Torsional behaviour of RC beams strengthened by NSM GFRP bars

Near-surface mounted (NSM) fiber-reinforced polymer (FRP) reinforcement is an innovative, serviceable, and beneficial method for strengthening reinforced concrete (RC) structures. Previous studies have proven that the epoxy-resin bonded NSM FRP method effectively achieved shear and flexural strengthening. However, only a few researches have employed this technique for improving torsion but, no research has been conducted on the use of NSM with GFRP bars and epoxy adhesives. This study investigates the behavior of RC beams subjected to combined torsion when they have been enhanced with NSM GFRP U-shaped and closed stirrups with different overlap sites. Four RC beams with dimensions of (150×250×2000) mm (width X depth X length) were considered in the experimental tests. One RC beam is not strengthened and used as a control beam (CB), while three beams are enhanced with NSM GFRP bars in different shapes and overlap configurations. The findings of the tests in terms of twist angle_applied torque curves, the torque of initial cracking, the ultimate torsional carrying capacity, and the ultimate angle of twist were all recorded and evaluated. It has also been shown that a considerable improvement in the performance of RC beams in twisting can be achieved by using NSM GFRP stirrups. From the different NSM strengthening configurations used in the tests, results have demonstrated that beams strengthened by closed GFRP stirrups at the opposite overlap with 100 cm spacing outperformed other strengthened beams of about 181.5% compared to the CB. In contrast, the beam, strengthened by U-shaped GFRP stirrups with spacing at 200 cm, had the lowest torsional strength of about 48.1% relative to the CB.


Introduction
Along with flexure and shear resistances, torsional resistance must be taken into account for different concrete members.These members must maximise their torsional capacity due to deterioration, structural damage, earthquakes, increased loading, and eccentricity.Torsional influence may be considered in some situations, such as with spandrel beams in structures and curved box girders in bridges.Thus, these elements should be enhanced by either the externally bonded reinforcement (EBR) method, which means putting the reinforcement materials on the concrete's surface, or the nearsurface mounted (NSM) method, in which pre-cut grooves in the concrete's cover are used to mount the reinforcing materials [1].The NSM strengthening technique is a suitable substitute for the EBR method [2].The most advantageous feature of the NSM method is the increased confinement supplied by the bonding agent and the concrete around it [3].Over the last few decades, research into the composite material of fiber-reinforced polymer (FRP) as a superior material for increasing strength [1] has been extensive.As a result of its superior strength properties, lightweight, corrosion-resistant nature, low ability to transfer heat, and flexibility, this composite material is ideal for use in construction applications.[1,3].Many studies have improved the flexural and shear strengths of RC beams using the NSM FRP enhancing method [4,5].Additionally, a few researchers have used the NSM technique for torsional improvement of beams.Some of them used this method with traditional steel bars [6,7].From these researches, Nasih et al. [6] compared the torque of initial cracking, the ultimate torsional moment, the ultimate angle of twist, and the torque intervals of a RC beam strengthened with NSM steel bars and CFRP sheet subjected to bending and twisting forces simultaneously.Experimental results have shown that CFRP sheets and NSM steel bars significantly improve RC beam torsional performance.However, the CFRP wrapping enhanced the torsional carrying capacity better than the NSM steel bar.Also, Nashi et al. [7] experimentally examined RC beams with NSM steel bars set up in different situations and applied them to flexure and torsion.The study has concluded that using NSM steel bars significantly improved the torsional capacity of RC beams.Results have also revealed that 90° NSM beams significantly outperform 45° NSM beams in various configurations.However, utilizing the FRP NSM technique for torsion enhancement has only been examined in a few studies [8][9][10][11].For instance, Al-Bayati et al. [8] conducted experimental tests on five concrete beams; four of the beams were enhanced in torsion with the NSM method by using CFRP laminates; besides that, one of the specimens served as a CB.Full wrapping and u-wrapping were tested as two alternative torsional strengthening configurations.Test results have demonstrated the impact of different enhancing configurations on the NSM torsional improvement of reinforced concrete beams.Strips of CFRP NSM laminate with three and four faces increased the beams' torsion capacities by 21.6% and 30.7% when compared to the CB, respectively.The enhanced beams have shown higher crack torques than reference beams.Further, the U-shaped reinforcement configuration outperformed the CB by a respectable margin.It has also been found that at the same torque, the laminate strips' corners developed critical torsional cracks that caused beam failure for beams enhanced on all faces.Al-Bayati et al. [9] also presented experimental tests on ten RC beam specimens, two of which were controls, while eight beams were laminated using NSM with CFRP on all four sides.Four of the NSM-strengthened RC beams utilised an epoxy adhesive as a bonding agent, while in the other four beams, the epoxy was replaced by a modified cement-based adhesive.The impact of groove spacing (the distance between each groove) was taken into account, so two different groove spacings were examined for each adhesive type.According to the findings of the experiments, it was found that the number of grooves could potentially limit the amount of strengthening that could be accomplished by employing the NSM method.The findings also indicate the ultimate torsion capacity of CFRP NSM laminates was enhanced by 28.2% and 35.9% when using epoxy and by 23.4% and 26.5% when using cement-based adhesive for 0.75D and 0.375D spacings, where D denotes the depth of the beam, respectively.Besides that, it was discovered that increasing the torsional strength by reducing the distance between the grooves and the changes in the rigidity of the beam coincided with a new crack's initiation or the yielding of an individual stirrup.In addition, Al-Bayati et al. [10] employed the same torsion-strengthening approach as Al-Bayati et al. [9], but the CFRP rope was wrapped across the cross-section of beams.The experimental programme includes using cement-based glue to strengthen four beams: two control beams and two strengthened with CFRP rope and epoxy to strengthen torsion.Results have shown that the ultimate torque and cracking rates had increased by 9.2% and 18.7%, respectively, relative to the CB, and the enhanced beams' ductility was higher than that of the CB.In addition, at the same torque level, reinforced beams have shown less strain, and torsional cracks at failure occurred in the groove where the concrete cover with CFRP rope delaminated.By Mashael et al. [11], a numerical study was presented to investigate the influence of NSM steel and CFRP bars on improving the strength of RC beams under pure torsional loading.The RC beams were improved using different NSM bar spacings and configurations.Results have shown that using NSM rebar increased the concrete beams' torsional strength, twist angle capability, ductility, and capacity for absorbing the energy of the concrete beams by redistributing internal stresses.It has also been shown that CFRP bars gave NSM beams a higher enhancement ratio than NSM steel rebar.Also, the enhanced beams have demonstrated improvements in the ductility of the twisted beams.From the above-mentioned review, it can be seen that a number of studies have enhanced RC beams' flexural and shear strengths using the NSM FRP enhancing method [4,5].In addition, several investigations [6][7][8][9][10][11] have concentrated mainly on torsion enhancement, On the other hand, there hasn't been any research done into using NSM GFPR with epoxy adhesives for torsional strengthening.Because of this and to fill this gap in the research, the purpose of this work is to investigate the behaviour and quantify the characteristics of RC beams strengthened by NSM GFRP bars in various configurations and shapes when they are subjected to the effect of a torque moment.

Geometrical properties and reinforcement details
The total length of the four RC beams is 200 cm, and their cross-sectional dimensions are 15 cm wide and 25 cm deep, with a clear span of 180 cm and an effective torsion span of 70 cm.They were prepared and cast using ready-mixed concrete; one of these RC beams was unstrengthened so that it could serve as a CB.The remaining beams were strengthened by NSM GFRP bars in different shapes and configurations, as shown in table 1 and figure 1.

Table 1. Characteristics of the near-surface mounted technique.
The RC beam specimens were designed in accordance with American Concrete Institute standards ACI 318-19 Code 19 [12] to intentionally fail by torsional failure at the central portion of the beams under a pure torsional moment.According to the design, all RC beam specimens are reinforced with three 12 mm diameter steel bars as tensile reinforcement at the bottom of the section and two 8 mm steel bars as compressive reinforcement at the top, with a 25 mm concrete cover shown in figures 1 and 2. Only one-steel stirrups with 8 mm bars were provided as torque transverse reinforcement.All RC beam specimens were subjected to a two-point loading system using a specially designed loading system, as will be described later.Further, the central portion length of the beam was 70 cm, which was its effective torsion span, which was suggested in this study to be adequate to permit the creation of at least one diagonal crack at a 45-degree angle with its longitudinal axes.Figure 1 illustrates the cross-sectional dimensions and the steel reinforcement details of the RC beams considered in the experimental tests.The minimum spacing of the transverse reinforcement was also purposefully increased beyond what it should have been and set to 350 mm, so the torsional failure can be observed without restriction from the stirrup [8].

No
Beam designation Stirrup shape Spacing(mm) Note

Material properties of the RC beam specimens
As shown in figure 2, four rectangular RC were cast using ready-mixed concrete with mixing proportions of 1:1.5:2.5.This concrete mixture comprised local Portland cement, fine aggregates, and crushed aggregates.According to the mix design, the corresponding equivalent weights of the cement, coarse aggregate, and fine aggregate were 475 kg/m3, 720 kg/m3, and 1190 kg/m3, respectively.In addition, 132 liters of water and 3 liters of Master Glenium 54, which is added as a high-performance concrete superplasticizer, were used.The average compressive strength of the concrete cubes at 28 days was about 41 MPa according to British standards [BS.EN 12390 (2,7:2000, 3:2002)].The RC beam specimens were cast using 20-mm-thick wooden moulds, with the reinforcement being placed at an adequate distance from the clear cover on all four sides.After lubricating the inner surfaces of the wooden moulds, the ready-mixed concrete was poured using a truck-mixed pump then the beam specimens were demolded after 48 hours and water-cured for 28 days.

Strengthening procedure using NSM GFRP bars
After 28 days of curing, the strengthening GFRP bars were mounted by cutting square grooves into the surface of the beam section through the concrete covers, as illustrated in figure 3. The grooves were cut around the cross-section of four beam specimens using a saw-cut machine.The dimensions of the square grooves were 15 mm in width by 15 mm in depth, as suggested by ACI 440.2R-17 [1], which must be greater than or equal to one and a half times the diameter of the NSM GFRP bars used.However, the grooves in the other two RC beam specimens were cut along both sides and the bottom of the beam section.A hammer drill was used to trim the groove's edges, and the lower surface of the grooves was roughened with a chisel.Any remaining concrete lugs were removed by a wire brush, and a high-pressure air jet was then used to clean the grooves.At this time, the 6 mm GFRP stirrups have been prepared, with an area of 31.67 mm2, an ultimate tensile load of about 28 kN, a guaranteed tensile strength of 896 MPa, a modulus of elasticity of 46 GPa, a weight of 77.4 g/m, and a transverse Planning and marking the grooves.

NSM GFRP bar installation
The grooves cutting and drilling filling with epoxy and NSM bar installation.
Figure 3. NSM strengthening procedure of the RC beams using GFRP bars.Two RC beams were strengthened using U-shaped GFRP stirrups, which were mounted at the bottom of the section and spaced at 100 mm for the first beam and 200 mm for the second.The last specimen was strengthened with two C-shaped GFRP stirrups mounted at the left and right sides of the beam section to obtain closed stirrups in a reverse position, as shown in figures 3 and 4.After that, an epoxy-based adhesive (Sikadur-30LP) was used to fill in the grooves in two stages around the GFRP bars.Table 2 displays the mechanical characteristics of the epoxy resin according to the manufacturer's datasheet.Half of the grooves were filled with epoxy glue, and then GFRP bars were lightly pressed into the grooves to ensure that the adhesive flowed around the GFRP bars.Then the remaining half was covered with glue all the way up to the top, making sure there were no gaps.The surface of the adhesive was levelled, and finally, all the beams were kept for a minimum of 14 days to ensure the epoxy would reach its full strength.

Configurations and shapes of the NSM GFRP stirrups
To determine the best strengthening configurations of the NSM GFRP stirrups in terms of economic, technical, and practical standpoints two types of stirrups were utilized in three beams of the present study, as illustrated in figure 5.A U-shaped GFRP stirrup was used to strengthen the two RC specimen beams, as shown in figure 5-a.While the last specimen was reinforced with a couple of Cshaped GFRP stirrups, which were mounted from the left and right sides of the beam section to obtain closed stirrups in reverse positions as shown in figure 5-b, the strengthened RC beams' grooves are oriented at right angles to the beam's longitudinal axis.Finally, all beam specimens were coloured with white paint to facilitate the detection of cracks.

Test setup and procedure
Figure 6 reveals the test setup and loading device (rig) used in the present study.The beam specimens were loaded using the loading frame accessible in the laboratories of the civil engineering department at Al-Qadisiya University.The load was applied by using a hydraulic jack.To generate a torsional moment, the applied main load was equally divided into two loads located 500 mm from the beam longitudinal axis and 650 mm from the beam ends using a clamping loading frame as shown in figure 7. The clamping frames have a spreader steel beam made of steel I-section with dimensions of 200 mm × 80 mm × 8 mm (height × width × thickness).The spreader steel beam rests on two lever arms made of two steel channels measuring 100 mm × 50 mm × 8 mm (height × width × thickness.As shown in figure 8, each arm was connected to the RC beam through a steel frame of four steel channels with dimensions of 100 mm×50 mm×8mm (height×width×thickness) surrounding the RC beam specimens.The lever arms provided 500 mm of eccentricity from the beam's longitudinal axis.
A hydraulic jack was used to apply a monolithic load, which was handled by a compression load cell with a 2000 kN capability and a 5 kN increment as shown in figure 8.At each load increment, digital Grooves @ 200 mm spacing dial gauges recorded the accumulated displacement, and crack meter to observe the cracks.To avoid compression and allow free rotation around its longitudinal axis, the beam's ends were supported on rollers.The beam specimens were positioned such that the distance between the supports is about 1800 mm, with a projection of 100mm outside the two supports.While, according to the designed loading mechanism described above, a 700 mm section over the specimen's mid-length was subjected to combined torsion and bending.Further, the twist angle over the torsion span was monitored and recorded using two dial gauges positioned beneath the beams to measure the downward displacement at the edge and at the center of the beam section at the connection between the RC beams and the steel lever arm.Another dial gauge was placed beneath the center of the beams to determine the central displacement, shown in figure 8.

Results and discussion
In this section, the experimental test results are presented and discussed in detail to gain an understanding of the performance and failure mechanisms of torsional-strengthened RC beams using NSM GFRP stirrups.The proposed design of the models was based on the assumption that the effect of loads and strengthening on the torsional behaviour of the beams is obvious and efficient.This is what this study is trying to find out.However, their influence on flexural behaviour was very minimal, so it was not further discussed in this paper.

The effect of NSM GFRP stirrups configuration
3.1.1.Crack pattern and failure mode.The crack pattern and failure modes of control and NSMenhanced beams with different configurations of NSM GFRP stirrups are depicted in figures 9, 10, and 11.It can be noted from these figures that the first hairline crack of the CB initiated at a torsional moment of 2.25 kN.m.However, the crack became more visible after the subsequent loading steps and propagated spirally around the RC beam section, resulting in the appearance of small inclined cracks.The diagonal torsion failure occurred at the maximum torque of 6.75 kN m and the twisting angle of 3.86, as shown in table 3. On the other hand, the first visible cracks appeared in the beam specimen GØ6@100-U-90˚ at the side face of the beam beside the twisting arm at the torque of 6 kN.m., with a 166.7% increase in the initial crack, compared to the CB.Also, a few diagonal cracks occurred during the loading and began to expand as the load increased.With increasing the load, additional diagonal cracks occurred on each side of the beams with different spacing and propagated towards the upper part of the beam However, the cracks extended from both sides of the beam section and expanded towards the upper face of the beam, causing torsion failure at the top of the section due to the lack of torsion reinforcement on this side of the section.
On the other hand, the first visible cracks of the beam GØ6@100-90˚ appeared at the torque of 8 kN.m. at the contact area with the lever arm in the opposite direction of the torsional moment, with a 255.6% increase of the initial crack relative to the CB, as shown in figure 11.Further, the effect of the torsional moment on the specimen increased when the applied loads increased with the appearance of many spiral cracks generated in the diagonal direction of the section.These cracks were repeated at various distances between the grooves and expanded and spread rapidly; the cracks on the lateral surfaces were more numerous when compared to those on the upper and lower surfaces.In expanded areas, cracks tend to spread at a faster rate than in those that are more constrained.Moreover, it was observed that one of these cracks spread faster than other cracks and expanded diagonally with increasing width until failure occurred at the ultimate torque of 19 KN.m.     3 and figure 12 show how the configuration of NSM GFRP stirrups affects the strengthened RC beam's torsional carrying capacity.The beam G6@100-90 strengthened by a closed NSM GFRP stirrup with opposite overlap at 100 mm spacing, has the highest ultimate torsional moment capacity of 19 kN.m.While the beam G6@100-U-90, strengthened by U-shaped NSM GFRP stirrups at 100mm spacing, has the lowest ultimate torsional moment capacity of 12.2 KN.m.However, other strengthened beams have torsional moments between these two extremes.Further, the ultimate torsional moment of the G6@100-90˚ beam is about 181.5% higher than that of the CB, while, in contrast, the ultimate torsional moment of the G6@100-U-90˚ beam is about 81.5% higher than that of the CB and 64.5% lower than that of the G6@100-90˚ beam.The top side's stiffness was lower than the other sides due to the lack of NSM GFRP stirrups, which meant the concrete alone was resisting torsional moments.However, the other three sides of the beam with GRRP stirrups could withstand the torsional moment's stresses.
The choice of U-shaped stirrups has both practical and economic importance, as their use represents an exclusive solution for some parts of the building that are difficult to reach to their entire perimeter of their beams, such as the girders in bridges, or where it is difficult to damage the ceiling of a floor in a commercial building.

The effect of NSM GFRP stirrups spacing
3.2.1.Crack pattern and failure mode.Figures 10 and 13 show the crack pattern and failure modes of the strengthened beams with varying NSM GFRP bar spacing; the first visible cracks appeared at the torsion zone of the beam specimen G6@200-U-90˚ at the cracking torque of 4.5 kN.m., with a 100% increase compared to that of the CB.Afterward, cracks spread and propagated as the load increased and finally caused diagonal torsion failure at the maximum torsional carrying capacity of 10 kN.m with a corresponding ultimate twist angle of 6.65 deg./m, representing a 48.1% and 72.4% increase of that of the CB, respectively.On the other hand, the first visible cracks appeared at a torque of 6 kN.m. at the torsion zone of the G6@100-U-90˚ beam specimen, with a 166.7% increase in the initial crack load compared to that of the CB.Moreover, the number of cracks in the strengthened beam G6@100-U-90˚ was significantly higher than that of the G6@200-U-90˚ beam.As the load increased, more cracks appeared, propagated, and eventually caused a diagonal torsion failure at the torsional carrying capacity of 12.25 kN.m., as shown in table 3.    14.
In comparison to the CB and the G6@200-U-90 beam strengthened by GFRP stirrups at 200 mm spacing, the ultimate torque capacity of the G6@100-U-90 beam is greater and it has more ductile behaviour, as shown in the figure below, which reported an increase in both the ultimate torque and the ultimate twist angle of about 81.5% and 92.1%, respectively, relative to the CB.It can be seen that the G6@200-U-90˚ beam displayed an increase in ultimate torque and ultimate twist angle of approximately 48.1% and 72.4%, respectively, with a corresponding cracking torque of approximately 100% relative to the CB.The above experiment confirmed that a beam's torsional resistance in the NSM GFRP stirrups enhanced RC beams increases with smaller GFRP stirrup spacing.

Conclusion
This research presented experimental results on the performance of RC beams under combined torsion when they were strengthened with NSM GFRP stirrups.Four RC beams were considered in the experimental tests; one RC beam was not strengthened, and three specimens were enhanced with NSM GFRP stirrups at various spacings and shapes of the stirrups.The test observations in terms of initial cracking torque, maximum torque, angle of maximum twist, and twist angle-the torque curves were all evaluated and discussed, and the following conclusions were extracted from these observations: • A higher torsional resistance was observed in all NSM GFRP stirrup strengthened RC beams in comparison to the CB.The configuration and spacing of the FRP's strengthening elements can have a significant impact on the FRP's torsional carrying capacity, which can increase or decrease accordingly.The best increase ratio for torsional carrying capacity was 181.5% relative to CB; it was achieved with closed NSM GFRP stirrups with a spacing of 100 mm.• It was possible to achieve an increase in the torsional carrying capacity of the NSMstrengthened beams by decreasing the distance between NSM GFRP stirrups.Results have shown that up to a 22.5% increase in torsional carrying capacity was achieved in the beam, which has Ø6mm NSM GFRP U-shaped stirrups with a 90˚ inclination angle and 200 mm spacing, when the distance between its stirrups was reduced from 200 mm to 100 mm, respectively.• Using U-shaped NSM GFRP stirrups is considered a beneficial and economical solution for strengthening the torsional moment for many structures, such as building beams and bridge
According to tensile tests [ASTM A615/A615M-09b] [12], the yield and ultimate strengths of the reinforcing steel bars used are 449 MPa, 480 MPa, and 547 MPa for bar diameters of 6 mm, 8 mm, and 12 mm, respectively.

Figure 2 .
Figure 2. Casting of the RC beam specimens.
150 MPa; these mechanical specifications were taken from the manufacturer's datasheet.The geometric shapes of the NSM stirrups used in the present study are shown in figure5.

Figure 6 .
Figure 6.Test setup and loading frame.

Figure 9 .
Figure 9. Crack pattern of the CB.

Figure 12 .
Figure 12.Torque versus twist angle of tested beam with closed and open NSM GFRP stirrups.

Figure 14 .
Figure 14.Torque versus twist angle of GFRP at various spacing.
in structures where it is challenging to reach the entire perimeter of their beams.Experimental results have shown that the ultimate torque capacity of the beam, which has Ø6mm NSM GFRP U-shaped stirrups with a 90˚ inclination angle and 200 mm spacing, is 48.1% greater relative to the C.B.

Table 2 .
Properties of the epoxy resin used in the current study.

Table 3 .
Experimental results of the tested beams.
Ultimate torque and torque-twist angle relationships.The relationships between torque and twist angle for the control and strengthened RC beams at various NSM stirrup spacings are shown in figure