bolted joint calculation of preload decrease as a result of members creep 1

[elinah34]

hello,

I am trying to find out how to evaluate by calculation the anticipated preload decrease as a result of a member creep in my joint.
Here is an illustration on the simplified clamped joint I tried to analyze:

I know that my ULTEM 1000 member is likely to creep and lead to a preload decrease in the bolted joint.
I am attaching a data of SABIC ULTEM 1000 creep in tension (although my case is a compression).
My initial preload is one that leads to a compressive stress of about 5946.5 psi in the ULTEM 1000 member.
I marked the relevant yellow graph which gives data regarding the creep (in tension, unfortunately not compression - but it might be good enough as a limit):

How can I calculate the preload decrease from that point?
I tried to calculate it by myself, but I got something that doesn't seem reasonable - losing the preload entirely.

I am going to detail what were my steps in the preload decrease calculation:
1. From the yellow graph I estimated a maximum of 25% creep (((0.015-0.012)/0.012)*100%).
2. I calculated the additional contract of the ULTEM 1000 member that is the result of creep = 0.25*(member initial strain)*(member thickness).
3. Assuming we maintain contact between the bolt head and nut and the members (and hence the clamping) - the bolt's contract will be equal to it.
This contract will be subtracted from the bolt's initial tension - and this what causes to preload decrease.
4. if we multiply the contract (as a result of creep) by the bolt stiffness, Kb, we get the preload decrease.

As I mentioned above, my calculation doesn't sound reasonable - losing all the preload.
In addition, after looking for a reference in which I can find out how to use the creep graph and calculate the preload decrease, I couldn't find any similar case and such a calculation.

your help will be blessed!

 

[desertfox]

Hi

This doesn’t help with the calculation as such but it talks about two stages of tightening to minimise creep in gasket materials

https://www.boltscience.com/pages/glorimermoreonrelaxationarticle.htm

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[desertfox]

Hi

Another site that might help but again it talks about using adhesive to combat relaxation and it also mentions allowing 10% on preload for relaxation

https://mechanicalc.com/reference/bolted-joint-analysis

“Do not worry about your problems with mathematics, I assure you mine are far greater.” Albert Einstein

 

[3DDave]

The spring rate of steel is roughly 30 psi per part per million (PPM) of strain. This relaxation is 15000 parts per million at 4000 psi.

To even this up you need to exceed an initial load of 30 psi/PPM * 15000 parts per million of strain or a tensile stress on the bolt of 450,000 psi, while maintaining that 4000 psi load in the Ultem 1000.

 

[MintJulep]

This is such a simple configuration.

Why don't you just do a simple test?

Get a sheet, drill some holes, put some bolts in. See what happens.

With such a short grip length losing all preload is not unreasonable.

 

[SnTMan]

You should add a flat washer.

Regards

Mike

The problem with sloppy work is that the supply FAR EXCEEDS the demand

 

[Stress_Eng]

Not knowing the purpose of the bolt joint or how flexible the design can be made, but will the design permit the use of washers and the ULTEM to be sleeved, thus permitting the preload to go through all steel parts?

 

[elinah34]

Thank you all for replying.
After reading all the comments I guess there is no simple way of predicting the preload decrease.
I anyway planned to have some experiments, but wanted to try to predict it by calculation.
Regarding the design comments-
Washer is indeed included in the real part. As mentioned above, the simplified model I posted isn't complete.
Unfortunately compression limiter isn't possible.

 

[dvd]

You might consider adding a belleville conical washer over a flat washer - using-belleville-springs-to-maintain-bolt-preload