Journal of Bridge Engineering

In current design specifications, a static procedure using code-prescribed loads has been commonly used to design bridges against ship collisions. However, the collision-induced response may be greatly underestimated as a result of the absence of the dynamic amplification effect in the static procedure. Hence, the safety of bridge structures exposed to ship collisions is still a concern, although they have been designed according to the current specifications. To this end, this paper aims to determine the dynamic ship-impact load on bridge structures and to evaluate its rationality. A simple approach to determining the dynamic ship-impact load was first developed based on the ship-bow force–deformation relationship (i.e., P–a curve) and the analytical model considering the structural stiffness. The equations for estimating the load duration and the time history of ship deformation were analytically derived. To evaluate the accuracy of the developed ship-impact loads and some other impact loads in current studies, the shock spectra corresponding to these dynamic loads were compared with the desired spectra. It was found that the shock spectra using the developed ship-impact loads are generally in good agreement with the expected results, indicating the rationality of the developed approach. The results obtained using the bilinear P–a curve and the sine pulse are often unconservative. In contrast, the trapezoidal pulse generally leads to a conservative spectrum. Finally, case studies of ship–bridge collisions further confirm the aforementioned observation from the shock spectrum comparison. The importance of shock spectrum approximation is clearly identified for the development of the dynamic ship-impact loads on bridge structures.

In this study, a track-bridge model was proposed by combining the direct stiffness method (DSM) and the mode superposition method (MSM) for analysis of the train-track-bridge interaction. The tracks were modeled by applying DSM, which can effectively display the nonlinear contact spring and the dynamic behavior of the tracks. The bridge was modeled by applying MSM, a method that can efficiently show dynamic behavior with a small number of modes. The analysis model was built based on a previously proposed train-track analysis model. Consequently, the train, track, and bridge were modeled as a coupled system that was interconnected by nonlinear Hertzian springs between the train and track, and springs and dampers between the track and bridge. Thus, the convergence is greatly improved as a result of the stabilization of the analytical system. Using the proposed train-bridge-track interaction analysis model, the impacts of the vertical displacements of the bridge resulting from thermal load on the dynamic behavior of the bridge, the train safety, and the passenger comfort were investigated for a steel arch bridge that will be constructed along the Honam high-speed line in Korea.

A novel damage-assessment method based on long-gauge strain sensors is presented in this paper. This method is not only suitable for rapid diagnosis of structures, but also it can be applied to long-term structural health monitoring of structures. First, theoretical derivation showed that the damage index strain time history area ratio (STHAR) of bridges under moving vehicular load was only in relation to the local bending stiffness, and it was independent of the weight, velocity, and axle numbers of the vehicular load. Second, the influence of vehicular weight, vehicular velocity, axle number, sensor gauge length, and measurement error on the method was investigated by simulation. In addition, a case study of a lab experiment was also examined. The results show that all damages can be localized accurately, and the mean relative errors between the estimated and actual damage extent were 0.7% and 12.3% for simulation and experiment, respectively. Finally, the robustness of the method was verified by field testing of a real bridge (in the undamaged state). The tendency of the damage index variation was predicted by finite-element method, and the mean relative error between the estimated and actual damage extent was found to be 12.6%.

Although integral abutment bridges (IABs) can reduce construction and maintenance costs compared with conventional jointed bridges, certain aspects of their structural behavior are still not well understood. Most prior IAB research was related to substructure behavior, and as a result, most limit states that have been considered in design guidelines have been based on substructure considerations. However, integral abutment construction also affects superstructure behavior and demands, and superstructure properties directly influence substructure behavior. This paper presents numerical simulations evaluating the behavior of IABs with composite steel I-girders subjected to temperature changes consistent with seasonal fluctuations in the state of Illinois. Bridge superstructures, abutments, piers, and pile foundations were modeled to determine various structural demands imposed by these temperature changes. A suite of nonlinear bridge models is introduced in which key bridge parameters were varied, such as overall bridge length, intermediate-span length, pile size, and skew. Results indicate that effective expansion length (EEL) has a primary influence on bridge longitudinal movement under thermal loads regardless of girder, abutment, or pile design. Also, results show that superstructure girder response (elastic) and substructure pile response (inelastic) to superstructure temperature change are influenced by parameters such as EEL, pile size, skew, and the rotational restraint that the superstructure imposes on the substructure. Results presented herein suggest that superstructure geometry should be directly or indirectly considered in IAB substructure design, and that thermally induced stresses and strains should be accounted for in superstructure and substructure design.

Despite the major advances in finite-element (FE) modeling and system identification (SI) of extended infrastructures, soil compliance and damping at the soil–foundation interface are not often accurately accounted for due to the associated computational demand and the inherent uncertainty in defining the dynamic stiffness. This paper aims to scrutinize the effect of soil conditions in the SI process and to investigate the efficiency of advanced FE modeling in representing the superstructure–soil–foundation stiffness. For this purpose, measured, computed, and experimentally identified natural frequencies of a real bridge were used. Field measurements obtained during construction were reproduced both in the laboratory and by refined FE modeling. In addition, to understand the physical problem more thoroughly, three alternative soil conditions were examined: rock, stabilized soil, and Hostun sand. Discrepancies on the order of 3–13% were observed between the identified and the numerically predicted natural frequencies. These discrepancies highlight the importance of reliable estimation of soil properties and compliance with the SI framework for extended bridges under ambient and low-amplitude vibrations.

This study addresses the feasibility of using a single-span bridge as a weigh-in-motion (WIM) tool to quantify the gross vehicle weights (GVWs) of trucks inexpensively with a small number of sensors and without using axle detectors. Field testing was performed on an interstate without any lane closures. Four preweighed trucks with different axle configurations traveled over a bridge at three different speeds and in two separate lanes. Measured strain data were used to implement bridge weigh-in-motion (B-WIM) algorithms and calculate the corresponding velocities and GVWs. A comparison was made between calculated and actual measured static weights, and between the calculated and specified speeds of the trucks. In addition to field testing, a finite-element (FE) model of the tested bridge was created and calibrated based on the measured strains at different locations. This calibrated FE model enabled the acquisition of the influence values for the bridge at any location (influence surface). Ten different points were selected to calculate the influence surface values, and a comparison was made between calculated strains from the influence surface and the actual response of the bridge recorded by strain gauges. The comparison showed that the updated FE model was capable of providing the influence surface values at different locations. The validated influence surfaces were then used to simulate the passing of common types of trucks with various weights and axle configurations over the bridge. The measured GVWs based on simulated strain measurements provided by the FE model and validated influence surfaces verified the applied method.

The widely used orthotropic steel deck system is subject to several durability problems: fatigue cracking of the steel, wearing of the surfacing, and corrosion. Thus, the composite deck system composed of an orthotropic steel deck and an ultrathin reactive-powder concrete (RPC) layer was proposed. In a prefabricated composite deck system, joints formed between the prefabricated RPC layer and the cast-in-situ RPC layer are defined as wet joints. The experimental study was conducted on the tensile behavior of RPC wet joints to ensure the integrity of the prefabricated deck system and the serviceability and durability of the RPC layer. Two series of tests were conducted, including seven tensile tests on panels and four negative bending tests on beams. In these tests, five types of new wet joint details were designed in addition to the integral casting detail and the conventional wet joint with a vertical, plane interface. To study the mechanical properties and crack resistance performance of different wet joint details, the load-deformation curves and the nominal RPC tensile stress–maximum crack width curves of the models in each test series were compared. Comparisons showed that the mechanical properties of different wet joint details were close to one another. However, their crack resistance performance differed considerably. The sawtooth wet joint, rectangular wet joint, and steel plate–enhanced wet joint had substantially better crack resistance performance than the conventional wet joint. In contrast, the inclined wet joint and the reinforcement-enhanced wet joint had poor crack resistance performance because of their plane interfaces. Additionally, test results suggested that the postcracking behavior of the RPC layer was improved by the steel fibers. Therefore, for more economical bridge designs, the durability-based allowable RPC tensile stress was recommended instead of the initial cracking stress.

Bridge scour is the number one cause of failure in bridges located over waterways. Scour leads to rapid losses in foundation stiffness and can cause sudden collapse. Previous research on bridge health monitoring has used changes in natural frequency to identify damage in bridge beams. The possibility of using a similar approach to identifying scour is investigated in this paper. To assess if this approach is feasible, it is necessary to establish how scour affects the natural frequency of a bridge, and if it is possible to measure changes in frequency using the bridge dynamic response to a passing vehicle. To address these questions, a novel vehicle–bridge–soil interaction (VBSI) model was developed. By carrying out a modal study in this model, it is shown that for a wide range of possible soil states, there is a clear reduction in the natural frequency of the first mode of the bridge with scour. Moreover, it is shown that the response signals on the bridge from vehicular loading are sufficient to allow these changes in frequency to be detected.

The development of corrosion as a result of tendon exposure to moisture or corrosive media often leads to wire fractures. Such fractures in unbonded tendons are expected to induce global strand and anchor response through the progression of prestress loss from the break to the end anchors. Radial pressure, interwire friction, and lateral confinement, however, affect the magnitude of prestress loss carried to the anchors and have important implications on breakage detectability by tendon monitoring methods that rely on anchor response. This paper presents an experimental investigation of tendon behavior after successive wire cuts in confined and unconfined conditions and demonstrates the effects of confinement on stress recovery. In addition to analysis of strand prestress loss and axial strain along individual wires, the strand’s dynamic postbreakage response was examined. Moreover, the effects of deviators on stress recovery were tested with a multistrand tendon specimen. A finite-element model of a seven-wire strand was developed and experimentally validated. Both experimental and numerical results showed significant prestress loss even at distant locations from wire breaks and confirm a correlation between wire breakage and modal properties, which can be a useful damage indicator.

Piles have been used to transfer superstructure (bridges, buildings, etc.) loads to deeper competent bearing strata to minimize settlements. Accurate predictions of design settlements of vertically loaded piles under working load are a critical issue in the design of pile foundations. Numerous comprehensive trials, built on a regression-based predictive formula that produces the axial settlement using elastic shortening as the input variable, were performed in this study, and a simple relationship was established between the elastic shortening of piles and the measured design settlements of 93 full-scale pile load tests installed in variable bearing materials. The calculated axial settlements of piles from the proposed method in this study were compared to the measured axial settlements of the piles. Accuracy and validity of the proposed method were verified through conventional statistical analysis in terms of mean of bias (λ_m) and coefficient of variation (μ). The proposed method yielded values of λ_m = 1.06 and μ = 0.15.

It is a very common practice to use longitudinal expansion joints to create dual-direction bridges in applications that would have otherwise resulted in bridges that are very wide. Unfortunately, the introduced longitudinal expansion joint can result in deterioration similar to that resulting from using transverse joints at bridge abutments and piers. The reason that longitudinal expansion joints are used in wide bridges is related to concerns about deck cracking resulting from constrained expansion and contraction. The main objective of this research was to evaluate the need for longitudinal expansion joints by assessing the relationship between bridge width (and other factors) and the propensity for deck cracking. To achieve this objective, an integral abutment bridge was selected to conduct live-load and thermal-load testing to provide field testing data for the calibration of a three-dimensional (3D) finite-element model (FEM) that would be used for a comprehensive parametric study. A combination of live load plus temperature and shrinkage loading was applied to validate the model. During the parametric study, three different width integral abutment bridge models were built to evaluate the need for the longitudinal expansion joint. Additionally, models with 45° skew and without an integral abutment were developed to study the relationship between bridge skew, abutment type, and the development of longitudinal and diagonal cracks. A review of historical deck-crack maps from actual bridge inspection reports was conducted for 40 common bridges. The results of this review of historical deck-crack development compared favorably with the parametric study results. The results of this study indicate that the use of a longitudinal deck joint will not reduce deck-end cracks. It was further found that the commonly observed longitudinal and diagonal cracking in the deck near the abutment on integral abutment bridges is a result of the restraint induced by the integral abutment combined with temperature and shrinkage differences between the abutment and the deck, and will occur for bridges of any width and any skew. It was also found that bridges that use conventional abutments (sometimes referred to as stub abutments) are not prone to develop this type of cracking, even for very wide bridges.

The antisliding safety of the middle saddle is a key issue for the design of three-pylon suspension bridges. Vehicle load is the main adverse effect, especially unsymmetrical vehicle load. A general mechanical model on the saddle sliding behavior is reviewed and proposed, and then characteristics of load responses of the main cable are investigated to explore a simplified live load computational method. Based on the deterministic approach from codes and a reliability-based approach, the antisliding safety of the middle saddle is discussed and assessed for the Taizhou Yangtze River Bridge, which has two symmetrical spans of 1,080 m. The study shows that geometric nonlinearity caused by dead load has a big effect on the load responses of the main cable, so it must be included in the design. Further, the maximum difference between the geometrically linear and nonlinear cases due to live load is up to 12.71%, whereas in all conditions, the balanced static friction coefficient is smaller without consideration of the geometric nonlinearity caused by live load. This means that the assessment of the antisliding safety of saddles is more conservative when live load geometric nonlinearity is not included, and the linear superposition principle between traffic load and influence surface can be directly applied. The results based on the deterministic approach from codes indicate that the antisliding safety factor of the middle saddle is larger than 1, which ensures that there is no sliding between the middle saddle and the main cable. However, the results from different codes differ because of the different requirements for the parameters. Therefore, based on a reliability approach, a probabilistic model of the saddle with balanced static friction coefficient is developed, which includes the effect of traffic by using weigh-in-motion (WIM) data and random synthetic vehicle flow simulation. In the paper, the maximum static friction coefficient model is obtained through the experimental data of cables, and the antisliding safety assessment of the saddle is made for different traffic conditions. All antisliding reliability indices are larger than 4.9, indicating that the saddle has good antisliding behavior in reserve.

The torsional behavior of hollow-core steel-concrete-steel (HC-SCS) columns is presented using finite-element (FE) and analytical approaches. The HC-SCS columns consist of a concrete shell sandwiched between two steel tubes. Software was used to develop a three-dimensional model of an HC-SCS column that was subjected to torsional loading. The FE results were validated against the experimental results collected from six HC-SCS columns tested under pure torsion. The average error from the FE analysis was 4.8%, compared with experimental results, when predicting the column’s torsion strength. The study revealed that the interaction between the steel tube’s stiffness and concrete shell’s thickness controls the behavior of the column. A parametric study was conducted to further analyze each parameter affecting the column’s torsion behavior. The parametric analysis concluded that the torsional behavior of the column mainly depends on the outer steel tube’s properties and the thickness of the concrete shell. A simplified equation was developed to predict the torsion strength of the member using a direct method of stress analysis. The proposed equation predicted the members’ torsion strength with an accuracy greater than 90%.

A new bridge system has been developed to (1) reduce on-site construction time by using precast components, (2) eliminate major earthquake damage by utilizing column rocking and confinement of the column ends with a steel tube, and (3) maintain the system functionality after a strong earthquake by minimizing residual drift through the use of pretensioned strands in the columns. Furthermore, it uses only conventional materials. This paper compares the shaking table performance of a quarter-scale, two-span bridge constructed through the use of the new system with that of a conventional cast-in-place bridge with similar geometry tested in 2005. The new bridge system was constructed in approximately 20% of the time needed for the conventional cast-in-place system. In tests, the conventional bridge suffered major concrete cracking and spalling, whereas in the new system, damage to the concrete was only cosmetic. In the conventional bridge, the longitudinal bars buckled, and both the longitudinal and spiral reinforcement fractured, whereas in the new system, the damage to the reinforcement was limited to longitudinal bar fracture, and that occurred only under excitations much larger than the design-level motion. Furthermore, bar fracture in the new system could be delayed by increasing the unbonded length of the bars. The residual drift of the new system was essentially zero for all motions, whereas one of the exterior bents of the conventional bridge was so badly damaged and out of plumb that some of the supplemental mass on the bridge had to be removed, and testing was stopped shortly thereafter.

As a result of heavy prestressing, anchorage zones of bridge girders frequently crack in consistent patterns at prestress transfer and cause durability concerns. These cracks, which appear during prestress release, continue to grow in width and length in the months following detensioning. This study investigated reasons behind crack growth after detensioning. Load effects created by differential cooling, creep, and shrinkage of bulb-tee cross sections were studied as potential reasons for crack growth. Fully coupled thermal-mechanical analyses included nonlinear concrete material properties. The analyses were verified and validated by analytical calculations and test data. A parametric study was conducted to understand the sensitivity of results to concrete thermal properties and ambient temperatures. Temperature changes resulting from curing and creep had a favorable and negligible effect on concrete strains, respectively. Shrinkage strains, although not sufficient to initiate cracking without prestressing, were determined to be a cause for increase in crack width and length after detensioning. Reinforcing bar stresses increased up to 20% in low ambient temperatures and for low-thermal-expansion concrete under curing temperatures and exceed the AASHTO limits.

Premature failure of expansion joints is a concern for bridge owners. After opening to traffic, the Jiangyin suspension bridge experienced frequent damage in expansion joints, and the problem remained unsolved even after replacement with new expansion joints and installation of viscous dampers at the girder end. This paper aims to clarify the failure mechanism of the expansion joints based on field test results. The field tests consist of three scenarios: static test, influence line test, and truck pass-by test. The tests aimed to investigate the behavior of expansion joints on both global and local scales. In the static and the influence line tests, it was found that strain response of the expansion joint is affected by the girder deformation, which reflects effect of global behavior of the bridge on the expansion joints. The truck pass-by test result shows that direct impact of heavy truck axles could induce high strain in expansion joints, which indicates that the expansion joints are affected by vehicle axle loading on a local scale. These results are in accordance with the premature failures observed in the expansion joints of the Jiangyin Bridge.