A Multi-step Method For Bridge Damage Identification

Along with construction of some long-span bridge, the technology of bridge damage identification became one of important problems for civil engineer. On the basis of collection and research of the technology about bridge model analysis and theory about artificial neural networks(ANN), a multi-step method for bridge damage identification has been brought forward. This method can identify the damage of continuous-beam bridge successfully. In details, the following aspects are investigated.1. This paper reviews the developing history and present status, and explores the method about bridge damage identification systematically, and point out that single method of damage identification has some shortage for bridge damage.2.The paper bring forward a multi-step method for bridge damage identification. This is based on the technology about bridge model analysis and theory about artificial neural networks. This method has two steps, the first is using technology of model analysis to detect the damage position, the second is using artificial neural networks to identify the damage degree.3.This paper sets up the finite element model of continuous-beam bridge tocompare of two index——curvature mode difference and flexibility curvaturedifference. In the paper we can find that flexibility curvature difference is an advanced and applied index.4.In the paper, a BP networks has been set up beads on the theory about artificial neural networks to identify the damage degree. At last, a example of continuous-beam bridge is given to demonstrate the validity of this method.