DAMAGE IDENTIFICATION IN REINFORCED CONCRETE BEAMS BY FINITE ELEMENT MODEL UPDATING USING PARALLEL AND HYBRID GENETIC ALGORITHMS

Abstract

Finite element (FE) model updating belongs to the class of inverse problems in mechanics and is a constrained optimization problem. In FE model updating, the difference between the modal parameters (the frequencies, damping ratios and the mode shapes) obtained from the FE model of the structure and those from the vibration measurements are minimized within an optimization algorithm. The design variables of the optimization problem are the stiffness reduction factors, which represent the damage. In this study, the Genetic Algorithms (GA), the Parallel GA, the local search algorithms, the Trust Region Gauss Newton, the Sequential Quadratic Programming, the Levenberg–Marquardt Techniques and the hybrid versions of these methods are applied within the FE Model Updating Technique for updating the Young's modulus of different FEs of a reinforced concrete beam. Different damage scenarios and different noise levels are taken into account. The results of the study show that the local search algorithms cannot detect, locate and quantify damage in reinforced concrete beam type structures while the GA together with the hybrid and the parallel versions detect, localize and identify the damage very accurately. It is apparent that the hybrid GA & Trust Region Gauss Newton Technique is best in terms of the computation speed as well as accuracy.