Fatigue tests on ‘double shear’ riveted specimens were carried out. Material was aluminium alloy, 7075-T73, thickness 3 mm. Specimen surfaces were protected by chromic acid anodization and epoxy primer. The mating surfaces of some specimens were sealed by a two-component manganese-dioxide cured, polysulfide compound. Specimens were pre-assembled by inserting a temporary spring fastener in a pilot hole. The holes were reamed to the final diameter after sealant curing; then the Hi-Loks were installed. Comparative fatigue tests were carried out on dry assembled specimens. Quite surprisingly, the fatigue resistance of sealed specimens was very low when compared with the results of dry assembled specimens. Hysteresis cycles measured in sealed and un-sealed specimens clearly indicated a lubricant effect of the sealing layer, which justified the results obtained. Additional tests were performed on specimens assembled by applying different clamping force during the sealant curing time. The results obtained indicated an independence of the fatigue resistance on this parameter. Different failure modes were observed in sealed and un-sealed specimens: fatigue cracks nucleated outside the hole at a location in the shadow of the rivet in the un-sealed joints, while fatigue cracks nucleated at both sides of the holes in sealed joints. Finite element calculations carried out under different values of the friction coefficient between the mating surfaces confirmed the different location of the fatigue critical areas in sealed and un-sealed specimens.
Read full abstract