Effect of Temperature on Bridge Health Monitoring Using Bayesian Bridge Weigh-in-Motion

Abstract

The importance of bridge health monitoring (BHM) has grown significantly due to its critical role in ensuring bridge safety and longevity with increasing traffic loads and aging infrastructure. Bridge weigh-in-motion (B-WIM) technology has emerged as a promising alternative to traditional methods for assessing the structural health of bridges. The bridge influence line (BIL), backbone of B-WIM system, contains substantial structural information, and any change observed in BIL can be used for BHM. Despite advancements, transitioning BHM from research to practice is hindered by environmental and operational conditions (EOCs) variability, which impacts BIL estimation accuracy and, thus, BHM effectiveness. Simultaneously, temperature effects stand out as a paramount concern for long-term BHM and are anticipated to introduce additional environmental variability, particularly through temperature fluctuations. Recent studies suggest that bridge deformation due to temperature-induced expansions and contractions can be equal to or greater than that induced by vehicular load. This can potentially affect the BIL estimation and mask the effect of damage, leading to false damage alarms. Several studies have explored various BHM approaches using B-WIM system in the recent decade. However, the specific impact of temperature on B-WIM-based BHM remains understudied. This highlights the need to evaluate the performance of B-WIM-based BHM methods under the influence of temperature variations. To address this, the study aims to evaluate BHM performance under temperature variations using temperature data of Patna city collected over the last two decades. This approach recognizes that substantial temperature variations can significantly affect bridge responses. Additionally, the study incorporates advancements in B-WIM systems, particularly using the evolving Bayesian framework, which is promising in improving B-WIM algorithms. Bayesian B-WIM (BB-WIM) incorporates prior knowledge with real-time observations to update probability distributions, enhancing accuracy and enabling uncertainty quantification. The methodology employs a 3D finite element (FE) model of a real bridge, incorporating various temperature conditions and damage severities. The FE model is validated by matching modal frequencies up to the fourth mode with those of the real bridge within acceptable limits. Different bridge response time histories (RTHs) caused by vehicular movements over the bridge are obtained at every quarter location under a longitudinal beam below the deck. The obtained bridge RTHs, such as strain, acceleration, and deflection, are then fed into the BB-WIM algorithm to estimate the BIL. Finally, gross vehicle weight (GVW) -based damage indicator values are calculated for performance assessment under temperature variation. The findings from this study are then summarized with a focus on highlighting the future aspects and practical implications for improving BHM systems.