Many damage detection algorithms based on vibration characteristics have been and are being developed for application to civil engineering structures. While each of them offers a degree of success at locating damage in a structure, each method also has certain limitations. To address this fact, a methodology which uses several methods combined to obtain a better and more consistent overall prediction is proposed. This paper presents the development of the combined methodology. The combined methodology uses the stochastic damage locating vector technique (SDLV), the Proportional Flexibility Matrix (PFM) technique, which are both based on the DLV technique, a flexibility-based method. A sensitivity-based model updating technique (SBMU) is then used to validate and improve the prediction. The methodology is applied to a 2D truss example and a 3D frame example. The results show that the combined methodology offers a more consistent and reliable prediction than either of the techniques individually.
展开▼
