A New Damage Detection Approach Under Variable Environmental or Operational Conditions

The basic idea of vibration-based damage identification approaches is that damage causes change in vibration response of structure. So monitoring the vibration response characteristics can be helpful in damage detection.  The main limitation in such methods is that these characteristics are also affected by the Environmental and Operational Variability (EOV) that can be incorrectly known as structural damage or sometimes cover actual damages. This paper aims to propose an innovative approach to detect and locate damage considering the EOV conditions. In this regard, an Independent Component Analysis (ICA) based Blind Source Separation (BSS) approach is employed to remove the EOV influences from the time history response of the structure. The beneficial of using the ICA-based BSS method is that there is no need to measure the environmental/operational conditions. Moreover, it is able to remove EOV influences using a limited group of response data monitored during different environmental and operational conditions. Time series analysis is then performed to extract damage-sensitive features. Finally, a statistical tool is employed to damage identification and localization by using EOV independent features. Two recognized benchmark structures are employed for verifying the accuracy of the proposed approach. Results indicate that the proposed method is a time-saving tool and efficiently successful in damage assessment of structures under EOV.