TORSION ENHANCEMENT OF REINFORCED SELF-COMPACTING CONCRETE BOX BEAMS USING INTERNAL FRAMED STEEL STIFFENING RIBS

Researches concerning internal torsional strengthening of box beams are very little, since few researchers may study such method of strengthening. This research adopts employing of internal Framed Steel Stiffening Ribs (FSSRs) to strengthen reinforced SelfCompacting Concrete (SCC) box beams subjected to pure torsion. These FSSRs were fabricated to be fully bonded with SCC through using of bolts as shear connectors. Four specimens (the first was a reference one with no strengthening and the other three were strengthened with one, three and five FSSRs, respectively) of reinforced SCC were tested under effect of pure torsion up to failure. All specimens have the same dimensions and reinforcement. The effect of FSSRs presence and their number on torsional capacity was studied. Also, applied torque-angle of twist relation at end, quarter and mid span were studied. Test results show that the ultimate torsional capacity is increased by 45.7 %, 75.5 %, and 122.4 % for specimens strengthened with one, three and five FSSRs, respectively, as compared with reference specimen. Presence of FSSRs tends to reduce deformations where deformations decrease with increasing of FSSRs number. It is concluded that the torsional capacity of a reinforced SCC box beams can be enhanced using internal Transverse FSSRs. The number of used internal FSSRs have main effect on the amount of enhancement.


Torsion in structural members
Structural members made of reinforced concrete are designed to resist flexural, shear, axial, and torsion forces according to loading conditions. Torsion forces are rarely act without bending, shear, or axial forces. Such torsion forces try to twist the member about its axis [1]. Torsional strength is of central feature in analysis and design of many structural members as curved bridges, eccentrically loaded box beams, edge beams in buildings, and helical stairway [1]. When a torque is applied to a member, shearing stresses on cross-sectional planes and on radial planes is developed extending from the axis of the member to its surface. In a circular member, the shearing stresses change, linearly, from zero value at the member axis to a maximum value at the outside of the member section. The maximum shear stress in circular section is constant along the circumference, while it changes from zero value at corners to a maximum value at the centers of the long sides of a rectangular member perimeter [2]. Torsional loading can be separated into two basic categories: equilibrium torsion, case of this research, where the torsional moment is required for the equilibrium of the structure, and compatibility torsion, where the torsional In this research, SCC is used because of small thickness of box section walls, presence of reinforcement as well as with FSSRs.
SCC is a mix with no need of vibration to expel air out of it and travels round obstacles, like reinforcement, to fill the formwork. Any mix has: i. passing ability, ii. Filling ability, and iii. Segregation resistance can be classified as SCC one. In addition, SCC overcomes poor accessibility near prestressing tendons and anchorages, and reduces noise level and health effects on operators using hand held vibrators. The only limitation is that the top surface must be horizontal [4]. Recognizing the lack of uniformity and complete compaction of concrete by vibration, researchers at the University of Tokyo, Japan, started in late 1980's to develop SCC. By the early 1990's, Japan had developed and used SCC that did not require vibration to achieve full compaction. By year 2000, the SCC had become popular in Japan for prefabricated products and ready mixed concrete [5].

Strengthening of Concrete members
Normally, concrete structures during their service life suffer in some way or another from strength deficiency; therefore they require strengthening or repair. Need of them may be resulting from errors in construction or design, changes of function, updates of design code, maintenance lack, structural system changing, traffic volumes increasing, blasts and explosions, accumulated damage over time or caused by accidental overloading, fires, or earthquakes [6]. Civil engineering infrastructure renewal has, and still, considerable attention over the last two decades, although the option of deficient or deteriorated structure replacement is desirable. Economic reasons make strengthening or repair to be the choice [7]. Purposes and strengthening techniques of structures are vary. Most strengthening systems consist of a composite strengthening one. The composite strengthening can be achieved by enlargement of section, external post-tensioning, steel elements bonded externally, Textile Reinforced Concrete (TRC), Fiber-Reinforced Polymer (FRP) composites, and Near-Surface Mounted (NSM) system. Anyone of these techniques or a combination of them can be used provided that the structure functionality must be kept throughout its designed service life [8]. Imran (2012) [9] made a review on existing literature related to strengthening of reinforced beams for flexure, shear and torsion so behavior can be better understood. This review showed that strengthening of RC beams using Carbon Fiber Reinforced Polymers (CFRP) are the most widely and suitable strengthening technique. They concluded that torsion strengthening is difficult and different studies suggest full wrapping which it is practically impossible from application point of view. However, several experimental studies revealed that 45° CFRP wraps are useful to enhance torsional strength of beams.

Previous studies
Al-Bayati and Al-Mahaidi (2013) [10] aimed to present a literature review concerning the torsional behavior of reinforced concrete beams strengthened with CFRP since both techniques, Externally-Bonded Reinforcement (EBR) and Near-Surface Mounted (NSM) were and still of widespread using. Among different strengthening layouts, full-wrap along the beam span seems to be the most effective layout for resistance and ductility improvement. Using such strengthening technique restricted propagation of cracks. Spacing and number of layers of CFRP affect considerably the beam torsional capacity. Anchors using produces a longer response and delays failure but crashing of concrete at unstrengthened areas may occur. Use of inclined 45° spiral strips seems to be the most effective layout but it has practical limitations. Reviewing of literature shows that no available data on torsional behavior using NSM technique to strengthen reinforced concrete beams. Khalil et al (2015) [11] investigated the behavior of ten reinforced concrete box beams, with and without web opening, under pure torsion. Beams were strengthened using External Prestressing Technique (EPT) horizontally and vertically. The strengthening using EPT enhance the torsional capacity by 58%, vertical EPT was more effective, and increased torsional capacity of beams with opening by about 13% were some of conclusions of the investigation related to present work. Shengqiang Ma et al (2016) [12] fabricated four RC beams of box section and tested under pure torsion till failure. First beam was a control beam while second, third and fourth one were externally bonded with CFRP. Second one was strengthened using U-wrap of one-layer, third one was strengthened using one-layer of both Uwrap and longitudinal strip and fourth one was strengthened using two-layers of U-wrap and one-layer of longitudinal strips. Ultimate torque is increased moderately by 16.6%, 15.9%, and 20.5% for second, third and fourth strengthened beam respectively. Torsional stiffness of strengthened box beams with longitudinal strips increased slightly, the cracking torque also increased. Aghara and Vora (2017) [13] studied torsional strengthening of RC beam with Glass Fiber Reinforced Polymers (GFRP) lamination experimentally and analytically. Six rectangular beams having rectangular section were casted; one of them was control beam and the others were wrapped with GFRP of different configuration. Non-Linear Finite Element (NLFE) was used to perform the task of numerical work. The work contained good review on previous studies of torsional strengthening for beams. They concluded that wrapping pattern at 45 o angel give more torsional strength than wrapping at 90 degree. Kandekar and Talikoti (2018) [14] studied the torsional behavior of reinforced concrete beams strengthened using aramid fiber strips to wrap the beams with different patterns. Twenty one reinforced concrete beams of rectangular section were cast and tested until failed in torsion. Different aramid fiber configuration effect on torsional capacity, twist angle and mode of failure of the beams is studied. It is seen that torsional moment carrying capacity increases when the beams was wrapped using aramid fiber strips. As strips spacing were increased, the torsional moment capacity decreases and the twist angle has small variations. In strengthened beams, Initial cracks occurred at higher torsional moments. Aziz and Hashim (2018) [15] studied, experimentally, the torsional behavior of six SCC box beams strengthened with transverse concrete diaphragms which may be opened or closed. First beam was the reference beam of hollow section, second and third were strengthened by two closed and opened i. Examining, experimentally, the significance of using the internal transverse FSSRs in strengthening of reinforced SCC box beams. ii. Studying the influence of number of the FSSRs on the amount of torsional strength enhancement. The new adopted method may be used to strengthen existent or prefabricated box girders according to cost study.

Experimental program
The experimental program comprised of pouring and test four reinforced SCC box beam specimens until failure by torsion only. The four specimens included one reference beam which has no strengthening and three beams strengthened with one, three, and five FSSRs of closed sections, respectively. The experimental program, also, contains a series of tests performed on control samples (cylinders and prisms) so mechanical properties of hardened SCC (compressive and tensile strength, modulus of elasticity) can be determined as well as many test to control SCC (slump flow, L-box, Vfennel, U-box and G-ring tests). Also, the reinforcing steel and steel plate from which the FSSRs are made were tested for yield and ultimate strength. During all specimens tests, clear span, concrete grade, and loading condition remain unchanged. Also, all specimens were simply supported attached on rotatable supports. Cracking and ultimate loads, mode of failure, and torque-angle of twist behavior are presented, and discussed in this research. Fresh SCC test methods adopted in this paper is based on EFNARC, Specification and Guidelines for Self-Compacting Concrete, [17].

Beam specimens description
All specimens have the same dimensions of (2100 x 300 x 300) mm for span length, width and height, respectively, and wall thickness of (60) mm. The longitudinal and transverse reinforcement were directly calculated based on ACI-318-14 code torsion requirements [18]. All beam specimens were of same reinforcement consisting of longitudinal bars (2ϕ12mm top and bottom) and transverse stirrups (ϕ8@50 mm at both ends of beam specimen to a length of 200 mm and ϕ8@130mm stirrups otherwise). Details of tested beam specimens are shown in Table  (1). Details of the reference beam specimen, 1NR, which are, except presence of FSSRs, typical for all tested beam specimens as explained in Figure (1).
As mentioned previously, the new technique of strengthening of box sections using internal transverse FSSRs (to enhance resistance against torsional stresses) is adopted in this paper. Therefore, transverse FSSRs of different number were used and located in their position before SCC pouring, see Figure (3). These FSSRs were fabricated to have square shape of dimensions, (180x180) mm as that of internal dimensions of the box beam specimens' cross section, closed cross-section, and shear connectors on all sides. The cross section of FSSRs was not of a standard type. It was made by assembling of its parts made of steel plates prepared previously, utilizing welding of (E6013) type. Details of FSSR are illustrated in Figure (2). -----1 3 5 FSSRs Type -----Closed section * First number refers to specimen number, N means nonprestressed, R means reference, second number refers to quantity of FSSRs, C means closed section, and F means FSSR. ** Reference beam specimen. *** Called wall thickness of beam section.

Materials
To make the required test beam specimens, the materials shown in Table (2) were used to produce the SCC as well as with longitudinal, transverse reinforcement and FSSRs in addition to styropore needed to form section hollowness. Figure (6) shows details of works of longitudinal and transverse reinforcement, FSSRs fixing in their final locations, and styropor fixing in proper position. All materials used in this research were conforming to applicable standards in Iraq. (Vol. 24, No. 04, July 2020) ISSN 2520-0917

Mix and pour of SCC
Rotary mixer, Figure (4), was used to mix the raw materials and produce the SCC for all specimens. Two steel formworks made of (2 mm) thickness steel plate stiffened with steel angle section members had been fabricated and used to cast the beam specimens. The molds had an internal dimensions of (2100x300x300) mm. Formworks was set on firm ground in horizontal level and well lubricated before casting, see Figure (5). The proportions of SCC mix are reported in Table (3).      a. Sand ratio = sand / (cement + sand + gravel + silica fume + limestone). b. Gravel ratio = gravel / (cement + sand + gravel + silica fume + limestone). c. As percentage replacement of cement. d. Superplasticizer ratio = superplasticizer / (cement content + silica fume + limestone). e. W/P = water/( cement content + silica fume + limestone).

Instrumentation and measurements.
Specimens of this research were tested for cracking and ultimate torques using hydraulic universal testing machine (produced by Maschinen Fabrik Liezen, MFL) with load capacity of (3000 kN), Figure (8). Angle of twist was measured at four locations; at both ends of specimen using two dial gauges for each end, at mid-span and at quarter span. All dial gauges which have accuracy of (0.01 mm/division), were attached to the bottom face of the specimen and located transversely at distance of (100 mm) from longitudinal axis. Twist angles were calculated, in degrees, depending on vertical deflection recorded for each step (5 kN) of loading at the locations illustrated above. Loading was applied gradually at constant rate.  Total longitudinal deformation was measured at both specimens' ends using one dial gauge for each end with (0.01 mm/division) accuracy fixed at the center of end cross section. Then horizontal deflection was calculated to find longitudinal elongation, in millimeters, for each loading stage. Figure (9) shows a photo of locations of used dial gauges.

Test procedure
Test setup of beam specimens is as shown in Figure (7). Beam specimen which to be tested is attached to machine supports which can set to be rotatable, so the specimen can rotate freely at ends about its longitudinal axis. The load is applied on the center of wide flange steel beam which transfers it equally to the loading arms as two concentrated loads out of the longitudinal axis of beam specimens under test by a distance of (500 mm). This eccentric distance, at the ends of specimen, is necessary to obtain torsional behavior through transfer of loading on the arms as concentrated torque at the ends of specimen and concentrated vertical force which transferred directly to the support. Applied loads of first visible crack (Pcr) and of final failure (Pu) were recorded. It is clear that this set up is symmetric about specimen's mid-span except (T) which has opposite sense at the ends.   Journal of Engineering and Sustainable Development (Vol. 24, No. 04, July 2020) ISSN 2520-0917 6. Properties of SCC

Fresh SCC Control Tests
According to EFNARC [17] standards, there are ten different tests to ensure that the produced concrete is of a SCC type. It is not necessary to make all of these ten tests on fresh SCC; so five tests were adopted in the present research as a control tests on fresh SCC, namely slump flow, J-ring, L-box, V-funnel, and U-box tests, see Figure (10-14). These five tests gave an indication for fresh SCC properties specially the important three characteristics of the SCC, filling ability, passing ability, and segregation resistance. Tests results indicated that used SCC conforms to the requirements of EFNARC [17].

Cracking and ultimate load
Beam specimens were tested under application of only torsion loading gradually. Loading process was continued until final failure mechanism occurs. Both cracking and ultimate loads (Pcr and Pu) were recorded, and then torsional moments (torques) were calculated, see Table (    percentages of (Tcr) increasing have little differences while they have considerable differences for (Tu). This may be resulted from the fact that at starting of loading, the applied load is resisted mainly by concrete and first crack initiates at midpoint of the external fiber of box section side. After that reinforcing steel begin to have its role in resisting the applied load. At final stages of loading, the applied load will mainly be resisted by FSSRs. On other point of view, it can be said that the (Tcr) value depends mainly on concrete contribution in resisting of applied load, while the (Tu) value depends on contributions of concrete, reinforcing steel, and FSSRs. As results, it can be concluded that the adopted strengthening technique gives a good enhancement in torsional ultimate capacity.

Angle of twist and longitudinal elongation
Two types of deformation were considered in present research, angle of twist θ and longitudinal elongation δL. The angle of twist was recorded at four locations; left end and right end which averaged to obtain end span angle of twist θES, quarter span θQS, and mid span θMS of the tested specimens. Longitudinal elongation δL is found by summing of two values recorded at left and right ends of the tested specimens. These deformation were recorded for both Tcr and Tu, Table (6). Figure (17-19) shows the applied torque-angle of twist curves and Figure  (20) shows the applied torque relation with longitudinal elongation. Note that θ is the total angle of twist in degrees and δL is the longitudinal elongation in millimeters. Subscripts ES, QS and MS means end, quarter and mid span, respectively.  Figure 15. Applied torque capacity, cracking and ultimate torques. Journal of Engineering and Sustainable Development (Vol. 24, No. 04, July 2020) ISSN 2520-0917

Mechanisms of failure
The propagation of cracks throughout length of beam specimen during the test gave good information regarding the failure mechanism. First visible crack of all specimens occurred near mid span at some point on mid height of front face of beam specimen which represents the weaker point and subsequently it increased gradually. As applied torque moment was increased, cracks extended to all sides gradually and finally took the spiral shape. Generally, major cracks, including first visible crack, were initiated and developed within the middle third of the span; then as cracks became closer to the ends of beam specimen, they had smaller width. This is may be due to that both ends of the specimen were solid and had transverse reinforcement more than that at entire span; also ends of the specimen were confined vertically by concentrated force at supports and laterally by forces due to loading steel arms fixation, and even that both ends were free to rotate. Accordingly, the final failure position of most beam specimens occur near mid span. Figure  (21) shows the failure mechanisms of tested specimens. It can be noted that all tested specimens failure was caused by diagonal cracks in concrete which had a spiral shape.

Cracks pattern
For reference beam specimen 1NR, which had no strengthening FSSRs, the cracks spread in an entire beam length. The cracks of middle third of the span was of width slightly more than that of the end thirds while near the supports the cracks had much smaller width. Generally, with increasing of loading the number of cracks increased and failure take place somewhere near the mid span. At failure, excessive parallel cracks at all sides along the span of the beam were observed. For strengthened beam specimens 5N1CF, 6N3CF, 7N5CF, the cracks spread with nearly the same number and grow slower at the strengthened zones where the strengthening FSSRs were provided since they reduce the rate of cracks growth. due to presence of one FSSR only that means section at mid span has stiffness more than that at ends. When more than one FSSRs were used, specimens 6N3CF and 7N5CF, the first visible crack tends again to be nearer to the mid span but still having distance more than that for specimen 1NR. This is may be due to that stiffness was increased at ends as at mid span. Finally, in all tested specimens, the cracks had the spiral shape around all faces of each specimen. The number of complete cracks of each tested beam specimen was approximately the same. This means that the strengthening process has no large effect on the number of cracks like its effect on values of cracking and ultimate torques as explained in section 6.1.

Conclusions
The following can be concluded depending on output results of the experimental work of this research and according to the discussion presented in section (7)