Stress analysis of fatigue cracks in mechanically fastened joints: An analytical and experimental investigation

The two historical fuselage failures, Comet in 1954 and Aloha in 1988, illustrate that similar accidents must be avoided which requires a profound understanding of the fatigue mechanisms involved, including analytical models to predict the fatigue behavior of riveted joints of a fuselage structure. The scope of the research project covers a variety of joint types and joining techniques for both monolithic and laminated sheet materials. The fuselage structure is a rather complicated system of parts consisting of skin sheets, tear straps, stringers, frames and doublers. These parts are interconnected by mechanically fastened and bonded joints, or a combination of both. The complex fuselage structure in the present research is reduced to specimen level size for laboratory testing and theoretical analysis.\par The major topics of the thesis are: Calculations of the combined tension and bending stress distribution in joints, which implies an extension of the so-called secondary bending model (Chapter 3). Find a direct and simple relation between the formed rivet head and squeeze force. (Chapter 4). Development of stress intensity factors for fatigue cracks in joints loaded under combined tension and application to fatigue crack growth results (Chapter 5). Fractographic observations with the scanning electron microscope of crack front shapes occurring in riveted joints under combined tension and bending (Chapter 5). Analysis of the residual strength of joints with fatigue cracks (Chapter 6). Neutral line model For mechanically fastened lap splice and butt joints in a fuselage structure, a dominant load is introduced by the Ground Air Ground (GAG) pressurization cycle. The hoop load is transferred from one skin panel to the next via the fasteners in the joint. The hoop load is offset by eccentricities in the load path, which causes secondary bending. The bending stress is a non-linear function of the applied tension load. The stress system in the joint then encompasses the membrane stress, the secondary bending stress and the bearing stress associated with the fastener loads on the holes. The secondary bending is highly depending on the magnitude of the eccentricity and the flexural rigidity of the joint between the fastener rows. The theory used to derive the bending stresses is based on the advanced beam theory. A further development of the neutral line model incorporates the internal moment, which is a useful representation of the load transfer occurring in multiple row joints. The calculation of the load transfer can be made for complicated lap splice and butt joints. With the developments of the present research, the neutral line model is still a very powerful tool to use in the early stages of joint design. It gives a good picture of the stresses in a joint. Riveting Solid rivets and more advanced fasteners are still widely used in aircraft fuselage design efforts. The fasteners are characterized by various parameters associated with the fastener material and geometry, sheet material and installation process. The present investigation focuses on solid rivets installed in aluminum and Glare. The expansion of a solid rivet in a rivet hole is important with respect to the fatigue properties of joints. The expanding rivet inside a fastener hole will create a compressive residual stress around the hole and this will delay fatigue crack nucleation. It is important to know the correct squeeze force used to form the driven head of a rivet. Measurements of the formed rivet head (diameter or protruding height) can be used to obtain information about the applied squeezing force. The riveting process is a non linear deformation process characterized by large plastic strains. Simple equations based on constant volume of the rivet and the Holloman model for uniform plastic deformation, were adopted to evaluate the riveting process. Useful results were obtained about the correlation between the rivet head deformation and the applied squeezing force. Stress intensity factors As a result of combined tension and secondary bending in a lap joint, fatigue cracks at the edge of a hole start at one side of the sheet only. Initially these cracks at the edge of a hole are growing as a part through the thickness corner crack, which later become a through the thickness crack, a so called through crack. But also for a through crack, the shape of the crack front is usually curved and the crack length measured at both sides of the sheet will be different. In view of fatigue crack growth predictions it then is necessary to obtain stress intensity factors for such slant and curved crack fronts. In the present investigation this problem has been explored for a simple configuration, which is an open hole in a sheet specimen subjected to combined tension and bending. Fatigue tests were carried out on specimens of AL 2024 T3 clad sheet material with three different thicknesses (1.0, 1.6 and 2.0 mm). In each specimen a single open countersunk hole was present. The development of the crack front is these specimens could be recorded because so-called marker load cycles were applied in these tests. It then was possible to observe the crack fronts in the scanning electron microscope, which still was a rather strenuous work. Reconstruction of the crack growth could be done for the larger part of the fatigue cracks.\par K-values were obtained for a large variety of crack front shapes and crack sizes. Comparison of the new calculated K values with existing solutions showed that the new solutions capture near the surface phenomenon more accurately than the previously published data. The improvement is a result of using an increased mesh density. For through the thickness cracks growing away from the countersunk hole, the normalized stress intensity factors approach the values of the normalized stress intensity factors for cracks emanating from a straight shank hole. Thus, the effect of the countersunk hole decreases with increasing crack length. The solutions for the pin loading b values show a dominant influence of the countersunk shape in the b/t values.\par Residual strength Static failure of a joint occurs when that joint is not able to carry the applied load anymore. The type of static failure in joints depends on the loading condition and the joint configuration. The most common static failure modes in monolithic aluminum joints are fastener shear failure, plate tension failure, bearing failure and plate shear failure. In Glare joints another failure mechanism, fastener pull-through, is often observed. This failure mechanism is related to the lower stiffness of Glare in thickness direction, leading to increased tilting of the fasteners and hence increased tensile stresses in the fastener. \par In the present thesis, a method is proposed to calculate the residual strength of joints of monolithic and fiber metal laminates. The method uses the remaining net section. For the fiber metal laminates the net section includes the remaining intact metal layers in combination with the intact fibers. The method starts with the blunt notch strength of the un cracked joint and the metal volume fraction for the fiber-metal laminates. The Norris failure criterion and the metal volume fraction are used to calculate the blunt notch strength for any possible Glare lay up. Secondary bending has a significant influence on the ultimate strength of both Glare and aluminum. The ultimate tensile strength reduces with increasing bending. Taking this into account, an empirically found reduction of 10% of the blunt notch values results in a more accurate representation of the stress system. Difficulties arise if significant plastic deformation occurs at the most critical fastener row, and further research using a finite element model is recommended