STUDY ON REINFORCEMENT OF FABRICATED HOLLOW SLAB BRIDGE BY POLYURETHANE-CEMENT COMPOSITE (PUC)

In this paper, a new polyurethane-cement composite (PUC) material is used to reinforce a 25-year hollow slab bridge. PUC material is composed of polyurethane and cement, which has good mechanical properties. After pouring PUC material at the bottom of the hollow slabs, the traffic can be restored in a short time. Ultimate bearing capacity was discussed based on the concrete structures. The bending failure mode of concrete reinforced T-beam bridges strengthened with PUC was controlled through PUC design strength. The main construction process was introduced, including concrete surface treatment, supporting template and pouring. To investigate the effectiveness of PUC in strengthening bridges, load tests were conducted before and after the reinforcement. The test results showed that PUC possibly removed the bridge load and increase the stiffness of the hollow slabs. The maximum carrying capacity and stiffness of the main girder were increased by about 20% and 28% after PUC material was poured. Meanwhile, the width of the cracks was reduced in different degrees.


INTRODUCTION
Concrete bridges are uninterruptedly unprotected to external environments. During their life spans, concrete bridges frequently experience situations that cannot be predicted during design. The durability of a concrete bridge is reduced by construction defects, overloaded vehicles, material characteristics, and environmental variations [1,2]. Replacing bridges to improve bridge performance can cause economic loss and inconvenient vehicle traffic [3]. Many researchers have studied proper repair and strengthening methods to solve the problems that occur in concrete structures [4,5].
To improve the working ability of concrete bridges, many techniques have been used in strengthening. The most common methods for strengthening beams have been the use of Carbon Fibre Reinforced Polymer (CFRP), steel plate bonding, high strength wire mesh reinforcement and others; these methods are widely used at present [6]. CFRP has good performance, high strength and light weight, but there are some disadvantages of CFRP reinforcement, the most important is Figure 1 shows a lateral view of Piao Fa Bridge. The bridge was built in 1995 and carries an average daily traffic of 800 vehicles. The bridge has 5 spans, each span being 20 meters of simply supported hollow slabs. Central supports consist of reinforced concrete bents, each supported by two circular columns. The total width is 12.0 m, and the width of the driveway is 11.0 m. The bridge has two lanes and carries one lane of traffic in each direction. The cross section of the superstructure consists of 9 hollow slabs with a width of 1.25m Figure 2.

Fig. 2 -Cross-section of the bridge
The bridge has different degrees of damage caused by aging, overload, surging traffic and environment. During routine inspection, salt infiltration was observed in the bridge superstructure. Many hollow plates have the phenomenon of alkali spreading. Many vertical cracks appear on the bottom of the hollow slabs across the middle area. Serious cracks occurred in the edge-beam, as water from the pavement cracks affected the durability of the girder. Reinforcing the structure was selected because the owner did not agree to replace the bridge or limiting traffic. PUC strengthening method was chosen based on its application of being the least intrusive with traffic and being the most practical.

MATERIAL CHARACTERISTICS
This project adopts polyurethane cement material with good mechanical properties with a density of 1500kg/m 3 to strengthen the hollow plate. The bending stress-strain curves of PUC with the same density are approximately the same. Bending test for PUC material is shown in Figure 3. The density of PUC mixed in the construction site was 1,524 kg/m 3 . The bending stress-strain curves of five groups of stirred PUC specimens at the construction site are shown in Figure 4. Ultimate strength ( Pu f )was about 42.3MPa. Design strength ( Pu f )was designed to be 3/4 of the ultimate strength, and the corresponding strain was 0.006 under the design strength. The minimum modulus of elasticity ( Pd E ) was 5MPa. Table 1 shows the properties of PUC. Basic properties such as the concrete compression strength and steel yield strength were assumed because the owner of the bridge did not allow coring to obtain the on-site strengths of the materials from the bridge girder. These basic properties c f s f were assumed as described by JTJ023-85 [16] for bridges of that age because no sample could be obtained onsite. The parameters of steel strand refer to the standard drawing of the same span in the same year. The strength and elastic modulus of materials such as steel bar, strand, concrete and PUC are shown in Table 1.

Material
Property Value

STRENGTHEN ANALYSIS
The service capacity of the inspected bridge is insufficient due to cracks, rain erosion and so on. Moreover, the owner of the bridge asked to complete the strengthening work with minimum interference to traffic. Therefore, the PUC material is designed to reinforce the hollow slab bridge. The hollow slab blending bearing capacity was analysed through the assumption that the plane section remained to be planar until failure. Blending bearing capacity was evaluated based on force equilibrium and strain compatibility until failure Figure 5.
The failure of the prestressed hollow slab strengthened with PUC material was slab bending failure because of the strong bonding performance between PUC material and concrete. The flexural capacity of the PUC -reinforced member depended on the concrete crushing or the PUC x represents the depth of the neutral axis and p  represents the strain of the centroid of the PUC section. When concrete crushing failed, the ultimate compressive strain for the concrete was assumed to be 0.003, and the steel strand has been in a yield situation.
According to strain compatibility, the strain of the centroid of the PUC section p ε can be expressed as follows.
(2) The PUC material and steel strand were treated as a nearly linear material, and the stress of the PUC material can be calculated by the following: The standard equilibrium equation is listed in Equation (5) where 0.8   [17]. Therefore, beam failure is not controlled by concrete crushing failure.
The ultimate strain of PUC is about 0.006 when the strength of the PUC material is design strength. The steel bar strain is larger than the yield sign strain and less than 0.01 according to strain compatibility.   (6) and (7).
The above mentioned analysis shows that the bearing capacity of the strengthened beam is controlled by the design strength of the PUC material. The flexural strength increased to 1543.7 kN.m through the pouring of the PUC material. The section of PUC is shown in Figure 6.

STRENGTHENING PROCEDURE
PUC strengthening is similar to the enlarging section method [18,19], but the curing time of the PUC strengthening method is significantly less than the enlarging section method. First, degenerate materials are removed from the concrete surface, and the visible cracks in the surface of the girders are closed through a chemical method. Second, the concrete surface is chiselled to ensure that its roughness meets the China Construction Code of Concrete Structure. Loose materials, which affect the bonding performance of the PUC material, are removed through highpressure water washing.
In achieving a better bonding effect, sufficient time must be ensured to dry the concrete surface naturally. The construction templates were made depending on the design of the dimensions of the PUC material. The template was installed in the location specified Figure 8. The mixing ratio of the PUC components (polyol: polyisocyanate: cement) was 1:1:2 by weight. The mixed PUC material was poured into the templates. First, a water drill is used to form a round hole with a diameter of 40 mm at the hinge joint position. Then, at the deck location, the PUC material is dumped through the circular hole. The templates were removed after two hours Figure 7c.

LOAD TEST
Before strengthening and two months after strengthening, load tests were performed to obtain the service ability of the bridge. These tests are repeated once a year, and they will continue over a period of three years. Three-axle trucks are used as live load. The details of the trucks are shown in Figure 10. The gross rail loads before and after strengthening are listed in Table 2. The bridge test was performed under four conditions. The two trucks were used at different locations for each condition and were assigned with unique letters A and B.   Figure 11. For condition 1, the two trucks were symmetrical with the centre of the bridge, and the distance from the outside wheels to the pavement edge was 2.95 m. For condition 2, the two trucks deviated from the centre line of the bridge, and the distance from the outside wheels to the pavement edge was 0.5m. For both conditions 1 and conditions 2, the third axle of the vehicles are in the middle of the span. For condition 3, the two trucks were symmetrical with the centre of the bridge, and the distance from the outside wheels to the pavement edge was 2.95 m. For condition 4, the two trucks deviated from the centre line of the bridge, and the distance from the outside wheels to the pavement edge was 0.5m. For both conditions 3 and conditions 4, the four axles of the vehicles are in the middle of the span.

STRAIN MEASUREMENT
Strain gages were placed and arranged at the bottom of the mid-span, as shown in Figure  9a. The letter n represents the hollow slab number, and the letter C represents the concrete strain gages. Strain gauges (Cn1) were installed on the bottom of the hollow slab before strengthening, as the bottom of the hollow slabs was covered with PUC after strengthening. The strain gages of the PUC surface are shown in Figure 9b. The letter P represents the PUC strain gages. The gage Pn2 and the gage Cn1 were installed at the same horizontal position. The strains were collected at a rate of two readings per second. Figure 12 shows the concrete strain on the bottom concrete surface under live load in condition 1 before and after strengthening. The strain was the largest in the middle slab. The maximum strain of slab 5 were 75 microstrains for the two trucks (Trucks A + B) before strengthening because of the damage and the weak interaction between the slabs. After strengthening, the maximum strain of slab 5 were 60 microstrains for the two trucks (Trucks A + B). The maximum strains of hollow slab 3 were 50 microstrains and 45 microstrains before and after strengthening, respectively. The strain was reduced to about 5 microstrains. The maximum strains of slab 1 were 31 microstrains and 25 microstrains before and after strengthening, respectively. Therefore, the concrete strain can be reduced after the PUC material is poured. For condition 2, the concrete strains on the bottom concrete surface under live load before and after strengthening are shown in Figure 12. The strain of beam 1 was the largest for the two trucks (Trucks A + B) because the two trucks approached the side of the bridge. The maximum strain of beam 1 were 88 microstrains before strengthening, and the strain became 69 microstrains after strengthening. Before strengthening, the strains of beam 3 and beam 5 were 83 and 76 microstrains, respectively, and after strengthening, the strains were 68 and 63 microstrains, respectively. In comparing the strains before and after strengthening, the concrete strain is reduced to some degree. DOI 10.14311/CEJ.2021.02.0029 401 Figure 13 shows the concrete strain on the bottom concrete surface under live load in condition 3 before and after strengthening. The strain was the largest in the middle slab. The maximum strain of slab 5 was 60 microstrains for the two trucks (Trucks A + B) before strengthening because of the damage and the weak interaction between the slabs. After strengthening, the maximum strain of slab 5 was 52 microstrains for the two trucks (Trucks A + B). The maximum strains of hollow slab 3 were 40 microstrains and 36 microstrains before and after strengthening, respectively. The strain was reduced to about 4 microstrains. The maximum strains of slab 1 were 25 microstrains and 20 microstrains before and after strengthening, respectively. Therefore, the concrete strain can be reduced after the PUC material is poured. For condition 4, the concrete strains on the bottom concrete surface under live load before and after strengthening are shown in Figure 13. The strain of beam 1 was the largest for the two trucks (Trucks A + B) because the two trucks approached the side of the bridge. The maximum strain of beam1 was 68 microstrains before strengthening, and the strain became 56 microstrains after strengthening. Before strengthening, the strains of beam 3 and beam 5 were 67 and 61 microstrains, respectively, and after strengthening, the strains were 56 and 52 microstrains, respectively. In comparing the strains before and after strengthening, the concrete strain is reduced to some degree.

-Concrete Strains and PUC Strains at Different Positions
The recorded strain values of gages Cn1, Pn2 and Pn1 for the five slabs in condition 1 and condition 2 are shown in Figure 14. In comparing the recorded concrete strains (gage Cn1) and PUC strains (gage Pn1 and gage Pn2), the concrete strain values are close to the PUC strain values on the same horizontal height. Therefore, PUC strain compatibility is well at the live load. In comparing the PUC strains (gage Pn1 and gage Pn2) with those recorded on the concrete (gage Cn1), the PUC strains are approximately the same as that of concrete strains. Strain compatibility DOI 10.14311/CEJ.2021.02.0029 403 leads to the higher PUC strain than the concrete strain because the bottom PUC is physically located below the bottom concrete in the beam section. Measured concrete strains of slab 1 and slab 9 in the flange plate position are listed in Table  3. Two trucks (Trucks A + B) were parked on the bridge in the condition 1 and condition 2. From this table, in comparing the strains for the given gages before and after pouring PUC, the compressive strains in the concrete were higher after the PUC was poured. In investigating this matter further, the axes were determined as shown in Table 4. Two trucks (Trucks A + B) were parked on the bridge in condition 1. From this table, as expected, the neutral axes of slab1 and slab5 migrated downwards by about 56mm and 55 mm, respectively, after the PUC was poured in condition1.

DEFLECTION MEASUREMENT
Displacement meters were installed on the bottom of the hollow slabs to measure the midspan deflection in the different conditions. The midspan deflection of all the slabs before strengthening in condition 1 is shown in Figure 15a. The deflection of slab 5 was the maximum under the symmetrical loads. The maximum deflections were 3.9 mm and 2.8 mm for two trucks (Trucks A + B) before and after strengthening, respectively. The deflection was reduced by 3mm compared with that before reinforcement. The deflection for slab 1 and slab 5 were 1.5 mm and 2.7 mm before strengthening, respectively. The deflection for slab 1 and slab 5 were 1.1 mm and 2.2 mm after strengthening, respectively. Figure 15b shows the midspan deflection of all the slabs after strengthening in condition 2. The maximum deflections of slab 1 were 4.6mm before strengthening. After strengthening, the deflection was 3.8 mm for the two trucks (Trucks A + B). As expected, a certain decrease in deflection after the application of PUC reinforcement could be observed.  Figure 16b show the midspan deflection of all the beams before and after strengthening in condition 3 and condition 4, respectively. In condition 3, the deflections of slab 5 were 3.0 mm and 2.5 mm for the two trucks (Trucks A + B) before and after strengthening, respectively. The deflection decreased by about 0.5 mm after strengthening. Before strengthening, the deflections of slab 1 and slab 3 were 1.7 mm and 2.5mm, respectively. After strengthening, the deflections of slab 1 and slab 3 were 1.7 mm and 2.2 mm, respectively. In condition 4, the deflections of slab 1 were 3.7 mm and 3.0 mm for the two trucks (Trucks A + B) before and after strengthening, respectively. The deflection decreased by about 0.7 mm after strengthening. Before strengthening, the deflections of slab 3 and slab 5 were 3.5 mm and 3.2 mm, respectively. After strengthening, the deflections of slab 1 and slab 3 were 2.2 mm and 2.8 mm, respectively.
These findings represent an initially good performance, although the long-term performance will be monitored throughout the next year.

CRACKS
In the load test before retrofitting, the width of the cracks increased with the test load, but no new cracks appeared. These cracks had affected the durability of the bridge and would have further reduced the performance of the bridge if no actions were taken. In observing the development of the cracks after reinforcement, a crack (2-2 measure point) in the side slab for slab 9 was selected to be the observation crack of the load test. The width of the crack on the bottom concrete is 0.15mm before strengthening. The depth from the observation location to the bottom concrete surface was 10cm. The width of the crack at the observation location is 0.08mm. Its width was unchanged for one truck (Truck A), 0.01mm for two trucks (Trucks A + B) in condition 2 and closed when the trucks were moved from the deck. Furthermore, periodical inspections were conducted to observe the development of the crack. The crack was found to be stable. considering four different parameters. Field testing could also provide information on live-load DFs for a given bridge type and geometry [21][22][23]. As the geometric size of all the hollow slabs is almost the same, the stiffness of every slab can be considered as having an equal value. The DFs can be determined from field. measurements using the following: where  i is the maximum static deflection in the i th slab, and . The deflection measurements shown in Figure 18 for Pia oFa Bridge in condition 2 were used to determine the live-load DFs according to Equation (7). From the horizontal hinged plate method [24], live-load DFs that can be compared with the measured DFs Figure 18 were provided. The slab 9 deflected by 2.5 mm for one truck (Truck A) in condition 2. The total deflection of all of the slabs was 8.5mm for a live-load DF of about 2.5/8.5 or 0.29. The live-load DF is 0.30 using the horizontal hinged plate method. The live-load DFs depend on slabs spacing, span, girder bending stiffness and girder torsional stiffness. The slab 9 deflected by 3.8 mm for the two trucks (Truck A + B) in condition 2. The total deflection of all of the girders was 14.3 mm for a live-load DF of about 3.8/14.3 or 0.27. The live-load DF is 0.28 using the horizontal hinged plate method. Thus, the horizontal hinged plate method is applicable for calculating the live-load DFs of T-beam bridges.

CONCLUSION
This paper uses an innovative method to strengthen a 25-year bridge with polyurethane cement composite material. The main construction process was described. Load tests were conducted before and after strengthening to evaluate the performance of this bridge and the effectiveness of the reinforcement method. Based on load test and analysis, the main results obtained in this study are listed below: The blending strength of the hollow slab bridge strengthened with PUC can be determined by force equilibrium, strain compatibility and controlled failure modes. The fabricated hollow slab bridge strengthened with PUC is governed by the PUC design strength.
The maximum carrying capacity of the main girder was increased by about 20% after PUC material was poured. The main girder deflection was reduced by about 28%. The width of the cracks was reduced in different degrees. Repeated load tests will be conducted to observe longterm performance. The PUC strain deformation complies with strain coordination under the static load. The neutral axis of hollow slab migrated downwards after the PUC was poured.
The transverse connection between slabs was proven by the hinged joints. The live-load DFs of the strengthened bridge calculated by experimental results were in good agreement with the modified eccentric-pressed method.