Data Analysis and Detection Methods for On-Line Health Monitoring of Bridge Structures

摘要

Developing an efficient structural health monitoring (SHM) technique is important for reducing potential hazards posed to the public by damaged civil structures. The ultimate goal of applying SHM is to real-time detect, localize, and quantify the accumulated structural damage. Traditionally, damage assessment techniques are primarily based on the observation of the changes of model parameters, such as modal frequencies and vibration shapes. However, several investigations have pointed out that this technique could be efficient only if very precise measurements are available or large level of damage exists. It is obvious that a more robust damage assessment algorithm must be developed. The objective of this research is to develop a new technique, named matched damage processing (MDP) technique, to perform damage detection, localization, and severity estimation for bridge structures subjected to traffic loading. The MDP technique builds upon the concept of matched field processing already developed in the underwater acoustics community. In short, the MDP technique compares measured responses collected from the physical structure with multiple modeled results calculated based on various damage locations and severity. The comparison is displayed in a contour or image plot with damage location and severity as the independent coordinates. The concept, effectiveness and robustness of the MDP technique are numerically demonstrated via a simple example where a simply supported beam is used to model a bridge and the traffic loading is modeled as a constant weight moving with constant speed across the bridge. A localized reduction of structural stiffness is applied to represent the occurrence of damage to the bridge.