• Ingen resultater fundet

Chapter 3 3.4 Discussion

3.4 Discussion Chapter 3 Paper II

3.4.2 CMOD

The correlation between the initiation and propagation of the main bending crack and the applied load, was well reproduced by the numerical model, cf. Figure 3.10. In general, the model underestimated the CMOD for a given load level. Comparing the experimental results for the three concrete compositions, it is seen that the crack was initiated for approx the same load level, viz. approx 15-18 KN. The numerical model was capable of fitting the crack initiation and the crack width increase as a function of the applied load for all the concrete compositions. Further, the reduction in CMOD for the SFRC specimens compared to the PC specimen at the same load seen from the experimental observations was captured by the numerical model.

It is underlined that any differences between the numerical simulations and the exper-imental results may be caused by in-accuracies in the fracture mechanical properties of the concrete given in Table 3.1 and Figure 3.6. Those values were determined in-dependently, from experimental observations, as previously described, and were not subject to fitting within the present numerical model.

3.4.3 Separation

The formation of separation at the concrete/steel interface 10 mm from the main bend-ing crack is seen from Figure 3.11. The relationship between the applied load and the slip observed from the numerical simulations was reproduced accurately by the nu-merical model for all three concrete compositions. However, higher deviations be-tween the numerical simulations and the experimental results were observed for the separation-load relationship for 0.5 vol.-% SFRC.

Comparing the experimental results for the CMOD-load and the separation-load rela-tionships for each concrete composition, it is seen that separation at the concrete/steel interface did not take place before the main bending crack was initiated. It is also seen from the same figure, that the numerical model simulated this phenomenon very well, although this relation had not been set as a prerequisite in the numerical model. This indicates that the numerical model reproduced the load-induced debonding formation physically correct. Moreover, the experimental and the numerical results indicated that the separation at the concrete/steel interface is controlled by the deformation of the specimen caused by the applied load. This is concluded from the observation that the initiation of the separation at the concrete/steel interface was only marginally af-fected by the addition of fibres.

Chapter 3 3.4 Discussion Paper II

3.4.4 Slip

The slip formation at the concrete/steel interface 10 mm from the main bending crack is shown in Figure 3.12. It is clear that the numerical simulations captured the exper-imental observations very well, though the relationship was slightly overestimated for PC, cf. Figure 3.12. The scatter in the experimental observations for 0.5 vol.-% SFRC is most likely due to the limited resolution of the experimental optical/digital tech-nique. Moreover, it is seen by comparisons of the experimental results concerning the slip at the concrete/steel interface and the CMOD-load relationships, that slip did not take place at the concrete/steel interface until the main bending crack was initiated.

The numerical model was capable of simulating the on-set of slip at the concrete/steel interface as well as the slip-propagation due to increased load applied.

3.4.5 Extent of Separation and Slip

It is seen from Figure 3.13a that the separation was not initiated for the same load level applied to the three concrete compositions, and that the separation propagates at approximately the same rate along the reinforcement in 0.5 % SFRC and 1.0

vol.-% SFRC whereas the separation rate, ie the growth of separation as function of the applied load, along the reinforcement was higher for PC. It can be assumed, based on those observations that separation at the concrete/steel interface did not initiate before the main bending crack reached the reinforcement, which was in agreement with the experimental results concerning the separation 10 mm from the main bending crack.

From the results presented in Figure 3.13a it may be concluded that the addition of fibres lead to a decrease in the separation rate along the reinforement. From Figure 3.13b it is seen that the addition of fibers results in a reduced separation for a given CMOD, and that the growth of the separation-length, is less steep for SFRC than for PC. However, it is noted, that contrary to expectation, the curve corresponding to 0.5 vol.-% SFRC is lower than the curve corresponding to 1.0 vol.-%.

The analyses concerning the length of slip along the concrete/steel interface as a function of the load revealed that the slip was initiated for the same load level for SFRC (0.5 vol.-% and 1.0 vol.-%) whereas the load level required to form slip in PC was slightly lower, see Figure 3.14b. Once the slip at the concrete/steel was initiated the propagation of the slip, as function of the load, along the reinforcement was lower for SFRC compared to PC. That may indicate that the addition of fibres prevented the slip formation at the interface. With regard to the slip-length along the concrete/steel interface as a function of the CMOD, see Figure 3.14b, it is seen that slip was initiated for the same CMOD for all three concrete compositions. However, the development of the slip-length for 0.5 vol.-% SFRC is much steeper than the curves for 1.0 vol.-%

SFRC and PC. Considering previously presented observations showing correlation between experimental observations and numerical simulations it may be concluded that the numerical model is physically-correct based. Therefore, the observations

3.5 Conclusions Chapter 3