Updating the Mathematical Models of Bridges Using Data-driven Techniques

摘要: Structural health monitoring (SHM) and damage detection based on vibrationbased techniques have attracted significant attention in the past. Many of these methods rely appropriate Finite Element (FE) or other mathematical model a structure updating using field data. There are number existing for such models identifying system parameters as stiffness mass dynamic response structure. These can be broadly categorized Physics-based data-driven models. Since, problem is an inverse problem, often provide nonunique solutions. On hand, like artificial neural networks (ANN) Genetic Algorithm (GA) utilizes differences data patterns structural order to update parameters. In this study, FE two bridges, Pre-Stress Concrete Box (PSCB) girder bridge Void Slab Bridge been constructed updated measured vibration It should mentioned that both bridges relatively new there no reported them. The were instrumented with accelerometers modal vectors frequencies determined ambient tests. aim research compare physics-based driven from While method was found more precise than GA NN case studies, latter flexibility when sets incomplete noisy. doi: 10.12783/SHM2015/156