Decentralized damage detection for nonlinear structures

Traditional damage detection methods for nonlinear structures are usually based on simplified models, such as mass-spring-damper models or shear-building models, and thus insufficient to predict the vibration responses of a real structure. Nonlinear finite element method is adopted in this paper to simulate the complex nonlinear behavior of structures accurately, and a decentralized damage detection method is later proposed for identification. The structure is divided into a number of small substructures according to its structural configuration. The structural parameters of each substructure are updated using the vibration responses of the substructure with Newton-SOR method. OpenSees is used to calculate the structural responses and the associated response sensitivity. A two-bay three-floor reinforced concrete structure subjected to earthquake loading is employed to verify the accuracy of the proposed method. The parameters in material constitutive model (such as compressive strength, initial tangent stiffness, and yielding stress) are identified successfully. It demonstrates that the proposed method is able to detect damage in nonlinear structures under seismic loading.