Journal of Bridge Engineering

In bridge health monitoring (BHM), a prominent goal is to rapidly deliver assessment metrics for these essential and aging urban lifelines when subjected to natural hazard. A vibration-based machine learning (ML) BHM paradigm has been established over the past three decades to allow near-real-time automated health state classification, with a particular focus on the tasks of feature engineering and ML damage identification. This paper presents the human–machine collaboration (H-MC) framework to address challenges of this paradigm, especially in the context of reinforced concrete highway BHM. These challenges include specification of strong motion events, data multidimensionality, and ML model interpretability. The H-MC framework for BHM employs the techniques of multivariate novelty detection and probability of exceedance envelope models with ordinal filter-based feature selection to maximize the use of available data from both recorded and simulated events while maintaining the statistical and physical significance of the results. The framework is demonstrated using a numerical example and two case studies. The findings show the effectiveness of the proposed method for efficient damage assessment to facilitate rapid decision-making.

Abstract

Steel plates have traditionally been the reinforcement of choice for conventional elastomeric bridge bearings. In addition, these bearings are often employed under fixed boundary conditions (bonded application) as seismic isolators. The main objective of this study is to develop a new type of elastomeric bearing with improved lateral flexibility and superior seismic isolation efficiency. The new bearing is a partially bonded mesh-reinforced (MR) elastomeric bearing. MR bearings employ high-strength steel mesh reinforcement layers instead of steel-reinforcing plates. Additionally, the bearing is utilized in a partially bonded application; that is, only a limited region at the central portion of the bearing contact surfaces is bonded to the top and bottom supports. Given this specific boundary condition and the bending flexibility of the mesh reinforcement layers, the MR bearing experiences lateral rollover deformations under shear loads. During lateral rollover deformation, the upper and lower surfaces of the bearing partially roll off the contact supports. This experimental study compared the cyclic lateral responses of bonded plate-reinforced (PR) bearings (as reference bearings) and their partially bonded MR-bearing counterparts. The elastomer material properties and geometrical characteristics of the two bearing types were identical. The experimental results suggest that partially bonded MR bearings are feasible, perform more flexibly in the lateral direction, and exhibit greater energy-dissipation capability than PR bearings.

Practical Applications

This study focuses on a new type of elastomeric bridge-bearing isolator designed to be more flexible horizontally and better at absorbing the input energy of earthquakes. The new bearing isolator was constructed using laminated rubber material layers that were bonded to high-grade steel mesh reinforcement layers instead of conventional steel plates. The mesh reinforcement is rigid when stretched in a flat direction; however, it can bend easily. The bearing isolator was only partially bonded to its supports, which allowed it to deform more freely in the horizontal direction. During horizontal deformation, the upper and lower surfaces of the bearing partially roll off the contact supports. In this study, the new bearing type was compared with the conventional (steel plate reinforced) type, and it was found that the new bearing type is more effective at absorbing seismic excitations and can provide greater flexibility in the horizontal direction. This new bearing type is called a partially bonded mesh-reinforced elastomeric bearing.

Concrete barriers are used for different purposes in the highway inventory. Different test levels have been specified to help design these barriers to function as expected for their intended purpose of use. Reinforced concrete (RC) Jersey barriers are commonly used to act as intervening structures protecting bridge piers from colliding vehicles during crash events. Thus, they are subjected to vehicular collision force (VCF). According AASHTO, RC barriers used to protect bridge piers should have a minimum height of 1,067 mm and survive test level 5 (TL-5). Many existing barriers that are currently used in roadways do not meet this requirement, hereby, they are referred to as “substandard” barriers. While these barriers are found to be inadequate to solely protect bridge piers against VCF, recent research proved that their presence to intervene in vehicle–pier collisions will contribute to reducing the severity of the crash effects and, consequently, reduce the VCF demand on bridge piers. Therefore, it is important to have a comprehensive understanding about the behavior of substandard barriers during high-level crash events. This will help in proposing methodologies to upgrade these barriers in an efficient and affordable way. In this study, the behavior of a representative substandard barrier, subjected to TL-4 and TL-5 collision events, was investigated through a matrix of dynamic finite-element simulations using the general-purpose finite-element analysis (FEA) program LS-DYNA, version R10.0. The study considered various boundary conditions of the barrier. The investigated quantities were the energy dissipation, velocity reduction, contact force absorption, and lateral displacement. The results showed that the maximum energy dissipation and velocity reduction were about 45% and 30%, respectively. In addition, the highest contact force demand and maximum lateral displacement were 580 kN and 800 mm, respectively. Specific reduction values are addressed in detail in the paper.

Understanding expected structural behavior enables the early identification of potential structural issues or failure modes, allowing for timely intervention and maintenance. Guided by this premise, this paper proposes the Bayesian dynamic regression linear model (BDRLM) tailored for predicting the real-time performance of cable-stayed bridges in the face of nonstationary sensor data. Drawing from local linear regression techniques, BDRLM integrates probability recurrence, exhibiting heightened sensitivity to structural behavior shifts. This capability fosters real-time behavior prediction and anomaly detection. Embracing a more pragmatic approach, the model treats the sensor measurement error as an unknown factor. This strategy, complemented by Bayesian probability recursion, refines the error's probabilistic distribution parameters, aligning the prediction process more congruently with field practices. Then, based on structural health monitoring (SHM) data of an actual bridge, the extreme stress of the main girder monitoring sections is dynamically predicted, and a dynamic warning threshold based on prediction updates is proposed. Finally, the time-varying reliability indices of the main girder are predicted and estimated. The effectiveness of the proposed method is validated through an actual application and comparisons of several other commonly used methods. This achievement can provide a theoretical basis for bridge early warning and maintenance with prediction requirements.

Direct shear strength (DSS) of the precast concrete dry joints (PCDJs) is one of the critical issues for ensuring the structural integrity and performance of the precast segmental concrete bridges (PSCBs). However, the effects of shear key configurations, such as size, shape, spacing, and number of the PCDJs, were studied within narrow value ranges and were not well considered in the widely used AASHTO formula, resulting in unconservative predictions of the formula for multiple-keyed PCDJs. This study investigated the effect of shear key configurations within wider value ranges and incorporated a reduction factor into the AASHTO formula through a newly established parametric finite-element (FE) simulation. The simulation included an automatic procedure of model generation and calculation to derive the DSS of PCDJs based on 13 input parameters of the joint and was validated by a comprehensive database including 111 DSS results of PCDJs collected from existing literature. A parametric study was then conducted on a customized joint, which allows for exploring the PCDJs with large key sizes or numbers. Finally, a reduction factor was proposed to the AASHTO formula based on the test and simulation data considering the effects of key configurations. Results demonstrate that the proposed FE simulation is able to consider the reduction effect from the key configurations and reaches an excellent agreement in predicting the DSS of PCDJs in the database with more accurate mean values (MVs) (0.95 and 0.98), higher R² values (0.63 and 0.69), and lower root-mean-squared error (RMSE) values (1.00 and 0.79 MPa) for both single-keyed and multiple-keyed joints compared with the results of the AASHTO formula. The parametric study revealed that the ASSHTO formula becomes increasingly unconservative with the increase of key numbers. The reduction factor needs to be included in the formula when multiple-keyed or large-size keyed joints are considered. The AASHTO formula with the reduction factor leads to a conservative prediction performance in the multiple-keyed joints, which is similar to that in the single-keyed joints.

The performance-based seismic design (PBSD) approach is implemented to achieve the desired structural performance over a wide range of seismic hazard levels. It requires a set of targeted performance levels and their corresponding limits to be defined. Because the current codes and guidelines do not prescribe these limits for different performance levels for old bridges with seismic deficiencies, such as inadequate ductility and low shear strength, this study aims to develop them. In this study, quantitative damage states that are expressed as drifts and damage indices (DIs) at various performance levels are developed using incremental dynamic analyses for retrofitted bents. Four retrofit options: (1) steel; (2) carbon–fiber-reinforced polymer (CFRP); (3) concrete; and (4) engineered cementitious composite (ECC) jackets are considered in this study. The concrete and longitudinal reinforcement of all bents cracked and yielded at limiting drifts of 0.06% and 0.38%, respectively. In addition, the ECC-jacketed bent experienced core crushing of the concrete at the highest limiting drift of 4.16%. In addition, a detailed example complements this study, which presents how retrofitting could be designed by considering the target seismic performance that uses the proposed damage states. The first-mode spectral accelerations of the bents were the optimum intensity measures (IMs) to study their relative performance for noncumulative and cumulative damage measures (DMs) at various hazard levels. Drift is considered noncumulative, and the DI that includes the combined effect of maximum drift and absorbed hysteretic energy is considered cumulative. The steel jacket was the most effective when decreasing the median maximum drift of the retrofitted bent, and the ECC jacket reduced the median DI of this type of bent the most.