Uni-Axial Tensile Creep Tests
The individual results of the uni-axial tensile creep tests are shown in Fig. 9 together with the average responses. The total time-dependant crack widening after 240 days is shown in Fig. 10. A similar trend can be seen as for the single fibre pull-out creep tests, i.e. the crack widening being proportional to the load level to at
Fig. 9 The individual results of the uni-axial tensile creep tests (a), with (b), the average responses
Fig. 10 The total time-dependant crack widening after 240 days for the uni-axial tensile creep tests least to the 50 % load level. It is also interesting to note that the time-dependant crack widening is around double the single fibre pull-out creep, which is to be expected as in the uni-axial tensile creep, the fibres creep from both sides. This confirms the mechanism of creep in the single fibre pull-out creep tests is similar to the uni-axial tensile creep tests.
Pre-Damaged Single Fibre Pull-Out Creep Tests
In order to look at the effect of the crack formation, which causes some initial fibre pull-out, on the single fibre pull-out tests, samples were pull-out out to 0.53 mm before commencing of the single fibre pull-out creep tests at a 50 % load level. The average responses of both the virgin 50 % tests and the pre-damaged creep tests are shown in Fig. 11, in both linear and logarithmic scales.
It is clear that the total pull-out creep is the same, whether the fibre was pre-damaged or not. However, it can be seen on Fig. 11b that the responses deviates substantially up to about 30 days. This can be explained using the postulation that a ITZ causes a significant part of the creep. For the pre-damaged fibre, the ITZ was crushed during the pre-test damage/pull-out and thus prevented the initial creep as was observed by the undamaged specimens. After time, this became insignificant and the results of the two sets of samples converged.
Fig. 11 The average responses of the single fibre pull-out results of the 50 % load level for both the virgin and the pre-damaged samples (a), and (b), the same results on a log scale