PSI - Issue 19

Fabian Becker et al. / Procedia Structural Integrity 19 (2019) 645–654 F. Becker et al. / Structural Integrity Procedia 00 (2019) 000–000

651 7

Fig. 6. State of visible specimen surface on the tension side at 0.7 of normalized cycles to failure

the specimen close to the neutral phase may be introduced by shear stress fatigue. However, no split cracks formed in these specimens. Additionally, the specimens show a plastic deformation after the experiments, which could not be seen in the reference testing series or at other strain levels. This plastic deformation was also reported by Paepegem et al. (2005) for bending fatigue of ± 45 ◦ woven glass-epoxy and may support the role of shear stresses in the matrix for the accumulation of plastic deformation. The specimen with the lowest equivalent bending strain (Specimen number 7 in table 1) showed a delamination growth at both sides of the clamping on the tension side and a split crack visible on the compression side. None of the spec imens showed a compression failure prior to reaching the residual sti ff ness criterion. However, it is expected that the interaction of fiber parallel stress and shear stress may lead to a failure on the compression side, if the bolt loads were adjusted to higher values (Pinho et al. (2006)). For a first estimation of fatigue life, both data sets were fit with a Basquin equation (Figure 10). Other fits may be used with a higher number of data including static failure values (Colombo et al. (2011)). Compared to the reference S-N data, the curve of the 5 kN shows a steeper slope. Due to the introduction of new failure modes compared to the reference, all lives are significantly shorter. The fatigue life estimation of a composite leaf spring with S-N data from pure three point bending fatigue experiments clearly would overestimate the life of the component.

4. Conclusion

A clamping system has been developed to introduce additional transverse stresses on the middle section of a glass fiber reinforced plastics material under three point bending loading. The desired pressure distribution is introduced through two surfaces by tightening two bolts, which are positioned on the left and right hand side of the specimen and monitored via strain gauges in the shaft sections of the bolts. The whole system aims to reproduce the clamping situation in a longitudinal leaf spring. The load in the clamping section leads to a di ff erent damaging and failure