UHPFRC Material Behaviour
The concrete mix used for this investigation is indicated in the Table 2. This concrete has a compression resistance between 150 and 190 Mpa. The steel fibres are manufactured by the company Krampeharex and are straight brass alloy coated steel wires. The water-cement ratio, under consideration of the water in the silica suspension, is 0.265.
In a compression test, the material has a quite elastic linear behaviour up to failure. The addition of fibre do not influence the load peak. The most common tests to characterize the tensile strength and the post cracking behaviour are the three and four point bending tests and axial tension tests with notched and unnotched specimens . The difference between the three and four point bending tests is presence of a notch in the three point test specimens, which fixes the point where the crack appears. Having only one crack in a defined position allows easier monitoring the width over the time.
Figure 2 shows reference tests on the adopted concrete with three different setups, all performed according the AFGC French guideline for UHPFRC . In general, the diagram starts with a linear elastic part, where the uncracked concrete matrix resists to the tensile forces, almost without influence of the fibres. With the formation of the first crack the tensile load is transferred from the concrete to the
Fig. 2 Reference tests, from left, 3 point bending, four point bending and centric tension tests
Fig. 3 Aramis image analysis, four point bending (left) and axial tensile test (right)
fibres that have to bridge the tensile forces over the cracked cross section. After the formation of the first crack, the applied load can be further increased. The tensile resistance of the concrete is exceeded then in several cross sections. The results is a multiple fine crack scenario difficult to observe with naked eye. In general, after the peak load, the failure occurs with crack localization: only one crack becomes larger and larger and the specimen slowly collapses.
To investigate in detail the evolution of the damage and the cracks progression in the cross section the optical measurement system ARAMIS was adopted. Figure 3 shows the analysis of the specimens surface with this system. The bending test, in comparison with the axial tensile test, shows a larger increase of tension load after the first crack up to failure and a larger number of cracks. These effects are caused by the internal force redistribution in the cross section and by the progressive activation of the fibres during the crack opening process. In the three point bending test, instead of the vertical displacement, the crack mouth opening displacement (CMOD) is measured with a clip gage. In this test only one crack appears. This observation allows to correlate the deformation of the cracked cross section with the internal stresses.
Fig. 4 Preliminary tests, vertical displacement on four point bending test
Preliminary Sustained Load Tests
Preliminary tests in a universal testing machine were performed in order to investigate the sensitivity of the material to tensile creep. A first result is the order of magnitude of the displacements. Specimens for the four point bending test according to  with a cross section of 70 x 70 mm and span of 210 mm were subjected to sustained load over 27 h. For the pre-damaging, the specimen were first loaded up to a deflection of 0.6 mm, as described above. This procedure provide the specimen the required crack pattern. Specimens were then unloaded and reloaded up to 95, 90 and 80 % of the load reached with the imposed deflection. In Fig. 4 the vertical displacement during the sustained load is shown. Since the time of observation was only few hours, a stabilization of the vertical displacement was not observed. Higher applied load results in level higher level of displacements. The tensile creep seems to be non-linear for such high load levels.