Research On Damage Identification Method For Bridge Structure Using Incomplete Parameter Information

Bridge is an important part of infrastructure, and plays important role in economic and society development. With the enhancement of comprehensive national strength, bridge construction has gain impressive achievements. However, the increasing vehicle load and material deterioration make the damage and failure of bridge unavoidable. The health monitoring and damage identification for existing old bridges have important practical significance to ensure the safety of the bridge operation.In vibration-based damage identification(VBDI) method, the damages are identified based on the change in dynamic parameters. However, damage identification methods are affected by uncertain factors including test noise, incomplete parameter information and missing basic model. These result the parameters used to identify damage incomplete and reduce the reliability in damage identification. How to identify structural damage with the incomplete parameter information becomes the key to promote the damage identification methods to be used in practical project. Therefore, aiming at the uncertain condition including test noise, incomplete parameter information and missing basic model, the damage identification method with better robustness is proposed based on fuzzy theory and data fitting reconstruction technology. The following work is carried out:(1) The research background and significance of bridge damage identification with incomplete parameter information are detailed introduced. The deficiencies of traditional damage identification methods in dealing with uncertain problems are analyzed. On the basis of analysis for the researches at home and abroad, the main research content in this paper is presented.(2) The basic fundamentals of particle swarm optimization and fuzzy C-means cluster algorithm are stated. Combing these two algorithms, the fuzzy C-means cluster optimized by particle swarm algorithm(PSO-FCM) is proposed. Considering the damage identification index is similar in similar damage condition, the PSO-FCM algorithm can be used to identify the damages in bridge. Therefore, the method and technological process for bridge damage identification are established based on PSO-FCM algorithm.(3) In order to solve the test noise and data missing in bridge damage identification, the damage identification method based on PSO-FCM algorithm is proposed. The Gaussian white noise added in the original mode data is used to simulate test noise, and mode data missing is used to simulate information incomplete. The proposed PSO-FCM algorithm is to identify damages in these two uncertain conditions. The identification results verify the validity and accuracy of PSO-FCM in damage identification with parameter noise and data missing conditions.(4) The uniform load surface curvature(ULSC) parameters of basic model are reconstructed by polynomial fitting. Uniform load surface curvature difference(ULSCD) established by ULSC parameters before and after damage is used to determine the damage locations. On the basis of damage location, the damage severity is identified by PSO-FCM algorithm. Finite simulation verified the accuracy of proposed method in identifying damage with basic model missing condition.