Torsional Strength Prediction of RC Hybrid Deep T-Beam with an Opening using Softened Truss Model

This paper presents an analytical study to predict the torsional behavior of reinforced concrete (RC) hybrid deep T-beam with opening and to compare to the experimental results. The RC hybrid deep T-beam was cast with normal weight concrete web and light weight concrete flanges. Based on the Softened Truss Model and Bredt’s theory, a new equation was derived and proposed to show the relationship between compression strut of light weight concrete and normal weight concrete. Based on the experimental results of cracking angles and curvature equation of light weight concrete and normal weight concrete, a new equation was also proposed to show the relationship between strain diagram of light weight concrete and normal weight concrete. The analytical results show that the prediction of maximum torque capacity of the beams are close to the experimental results, except the prediction of twist angle which are larger than the experimental results.


Introduction
In order to pass utility (ducts and pipes) and also to save headroom, an opening is often required in the beam. Beam with opening subjected to pure torsion is very rare. Normally it is subjected to combination of bending, shear and torsion. However, it is important to study the beam under pure torsion as the basis of understanding their complex behavior under combine loadings [1]. Researches in torsion of shallow beam were extensively carried out [2][3][4][5][6][7][8], however, research in torsion of deep beams is rare and attracts little attention. Beams with span/depth ratios about four or less are categorized as deep beams [9]. Akhtaruzzaman and Hasnat [10] conducted investigation on the torsional behavior of deep beams with and without opening, while Samman et al [11] reported their investigations on the torsional behavior of rectangular plain concrete deep beams without opening.
Lisantono et al. [12] has reported preliminary investigation on the effect of light weight concrete (LWC) flanges and web opening on the torsional capacity of RC deep T-beams. Lisantono et al. [13,14] also investigate the effect of web opening locations, in horizontal as well as in vertical direction, on the torsional behavior of RC deep T-beam. To obtain a clear understanding on the effect of web opening dimension on the torsional behavior of RC hybrid deep T-beams, Lisantono et al [15] conducted an experimental program. This study aims to develop a method for predicting the torsional capacity of RC hybrid deep T-beam, especially deep T-beam with web opening.

Softened Truss Model Theory
A beam subjected to pure torsion will behave as an analogous thin wall tube-space truss [2][3][4][5][6]. After the beam cracks, concrete strut will be subjected to compression force and the reinforcement acts as a tie. Hsu [16] stated that three equilibrium conditions of Truss Model satisfy Mohr's stress circle and by assuming that the steel bars can only resist axial stresses, superposition of concrete stresses gives: According to the Bredt's theory which stated the shear flow should be constant along the centerline of the shear flow zone and can be related to the torque, T. Hsu [4] adding one more equilibrium equation: where T = torque A0 = the area within the centerline of the shear flow td = the effective thickness of shear flow Hsu [16] showed that from the compatibility condition of the truss model, the average strains satisfy the Mohr's strain circle and gives the following equations: where l, t = average strains in the l and t directions, respectively (positive for tension) lt = average shear strains in the l-t coordinate (positive as shown in Figure 1) d, r = average principal strains in the d and r directions, respectively (positive for tension) The stress and strain of concrete in the d direction follows material law proposed by Vecchio and Collin [17] for the softened concrete: where d = compression stress at the concrete strut  = softening coefficient  = Poisson rasio 0 = strain at the maximum compressive stress of nonsoftened concrete and can be taken as -0.002 [4].
Hsu [7] used the softening coefficient as proposed by Belarbi and Hsu [18] as follows: The stress-strain relationship in the r direction can be expressed by where Ec = modulus of elasticity of concrete cr = strain at cracking of concrete, taken to be c cr E f fcr = stress at torsional cracking of concrete is assumed to occur when the principal tensile stress reaches the tensile strength of the concrete in biaxial tension-compression and can be taken as The stress-strain relationship for longitudinal and transverse steel bars are assumed to be elasticperfectly plastic.

Softened Truss Model for RC Hybrid Deep Tbeam
According to CEB-FIP Model Code 1990 [20] for beam consisting of several rectangular sections, the torque of the beam can be assumed to be distributed on each rectangular section as: where T = total torque Ti = torque that assumed to be resisted by the-i th section Xi = the smaller size of the-i th section Yi = the larger size of the-i th section The RC hybrid deep T-beam in this research was cast of normal weight concrete web and light weight concrete flanges as shown in the Figure 2.a.
If the flange portion and web portion are denoted as I and II, respectively, the proportion of each torque becomes: Considering the web opening in the beam ( Figure  2.b), Equation 12 and 13 become: (a) Torque proportion without opening (b) Torque proportion with opening  (15) where T = total torque TI = torque resisted by flange (light weight concrete) TII = torque resisted by web (normal weight concrete) bf = width of flange hf = thickness of flange bw = width of web hw = height of web d0 = opening diameter To derive the torque capacity of a RC hybrid deep Tbeam after cracking the beam is assumed to be a space truss-thin wall tube analog as shown in Figure  3. Figure 3 shows that the shear flow of flange section (light weight concrete) is qI and the effective thickness of shear flow zone is tdI, while the shear flow and the effective thickness of web section (normal weight concrete) are qII and tdII, respectively. Taking a small element A at flange section and B at web section of the RC hybrid deep T-beam as shown in Figure 3, the shear stress of element A will be: According to the Bredt's theory [4] the shear flow q must be constant along the shear flow center line. Therefore, the relationship of shear flow of light weight and normal weight concrete can be obtained as: Hsu [5] stated that as long as the tensile strength (r) of the concrete is very small compared to the compressive strength (d), the effect on the torque will be small too. Therefore, the tensile strength (r) can be neglected in the analysis. Substituting rI = rII = 0 into Equations 22 and 23, gives illustrates that the compressive force at the strut of light weight concrete must be in equilibrium condition with the compressive force at the strut of normal weight concrete. Therefore, if the characteristics of light weight and normal weight concrete are not the same, the effective thickness of shear flow zone or td must not be the same either.
Lampert and Thurlimann [21,22] proposed that due to warping, the strut concrete is not only subjected to compressive force but also flexure at the tube wall. This phenomenon can be stated in curvature equation as: where,  = curvature of concrete strut According to the flexural phenomenon at the tube wall, the curvature of concrete strut at each light weight and normal weight concrete can be obtained as: , it can be obtained that (see Figure 4): It is to be noted that program was terminated when εdI has reached the value of 0.00175 or εdsI = 0.0035, where εdsI is the compression strain of concrete at the outer surface of diagonal concrete strut, while εdI is the compression strain of concrete at the middle height of the effective thickness of the diagonal concrete strut (see Figure 4).

Experimental Program
To verify the theory, Lisantono et al [15] conducted experimental program of torsional RC hybrid deep-T beams. Four beams namely B4HS; B4HOD1; B4HO; and B4HOD3 were prapared. The first beam (B4HS) was reinforced concrete hybrid deep T-beam without opening, cast of normal weight concrete web and light weight concrete flanges.
The second, third and fourth beams (B4HOD1; B4HO and B4HOD3) were reinforced concrete hybrid deep T-beams with web opening diameters of 100 mm; 200 mm and 300 mm, respectively. The circular openings were located at mid span and mid depth of the beam. All beams have the same span of 2000 mm and the same nominal cross-sectional dimensions. Details of the beams are shown in Figures 6, 7, 8, and 9.  Table 1. The test setup of the specimen is illustrated in Figure 10. The beams were instrumented for measurements of prevailing deflections and rotations. The deflections and rotations due to the torque force were measured using LVDT and inclinometers respectively. The electrical resistance strain gauges were used to measure strains in reinforcements and concrete.

Results and Discussion
Comparison between the experimental results and the softened truss model theory of the beam B4HS, B4HOD1, B4HO, and B4HOD3 can be seen in Figures 11, 12, 13, and 14, respectively.

Figure 10. Test Setup of the Specimen
The experimental results show that generally the strength characteristics of the tested beams up to first cracking behave essentially linear. After first cracking, preceded by a small drop in torque, the curves increased non-linearly with increasing twist up to their ultimate torque. It was observed that the small drop in torque occurred when the crack propagated a short distance along the corner line of the web-flange interface. After reaching their ultimate torque, the curves decreased non-linearly with increasing twist, and proceeded with a section of an approximately horizontal plateau, indicating a state of yielding prior to collapse [15]. Figures 11,12,13,and 14, show that the theory can predict not only the torsional strength of the RC hybrid deep-T beam, but also the angle of twist throughout the post-cracking loading history. Comparing to the experimental results, it can be seen that generally the stiffness of the tested beam is stiffer than the curve predicted by the theory. To evaluate the accuracy of proposed theory, the torque capacity and twist angle predicted by the proposed theory were compared to the experimental results. Comparison of the maximum torque capacity based on the proposed theory with the experimental results can be seen in Table 2.   The comparison of the twist angle at maximum torque obtained using the softened truss model with the experimental results can be seen in Table 3.  Table 3 shows that the maximum twist angle predicted by the theory is higher than the experimental result. It means that twist angle predicted by the theory gives more excessive twist angle than the experimental results, especially for the beam with the large opening. It is suspected, this softer behavior is due to the fact that the proposed method was derived based on the assumption that the concrete of the beam was already crack, while in reality the beam has two behavior condition throughout the loading history that is before first cracking and after cracking of the concrete.

Conclusion
Based on comparison of the results between analytical method and experimental program, the following conclusions can be drawn. The softened truss model can predict not only the torsional strength of the RC hybrid deep-T beam, but also the angle of twist throughout the post-cracking loading history.
Prediction of the maximum torque capacity of the RC hybrid deep T-beams based on the softened truss model is close to the experimental results. However, the twist angle predicted by the softened truss model is higher than to the experimental results, especially for the beam with large opening. The excessive prediction of the twist angle is due to the fact that the softened truss model was derived based on the assumption that the concrete of the beam was already crack, while in reality the beam has two behavior condition throughout the loading history, that are before first cracking and after first cracking of the concrete.