Abstract
Staggered widening refers to a rehabilitation scheme in which new bridge piers are not aligned with existing ones due to constraints from current river embankments or road boundaries. To date, research on staggered widening bridges remains limited. Owing to the staggered span layout, girders are supported by piers of adjacent bridges at the mid-span region, leading to a vehicular load distribution mechanism distinct from that of conventional widening bridges, especially for interior girders near the closure slab. This study investigates the load distribution factor (LDF) for bending moment in staggered widening bridges. First, a simplified analytical model is developed using the rigid-jointed girder (RJG) method to derive transverse influence lines. The boundary conditions in the conventional RJG method are modified to capture the mid-span supporting effects provided by adjacent piers. The governing equation of the analytical model is derived based on the updated model configuration. The accuracy of the simplified model is then verified against finite element analysis results. Transverse load distribution characteristics of staggered widening bridges are compared with those of conventional widening schemes. On this basis, a parametric study is performed to examine the effects of key parameters, including diaphragm rigidity, flange rigidity, girder rotational stiffness, closure slab width, girder width, and the number of girders, on the LDF. Finally, based on the parametric analysis results, a generalized formula is proposed for predicting the girder LDF of staggered widening bridges.
Keywords: Staggered widening bridge, Load distribution factor, Transverse vehicular load distribution, Rigid-jointed girder method, Simplified analysis model, Supporting effects, General formula
Subject terms: Engineering, Materials science
Introduction
Bridge widening is a commonly adopted strategy to enhance the traffic capacity of congested bridges1–7. Due to constraints imposed by existing reiver embankments or road boundaries, the piers of adjacent bridges may not be aligned, resulting in the so-called staggered widening8. Since the demolition of existing bridges leads to increased project costs and carbon emissions, the implementation of staggered widening offers significant economic and environmental benefits.
Figure 1 presents a typical span arrangement of staggered widening bridges. Owing to the restriction of existing road boundaries, the piers of the new bridge have to be offset far from those of the existing bridge. In this case, the mid-span segment of the new bridge is spliced with the girder end segment of the existing bridge. This unique planar layout gives rise to two main engineering problems that are not encountered in conventional widening bridges. First, the mid-span segment of the new bridge deflects under vehicular loads and foundation settlements, whereas the girder end segment of the existing bridge dose not deflect due to the vertical restraint of piers. Under such circumstances, the deck of the existing bridge is subjected to an additional transverse hogging moment, which adversely affects the transverse loading condition of the bridge deck. Second, the mid-span segment of the new bridge is supported by the pier of the existing bridge. Consequently, the vehicular loads borne by the girders can deviate significantly from those of conventional widening bridges, especially for the interior girder adjacent to the existing bridge.
Fig. 1.
Illustration of a typical staggered widening bridge.
Recently, the first engineering problem related to the transverse bearing capacity of the bridge deck has been investigated by Yang and Zhang8. A novel segment model was developed to reproduce the transverse loading condition of the bridge deck in staggered widening bridges. On this basis, the transverse bearing capacity of the existing bridge deck was evaluated through segment model tests. The test results indicated that the fundamental design principle for staggered widening bridges is to minimize the differential deflections between adjacent bridges. Ultimately, several recommendations were proposed for the design of staggered widening bridges.
Despite the recent study by Yang and Zhang8, concerns remain regarding the influence of the supporting effects provided by the existing bridge’s piers on the vehicular loads borne by the new bridge. Staggered widening is usually applied to low-to-medium span bridges, in which the live-to-dead load ratio is substantially high. Thus, the vehicular load distribution in each girder is directly related to the prestress and structural dimensions of each girder, both of which are closely associated with the structural safety and economic efficiency. However, the existing methods for calculating vehicular load distribution in conventional bridges9–24 may be inapplicable to staggered widening bridges due to their unique planar layout. Therefore, the vehicular load distribution in staggered widening bridges requires in-depth investigation, which constitutes the topic of this study.
To quantify the vehicular load distribution among multiple girders, the concept of the load distribution factor (LDF) has been introduced in previous studies9–24. The LDF is defined as a dimensionless factor representing the proportion of vehicular load distributed to an individual girder in a multi-girder bridge system, reflecting the transverse load-sharing behavior of the deck-girder system. To determine the LDF for each girder, a transverse influence line is first established to quantify the fraction of live load borne by each girder when a unit live load is applied at different transverse locations across the bridge cross-section. The design traffic load is then positioned transversely to induce the extreme load effects, which yields the LDF for each girder.
The existing methods for calculating the load distribution factor (LDF) for multi-girder bridges can be roughly classified into two categories: numerical models and simplified analytical approaches. Numerical models, such as the grillage model22, finite strip method25, and finite element model20–23, can offer relatively more accurate simulation results of the LDF at the cost of increased model complexity and computational time. Thus, simplified analytical methods are more practical for trial-and-error design and parametric studies of multi-girder bridges. According to a comprehensive review of existing simplified analytical methods16,24, the rigid-jointed girder (RJG) method26 features clear physical significance, fewer parameters, and moderate computational cost. Furthermore, the governing equations of the RJG method can be expressed in dimensionless form to characterize the effects of global and local stiffness of multi-girder bridges on transverse live-load distribution. Therefore, the RJG method is adopted in this study to calculate the load distribution factor (LDF) for staggered widening bridges.
In this study, the load distribution factor (LDF) for bending moment in staggered widening bridges is investigated. Firstly, the boundary conditions of the conventional rigid-jointed girder (RJG) method are modified to consider the mid-span supporting effects provided by piers of the existing bridge. The governing equation of the analytical model is derived based on the updated model configuration. Based on the modified model, the governing equations are derived for calculating the transverse influence lines of the new bridge is staggered widening configurations. The proposed simplified analytical model is then verified through comparisons with finite element analysis (FEA) results. The transverse vehicular load distribution characteristics of staggered widening bridges are also compared with those of conventional widening bridges. Subsequently, a parametric study is carried out using the simplified analytical model to examine the effects of key parameters, including diaphragm rigidity, flange rigidity, girder rotational stiffness, girder width, number of girders, and closure slab width, on the girder LDF. On this basis, a generalized formula is proposed to predict the LDF of staggered widening bridges. The accuracy of the derived formula is further validated against the LDFs obtained from the simplified analytical model.
Project overview
The span arrangement and cross-sectional dimensions of a typical staggered widening bridge are depicted in Fig. 2. The span combinations consist of
m for existing bridges and
m for new bridges. These bridges were constructed as simply-supported prestressed concrete T-girders with link slabs27 to improve both durability and driving comfort. To preclude the transverse transfer of dead loads, the closure slab was cast after the new bridges and the pavement were completed. As reported in the study by Yang and Zhang8, the transverse connection between adjacent bridges was designed as a flexible joint without a diaphragm. The design objectives are outlined as follows: (1) A flexible joint can reduce load transmission from adjacent bridges, thereby improving the transverse loading conditions of the deck. (2) A flexible joint can also enhance the ductility of the deck and facilitate a ductile failure mode, ensuring structural safety3. Additionally, as depicted in Fig. 3, if a rigid connection were constructed, the mid-span section of the girders would experience significant shear force due to the vertical support from adjacent bridges under vehicular loads and foundation settlement.
Fig. 2.
Span arrangement and cross-sectional dimensions of a prototype bridge.
Fig. 3.
Large shear force in the mid-span section of the interior girder resulting from the rigid transverse connection between adjacent bridges.
For the new bridges, the types of transverse connection systems can significantly influence the transverse vehicular load distribution among the girders28–32. Therefore, in the prototype bridge, various transverse connection systems were considered, including a flexible connection without a diaphragm, a semi-rigid connection with a K-shaped diaphragm and a rigid connection with a solid diaphragm. Moreover, due to the substantial changes in the transverse boundary conditions of staggered widening bridges, significant discrepancies exist among the vehicular loads borne by each individual girder. To facilitate description, the girders of the new bridge in the prototype were labeled as exterior-, intermediate- and interior-girders according to their transverse positions, as illustrated in Fig. 2(b).
Simplified analysis model for calculating transverse influence line
Model motivation
The load distribution factor (LDF) is used to characterize the proportion of vehicular load borne by each girder33–41. To achieve this, a transverse influence line should be obtained for each girder firstly42,43. Comprehensive reviews of the relevant methods for calculating the transverse influence line can be found in the studies by Kong et al.16 and Chen et al.24. In this study, the rigid-jointed girder (RJG) method16,24,42,44,45 was employed because of its clear physical significance and relatively low computational cost. In the RJG method, the vertical deflections and bending moments of a simply-supported girder are zero at both ends. From a mathematical perspective, the deflection curve of the girder can be approximated using a Fourier sine series. For engineering applications, the half-wave sinusoidal profile has been proven sufficiently accurate (with an error of approximately 1%) in representing the actual load and deflection distributions16. Therefore, as illustrated in Fig. 4(a), the RJG method assumes that vehicular loads (including uniform loads and multiple concentrated forces) can be approximated by a half-wave sinusoidal load to simplify the subsequent derivation. Given the direct correlation between girder deflections and internal forces, which is established through girder stiffness, the transverse distribution of vehicular loads can be determined by the ratios of girder deflections. Meanwhile, since the girders deflect in the same sinusoidal pattern but with varying magnitudes along their longitudinal axis, the LDFs of each cross-section are considered identical. Consequently, the mid-span section is commonly chosen to study the LDFs of the girders16.
Fig. 4.
Model motivation based on the rigid-jointed girder method for staggered widening bridges.
Based on the above-mentioned basic assumptions of the RJG method, the mid-span section of the new bridge with a unit width was selected to calculate the vehicular load distribution in staggered widening bridges. Moreover, a fixed-end boundary condition was imposed near the closure slab to account for the supporting effects of the adjacent piers of existing bridges. As a result, a simplified analysis model, as shown in Fig. 4(b), was developed to determine the transverse influence line for staggered widening bridges.
Governing equations of the simplified analysis model
The force equilibrium equations for the simplified analysis model depicted in Fig. 4(b) were derived herein. To this end, the free-body diagram of the simplified analysis model is illustrated in Fig. 5(a). The RJG method assumes that connections between adjacent girders can transfer shear forces and bending moments. Considering the global and local deformations of the girders under redundant forces and external loads, as shown in Figs. 5(b)-(h), the force equilibrium equations were derived as follows:
![]() |
1 |
here
is the flexural matrix,
is the redundant force vector and
is the external displacement vector.
Fig. 5.
Mechanical analysis diagram of the simplified analysis model for vehicular load distribution of staggered widening bridges.
Specifically, the flexural stiffness matrix
is expressed as:
![]() |
2 |
where,
is the mid-span deflection of the girder under unit shear force applied at the end of the flange;
and
are the rotational angle under unit shear force and unit bending moment applied at the end of the flange, respectively;
is the flange deflection due to unit shear force;
and
are the flange rotational angles due to unit shear force and unit bending moment, respectively;
and
are the rotational angles of the closure slab due to unit shear force and unit bending moment, respectively;
is the closure slab deflection due to unit shear force.
The above-mentioned deformations are calculated as:
![]() |
3a |
![]() |
3b |
![]() |
3c |
![]() |
3d |
![]() |
3e |
![]() |
3f |
![]() |
3g |
![]() |
3h |
![]() |
3i |
where,
is girder span length;
is girder width;
is closure slab width;
is deck thickness;
is elastic modulus of concrete;
is shear modulus of concrete;
and
are the moment of inertia and polar moment of inertia of the girder section.
The redundant force vector
is expressed as:
![]() |
4 |
where,
,
and
are redundant shear force;
,
and
are redundant bending moment. The positive directions of the redundant forces
are illustrated in Fig. 5(a).
The external load vector
is calculated according to the position of unit external load
. If the unit load is applied to the exterior girder,
is calculated as:
![]() |
5 |
where,
and
are the girder deflection and rotational angle due to the external load, respectively;
and
are the flange deflection and rotational angle due to the external load, respectively;
denotes the external load position.
If the unit load is applied to the intermediate girder,
is calculated as:
![]() |
6 |
If the unit load is applied to the interior girder,
is calculated as:
![]() |
7 |
If the unit load is applied to the closure slab,
is calculated as:
![]() |
8 |
where,
and
are the closure slab deflection and rotational angle due to external load, respectively.
The abovementioned deformations due to external load,
,
,
,
,
and
are calculated as:
![]() |
9a |
![]() |
9b |
![]() |
9c |
![]() |
9d |
![]() |
9e |
![]() |
9f |
Insights into the governing equations
As previously elaborated, Eq. (1) presents the governing equations for calculating the LDF in staggered widening bridges. To incorporate the influence of diaphragm rigidity, a more general form of the governing equations is derived. For this purpose, the deck thickness
is distinguished as
and
for regions with and without in-span diaphragms, respectively, as illustrated in Fig. 6(a). The thickness
denotes the effective thickness that accounts for the effects of diaphragm rigidity. In particular,
for bridges without in-span diaphragms, and
is taken as infinitely large for bridges with rigid in-span diaphragms. By dividing both sides of Eq. (1) by
, the following general form of the flexural matrix,
is obtained:
![]() |
10 |
Fig. 6.
Illustration of
and
for cases with different in-span diaphragm rigidities.
where,
,
and
are dimensionless parameters that expressed as:
![]() |
11a |
![]() |
11b |
![]() |
11c |
Note that the parameter
denotes the ratio between the transverse connection (including the deck and in-span diaphragms) deflection and girder deflection;
denotes the ratio between the flange deflection and girder deflection;
denotes the ratio between the torsional angle and deflection of the cross section. Thus, taking girder deflection
as a benchmark, parameter
reflects the influence of diaphragm rigidity,
reflects the influence of flange rigidity,
reflects the influence of torsional stiffness of the cross section.
Moreover, it can be concluded from Eq (11a) that
increases with decreasing in-span diaphragm rigidity. As illustrated in Fig. 6(b), three typical cases are identified: (1) For bridges without in-span diaphragms,
and
. (2) For bridges with rigid in-span diaphragms,
tends to infinity and
. In this case, the bridge cross-section experiences deflection and rotation as a rigid body, and the transverse deformation remains planar. This assumption is valid when the in-plane diaphragm rigidity is sufficiently high and the spacing between in-span diaphragms is sufficiently small. Under these conditions, the RJG method degenerates into the rigid transverse beam (RTB) method16,42. (3) for bridges with semi-rigid in-span diaphragms (e.g., K-shaped diaphragms),
.
Verification and comparison study of the calculated results
Verification by finite element analysis
To validate the simulated results from Eq. 1, three-dimensional finite element models of the prototype bridge depicted in Fig. 2 were built into the software ABAQUS 2021. To investigate the influence of diaphragm rigidity, different types of diaphragms were incorporated into the finite element model, as shown in Fig. 7. In-span diaphragms were placed at the
,
and
cross sections. The bridge was discretized using C3D8R solid elements. The elastic modulus and shear modulus of concrete were set as
and
, respectively. To simulate the vehicular load effects, a sinusoidal half-wave load with unit amplitude was applied to the new bridges, as illustrated in Fig. 8. The sinusoidal half-wave load,
, was expressed as:
![]() |
12 |
Fig. 7.
Three-dimensional finite element model for staggered widening bridges with different transverse connection system (from left to right: without in-span diaphragm, K-shaped diaphragm and solid diaphragm).
Fig. 8.
Sinusoidal half-wave load applied to the new bridge.
where
is the longitudinal coordinate along the bridge span.
By moving the sinusoidal load transversely across the deck, the ordinates of the transverse influence line can be calculated based on the mid-span deflection as:
![]() |
13 |
where,
is the ordinate of the transverse influence line for the ith girder,
is the mid-span deflection of the ith girder calculated using finite element analysis.
The dimensionless parameter
was determined for different diaphragm rigidity cases as follows:
(1) For the new bridge without in-span diaphragms,
was calculated as
according to Eq (11a).(2) For the case with solid diaphragms, a rigid in-span diaphragm was assumed, and
was set to
;(3) For the case with K-shaped diaphragms (i.e., the semi-rigid diaphragm case),
.
In engineering practice,
of the case of K-shaped diaphragms can be determined using the modified RJG method proposed by Chen et al.24, which was developed for LDF calculation in T-girder bridges with truss diaphragms. Further discussion of this modified RJG method is beyond the scope of the present study.
In this work,
was calibrated to
based on parameter sensitivity analysis to ensure that the results from the RJG method agree well with those from the finite element model. As noted in Section “Selected parameters for comparison”,
is adopted as the demarcation point between the semi-rigid diaphragm cases and the cases without in-span diaphragms.
The calculated results using the simplified analysis model are compared with the finite element analysis (FEA) results in Fig. 9. From Fig. 9, it is observed that the proposed simplified analysis model simulates the transverse vehicular load distribution in staggered widening bridges with reasonable accuracy. Moreover, the numerical analysis effectively reflects the influence of different types of in-span diaphragm.
Fig. 9.
Comparison between the transverse influence lines calculated by the simplified analysis model and finite element analysis.
Comparison with conventional widening bridges
Based on the validated analytical model shown in Fig. 5, the transverse vehicular load distribution in staggered widening bridges was compared with that in conventional widening bridges, as presented in Figs. 10 and 11. The span arrangement and cross-sectional dimensions of the staggered widening bridges are illustrated in Fig. 2. For the conventional widening bridge, the new bridge adopts the same span arrangement as the existing bridge shown in Fig. 2(a), and its cross-section is consistent with that of the staggered widening bridges depicted in Fig. 2(b). Cases with flexible and rigid transverse connections were selected for comparison. As observed from Figs. 10 and 11: (1) the staggered pier arrangement significantly affects the transverse vehicular load distribution among the girders, particularly for the interior girder; (2) due to the supporting effects provided by the existing bridge’s piers (shown in Fig. 1(b)), the vehicular load borne by the interior girder in staggered widening bridges is significantly lower than that in conventional widening bridges. In this case, designing the girder using commonly adopted LDF equations9–24 would result in over-conservative prestress dosage and structural dimensions. Further parametric analysis on the LDF of staggered widening bridges was conducted in the section “Parametric analysis of the LDF for staggered widening bridges”, based on which a formula for predicting the LDF of staggered widening bridges was proposed.
Fig. 10.
Comparison between the transverse vehicular load distribution in staggered widening bridges and conventional widening bridges without in-span diaphragm.
Fig. 11.
Comparison between the transverse vehicular load distribution in staggered widening bridges and conventional widening bridges with rigid (solid) in-span diaphragms.
Parametric analysis of the LDF for staggered widening bridges
Because staggered widening bridges exhibited significant differences in the transverse vehicular load distribution compared to conventional widening bridges, a further parametric analysis of the LDF was conducted here for staggered widening bridges. On the basis, a formula of LDF was proposed for staggered widening bridges.
Calculated LDF via the transverse influence line
Based on the transverse influence lines calculated by the simplified analysis model, the LDF was calculated following the proposed method by Kong et al.16. This method consists of rough- and refined-calculations for the LDF. First, the vehicular loads were moved as a whole along the transverse direction of the deck to obtain the maximum load effects for each girder. Then, the lanes and vehicular loads were placed according to the specification requirements to obtain a refined calculation result. The vehicular load utilized in this work is the design truck load suggested by the AASHTO LRFD standard9. The transverse spacing of the wheels is defined as 1829 mm (6 ft.). The traffic lanes width 3658 mm (12 ft.) and the truck loads were assumed to occupy 3048 mm (10 ft.) within a design lane. The wheel loads should be placed not closer than 305 mm (1 ft.) from the face of the curb and 610 mm (2 ft.) from the edge of the design lane. The number of lanes,
is determined by the integer part of the ratio of deck width in millimeters to 3657. The multiple presence factor,
that incorporate the probability of simultaneous lane occupation is defined in Table. 1 according to the AASHTO LRFD standard9.
Table 1.
Multiple presences factor according to the AASHTO LRFD standard.
| Parameter | Number of lanes | Multiple presence factor |
|---|---|---|
| Notation | ![]() |
![]() |
| Value | ![]() |
![]() |
![]() |
![]() |
|
![]() |
![]() |
|
![]() |
![]() |
Selected parameters for comparison
As elaborated in the section “Insights into the governing equations”, the general form of the flexural matrix,
consists of five independent parameters, in which
is related to the diaphragm rigidity,
is related to the flange rigidity,
is related to the girder rotational stiffness,
and
are widths of the girder and closure slab, respectively. Thus, the parameters
,
,
,
and
are considered in the parametric analysis. Besides, the number of girders,
is directly related to the number of lanes,
. Thus, the girder number,
is also considered.
To determine the reasonable regions for the selected parameters, a series of prestressed concrete T-girders that designed by CCCC Second Highway Consultants Co. Ltd. were collected in this work. The span length of the T-girders ranges from
to
m. The determined reasonable regions for the T-girders are shown in Table. 2. Note that for the parameters
and
, only the upper limits were determined according to the bridge collection. The lower limit of
is set as zero to account for the rigid diaphragm cases. Because the deck of the prototype bridge that shown in Fig. 2 has been designed relatively flexible with length of
and thickness of
between adjacent girders, the calibrated
value of
for the K-shaped diaphragm case was adopted as the upper limit of
for the semi-rigid diaphragm cases.
Table 2.
Reasonable ranges for the critical parameters.
| Parameter | Diaphragm cases | ||
|---|---|---|---|
| Without diaphragm | Semi-rigid diaphragm | Rigid diaphragm | |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
||
![]() |
![]() |
||
![]() |
![]() |
||
![]() |
![]() |
||
![]() |
![]() |
||
The units of
and
are millimeter.
Parametric analysis results
Influence of girder position, diaphragm rigidity
and girder number
The calculated LDFs with different
and
for the exterior-, intermediate- and interior-girders,
,
and
, are shown in Fig. 12. The LDF for the intermediate girder, denoted as
, is defined as the maximum LDF among the remaining girders excluding both exterior- and interior-girders. Note that the parameters
and
are set as
and
, respectively. The girder width
and closure slab width
are set as
and
, respectively. For the new bridge with two girders (
), there are only exterior and interior girders. From Fig. 12, it is observed that: (1) the girder position significantly affects the load distribution factor (LDF) of each girder, and the LDF decreases from the exterior girders to the interior girders. (2) for the LDF of exterior and interior girders, the cases with two girders are significantly different from the other cases; For the LDF of intermediate girders, the case with three girders stands in marked contrast to the remaining cases. These disparities can be attributed to the alterations in the structural system of the bridge. (3) the LDF of the interior girder is much less sensitive to the variation in girder number
than that of the exterior girder. (4) the LDF of interior girder increased as
increased (diaphragm rigidity decreased). (5) the LDF of the girders became more sensitive to the variation in girder number as
decreased, especially in case of exterior girder. Note that
is inversely proportional to diaphragm rigidity. As
decreased, the RJG method26 approached the RTB method16,42, as elaborated in the section “Insights into the governing equations”.
Fig. 12.
Influence of diaphragm rigidity and girder number on the LDF of: (a) exterior girder; (b) intermediate girder; (c) interior girder.
Therefore, it can be concluded that the LDFs of girders in staggered widening bridges are significantly influenced by the girder number, diaphragm rigidity and girder positions. Consequently, in the subsequent comparative studies, the calculated LDFs should be categorized into distinct cases based on these three factors: girder number, diaphragm rigidity and girder position.
Influence of flange rigidity
The calculated LDFs with different
are shown in Fig. 13. The calculated LDFs have been categorized into cases with different diaphragm rigidity, girder number and girder position. Note that for the cases with semi-rigid diaphragm, the upper limit of
as shown in Table. 2 is adopted. The
value for the cases without in-span diaphragm is set as
For the investigated cases in Fig. 13, the parameter
is set as
. The girder width
and closure slab width
are set as
and
, respectively. From Fig. 13, it is observed that: (1) the LDF of the interior girder is more sensitive to the variation in
(flange rigidity) than that of the exterior girder. (2) the variation in
dose not significantly influence the variation in LDF for the exterior girders when
and intermediate girders when
. (3) the LDF of interior girder increased with increasing
(decreasing flange rigidity). (4) for the exterior- and interior-girders when
and intermediate girder when
, the LDF increased with increasing
(decreasing flange rigidity).
Fig. 13.
Influence of flange rigidity on the LDF for the cases of: (a) without in-span diaphragms; (b) semi-rigid diaphragm; (c) rigid diaphragm.
Influence of girder rotational stiffness
The calculated LDFs with different
are shown in Fig. 14. The calculated LDFs have been categorized into cases with different diaphragm rigidity, girder number and girder position. The parameters that remain unchanged, namely
,
and
, are in accordance with those depicted in Fig. 13. The parameter
related to the flange rigidity is set as
. From Fig. 14, it is observed that: (1) the LDF became more sensitive to the girder number
as the diaphragm rigidity increased, as elaborated in the section “Influence of diaphragm rigidity
, girder number
and girder position.” (2) in most cases, the LDF increased as the rotational stiffness decreased (
increased). (3) the LDF became less sensitive to the rotational stiffness
as the diaphragm rigidity increased.
Fig. 14.
Influence of rotational rigidity on the LDF for the cases of: (a) without in-span diaphragm; (b) semi-rigid diaphragm; (c) rigid diaphragm.
Influence of closure slab width
The calculated LDFs with different closure slab width
are shown in Fig. 15. The calculated LDFs have been categorized into cases with different diaphragm rigidity, girder number and girder position. The unchanged parameters including
,
, and
are consistent with those shown in Fig. 14. The parameter
related to the girder rotational stiffness is set as
. From Fig. 15, it is observed that: (1) the LDF of interior girder is the most sensitive to the variation in closure slab width. (2) a greater width of the closure slab led to a diminished supporting effect provided by the adjacent piers. Consequently, the LDF of interior girder increased with increasing closure slab width. (3) For exterior- and interior-girders when
and intermediate girder when
, the LDF increased with increasing closure slab width. (4) For exterior girder when
and intermediate girder when
, the LDF is not sensitive to the variation in closure slab width.
Fig. 15.
Influence of closure slab width on the LDF for the cases of: (a) without in-span diaphragm; (b) semi-rigid diaphragm; (c) rigid diaphragm.
Influence of girder width
The calculated LDFs with different girder width
are shown in Fig. 16. The calculated LDFs have been categorized into cases with different diaphragm rigidity, girder number and girder position. The unchanged parameters including
,
, and
are consistent with those shown in Fig. 15. The closure slab width
is set as
. From Fig. 16, it is observed that: (1) for the cases with two girders (
), the LDF of interior girder decreased whereas the LDF of exterior girder increased as girder width increased. (2) the LDFs of the exterior- and intermediate-girders increased as girder width increased. (3) for the cases without in-span diaphragm, the LDF of interior girder increased as girder width increased due to the increase in tributary area of the deck to each girder, as elaborated by Kong et al.16. (4) when the diaphragm rigidity increased, the transverse connection system of the bridge changed, enabling the girders to collaborate more effectively in sustain vehicular loads. Moreover, the boundary conditions of the interior girder in staggered widening bridges differed markedly from those in conventional widening bridges. Thus, the LDFs of interior girder in staggered widening bridges do not always increase as girder width increased. (5) the abrupt changes in the LDF when
and
is due to the sudden variations in the number of lanes,
, with increasing girder width, as shown in Fig. 17. (6) the exterior girder is less sensitive to the variation in the number of lanes,
than the other girders.
Fig. 16.
Influence of girder width on the LDF for the cases of (a) without in-span diaphragm; (b) semi-rigid diaphragm; (c) rigid diaphragm.
Fig. 17.

Number of lanes corresponding to different girder number and girder width.
Proposed formula of the LDF for staggered widening bridges
Based on the above parametric analysis, a general formula of the LDF for staggered widening bridges are proposed here. The independent variables of this formula are
,
,
,
and
, all of which are within the reasonable ranges presented in Table. 2. The proposed formula is expressed as:
![]() |
![]() |
14 |
where
to
are empirical parameters calibrated according to sample points. The units for parameters
and
are
.
To obtain accurate calculation results of the LDF, a series of sampling points of the independent variables were obtained through Latin hypercube sampling (LHS) method46. To ensure that the sampling points can represent the typical mechanical characteristics of a staggered widening bridge, the variable
was categorized into the cases of rigid diaphragm (
), semi-rigid diaphragm (
) and without diaphragm (
). Then,
samples of the variables,
were obtained via the LHS method for each case of
. The LDFs calculated using the formula with sampling points were then compared with those calculated by the simplified analysis model. The least square method was utilized to optimize the empirical parameters,
to
in the formula. According to the abovementioned parametric analysis, the exterior-, intermediate- and interior-girders exhibited different transverse vehicular load distribution characteristics due to the supporting effects of the adjacent bridges. Meanwhile, the bridges with a smaller girder number (
for exterior- and interior-girders and
for intermediate girder) exhibited different transverse vehicular load distribution characteristics compared with other girder number cases. Therefore, the parameters
to
in Eq.
were calibrated separately for the cases with different girder positions and girder numbers. The final calibrated parameters are presented in Table. 3.
Table 3.
Calibrated parameters for the cases with different girder positions and girder numbers
| Parameter | Investigated cases | |||||
|---|---|---|---|---|---|---|
| Exterior girder | Intermediate girder | Interior girder | ||||
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
|
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
The units for psssarameters
and
are
.
Using the calibrated parameters, the LDFs of each girder calculated by the formula are compared with those obtained from the simplified analysis model, as illustrated in Fig. 18. Meanwhile, the
error bars are also presented in Fig. 18. From Fig. 18, it can be concluded that the proposed formula predicts the LDFs of girders in a staggered widening bridge with reasonable accuracy, satisfying engineering requirements.
Fig. 18.
Comparison between the LDFs calculated for each girder using the simplified analysis model and the proposed formula: (a)
; (b)
; (c)
; (d)
; (e)
; (f)
.
Conclusion and discussion
Conclusions
This study mainly concerns the transverse vehicular load distribution in staggered widening bridges. First, a simplified analysis model was developed based on the RJG method to calculated the transverse distribution of vehicular load for staggered widening bridges. The accuracy of the simplified analysis model was then verified by comparison with finite element analysis results. Then, the transverse influence lines of the girders in a staggered widening bridge were compared with those in a conventional widening bridge. The results indicated that:
staggered widening significantly affects the transverse distribution of vehicular load in the girders, particularly for the interior girder;
due to the supporting effects that shown in Fig.1, the vehicular load borne by the interior girders has been significantly reduced, which indicated that the commonly used LDF equations may prove inapplicable to staggered widening bridges.
Based on the simplified analysis model, parametric analysis was conducted. The influence of critical parameters, such as diaphragm rigidity, flange rigidity, girder rotational stiffness, girder width, girder number and closure slab width on the LDF of girders in a staggered widening bridge was investigated.
Based on the parametric analysis results, a general formula of the LDF was proposed for staggered widening bridges. This formula consists of five independent variables, such as
,
,
,
,
and nine empirical parameters. To verify the accuracy of this formula, a series of sampling points of the independent variables were obtained through Latin hypercube sampling (LHS) method. The LDFs calculated using the formula with sampling points were then compared with those calculated by the simplified analysis model. The least square method was used to optimize the empirical parameters based on cases featuring different girder positions and girder numbers. Comparison between the calculated LDFs using the formula and the simplified analysis model confirmed the applicability and accuracy of the proposed formula.
Applicability and scope of the proposed method
The applicability and scope of the proposed method in this study are elaborated as follows:
First, the simplified analytical model shown in Fig. 4 is valid for large-offset staggered widening bridges, where the existing bridge’s pier is located right at or in the vicinity of the new bridge’s mid-span segment. For small-offset staggered widening bridges, this model may not be applicable, and the complex spatial mechanical behavior of the bridges should instead be evaluated using the three-dimensional finite element method.
Second, from the perspective of engineering practice, large-offset staggered widening bridges have attracted more attention than small-offset ones, mainly for two reasons: (1) The differential deck deflection between adjacent bridges increases with pier offset; thus, for large-offset cases, the bearing capacity of the bridge deck should be evaluated, and the existing bridge deck should be strengthened if necessary8; (2) The supporting effects provided by the existing bridge’s piers also increase with pier offset, so the prestress dosage and structural dimensions of the new bridge’s interior girders should be designed considering these effects to improve economic efficiency.
Third, based on the findings of this study, the implementation of staggered widening reduces the vehicular loads borne by the interior girder. Therefore, the bearing capacity of the existing bridge’s girders does not need to be evaluated, as the supporting effects have reduced vehicular loads borne by the existing girders. In contrast, the prestress dosage and structural dimensions of the new bridge’s girders should be designed considering these supporting effects to enhance economic efficiency. For small-offset staggered widening bridges, insignificant supporting effects lead to no notable reduction in prestress dosage or structural dimensions, so the new bridge’s girders can be designed like those of conventional widening bridges to facilitate engineering practice.
In addition, the proposed simplified analytical model is formulated based on the RJG method26, which is a general approach for multi-girder bridge systems and is not restricted to a specific number of girders or girder type. In the validation section, the prototype bridge with three T-girders was adopted as a representative example to verify the accuracy of the model. The governing equation (Eq. 1) can be readily extended to a general form covering a wide range of girder numbers, following the approach proposed by Kong et al.16 and Chen et al.24. Furthermore, the RJG method is also applicable to other bridge types, such as multi-box girder bridges42. Therefore, the validation results are not affected by the number of girders or bridge types.
Future work suggestions
In this study, the calculation results of the simplified analytical model are compared with those of the three-dimensional finite element model, which is commonly adopted method for LDF calculation in previous studies20–23. Thus, the rationality of modifying the configuration of conventional RJG method to account for the supporting effects provided by the piers of existing bridge has been validated at the theoretical level.
However, it should be acknowledged that this validation belongs to model-to-model numerical verification rather than experimental or field-based verification. The proposed formula is calibrated based on the results from the simplified analytical model. Its accuracy has not been validated through laboratory tests24 or independent field test benchmarks14–19. Thus, the applicability of the proposed method is limited to the theoretical and numerical scope within the parameter range considered in this study.
Fortunately, large-offset staggered widening bridges have been applied in the design of the Guangzhou–Shenzhen Highway expansion project. In future work, field tests will be conducted to collect monitoring data related to the mechanical characteristics of staggered widening bridges, which will be used to further validate the effectiveness and accuracy of the proposed model and formula.
Author contributions
Changxi Wang: Conceptualization, Formal analysis, Writing – original draft; Zixiang Zhang: Methodology, Validation, Writing - original draft; Yu Bao: Formal analysis; Xing Wei: Validation, Writing - review amp; editing; Yu Zhu: Writing - review amp; editing.
Funding
This study was financially supported by the National Natural Science Foundation of China (52078424).
Data availability
Some or all data models, or code that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Kwan, A. & Ng, P. L. Effects of traffic vibration on curing concrete stitch: Part Ⅰ —— Test method and control program. Eng. Struct.29(11), 2871–2880 (2007). [Google Scholar]
- 2.Ng, P. L. & Kwan, A. Effects of traffic vibration on curing concrete stitch: Part Ⅱ —— Cracking, debonding and strength reduction. Eng. Struct.29(11), 2881–2892 (2007). [Google Scholar]
- 3.Chai, Y. H. & Hung, H. J. Waiting period for closure pours in bridge widening or staged construction. J. Bridge. Eng.21(5), 04016006 (2016). [Google Scholar]
- 4.Guo, J., Deng, K., He, M., Zhao, C. & Li, W. Experimental study on the construction stages of an RC closure pour in bridge widening. J. Bridge. Eng.22(12), 06017007. 10.1061/(asce)be.1943-5592.0001155 (2017). [Google Scholar]
- 5.Weatherer, P. J., Fargier-Galbadon, L. B., Hedegaard, B. D. & Parra-Montesinos, G. J. Behavior of longitudinal joints in staged concrete bridge decks subject to displacements during curing. ASCE’s J. of Bridge. Eng.24(7), 04019067 (2019). [Google Scholar]
- 6.ElSafty, A., Okeil, A. M., Torres, K., Tawfiq, K. & Fallaha, S. Investigation of empirical deck design in bridge widening. J. Bridge. Eng.25(10), 04020079. 10.1061/(asce)be.1943-5592.0001617 (2020). [Google Scholar]
- 7.Zhang, R., Warn, G. P. & Radlińska, A. Assessment of transverse deck cracking in bridges during staged construction. ASCE’s J. Perform. Con. Struct. Facil.38(2), 04024003 (2024). [Google Scholar]
- 8.Yang, Z. & Zhang, Z. Experimental evaluation of staggered widening bridges. J. Bridg. Eng.30(6), 05025004 (2025). [Google Scholar]
- 9.AASHTO. AASHTO LRFD bridge design specifications 8th ed. (AASHTO, 2017). [Google Scholar]
- 10.Kim, S. & Nowak, A. S. Load distribution and impact factors for I-girder bridges. J. Bridge. Eng.2(3), 97–104 (1997). [Google Scholar]
- 11.Barr, P. J., Eberhard, M. O. & Stanton, J. F. Live-load distribution factors in prestressed concrete girder bridges. J. Bridge. Eng.6(5), 298–306 (2001). [Google Scholar]
- 12.Huo, X. S., Wasserman, E. P. & Zhu, P. Simplified method of lateral distribution of live load moment. J. Bridge. Eng.9(4), 382–390 (2004). [Google Scholar]
- 13.Hughs, E. & Idriss, R. Live-load distribution factors for prestressed concrete, spread box girder bridge. J. Bridge. Eng.11(5), 573–581 (2006). [Google Scholar]
- 14.Huang, J. & Davis, J. Live load distribution factors for moment in NEXT beam bridges. J. Bridge. Eng.23(3), 06017010 (2017). [Google Scholar]
- 15.Terzioglu, T., Hueste, M. B. D. & Mander, J. B. Live load distribution factors for spread slab beam bridges. J. Bridge. Eng.22(10), 04017067 (2017). [Google Scholar]
- 16.Kong, S., Zhuang, L., Tao, M. & Fan, J. Load distribution factor for moment of composite bridges with multi-box girders. Eng. Struct.215, 110716 (2020). [Google Scholar]
- 17.Jiang, D., Terzioglu, T., Hueste, M. B. D., Mander, J. B. & Fry, G. T. Experimental study of an in-service spread slab beam bridge. Eng. Struct.127, 525–535 (2016). [Google Scholar]
- 18.Rambo-Roddenberry, M., Kuhn, D. & Tindale, R. G. Jr. Barrier effect on transverse load distribution for prestressed concrete segmental box girder bridges. J. Bridge. Eng.21(6), 04016019 (2016). [Google Scholar]
- 19.Seo, J., Phares, B. & Wipf, T. J. Lateral live-load distribution characteristics of simply supported steel girder bridges loaded with implements of husbandry. J. Bridge. Eng.19(4), 04013021 (2014). [Google Scholar]
- 20.Eamon, C. D., Kamjoo, V. & Shinki, K. Design live-load factor calibration for Michigan highway bridges. J. Bridge. Eng.21(6), 04016014 (2016). [Google Scholar]
- 21.Ravazdezh, F., Seok, S., Haikal, G. & Ramirez, J. A. Effect of nonstructural elements on lateral load distribution and rating of slab and T-beam bridges. J. Bridge. Eng.26(9), 04021063 (2021). [Google Scholar]
- 22.Puckett, J. A., Huo, S. X., Jablin, M. & Mertz, D. R. Framework for simplified live load distribution-factor computations. J. Bridge. Eng.16(6), 777–791 (2011). [Google Scholar]
- 23.Dicleli, M. & Yalcin, O. F. Incorporation of skew effects in live-load distribution factors developed for typical integral bridges. J. Bridge. Eng.23(2), 04017135 (2018). [Google Scholar]
- 24.Chen, C., Yang, C., Zhang, K., Wang, W. & Dong, L. Structural behavior and load distribution factor of a T-girder bridge with various truss diaphragms. J. Bridge. Eng.28(12), 04023094 (2023). [Google Scholar]
- 25.Loo, Y.-C. & Cusens, A. R. The finite strip method in bridge engineering (Akademiai Kiado, 1978). [Google Scholar]
- 26.Kuzmanovic, B. O. & Sanchez, M. R. Lateral distribution of live loads on highway bridges. J. Struct. Eng.112(8), 1847–1862 (1986). [Google Scholar]
- 27.Pearson, A. N., Birely, A. C., Yarnold, M. & Hurlebaus, S. Field study of bonded links slabs subjected to ambient live and thermal loads. J. Bridge. Eng.29(6), 04024029 (2024). [Google Scholar]
- 28.Abendroth, R. E., Klaiber, F. W. & Shafer, M. W. Diaphragm effectiveness in prestressed-concrete girder bridges. J. Struct. Eng.121(9), 1362–1369 (1995). [Google Scholar]
- 29.Cai, C. S., Chandolu, A. & Araujo, M. Quantification of intermediate diaphragm effects on load distributions of prestressed concrete girder bridges. PCI. J.54(2), 48–63 (2009). [Google Scholar]
- 30.Hou, P. et al. Experimental and simulation studies on the mechanical performance of concrete T-girder bridge strengthened with K-brace composite trusses. Structures43, 479–492 (2022). [Google Scholar]
- 31.Lin, C. S., & Vanhorn, D. A. The effect of midspan diaphragms on load distribution in a prestressed concrete box-beam bridge. In Philadelphia bridge. (Bethlehem, PA: Lehigh University, 1968).
- 32.Li, L. & Ma, Z. J. J. Effect of intermediate diaphragms on decked bulb-tee bridge system for accelerated construction. J. Bridge. Eng.15(6), 715–722 (2010). [Google Scholar]
- 33.Barth, K. E. et al. Enhancing live load distribution predictions for press-brake-formed tub girder bridges for moments in interior girders: A statistical approach. J. Bridge. Eng.30(6), 04025025 (2025). [Google Scholar]
- 34.Crabtree, B., Ross, B. E., Cousins, T. E. & Ziehl, P. Live-load testing of flat precast slab bridge to determine joint efficiency and distribution factors for moment. J. Perform. Constr. Facil.35(1), 04020134 (2021). [Google Scholar]
- 35.Firoj, M., Yang, D. Y., Wood, T. A. & White, K. Uncertainty analysis and quantification of the AASHTO live load distribution model for moderately buried-reinforced concrete box culverts. J. Bridge. Eng.10.1061/JBENF2.BEENG-7393 (2025). [Google Scholar]
- 36.Guo, K., Bairan, J. M. & Liu, Z. Simplified analytical approach for calculating the transverse load distribution of precast slab bridges with and without concrete overlays. J. Bridge. Eng.29(7), 1–14 (2024). [Google Scholar]
- 37.Hedegaard, B. D., Dymond, B. Z., Linderman, L. E. & Hill, K. A. Load distribution in reinforced concrete slab span bridges. J. Bridge. Eng.29(9), 04024065 (2024). [Google Scholar]
- 38.Mensah, S. A. & Durham, S. A. Live load distribution factors in two-girder bridge systems using precast trapezoidal U-girders. J. Bridge. Eng.19(2), 281–288 (2013). [Google Scholar]
- 39.Ndong, A. K., Sherif, M., Kassner, B., Harris, D. K. & Ozbulut, O. E. Potential improvement in rating factors of concrete T-beam bridges through refined analysis: Evaluation of distribution factors. J. Bridge. Eng.27(9), 04022081 (2022). [Google Scholar]
- 40.Ndong, A. K., Sherif, M., Kassner, B., Ozbulut, O. E. & Harris, D. K. Refined analysis of steel girder bridges for computation of live load distribution factors considering effects of freight and emergency vehicles. Eng. Struct.293(15), 116630 (2023). [Google Scholar]
- 41.Pinkham, J., Davids, W. G. & Schanck, A. Assessment of moment live load distribution in FRP composite tub girders for highway bridges. J. Bridge Eng.29(5), 04024024 (2024). [Google Scholar]
- 42.Yu, Z., He, S., Ren, L. & Zhang, L. Calculation of transverse load distribution of multi-beam short box girder bridge considering influence of torsional resistance. China J. Highw. Transp.33(7), 105–113 (2020) (In Chinese). [Google Scholar]
- 43.Cui, H. Y. et al. Individual girder-level evaluation of assembled multigirder bridges using load transverse distribution identified from moving load test data. Eng. Struct.343(15), 121289 (2025). [Google Scholar]
- 44.He S.H. & Xie R.W. Calculation method of transversal distribution for highway bridge load. (Beijing, China Communications Press 1996).
- 45.Shao, X.D. & Gu, A.B. Bridge Engineering 2nd eds (Beijing, China Communications Press, 2007).
- 46.Shields, M. D. & Zhang, J. X. The generalization of Latin hypercube sampling. Reliab. Eng. Syst. Saf.148, 96–108 (2016). [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Some or all data models, or code that support the findings of this study are available from the corresponding author upon reasonable request.












































































































































