Can a Torqued Bolt Become Loose?

The answer to this question is, "Yes it can." But we need to qualify this answer. First lets eliminate the possibility of a nut rotating, thus loosening the torque. So now we are left with “how else can a torqued bolt become loose?”. OK here is one way that a bolt, after it has been properly torqued, can become loose. This phenomenon is call “embedment relaxation”.

Embedment relaxation is a phenomenon in mechanical engineering in which the surfaces between mechanical members of a loaded joint embed. It can lead to failure by fatigue and is of particular concern when considering the design of critical fastener joints such as in the through-bolts in the crankcase supporting the engine main bearings.

In critical fastener joints, embedment can mean loss of preload or clamping force. Flattening of a surface allows the strain (or preload) of a bolt to relax, which in turn correlates with a loss in tension and thus preload. Therefore, embedment can lead directly to loosening of a fastener joint and subsequent failure.
In bolted joints, most of the embedment occurs during torquing. Only embedment that occurs after installation can cause a loss of preload, and values of up to 0.0005 inches can be seen at each surface mate, as reported by SAE.

As each surface is pressed together, the high spots are crushed and deformed to form a surface capable of supporting the load.

Surface Finish:
There are multiple surfaces in the joint:

1. between the head of the bolt and the washer,
2. between the bolt washer and faying surface (faying surface is the surface of the object being fastened together),
3. between the two faying surfaces,
4. between the faying surface and nut washer, and
5. between the nut washer and the nut face.

Each of these surfaces squeeze together. Any paint, sealants, nicks, or alignment errors are gradually crushed down to support the load. Bolt threads also embed. Threads are pulled in shear, slightly increasing the thread pitch. Nut threads are compressed and lose a little pitch. As embedment occurs, the surfaces press further together and reduce the bolt's clamping force (preload). Note that relaxation occurs without any off-rotation of the nut.

Under optimum joint conditions in a lab one can expect between 1 and 11 percent. Obviously then, in the field, more clamping force can be lost. Lockheed did a lab test, Report No. LR 25049 where they tested 1 inch by 12 UNJF thread size L-1101 engine pylon bolts. They used lab conditions, with hardened steel bushings. In a static joint with no load fluctuations, between 1 and 11% preload was lost. The greatest loss occurred in the first eight hours after installation.

I have also read that the SAE did a study on this subject and they determined that as much as 20% of the preload was lost due to embedment. Some aircraft engine overhaul shops recommend that all the critical bolts on the engine be re-torqued after the first 25 hours of operations. This is most likely when the embedment has settled into a position where the surfaces are supporting themselves.

This may explain why some aircraft engine manufactures recommend that a complete 100 hour inspection be performed after the first 25 hours of operations after an overhaul or new engine is installed.