Multiple site damage of aeronautical riveted joints

Among all aeronautical structures prone to develop MSD, riveted lap joints in the fuselage have been identified as being the most susceptible. Recent recommendations by regulators to avoid MSD threat stipulate an Inspection Starting Point (ISP) and a Structural Modification Point (SMP) in the life of aircraft. These points can be defined in terms of MSD analysis and the capability to accurately calculate service life to MSD onset becomes of considerable importance. To investigate this failure mode, a probabilistic model for MSD assessment considering both fatigue crack initiation and crack propagation as random variables was proposed. The fatigue crack initiation stage of the model differed from other published models for incorporating continuing damage assumption instead of a damage accumulation technique for re-initiation of fatigue cracks at crack-free fatigue critical locations. The crack propagation stage of the model was firstly performed deterministically by means of a dual boundary element computer code; and, secondly, probabilistic crack growth treatment was incorporated in a simple way. The results from the MSD assessment model provided good agreement with published experimental work on fatigue of lap-splice joints and with other model for the same geometrical configuration. In order to fulfil a lack from the literature, the probabilistic MSD model was employed to investigate variables influencing MSD. The variables were related to possible changes in the standard deviation for fatigue crack initiation, uniform and non-uniform hole pin-loadings, nominal stress level and high rivet squeeze force. The parametric study showed that the ISP and the SMP can be considerably affected. Further results include the identification of a conflict between two different structural safety criteria and the proposition of a new one; the use of uniform pin-loading distribution with peak loads from a non-uniform pin-loading distribution was suggested to avoid non-conservatism specially at low cumulative probabilities of failure; a clear tendency to decrease the mean time for the lead crack propagation as the number of MSD-like scenarios increased was verified, but not always a bigger number of nucleated cracks per damage scenario gave the smallest time for crack propagation and the crack nucleation sequence was more important than the number of nucleated cracks per damage scenario. To investigate the effects of high rivet squeeze on MSD, an experimental work was carried out to obtain input S-N fatigue data. Recent findings from the literature established the benefits that high squeeze force can provide for the mean time to fatigue crack initiation, but no probabilistic analysis was undertaken comparing different squeeze force values. The results of such analysis reviewed that the whole MSD failure process was retarded and the number of MSD-like scenarios considerably reduced, demonstrating that high rivet squeeze force is extremely beneficial for MSD prevention. Finally, the probabilistic model was employed for comparison to in-service MSD data from pressurized fuselage panels. During the preliminary modelling stage, it was found clear evidences that S-N input data obtained from good quality riveted test specimens could not be used for MSD assessment of real pressurized fuselage panels, and it was assumed open hole quality input S-N fatigue data. The results demonstrated to be rationally conservative compared to in-service findings. In-service data indicated that both the ISP and the SMP were well established from the simulations, and failure due to MSD occurrence would not threaten structural safety during the monitoring period.