Open Access
Subscription or Fee Access
Understanding Environmental Effect on Building Vibration for Structural Health Monitoring Using Event Detection and Causal Analysis
Abstract
Structural damage diagnosis algorithms often involve analyzing vibration responses to extract information about dynamic characteristics closely related to integrity. The vibration data are, however, often influenced by various environmental effects, which may degrade diagnosis performance and result in erroneous decisions. This paper focuses on understanding the consequence of significant environmental effects, such as trains passing nearby, on building vibration data using event detection methods and causal analysis. The event detection methods identify significant events using wavelet analysis, which is effective for decomposing non-stationary signals. The causal analysis allows us to investigate wave propagation patterns in structures by quantifying causal dependencies between measurements collected from different locations using directed information. These methods are applied to acceleration data collected from 40 accelerometers deployed to an 11-story office building located next to railway. The results show clear patterns of causal dependencies among vibration data from different locations in the building, and their patterns change under different environmental conditions. It provides insights about environmental loading effects on building vibration that can be helpful for improving the accuracy of damage diagnosis