Mixed Mode Fracture

When considering friction stir welding (FSW) as a replacement joining process in aerospace structures, residual strength assessment of FSW components is an important aspect of the decision-making process. Since most aerospace joints will be subjected to a combination of tension and shear loading, the response of a flawed, FSW joint to mixed mode loading conditions is needed. To address this issue, a comprehensive, mixed mode I/II test program has been initiated.
In this work, initial results from our mixed mode test program are reported. Using a large Arcan test fixture to obtain a wide range of mixed mode I/II conditions, fatigue pre-cracked FSW joints in 6.35 mm thick, 2024-T3 aluminum plate materials were tested for several combinations of mixed mode loading. Results from these tests are compared to companion, mixed-mode, crack growth data from fatigue pre-cracked base material specimens to determine the effect of the FSW process on the crack growth response of the material.
Data from the experimental program indicates that (a) obtaining initial fatigue crack growth in the FSW specimens was difficult, most likely due to the effects of residual stress and heterogeneous material behavior, (b) base metal fatigue cracks grew in the same manner that has been observed for 2.3 mm thick specimens (kinking away from the initial crack line for predominantly mode I loading, straight shear crack growth for high mode II loading), (c) mode I loading of FSW cracks resulted in an apparent minimum in the residual strength, (d) under all levels of mixed-mode loading, FSW cracks tended to grow across the FSW (towards the HAZ of the FSW) before re-kinking and growing along the original crack line direction.


Preliminary Studies of Mixed Mode Fracture in 2024-T3 Friction Stir Welds

Mixed-mode I/II test fixtures and specimen

Loading and relationship to Mode I and
Mode II directions

Mixed mode specimen undergoing Mode I loading and crack growth. To stabilize the crack growth process, an MTS clip gage was attached to specimen and used for displacement control of the loading process process.

All flaws were edge cracks. Coordinate system was positioned at crack front on the edge of the specimen on the crown side. The X-direction is along the crack. The Z-direction is perpendicular to the specimen and oriented outward. The Y-direction is perpendicular to the crack.




Uniaxial stress strain curves for base metal and center of FSW





MODE I LOADING (0°)

Crack growth direction was similar for both base material and FSW under Mode I loading. The direction was nearly straight ahead in both cases. The fracture surface for base material was uniformly void growth. The fracture surface for FSW was highly unusual. There was a periodic array of surface protrusions on the fracture surface at a spacing of approximately 2mm, indicating that fracture involved the separation of material at a clearly defined and nearly uniform spacing within the FSW. This is most likely due to the joining process.





Crack Growth Process, FSW, Mode I Loading

Strain ahead of crack caused an apparent fracture or debonding of specific extruded flow fingers, which provided a path of least resistance for crack extension. This process continued throughout the fracture process and the crack appeared to extend one extrusion finger at a time. Thus, the damage process in the friction stir weld appears to be is related to a combination of factors; delamination/debonding of extruded metal along interfaces; void growth; crack extension. Results also indicate that the Mode I crack growth occurred under the lowest loading. Given the lower load (and KI) for crack growth in FSW, it appears that bonding of the extruded material is the weak link in the fracture process.


MIXED MODE LOADING (30°)

Base metal crack growth is consistent with previous observations in 2024-T3 aluminum. Kinking occurs to maximize Mode I COD . FSW crack path shows rapid, sharp kinking away from center of weld towards HAZ and pin diameter region. Nearly flat fracture in middle, some shear lip formation along root with much smaller shear lips or no shear lips along crown region. Fracture at approximately the pin diameter location initially has minimal appearance of microstructural features. Reasons could include fact that this location is just outside of swirl processing zone. Later, this changes to a pattern similar to that observed in Mode I fracture with clear periodic features on the fracture surface.





Initial Crack Growth Process, FSW Mixed Mode Loading

Process seems to be quite similar to that observed for Mode I. Strain ahead of crack caused an apparent fracture or debonding of specific extruded flow fingers, which provided a path of least resistance for crack extension. This process continued throughout the fracture process and the crack appeared to extend one extrusion finger at a time. Thus, the damage process appears to be is related to a combination of factors; delamination/debonding of extruded metal along interfaces; void growth; crack extension.


MIXED MODE LOADING (60°)

Base metal fracture is typical of what was observed in 2.3 mm thick 2024-T3 aluminum, with kinking along direction of maximum COD. FSW has shear lip on root side and crown side. The center is relatively flat, with muted microstructural features and no obvious vertically debonded extruded material. At this location in the FSW, which is outside of the pin diamater but inside of the shoulder diameter, it seems to be easier to fracture the process-induced features than debond them. Note that there is a dramatic shift back to the center of the FSW (see end of fracture surface in picture). This occurred immediately after we rotated the specimen and applied Mode I loading to the specimen.





Crack Path and Microstructure

The crack paths for 30° and 60° paths are approaching the interface region between base metal and weld. However, most if not all of the fracture surfaces are still in the finer microstructure region of FSW nugget and HAZ.





Conclusions


Student: Ning Yuan
Advisor: Dr. Michael A. Sutton