures using Vibration Parameters and Artificial Intelligence Techniques

Abstract

In the present analysis, special attention has been focused for detecting the damages present in Al, composite and steel beam structures by comparing the characteristics of damaged and undamaged state of the structures. In the current research, damage detection of damaged cantilever and fixed-fixed beam is carried out using numerical, Finite element analysis (FEA), fuzzy logic and neural network techniques. Numerical analysis has been performed on the cantilever beam & fixed-fixed beam with damage in the transverse direction to obtain the vibration parameters of the beam members utilizing the expression of strain energy release rate and stress intensity factor. The presence of damage in a structural member introduces local stiffness that affects its dynamic characteristics. The local stiffness matrices have been determined using the inverse of local dimensionless compliance matrix for finding out the deviations in the vibrating signatures of the damaged beam structures from that of the intact beams. Finite Element Analysis has been carried out to derive the vibration indices of the damaged structures using the overall stiffness matrix, total stiffness matrix, stiffness matrix of the intact beams. It is concluded from the conducted research that the performance of the damage diagnosis techniques depends on several factors for example, the material type, the number of sensors used for acquiring the dynamic response, position and severity of damages. Different artificial intelligent model based on fuzzy logic, neural network have been designed using the estimated vibration signatures for damage diagnosis in beam structures with higher precision and remarkably low calculating time.