This study investigates utilizing raw strain measurements from Bridge Weigh-In-Motion (BWIM) systems for long-term structural health monitoring of highway bridges. Using one year of continuous traffic data collected from an instrumented bridge on the Trans-Canada Highway in New Brunswick, the research demonstrates that select classes of vehicles can serve as consistent probes of bridge response, even without prior knowledge of vehicle weights.
Local deck strain signals were processed to detect vehicle presence and axle configurations using a combination of thresholding, peak detection, and a supervised Support Vector Machine (SVM) classifier. Initial analyses focused on 7-axle trucks; however, due to observed variability in their strain responses, a Leveraging BWIM System Traffic Data for Long-Term Structural Health Monitoring of Highway Bridges more refined selection of log trucks—a specialized vehicle class with highly uniform configurations—was undertaken. The results showed that log trucks produced strain responses with significantly lower variability, confirming the critical role of meticulous vehicle selection in achieving reliable monitoring.
Temperature effects were also evaluated by correlating monthly average strain responses with ambient temperature data. A moderate and statistically significant positive correlation was observed, highlighting the necessity of accounting for environmental factors when interpreting structural response trends over time.
Overall, the findings suggest that BWIM installations, traditionally used for traffic weight estimation, can be repurposed for efficient, continuous bridge monitoring. Strategic selection of vehicle types and environmental correction procedures are essential for maximizing the reliability of such monitoring frameworks. This approach offers a scalable, low-cost alternative to conventional Structural Health Monitoring (SHM) systems through utilizing existing BWIM systems.