Hi ,

I request the forum members to throw some light on the frictional reaction of knuckles in an expansion bellow.
Or please suggest me a reference.
Thankyou.

ask the vendor

Hello PIPING STRESS ANALYST,

It is not clear from reading your posting what specific information you are asking for or what specific topics you would like to discuss. Will you please go to greater length in rephrasing your question?

John Breen Sir,

I am unsure how to represent the frictional reaction in a caesar model.
I read in the Applications Guide that a 30% increase in lateral stiffness is sometimes used to compensate for these effects for lack of a better value.
Is there an alternative?

Many thanks to you.

I have added two additional lateral restraints across the expansion joint to address this break-free load. Use a bilinear restraint across the joint (by using Node and CNode) with a very large K1, a very small K2, and an FY equal to the break-free load. (You might change the "Connect Geometry thru CNodes" to have it plot OK.) If the lateral load is less than FY, the joint will not deflect; if the lateral load is greater than FY, it will deflect (and carry the break-free load).

But is this what really happens? No.

Once the lateral load overcomes this break-free load, the ends move and the lateral load caused by this friction is released. Once the load is released, that friction load "disappears" and the joint works as if there was no friction in the first place.

So, all this fancy modeling is appropriate if the internal loads exceed the friction. And it would only calculate the "instantaneous" maximum load transmitted by the joint. But if the joint doesn't displace, why is it there?

I think a quick check might be needed to assure that the joint will deflect. Run the analysls with no added lateral flexibility across the (now ineffective) joint, or even without the joint. If the lateral load across this element is less than the break-free load, the joint will not move and your design needs review. If the internal load across the joint is greater than the break-free load; ignore that load in the analysis unless you want to get an idea of what that temporary break-free load will do to your system.

I'm aware of a lot of concern for this load in Europe but not so much here in the States. At least one European XJ manufacturer specifies these break-free loads but I don't see them over here.

If you know any XJ manufacturers who have differing opinions, please point them to this thread if they care to comment.

Hi Dave,

Thanks for your input.

I did not realize that we were discussing telescoping "slip joints" until the reference to the Ap Guide. The "break away" load SHOULD be important to everyone as it causes impact loading to be applied to the restraints and anchors. You have certainly provided the definative model for this application, I hope it goes into the next Ap Guide revision (?).

I am sure you remember fielding a similar topic that was posted earlier by Pete Chandler, so I will give the reference to readers in case they would be interested in the background discussion.
http://www.coade.com/cgi-local/ultimatebb.cgi?ubb=get_topic;f=1;t=001354

Regards, John.

John,

I recall that thread. That impact load you mention is something that lies beyond our clear capabilities in CAESAR II. It points out that field experience cannot be replaced by analytical tools.

No, here we're talking about lateral resistance to sliding, not axial. (But point it in the axial direction and you could use that bilinear restraint with a CNode for the slip joint too.)

Hi Dave,

Ok, now (finally) I am with y'all.

Regarding the "breakaway" load (" that lies beyond our clear capabilities in CAESAR II") - what do you think of using your submodel for the slip joint application as you describe it (pointed in the axial direction) and just tell the structural guy to use an appropriate dynamic amplification factor (DLF) as described in B31.1 Nonmandatory Appendix II, paragraph II-3.5.1.3?

Sorry to go "off-topic" with that proposal.

Regards, John.

A little more off-topic...

My comment on the capabilities of analysis was focused on the interaction between several of these events. The first one may pop based on pressure and thermal strain but then the others will react due to those (static) loads and the transient response from the first pop. While CAESAR II could generate numbers for the subsequent events, I wouldn't say that it must happen that way.

I agree that your DLF=2 is a good way to go for a single impact load. (I see a similar reference in ASCE 7-05 in note b to Table 13.6-1. It's seismic but it addresses gaps.)

...if this goes any further, we should move it to a new thread.