What is relaxation and what are the mechanics of it?

Relaxation is the phenomena of clampload loss after the final tightening has been accomplished! There are many causes so we will only look at a few. For now!!

A)   Gasket stiffness (soft gaskets creep and lose more clampload than the stiff ones. A gasket manufacturer once told me that his gaskets had a residual bounceback, that is, the gasket would regain some of its thickness, when you take the load off. I thought, “So what, that does me no good to reclaim the lost clampload.” He just wouldn’t understand creep and crush and its affect on the joint.

B)   Dissimilar metal joints; i.e.: Aluminum/Cast Iron. The aluminum will flow under stress more than cast iron and therefore gives up some of the clampload. Aluminum has a modulus of ~10,000,000 PSI compared to Cast Irons ~26,000,000 PSI so the stiffness ratio is ~1/2.6.

C)   Surfaces with paint and/or corrosion protection---these flow under stress and lose clampload fast. Beware of painted surfaces and gaskets, they can lead to disaster.

D)  Metal surfaces with micro-asperities- super fine irregularities that resemble very fine threads that almost match but not quite. The stress on a given point will often be many times the crush strength of the material and the point will collapse and then there is a cascade of crushing of the points until equilibrium is attained. At this time the clampload can be as much as 15-25% less than the clampload at the final torque. (From Bolt Science: This type of relaxation is often referred to as 'embedding'. The image below shows the 'smooth' surface of an electro zinc plated washer prior to tightening. The clamp force loss from embedding is due to the deformation of these asperities following the completion of the tightening process. The image was produced by a MicroXAM surface mapping microscope.)

E)   Any dampness (such as oil film or dampness caused by bringing cold parts inside) that gathers on the surface of the mating metal parts can cause creep and clampload loss. At tightening, because of the crush strength of the fluid (the basis of all hydraulic force usage) there will be no readily noticeable joint problem, later on because the fluid is under great stress it will seek a path to escape and will between the micro-asperities. If this joint is heated in service it will lose more clampload and will need to be retightened a second time.

Some of the ways to overcome creep and relaxation are to examine the joint for usage and to harden the joint surface if possible and feasible. You can flame harden some surfaces and heat-treat others but this is time consuming and costly. So you may have to consider a tightening strategy that overrides the creep factor.

1)    A gasketed joint can be ‘conditioned’ by tightening it to a given torque and turn of the nut (angle) and then releasing the load and immediately retightening to the same specification. This method retains more clampload and allows less gasket creep to occur. The reason this works is that the gasket after the first tightening is at the height of the clampload attained during the first tightening, when you perform the second tightening you are tightening on a ‘harder’ gasket. The loss will be ~10% as compared to ~25% or more.

There are many ways to make sure the clampload loss is minimal and most of them are in the design stage of the joint. The materials, the usage, the heat range of service, the expected service loads --- all of these and more are very important to make the joint work as designed and to live the expected life. And probably the most important factor is the tightening strategy for the joint. Some joints may require a small fastener and a fairly large clampload with little or no relaxation, consider hard surfaces and tighten a high grade (9.8, 10.9) to the yield point.  This way the joint will get the maximum clampload from a given fastener and give up little clampload from relaxation.

The hardest (not difficult) joint I ever worked on was the ring gear to axle cage. After I tightened the joint and measured the clampload with an ultrasonic bolt tester, I placed the assembly in an oven for four days at ~300 deg. F. After removal from the oven and cooling for several hours, I re-measured the clampload and the loss was less than ~2-3%. It was an extremely hard joint and was well engineered for fastening. Live long and tighten well!!  George