PI: Shamim Pakzad
Co-PI(s): N/A
University: Lehigh University
Industry partner: Specialty Engineering
Fatigue life estimation of bridges is critical for health prognosis and maintaining a resilient network of highway bridges. This typically entails studying the number of loading cycles that a bridge is subjected to, through the use of rain flow diagrams constructed from strain measurements. The traditional approach to accomplish this involves deploying a fixed strain sensor network on a bridge for continuous monitoring. The key drawbacks of this system are (a) the effort and cost associated with the installation and maintenance of a fixed strain monitoring system, and (b) the limited resolution and scalability of fixed sensor networks. To overcome the first issue, strain is estimated from the sensed acceleration response on a structure. However, this approach is still limited to the development of a relationship between responses acquired from a fixed accelerometer and strain sensor networks. Mobile sensing is a pathbreaking and emerging technology that addresses the scalability and resolution of continuous monitoring systems for bridge structures. This entails recording acceleration responses on smartphones from inside moving vehicles that traverse over a bridge. This facilitates a significantly enhanced response resolution along with the potential for using crowd-sensed data. However, the recorded response is a combination of bridge vibrations and vehicle dynamics. In this project, we will harness the power of mobile sensing for strain estimation and subsequent fatigue analysis of a bridge. In particular, we plan to develop a machine-learning framework that can extract the bridge acceleration portion from mobile sensed responses and estimate strains that can be used for fatigue life estimation. We will demonstrate the efficacy of the proposed framework using field data collected from bridges in the Lehigh Valley. For these field experiments, we rely on our industrial partner to provide data from a fixed sensor network as a reference for assessment of the quality of the estimates.