1 Introduction
In general, steel-normal concrete (NC) composite beams are widely used in bridge engineering because they make full use of the steel tensile strength and NC compressive strength along with advantages of light self-weight, convenient construction, and economic performance. However, in a continuous composite beam (CCB), cracking of the NC flange plate may easily occur at a low tensile load due to the low tensile strength of NC at the middle support or any pier supports. Accordingly, these cracks allow deicing chemicals to corrode reinforcing steel and spall concrete of the NC flange plate in the serviceability limit state. As a result, the durability of the composite beam is significantly reduced. In addition, severe cracking in the NC deck slab reduces the stiffness of the CCB, resulting in a large deformation of the CCB and large tensile stress of the steel beam at the mid-span section. The negative bending region becomes the weakest part of CCB bridges and restricts CCB bridges’ application in practical engineering.
At present, improving the cracking resistance of the NC deck slab in the negative moment region is one of the main tasks in designing CCB bridges. Several practical approaches have been proposed to enhance the cracking resistance. The strategies mainly include increasing steel reinforcement, applying prestressing force [
1,
2], grouped stud-shear connector [
3–
5], optimizing construction process [
6–
8], and partial shear connection [
9]. As new types of concrete materials with superior tensile performance have been developed, such as steel fiber reinforced concrete (SFRC), engineered cementitious composite (ECC), and ultrahigh performance concrete (UHPC). These concrete types might become an effective alternative to improve the cracking resistance of the NC deck slab in the negative moment region, solving the cracking problem in steel-NC CCBs [
10–
12]. Mainly, UHPC is a new type of cement-based composite material (compressive strength greater than 150 MPa, and post-cracking strength greater than 5 MPa). Uniformly distributed steel fibers with a volume fraction of about 1%–4% in the UHPC matrix significantly improve the tensile performance and toughness of the material. Thus, UHPC exhibits a tensile strain-hardening behavior and a superior ability to limit the crack width [
13–
15]. Additionally, UHPC has good compaction, ultra-low permeability, and excellent durability [
16]. Moreover, UHPC has good bond strength with steel rebars, and the anchoring length of the steel rebar can be reduced to less than ten times the diameter of the steel rebar [
17].
Past research has extended the application of the UHPC materials to steel-concrete composite structures and carried out experimental studies with respect to the structural performance of steel-UHPC composite beams. Wang et al. [
18] fabricated nine steel-UHPC composite beams with different concrete-steel beam interface treatments, including UHPC directly cast on different plate textures, interface with an epoxy-based adhesive, and traditional headed studs. The tested results indicated that the pure connection between the UHPC and the plate was extremely weak, while the adhesive and traditional-headed studs performed well. Yoo and Choo [
19] proposed a new type of steel-UHPC composite beam, composing a UHPC slab and a steel girder without a top flange (i.e., inverted-T steel girder). The studs were welded on the web of the inverted-T steel girder to get a good bond strength between the UHPC and slab. Zhang et al. [
12] recently conducted negative bending moment tests on steel-UHPC composite beams with stud connectors and bolt connectors. Test results indicated that the steel-UHPC composite beams exhibited excellent cracking and flexural performance under the negative moment. Also, Zhu et al. [
20] experimentally investigated the flexural behavior of steel-UHPC composite beams with joints under a negative bending moment. Results showed that T-shaped joints in intermediate connections under the negative bending moment presented a more reliable mechanical performance than rectangular joints with the same joint bottom width. Hu et al. [
21] investigated the structural behavior of steel-UHPC composite beams with a shear pocket under the positive bending moment, whereas Qi et al. [
22] studied the structural performance of steel-UHPC composite beams with a shear pocket under the negative bending moment. Both studies showed that the use of UHPC significantly improved the flexural capacity and stiffness of the composite beam, and the full composite action between the UHPC and steel was confirmed. Wang et al. [
23] proposed a fully dry-connected prefabricated composite beam and experimentally investigated its flexural behavior under the negative moment. Results showed that the fully dry-connected composite beam improved cracking resistance and stiffness under the negative bending moment.
Research to date on using UHPC materials in composite beams was mainly restricted to studying the structural behavior of steel-UHPC composite beams. There are limited studies using UHPC to strengthen the cracking resistance of negative bending moment regions in steel-NC composite beams [
24]. Moreover, no published literature reports the actual application of UHPC in the negative moment regions in CCBs. Based on the above background, the primary objective of this study was to propose a new structural connection that used UHPC to strengthen the cracking resistance of negative bending moment regions in steel-NC CCB bridges. A full-scale CCB bridge field test was conducted to verify the feasibility and effectiveness of this new configuration with UHPC in the negative bending moment region. In order for the outcomes of the new structural connection configuration with UHPC to be used in the design and analysis, the design philosophy and structural characteristics were described. Also, the advantages of the new connection configuration were summarized and compared to conventional steel-concrete connections in terms of cracking resistance, durability, constructability, and economy. The effect of the new configuration with UHPC in the negative bending moment region on CCB behavior was also investigated through a validated 3D finite element bridge model using the strain and deflection results reported in the field investigation.
2 Experimental project program
2.1 Project description
The superstructure of Changsha−Yiyang Expressway Expansion Project in Hunan Province adopts steel-concrete composite beams with a standard span of 30 m, and the structural system during construction transforms a simple-supported beam into a continuous beam. Among them, the right side of the bridge is the first continuous unit (namely Jinling Road Viaduct) with a span of 4 ×30 m. The elevation of the Jinling Road Viaduct is shown inFig.1.
The width for the right side of the bridge is 16.5 m, and the standard cross-section is shown in Fig.2. The concrete bridge deck, with a thickness of 250 mm at the transverse mid-span section and 400 mm at the top of the steel beam section, is made of C50 concrete (characteristic cube compressive strength of 50 MPa). With an 1180 mm height, the steel beam is set at the bottom of the concrete bridge deck and fabricated by the Q345 steel. The thickness and width of the top steel beam flange plate are 16 and 400 mm, respectively, and the thickness of the web is 16 mm. The width of the bottom flange is 600 mm, with thickness ranging from 24 to 36 mm. At the pier section, the steel beams between the adjacent two spans are connected through the transverse ordinary concrete beam, with a width of 1200 mm. The total height of the composite beam at the mid-span section is about 1580 mm, and the total height near the pier section is increased to 1800 mm.
2.2 Connection design
In the negative moment region of the Jinling Road Viaduct, a novel UHPC thin layer strengthening scheme is adopted and used at pier-1, pier-2, and pier-3, as shown in Fig.3(a). In order to reduce the cracking risk of the NC deck slab in the negative moment region of the CCB, a UHPC thin layer was used to replace ordinary concrete partially. The UHPC thin layer was cast on the top of ordinary concrete to improve the structural and durable protective layers. Considering the economy of the scheme, the UHPC thin layer is only used to replace the ordinary concrete at the top of the concrete bridge deck with high tensile stress in the negative moment region in the CCB. The thickness and longitudinal length of the UHPC layer are 0.11 and 6 m, respectively. In addition, the micro-expansive NC is used to cast the transverse fulcrum beam (longitudinal length of 1.2 m and height of 1.69 m) to avoid cracking due to shrinkage, which easily appears in the NC under the constraint of steel structure. The micro-expansive NC and steel beam are connected by shear studs and perfobond rib shear connectors. The diameters of steel rebars in the UHPC thin layer and the bottom NC deck slab are 20 and 25 mm, respectively. Both have a transverse spacing of 100 mm. The strength grade of steel rebars is HRB400. The longitudinal rebars in the UHPC thin layer are bounded with an overlapping length of 400 mm (i.e., 20
d), while the longitudinal rebars in the ordinary concrete deck slab are welded on one side with a weld length of 250 mm (i.e., 10
d) [
25].
Fig.3(b) shows an original design scheme (NC reinforced with heavy steel rebars). Two layers of longitudinal rebars with a diameter of 28 mm and one layer of longitudinal rebars with a diameter of 25 mm are used, and their transverse spacing is 100 mm. From the viewpoint of cracking resistance, economic performance, durability, and constructability, the UHPC thin layer strengthening scheme is compared with the original design scheme, as shown in Tab.1. Compared to the conventional scheme, the UHPC thin-layer scheme exhibits higher cracking resistance, and the consumption of steel rebars is reduced by 19.0%. In addition, the UHPC thin layer scheme exhibits equivalent economic performance and better durability. Moreover, the weld length is reduced by 36.5% [
25], which is convenient for accelerated bridge construction.
2.3 Construction procedures in the field
The UHPC used in the actual bridge was made of a standardized UHPC premix provided by a factory and mixed in the bridge site. The components of UHPC premix included Portland cement, quartz sand, quartz powder, silica fume, fly ash, steel fiber, and high-range water reducing agent. The standardized matrix design of UHPC is shown in Tab.2. Among them, straight steel fibers have a diameter of 0.12 mm, a length of 8 mm, and a volume fraction of 2.5%. During construction of the UHPC thin layer, three 100 mm × 100 mm × 100 mm cubic specimens, three 400 mm × 100 mm × 100 mm prism specimens, and three 300 mm × 100 mm × 100 mm prism specimens were fabricated to test compressive strength, flexural strength, and elastic modulus, respectively, in accordance with the code [
26]. These specimens had the same curing conditions as the proposed structural configuration for the negative bending moment region in the bridge site. The compressive strength, flexural strength, and elastic modulus of the UHPC at 3, 7, 14, and 28 d were tested in accordance with the code [
26], as shown in Tab.3.
The first step in the construction procedures for the UHPC thin layer is to weld and fix the steel rebars with a reinforcement ratio of 3.5% in the design position, as shown in Fig.3(a). Then, the micro-expansive NC was poured to form the transverse fulcrum beam. After 40 d of natural curing of the transverse fulcrum beam, concrete on the upper surfaces of the precast concrete deck slab and the transverse fulcrum beam was chiseled. It should be noted that a complete NC detail (see Fig.3(b)) would have resulted in faster construction of the bridge system. Thirdly, the steel rebars with the reinforcement ratio of 1.4% in the UHPC thin layer were bound. Then, before pouring UHPC, the upper surfaces of the precast concrete deck slab and transverse fulcrum beam need to be saturated. After that, the UHPC was cast with an intelligent casting machine. Finally, the UHPC was cured with adequate moisture for no less than 7 d. The images for the main construction procedures are shown in Fig.4.
2.4 Instrumented bridge sections
The mechanical performance of the negative moment region at the top of the first pier for Jinling Road Viaduct was tested. As shown in Fig.5, the concerned sections are section A (center of the pier), section B (micro-expansion NC-NC interface), and section C (NC-UHPC interface), whose distances away from the centerline of the first pier are 0, 0.6, and 3 m, respectively. The girders #2 and #7 in Fig.2 are selected as the test girders. Electrical resistance strain gauges are installed into the concrete, steel beam, and steel rebar surfaces to obtain the strains in the negative moment region of the actual bridge. The layout of measuring points for the strains in five sections (I, II, III, IV, and V) is depicted in Fig.5. The measurement results of the strain gauges had considered the thermal effects during the data acquisition. In addition, the Level measurement is used to record the deflection of the mid-span sections of the first and second spans in girder #2. Moreover, visual inspection for the top surfaces of the UHPC thin layer and concrete deck slab is performed during the loading process.
2.5 Load vehicles and load cases
With a total weight of 300 kN, the three-axle truck was used for the field load testing, as shown in Fig.6(a). The wheelbase and axle load of the load vehicle are shown in Fig.6(b). Before loading, each truck was weighed and checked to make the axle loads meet the test requirements and ensure the reliability of the axle load of each load vehicle during the test. The transverse and longitudinal distributions of loading vehicles are shown in Fig.7 and Fig.8, respectively. The distribution of longitudinal wheel loads is determined according to the influence line of the tensile stress of the concrete deck slab for the three concerned sections (A, B, and C) in the first span.
In the loading test, each load case is divided into three loading levels. The loading sequence and vehicle numbers are shown in Tab.4. In Tab.4, Level 1 and Level 2 represent applying trucks 1&4 and trucks 2&5, respectively; Level 3 represents applying trucks 3&6 (truck six does not appear in A-DL and C-DL), corresponding to reaching the design load (DL) value or the overload (OL) value. The vehicle numbers are shown inFig.8. The holding time for each loading level was longer than 15 minutes, and measuring data was recorded after the mid-span deflection by the Level reading was stable. The bridge load testing was carried out after UHPC was naturally cured for 38 d. The image of the actual bridge test is shown in Fig.9.
3 Three-dimensional finite element modeling
Because of the limitation in the number of large-scale specimens as well as time and costs, finite element analyses were conducted. The finite element model of the composite deck system was built in ABAQUS, as shown in Fig.10. All four spans’ superstructures were simulated. 8-noded linear-displacement brick elements C3D8R with hourglass-control reduced integration were used for the concrete deck. Each node of C3D8R elements had three translational degrees of freedom. C3D8R elements were well suited for large deformation, large strain, and nonlinear problems. 4-node general-purpose shell elements S4R (reduced integration with hourglass control and finite membrane strains) were used to simulate the steel beam. 2-node linear displacement truss elements were used to simulate the steel rebar. A convergence study was performed to choose the appropriate mesh that would provide reliable results and require less computational time. The global mesh size of the composite system was 0.1 m, based on the mesh-sensitivity analysis.
3.1 Material and interface properties
All steel rebars were embedded into the concrete using embedment technology in ABAQUS. The embedment technology ensured that the translational degrees of freedom of the nodes on the rebar element were restrained to the interpolated values of the corresponding degrees of freedom of the concrete elements. Other interactions (surface to surface) between components were described by TIE technology in ABAQUS, which forced the surfaces to share the same degrees of freedom. Because the structure was still in a linear elastic state during test loads, the bond-slip effect and material non-linearity were neglected in the finite element analysis. Linear elastic material properties (concrete and steel) were defined in the FE model, and the elastic modulus of each material was shown in Tab.5.
3.2 Loading and boundary conditions
The applied load on the superstructure included the vehicle load, self-weight load, and guardrail load. A steel block simulated the vehicle load (i.e., wheel load) with dimensions of 0.2 m × 0.6 m. The steel blocks’ position on the bridge deck was consistent with the actual wheel position under each load case in the field test. In each load case, three steps were simulated in ABAQUS to replicate the three load levels applied in sequence. Line constraints (UX = UY = UZ = 0) were applied on the most left and right boundaries. Line constraints (UX = UY = 0) were applied to the intermediate supports. Both the boundary conditions applied at the end supports and the intermediate piers were pinned supports.
4 Results and discussion
4.1 Experimental results
4.1.1 Concrete, steel beam, and rebar strains
The concrete tensile strains at the top surface of the negative bending moment regions of the girders #2 and #7 are measured by strain gauges arranged in the bridge deck. Fig.11 shows concrete tensile strains at different sections (I, II, III, IV, and V) under different load cases. The five sections have different distances from the pier centerline, which are presented by the horizontal axis. Fig.11(a) is provided to show these sections’ positions. For girder #2, the concrete tensile strains at the different load levels for each load case are plotted in Fig.11. The concrete strains in girder #2 almost linearly increase with the load level for the design load cases (A-DL, B-DL, and C-DL) when the strains versus the load level (Level 1, Level 2, and Level 3) figure is plotted for each section. A significant jump in the strain at Section II (600 mm away from the centerline of the pier) is occurred due to the discontinuity of the NC deck. The same reason applies to Fig.12.
Furthermore, the maximum measured concrete tensile strains under each load case at the concerned sections (I or A, II or B, and III or C) in the girder #2 shown in Fig.11 approach the design stresses in Tab.6. No visible cracks were observed under each load case, which shows a high cracking resistance of the UHPC-jointed structure. From Fig.11, the concrete tensile strains of two girders are different in the same section. Under each load case, the concrete tensile strain is more significant in girder #2 as the girder is closer to the centerline of load vehicles than girder #7. Meanwhile, two cases for each concerned section are considered: the design load case (the experimental tensile stress on the surface of the concrete deck slab at a specific section reaches about 1.0 times the design tensile stress) and the overload case (the experimental tensile stress on the surface of the concrete deck slab at a specific section reaches about 1.15 times the design tensile stress). Thus, the testing load cases are divided into six categories, as shown in Tab.6.
Strains at the bottom flange surface of the steel beam of girders #2 and #7 are measured by strain gauges. Fig.13 shows bottom flange compressive strains at different sections in the negative bending moment regions of the girders #2 and #7 under the different load cases. From Fig.13, the maximum compressive strain at the bottom flange in the girder #2 is 516.5με under the load case of A-OL, which is much smaller than the yielding strain of steel (around 0.002). This indicates that steel beam stress under the overload case still has a high safety factor. The bottom flange compressive strain distribution along the transverse direction is similar to the concrete strain at the top surface.
Longitudinal rebar strain inside the UHPC layer in negative bending moment regions of the girders #2 and #7 are measured by strain gauges. Fig.12 shows rebar strains at different sections of the girders #2 and #7 under the different load cases. This figure is similar to Fig.11 and Fig.13. It can be seen that the maximum tensile rebar strain is 243.7με which is much less than the yielding strain steel. As is well known, the bridge under operation may be subjected to various loads, and these long-term and short-term loads have an adverse effect on the joint performance of the bridge. However, the concrete stress, steel beam stress, and rebar stress are at a low level from this field test. Although this field test was conducted before the operation, a high safety factor is produced for the bridge without long-term loads considered. Therefore, it is inferred that the high safety factor of this jointed bridge would compensate for the adverse effect of the various loads.
4.1.2 Deflection
The deflection under the different load cases was recorded. Fig.14 shows the mid-span deflection of the girder #2 (first and second spans) versus load levels. In the first two levels of all load cases, the mid-span deflection increases in both first and second spans. However, for the third level of the A-DL and C-DL load cases, the mid-span deflection in the second span decreases due to two vehicles in the second span and three vehicles in the first span. This unsymmetrical loading decreases the mid-span deflection in the second span at the third level. For the third level of the A-OL and C-OL load cases, the mid-span deflection at the second span increases. Meanwhile, their mid-span deflection at the first span is slightly smaller than in the A-DL and C-DL load cases. Overall, the maximum mid-span deflection is 18.1 mm, which is less than the limited value (i.e.,
L/500 = 30000/500 = 60 mm) defined in the code [
27].
Based on the limited use of the UHPC thin layer cast on the top of ordinary concrete to improve both the structural and durable protective layers, the detailed structural characteristics of the UHPC thin layer scheme are warranted. In order to improve the cracking resistance and durability of CCB, the excellent mechanical properties and durability of UHPC can be well utilized, which can potentially reduce the maintenance cost in the long term.
Improving cracking resistance in the negative moment region, UHPC has a dense microstructure mixed with plentiful steel fibers and exhibits a tensile strain-hardening behavior. Due to the bridging effect of steel fibers, UHPC shows an excellent ability to limit the crack width. Furthermore, the cracking resistance of UHPC can be further improved by configuring appropriate steel rebars [
28]. On the other hand, due to the excellent cracking resistance of UHPC at the upper layer, the crack width expansion of ordinary concrete at the bottom layer can be effectively limited, and thus improving the cracking resistance of concrete deck slab subjected to the negative moment in the CCB bridge [
29].
Simplifying the on-site construction process, when the concrete deck slab in the negative moment region in the CCB is strengthened by the UHPC thin layer, the excellent tensile properties of UHPC can be fully utilized to resist the negative moment. Therefore, the prestressed tendons applied in the ordinary concrete in the negative moment region can be canceled. On the other hand, the anchorage length of the steel rebar can be reduced to 10 times the rebar diameter due to the excellent bond performance between UHPC and steel rebar [
17]. It is possible that the welded connections for on-site steel rebars can be avoided, and thus greatly simplifying the construction process in the negative moment region in the CCB.
Extending durability and reducing the life cycle cost of the structure, due to the excellent durability of UHPC, the UHPC layer does not show common concrete bridge diseases such as concrete carbonization, chloride ion erosion, and alkali-aggregate reaction [
30]. Meanwhile, the UHPC layer can be used as a durable protective layer to ensure the structural durability of ordinary concrete, studs, and steel beams. Hence, the durability of composite beams subjected to the negative moment is significantly improved, and the maintenance cost during operation could be effectively reduced. Generally speaking, the application of the UHPC materials to the concrete bridge deck in the negative moment region in the CCB can not only improve the tensile strength and cracking resistance of the bridge deck but also simplify the construction process, improve the durability, and reduce the overall lifecycle cost of the bridge deck in the negative moment region.
4.2 Three-dimensional finite element model validation
The finite element analysis results in terms of the tensile strain in the bridge deck, the tensile rebar strain in the UHPC layer, and the compressive strain in the bottom flange of the steel beam are compared with the test results. Fig.15–Fig.17 show the comparison between them in sections I, II, III, IV, and V. The distances of these sections to the center of the transverse beam can be seen in Fig.11(a). The numerical strains show a similar distribution trend to the experimental strains in the negative bending moment region. The FE curves in Fig.15–Fig.17 for the DL and OL cases are very close to each other, whereas the experimental curves for the same load cases show the disparity. Also, in some cases, the FE curves are nearly half the values for the experimental results. The strain predictions are unconservative. Tab.7 compares numerical and experimental mid-span deflection of girder #2 in the first span. It can be seen that the numerical mid-span deflection in the first span is slightly higher than the experimental one. These differences in strains and mid-span deflection are probably because of measurement errors, assumption of the material constitutive relationship, and unreasonable surface interactions between components in the finite element model. However, the developed finite element model provides a conservative prediction for the structural deflection under the field loads. Fig.18 shows numerical stress results of the cast-in-place UHPC slab on the top of the pier 1, prefabricated concrete slab, and steel structure near the top of the pier 1 under the load case of C-OL. It can be seen that all concrete stresses are within allowable ranges as specified in Tab.6 and and steel stresses are within elastic range.
4.3 Stress distribution in negative bending
As UHPC has high cracking resistance and post-cracking performance, a short discussion on the stress redistribution phenomenon should be interesting to the readers. The concrete cracking in the negative bending moment region results in the stiffness reduction of the steel-concrete composite section. The negative bending moment near the support region is decreased, and the positive bending moment in the middle span is increased. This is called the stress redistribution phenomenon after concrete cracking. After stress redistribution, the concrete cracking in the negative bending moment region is aggravated, and the composite section stiffness is further reduced. In turn, stress redistribution repeatedly occurs, which has an adverse effect on the structure. The steel-concrete composite section near the pier in the present study is shown in Fig.19. Fig.19(a) shows a UHPC-NC composite section at the top of the steel beam, and Fig.19(b) shows an assumed NC composite section at the top of the steel beam. Based on Eurocode 4 [
31], the composite section belongs to the first type I section, which exhibits good plastic deformation capacity and stress redistribution. Here, three cases are considered, including uncracked NC, cracked NC, and cracked UHPC-NC, to investigate the stress redistribution after concrete cracking in the negative bending moment region. The steel-concrete composite section stiffness of the three cases is calculated first [
25]. Then, the calculated section stiffness is incorporated into the finite element model under the load case of A-DL.
Fig.20 shows the calculated concrete tensile stress distribution in the negative bending moment region (i.e., five sections in the first span) of girder #2 for the three cases. Compared with the uncracked NC case (EI0), the concrete stresses in the cracked NC (EIcr) and UHPC-NC (EIucr) cases are reduced. Compared with the cracked NC case, the concrete stress in the cracked UHPC-NC case shows a minor reduction. This indicates that a small stress redistribution, high stiffness, and good ability to control cracks are observed in the negative bending moment region with the cracked UHPC layer. This is attributed to the fact that UHPC has stable post-cracking performance due to the fiber bridging effect and good bond strength with steel rebars, while NC does not work after cracking.
4.4 Limitations and future work
The premature bond failure (i.e., cracking) between the two different material layers might affect the durability and stress redistribution of the structure. However, because of the chiseling treatment at the interface, the bond strength from the direct tension test and interface shear strength from the double-sided shear test between the UHPC layer and NC layer can reach 5.63 and 6.55 MPa, respectively [
32,
33]. Both studies confirm that the bond between the UHPC and NC substrate increases with the roughness degree, causing the appearance of the failure in the NC rather than at the bond line for the rough surface. Although the chiseling treatment in practice increases the construction difficulty, the construction time, and the possible damage on the NC substrate, this measure is a reliable interface treatment method to avoid premature bond failure based on the previous studies. Therefore, in the finite element simulation, the perfect bond between the UHPC and NC was considered, and the TIE technology available in the ABAQUS was utilized to achieve the perfect bond. Indeed, there are limited strain results near the bonding surface reported in the present study. As shown in Fig.5, section C has strain gauges on the top UHPC surface near the interface between the UHPC and NC. From the strain gauge data in this position (see Fig.11(b) with the load cases of B-DL and B-OL), the maximum values are less than the allowable tensile cracking value. In fact, this strain data represents the tensile UHPC surface strain, although the strain gauges are near the interface. To investigate possible interface slip (i.e., shear failure), a desirable test setup needs to be reasonably designed, and the strain gauges (parallel or perpendicular to the interface) need to adhere to both UHPC and NC surfaces near the interface at least. However, as this newly built bridge will be subjected to live load (such as vehicle load) and environmental load (such as seasonal temperature change), the bridge conditions will be inspected in a fixed period to check the potential failure problem (such as interface cracking between the two material layers). Also, the fatigue performance of the UHPC enhanced negative bending moment region will be investigated by conducting tests at the lab and field and these data will be instersting for UHPC structure research community.
On the other hand, in the durability of the structure, it has been reported that the UHPC has been used to strengthen the corroded steel beam [
34–
36]. The corrosion damage was encased in the UHPC along with welding the shear studs to the undamaged web portion, changing the load transfer from the girder (i.e., corroded web and stiffener) to the local UHPC. This repair successfully restored the original design capacity of a corroded girder. Moreover, through the accelerated electrochemical corrosion with the penetration of chlorides to the losing bond between the UHPC and web, the mechanical performance of the embedded shear studs did not weaken because the ions did not penetrate the concrete. These studies demonstrated the UHPC application for improving the performance and durability of the corroded beam. Therefore, the long-term durability of this bridge will be monitored, and the durability issue for the interface between the UHPC and NC will also be studied and reported in future work.
The finite element model results show relatively large discreteness from the field test results. Because the perfect bonds among the components’ interactions and the TIE technology available in ABAQUS were adopted, and the non-linearities of the materials (such as UHPC) were not considered in the finite element model as the field test results were within the elastic stage. However, the finite element model results show a similar, varying trend to the field test results. Meanwhile, the concrete stiffness adjustment to consider the stress redistribution after cracking in the hogging moment region is simplistic. However, this assumption provides consideration for the advantage of the UHPC strengthened hogging moment region after cracking. Further studies are needed to highlight this advantage by the excellent post-cracking performance of UHPC.
5 Summary and conclusions
This study proposed a steel-concrete composite continuous girder bridge with a UHPC thin layer applied in the negative bending moment region. A field loading test was conducted on a full-scale bridge, and a finite element model for the bridge superstructure was established to replicate the mechanical response of the newly UHPC jointed structure. Compared to conventional steel-concrete composite beams, the proposed scheme has advantages in cracking resistance, economy, durability, and constructability. The construction procedures for the proposed scheme in the full-scale bridge were described. From the field load testing, the concrete strain on the bridge deck, the rebar strain, and the bottom flange strain of the steel beam were measured, and these values were within the elastic range of the materials. No cracks were observed on the surface of the concrete bridge deck during the test. The maximum mid-span deflection measured among different load cases was smaller than the limited value defined in the code [
27]. The finite element model shows a conservative prediction in the mid-span deflection. Further discussion on the stress redistribution in the negative bending moment region after concrete cracking highlights the advantage of UHPC in terms of the post-cracking performance. The finite element model and the field load testing demonstrate that the UHPC thin layer can be securely and successfully applied in the negative bending moment region (i.e., over pier) of a full-scale steel-concrete composite continuous girder bridge.
PDF
(10135KB)
