Today, we conclude this discussion. If you missed anything, simply look for part 1 and part 2.

6.) If you have fracture as the primary damage category, you now need to determine two things.
    A.) The loading geometry – tension, compression, bending, torsion, contact or direct shear
    B.) The material behavior (not inherent tendency, which comes later in many cases) unless deformation is obvious in the fragments. This is a choice of macro ductile or macro brittle.

7.) Since macro-ductile torsion (in a cylindrical shaft at least) will have identical fragment shapes to macro-brittle tension and also macro-brittle bending, you have to go deeper to make sure you know whether it’s macro brittle or ductile. So you’ll end up going in a circle of loading geometry versus material behavior.
    A.) If it’s ductile torsion, there MAY be smear marks on the fracture surface, or you may be able to detect some limited deformation of splines or longitudinal grinding marks near the crack edge.
    B.) If it’s brittle tension, you may be able to detect a fan pattern or ratchet marks pointing back to the initiation. Remember that fatigue is always macro brittle.
    C.) In the case of macro-ductile bending, the crack will be 45 degrees to the flat plane that macro-brittle bending would have created. The 45 degrees can be in either of two orientations. The Charpy coupons of Figure 1 have macro brittle in the center and macro ductile “shear lips” along the sides. The crack direction is going from top to bottom on the upper portion of each fragment pair. Note that the macro-ductile features (shear lips) are more prominent on the part that was tested at a higher temperature.
 

The round bar of Figure 2 was saw-cut when soft, hardened, then broken off to have a convenient test coupon. Even though the part was hard, the crack was macro ductile. Figure 3 shows a paired tensile test (axial tension, macro ductile along the cup edges, probably macro brittle in the center, but micro ductile in that same flat center).

8. ) Note also that stress states are different from loading geometry.
 

For a coil spring with hooked ends, intended to be loaded by pulling on both ends, the spring itself is loaded in tension. However, the part experiences torsion in the coils, and at the end of the coiled area as it blends into the hook we have torsion plus bending.

So, I don’t know what to tell you about what to call this ... THE COMPONENT LOADING IS TENSILE, but there’s nowhere on the PART that experiences TENSILE LOADS. There are tensile stresses due to the torsion and loading! You have to look at the situation and figure it out. In a way, all categories are artificial!