Performance of High Strength Self-Compacting Concrete Beams under Different Modes of Failure

Twelve beams were tested as identified in Table 1. The beams are named according to their mode of failure: flexural beams, shear beams or bond beams. For each of the three modes, four beams were tested with two identical replicates made of SCC or VC. Replicates were used to validate the test results. The comparison of the beam specimen behavior was conducted in terms of two variables. The first variable is the type of concrete used in the placement of the tested beams: SCC denotes that third generation (PCE) was incorporated in the production of the concrete mix, while VC indicates that the concrete mix was made using second generation (SNF) superplasticizer. The second variable is the mode of failure of the tested beam (flexural, shear or bond splitting).

Accordingly, the beams are identified by a three part notation system. The first term indicates the type of concrete mix used in the casting of the beam (SCC or VC). The second term specifies the preset mode of failure (F for flexure, SH for shear, and B for bond splitting). The third term designates the listing number of the two replicates (B1 or B2).

The flexural beams were properly designed and reinforced to avoid shear failure before steel yielding. The shear beams were under-reinforced in shear to prompt brittle shear failure. The bond beams had their bottom tensile bars spliced at midspan, the lap length was chosen to be the minimum recommended by the ACI building code to avoid the yielding of the bottom bars and to ensure a concrete splitting mode of failure in the splice region. Longitudinal and cross sectional details are shown in Figs. 1, 2 and 3.

The beam specimen was 2,000 mm long with a distance of 1,800 mm between supports. The width of the beam was 200 mm and the depth was 300 mm.

The reinforced concrete beams were tested using an MTS (Material Testing System) closed loop servo-hydraulic machine with a dynamic actuator having a capacity of 1,000 kN. The span between the centerlines of the supports was taken to be 1,800 mm. The beams were tested in positive bending. The loading system was designed to produce a constant moment region (zero shear) in the middle of the beam specimen. Two concentrated loads were applied continuously at a distance from the two supports equivalent to one third of the span length (600 mm). A steel plate was placed under each point load to distribute the load evenly over the 200 mm beam width. The vertical deflection was monitored at midspan using an LVDT sensor. The two strain gages mounted on the bottom tensile reinforcing bars were connected to a computer system to record the steel elongations inside the concrete. The loads were applied in increments of 10 kN until failure. At each load increment, crack width readings were taken using a crack comparator. In the flexural and shear beams, cracks that have initiated below the two concentrated loads and at midspan were observed. In the bond beams, cracks located at the end of the lap splices were also checked. In all beams, the propagation and widening of the shear cracks were examined attentively. An actual view of the testing setup of beam specimen is shown in Fig. 9. A schematic is also shown in Fig. 10.

The following sections present the equations for the modulus of elasticity, the splitting tensile strength and the modulus of rupture that were found to best represent the experimental outcomes.

The proper performance of reinforced concrete members in flexure or direct tension comes as a result of adequate force transfer between reinforcing bars and concrete. In their studies on the behavior of full-size reinforced concrete elements in bond splitting modes of failure, researchers opted to use splice specimens that appeared to be effective in providing realistic experimental data. Accordingly, four beam specimens were cast to test and compare the bond strength of steel reinforcement in SCC beams.

Results of the tests on bond beams are listed in Table 7. The results include the ultimate load at bond splitting failure, the number of cracks and the maximum flexural crack width at the end of the splice region. The crack widths listed in Table 7 and the cracked beam sketches, associated with SCC and VC bond beams displayed in Figs. 11–20 revealed similar cracking patterns of replicate SCC and VC beams.

Figure 21 displays the response of bond beams to equal load increments. The load deflection curves of SCC-B-B1, SCC-B-B2, VC-B-B1 and VC-B-B2 demonstrated similar stiffness characteristics. The average load resulting in the bond splitting failure was approximately 90.2 kN in reinforced SCC beams and 89.1 kN in VC beams. This experimental result has proven compliance with the equation suggested by ACI committee 408 (2003) (P _SCC  = 75.7 kN and P _VC  = 74.3 kN) and Darwin et al. (ACI Committee 408R 2003). (P _SCC  = 70.7 kN and P _VC  = 69.4 kN) that provide conservative predictions on the bond strength of SCC and VC beams with factors of safety ranging from 1.2 to 1.3. The equation proposed by Esfahani and Rangan (ACI Committee 408R 2003) for high strength concrete overestimates the capacity of reinforced SCC and VC beams (P _SCC  = 107.1 kN and P _VC  = 103.2 kN). The theoretical equations considering the contribution of transverse reinforcement in the transfer of bond stresses were disregarded.

The test results of the bond beams indicate that the high fluidity characteristics of SCC mixes have no impact on the bond characteristics of reinforcing bars, a statement that refutes the hypothesis presented earlier in this paper.

The shear strength of SCC is designated by the ultimate load that triggered the appearance of the first diagonal crack and was evaluated through the close monitoring of the development of this crack. Since the concrete shear strength in beams is dependent on the tensile characteristics of concrete and is independent of the area of transverse reinforcement allocated for a concrete section, the concrete shear strength results of the flexure and shear beams were considered. In the analysis of the test results, the experimental values were compared to the theoretical estimation of the concrete strength in shear. The maximum shear capacity carried by concrete was computed using Eqs. (11-5) of Sect. 11.2.2.1 of the ACI Building Code ACI 318-11 (2011). The ACI equation for the shear strength considers the effect of the longitudinal reinforcement and the applied moment on the shear resistance of reinforced concrete beams.

Test results of the shear beams are listed in Table 8. The results include the load at first diagonal crack and the ultimate shear load at failure. Also listed in Table 8, are the number of cracks and the maximum shear crack width. The crack patterns of SCC and VC shear beams beam are shown in Figs. 22–25.

The experimental results revealed an average concrete shear capacity of 50 kN in SCC beams and 42.5 kN in VC beams compared to a theoretical concrete shear load of 55.5 kN for SCC beams and 53.8 kN for VC beams. A reduction factor ϕ of 0.75 was used in the computation of all the theoretical concrete shear capacities.

Considering identical mix designs for SCC and VC, SCC demonstrated better shear resistance than its VC counterpart. Consequently, the high consistency of self-consolidating concrete has little effect on the concrete capacity in shear.

On the other hand, both SCC and VC beams have shown a concrete capacity that is lower than the nominal capacity computed using Eq. (11-5) of ACI 318-11 (2011). The ACI report on high strength concrete (ACI 363R 2010) states that with increasing concrete compressive strength, the actual concrete contribution in the shear resistance reveals lower values than the ones predicted through the more complex ACI equation for V _c due to the reduction in the aggregate interlock for HSC.

In the case where the concrete shear resistance was computed using the simplified Eq. (11-3) of the ACI 318-11 (2011), the theoretical concrete shear capacity of SCC becomes 50 kN compared to 48.2 kN for VC beams. The simplified equation has proven to be the best fit equation for the estimation of the SCC shear capacity and can be considered as reliable in the simulation of the behavior of SCC beams in shear. The adoption of Eq. (11-3) appears to conservatively cover the effect of the reduction in the aggregate interlock in SCC beams through providing theoretical estimations that are equal to the experimental values. In contrast, this same equation is not representative of the behavior of VC beams.

The loads at ultimate shear failure for the four shear beams, consisting of the combined concrete and transverse reinforcement shear capacity, are listed in Table 8. The load–deflection curves shown in Fig. 26 indicate similar load–deflection history for the VC and SCC beams.

The results of the shear failure of SCC and VC beams revealed an equal average ultimate shear capacity of approximately 120 kN in both beam types. The theoretical ultimate shear capacity was found to be equivalent to 93 kN in SCC beams and 91 kN in VC beams.

Crack patterns were very similar for replicate identical shear beams and were very similar for the VC and SCC beams. The diagonal crack widths constant for identical shear and flexural beams indicate the consistency of the test results.

These outcomes refute the hypothesis statement declaring that the high consistency of SCC will negatively affect the shear strength of SCC members. Also, the effect of the aggregate interlock on the concrete shear capacity was found to be more pronounced in VC beams than its is in SCC beams. This difference can be associated to the enhanced hydration of cement and the improved cohesiveness of the concrete mix.

Test results of the flexural beams are listed in Table 9. The results include the load at first diagonal crack, the load at yielding, and the ultimate load at failure. Also listed in Table 9, are the number of cracks and the maximum width of flexural and shear cracks.

The cracked beam sketches corresponding to SCC and VC shear beams are shown in Figs. 27–30.

The number of cracks and the crack width measurements disclosed similar values for the VC and SCC beams.

The loads at yielding and at ultimate for the four flexural beams are listed in Table 9.

The average yielding load for the two replicate beams was similar for SCC (P _y  = 99.5 kN) and VC (P _y  = 91.5 kN).

It was noticeable that the flexural crack height was greater in the SCC and VC beams than it was predicted using cracked section analysis. According to the ACI report on high strength concrete (ACI Committee 363R 2010), this behavior can be foreseen in HSC beams where shallower compression zones are required to maintain equilibrium in flexure.

In reference to Eqs. (7)–(12) the ductility index was taken as the ratio of the deflection at failure to the deflection at the load producing reinforcement yielding.

Beyond the yield load, the flexural beams exhibited a shear mode of failure with an average maximum load in SCC beams of 154 and 156 kN in VC beams. These results confirm the findings of the shear beam experimentation analysis.

The load–deflection curves shown in Fig. 31, indicate very similar load–deflection history for the VC and SCC flexure beams. Based on the displacement ductility index, SCC and VC beams have also shown similar ductility in bending.