Railway simply supported steel truss bridge damage identification based on deflection

To propose a novel damage identification method, this study firstly used deflections and SVM to establish the damage identification model. It has significant theoretical significance and practical value to timely master the bridge structure's health condition and identify the damage location and damage degree. There are five load cases, such as, one locomotive run on the bridge, two locomotives coupling run on the bridge, three locomotives coupling run on the bridge, a train with one locomotive run on the bridge, a train with two locomotives run on the bridge. When the load cases respectively act on the 64 m railway simply supported steel truss bridge, the change percentages of the lower chord panel points maximum deflections and the beam end maximum displacement are calculated. The percentages are independent variables, the damage location and the damage degree are dependent variables, the identification models are established respectively using C-SVC and Ε-SVR to identify the damage location and the damage degree. These two models all have good anti-noise ability and good generalization. © 2013 Asian Network for Scientific Information.