Analysis of static response of RC beams with NSM CFRP/GFRP rods

A BSTRACT . In this paper experimental results of investigation on reinforced concrete (RC) beams strengthened with the near surface method (NSM) are analyzed considering the response under bending tests on two beams. One of the RC beams was damaged by bending until the yield of reinforcement and successively strengthened with carbon fiber polymer (CFRP) rod, while the second beam was strengthened with glass-FRP rod. Both the beams have been subjected to bending tests until failure. Experimental diagrams and discussion on static response are presented in the paper. It also places a particular emphasis on the non-linear response of RC sections strengthened with CFRP and GFRP rods under bending moment beyond the first elastic behavior.


INTRODUCTION
he damage of reinforced concrete (RC) structures both for civil building and bridges is an important duty of structural engineering. Rehabilitation of structures requests innovative techniques with new material [1][2][3][4]. In the last decades the use of fiber reinforced polymers (FRPs) has been growing and, in general, two methods of strengthening have been adopted. First it was the FRP strips or laminates glued to surface of RC beams [2,3]; this technique may be indicated as external bonded (EB) strengthening. It appears an available method to strengthen RC beams, but it may be loss validity for impact or due to fire because the FRP strips or laminate are non-sufficiently protected. A second technique is the near surface mounted (NSM) that foresees to insert FRP rods in grooves on the concrete cover [4][5][6][7][8][9][10]. This technique appears available in many experimental studies thought the bond between concrete and FRP rods is a yet open problem to analyze [10][11][12][13]. In fact, the maintaining of bond is the most important condition that must be monitored in the beams strengthened with NSM FRP rods [14][15][16][17][18]. Experimental works have been developed by researchers to analyze the static [14,15] and vibration responses of strengthened RC beams in real scale or in small scale [19][20][21][22]. Other aspects that may influence the response under loading are: shape of FRP rods, circular or rectangular; distance between rods and surface of grooves; roughness of FRP rods' surface; tensile strength of FRP rods [23][24][25]. In literature, investigations on the behavior of RC beams strengthened with CFRP and GFRP rod are present [26][27][28][29][30] also considering new methods [30]; further Codes of Practice [31] have been developed and they may be utilized by engineers. Unfortunately, many aspects of NSM technique T must be develop more and other experiments are necessary to define modelling and responses both in service and ultimate conditions. This paper deals with the investigation by static tests on RC beams strengthened both with CFRP and GFRP rods. A couple of beams in real scale with one RC beam subjected to bending tests at different damage degree is analyzed: one beam, damaged by bending and strengthened with NSM Carbon-FRP rod, has been tested; another one RC beam strengthened by NSM GFRP rod, has been experimentally studied. Static results are shown and discussed below considering the nonlinear behavior of RC beams under loading.

Set-up and strengthening material
xperimental tests were carried out on two RC beams, B1 and B2, with a cross-section of 120·160 mm and length of 2200 mm, the steel reinforcement consisted of four longitudinal steel bars of 10 mm diameter and stirrups of 6 mm diameter. A groove of 20·20 mm was created at the bottom of each beam to locate an NSM reinforcement bar. After static tests, beam B1 was strengthened with a CFRP rod measuring 9.7 mm in diameter and 2000 mm in length, while beam B2 was strengthened with a GFRP rod of 9.5 mm diameter and 2000 mm length. Fig. 1 depicts geometric configuration of the specimens. The two beams tested in laboratory were characterized by concrete with a tested average cylinder compressive strength equal to fc=44.31 MPa and Young's modulus Ec=34492 MPa; steel bars with a yielding stress equal to f y =450 MPa and Young's modulus E s =210 000 MPa.  (a) (b) Figure 2: (a) Specimens of CFRP and GFRP rods for tensile tests; (b) failure of specimens after tensile tests. E Company MAPEI produced the CFRP and GFRP rods used in this research. Tensile tests on three specimens for each type of fiber were used to determine the mechanical and geometrical properties of the C-GFRP rods (Fig. 2a); all the samples' failures were in the XGM category -Explosive, Gage, Middle -as defined by ASTM D-3039 [32] (Fig. 2b). Main results obtained by tensile tests are summarized in Tab. 1. A two-component epoxy structural adhesive (Eres=5130 MPa) was adopted for the bonding of the FRP rod to the concrete surfaces into the notch.

Static response of beams damaged and strengthened
The first phase of static bending tests involves experimentation on the un-strengthened RC beam B1. The apparatus was design to reproduce a simply supported condition with hinge restraints at the extremities, as presented in Fig. 3. The external load was applied in two points placed at a distance of 300mm through vertical jacks; the compressed concrete's and the tensile steel's strains were monitored using electronic strain gauges positioned at the centreline (Fig. 4). The beam's deflection was recorded by a Linear Inductive Displacement Trasducer (LVDT) applied at the midspan section. The beam without strengthening was subjected to static tests using an increasing bending loading path. Three cycles of loading were adopted. For each cycle of loading P i , a damage level D i with i=1,2,3 was identified, as shown in Tab. 2. Results of static bending tests, in terms of deflection, concrete and steel strain and curvature, are summarized in Tab. 2. The experimental diagrams displayed in Fig. 5 were developed by measurements taken in terms of deflection and strains at the midspan of RC beam B1. Therefore, it is possible to characterize the static behavior of beam B1 before the application of NSM CFRP strengthening, based on these data. After reaching a consistent state of concrete's cracking, a CFRP circular rod was positioned in the notched and then it was filled by adhesive epoxy resin. The strengthened beam B1 was subjected once again to cyclic bending loading. The first three damage levels were the same as for the previous unreinforced case. Two additional cycles of loading, D4 and D5, were adopted, for a total of five degrees of damage.  After the fifth damage level, the strengthened beam B1 was subjected to increasing load until the collapse. The maximum load reached during test, that lead to the specimen's failure, was equal to P u = 49.06 kN. As previously described, the recording of the strains achieved during the test, has been entrusted to one strain gauges placed on the CFRP rod, as shown in Fig. 5. Tab. 3 contains a summary of the main data acquired from the instrumentation used for each loading step and the curvature at the midspan section evaluated for the strengthened beam B1.   In Fig. 7 the diagrams of moment, M, versus curvature, χ, refer to the cross-sectional area, obtained by static bending tests are illustrated: the first one is referred to un-strengthened specimen B1, where χ is calculated from values recorded on compressive concrete edge and on steel bar; the second one is referred to strengthened beam B1, where χ is obtained from strain values measured on compressive concrete edge and on CFRP rod. Results of static bending tests obtained for beam with NSM CFRP strengthening are shown in Fig. 6 in term of loaddeflection (P -δ) and load -strain of tensile CFRP rod (P -εCFRP) diagrams evaluated at midspan section of beam. During the tests, the propagation of cracks was also visually observed: after the first load cycle, with P1 = 4 kN, the cracks on the un-strengthened specimen, was almost absent; when the load was increased, the crack pattern followed the trend of a typical RC beam, with vertical cracks in midspan and oblique cracks nearby the supports, as shown in Fig. 8.
In the NSM CFRP strengthened beam, the crack pattern developed following the fractures occurred during the previous test on the un-strengthened specimen; only for the last two loading cycles the cracks increased in depth and width (Fig. 9). If we analyze the collapse mode of beam B1 strengthened with NSM CFRP, it can be seen that the specimen reached failure for the crushing of compressed concrete and the debonding of the CFRP rod that began at the maximum moment region region and propagated to an extremity of beam. In particular, the debonding between adhesive and the surrounding concrete was recorded at midspan; moving away from the midspan section, also part of concrete cover was interested (Fig. 10).    Following the same strategy, RC beam B2, strengthened with NSM GFRP rod was tested statically, adopting the same setup and the same instrumentation. Static bending tests were carried out on the beam applying four cycles of bending loading and then increasing the load until the collapse of the sample. In this case as well, the steps of loading were related to the degree of damage D i with i=1,2,3,4. Tab. 4 presents the main level of loading adopted durin tests together with the main data achieved from experimentation. Fig. 11 presents the trace of the deflection, δ, recorded at the centerline of beam in relation to the increasing of bending load, P; in addition, the development of strain on tensile GFRP rod, ε GFRP , is traced. The evolution of the curvature, χ, relative to the cross-sectional area and computed from data measured on compressive concrete and GFRP rod, in relation to the increasing of moment, M, is depicted in Fig. 12  In terms of fracture's propagation, a condition like the previous one occurs during the test, as shown in Fig. 13. After attaining a load value of Pu= 38.40kN, the specimen B2 collapses, resulting in the crushing of compressed concrete and total debonding of the NSM GFRP rod. The strengthening's debonding interested the portion of beam from the midspan section to the end section, leading the detachment of the concrete cover.

DISCUSSIONS OF EXPERIMENTAL RESULTS
his experimental research performed on RC beams strengthened with NSM CFRP and GFRP rod permits the highlighting of numerous aspects that could be helpful in the civil applications for the structural repair of damaged elements. The effectiveness of the Near Surface Mounted approach is the first outcome that should be highlighted. This technique allows for beams to be strengthened until their collapse under bending, preserving the FRP rods' connection without any separation. The strengthened elements reached failure for the attaining of the ultimate strain of compressive concrete. An improvement in the rigidity capacity of RC beams with NSM CFRP and GFRP during bending tests was verified. If we examine the load versus deflection experimental diagrams for the models with strengthening, it can be seen how ductility and ample deflections typify the response of strengthened elements until failure condition. In Fig. 14 the behavior of two beams under bending loading cycles until failure are compared. We can see that the strengthening of beams with FRP rod is adequate for both beams. Stiffnesses of beams B2 is lower than that of beam B1 and this is a direct result of the mechanical properties of the strengthening bar of CFRP respect to GFRP rod being the Young's modulus of GFRP is much lower than that of the CFRP while the area of section is almost equal between CFRP and GFRP rods. This result is also reflected on the ultimate capacity in terms of load P u which is minor of 30% for the beam strengthened with GFRP rod. Moreover, another important aspect emerging from experimental campaign, is the impossibility to apply the Bernoulli's hypothesis in the study of RC sections of beams strengthened with NSM FRP rod; this is due to the presence, under bending, of a FRP stress-strain lag that which makes it impossible to consider the section as plane. The entity of strain collected at the midspan section at damage degree Di has been diagrammed and depicted, respectively, for the un-strengthened beam B1, T in Fig. 15, for the NSM CFRP strengthened beam B1, in Fig. 16 and for the NSM GFRP strengthened beam B2, in Fig. 17. The measurement of strains obtained for the edge of compressive concrete and at the level of steel bars underlines the maintaining of plane section in the case of un-strengthened RC beam B1 (Fig. 15). In Fig. 16 the point of compressive edge of concrete, tensile steel and tensile CFRP rod is considered for beam B1 strengthened with CFRP rod. In this case, it is noted the non-linear distribution of strains through the full depth of the beam; in particular, the strains on the CFRP rod aren't linearly congruent with the strains of steel and of the compressed concrete fiber and are affected by a stress-strain lag. It means that the hypothesis of preserving the planarity of the bending section isn't satisfied. Also, in the case of beam B2 the non-planarity of section appears at midspan since the first load cycle D1=4kN, as Fig. 17 shows.   The entity of stress-strain lag can be estimate considering the ratios (1) and (2)