FATIGUE CHARACTERIZATION OF BITUMINOUS MIXTURES USING THE FOUR-POINT BENDING TEST

This thesis aims to define a proper reliable procedure to characterize the fatigue resistance of bituminous mixtures according to European standards. This is achieved by using the four-point bending beam test device under strain-controlled sinusoidal loading configuration at different strain levels, with a reference frequency of 10 Hz, and a constant temperature of 20 °C. By analysing the vast literature on the subject, five failure criteria were selected and compared to come up with the best path to follow for fatigue resistance characterization. The first two failure criteria are based on the modulus reduction: 50 % reduction in the initial modulus taken at the 100th cycle, and 50 % reduction in the modulus extrapolated from the phase of linear reduction of the modulus (Intermediate phase). The third failure criterion is the peak in the phase angle. The last two criteria are based on energy concepts, the first one is the energy ratio, and the second one is the dissipated energy approach. In consequence, five dissimilar fatigue curves have been produced, using three strain amplitude levels (140 µm/m, 200 µm/m, 300 µm/m) and a minimum of six replicates at each strain level. Some additional relationships have been introduced like the relationship between the visco-elastic-related modulus versus the void ratio of the beams, and fatigue-related modulus versus the void ratio. Also, the positions of the beams derived from the wheel compacted slab were drawn against the corresponding air voids. Statistical analysis is carried out, dealing with the expected non-homogeneity of the results. Consequently, a sensitive pavement design was carried out using KENLAYER non-linear analysis software. The design is based on a fixed pavement configuration, adopting the mechanistic empirical pavement design fatigue model (Transfer function). The purpose was to see how pavement design is affected by variability in fatigue resistance characterization.