I didn’t post this earlier because I didn’t want to speculate, but the FAA/NTSB has this excellent page on the AA191 accident lessons learned, which goes into depth on the design of the DC-10/MD-11 engine 1 and 3 pylon:
To be honest I was a little surprised to learn how the DC-10/MD-11 pylon attachment works, because it’s a relatively uncommon load path compared to Boeing and Airbus pylons which use a more truss strut-braced style.
On the DC-10/MD-11 pylon, the forward attach fitting is a single integral frame/bulkhead that penetrates up into the wing box and picks up a sort of triangular shaped fitting that bolts into the forward wing spar. This presumably reacts vertical loads and side loads and moments. There looks to be a fwd-aft oriented linkage that presumably takes thrust loads from the engine, but it only shows up in one picture. The aft attach fitting is a lug and clevis style fitting, with the lug on the pylon side and the clevis on the wing side. The way it’s oriented, it would primarily be taking vertical load and reacting the pitching moment due to engine thrust along the spanwise axis of the wing.
Pylon attach cutaway:
View attachment 86905
Stand alone pylon structure:
View attachment 86906
Pylon attach details with a crudely drawn aircraft coordinate system. Looks to me like point C takes thrust load (force in the X-direction, Fx), A and B and their associated bolts on the wing side to the forward spar take forces in the y and z directions (Fy, Fz) and moments Mx (B probably takes most of the load with A being there to couple out Mx), and the aft attach fitting can take forces in the y and z directions (Fy, Fz) which working together with the forward fitting reacts moments My, Mz. My educated guess would be that aft fitting is primarily loaded in the Fz direction and couples out the My moment due to the thrust line of the engine offset from the wing:
View attachment 86907
Finally, the lessons learned document has this view of the aft lug and clevis attach fitting, noting that the forward and aft surfaces of the lug need to be inspected for cracks:
View attachment 86908
These forward and aft faces of the aft attach lug on the pylon side are exactly where the failure occurred on both the AA DC-10 and the UPS MD-11! To me, this seems a bit like lightning striking the same place twice…
Note that lug and clevis fittings are extremely common in aerospace on everything from control surface hinges to main landing gear trunnions. The lug side of the fitting gets a spherical bearing to allow for some misalignment in the assembly without transferring bending moments:
View attachment 86909
Note that simply pulling the bolt and putting a borescope or eddy current probe in the hole is NOT ENOUGH, because the spherical bearing is in the way. The two halves of the assembly must be physically separated to be able to fully inspect the faces of the lug for cracks or damage. This means unless the gaps were large enough (which I find highly doubtful), the pylon had to physically be removed from the airplane to fully inspect the lug surfaces. Further, there could be cracks in the bore between the spherical bearing and the lug that you would only be able to see by pressing out the spherical bearing and replacing it, which causes further opportunity to inadvertently damage the lug. It sounds like this is what happened with AA191 - they replaced the spherical bearing on the aft pylon attach lug, then accidentally caused damage to the freshly repaired lug while re-installing the engine/pylon with the forklift, and weren’t able to inspect the area and see the problem after installation.
Finally lugs fail in a couple common ways. The failure shown in the NTSB photos from the UPS accident is text book net-section (net-tension) failure, likely due to cracks that originated at the edge of the hole and worked their way to the edges of the lug.
View attachment 86911
I think
@Roger Roger is right on with his questions. Is this going to end up being another case of inadvertent damage caused during removal and replacement of the lug spherical bearing similar to AA191, or is this aft pylon attach lug truly undersized in MD/Boeing’s fatigue analysis, requiring a re-design of the part? Either way, I find it HIGHLY suspicious that the same part failed in exactly the same way in 1979 and 2025. Design for inspectability and maintainability are both important factors in addition to fatigue. Quoting
@knot4u for his past NDT inspection experience, it really looks to me like the upper faces of the lug sandwiched inside the wing clevis are in an uninspectable region, unless that gap is large enough to get a borescope or NDT probe in.
Given the extreme prevalence of lug and clevis designs in aerospace, this DC-10/MD-11 pylon design seems to be an outlier in a bad way, and I would be highly suspicious of all of them until we get more information on the status of the whole fleet. Thinking through some possible solutions, you could thicken the lug to increase the fatigue life, but this would cause a redesign of the clevis and a very costly retrofit of the wing structure to remove and replace the clevis fitting in addition to retrofitting the pylon. You could change to a higher strength material on the lug (from aluminum to titanium or steel for example), paying a weight penalty. This would require removal of the engine and pylon and retrofit of the pylon. Or you could increase the frequency of inspection, possibly resulting in an AD that removes the engine and pylon and potentially presses out the spherical bearing to inspect the bearing surface in the lug, with increasing frequency. None of this bodes well for the continued operation of the MD-11 fleet, but I guess we’ll wait and see.
