Co-authors​ :

 Koerwien THOMAS (GERMANY), Marco HOFFMANN , Benjamin GREBING (GERMANY), Gregor DIEZ  

The certification of structural bonding of CFRP structures is highly regulated. These regulations are a result of incidents involving failed bonded joints due to various reasons such as e.g. faulty material, bad processing or flawed design. Nowadays, certification of bonded joints require the limitation of crack growth of possible local damages through design features in order to maintain limit load capability, or a repeatable measurement of strength, or proof load testing up to the limit load of each bonded joint. The determination of the actual bondline strength is not within reach and proof load testing applicable for prototyping the way to certify a bonded joint is to use design features for damage containment of local damages. Current applications of bonding technology are typically used in stiffened panels as well as small bonded repairs. The latter is limited in size by the bonding repair size limit assuming that with a disbonded repair patch the surrounding structure still has limit load capability. In order to leverage the full potential of bonding technology a design and sizing methodology is required to design for damage tolerance. This involves the modeling of e.g. impact damages, virtual fatiguing and approximation of the residual strength. In order to validate the developed methodology a series of experimental tests have been carried out. This involves two types of tests, first the single lap tensile after impact (SLTAI) and second the wide single lap shear (WSLS) specimens. In these specimens local damages have been incurred. These were either barely visible impact damages (BVID) or local disbonds aided by release foils approximating the BVID damage size. Besides, the stacking sequence has been up for variations as well. In order to investigate arrested growth, rivets have been used as design features by default. The samples were tensile fatigued with periodic interruptions for non-destructive inspections to determine possible crack propagation. Finally the samples were statically tested to determine the residual strength.
The experimental data obtained give valuable input for the developed simulation method. Both specimen types featuring artificial damage proved very suitable to monitor crack growth under fatigue. For impact damage WSLS is more suitable. For SLTAI it proved very difficult to determine the appropriate impact energy. 1 J made the difference between almost complete failure and no damage at all. The clear influence of parameters like stacking sequence, damage type and design features on crack progression and failure modes will be discussed.