Discussion of Test Procedure
In comparison to panel testing, only a small amount of fibres is oriented in the main stress direction, when using beams. It is therefore necessary to use beams with a large cross section to reduce scatter in results. The span of the beams should not be too small, otherwise you test vault effects instead of real bending.
Notched beams, using centre point loading, have the advantage of direct correlation between load and crack mouth opening displacement, but give higher results (up to 15 %) than tests on unnotched beams in four-point loading. Furthermore the cross section is reduced compared to unnotched beams. As the creep is assessed as a percentage of load at a given deformation, the creep load is effected, too. For practical reasons (stability) many creep tests on notched beams have been performed in four-point loading.
Bending tests on steel fibre reinforced beams often result in different shape of the load-deformation graphs in comparison to beams with synthetic fibres. Both rise and reach a peak when the “first” crack occurs. The graphs of the samples with steel fibres then fall steadily, when the fibres are pulled out. Opposite to this behaviour, samples with synthetic fibres fall sharp before they recover. They reach a second peak before they start falling again (Fig. 2). The deflection, at which they hit a low, often can be 0.1—0.6 mm (tests with 450 mm span). This is in the area of deflection, many creep tests use as starting point. The post crack parameters used for design may be nearly the same for both materials. But, when one compares the behaviour of beams with different fibres, a creep load of the percentage of the residual load at defined deflection is used. The load of beams with steel fibres is much higher in this area of deflection than that of beams with synthetic fibres. The relative elongation of the synthetic fibres is quite low, they are at this period of the test not loaded to the same extend as the much stiffer steel fibres. This may result in misleading conclusions.
In flexural creep test a mixed deformation is monitored resulting from deformations in tensile and compression zones of the cross section. As the creep deformation
Fig. 2 Typical load-deflection graphs of FRC with steel fibres and synthetic fibres
measured by crack mouth opening or net deflection is a mixture of creep and shrinkage of the concrete and the creep and slip out of the fibres, additional testing or calculations for these properties would be necessary, but are often not available. The environment during testing is mostly kept at ordinary lab conditions, some tests have been done at elevated temperature; other tests cover the beams by aluminium sheets to prevent evaporation. Temperatures higher than 40 °C will clearly increase the deformations, when synthetic fibres are used. The slip of all fibres will be influenced, when the humidity of the concrete matrix changes between water saturated and completely dry conditions.
As quite small deformations are considered, it is of importance that the supports are allowing these deformations with as little friction as possible. As many test rigs are produced as low cost prototypes, these requirements are often not fulfilled completely.
Due to limited access to test rigs and limited space in climate chambers, often several beams are loaded together and the load level is increased stepwise during time. This results in imprecise load levels and lack of data for creep models, which need data from uniform loading at different load levels. Due to lack of time, many tests end too early.