Study of the behavior of reactive powder concrete RC deep beams by strengthening shear using near-surface mounted CFRP bars

: The applications of modern materials especially reactive powder concrete for improving concrete structures have been signi ﬁ cantly growing in recent years. Great conduct properties creep, shrinkage, little perme-ability, ultra-high strength, and expanded safety against corrosion are the important features of Reactive Powder Concrete. In addition, the use of the Near-Surface Mounted technique in recent decades has helped strengthen and repair shears-reinforced concrete (RC) for deep beams using carbon ﬁ ber replacement reinforcing bars. The para-meters studied in the present research investigated the impact of the maximum load, de ﬂ ection, stress – strain curve of concrete, ﬁ rst shear crack, crack pattern, and crack width. Considering the aforesaid cause and objective, one specimen of Reactive Powder Concrete RC deep beams has a rectangular cross-section of 150 mm in width, 500 mm in depth, and a total length of 1.2 m. One control specimen was tested for comparison. In addition, 12 control specimens (cylinder and cubes) are used for experimental investigation on the mechanical properties of normal and Reactive Powder Concrete deep beams. Following the specimens ’ processing, they were subjected to one concentrated load pressure test through a hydraulic jack. Moreover, six core drill specimens were taken from those deep beams to obtain the real mechanical properties of those beams, including maximum stress, modulus of elasticity, density, stress – strain curve, and Poisson ’ s ratio, after subjecting them to the pressure machine. Depending on the results, the ultimate strength, de ﬂ ection, and ﬁ rst shear crack capacity for the specimens (RPCDB1P-4NSM & RPCDB1P-8NSM) have increased by (21, 25), (47, 27), and (133, 150)%, respectively, compared with (CDB1PC20). Moreover, the specimens above have reduced the ﬁ rst shear crack width by (9, 33.25)% respectively compared with (CDB1PC20) at 65% of the ultimate shear load.


Introduction
Shear is one of the most dangerous types of failure in reinforced concrete (RC) structures because it happens suddenly and without warning, like the shear failure that occurs in deep beams. Recent studies have found that current RC beams suffer from shortcomings in shear and require strengthening. Deficiencies can occur due to several factors such as insufficient shear reinforcement resulting from design or construction errors or use of outdated codes, reduction in the steel area due to corrosion, and increase in demand for service load (Khalifa et al.) [1]. The use of FRP materials has been rapidly expanding throughout the world in order to achieve the modern reinforcement and repair of existing concrete structures (Katsumata et al., Grace and Sayed, Rostasy et al., Al-Mahmoud et al.) [2][3][4][5]. An effective technique for Fiber Reinforced Polymer (FRP) with Near Surface Mounted (NSM) technology has been used to improve the bending and shear behavior of existing or newly reinforced structural elements (Hatem et al.) [6]. A widely accepted solution is the use of fiber-reinforced polymers (FRP) to strengthen RC structural elements in bending and shearing areas due to this technology's advantages over other traditional strengthening methods (Saadah et al.) [7]. Moreover, the use of highperformance concrete, such as reactive powder concrete (RPC), has been observed, which has high-performance properties that will limit the failure of shear to contain a few steel fibers in its mixture.
In addition, with the advent of a new technique such as NSM, which will require investigation into its potential applications, for example, the use of NSM CFRP bars with other types of concrete such as ultra-high performance concrete RPC as a strengthening or repair of the shear for deep beams. Also, until recently, no researchers have conducted the strengthening or repairing of RPC deep beams using NSM CFRP rods.
Today, it is considered good to adopt the use of promising modern materials and technologies and compare their performance with the maintenance methods used in the past, medium to a long period. This will help close the research gap for applying the NSM technique in controlling bending and shear failure in the long term.

Experimental program
The testing matrix contained two simply supported RC deep beam specimens with and without reactive powder concrete (RPC), with a total span (L) of 1,200 mm, overall depth (H) of 500 mm, width of 150 mm, and shear span to effective depth ratio (a/d) equal to 0.77. All beams were tested under a three-point loading configuration as shown in Figure 1. Three Ø16-mm deformed bars were provided as longitudinal tension reinforcement for the bottom zone, two Ø12-mm as longitudinal tension reinforcement for the top zone, and Ø12-mm @200 mm c/c for stirrups. The concrete cover was 40 mm.

Material properties
The specifications of materials used in this research and their suitability for making the RPC are explained here.

Cement
The type of cement is ordinary Portland cement of the KAR Group factory in Al-Najaf. It conforms to Iraqi Standard Specification I.Q.S. No. 5 (IQS) [18]. The physical and chemical tests of the cement were conducted at Kufa University and the laboratories of the engineering consulting bureau.

Fine aggregate
Natural sand locally available from the Al-Najaf region was used for concrete mixes in this research. In reactive powder concrete, all large size of particles must be eliminated, and very fine sand aggregate is used in order to produce a very dense form. The fine aggregate was passed from sieve No. 40 (450 μm) maximum size and remained on sieve No. 100 (150 μm) (single grading) (Richard and Cheyrezy) [19]. The chemical and physical tests of fine aggregate were conducted at Kufa University and laboratories of the engineering consulting bureau.

Silica fume
Silica fume is a highly reactive material that can be used in comparatively small quantities to improve the characteristics of concrete ACI 234R (ACI) [20]. Also, silica fume is described as "very fine non-crystalline silica produced in electric arc furnaces for silicon or alloys containing silicon." It is usually a gray-colored powder, slightly like Portland cement or some fly ashes. One of the most substantial components of reactive powder concrete is using pozzolanic material. It is not only used to give extra binders but also used as a filler. The present manuscript has used densified silica fume (Sika ® Fume S 92 D). The micro silica used in this research was based on the physical and chemical limitations of ASTM C1240 (ASTM) [21].

Superplasticizer
GLENIUM® 54 is a third-generation superplasticizer for concrete and mortar. It achieves the limitations for superplasticizers regarding ASTM C494/C494M (ASTM) [22]: types A and F. It was used as a water-reducing admixture with a nominal dosage of 2% of cement weight. This mixture gives high water shorthand (causing high strength and density), excellent workability without bleeding or segregation, and very high early and high ultimate strength concrete with minimal voids and therefore optimum density and improved water impermeability and surface finish.

Steel fibers
RPC contains small steel fibers instead of coarse aggregates. Steel fibers are used in reactive powder concrete to enhance some properties and improve ductility. The steel fiber has a length of almost (13 mm) and a diameter of about (0.2 mm) based on ASTM A820/A820M (ASTM) [23]. The steel fibers used in this research were straight steel fibers manufactured by the Bekaert Corporation with a volumetric ratio of 0.7%. The properties of steel fibers obtained from China are given in Table 1.

Steel reinforcing bars
Shear design is a very required factor when starting the design of a deep beam, especially the ACI code 11.8. A deep beam has a clear span of a depth ratio of less than five for divided overloading requirements [24]. Reinforcing steel used as longitudinal bars and web reinforcement in RC deep beams met the ASTM A615, 1990 [25]. Two sizes of bars (Ø 16 mm and Ø12) were used as longitudinal and stirrup reinforcement to reinforce all deep beams tested the specimen of the Ukrainian reinforcement steel bars was tested in tension and shear to determine its mechanical properties. The results are shown in Table 2. The tensile tests of the specimens were executed by the testing machine available at Kufa University and laboratories of the engineering consulting bureau.

Specimens' preparation
For this study, 20 mm thick four-plywood formworks were used to cast three beams in each casting process. Next, all formworks were coated with formwork oil at the inner surface of the formwork so that the detaching process of the beam afterward would be easier. A mixer was used to mix the concrete. The rebar cages were placed inside the forms and held in place by a 2.5 cm plastic chair to ensure that the proper cover was maintained. Before mixing, all the quantities were weighed by a sensitive electronic to ensure accurate measurement of quantities and packed in a clean metals container prior to starting the rotation of the mixer materials. First, the mixer was operated, and then, the fine aggregate, cement, silica fume, and water were added while the mixing was running some of the mixing water and Superplasticizer were added. finally, the fiber steels were added to all mixed concrete. The concrete mix time was about 30 s after all ingredients were put in the mixer. Twelve cylinders (100 × 200) for each series were cast to specify the concrete properties. The concrete was then poured into six layers to make sure the compaction process was sufficient and not overly done. Insufficient compaction could cause incompact concrete and the form of a honeycomb in the specimens. All things mentioned above are shown in Figure 2.

Concrete mix design
Before starting the concrete mix, a slump test was taken for the concrete mix. Many trial mixes were originally used in this scientific article before reaching the required performance of the RPC. The dominant mix proportioning (by weight) for all beams and their reference specimens (cubes, cylinders, and drilling cores) was 1∶1∶0.08 (cement: fine sand: silica fume) with a water-to-cement ratio of 0.25. The superplasticizer was used with a proportion of 2% by the cement weight, and steel fiber with a volume fraction of 0.7% was used in all mixes ( Table 3).

Strengthening procedure
The preparation of the strengthened beams involves cutting the grooves on the sides of the beams, and application of the NSM CFRP rods. After the beams had been cured properly, the concrete was cut using a special machine to make grooves, and all the grooves had square cross-sections, with a depth and width of 10 mm as shown in Figure 3. Then, the grooves were cleaned with jet water to remove the concrete powder produced during the cutting process as shown in Figure 4. After that, the epoxy paste was prepared by mixing resin and hardener in 2:1 proportion by volume with a power mixer as shown in Figure 5. The groove was filled halfway with the paste. Next stage, the rod was then placed in the groove and pressed lightly. The groove was then filled with more paste to the surface level as shown in Figure 6. The specimens remained in the laboratory environment for two weeks before being tested. After completing the CFRP rods installation within two days, all apparent concrete surface specimens were painted white to easily detect the crack patterns (growth of crack) before testing as shown in Figure 7.

Testing procedure
After the beams were removed from the water at the age of 28 days. Also, the beam specimens were cleaned and painted in order to show crack diffusion. All beam specimens were tested under a static one-point load test to study their behavior by a universal testing machine as shown in Figure 8. The test was stopped at every 10 kN to register the evolution of cracks, strains, and deflections along the beam, and the specimens were positioned in the universal testing machine for midpoint loading. The deflection dial   [29] and which it calculates from stress-strain curve by slop = Rise/Run. *Ec(Equation (1)) modulus of elasticity of ordinary concrete based on (ACI-318M-14) [30] and which it computes by equation (1), = EC 4,700 fc′ . *Ec(Equation (2)) modulus of elasticity of ordinary concrete based on (ACI 363R-92) [31] and which it computes by equation (2), = + EC f c 3,320 ′ 6,900. *ᶹ Poisson's ratio based on (ASTM C469-02) [29]. CDB* -Concrete deep beam. RPCDB* -Reactive powder concrete deep beam. Note: (1) The values of (f′ c , f′ cu , f′ dc , and ffu) were calculated for an average of three specimens. (2) We can note that the above two equations are for calculating (Ec) for ordinary concrete only, and there is no equation for calculating (Ec) for reactive powder concrete until now (3). We can observe that the value of the drilling core compressive (f′ dc ) is a real and very different value compared with the two values of f′ c , f′ cu . For the CDB specimen, where f′ dc differs with (f′ c , f′ cu ) by (30, 55%), respectively. For the RPCDB specimen, where f′ dc differs with (f′ c , f′ cu ) by (12%, 20%), respectively. gauge was placed at its marked position to touch the bottom face of the beam center. When the beam showed deterioration with increasing deformation, the ultimate load was registered, and the load was removed to take     photographs of the final crack pattern. In addition, after the failure, (6) core drill specimens were taken from undamaged areas of deep beams as shown in Figure 9, which will be divided into three specimens from the deep beam with ordinary concrete and three specimens from the deep beam with reactive powder concrete, which will be   subjected to examination by the compression machine using a digital caliper and mini digital thickness gauge which have been manufactured manually and using them with the compression machine as shown in Figure 10 to get the real mechanical properties especially stress-strain curve.

Results and discussion
The obtained experimental results are summarized in Table 4, including the first cracking load, ultimate load, ultimate shear strength, and crack patterns-failure mode. In addition, comparisons of load-deflection behavior, shear crack width, and stress-strain curve of these beams are shown in Figures 11-18. The specimen control deep beam one-point load compressive strength with 20 MPa (CDB1PC20) failed to shear. The first flexural crack was visible at approximately 16% of the 80 kN load, followed by another crack at the 120 kN load. With increasing load, the flexural cracks spread more but stopped at 280 kN loading. On the other hand, the first shear crack was observed at 120 kN loading from the inner edge of the support. When the load increased to 260 kN, another primary shear crack appeared in the center of the support. By increasing the load, bending and shear cracks spread and mature. Upon reaching the failure load, the main diagonal slit occurred at an angle of 53°as shown in Figure 19. The specimen CDB1PC20 failed when the load reaches 480 kN or the shear load is 240 kN.
From observation, the beam suddenly failed, and the failure mode was due to the failure of the shear pressure. The specimen reactive powder concrete deep beam onepoint load using four CFRP bars (RPCDB1P-4NSM) failed in the flexural region and did not fail to shear due to the use of the near-surface mounted strengthening technique. The first flexural crack appeared in the middle of the beam when the load was about 100 kN. Other small cracks appeared widely between the two supports starting from below toward the bearing points as the load gradually began to increase. Other oblique shear cracks (in both shear periods on the left and right sides of the beam) initially formed at different points at a load of 280 kN and spread towards the loading points. With overloading, these cracks quickly spread and widened. The beam failed to flexural, and the failure occurred in the middle of the beam, as shown in Figure 20. RPCDB1P-4NSM failed at loading up to 580 kN or a shear load of 290 kN. From observation, the packet failed after a period of loading and the failure mode was due to the failure of the flexural compression. Also, from the significant decrease in the shear capacity of the CDB compared to that of the RPCDB1P- 4NSM beam, the presence of the polymer rods used with the specimen with reactive concrete powder reduces the shear ability. The shear capacity of the RPCDB specimen decreased by 20% compared to the shear capacity of the CDB specimen.
For the specimen reactive powder concrete deep beam one-point load using four CFRP bars (RPCDB1P-8NSM) failed in the flexural region and did not fail to shear due to the use of the Near-Surface Mounted strengthening technique. The first flexural crack appeared in the middle of the beam when the load was about 100 kN. Other small flexural cracks appeared for a period of loads ranging (180-260) kN. When the load began to gradually increase, shear cracks were formed, but they remained at a short length at the abutment and did not extend to the loading points due to the strengthening NSMCFRP rods technique close to the surface, where the first shear crack was formed at the load of 300 kN and stopped until the gestation period of 340 kN and equally from both sides of the threshold at the supports. The beam failed in flexural, and the failure occurred in the middle of the beam, as shown in Figure 21.
The deflection profile for beams CDB1PC20 and RPCDB1P-4NSM is clearly shown in Figure 11. The deflection was recorded at every load increment until failure. Based on the observation, a deflection value of 9.73 mm was recorded for beam RPCDB1P-4NSM at the highest peak but only 6.59 mm for beam CDB1PC20. However, beyond the failure load, the post-peak behavior for RPCDB1P-4NSM and CDB1PC20 showed similar trends, that is, a reduction in load but with a gradual increase in deflection. The post-peak increment in deflection was more gradual for beam RPCDB1P-4NSM compared to CDB1PC20. Prior to total failure, a maximum deflection of 11.40 mm for RPCDB1P-4NSM and 8.45 mm for CDB1PC20 was recorded.
In addition, the deflection profile for beams CDB1PC20 and RPCDB1P-8NSM is clearly shown in Figure 11. The deflection was recorded at every load increment until failure. Based on the observation, a deflection value of 8.40 mm was recorded for beam RPCDB1P-8NSM at the highest peak but only 6.59 mm for beam CDB1PC20. However, after the failure load, the subsequent peak behavior of (RPCDB1P-8NSM) and (CDB1PC20) showed similar trends,   that is, a decrease in load but with a gradual increase in deviation. The post-peak increment in deflection was more gradual for beam RPCDB1P-8NSM compared to CDB1PC20.
Prior to total failure, a maximum deflection of 10.70 mm for RPCDB1P-4NSM and 8.45 mm for CDB1PC20 was recorded.
On the other hand, shear crack width was measured and recorded during every load increment. For control beam CDB1PC20, the first shear crack appeared at 120 kN, and the crack width value was measured at 0.05 mm. The crack width continued to increase with every load increment, and at 65% of the ultimate shear load (312 kN), the value for the diagonal crack had increased to 2 mm. Moreover, the first shear crack of the RPCDB1P-4NSM beam occurred at a load of 280 kN, and the crack width value was measured at 0.15 mm as a load was increased; the shear diagonal crack at 1.8 mm was recorded at the load of 377 kN. Figure 13 shows the shear crack width development for specimens CDB1PC20 and RPCDB1P-4NSM loading.
In addition, shear crack width was measured and recorded during every load increment. For control beam CDB1PC20, the first shear crack appeared at 120 kN, and the crack width value was measured at 0.05 mm. The crack width continued to increase with every load increment, and at 65% of the ultimate shear load (312 kN), the value for the diagonal crack had increased to 2 mm. Moreover, the first shear crack of the RPCDB1P-8NSM beam occurred at a load of 300 kN, and the crack width value was measured at 0.15 mm; as the load was increased, the shear diagonal crack at 1.5 mm was recorded at a load of 390 kN. Figure 14 shows the shear crack width development for specimens CDB1PC20 and RPCDB1P-8NSM loading.
From observation, the beam (RPCDB1P-4NSM) delayed the appearance of the first shear crack by (133%) with respect to the load compared with the beam (CDB1PC20) due to the higher shear capacity of the specimen (RPCDB1P-4NSM). Moreover, the beam (RPCDB1P-4NSM) reduced the first shear   crack width by (9%) compared with the beam (CDB1PC20) at (65%) of the ultimate shear load.
In addition, the beam (RPCDB1P-8NSM) delayed the appearance of the first shear crack by (150%) with respect to the load compared with the beam (CDB1PC20) due to the higher shear capacity of the specimen (RPCDB1P-8NSM). Moreover, the beam (RPCDB1P-8NSM) reduced the first shear crack width by (33.25%) compared with the beam (CDB1PC20) at (65%) of the ultimate shear load.
From the drilled core test based on ASTM C42/C42M-18a [28] applied on beams (RPCDB1P-4NSM) and (CDB1PC20), real and realistic data for those deep beams were obtained, which were translated in the form of stress-strain curves which including longitudinal and lateral stress-strain curve for concrete compression zone of beams (CDB1PC20) and (RPCDB1P-4NSM) as shown in Figures 15 and 16. The longitudinal stress-strain curve for the concrete compression zone for (CDB1PC20) and (RPCDB1P-4NSM) specimens was combined with one aggregation curve to compare the results between these two specimens as shown in Figure 17.
In Figure 17, the stress-strain curve of the control specimen is shown in the (CDB1PC20) specimen, the shape of the sample is evident, and the drop is also seen at the end of the loading. However, the behavior of the (RPCDB1P-4NSM) specimen is very different compared to that of the (CDB1PC20) specimen of the higher stress and decreased strain of the reactive powder concrete deep beam, and this is evident by the high peak and little extension of the curve.
Young's modulus increased from 27,932 MPa for reference mixings to 43,029 MPa for mixings that have 10% silica fume powder. In addition, the flexural strength rises by rising silica fume to 12.5%; then, after this value decrease [26].
From the split-cylinder test based on ASTM C496 [27], the stress-strain results were obtained for the concrete tensile zone for CDB1PC20 and RPCDB1P-4NSM specimens as shown in Figure 18. In Figure 18, the stress-strain curves for the stress in the tension zone for CDB1PC20 and RPCDB1P-4NSM specimens; it can be observed that the tension stress of beam RPCDB1P-4NSM is higher than beam CDB1PC20. Moreover, the strain of specimen beam RPCDB1P-4NSM is higher than that of specimen CDB1PC20. The increase in tensile stress for the RPCDB1P-4NSM specimen at maximum stress compared with the CDB1PC20 specimen is 10%, while the increase in tensile strain for the RPCDB1P-4NSM sample at maximum stress is 29%.
All the stress-strain results for specimens RPCDB and CDB were compared based on the results of previous studies as shown in Figure 22. It showed an acceptable agreement with those results, which proves the validity of these laboratory results.