Automatic uncoupling of massive dynamic strains induced by vehicle- and temperature-loads for monitoring of operating bridges

Abstract The massive dynamic strain of operating bridges, measured by a structural health monitoring (SHM) system, is naturally the coupling of different strain components associated with various sources, among which vehicle- and temperature-loads are the major contributors to strain components. Real-time uncoupling of vehicle- and temperature-induced strains is considerably needed when processing massy dynamic strain data acquired by SHM systems for monitoring of bridges relying on specific strain components. Currently, the empirical mode decomposition (EMD) is a typical method to uncouple the vehicle- and temperature-induced strains. Nevertheless, this method is capable of uncoupling these strains in a non-real-time manner due to involvement of manual intervention for setting certain parameter values in implementing the method. The key to real-time monitoring is to achieve uncoupling in an automatic way. This study proposes an enhanced EMD method, termed Auto-EMD method, based on a progressive utilization of EMD, Hilbert marginal spectrum, and Gaussian mixture model clustering. The Auto-EMD method embodies the predominant feature of automatically uncoupling of vehicle- and temperature-induced strains of operating bridges. The effectiveness of the proposed method is verified by numerical models of bridges subject to both vehicle- and temperature-loads, and the robustness to measurement noise is also demonstrated. Furthermore, the applicability of the proposed method in engineering practice is validated using dynamic strain data captured from the Sutong Yangtze River Highway Bridge. The results show that the proposed method can uncouple in real time the temperature- and vehicle-induced strains in a superior intelligent mode compared to that of the existing EMD. The Auto-EMD method provides a viable paradigm of uncoupling massive dynamic strain data for SHM applications of operating bridges.