PSI - Issue 11

Petr Stepanek et al. / Procedia Structural Integrity 11 (2018) 12–19 Petr Stepanek at al. / Structural Integrity Procedia 00 (2018) 000–000

18

7

to the cyclic load and after execution of 100 cycles. No reduction of the modulus of elasticity was observed for any of the tested samples (on the contrary, there was a slight increase at all levels). It can be assumed that the loading did not apparently result in a failure of the fibres, which would result in a decrease in stiffness. All samples were then divided into 6 pieces of 60 mm for the inter-laminar shear test. After 100 cycles at all tested levels of cyclic loading, there is already a significant degradation of the specimens (see Table 4). Table 4. Comparison of mechanical properties before and after 100 cycles.

Modulus of elasticity before the cyclic test [GPa]

Average strength in interlaminar shear at monotonic test [MPa]

Average strength in interlaminar shear after 100 cycles [MPa]

Modulus of elasticity after 100 cycles [GPa]

Load level

40% f ult 50% f ult 60% f ult

53.43 52.71 52.51

53.65 (+ 0.41%) 53.48 (+ 1.46 %) 53.19 (+ 1.30 %)

60.26 (- 4.83%) 59.80 (- 5.89%) 57.5 (- 10.03%)

63.32

The effect of a stress range is determined at three load levels with the adjustment of the maximum f max and f ult minimum stress in the cycle while maintaining the stress ratio R= 0,1. After plotting, it is obvious that the logarithm of fatigue life of the GFRP reinforcement increases linearly with decreasing maximum stress. The resulting S-N dependence can be expressed for bare bars according to equation (1) and for bars encased in concrete, according to equation (2). Obviously, these relations are valid for the tested GFRP reinforcement and the configuration of the experiment f max / f ult = 1.123 – 0.182 log N, (1) f max / f ult = 1.026 – 0.135 log N. (2) From Fig. 6 and the equation (1) and (2), it is clear that the fatigue life of the samples encased in concrete is higher in the entire load range tested and the difference increases with a decreasing load level. In the macroscopic scale, in the majority of cases, the encased samples did not show signs of failure of the bar surface caused by friction against the surrounding concrete (but this must be further refined by observing the meso and micro levels). In addition, the surrounding concrete prevented the reinforcement from transverse expanding during failure. The resulting fatigue life of bare bars is likely to significantly affected (reduced) by the anchoring method. This is particularly noticeable at 40% f ult, load levels. At this load level, at same samples undesirable damage occurred directly in the anchor, or longitudinal cracks reaching all the way to the anchor occurred. However, results are likely to be affected at all levels tested. 4. Conclusions The influence of the alkaline environment on the tensile strength of GFRP reinforcement was demonstrated. With increasing temperature and exposure time the tensile strength decreases. This behaviour is expected and in accordance with the results of other studies (Robert at al. (2010)). The temperature and alkalinity of the environment does not significantly affect the value of the elasticity modulus of the GFRP reinforcement. However, the observed degradation of the surface layer of the specimens is very important, as during the 365 days long immersion in the 60 °C solution the adhesive layer of silica sand is lost. Thus, ensuring the good bond of such degraded specimens with concrete is highly problematic and, therefore, it should be given increased attention. The behaviour of GFRP bars subjected to fatigue loading was presented. S-N curves were constructed, which were then compared with other available results. Throughout the tested range, the fatigue life of axially loaded bars encased in concrete is higher than that of the bare bars. This is inconsistent with the findings of the study by Adimi & al. (2000), in which the impact of friction between the reinforcement and concrete was quite apparent. For a complete description of fatigue life, it is necessary to extend the S-N curve with additional (lower) loading levels up to a fatigue life of 2 million cycles of encased samples. In the case of bare samples, new fatigue tests across the S-N curve range will be performed after the anchor have been modified. As soon as after the first 100 cycles, all tested levels show the obvious failure of matrix or matrix/fibre interface