Long-Term Modal Analysis of Wireless Structural Monitoring Data from a Suspension Bridge under Varying Environmental and Operational Conditions: System Design and Automated Modal Analysis

Abstract

Structural monitoring systems installed on cable-supported bridges have the potential to generate large data repositories from which a deeper understanding of bridge behavior can be obtained. The core focus of this study is the use of response and environmental data collected by a permanent wireless monitoring system operating on the New Carquinez Bridge (Vallejo, California) since 2010. Given the large amount of data available, the study proposes an automated stochastic subspace identification approach for the extraction of bridge modal properties. The study fits logistic distributions to the extracted modal frequencies and modal damping ratios to provide statistical models for these important bridge properties. Low levels of modal damping well below 0.8% are reported for the majority of structural modes. Bridge modal properties exhibit sensitivity to the environmental and operational conditions of the bridge. Ridge regression and Gaussian process regression (GPR) are used to model the dependency of modal frequency on bridge environmental and operational conditions. The GPR models are shown capable of accurately modeling the relationship between modal frequency and bridge environmental and operational conditions.