Structural damage identification based on change in geometric modal strain energy–eigenvalue ratio

Abstract

This study presents a new damage identification method to locate and quantify damage using measured mode shapes and natural frequencies. A new vibration parameter, ratio of geometric modal strain energy to eigenvalue (GMSEE), has been developed and its change due to stiffness reduction has been formulated using a sensitivity matrix. This sensitivity matrix is estimated with measured modal parameters and basic information of the structure. For damage identification, firstly, the locations of damage and the correlative damage extents are identified by maximizing the correlation level between an analytical GMSEE change vector and a measured one. Herein, the genetic algorithm, which is a powerful evolutionary optimization algorithm, is utilized to solve this optimization problem. Secondly, the size of damage can be estimated using the proposed GMSEE technique and compared with a conventional technique using frequency change. A numerical 2D Truss bridge is used to demonstrate the performance of the proposed method in identifying single and multiple damage cases. Also, practicality of the method is tested with a laboratory eight degree-of-freedom system and a real bridge. Results illustrate the high capability of the method to identify structural damage with less modeling efforts.