Journal of Bridge Engineering

Conventional methods to identify influence lines, which are essential in design and evaluation of bridges, use contact sensors involving high upfront and operational costs. This paper presents an approach to identifying influence lines based on computer vision measurements. The approach integrates vision-based identification of vehicle types, estimation of vehicle loads, bridge displacement measurement, and Bayesian parametric estimation. A you only look once version 4 (YOLOv4)—a real-time object detector—with a convolutional block attention module is trained to identify vehicle types and estimate vehicle loads. Bridge displacement measurements provide dynamic deflections, which are then used to analyze the influence line through Bayesian parametric estimation. The performance of this approach was evaluated through laboratory and field experiments with different types of vehicles and driving speeds. The results show that the errors were up to 4.88% for laboratory experiments and up to 11.48% for field experiments. This research provides findings that will help with the practices of condition monitoring and assessment of highway bridges.

The use of hollow concrete columns (HCCs) as piers and piles for bridge applications is widespread due to their higher load-carrying capacity, stiffness, and strength-to-mass ratio compared to the solid section. This study aimed to examine the behavior of HCCs reinforced with glass fiber–reinforced polymer (GFRP) bars and spirals under different loading conditions, analyze the impact of various parameters on their load-carrying capacity, and expand the research database with numerous load–moment interaction diagrams. Ten large-scale GFRP-HCCs, which had a height of 1,500 mm and inner/outer diameters of 113/305 mm, were tested under different levels of eccentricity (concentric, 8%, 16%, 33%, and 66%). A parametric study was conducted to examine the effects of the hollow ratio, longitudinal reinforcement ratio, bar compressive strength, longitudinal reinforcement type, and concrete compressive strength on HCC behavior. The study highlighted the importance of considering the compressive strength of the longitudinal GFRP bars because neglecting it underestimated the axial load and bending moment capacities of the HCCs. The results revealed that initial eccentricity had a greater impact on bending moment than second-order effects.

Cross-frames, which primarily serve as stability braces in steel I-girder systems during erection and deck construction, are susceptible to load-induced fatigue problems in finished composite structures if not properly detailed. Due to eccentricity in the connections, cross-frames are often subjected to significant in-plane and out-of-plane bending effects that can affect the fatigue resistance. This paper documents the results of a full-scale laboratory study that evaluates the fatigue performance of common welded cross-frame configurations and details. Eighteen unique cross-frame panels, including various typical and proposed connection details, were fabricated and tested for failure under cyclic loading that simulated the effect of live load traffic on composite girder systems. Based on the results of this experimental study, many details that are prevalently used in practice were shown to exhibit poor fatigue characteristics due to (1) the presence of localized stress concentrations near welds and/or (2) significant bending stresses not explicitly considered in analysis and design. Conversely, the proposed details mitigated these effects and were subsequently found to improve the fatigue life of cross-frame systems.

In this study, the investigations on the mechanism and aerodynamic countermeasures (ACs) of vortex-induced vibration (VIV) for a narrow Π-shaped girder (NG) section with the aspect ratio of B/D = 5 (B is the width and D is the height of the main deck) are conducted. First, the wind tunnel tests on VIV of the NG section with/without ACs are carried out. Furthermore, numerical simulations of VIV of the NG section with/without AC-VII, namely combined AC of modified fairings and inclined deflectors, for wind attack angle of +3° are conducted via fluid–structure interaction, respectively. Moreover, the mechanism of VIV of the NG section is analyzed from the characteristics of vorticity, pressure, and vortex-induced force (VIF) distributions, respectively. The research results show that the original NG section exhibits obvious vertical and torsional VIV for wind attack angles of +3° and +5°, respectively. Moreover, the ACs of low vertical stabilizers and horizontal separators, which are commonly applied to suppress the VIV responses of the wide Π-shaped girder section with aspect ratio B/D = 10–14, cannot effectively suppress the VIV responses of the NG section with B/D = 5. However, the AC-VII can effectively suppress the VIV responses. For the original NG section, the shear layer separation vortices that are generated on the windward side will extend to the leeward of the section and form the periodic shedding vortex, which is the mechanism of the VIV of the NG section. However, the AC-VII effectively suppresses the scale and vorticity intensity of the upper dominant vortex, avoiding the appearance of the vortex acceleration region at the leeward of the NG section. Therefore, the vorticity intensity of the upper and lower dominant vortexes is reduced during the shedding process, and the peak value of the fluctuating pressure in the VIV’s driving region is reduced by 86%.

Overheight truck collision with a bridge superstructure can cause extensive structural damage and result in severe traffic congestion. Placing protection beams in front of the fascia girder is one of the common countermeasures to guard against such events. However, little research has been done on this topic in the past and there are no standard guidelines for designing these protection beams. High-fidelity numerical simulations are conducted to formulate a demand model for protection beams subjected to overheight truck impact scenarios involving three different trailer models (representing soft, semirigid, and rigid impactors), a range of truck weights and velocities, and a number of beam sections and lengths. A validated demand model is proposed to represent the loading pulses caused by the different types of impactors. Two performance levels for the protection beam are suggested. Simulation data are used to propose a reliable machine-learning (ML) model for the classification of beam performance as a function of the parameters of the impacting truck and protection beam. The ML model and demand function are complementary to each other and enable rapid and accurate assessments, respectively, of the performance of protection beams during impact scenarios.

This paper presents built-up press brake-formed tub girders (PBTGs) as a new approach for the fabrication of resilient, short-span steel bridges. In contrast to existing PBTGs that are cold bent from a single steel plate, this is a built-up system in which webs are cold bent and bolted to separate, flat bottom and top flange plates. It offers enhanced versatility, as this is a “kit-of-parts” where different sizes of components can be used to achieve required capacities or span lengths (e.g., thicker bottom flanges for negative moment regions, deeper webs for longer spans). Using built-up sections joined by bolts provides internal member redundancy, as crack propagation between components is arrested. The objectives of this paper are to: (1) introduce the system including design details; (2) numerically demonstrate its feasibility through the finite element modeling of two demonstration bridges; and (3) investigate the internal redundancy by evaluating behavior when a flange or web is abruptly lost. Results indicate the viability and redundancy of the system for simple and two-span continuous bridges.

Accurate assessment of temperature characteristics of flat steel box girders (FSBGs) is of great significance for the design, construction, and maintenance of long-span bridges. However, the inflow wind speed has been routinely adopted in the conventional thermal analysis on FSBGs; this ignores the actual wind field around the FSBGs and might result in inaccurate assessment of temperature characteristics of FSBGs. In this study, a novel thermal analysis approach is proposed to evaluate the temperature characteristics of FSBGs, which accounts for the nonuniform wind distribution at the thermal boundary layer thickness (U_a) around the FSBGs. To achieve this, a wind adjustment factor derived via computational fluid dynamic simulations is employed to correlate the inflow wind speed and U_a. The feasibility and accuracy of the proposed thermal analysis approach is validated through field measurements. Subsequently, four indices, i.e., effective temperature (ET), vertical temperature difference (VTD), transverse temperature difference (TTD), and local temperature different gradient (LTDG), are adopted for evaluating the temperature field characteristics of FSBGs of a prototype bridge. Finally, a parametric study is performed to investigate the influence of several key parameters on the temperature field characteristics of FSBGs. These research methods and conclusions can provide valuable references for the thermal design, monitoring, and control of long-span bridges employing FSBGs as the main girders.

Abstract

A diaphragm is an essential component of a T-girder bridge. Evaluating the influence of various truss diaphragms (TDs) on the structural behavior and load distribution factor of T-girder bridges assists in bridge design and strengthening. In this study, a series of experiments and simulations were conducted to investigate the strengthening effect of two types of TDs (i.e., triangle-TDs and K-TDs) based on a small-scale T-girder bridge model. Formulas for the flexural rigidity of the two types of TDs were proposed and verified by using rigid-joint girder methods, experiments, and simulations. Then, taking the K-TD as an example, the calculation method for the stiffness of the truss was analyzed and derived based on rigid-joint girder and graphic multiplication methods. The results showed that K-TDs had a better strengthening effect than triangle-TDs. The deflection and strain of the K-TD-strengthened T-girder bridge were reduced by 21% and 16%, respectively, compared with those of the triangle-TD-strengthened bridge. The formulas for flexural rigidity were proposed and used to calculate the load distribution factor. The maximum error of the calculated load distribution factor was 16% compared with the simulation and experimental results. Moreover, the calculation method for the stiffness of the K-TDs was analyzed and obtained.

Practical Applications

Deterioration of multigirder bridges is getting worse due to the increasing traffic load and insufficient maintenance. Adding diaphragms to the existing bridges can improve the load distribution of multigirder bridges and reduce the maximum load of the girders. This work evaluated the influence of two types of truss diaphragms (triangle truss diaphragm and K truss diaphragm) on the load distribution of a T-girder bridge, and the results showed that the two types of truss diaphragms can reduce the maximum deflection and strain in the T-girder bridge effectively. The flexural rigidity formulas for the two types of truss diaphragm were proposed and proved to be accurate by comparing with experiments and finite-element methods. Moreover, the truss stiffness of the K truss diaphragm was derived by a graphic multiplication method. These formulas can be used by bridge engineers for new bridge designing or existing bridge strengthening in practical engineering.

Many of the procedures used in bridge design are force-based and may be considered reasonable design approaches that will lead to safe structures. But they do not directly address performance criteria at the initial stage of the design. On the other hand, the previously presented displacement-based design procedures of isolated bridges only considered the displacement of the deck, which means the pier’s seismic performance and the substructure mass effects are ignored. This paper presents a new performance-based design method for isolated RC bridges considering the contribution of substructure mass. This method allows the designer that selects the performance level of the superstructure and substructure at the beginning of the design procedure. In this respect, the two degrees of freedom analytical model of the isolated bridge containing substructure mass is presented to quantify its effect on the performance of the seismically isolated bridge. The proposed design method is applied to two continuous-span bridges with regular and irregular substructures, and different target performance levels, including elastic and yielded substructures. The results of the proposed design method have been compared with nonlinear time history analyses, AASHTO, and National Cooperative Highway Research Program simplified methods. The results showed that the seismic performances are close to the results of the nonlinear time history analyses. However, the other design approaches, which ignore the substructure mass, underestimate the responses of the substructure.

The incorporation of corrugated steel webs in composite box-girder bridges has greatly improved their structural performance. However, the low axial stiffness of the shear-deformable corrugated steel webs also affects the structural behavior under the commonly encountered eccentric loading. In this study, experiments were carried out to investigate the behavior of this type of bridge under eccentric loading. Based on the experimental study, the basic assumptions in existing torsion theories are examined. In addition, the proposed formula to estimate the torsion constant and amplification factor considering the additional sectional normal stress due to distortion and warping under eccentric loading in design codes is also checked. From the study, some design recommendations are provided.

Integral components of suspension bridges are anchor spans, consisting of separate strands splayed from the main cable between the splay saddle and the front anchor plane. The main cable is separated into several strands and anchored to the anchorage through the twisting and dispersion of the splay saddle. Tensions in the anchor span strands are related to the overall structural safety of the suspension bridge, the geometric form, and internal stresses in the main cables generated in the completed bridge state. Therefore, the calculation and analysis of anchor span tension are particularly significant. Meanwhile, analyzing tensions in the anchor span strands is quite problematic, insofar as each strand in the saddle groove of the splay saddle is horizontally and vertically bent and then into a spatial catenary, while individual strands form a complex spatial system. This study first determines the spatial orientation of the anchor span strands passing over the splay saddle and their positions relative to the splay saddles. Analytical solutions are found by assuming that the strand tensions are equal at the point of tangency in each strand in the completed bridge state. Therefore, it can be ensured that the tensions in the anchor span strands are equal, which further leads to a sufficiently high safety factor for the cable strands. Given that the strands are separated from each other, an analytical approach is adopted for geometric coordination and mechanical analysis. Tensions in all anchor span strands, tangent position, and angle of each strand in the splay saddle can be determined quickly and precisely. A regular distribution pattern about the calculation results is observed according to the relative position of strands. A comparative validation is performed for strand tensions calculated theoretically against the measured values in a calculation example, proving the feasibility and high accuracy of the proposed method.

There has been a growing interest in using large-diameter strands to alleviate congestion by using fewer strands, reduce the total number of girders by increasing girder spacing, increase the span length, and allow shallower girders. While the 2020 AASHTO LRFD Bridge Design Specifications included 17.8-mm (0.7-in.) strands by reference to AASHTO M203, these larger-diameter strands were not used in bridge construction primarily because AASHTO LRFD was silent about the design aspects of members reinforced with 17.8-mm (0.7-in.) strands. The presented multifaceted, multiyear research involved an extensive parametric design case study, nonlinear finite-element analyses, material characterization, component tests, and full-scale girder experiments. This paper focuses on the full-scale girder tests, which were used to examine development length, detailing requirements, and flexural and shear behavior and strength. Experimentally determined development lengths were found to be shorter than those prescribed by the AASHTO LRFD Specification. Flexural and shear strength could be determined using established procedures. The current minimum required amount of confinement reinforcement was found to be sufficient to confine 17.8-mm (0.7-in.) strands. The extension of bottom flange confinement reinforcement was found to be inadequate for cases with partially debonded 17.8-mm (0.7-in.) strands, but extension of bottom flange confinement reinforcement to 1.5d beyond the end of the girder was adequate for cases with no debonded strands. The minimum bottom flange confinement reinforcement required by the AASHTO LRFD Specification must be extended to at least 1.5d beyond the termination of the longest debonded length of 17.8-mm (0.7-in.) strands.

The rapid and accurate identification of steel strand tension is critical in ensuring the safety of load-bearing structures and structural members, such as bridge cables. Ultrasonic-guided waves (UGWs) have a reasonable propagation distance and reliable detection efficiency, and they are promising for ensuring the safe monitoring of long-distance structures. The L(0,1) waveguide mode of the steel strand has a missing frequency band, and the center frequency of the missing band (notch frequency) positively correlates with axial tension. Based on this feature, a method for detecting the axial tension of steel strands was developed in this study based on notch frequency analysis. First, normalized complex frequency analysis (NCFA) was applied to investigate the trailing and amplitude attenuation of UGW signals generated by the missing frequency band to obtain the characteristic parameters related to the notch frequency under different tensions. Subsequently, a multiple nonlinear regression model was developed to predict the strand tension using the parameter samples. The experimental results showed that the proposed method can overcome the resolution limitation when using only frequency as the parameter. Moreover, the proposed method had an improved parameter and noise robustness. The average error of the prediction results was 1.843 kN within a loading force range of 50–150 kN, demonstrating the potential for practical engineering applications.

Recently, bridge collapse accidents have become increasingly frequent during the flooding season, and foundation scour is one of the main reasons. Tracking scour evolution accurately is a key premise for preventing and controlling hydrological damage. Scour identification according to the changing dynamic characteristics during the scouring process tends to be one of the top technical methodologies in scour monitoring. Although efforts have been made to investigate the dynamic identification of foundation scour, the investigations have mainly focused on the qualification of foundation scour utilizing numerical simulations. Quantitative analysis and validation through laboratory experiments with large-scale water flumes are still lacking. To bridge the gap, this study performed physical modeling experiments with a large-scale water flume for foundation scour to investigate the relationship between structural frequency and scour evolution. The research sought to validate the feasibility of using structural frequency in dynamic identification. First, scour experiments with a large-scale water flume for three piers were performed to collect the time history of acceleration signals. Then, the acceleration signals were processed to recognize the temporal evolution of structural frequency during the scouring process. Finally, the relationship between the temporal scour depth and frequency was assumed to be linear and nonlinear to fit the time history of structural frequency. The results indicated that the frequency and the square of frequency can be taken as the dynamic fingerprint in scour identification according to the frequency range. Based on the validation of large-scale flume experiments, the proposed nonlinear temporal models of frequency in the study demonstrated a good indicator for predicting scour depth. The methodology can greatly enhance the practicality and convenience of bridge scour dynamic identification.

This study offers a probabilistic framework for the seismic life-cycle cost (SLCC) assessment of aging highway bridge structures in complex seismogenic zones, followed by a case-study application on a popular bridge class in Seattle (United States). The proposed framework efficiently handles the implications of differing seismic hazard scenarios (crustal and subduction earthquakes) along with time-dependent bridge deterioration mechanisms on seismic fragility and lifetime economic losses. These losses, stemming from the seismic damage-induced repair of critical bridge components, may be uncertain depending on the repair technique, availability of resources, or decision of the stakeholder. Unlike previous works that have evaluated seismic losses under a single hazard source and deterministic repair costs, this study aimed for a better approximation of the accumulated lifetime losses for aging highway bridges under simultaneous exposure to differing hazard sources, while accounting for repair uncertainties. The case-study results revealed that the SLCCs using deaggregated hazard sources incurred significantly higher contributions from subduction earthquake sources than shallow crustal earthquake sources for both as-built and aging bridge structures. Lastly, based on the proposed framework, a statistical model is proposed for easy adoptability by stakeholders and decision-makers for the SLCC computation of bridge structures in the case-study region.

Abstract

Recently, the construction of composite steel plate girder bridges that uses precast concrete deck planks has significantly increased due to their ease of construction, rapid installment, relatively low cost, and lightweight. Recent structural failures have been reported for this type of bridge system, which include the total collapse of one section of the Mexico City Metro Overpass in May 2021. The bridge collapse killed 23 people, and dozens were injured. This study provides an overview of the causes of the failure of the Mexico City Metro Overpass. It highlights important structural deficiencies that are currently overlooked in design codes and could cause future catastrophic collapses in existing bridges that utilize the same type of bridge. Therefore, a detailed finite-element model of a composite steel girder bridge that uses precast deck planks was established. The bridge dimensions, properties, and loading conditions that were used in this study were obtained from the Mexico City Metro Overpass parameters with some modifications. The performance of the bridge structure was investigated under different conditions with consideration of joint deterioration between the precast concrete planks. The results indicated that the deterioration in the transverse joints due to long-term exposure to moisture, environmental conditions, and vibrations affected the amount of force that was transferred between the concrete deck and the bare steel girder and, therefore, hindered the composite action. The gradual loss of composite action could cause a significant increase in the demands that are imposed on the bare steel section and result in high deflections in the bridge deck. Further deterioration in the transverse joints could be ongoing due to the repetitive vehicle loading on the compromised composite girder. Vibrations and exposure of the cracked joints to moisture and weather could worsen these conditions. Catastrophic failure of the bridge could occur if the deterioration issues were not observed and addressed in a timely manner. This study concludes by providing recommendations for design considerations to improve bridge performance, extend service life, and prevent such failures.

Practical Applications

The collapse of the Mexico City Metro Overpass killed 23 people and injured many others and inspired this study. The system used in this bridge is a composite plate girder with full-depth precast concrete (FDPC) deck, a recent system that allows rapid construction. The best way for engineers to prevent repeating these accidents is to understand the factors that lead bridges to fail. This study starts by exploring the causes of failure in bridges when focusing on the expected deterioration effects in the long run. Simulation techniques are employed to build a model for a bridge that is similar to the system that collapsed in the Mexico City Metro Overpass. It was used to understand the factors that led to the collapse under the moving train load. The results showed that adhering to current design codes and guidelines is not sufficient to prevent the failure of a bridge. This is a serious issue, because the codes are expected to provide the minimum requirements that guarantee the safety of structures during their service life. Therefore, this study concludes by providing recommendations for design considerations to improve bridge performance, extend service life, and prevent such failures.